PFIN Technical Note 85-12
A BENEFIT-COST ANIALYSIS OF NUTRITIONAL INTERVETIONS
FOR ANEMIA REDUCTION
by
Henry M. Levin
Stanford University
July 1985
Population, Health and Nutrition Department
World Bank
The World Bank does not accept responsibility for the views expressed herein
which are those of the author(s) and should not be attributed to the World
Bank or to its affiliated org7anizations. The findings, interpretations, and
conclusions are the results of research supported by the Bank; they do not
necessarily represent official policy of the Bank.    The designations employed,
Lhe presentation of material, and any maps used in this document are solely
for the convenience of the reader and do not imply the expression of any
opinion whatsoever on the part of the World Bank or its affiliates concerning
the legal status of any country, territory, city, area, or of its authorities,
or concerning the delimitations of its boundaries, or national affiliation.



TABLE OF CONTENTS
PAGE
ABSTRACT . .   .  .  .   . . . . . . . . . . . ..
SUMMARY AND CONCLUSIONS ..........................ii
I           INTRODUCTION .......................   ...... 1
II          PREVALENCE AND TREATMENT OF ANEMIA .         4
III         BENEFITS OF ANEMIA REDUCTION . .   .......... 17
IV          COSTS OF INTERVENTIONS . .    . . . . . .  .  .  38
V           CALCULATING COSTS ....................... ..... 43
VI          CALCULATING BENEFITS .   .... . . . . . 59
VII         CALCULATING BENEFIT-COST RATIOS. . .......   70
BIBLIOGRAPHY . . . .   .   . . . . . . . . . .  86



LIST OF TABLES
TAB LE                                                      PAGE
ONE       POPULATIONS AT RISK: ESTIMATED PERCENTAGE           6
WITH HEMOGLOBIN CONCENTRATION BELOW THE NORM
FOR NON-ANEMIC SUBJECTS
rWO       LRON INTERVENTLONS AND CHAN&GES IN HE4OGLOBIN      18
LEVELS
T'HREE    HEMOGLOBIN LEVELS AND MEASURES OF WORK OUTPUT      27
FOUR      ESTIMATED ANNUAL COST FOR DELIVERING 'MEDICAL      55
SUPPLEIMELNTS Tro REDUCE ANEMIA (BASED ON SERVICE.
FOR I ,000 PERSONS)
FIVE      ESTIMATED IMPACT OF IRON INTERVENTLONS ON          63
WORK OUTPUT
S t.K     ESTIMATED BENEFITS PER CAPITA OF ANEMIA            72
INTERVENTIONS IN INDONES IA
SEVENI    ESTIMATED BENEFITS PER CAPITA OF ANEMIA            73
INTERVENTIONS IN KENYA
EIGHT     ESTIMATED BENEFITS PER CAPITA OF ANEMIA            74
INTERVENTIONS IN MEXICO
NIINE     ESTIMATED COSTS OF ANEMIA INTERVENTIONS            75
PER CAPITA
TELNI     PER CAPITA BENEFITS AND COSTS OF FORTIFICATION     79
ELEVEN    PER CAP ITA BENEFITS AND COSTS OF SUPPLEMENTATION  80



PHN Technical Note 85-12
A BE1NEFITt-COST ANALYSIS OF NUTRITIONAL INTERVENTIONS
FOR ANEMIA REDUCTION
A B S T R A C T
Iron-deficiency anemia is one of the most prevalent nutritional
disorders in both the industrialized and less-developed countries. The
purpose of this paper is to evaluate potential interventions for reducing
anemia with a benefit-cost analysis. The study begins with a discussion of
the origins and prevalence of anemia and proceeds to some of thie
consequences of anemia on work capacity, work output, learning, and other
outcomes. Specific estimates are made of the effects of reducing anemia on
work output.
Both medicinal supplementation and fortification of food with
iron are considered as interventions. Each intervention is evaluated for
the costs of specific strategies in less-developed societies, with
particular emphasis on cost estimates for Indonesia, Kenya, and Mexico.
Estimates of benefits are calculated for the value of additional work
output in labor-surplus societies as well as assessments of other
benefits. The results of the cost and benefit analysis are used to
construct benefit-cost ratios for evaluating the investment potential of
anemia interventions.
Under a wide range of assumptions, the benefit-cost ratios are
found to be substantially greater than one.    This is especially true of
dietary fortification where the benefit-cost ratios were found to be
between 7 and 70 for the three illustrative countries. But even dietary
supplementation was found to show a range of benefits to cost of from 4 to
38 for the most reasonable set of assumptions. The study concludes that
benefit-cost ratios of nutritional interventions to reduce anemia appear to
be large, and field trials should be carried out in specific settings to
see if the overall findings of this study are supported in particular
cases,
* * ** *** * * *** *
Prepared by: Henry M. Levin, Consultant to the World Bank
Stanford University
July 1985



SUMMARY AND CONCLUSIONS
Although anemia is a serious nutritional problem around the
world, it is especially severe in the tropics where regional studies
have commonly found more than half of the population to be anemic.
Populations that are especially at risk are infants, children, and
women and especially pregnant or lactating women. Nutritional.
interventions focus on increasing iron intake (and folate where
indicated) as well as increasing the dietary intake of foods or agents
that vwill increase iron absorption.      Especially important in
increasing iron absorption from existing sources is the addition of
ascorbic acid to the diet and foods that are rich in heme sources of
iron such as meat, poultry, and fish.
Such interventions can take the form of medicinal supplementation
through orally taken or injected iron compounds or dietary
fortification through the addition of the appropriate compounds to
food vehicles found in the normal diet. The latter is generally to be
preferred because it has a lower distribution cost and does not
require    changes in behavior in order to improve Hb.          Such
interventions have been shown to raise Hb levels dramatically,
especially among those with severe anemia.
Thera are- many studies on the. effects of H:b on humar. behavior, but the
evidence seems to be most substantial in the. area of work capacity and
work output. Work capacity refers to various physiological indicators
of the capability for doing work such as maximum oxygen uptake and the
heart rate and production of lactates for any particular level of work
- 11 -



effort. Experimental studies have shown that when Hb of anemic sub-
jects is raised through nutritional interventions, the maximum oxygen
uptake rises; and heart rate and lactates associated with a given work
effort decline. Quantitative estimates are provided in the report.
Summaries of a large number of studies that relate Hb and changes
in Hb to work output show similar conclusions.    Among a wide range of
national settings and measures of work, Lt appears that a 1 percent
rise in Hb is associated with between a I and 2 percent increase in
work output.     This finding is remarkably robust among different
investigations. Other studies show the relation of Hb to intellectual
growth and school achievement, morbidity, infection, and so on.
Given an understanding of the nature of interventions and the
benefits that might result from them, the report attempts to estimate
the monetary values of the costs and benefits. Costs are estimated for
medicinal supplementation and for dietary fortification.        In both
cases, the costs are estimated by first stipulating the resources that
would be required for the interventions such as personnel, facilities,
transportation,    supplies, and iron compounds or ascorbic acid.   The
costs of these resources are estimated under different assumptions
regarding the markets for them in different settings.      In addition,
the cost estimates take account of additional caloric requirements of
workers with higher work output as a result of anemia interventions.
Benefits are estimated by assessing the value of additional work
output for agricultural workers in societies characterized by labor
surpluses.    They are predicated on the assumption that the additional
output of a worker is equal to about half of the average wage of that
- iii -



worker in such a context.      Benefits are also estimated for the
non-labor market effects of reductions in anemia. The costs and
benefits are calculated for agricultural workers in three countries:
Indonesia, Kenya, and Mexico.
The estimated monetary value of benefits exceeds costs by
substantial margins for all three countries for both supplementation
and fortification.    Under the most "reasonable" sets of assumptions,
the benefit-cost ratios for fortification are estimated to be about 7
for Indonesia, 43 for Kenya, and 71 for Mexico.  For supplementation,
the comparable benefit-cost ratios are 6, 34, and 56.      These high
benefits relative to costs seem to hold under a wide variety of
different assumptions regarding the calculations.
The overall conclusion is that nutrition        interventions for
reducing anemia appear to represent social investments that are highly
productive and that ought to be considered seriously by the Bank.  In
addition,     it would be useful to carry out field trials of
interventions in which costs and benefits were estimated directly in
order to see if the global results found in this study are supported
by specific interventions at particular sites.
- iv -



A BENEFIT-COST ANALYSIS OF NUTRITIONAL INTERVENTIONS
FOR. ANEMIA REDUCTION
L. INTRODUCTION
One   f t.he imost prevalent riutritional disorders .n both industrialized
and less-deve IDoped countries (LDC' s) is iron-deficiency anemia (Baker and
DeMaeyer L979: 388-92; Charlton and Bothwell 1982; Fleming 1982; Masawe
1981). Anemia refers to a condition in which the hemoglobin concentration
in the blood   is considered to be be low some normal value for a given
populat ion. Although anemia may be caused by other factors such as disease
or blood loss,   the most common cause is a deficiency of iron (Charlton and
BOathwell 1982 : 310-16). Such iron deficiency is typicaLLy a result of an
inadequat.e intake of absorbable iron, relative to the needs of the body for
forming hemoglobin and meeting other iron needs.
The hemoglobin level is particularly important, since it provides the
oxygen-transport mechanism for the body. At low hemoglobin levels, the
b lood is restricted in its capacity to carry oxygen to the celLs, limiting
the ability of the body to produce energy and meet other functional needs.
From a health-related perspective, the anemic person feels weak, listless,
and may be more susceptible to infection. Work capacity is also impaired,
and anemic chiLdren perform less welL in school. Many of these behavioral
outcomes of anemia have been summarized in reviews on the subject (e.g.
PoLlitt, V'iteri, Saco-Pollitt, and Leibel 1982; Pollitt and Leibel 1976;
Read 1975; Scrimshaw 1984).
The purpose of£ this study is to evaluate potential interventions for
reducing iron-deficiency anemia in LDC' s from the perspective or an



2
inves tment in human resources as evaluated by a cost-benefit analysis
(Sorkin 1976: 33-9). All investments have a cost which can be defined as
the value of the resources ucilized for the intervention (Levin 1983). The
benefits are equivalent to the value of the outcomes that are produced by
the intervention.    In the case ot interventions to reduce iron-deficiency
aneMia, the costs derive from the dietary iron supplements or fortification
and the system for delivering them to insure that they are consumed by che
appropriate populations. Potential benefits are associated with the
improved feeling of well-being of the populations, improved fetal and child
growth, Lower morbidity and mortality, higher productivity both inside and
outside of the workplace, more enjoyment of lei~sure, and more effective
learning among students.
The case for using a cost-benefit analysis for evaluating programs for
alLeviating iron-deficiency anemia is straightforward.           LDC's are
characterized by a large number cf challenges such as unemployment, health
probleias, nutritional deficiencies, poor education, inadequate housing and
transportation.    The potential responses to these problems are also many.
Investments in health, nutrition, education, housing, transportation, water
resources, and agricultural and industrial development all represent
potential paths for improving the welfare of the population. But, it is
likely that some investments will be relatively more productive in their
impacts than o thers for any aiven resource out lay. The purpose of
cost-benefit analysis is to ascertain if the benefits of a particular
s trategy exceed its costs and by ho; much.         In this way one can
compare the cost-benefit status of one alternacive with others and can



3
choose those alternatives which are likely to maximize the benefits to the
society relative to their costs.    In its ideal form, cost-benefit analysis
provides a guideline for ctioosing investment priorities when resources are
Limnited relative to the needs that they must address (Mishan 1976).
Ideally, this exploration would proceed from a case studv or series of
such studies in which interventions were undertaken under different sets of
representative conditions.   Precise measures would be available for all of
the factors that were pertinent, and one need only place them into a
bene fit-cost framework to provide the necessary result.s.  Unfortunately, no
such case studies exisc that provide systematic data for alL of the
pertinent relations. However, a large variety of studies exist that can be
used to "construct" a picture of the magnitudes of benefits and costs of
anemia interventions.     The purpose of this study is to use available data
to establish the overall parameters of benefits and costs for strategies to
reduce nutritional anemia.       A major emnphasis will be on building a
methodology tthat can be applied to new data as they arise or to specific
field trials.     Lii constructing this framework, a variety of assumptions
wilL be used to establish linkages where more precise data are lacking.
The attempt will be to make the methodology and assumptions transparent so
that other assumptions can be imposed in order to see if they would modify
the conc lu s ions.  This study should not be viewed as a substitute for one
based upon f ield trials in specific settings, but only as an overall
benefit-cost guide to anemia interventions.
The remainder o f this report will be organized in the following way.
Section II will discuss the prevalence and treatment of anemia. Section III
wiLl develop the potential benefits of anemia reduction, and Section IV



4
wi L L present the costs of interventions to reduce iron-deficiency anemia.
Seccions V and VI will develop specific calculations for coscs and benefits
raspectively.    The final section will integrate these results into a
benefit-cosr frainework.
LI . PREVALENCE AND TREATMENT OF ANEMIA
In this section we will present information on the prevalence of
anemia and its treatment.  Diagnosis of anemia is usualLy made on the basis
of an evaluation of the hemoglobin content of the b.lood. Hemoglobin is a
substance of iron (heme) and protein (globin) found in the red corpuscles
Of ciie blood that carries oxygea fromn the Lungs to the tissues and some of
che carbon dioxide froa the tissues to the lungs. Each molecule of
hemoglobi,n can carry four molecules of oxygen. In a non-anemic person,
elach liter of blood contains between 110 and 160 grams of hemoglobin.
Reinotglobin accounts for about two-thirds of the 3..5 grams of iron in the
healthy adulc male.     The standard test for anemia is to assess the
concentration of hemoglobin, which is usually evaluared in grams per
deciliter of blood (g/dl).      An alternative measure is hematocrit which
requires only a minute amount of blood and can be done in a small clinic or
office.   The hematocrit is about equal to the hemoglobin (Hb) concentration
multCiplied by 3, however it is a less reliable means of diagnosing anemia
(DalLman 1982: 67).
Although groups like the World Health Organizacion (WHO) have
established general criteria for determining if a person is anemic, the
an o ra"    l a I  e v   of hemoglobin, at sea-le vel,  w i 1 L d i F f e r f r o m



5
person-to-person and population-to-population. For eKL=-c&, WHO (1968)
defined the following levels of hemoglobin concentration in g/dl below
whiich anemia is likely to be present: children 6 months to 6 year.s, 11;
children 6 to 14 years, 12; adult males, 13; adult females, non-pregnant
L2; and adult-    emales, pregnant,ll.  However, these are conv.idered to be
zeneral indicators of anemia rather than precise criteria.   Likewise, iron
requirelnents also differ among individuals and groups. Daily requirements
o f iron that mus t be ab sorbed to maintain homeostasis are estimated to
range from .7 mg for infarnts and .9 mg 'or men to about 3 mg for women in
the second half of pregnancy (Baker and DeMaeyer 1979:375).
One mne thod of ascertaining if an individual is anemic is to establish
an initial hemoglobin or hematocrit level followed by dietary supplement
with irorn and a subsequent measurement of hemnogliAin or hematocrit. Those
persons whose hemoglobin or hematocrit levels rise as a response to
supplementation are considered to have been anemic. The larger the
response, the more serious the anemia. For any population it is possible
to relate different initial Levels of hemoglobin concentrations with
response rates to determine a "cutoff" value for Hb that would predict that
persons be low that level were anemic (Leibel, Pollitt, Kim and Viteri
1982). Cook and Finch (1979) have shown that alternative laboratory
measurements of iron status may be more useful than Rb concentrations among
some populations, particular those with mild iron deficiencies, and that in
field t.rials it may be desirable to use multiple measures.
Anemia is a serious nutritional problem around the world, bur
especially in the tropics (Fleming 1977 & 1982; Slasawe 198'1; Woodruff



6
TABLE ONE
POPULATIONS AT RISK: ESTIMATED PERCENTAGE WITH HEIOGLOBIN CONCENTRATION
BELOW TiE NORM FOR NON-ANEAMIC SUBJECTS
Countrv          Date               A&e             Sex             Urban/Rural      % Anemic
3angLadesh       19762              adult           F-pregnant      Urban            66a
3urma            19761              adulc           F-pregnant      Urban            82a
Surina           19722              adult           F-pregnant      Urban            ,-4ia
Burma            19724              pre-school     M/F                               3-27c
Fiji (Indian)    19704              adult                                            80C
North India      19681              adult           F-pregnant      Rural            80a
A;orth India     19734              children        M/F                              9Qb
adult          M/ F                              48g/ 84c
South India      1968/19731         adult           F-pregnant      Mainly Urban     57.4a
Souch India      19752              children       M/F              aural            76c
adult          M2               Rural            56c
adult          F                Rural            SIc
India            19752              adult           F-pregnant      Urban/Rural      88a
Indonesia        19802              adult          F-Pregnant       Rural            37a
(East Java)                                      F                                30c
:ndonesia (West 19732               adult           F-pregant       Rural            65a
& Central Java, Bali)
Jatnica          L9794               pre-school     M/F                              76c
Kenya            19573               adult          M/F             Rural            32.3e
Latin America    19711              adult           F-pregnant      Mixed            26.5a
Latin America    19714              adult          M1 Mixed                          4c
.Maur i t ius    19603              pre-school      / F                             5 Od
Malaysia         19642               adult          F-pregnant      Urban            75a
:'e:ico          19681              adult           F-pregnant      Rural            26.6a
.;eoai           19772               adult          F-pregnant      Urban            35a
Pakistan         19702              adult           F-pregnant      Urban            73a



7
T     BLE L:  (continued)
C5uncrv          Date                Age            Sex              Urban/Rural      % Anemic
* .ilippines      19712              adult          F-pregnant       Urban            63a
: i1ippines      29762              adult           F-pregnant      Urban            -2a
;I'?ppines     19764               pre-school     M/F                               4c
adult         M'                                 7c
adult          F                                 37:
-o Land          19681               adult          F-pregnant       Urban            21.3a
z6uch 'iZrica     19763              adulc          14                                44. 3c
"'a tal)
adult          F                                 33.lc
r,apre          19722              adult           F-pregnant      Uirban Indian    205
Urban Ma lay      215
'Jrban Chinese Gb
iri Lanka        19572               adult          F-pregnant       Rural            5Oa
3,.i Lanka       29744               adult          F                                  Oc
adult          14g
.-rzania        19733               adult          M4/F             RU raL           37.3f
.ai Land        19802               adult          F-pregnant       Urban            BOa
I.Ban;kok)
_ aiLand (Ubal)  l9712               adult          F-pregnant       Rural            48a
.;iailand        19794               pre-school     14/F                              45c
adult          M                                 35c
adult           F                                45c
Thailand         19802               pre-school     M/F                               L5c
children       M/F                               33C
15-49          M1                                1Sc
15-49           F                                18c
over 49        14                                34c
over 49        F                                 Si c
:ricerion:  a) Hb<ll    b) Hb<lO.5   c) WHO criceria--pre school1l1, schooL children<12,
adult mnales<13, adult Eemales<12, pregnant females<11    d)Hb<10.3   e) Hb<8
f) Hb<10   g) HbK12
'eference:   L) Baker & DeMaeyer, 1979   2) Baker, 1981    3) Masawe, 1981   4) Fleming, 1982



1972 & 1982).     Table One shows populations at risk based upon a number of
studies around the world. An attempt has been made to indicate the age,
sex and pregnaacy s tatus, and urban-rural origin of the populations.    In
some of the studies, 80-90 percent of the population were found to be
anemic. c  As expected, pregnant females seem especially at risk, but for
adult males and females and children there are studies represented in the
Tab le showing 80 percent or higher rates of anemia. There seems to be no
particular pattern between urban and rural areas, with high incidences of
anemia found in both.
A major recent survey of nutritional anemia among women in developing
countries has been done by Royston (1982).          Royston's review of the
literature and summary of the statistics is unusually comprehensive with
attempts to incorporate virtually all available data on-hemoglobin
concentration for women in developing countries according to pregant women,
lactating women, non-pregnant women, and all women. His data suggest that
it is common to find at least half the women below the norm on Hb for any
particular classification, and in some cases the entire population is
considered to be anemic.
Causes of Iron Deficiency
The treatment of anemia obvious depends upon its etiology.     Fleming
(1977) has delineated four factors that determine the iron status of the
body: (1) iron intake; (2) absorption of iron; (3) physiological demands
including growth, menstruation, and pregnancy; and (4) pathological loss
through hemorrhage.     With respect to iron intake, there are many organic
and inorganic sources.      For example, many typical foods have high iron
contenc (Bogert, Briggs, and Caltoway 1973: 269). Although a study in



Mauritius found an iron intake of only 5 ig a day and in India only 15-30
mg a day, a study in Ethiopia showed an intake of iron of as much as 180 mg
a day (Fleming 1977: 316). Not only does the intake of iron vary widely
among countries and regions and populations within countries, but the
sources and forms of iron intake also vary considerablY.
The absorption of iron fromn food depends on the source of the iron as
well as other factors. Nutritionists distinguish between heme and non-heme
sources of iron. Heme sources include meat, fish, and poultry. Iron is more
readily absorbed from heme sources, than from the non-heme cereals and
legumes that comprise the staple diec for most of the world (INACG 1982).
Layrisse et al.(1969) and Layrisse and Martinez-Torres (1971) found
that the mean absorption of iron in three major grains that often dominate
the diet in developing societies (wheat, rice, and maize) was within the
range of 1-7 percent, while for fish and meat the range was 12-20 percent.
Even if heme formns of iron represent a small part of iron intake, their
high levels of bio-availability mean that they are often the dominant
source of usable iron.
Heme sources of iron (such as meat) are also important because they
enhance iron absorption from non-heme sources. Thus, even small amounts of
meat when added to cereal and legume diets can have a significant role in
bio-availabiLity of iron. Ascorbic acid is another significant facror Ln
enhancing iron absorption along with other organic acids (Hallberg 1981).
.Many comTpounds have been found to inhibit absorpcion including carbonates,
oxalates, phosphates, and phytates. Consumption of egg yolk and Indian tea
are also important obstacles to absorption. Beyond these there will be
different absorption rates among individuals because of diSferences in body



10
chemis try as we lL as di f ferences from clemical changes in food due to
processing, scorage, cooking, and interaction with ot.he- foods in a meal.
Finally, absorption is highest when body stores of iron are most deficient
and Lower when storage iron is repLenished or adequate.
Physiological factors affecting iron deficiency are those associated
with the iron requirernents of the body. There are special growth needs in
infancy and childhood (Burman 1982) and adolescence (Lanzkowsky 1982).
Pre-menopausal women need more iron than men because of the iron lobs
during menstruation, and pregnant women also have higher iron needs to meet
the requirements of the fetus and placenta as do lactating women (Bothwell
and Charlton 1981).
Pathological factors are those which cause bleeding with a resultant
increase in iron requirements.       The most common pathological factor in
developing countries is that of hookworm infestation which has been
estimated to affect almost half a billion people around the world (Chariton
and Bothwell 1982:313). The number of worms and type will determine the
blood loss and iron needs. Other types of parasites may also contribute to
blood los.s as well as any form of internal bleeding such as duodenal
ulcers.
Dietary Interventions
In situations where there is a high prevalence of hookworm and other
oarasites that cause gastrointestinal bleeding, the aLleviation of anemia
will be problematic through nutritional interventions alone. Those
situations may reequire investment to reduce parasite infestations such as
ant i-he l;nintn drugs and sanitary water supplies as well as
hygiene-education     for the affected populations.  However, even in these



cases, d ietary interventi ons may be in order. The purpose of dietary
in tervent ions is to increase the amount of iron intake and especially that
of bio-avai lable iron to replenis t andl maintain adequate hemoglobin
co ncentrations.
D)ietary   LaterV/e ations ca ke two 'orms, supplemerntation and
for t i f icat ion (Baker and De Maeyer 1979: 393).  Supplementation refers to
the provision of an extra amount of nutrient in medicinal form. In this
case, iron and other substances that enhance iron absorption can be given
orally or by injection.    Fortification refers to the addition of nutrients
to foods to maintain or improve the quality of diet.
Callender (1982), INACG (1977), and Bothwell and Charlton (1981) among
others have provided      comprehensive reviews of the various forms of iron
used for supplementation.       Callender (1982) stresses that among the
bewildering number of iron preparations available, the preferred choice is
one that is highly absorbed and tolerated as well as having a low cost.
With respect to absorption, reduced forms of iron are best.     Also, it is
generally accepted that the commonly used iron compounds (sulphate,
Lactate, fumarate, gluconate, glycine sulphate, and glutamate) with the
same iron content given under standard conditions liave about the same
absorption rates (Brise and Hallberg 1962) . Although "slow-release"
oreparations are designed to improve tolerance to iron, Callender (1982:
330-331) maintains that they have never been shown to have any signiFicant
advantage over simaple iron preparations. All Forms of iron are astringent
and ingestion in large doses is frequently accompanied by nausea, abdominal
pain, diarrhea, and constipation.



t2
The level of supplemetitation mus t depend upon the degree of iron
depletion, iron needs, bio-availahIe        iron from other sources, and
absorption races of supplemental intake.      These factors wilL vary among
populacions, and clinicaL trials are usually recommended to determine
op°Cimal interventiions (INACG 1977: 8-12; W4H0 1975).        A standard
supplementary intervencion for iron deficiency is 120-200 mg a day of
ferrous sulphate or ferrous fumarate per adult with a smalLer dosage for
in fants and children (Callender 1982: 332-334). Ferrous sulphate is one of
the cheapest of all nutritional supplements, with costs estimated at only
68 cents a kilo    in 1977 with other iron compounds ranging from two to
seventeen times as costly as ferrous sulphate after adjusting for
bio-availability (INACG 1977: 15; Bothwell and Charlton 1981: 44).   It is
imnportant to note that while costs are still relatively low,. they have
rtisen by about 30 percent since 1977. Adding ascorbic acid to increase iron
absorption fromn meals would increase costs, with both stabilized and
unstabilized ascorbic acid estimated to cost between $10 and $11 per
kilogram in 1982 ( INACG 1982: 31). However, the additional cost may be
jus tified.   According to Bothwell and Charlton (1981: 52) the addition of
200 mg or more of ascorbic acid increases iron absorption by at least 30
percent.
In some cases, folace deficiency is also a cause of anemia, often in
conjunction wich iron deficiency. This is especialLy true for pregnant
women n   4WHO has recommended that pregnant women with folate deficiencies
should receive 2.5 to 4.0 milligrams it- folic acid as a supplement (INACG
1977:21-22).    The cost to UNIICEF of 1000 tablets containing 60 mg of
elemental iron in the the form of iron sulphate and 0.5 mg of folate was



13
about $l.00 in 1981 of which the Lolate added onlv about 2 percent to total
2ost (DeMaeyer 1981: 366).
Dietary supplementation with iron is a short-run strategy designed to
replete iron scores in a population.   Its major weakness is the problem of
gettzing :otnpliance among the population to aoLLow a regimen in which
lnedicinal iron is taken on the recommended schedule and for the full period
of supplementation.      In contrast, dietary fortification is designed as a
long-term strategy for maintaining ironi status while requiring no special
behavior on the part of its recipients.
The advantage of fortification of the existing diet is that the
logi3tics and cost-of delivery mechanisms for therapeutic suipplementation
are avoided while the population obtains iron through its daily diet. A
major challenge is the choice of a food vehicle for fortification. WHO has
set out the following criteria for choosing such a vehicle:
1- It is consumed by the vast majority of the target population in
adequate amounts.
2- It is available for fortification on a large scale and at
reLatively few centers, so the fortification process can be adequately
supervis ed.
3- The resultant product is stable under the extreme conditions likely
to be encountered in storage and distribution.
4- The palatability of the vehicle, or other foods that the veihicle
may be mixed with (e.g., in cooking) is unchanged (Baker and DeMaeyer 1979:
393-394; WHO 1975: 25-29).
In adddition, WHO has suggested that iron compounds used in
ror.ification must be readily absorbed when inixed with the vehicle and when



14
t'he vehic le is added to the diet, must not cause changes in the color or
taste off food, and must be stable under conditions in whi-h it is used or
s cQrad ( Baker and DeMaeyer 1979:394) . Fortification strategies must begin
with an anaLysis o f the diet of the population and an evaluation of the
approoriate vehicle and iron compound that are consistent with that diet.
The ideal vehicles for fortificacion may vary from country-to-country
and perhaps, eveni region-to-region within country. In industrialized
countries like the U.S. and Sweden, the fortification of wheat flour is
common, improving the iron content of baked produnts and cereals. In the
U.S. such en.richment has tak-eni pLace since 1941 with iron levels in
enriched wheat varying from about 29-36 mg/kg (INACG 1982: 14). However,
in less-developed countries wheat is less likely to be centrally-milled or
a dietary stapLe. The more common vehicles in the IDC's are salt, sugar,
infant foods, condiments, and skimmed milk (1WHO 1975: 27-28).
Because salt is consumed by all populations and is processed in
relatively few centers, it has two important characteristics of a good
vehicle.   Perhaps its major drawback is that powdered iron and iron
compounds tend to discolor salt, a process that is accelerated under high
humidity.    Also, the amount of salt intake may have to be high to provide
adequate iron availability at a typical fortification level of I mg of iron
per gram of salt (Food and Nutrition Board and UNICEF 1981). The addition
of ascorbic acid or sodium hydrogen sulfate seems to reinforce absorbancy,
perhaps o ffsetting this concern (WHO 1975: 27). Fortification of sugar is
also attractive because it is widely consumed and often refined in just a
few centers while not being discolored by certain iron compounds. However,
fortified sugar does discolor tea, and very little refined sugar is
consumed in many rlDC's.



15
Since cereals and legumes represent the dietary staples in LDC's, the
possibiLity of fortifying those must be considered (INACG 1982). Rice,
wheat, and maize and other cereal 3rains are the most important source of
energy for the world' s populations (INACG 1982: 12-22). The problem with
ce rea Ls is that their iron concent is characterized by low bio-availability
and absorption in the absence oL heme forms of iron in the diet.
Enrichment of such grains is possible, but there are a number of practical
shortcomings.     Firs t, much of the production of these grains is done for
,home consumption or local markets, making         central processing and
d,istribution a difficult challenge. Second, although rice is the major
staple for more than FialE of the worLd's population, it has noc yet proven
to be a satisfactory vehicle for fortification.  Rice can be coated with an
Lron compound, but such a coating will be washed away during the rinsing
and cooking process. Recent progress in fortifying rice may itmprove its
feasibi Lity as a future vehicle (Cook and Reusser 1983).  Iron seems to be
better absorbed in meals where wheat and rice are the main source of energy
than in maize meals, although the form of the cereal will affect these
results. For example, rolls made from maize starch do considerably better
in iron bioavailability than maize porridge (INACG 1982: 19).
In general, it is agreed that the addition of small amounts of heme
sources to the diet such as meat or fish (wilich unfortunately are either
costly or unavailable in many instances) as well as foods with ascorbic
acid may be more effective than iron fortification of cereals. The
addition of iron to common spices and condimnents is also a possibility,
Typical cereal-based diets are bland, requiring the addition of sauces and
ocher condiments to give them character. Many countries and regions of



16
countries use specific sauces and condiments to flavor the diet. For
example, in Thailand a fish sauce was used as a vehicle (Garby and Areekul
1974).   The major drawback is that such sauces are often prepared in the
household rather than in a cencral processing plant for market
distribut ion.
Fortirication of infant foods with various forms of iron and/or
ascorbic acid is common. Specific vehicles include cereaLs, milk powders,
and in f an t formulas.  WHO recommends adding both iron salts and ascorbic
acLu Lo Lnfant formu'Las in a ratio of iron:ascorbic acid of at least 1:l'
(WHO 1975: 28). SkimLmed nilk fed to preschool children has also been used
as a fortified vehicle.
Fortification with ascorbic acid is an important means for promoting
iron absorption. Studies have found that salt fortified with ascorbic acid
had the effect of increasing the absorption of the intrt 'sic iron in maize
porridge or rice by 2-4 times (WHO 1975: 28). Indeed, even the addition of
small amounts of ascorbic acid such as 25 mg to a simple maize meal tripled
the absorption or iron, and 200 mg increased iron absorption six-fold
(HalLberg 1981: 53-54). Dorman et al. (1977) found that the addition of
50 mg of ascorbic acid through fortified sugar increased the absorption of
iron nine-foLd from maize-meal porridge. Finally, fortification of foods
with folate should also be considered when its need has been established
(WHO 1975: 29-30).
Table Twqo shows the results of a number of studies of iron
supplementation and fortification. There are far more sr-tldies on
supplementation than on fortification for a number of reasons.
Supplementation can be done more easily in small trials than fordification,



17
and the measurement of iron intake is more easily observed.     In general,
the following patterns seem to hold. When initial hemoglobin levels are
Low, iron supplementation is associated with very substantial rises in
hemoglobin, even on the order of 60-100 percent. Relatively short periods
or supplementation, for example 2-3 months, produce strong effects.
However, it is likely that some dietary fortification would have to be
present to sustain these gains.
The last two entries in the table show attempts ac fortification. The
Indian study used salt as the vehicle, while the study for ilauritius used
wwheat f Lour baked into bread.   In the Indian study, results are available
ror three sites and several population groups. Only the results for 25-44
year olds are shown here, but they are also fairly representative of the
o ther groups.   Again, the general pattern that holds among the three sites
is that the largest rises in Hb associated with fortification are found
among populations with the most severe anemia.    For exaimple, the Calcutta
females with mean initial Hb of 8.5 showed a 35 percent increase to 11.5,
while the males in the Calcutta sample indicated an increase of similar
magnitude from an Hb of 9.7 to 12.8 at the end of 12 moinths. Changes were
considerabLy smalLe- at the two othier sites where mean Rb levels were
suf ficiently h igh that the incidence of anemia--and especially severe
anemia--was smaller.
III-- 3ENEFITS OF ANEMIA REDUCTION
The purpose of this section is co provide a summary of the benefits of
anemia reduc c ion.   Such benefits can include a reduction in both maternal
and infant mortality, in stunted growth and dievelopment, and in infection,



18
TABLE TWO
IRON INTERVENTIONS AND CHANGES IN HDIOGLOBLN LEVELS
STUD.Y                  FORM OF IRON            AMOUNT        TIME PERIOD      Hb CHANGES
Sri Lanka               Ferrous-Sulphate tab.    200mg/day    2months          10.8-12.8= 2
mnales/Females age 20-60                                                       (19%)
igerton, ec al, 1979
Sri Lanka               Imferon I.V.             30-50/ml     I week           Av. 5 aroups
-iiales/ fema les age 39-54                                                    Range=Before
Ohira, eC al, 1981                                                                    6.4-14.1
After
7.6-14
= .66 (9%)
Norway                  Bivalent iron tab.       60mg/3Xday   9 months         11.8-13.4  1.6
,,a l-s/ fema leS                                                              (14%)
adolescents                                                                    (females only)
Vellar & Hermanson, 1971
lr:eLand                Ferrous-carbonate tab.  150mg/day    2 months          10.5-12.5= 2
rernales age 20+                                                               (19%)
Elwood & Hughes, 1970
:alonesia               elemental iron          100mg/day    2 months          12-13.5= 1.5
adult maLes                                                                    (13%)
Basca & ChurchiLL, 1974
Inidonesia              elemental iron          100mg/day    2 months          12-13.3= 1.3
adult males                                                                    (11%)
3asta et al, 1979
Tanzania                oral iron                200mg/day   3 months          7.8-13.4= 5.6
iales age 18-35                                                                (71%)
Davies & VanHaaren, 1973
Venezuela               iron dextron inject.                 80 days       (F) 7.7-12.4= 4.7
males/females age 17-46                                                       (61%)
Gardner, ec al 1975                                                       (.Y) 7.1-14= 6.9
(97%)



TABLE TWO (continued)                           19
STUDY                    FORM OF IRON             AMOUNT       TIME PERIOD       Hb CHANGES
Guatemala                elemental iron as        100mg/day-   6 months          9.5-14= 4.5
a:iult -nen              ferrous sulphate         iron:5mg/day folic acid        (47%)
',iceri & Torun, 1974
:ndia                    B12, folate & iron tab   100mg 312    3 months          9.58-10.84= 1.26
pregnant femnales (22 weeks)                      fortnight                      (13%)
-ocd at at, 19975                                5mg folate-daiLy
120mg iron-daily
WI La                    B12, folate & iron tab  100mg 312     3 months         9.43-10.82= 1.39
?regnant females (22 weeks)                       fortnight                      (15%)
Sood et al, 1975                                  5mg folate-daily
240mg iron-daily
Mauritius'               ferrous sulphate-bread  extra 10 mg! 4 months           14.7-16.0= 1.3
addiLt mnales                                     day                            (9%)
Stoct, i960
.i dia                   elemental iron in salt   Lmg/g salt   12-18 months    'Madras (12mo)
ma tes, females ages 25-44                                                   (M) 15 .4-15 .9= .5
?od. & Nlutrition Board                                                          (3%)
ania UNICEF 1981                                                             (F) 12.4-12.9 = .5
(4%)
Calcutta (l2mo)
(M) 9.7-12.8 = 3.1
(32%)
(F) 8.5-11.5= 3
(35%)
Hzderabad (18mo)
(M ) 13.7-14.4= .7
(5%)
(F) 11.1-11.9= .8
(7%)



20
and rises in worker productivity of both mar'ket and home production, in
:eelings of welL-being and in intellectual functioning.         Clearly, a
cost-benefit analysis should attempt to incorporate the value of all of
threse into a measure of benefits. However, some of these categories lack
quantitative data on their relations to anemia (e.g. reduction in
inrect ion), and ochers are difficult to convert into monecary values even
if the relational data were available.       (e.g. feelings of wellbeing).
Accordingly, most of the emnphasis in this section will be on the relation
betweer, anemia on the one hand and work output and cognition or school
attendance and progress on the other. The final section of this report
wilL make an attempt to incorporate an. estimate of all of the benefits in
the construction of benefit-cost ratios.
Work CaDacity
It is in the area of work capacity and work output that the reduction
of anemia has the greatest demonstrable benefit. Both the underlying
unders tanding of the effect of anemia on work capacity and the empirical
Literature provide strong support for this relation. Muscle cells need
oxygen in order to function. Oxygen is carried by the hemoglobin through
the bloodstream to the cells.     If hemoglobin levels are low, then oxygen
transport is impaired with a resultant limication in human workl- capaci,'y.
One measure of the effect of anemia on work capacity is its effect on
maximat oxygen uptake (Astrand and Rodahl 1977: 334-344) . Maxiarum oxygen
uptake is the largest amount of oxygen that a person can take in during
exercise.    It is an indicator ofc the ability of a person to transport
oxygen to the tissues, and it is considered to be a good indicator of



21
fi tness.   The average middle age male hlas a maximal oxygen Uptake of -about
35-40 ml per kg of body weight per minute or about 2.5 liters of oxygen per
minute (Bogert, Briggs, and Calloway 1973:482). Other indicators of
fitness include the actual level of oxygen uptake and the heart rate for
any e:ercLse l2evel (As trand and Rodahl 1977: 344-347). The larger The
actuat Level of oxygen uptake for any exercise level, the less that the
body incurs an oxygen debt and accumulation of lactic acid. After the
exercise is completed, an individual must take up additional oxygen to
metabolize any accumulation of lactic acid. The lower the heart rate at any
exercise level, the more fit is the individuaL to deliver oxygen to the
tissues.
E,xperimnental studies have found that increasing Hb concentrations
through iron supplements is associated with an increase in oxyrgen uptake
and lower heart rates.     For example, Davies and Van Haaren (1973) divided
maLe subjects between 18-35 years of age in Tanzania into two groups, those
with low hemnogLobin leviels (Hb average = 7.8 g/dl) and those with high
hemoglobin concentrations (Hb average = 13.7 g/dl). The subjects ingested
200 mg day of oral iron for three months with a resulting increase in Hb in
the low group to 13.4 or 71 percent and a much smaller increase in the high
group to Hb =14.1. A regimen of exercise was administered to both groups
both before and after the iron supplementation.
Af ter iron therapy both groups showed improved oxygen uptake (V02) and
maximal oxygen uptake (VO2 max). Thie latter was predicted from V02 by using
standard teciniques.      'dear: races were also lower, indicating that the
h iher concentration of hemoglobin permitted the heart to work less hard to



22
deliver a given ar. zunt of oxygen.      The low Hb group had increased its
maximal oxygen uptake by 26 percent, while its average heart rate had
declined by 15 percent. Changes were mnuch smaller for the high group.
Elas ticities were calculated to provide a s tandard measure for
assessin-g the response of the various physiological neasures of work
capacity to changes in Rb .     Al though the concept and application of
elasticities wiilL be familiar to the e-coiomist, the non-economist should
refer to any elementary text book on economics for a discussion, for
example, HReiLbroner and Thurow (1975: 105-110); In this context, an
e las tic i ty wi ll represent the percentage change in a work capacity or work
ouc put variab le in response to a one percent change in Rb. An elasticity
of l means that for each one percent change in Rb, there wili be a I
percent change in a specific .,easure of work capacity or work output. An
elas ticity of .5 means that a I percent change in Hb is associated with a
.5 percent change in the pertinent measure of work capacity or output.
The advantage of us ing elasticities is that they represent a
s tandardized method of characterizing responses to changes in Hb by
representing them as the percentage increase (decrease) in a phenomenon
associated with a one percent change in Hb. Although the elasticity is
presented here as a constant or linear bi-variace relaticn, it is actually
an estimate of a curviLinear relation between the two variables because it
is a linear approximation of a logarithmic relation. That is, it is a
linear estimnate of the log of the dependent variable (work capacity or work
outpu-t) and the log of the independent variable: (hemoglobin). T-he
elasticity assumes that when initial Rb is high, it will require a larger



Z3
change in Rb to get an equivalent change in work capacity or work output
than when initial Hb is low.
Elasticity = percentage change in work capacity or output/ percentage
change in Hbb.
In applying the use of elasticities to the results from Davies and Van
Haaren (1977) presented above, each elasticity represents the percentage
chanae in V02, V02 max, and heart rate for each one percent change in fib.
For che low Hb group the elasticity for V02 max and heart rate were .36 and
-.20, indicating that a one percent increase in Rb was associated with a
.36 percent increase in V02 max and a .20 percent decrease in heart rate.
The comparable elasticities for the high group Hb group 'ere .93 and -.24.
The elasticities are actually Larger for the subjects who had high initial
[Lb because the physiological changes are associated with very small
improvements in Rb concentration.
Gardner et al. 1975 administered iron dextron injections to highly
anemic males and females in Venezuela. Initial Rb levels were 7.1 for
-nales and 7.7 for females.    After 80 days they had risen to 14.0 for the
males and 12.4 for thne females. increases of 97 percent and 61 percent
respectively.    Heart cates for a given exercise regimen declined fromn 155
to 1 13 for males and 152 to 123 for females, reductions of 26 percent and
15 percent respectively.   The elasticities of heart rate with respect to Hb
concentrations were -.44 for males and -.20 for females.



24
in cross-s ectionaI scudies by Davies (1973b) of males age 9-16 and by
Davies, Chukweumeka and Van Haaren 1973 of males age 17-40, it was found
that V02, V02 max, and heart rates varied in the expected ways with 'b.
Comparisons of males with average Hb of 9.2 and those with Hb of 14.5
showed elasticities of .63 for 'V02 max and -.36 for heart rates.
Work Output
The evidence suggests a consistent relation between Hb and various
measures of work capacity.      But, work capacity is not the same as work
output.   As the names imply, w4ork capacity refers to the capability for
performing work, while work outout refers to the actual work that is
achieved.   Work capacity and work output may differ for a number of
reasons.
First, work capacity is a limiting factor on the maxiuum amount of
work that can be performed.    If a job requires individuals to perform at
sub-maximaL Levels, low Hb may not be a hindrance. The types of jobs that
are likely to draw upon maximal aerobic capacities are those that are
hiighly physical and require continuous exertion, jobs that require stamina.
These jobs include much agricultural work, jobs in labor-intensive
manufacturing t-hat are typical of developing sXc cieties, jobs in
infra-s tructure industries such as construction and mining, and various
service jobs such as loading and unloading vehicles, transportation
services by foot or foot-driven vehiicles, and cleaning activities. They
also include maky household tasks such as cutting and carrying firewood,
drawing water, and hand-grinding of grain. These jobs represent a very
high proportion of work activities in developing SQcieties. In contrast,



25
such service occupations as office workXers, cashiers, or sales clerks are
less Likely to draw upon the upper limits of oxygen uptake.
Of course, we should aLso bear in mind thac even at lower levels of
exercise, a f it ter worker may be more proficient and produce more work
output.   The entire cardio-vascular system can wor' ac a lower level of
ef 'ort for delivering oxyuen to the tissues, with less fatigue for the
worker.   But, a low Hb concentration is likely to be a limiting factor
primarily for workers whose jobs are physically arduous.
Second, work output depends not only upon work capacity, but upon a
Large nurmber of other factors that will determine the worker's activity.
Among these are the mental and physiological capabilities of the worker
such as intelligence, skilLs, motivation, size, strength, and stature.
Other factors include the availability of work, access to tools and
equipmenc, incentives to work, and supervision. Finally, outdoor work is
especialLy affected by the weather so that rainfall can influence the
amount of work that is achieved (Popkin 1978).
However, even given these differences, research has shown a close tie
between Rb-related measures of work capacity and work output.  For example,
Davies (1973a) studied the relatioLn of V02 max to both productivity and
absenteeism among 78 cane cutters aged 18-50 years old in Tanzania.  Output
per day was found to be positively related to V02 max, and involuntary
absences from work were found co be negatively related, both at significant
statistical levels.    When the levels of V02 max (adjusted for body weight)
on the one hand and daily productivity and absenteeism on the other were
compared among groups of workers differentiated according to productivicy,



26
elasticities were very high.     For example, the apparent elasticity of V02
max on daily production between high and medium producers was over 2 and
between medium and low producers was about 10. Comparable elasticities for
the number of days voluntarily absent were 6 and 26.           It should be
emphasiized that these elasticities are probably overstated, since other
diffperences between the groups are not controlled in these analyses.
Nevertheless, the potential relation is impressive.
Spurr ec al. 1977 studied the nutritional status, bodily stature,
and produccivi,ty of 46 sugar cane cucters between 18 and 34 years of age in
Co lombia. Iron status among these workers was relacively high with a range
of Ub = 12.0-16.5 and a mean of Hb = 14. 1.         Differences in daily
productivity were explained statistically by a regression equation in which
V02 max, percent body fat, and height were the explanatory variables. The
V02 max variable was statistically significant and an important determinant
off productcivity.
Tab le three shows the relation between hemoglobin levels and measures
of work output for six studies that provide such data.   Several measures of
work output are used.      These include the Harvard Step Test, Progressive
Tread Mill, and measures of work output for latex tappers and weeders. The
Harvard Step Test is a measure for selecting individuals according to their
physical fitness that was developed during the Second rWorld War. It
requires an individua l to repeatedly step up and down a 50 cm, st)p at a
particular rate desiganed so that only about one-third of the subjects



TABLE THRiEE
hlEMOGLOBIN LEVELS AND MEASURES OF WUORK OUTPUT
STnDY                                 lib LEVELS                 IIARVARD SILP TEST (elusL       PRI'RL SSIvE TItEADMILL (elast)    J(lB AcTIVITIES (elast)
1.  Iiidjiesija                       Before   After                   69.41   (0.47)                                               I.jtex T.appers (kg/day)
dJitE I11.s1es                      12         13.5                82.20                                                            20.94   +(3.38)
tasca L Churclii1 . 1974,                                                                                                           29.78
basca Gj al. 1979                                                                                                               We!der_ sg.vierer/day)
91      +(t.84)
112
2. Sri Ldnka                          Before  AfLer
IU isoles, 35 female.,     L-L       6.4        7.6                                            6.2     9.5  *(2.83)
age! 39-54                 L-Ul      7.5        8.8                                            11.1   16.6  -(2.55)
Uniira, er al. 1981        H-L      11.8       12.1                                            14.2   17.4  +(9)
tI-Il    11.9       12.3                                            15.9   16.5  +(1.12)
if-ll    14.1       14.0                                            14.9   17.4  A
3.  Guatemasl                         Before   Afrer
B3ItIL Ute.11s                       9-5       14                  47       +(1.21)
ViEeri b Toruii, 1974                                              74
4.  Sri Lalnk                         Before   After                                                                                Tea (kdiday)
i-ia.Iale  age 20-60                 7.3       11.5                                                                                 1;.6    -(.22B
Edg.rcon. tL al. 1979                                                                                                                7 17.5
(arooghr season)
5.  111udelsia                        Rentang
via1es, age 25-28                 Low     8-6                      64.0     +(0.38)
9;ary.ardi 6 Bsrsa, 1973          High= 15.2                       82.3
Sa I ada ritea
Lou-   8.7                      5141
iliDh= 15.6                      74.4     *(U.58)
Hali5£
1Ln r-  8.0                      38.o
hligbl 15.6                     70j. 5    -(1.03)



TAULE !!lill E (cotnciguued)
iTIIDY                                 llb l.EVEl.S               JIAJiVABD STEP TEST (la.ti)     PIWOCESSIVE TItEAD.1IM (I I .4.)   JOB ACTIVITIES .elasc)
0.  Sri Linka                          Bierore  Atter
rnwd  age 22-6S          L       b.5        d.5                                            10.4    1 3.7  +(1.ol)1
l .i .e1. 1917               9.5       11.5                                            14.5    14.5  '(0)
11      12.5      13.5                                            16      18     *(I.St)
.WL t  cu a Eed



29
should be ab Le to perform the tes t for a 5 minute period (Astrand and
RodahI 1977: 344).     The definition of HUST scores is: greater than 89,
excelLent; 80-89, above average; 65-79, average to high average; 55-64, low
average; less than 55, poor.
The progr-essive treadmill reauires the subject to walk or jog at .ne
speed of a treadmill which is set eor a particular regimen that moves
progressively faster and at a higher grade of climb, the longer the
duration of the exercise.       The mneasure of work output is the number of
minutes that the individual is abLe to continue to exercise. However, since
che lonaer the individual continues to exercise, the fas.ter the speed, the
amount of work output increases at a faster rate than is reflected in the
length of the exercise period. This means that additional minutes for a
per formance of longer duration are mnore demanding than additional minutes
beyond a shorter performance.     Measures of job activity are determined by
the activity   itselE.    These include latex tapping, weeding, and tea
picking.
The 'db levels in table three include before and after comparisons
referring to Hb both before and after iron interventions for all cases
except 5. In the case of study 2 and study 6, ttue subjects are also
sub-divided into groups according to initial Hb levels.    Study 5 compares
separate anemic and nion-anemic groups rather than the before-after
intervention design.
Most of our attention will be devoted to the elasticity of work
outputs with relation to the changes in Rb. In this case, each elasticity
will refer to the approximate percentage change in work output for each
pe.rcent change in 'Rb. Elasticities are shown in parentheses.    Studies 1



30
and 3 show elasticities of over I for the Harvard Step Test, implying that
percentage increases in Hb are associated with even greater percentage
increases on the HST. For example, according to these two studies, a 10
percent increase in Rb would provide about a 15 percent increase in HST
per formance in the Indonesian sample in study 1 and a 12 percent increase
in the Guatemalan sample in study 2. Since these represent the results
c oinparing individuals both before and after iron supplementation, we can
have a higher confidence in these results than in the somewhat lower
eLasticities generated in study 5 which compare anemic and non-anemic
groups in three locarions in Indonesia. However, it is also important to
note that an evaluation of HST performance for individuals with different
levels of RSB in Guatamala (Viteri and Torun 1974: 614) yields an estimated
elasticity of about 1.6.
These reLatively high elasticities are also supported by results for
the progressive treadmill in studies 2 and 6 and the job activities in
s tudy I.  Study 2 breaks down the subjects into different initial Rb
ranges.   Although many of the elasticities are very high for the treadmilL
test, one should bear in mind that some of the result may be due to
proficiencies gained in taking the test initially.  In fact, the high-high
Rb group shows a substantial improvement in the progressive treadmill
without a change in Hb, suggesting results of practice, possible Hawthorne
effects, or a combination of these and non-hematological effects of iron
repletion   that are associated with improvements in iron status. This
latter exo lanation will be discussed below. Since additional minutes on
the progressive treadmill are more arduous than previous minutes, the



31
elas ticities for the higher 'Rb groups are understated relative to the lower
Hb groups.
Study 1 provides estimates for the same subjects for both the HST and
for two job activities, latex tapping and weeding. The elasticity for the
H ST is less than iialE tiat for latex tapping and slightly lower than that
tor weeding.     It may be that the HST task is too limiting to reflect the
wider range of differetices in job activities themselves.      In any event,
this evidence suggests that elasticities for the HST may understate what
the elasticities would be for jobs for the same subjects. Although study 4
shiows a very Low elasticity of work output for tea pickers in Sri Lanka,
the resuLts wera obtained during a drought when there was little tea to
pick, regardless of work effort.
What is particularly interesting is the pattern of elasticities for
grouos with different initial Hb. One might expect that the eLasticities
woul d be higher for improvements in Rb among highly anemict populations, but
ao such pattern appears in the data.          That is, the elasticities of
responses in work output to changes in hemoglobin concentration seem to be
similar for different levels of initial hemoglobin, even given that they
are understated for higher levels of performance on the progressive
treadmill.    This pattern is probably not applicable to those populations
with very appreciable iron stores, since it is unlikely that additional
.ron would improve their performance.
The overall results in Table Three suggest elasticities of work output
in relation to increases in Hb of between 1 and 2. This suggests that a
rise in Hb of 10 percent is associated with a rise in work output of
between 10 and 20 percent. This range is further supported by an elaborate



32
s tat is tical s tudy of road construction workers in the Philippines (Popkin
1978).   Hemoglobin concentrations and other pertinent variables were used
co explain differences in productivity with regard to the average daily
output of workers in loading, unloading and tamping soil.  The productivity
me.asure was the number of cubic meters of fir,n soil per day achieved byo the
worker through che loading, unloadaig, and tamping process.     Fifty-eight
percent of the workers liad hemoglobin levels of Less than 13 with an
average level -or all workers of 12.4. A doubLe-logarithnic regression in
which daily soil output per worker was the dependent variable showed an
eLasticity of hemoglobin on work output of 1.45.   The study also estimated
non-logged regressions for days mnissed in the previous six weeks, the
average time worked per day (based on time and motion evaluations), and che
proportioni of time worked in the day. Estimates of elasticities based on
2 3 p k i n' s d a t a were: days m i ssed, -3.55; time worked, 2.08; and percent o f
t uine worked, 1. 79.  Since the coefficient for days missed was not
statistically significant, its precision is open to question; but the other
coefficients were s tatisticaLly significant.       The range of elasticity
results are consistent with those of the iron intervention studies in table
th ree.
The only study that I am aware of that shows a lower elasticity on job
activities is that of Kenyan road constr-.:tion laborers (W;olgemuth, Latham,
Hall, and Chesher 1982).      These workers had unusually high levels of
hemoglobin concentration for a rural region in an LDC, an average Rb of
13.3 and 13.0 for the cwo female samples and 14.71 and 14.53 t     tne two
male samples.     Although this study found that the impact of a diecary
supplement which included caloric, protein, and iron inputs increased



33
worker productivity by about 12.5 percent,      the separate effects of the
ironi intervention on prodLuctivity were not isolated by thle analysis.
However, Rb was included as an explanatory variable in a regression
equa t ion on productivit y for the pre-intervention sample. On the basis of
the regression coefficient Eor Hb, my c3mputed elasticity coe r ficient was
.64.   In interpreting this result, the very high average Hb and the high
altitude for the sample should be taken into account.
With the exception of the Kenyan study and the Sri Lanka tea pickers
in which the drought had intervened, the range of elasticities seems to be
about I to 2 with 1.5 being the best overalL approximation. But, how can
one reconcile these. high elasticities with the much smaller ones relating
Hb and such physiological irndicators as V02 max or heart rates?
Such dif ferences are perfectly consistent if one acknowledges that
some of the effect of iron status on work output is likeLy to be correlated
with, but independent of, Hb. With an improvement in iron status of the
individual, both hemotological and non-hemotological factors may improve
work performance so that measures of Rb and oxygen uptake may reflect only
a partial picture of the underlying relations (Edgerton, Ohira, Gardner,'
and Seriewiratne 1982: 146-150). To the degree that increases in iron intake
and Rb are correlated, rises in Rb will serve as a statistical surrogate
for other iron-ind'uced improvements in work output.
In this respect, it is important to note the recent research focus on
che non-hemoto logical effects of iroti deficiency (Mackler and Finch 1982;
Jacobs 1982; Oski 1979). Support for an additional non-Rb effect of iron
s tatus on work output is provided by Ohira, Edgerton, et al. (1981) and
Ed-erton & Oh ra (1981) who found that persons macched on Hb, but with



34
higher levels of serum iron had substantially higher performances on a
treadmill.    Also, Baker and DeMaeyer (1979: 386) cite a double-blind study
by Ericsson in which subjects received an oral supplement of 120 rng of iron
a day for three -months resulting in increased work perfornance as measured
on a bicycle er-omneter, despite no increase in hemoglobiA concentration.
They concLude that this effect may be related Co evidence in rat studies
that iron deficiency is related to striated muscle dysfunction which is
reversible with iron therapy. Effects of tissue iron deficiency are also
summarized by Dallman (1981).
The presence of iron-induced effects on work performance that are not
a result of elevated Hb is also consistent with the findings that the
elasticities relating Rb to work output are considerabLy greater than those
for V02 max.     If some of the improvement in work output associated with
improved ironi status is due to factors other than the effect on oxygen
uptake--eg. effects on the muscles, brain, or nervous system--, then
changes in Rb may also serve as a proxy for the correlated, but unmeasured,
physiological and mental changes that affect work performance.         This
explanation   is aLso supported by the very high elasticities that we
calculated for V02 max and both daily productivity and absenteeism in the
Davies (1973) study. In that stud7, the elasticities between V02 max and
productivity were estimated co be 2 and L0 and those for V02 max and
absenteeism were estimated to be 6 and 26 for the different comparisons.
Finally, since our estimates of work output refer only to output on a
particular activity (HST Or progressive treadmill) or daily work output ac
a particular job, they do not take account of effects of Rb in raising work
output by    reducing absenteeism. '4orkers with higher Hb will have higher



35
daily outputs, and they wilL aLso work for -aore days of Ltie month or year
for reasons discussed in Davies (1973a). Thus, the elasticities for daily
output would .understate the elasticities of [ib with respect to monthly or
annual work output.   It is reasonable to conclude that an elasticity range
Of 1-2 is ptDbably a conservative estitnate of thie relation beuween hb and
annual work output for the physicalLy arduous tasks reflected in table
t hree and that are so representative of work in labor-intensive agriculture
and other work activities of LDC's.
Cognition and Schooling
A variety of studies has found effects of anemia on learning in
infants, children, and adults (Leibel, Greenfield, and Pollitt 1979:
400-410).    However, prior to very recent work, each of the individual
s tudies has been open to reservations because of methodological
shortcomings and inconsistencies in results. Moreover, differences in
study populations, designs, and measures have produced a wide range of
results so that a 1979 summary by noted authorities concluded that:
"Des pite strongly held clinical impressions and firm lay-person acceptance,
there exists no unequivocal demonstration of an adverse effect of iron
def iciency on intelligence, learning, actention, motivation, or general
sense of well-being (Leibel, GreenfieLd, and PolLitt 1979: 431)."
However, subsequent work by the same authors with 3 to 6 year old
chi ldren ( Pol Litt, Greenfield and Leibel 1982) as well as assessments of
the findings of Lozof f et al. (1982a) and Oski and Honig (1978) and
Oski (1979) have led to the conclusions that iron deficiency '". has
adverse e ffects on cognition, and that these are reversible following iron
rep Letion.  The effects are mild and most probablv located at the level of



36
informat ion reception (Po Ilitt, Viteri, Saco-Pollitt, and Leibel 1982:
297).'
More recently, Popkin and Lim-Ybanez (1982) have published an
extensive analysis of the relation between nutrition and school achievement
for 1L32 chiLdren, ages 12-14, in three rural and three urban schools in the
Greater Manila area of the Philippines. A significant positive relation
was found beteen hb and the language test score, with no statistically
siginificant relation between Hb and science or mathematics scores. Hb
seemed to have no relation to student ability to concentrate or student
participation in extra curricular activities. However, it showed a
significant, but slight, negative relation to the number of days absent.
Moock and Leslie (1982) studied childhood malnutrition and schooling
among about 400 school-age children from subsistance farm families in
Nepal.   They attempted to explain the probability of a child being enrolled
in schoo l and the progress of a child in terms of the grade leveel attained
by us ing a host of parental, family, comrmnity, and child characteristics
including Hb of the child.      ALthough hemoglobin was not found to be
statistically related to either school outcome, two anthropometric
indicators of nutritional status, height for age and weight for height were
important determinants.   The more important of these variables seems to be
height for age, and Jamison (1981) found that height for age was an
important determinant of grade attainment in China. Since stunting may
occur through anemia, it is quite possible that anemia has an indirect
effect an school enrollment and grade attainment through its retarding
effects on growth.     However, the evidence on the relation between Hb and



37
school success or intellectual growth is not as consistent as that relacing
'b to work capacity anid work output.
Other Areas of Potential Benefit
Anemia is associated with a number of other areas of debilitation that
night be reduced through approprLate intarventioas         Summ aries of the
e f tec ts o f s evere anemia during pregnancy have established an association
with increased risk of both maternal and fetal mortality and morbidity
(1rLNACG 1977: 2).   Even milder anemia has been associated with premature
delivery and low birthweight (INACG 1977: 2; Baker and DeMaeyer 1979:386).
There is also some evidence that iron deficiency is associated with lower
weight gains amonig infants and children (Baker and DeMayer 1979: 384;
Burman 1982; Oski 1979). Many researchers believe that iron deficiency may
increase susceptibility to infection, although the evidence is not
straightforward    (Baker and DeMaeyer 1979: 384-385; INACG 1977: 2-3;
Nutrition Reviews 1975: l03-105), and there is some support for the view
cltat iron deficiency is a defense against certain types of infections and
is a factor reducing the probability of heart disease (Callender 1982:
327).   Symptoms commonly associated with anemia are fatigue, headaches,
weakness, lightheadedness, and irritability.   Such pheniomena are difficult
to define and measure, and various studies have found no evidence between
the ectent of these symptoms and the severity of anemia (Leibel,
Greenfield, and Pollitt 1979:399).
This section has discussed a number of potential benefits of anemia
reduction and focussed in somne detail on the decreases in work capacity and
output associated with iron-deficiency anemia.   The remaining parts of the
report wiLl address the calzulations of costs and benefits.



38
IV COSTS OF INTERVENTIONS
Thu s far we have reviewed the types of benefits that one mi,ght expect
from reducing iron deficiency anemia as well as their magnitudes.  Special.
attention was devoted to those associated with work, output.       In this
section, w7e wiLl address     the costs o. potential inLterventiLons for
addressing iron-deficiency anemia, and in the following sections we will
attempt to calculate the monetary value of both costs and benefits.
There are two major issues when addressing costs.  First is the macter
of what constitutes a cost and how to measure it.   Second is the question
of what Eactors are likely to determine the costs of interventions.     In
generaL, the term cost is used to refer to the sacrifice of a valued
a l ternative. WAheo resources are used to reduce anemia, the cost to society
is the value of what is being given up in the best alternative use of those
resources.    This criterion is important because it distinguishes the
economic definition of cost from a bookkeeping or accounting definition and
from the issue of funding. Budgetary data typically understate the true
cost of an intervention by omitting the value of resources that are
contributed (e.g. facilities or volunteers) or understating the value of
resources that are subsidized.
A straightforward approach that has been developed to estimate the
costs of interventions is the ingredients method (Levin 1983). This
approach requires an identificacion of the specific ingredients or
resources that are likely to be required. Once the ingredients are
identified and described in adequate detail, the value or each is
determined by using standard costing methods.   These costs are aggregated



39
to determine the cost of the overall intervention, and they are analyzed
according to whether one is concerned with the average cost per client or
unit of service or some other criterion.   Finally, the data can be used to
to ascertain which constit.uencies are bearing the costs of an on-going
project.
The costs of the intervention for any service level will be influenced
by several types of factors. The most important are obviously the types of
resources chat are required, the costs of those resources, and the
productivity of the intervention in providing services.  In the ideal case
we wouLd obtaLn intormation frown field trials of a range of proposed
interventions in a variety of settings in order to establish re.2ource
requirements, costs, and productivity. In the absence of these data, one
can rely on documentat ion from     previous: s tudies.  However, little
systemnatic collection anEid analysis of cost data is avaiLable on anemia
in tervention programs. In part, cost studies of other health interventions
can be used as a guideline for evaluating costs (Robertson 1984)) and
particularly those studies that are somewhat analogous to supplementation
such as immunization programs (Creese et al. 1982) or anti-helminth
drug   programs    ( Stephenson et al. 1983).     In the case, of iron
fortif ication programs there seems to be at least some direct evidence on
costs (Cook and Reusser 1983).
Required Ingredients
Based upon these s tudies as well as the obvious requiremnents of
nedicinal supple:.entation of iron compounds, supplementation would require
the following basic ingredients:



40
(1 ) Personnel--The main personnel costs at the site level are those
associated with che distribution of the iron supplements; record keeping;
and health education.     In a smualL village these responsibilities might be
incorporated inco a fulL-time community health worker. In a larger secting
tiere might be a division of Labor with supervisory personnel, communicy
heal th d orkers, c lericals, and warelhouse personnel. However, the latter
might be considerably less costly per client served because of the
economies of scale from a high client density.
(2) M4edicinal Supplements--Clearly, a central ingredient is that of the
iron compounds, Eolic acid, and absorbency enhancers such as ascorbic acid.
(3) Trallsportation--In order to distribute the supplements and provide
information or instruction to clients, a system of transportation is
needed.   This can vary from public transportation and bicycles in urban
areas to motorbikes o.r motorcycles in outlying areas with reasonable roads
to four wheel vehiicles or even animal transport in extremely remoce areas,
Equipment, maintenance, and fuel requirements must be taken into account.
(4) Facilities--At the village level the facilities requirement is likely
to be minimal with a small office and storeroom being sufficient. The iron
compounds and other supplements do not normally    require refrigeration or
other special treatment, although extremne heat and humidity may require
soecial arrangements.      In urban areas, the intervention could utilize a
small portion of a larger health Eacility. In addition to the facility,
such related ingredients as supplies, maintenance, and energy needs -nust be
inc Luded.
Within these general categories, it is the precise ingredients
required for an intervention that will deter.aine costs. Factors thac nusc



41
D e t aken  in to  account on the natuire of the intervention (e.g.
supplementation or fortification approach) include the severity of anemia
and its specific causes;      cultural and dietary ciharacteristics of the
population; and the degree to which an interlvention delivery mechanism
already exists.
Obviously, the iron supplementation or absorption enhancers used must
take. account of the etiology and severity of the anemia. C u 1 t u r a 1 a nd
dietary differences may affect the receptiveness of the population to
different forms of supplementation. For example, some populations may
require an educational component to inform and encourage the seLection of
iron-rich food sources or to increase the likelihood that the dietary
supplement is taken on a daily basis for the entire regimen. If a health or
nutritional delivery mechanism already exists, few additional resources may
be required to provide mediciaal suppLemenitation for reducing anemia. The
costs of adding iron compounds to an existing fortification program tmiay
also be low in comparison with developing a unique program for iron
foL tification. In both of these cases the marginal costs of "piggy-backing"
the delivery of iron compounds or absorbancy enhancers on existing programs
will be low because few additional ingredients will be needed.
Cost of Ingredients
A second factor influencing intervention costs is the cost of each
ingredient.    The same ingredient may be characterized by widely different
costs in different societies, based largely on considerations unique to
each setting.    For example, Lte relative scarcity of different types of
labor in different labor markets will affect its costs. Creese et al.
(1982:624) found that tLe daily salaries of capable vaccinators in



42
Indonesia and ThaiLand in 1979 differead by more than 250 percent.
Facilities, equ ipment, and transportation costs 4ill be heavily conditioned
b v the avaiLability and quaLity of roads or other thoroughfares. The
design and construction of facilities requirements may vary according to
climate or the need for security.        Even regional differences vi:hin a
society such as urban-rural distinctions in labor markets, transportation,
or weather conditions can be important sources of cost differences.
Final Ly, the cost oE pharnmaceuticals may vary according to whether they are
imported or manufactured domestically, government policies on imports, and
the competitive structure of pharmaceutical markets.
Productivity of Service Deliverv
The costs for servicing a given population will depend upon the
efficiency with which a supplementation unit or fortification program
functions as welL as the ability to reach the target populations. The
discussion of factors affecting organizational efficiency is beyond the
scope o f this paper. However, a major determinant of productivity, the
number of clients that the delivery system can efficiently serve, is
closely tied to the size and density of the geographical area as well as
its transportation facilities.   The high population deasity and relatively
good transportation found in many urban areas enables a larger population
to be served by a given configuration of ingredients than in rural areas
and especially rural areas characterized by great distances among the
population and poor roads. Clearly, if an organizational model with a
g,vien set of resources can serve 10,000 persons annually in some areas, but
only 200-300 in other areas, costs will vary enormously.



43
V- CALCULATING COSTS
Wi th this background on costs, it is possible to estimate the costs of
hypochetica L intervencions for reducing iron-deficiency anemia. We will
consider separatelIy the costs of fortification, supplementation where the
.eL very system e xists, and s u oLeme ntatLon in tiie absence of a deLivery
s vs tea.  3ased upon the previous discussion, it is impossible to determine
costs for a generic intervention because of the range of different
circumstances that are likely to be encountered. Accordingly, the emphasis
wiLl be on the establishment of a range of costs.          In each case the.
assumptions wilL be stated and their consequences so that the reader can
modify them for any particular population or situation to bring them in
line with other realities.
Costs of Fortification
Although supplementation with therapeutic doses of iron is often
recominended as a short-term measure for imnproving iron status in a
population, over the long run it is crucial to improve che dietary intake
and absorption of iron through fortification.    Daily requirements of iron
that must be absorbed to maintain homeostasis are estimated to be from
about .7 mg for infants and .9 mg for men to about 3.0 mg for women in the
second half of pregnancy (Baker and DeMaeyer 1979: 375). But, some iron
needs will be met through the normal diet, so fortification need only meet
the daily sliortfall.  For example, if 75 percent of iron needs are muet from
::onventional sources, only 25 percent need be met through fortification.
Given the absorption rates of 15-20 percent associated with iron fortified
sugar administered through different beverages (Layrisse et al. 1976),
Chis means that che additional dietary input of iron that would nave to be



44
satisfied by this vehicle would be about 1.5 mg for men and about 4.3 mg
for women in the second half of pregnancy.  Accordingly, the maintenance of
adequate iron status through dietary fortification requires relatively
small amounts of iron in comparison with therapeutic supplementation for
repLetion of iron under condition-s of severe iron deficiency.
Iron fortification is believed to be the optimal approach for reducing
iroa deficiency anemia because it requires virtually no special effort on
the part of the population and has very low costs. The low costs are due
to the fact that fortification simpLy entails the addition of such-
substances as iron or ascorbic acid to a. food vehicle that is widely
consumed.    The only cos ts beyond those of the food vehicle which would
normal ly be consumed are those associated with the costs of the fortifying
nutrients and s tabi lizers and their processing into the food vehicle as
well as any special packaging or distribution requirements.
The relatively low cost of general fortification programs is
i llus trated by the case of wheat flour fortification in India (Bender 1979:
168).   Beginning in 1 9 70, each ton of wheat or atta has been fortified
with: edib le grade groundnut flour (45-50% protein) 50 kg, retinol 9.2 g,
ribo f lavin 1.38 g, nicotinic acid 7.6 g, thiamin 1.5 g, calcium diphosphate
800 g, ferrous sulphate 96 g, and calcium carbonate 800 9.           This
fortification added only 4 percent to the cost of wheat Elour. However, a
serious deficiency was the fact that only about 15 percent of the fLour
that was consumed each year was processed by the large mills where
fortificacion couLd takae place, with the rest ground in small, hand
operated mills for local consumption.        In fact, that is the central
^hal lenge o f fortification, finding a vehicLe that is consumed by most of



45
the target population in adequate amounts; capable of being fortified in a
few processing centers; palatable and relatively unchanged in appearance as
a resulc of fortification; and that is stable under conditions of storage,
use, and distribut ion.
Sugar and salt tend to be centralLy processed so that 'ortification of
3 r e latively high proportion of consumption is feasible.  In addition, they
are excellent vehicles for appropriate iron compounds with respect to
appearance, taste, stability, and effectiveness (Report of the Working
Group on Fortification of Salt With Iron 1982; Layrisse et al. 1976).
In addition, i t has been shown that both salt and sugar can be used to
carry both iron compounds and ascorbic acid as an absorption enhancer w.ith
good results on. ironE absorption (Sayers et al. (1974) and Derman et
al. (1977).    We will focus on the costs of salt and sugar fortification
because these seem to be the most widely tested       vehicles, evidence on
their effectiveness is available from field trials, and cost data exist.
In contrast, other promising fortification vehicles are still in the
developmental stages or need field trials to judge their effectiveness and
costs (Cook and Reusser 1983).
We will base our cost analysis of salt fortification on the actual
costs cited in the Report of the Working Group on Fortification of Salt
with Iron (1982) in large scale field trials over 12-18 months at three
rural sites and an urban one in India, each site covering 4000 to 6000
inhabitants.   Fortification consisted of 3.5 grams of ferric orthophosphate
added to each kilogram of common salt.  This level was estimated to provide
an additional 10-15 mg of iron intake a day in adults (at about I mg of
elemental iron per grain of salt).   Statistically significant increases in



46
Hb were found in the three sites for which valid data were obtained--the
eva luation at one site encountered operationial difficulties.  The strongest
e f fec ts we re found in Calcutta where increases of about 3 a/dl in Hb over
initial levels were found in contrast to little or no change in control
groups.   Changes in 1b were smaller at other sites, in about the range of
.8 g/dl in Hyderbad and about .5 g/dl in Madras, wich important differences
by gender and age.     As we noted in the discussion of Table Two, the
severity of anemia was much greater in Calcutta, almost certainly
accounting for the substantially greater Hb resoonse to Lortification.  The
fortification effort added about 20 percent to the cost of the salt
(Working Group on Fortification of Salt with Iron 1983: 1450) or an
estimated $0.07 per person per year (Cook and Reusser 1983:652).
The analysis of sugar as a fortification vehicle draws upon the fieLd
trials carried out in Guatemala by Viteri et al. (1981) in which two
communities in thie highlands and one comimunity in the lowlands received
fortified sugar over a 31 month period. The changes in the prevalence of
anemia were compared at the beginning and end of the fortification trials
and were compared with two communities that had been selected as controls.
Sugar was fortified with NIaFE EDTA at a ratio of 13 mg of iron per 100 g.
sugar.   Mean daily consumption of sugar was 40 g., so that the approximate
daily iron intake from the fortification source was over 5 mg. The
Lncidence of anemia was reduced substantially in alL of the communities
receiving forcirication in contras'. with the controls. Biochemical indices
of iron status also subst antiated the changes. According to Cook and
Reusser (1983: 653), the iron fortification added about 1-2 percent to the
cosc of sugar or about $0.10 a year per pecson.



47
In summary, studies of iron fortification of sugar and salt have both
concluded that costs are less than $0.10 a year per person. However, we
hlave been unable to obtain any systematic documentation that would account
for, in detail, the processing and distribution costs, although some of
these costs for salt fortification are discussed in Working Group on
Fortification of Salt With Iron (1982: 1450) and Food and Nutrition Board
and UNICEF (1981: Annex 3). Accordingly, we will make the conservative
assumption that these costs do not fully account for the resources required
for the intervention and will assume that they represent a lower limit with
$0. 30 a year per person as an upper limit and $0.20 a year per person as
the intermediate or "best" estimate of the costs of iron.fortification.
Experimental studies have also b,een carried out in which ascorbic acid
was added to rice meals along with ferrous sulphate (Sayers et al.
1974). In one of the experiments, 4 mg of ferrous sulphate was added to a
r ice meal and compared with the effects of adding 4 mg of ferrous sulphate
and 60 mg o f ascorbic acid. The apparent effect of adding the ascorbic
acid was to raise iron absorption from 4.2 percent to 12.2 percent. The
study also used commorn salt as a carrier for both the ascorbic acid and
ferrous sulphate and found no evidence of discoloration or change of taste
in a temperate climate, although it stated that this could change under hot
and humid conditions.
Derman et al. (1977) have reported on adding ascorbic acid to cane
sugar.   The ascorbic acid was dissolved in distilled water and sprayed onto
che dampened sugar which was subsequently dried under warm air. They found
that the process did not alter consumer acceptability. The addition of 50
mg of ascorbic acid through this carrier was shown to improve the



48
absorption of iron nine-fold       from maize-meaL porridge. The authors
concluded that fortification with ascorbic acid alone may be highly
desirable when iron deficiency results from low absorption from a primarily
cereal-based diet.    The cost of stabilized ascorbic acid  was estimated to
be about $ 10.70 per kilogram in 1982 (INACG 1982: 31) representing a cost
of about $0. 39 a year per person for 100 mg daily (50 mg of ascorbic acid
in each of two daily meals). There would also be an additional cost for
fortifying the 3.65 kg. of sugar used as the carrier. While no costs are
provided, it wouLd be surprising if this relatively simple fortification
process exc eeded $0. 20 for this small quantity. Accordingly, we estimate
the total cost to be less than $0.60 a year per person.
Costs of Supplementacion
There are two issues regarding supplementation that need emphasis.
First, medicinal iron supplementation should be viewed as a short-run
therapeutic s trategy to raise the iron status of a population to normal
levels in a relatively short period of time (e.g. three months). Once this
is done, dietary fortification is the proper long-run strategy to maintain
appropriate LLon levels. Accordingly, a benefit-cost analysis should not
view supplementation as an a lternative to fortification, but as a
complement.      Even in the absence o E medicinal supplementation,
fortification can improve iron status immenselv as the Indian salt study
slhowed for the Calcutta sample, raising Hb from 8.3 for females and 9.7 for
males to 11.5 and 12.8 respectively.
Second, the establishment of a delivery mechanism for a single dietary
intervention does not make a great deal of sense for populations that are
typicalLy suffering from several dietary deficiencies or health problems.



49
Medicinal iron supplementation should be considered as one of a number of
nutritional or health interventions provided by an overall system for
delivering such services. There is a large tixed cost for establishing a
health care or nutritional supplementation delivery system, a cost that can
az shared among many interventions.       For examnple, the marginal cost of
providing an additional medicinal supplement under an existing delivery
sys tem may be little more than the cost of the suppLement.
The construction of a delivery system to be used exclusively for iron
supplemen tat ion would be both costly and. wasteful, given the
underutilization of its capacity to deliver joincly a variety of
nutritional and health services. Accordingly, the most reasonable basis
for making cost estimates for supplementation is to assume a "shared"
delivery system in which the marginal or average cost per intervention is
the pertinent one.     At best, the single service delivery system for iron
supplements should be viewed as an upper limit on costs.
The estimates of costs for supplementation will be based upon two
different overall assumptions.  In the first case it will be assumed that a
delivery system exists that requires only the marginal addition of the
dietary supplements in which case it is only the cost of the supplements
that wiLL be included. In the second case it will be assumed that a shared
deliverv system is used for at least four different dietary supplements or
lhealth services.   We will attribute one-fourth of the overall costs of the
delivery sys tern to the anemia intervention as well as all of the costs of
the iron and ascorbic acid supplements.
The strict marginal cost assumption for imedicinal supplementation
assumes that a nutritional or health care delivery system is already in



50
p lace that provides a capability for delivering an additional service at
only the additional cost of the supplement.   In both the workplace and che
commun i ty, such delivery mechanisms exisc. For example, in both factories
and farms, meals are sometimes provided to workers. Providing iron or
ascorbic acid as medicinal supplements or in fortified meals would entail
mainly the cost of the nutritional suppLements. Likewise, the existence of
community health delivery systems will enable the provision of an
additional nucritional service along with c ther dietary and health
interventions. In these cases, the marginal cost of providing medicinal
supplements would be limited to the costs of the supplements themselves.
Based upon the interventions in Table Two, a typical intervention
would provide 100-200. mg of ferrous sulphate a day for 2-3 months. Such a
supplement would be expected to increase Rb by from 20-50 percent,
depending upon the initial Hb, the specific populations, and the existence
oF parasices as welL as other pertinent faccors. In 1981 the cost to UNICEF
of 1000 tablets of 60 mg of iron sulphate with .5 mg of folate was about
$1.00 (DeMaeyer 1981:366) or about $1.10 a year for 180 mg per day. The
U.S. governmenit depot from which the U.S. Public Health Service obtains
pharmaceuticals was charging $7.80 for 1000 tablets off 500 mg of ascorbic
acid in March 1984 or a cost of less than $3.00 a year for one tablet a
day.   Presumably, the cost would be somewhat lower for a regimen of three
.ablets of 100 mg, a day of ascorbate taken with meals to increase iron
absorption.    Or course, all ascorbic acid costs might be higher in
developing sociecies unless the pharmaceuticals were purchased in large
quan t ities and distributed bv a multi-national organization such as UNICEF.



51
The supplemental intervention rwould be based upon using either
medicinal iron or ascorbic acid.     In the case of inadequate iron intake,
the iron supplement would be the likely choice.  In the case of inadequate
absorption of iron, the ascorbic acid might be chosen. The massive iron
s! pplemenit woul A1 not require absorption enhancers, and there is little
evidence that enhancers can improve absorption of medicinal iron.
Accordingly, the marginal cost of supplementation with an existing delivery
mechanism would cost about $1.10 a year per person for iron with folate to
about $3.00 a year per person with ascorbic acid.
The development of a delivery system for anemia interventions and
other purposes requires more discussion. lWhile we will refer generally to
the requirements and costs of sucth a system, we will be particularly
concerned with the approximate costs of the system when applied to
Indlonesia, Kenya, and '4exico. These countries will serve as illustratiorns
For the benefit-cost analysis.  The delivery system could be built around a
community or village-based health care approach (Djukanovic and Mach 1975;
Hetzel 1978; PAHO 1973; WHO 1979). Such a system makes heavy use of
community resources as well as health auxiliaries or community health
workers.   The model is a general one with vastly different forms of
implementation and cost implications in different societies (Robertson
1984) .
Health auxilLaries are persons who have completed all or most of
primary    school and are literate     in basic reading, writing, a.-d
comnputational skills.   They are typically drawn fromn the local comnunity,
so that they will relate well to the populations whom they are serving.



52
They can de live nutritiornal supplements and proviTe infornacion on their
use and che importance of taking the entire regimen. They are also able to
provide innoculations and other health services.
Such personnel are given short training prograins to assist them in
Learning the healch functions thac they wilL serve, the specific tasks that
they wiLL perlorm, and the information that they wilL need co answer basic
questions and to provide health education. While they may work directly
under supervisory personnel in larger centers, health auxiliaries in rural
areas wilt have only occasional and intermittent contacc with more highly
training personnel or administrators.
The basic model that will be used here assumes that a health a.xiliary
can serve a large village of 1000 inhabitants or several smaller ones that
add up to 1000 (e.g. two adjacent villages of 500 inhabitants each).
Creese e- al. (1982) has reported daily wages in 1979 of midwives and
sanitarians in Indonesia, the Phillipiaes, and Thailand. Pay rates vary
from $ 2. 24 a day for a vaccinator in Indonesia or $560 a year for 250 days
to $5. 90 a day for a midwife in Thailand or about $1500 a year. We will
assume chat it is the higher figure that is necessary to obtain the skilLs
and experience required. This f igure is probabl        somewhat high for
Indonesia where the cost of an inoculator for 250 days was less than $600 a
year according to Creese' s f iaures and for Kenya where according to
Stephenson et al. (1983: 183), community field workers were used to
provide antihelminch medications to children at a cost of $2.50 a day or
about $625 for a 250 day year.     In contrast, it may be low for Mexico, a
factor that is taken account of in Section VII where a sensitivity analysis



53
is done us ing the assumption of a very high cost of auxiliary health
personnel, $4,500 a year.
Facilities tor a single health auxiliary are likely to be minimal,
requiring a small off ice with storage facilities.      In rural areas the
faciLicy is l ikely to be constructed from local materials and by local
Labor at very Low cost. We assume that in urban areas the costs will be
higher.   Accordingly, we estimate the cost of the facility at about $2,500
in low cost rural areas and about $10,000 in high cost urban areas, with
about $5,000 in the middle range.       Of course, in many cases the space
requirements will be met by a room in a larger facility such as a comaunity
health center in an urban area or a school, church, or home in a rural
.area,  If i t is part of a larger facility in an urban area, the marginal
Cos t o f the s pace should certainly be less. Using a 10 percent interest
cate, the annua lized cost of a $10,000 facility with a 20 year life is
abou t $1, 175 and the annualized costs of $5 ,000 and $2,500 are $588 and
$294 respectively (Levin 1983: 70).
Transportation can be provided by public systems, bicycle, moped,
motorcycle, automobile, or four-wheel drive vehicle.       In urban areas,
public transportation, bicycles, and mopeds are feasible, while in rural
areas the road conditions will determine the appropriate means of
transport. Based upon current prices, we wiLL assume the following purchase
costs:   bicycles $200; moped $800; motorcycle $2,500; smaLl automobile
$6,000; four wheel drive vehicle S10,000. Assuming a 6 year life and 10
percent interest rate for each, annual costs are: bicycle $46; moped $184;
motorcycle $574; automobile $1,378; and four wheel drive vehicle $2,296.



54
'4e wi L assume operation and mnaintenance costs for each are about equal to
t he annua l ized cost of the vrehicles so that the total annual cost wilL be:
' icycle $92; moped $368; motorcycle $1148; automobile $2,756; and four
wheel drive vehicle 54,592.
Although it is difEficult to estimate the cost of materials and
suppLies (exclusive of medicinal supplements), it would seem that $1,000
would be sufficient.     This would be used for records, comriunications,
written information for literate clients, office supplies, and energy.
Table Four provides estimates of the annual cost of delivering
medicinal supplements to reduce anemia.         The ingredients for service
deelivery are estimated on the basis of low cCost, medium cost, and high cost
assumptions.   Personnel costs and suppLies are simiLar-in all three cases,
but different assumptions are made on facilities and transportation costs
as discussed above.     The total cost per year is about $3,200 for the low
cost case, $5,800 for the medium cost case, and $8300 for the high cost
case.   Since it is also assumed that there are at least three ocher
nut rional or health interventions that would be carried out under this
approach, we divide the total costs by four to obtain the average costs for
che anemia intervention.    Thes3 values are diLvided by 1000 inhabitants to
obtain per capita costs of service delivery. Clearly, the estimates would
be lower if this service delivery model ccould cover more clients as in
urban areas, and it would be more costly in very sparse areas where the
population is too dispersed to be able 'co serve 1000 persons by this
mod el.



55
TABLE FOUR
ESTIMiATED ANNUAL COST FOR DELIVERING MEDICAL SUPPLEMENTS
TO REDUCE ANEMIA (BASED ON SERVICE FOR 1,000 PERSONS)
Engredients                 Low Cost         Medium Cost         Hiih Cost
Personnel                     $1,500            $1,500             $1,500
Facilities                       294               588              1,175
Transoortation                   368             2,756              4,592
(moped)            (auto)          (4 wheel drive)
Supplies                       1, 000            1,000              l,000
Total Cost          $3,162            $5,844             $8,267
Average Cost                  $  791            S1,461             $2,067
(Total Cost . 4)
Per Capita                    $    0.79          $   1.46          $    2.07
(-i I,000)
Including Ferrous SuLlphate   $    189           $   2.56           S   3417
Including Ascorbic Acid       $    3.79          $   4.46           $   5.07



56
PopuLation density has been a very important factor in explaining costs of
immunization programs (Creese et al. 1982:629).
The per capita costs of service delivery are less than $1.00 for the
low cost assumptions, about $1.50 for the medium cost assumptions, and a
bi-, over $2. 00 for the high cost assutnptions.   To these we .must add the
costs of the medicinal suppLements resuLting in costs of almost $2.00 a
person with ferrous sulphate or almost $4.00 with ascorbic acid for the low
cost model; about $2.50 with ferrous sulphate and $4.50 with ascorbic acid
ror the medium cost model; and about $3.00 with ferrous sulphate and $5.00
with ascorbic acid for the high cost model.
Hizher Caloric Needs
The final identifiable cost 'or both aortification and suppLementation
is associated with the higher energy needs for non-anemic persons who are
engaged in strenuous activities. Supplementation and fortification results
in Table Two suggest increases in Hb of 7-50 percent or more from
interventions.    Using an elasticity of Rb on work output of 1.5, these 'db
changes translate into potential increases in work output of between 30 and
75 percent.    The reader is reminded that the elasticity represents the
precentage change in work output associated with a onie percent change in
Hb.   Thus, an elasticity of 1.5 suggests that each one percent increase in
Rb is associated with a 1.5 percent increase in work output. Clearly    ch
increases in work output would require an increase in energy to suscain
over the long term (Viteri et al. 1971). This issue has been reflected
i-n s tudies that have attempted to ascertain the conditions under which the
cost of a higher caloric input for workers is justified by the valle oz the
higher agricultural output thac will be produced as welt as the



57
impplications of tie phenonmenon for rural labor markets (Immink and Viteri
1981 a & b; Mirralees 1976; Leibenstein 1958; Stiglitz 1976).
The obvious chalLenge is to ascertain what the additional caloric
needis would be under different assumptions about gains in Hb and increases
itn work exertion.    Although- studies exist on how eiergy requirements
increase among work activities of different intensities, our data on work
output refer to higher Levels of output for the same work activityI  Energy
needs depend also upon the temperature of the work environment and weight
oE the individual worker.       For thiese reasons, generalization on the
re lat ion be cween higher work output and higher energy needs for the many
di f ferent s itua t ions and populations in LDC' s is problematic.  Bogert,
Br iggs, and Ca  loway ( 19 73: 44): provide data from a Canadian study that
evaluated the energy requiremnents of different occupations among men and
women and categorized them according to the intensity of activity and by
weight of subjects. The requiremnents in kcal per day for persons in the
50th percentile according to weight, for men and women, were 2300 and 1900
respectively for sedentary activity; 2850 and 2400 for light activity; 3650
and 3000 for moderate activity; and 4250 and 3550 for heavy manual work or
athletic training.     A shift fromn sedentary activity to heavy manual work
antailed an increase in daLly requirements of 1950 calories for men and
1650 for women.
However, the average Canadian man at 72 kg and Canadian woman at 56
ki lograins were considerably heavier than the average member of the ac-risk
populations in developing societies. While we hnave no overall weight "actor
for the latter groups, 60 kg; would seem to be a more reasonable weight for
-nen, a Level about 17 percent less than the Canadian average.       Tn the



58
Canadian data, a reduction in weight of 17 percent was associated with a
reduction in the need for additional calories of about 14 percent.
Accordingl y, we might expect the increase in calories for workers in LDC' s
wwho shift fron sedentary to heavy manual rIorK to increase by about 1710
calories for -nales and bv about 1450 for womnen. For a workforce that is
about two-thirds maLe, the average increase would be about 1600.
Finally, this increase must be related to percentage increases in work
output to be cons is tent wich our measures of the effects of increases in
Hb.   Accordingly, we will assume that a shift from sedentary work to heavy
-manual work is equivalent co a 100 percent increase in work output in our
data.   This means that for every 10 percent increase in work output, there
will be an additional daily caloric requirenent per worker of 160 calories.
En order to provide an approximace cost for additional caloric intake,
we will use the calorie content of a major food staple and its price.
According to Wa tt and Merrill (1963:52), uncooked rice and corn meal both
have a caloric content of about 365 calories/100 g. The unsubsidized price
(reflecting    the full cost) of rice in Indonesia in 1980 was about U.S.
$0.34 per kg (Mears 1981:551) or a daily cost of about U.S. $0.0093 daily
for an amount that would be expected to provide abouc 100 more calories.
Corn meal was in the same range in Mexico. On the basis of 200 days of work
a year, the additional cost of LOO calories would be about $1.86 a year,
and for 300 days it would be $ 2.79. This means that for a daily increase
of 400 calories--the amount associated with a 25 percent increase in work
output--the anaual cost would be almost $7.50, and for 800 additional
calories, almost $15.00 for 200 days of work.



59
However, we should bear in mind that unlike the estimated costs of
fortification and supplementation, these costs for additional energy
requiremnents are per worker costs, rather than per capita ones. They mnust
be divided amona the entire population to make them consistent with the
other cost estctnates. Further, we wilL also show tiiat soLte of the
additional energy intake required for higher work output of a given worker
wilL be offset by lower requirements for workers who are displaced in a
lab-or surplus -situation. Both of these adjustments will be discussed and
imp Lemen ted in Section VII. En the next section we wilL make estimates of
benefits of the interventions, and in the final section we will combine
them with costs to obtain benefit-cost results as well as addressing their
policy consequences.
VI CALCULATING BENEFITS
Section III provided a survey of potential benefits of anemia
reduction.    Specific attention was devoted to the benefits of increased
work output because the evidence on this dimension was substantial and
consistent, its value can be estimated in labor markets, and it is clearly
a major social benefit of reducing anemia.         In ti;is section, we wilL
calculate the value of the benefits. While most of the attention will be
devoted to the benefits of additional work output, we will also estimate
the value of the other benefits of anemia reduction.
Be fore making these calculations, it is important to point out the
labor market context Eor whichi benefits will be estimated. To a larae
de,ree labor markets in developing societies are composed predominantly of
a gricul tural work ers.  For example, in 1979 about 71 percent of the labor



60
force was engaged in agricultural activities in low-income countries as
de f ined by the Woorld Bank, and. 43 percent of the labor force was in chis
sector in the middle-incomne countries (World Bank 1981:170-171). In the
A.
Low income countries, about equal proportions--14-15 percent--were engaged
Ln m.anufacturing aad services.      .n t he middle-income countries, the
proporcion in manufacturing was about 23 percent and in services was about
34 percent.
These countries were generalLy characterized by large labor surpluses
and high rates of rinemployment.      They also had many localized labor
markets, each characterized by different wages according to region of the
country, urban-rural distinctions, season, and industry composition. Much
of agricultural, handicraft, and hzousing output is produced for home
consumption, and many markets do not approach the perfectly compecitive
model as large agricultural estates or factories dominate particular labor
markets.
It is important to consider the implications of these labor mar'kets
for estimating benefits of higher work output. Raising iron status of
workers was shown to be associated with higher work capacity and output.
Although the relation is pertinent to even sub-maximal human activity, it
is particularly pertinent to activities that make heavy and continuous
physical demands on workers such as those in agriculture, construction,
road-building, and many of the other activities of the primary labor Lorce
in developing societies.
WhiLe these findings might enable us to predict the increase in work
output associated with -an increase in Hb, the placing of a valuation on
additionaL work output is more problematic. For example, Droductivities and



61
wages di f fer substantially both within and among societies, so there is no
possibility or a single estimate for valuing such added work. output. In
addition, the ex;istence of surplus labor and high levels of unemployment
wi LL mean that some of the additional rwork output of the employed Labor
force will be translated     into a need for fewer      aorkers and greater
unemp Loyment.
Keeping these factors in mind, we can set out a number of steps for
es timating benefits of anemia reduction.   First, we need to stipulate the
probable effect of specific types of interventions on Hb and their likely
impact on work output. Second, we need to estimate the pecuniary value of
the additional work output. This will require ascertaining an appropriate
measure for assessing the value of additional output for any particular
t ime unit (e.g. hour or day) and multiplying that by the number of hours or
days of work per year to obtain an annual estimate. Third, the value of
additional work output per person will have to be adjusted for the number
of persons who are not in the productive labor force but are benefitting
from the intervention to obtain a per-capita benefit. This will provide an
overall estimnate of the benefit per capita of the additional work output
associated with a particular fortification or supplementation approach.
Finally, an estimate must be made of the value of non-labor market benefits
such as higher Levels of home production, Lower morbidity, improved
physical stature and learning, and reduced mortality (particularly infant
and maternal) which when added to the value of additional work output will
provide a total benefit per capita.
-ffects of Interventions on Hb
Based upon the various interventions described in Table Two and in the
text, we wish to estimate the probable range of effects of interventions on



62
Hb.   Effects range from about 5-30 percent for the different sites of the
Ind ian salt s tudy, based upon my reanalyses of the appropriate appendix
t ab les    Food and Nutrition Board and UNICEF (1981). However, the higher
figure for the Calcutta site seems so far above the other estimates that we
wi11 assume an upper value of 20 percent and a most likeLy value of 10
percent.   We should bear in mind that the expected incremental change in Rb
refers to the difference in Rb concentration expected in the absence of
fortification versus that with fortification.
With respect to iron supplementation studies, most of the impacts
on Hb are in the range of 10-50 percent although several are higher and at
least one approaches 100 percent.   It shiould be noted that supplementation
Ls a short-run strategy to re plenish iron stores among severely anemic
persons.   Low initial Rb in these populations and high iron intake are
iecessarily associated with a larger 1H response than in the less anemic
populations receiving iron fortification.  We will use 10 and 50 percent as
the high and low values respectively for estimating the incremental change
in Hb associated with iron supplements and 25 percent as the most likely
va lue.  We do not have evidence from field trials on the probable effects
of ascorbic acid interventions.
Effects of Chan es in Hb on Work OutDut
Based upon the results presented in Table Three and the text, we will
assume that the elasticity of Hb changes on workc output will be between L
and 2 with the most likelyr value being 1.5. That means that for every
increase of 1 percent in Rb, there will be an expected increase in work
output of between 1 and 2 percent with the most likely increase being 1.5
percent. Wnen these elasticities are applied to the exoected changes i"n Hb
resulcing fromn the interventions, we obtain the estimated impact of the



63
TABLE FIVE
ESTIMATED IMPACT OF IRON INTERVENTIONS ON WORK OUTPUT
HIb/Work Outpuc Elasticity
1                1.5                  2
Fortifica tion:
, Hb =   5%                 5%               7.5%                10%
AHb = 10%                  10%              15%                  20%
Hb = 20%                20%              30%                  40%
Supplementation:
. Hb = 10                  10%              15%                  20%
H Hb = 25'                 25%              37.5%                50%
Hb = 50%                  50%              75%                 100%



64
interventions on work Qutput as shown in Table Six. This table shows the
estimated increase in work output for arduous, physical occupations (e.g.
Labor intensive agriculture, road-building, construction) for the different
assumptions regarding both Rib changes and the elasticities of 'i-b on work
out put. .Many of the assumptions suggest dramatic effects.   For exaimple,
even an increme.nt in ib of 20 percent and an elasticity on work output of
1.5 would suggest more than a 30 percent increase in work output.
Pecuniary Value of Work Output
The value of rwork output in different societies and in different parts
of the same sociecy wilL differ enorrmously.  Factors determining the value
include the organization of work, capital intensity, technology, and
economic structure, and overall level of economic development as it affects
the value of labor, goods, and services.  In a perfectly competitive market
economy that meets all of the textbook assumptions of large numbers of
buyers and selLers, perfect factor mobility, flexibLe prices and wages,
perfect information, and full employment of all resources, the equilibrium
wage is assumed to approximate the value of worker productivity at the
margin.   However, labor markets in developing societies do not appear to
approach this standard.      Especially in agricultural production, much of
out put is produced on traditional family or community farms with very small
holdings and no hired labor. 'dired labor is found on the Large plantations
and estaces, with labor markets dominated by a single employer or at the
mnost a few employers.     Traditional attachments to villages and poor
transportation reduce labor mobility, aLid poor access to capital markets
limits capical mobility.      Wich a rapidly growing population and Labor



65
supply, there is a labor surplus even at subsistence wages or wages
slightly above subsisteace levels.
Now consider the social value of additional output.        For the
self-contained traditionaL Earm, additional output can increase home
c.)nsulnptioa and raise the standard oL living.      But, if there is
u-nderemnp Loymen t among f ami ly members comprising the work force of the
traditional farm, the higher capability to produce output may not be fully
reali-zed.  This is especially likely to be so during the seasonal lulls in
agricultural activity created by weather or crop cycles. Of course, during
periods of high employment (e.g. harvesting), the additional capabilities
of workers can be put to full use.      The same is true of hired labor.
Duriang periods of underemployment and unemployment, the higher output of
less anemic workers may only serve to displace other workers who would have
had more employment.     However, during periods of peak labor demand, the
higher work output attributible to anemnia reduction should translate into
higher social output. At one extreme, when there is a slack demand for
labor and high unemployment or underemployment, increases in the
productivity of one group of employed workers will just displace other
workers who would have been employed, resulting in no net increase in
social output.
At the other extreme, during periods of peak labor demand, the
additional productivity of any worker will also increase total social
output. Thus, the social value of additional output of a particular worker
over a year is likely to be greater than zero, buc less than his or tier
average productivitv over the year (Gittinger 1982: 258-63; McDiarmid
l)77) .



66
Any es timate o f the value of increased output associated with higher
work capacity must take these employment effects into account.  One method
of es timating the additional output of a worker whose iron status has been
improved is to begin by estimating 'Lhis or her productivity as a reflecting
of earniings. The gross value of the additional output due to greater work
capacity can be estimated by applying the appropriate percentaces in Table
Five to expected earaings for that type of labor. But, this method does
not take account of the displacement of other workers in a labor surplus
situation, when worker productivity is increased.   In order to 'adjusC for
chis effect, economists attempt to assess the social opportunity cost of
the `mar,ginal" wor'ker by asking how much of his/her output r.presents a
gain in ou.tput for, the society.    That is, if such a worker were removed
from production, to what degree would the employment of another--unemployed
or unde-employed--worker compensate for the loss of output of the first
worker.
Extensive analysis of the Indonesian labor market has suggested that
the marginal opportunity cost of rural agricultural labor is about
one-third of the going agricultural wage for the year as a whole (The World
Bank 1983: 127-8).     However, in the LDC's, hired agricultural Labor is
typically emploved on large estates by a single employer or a few large
employers in any local Labor market. Given the monopsonistic nature of
such Labor markets, it is likeLy that wages are below the competitive
equilibrium of the classical labor market in which waves are assumed to be
equal to the productivity of the marginal worker.          Accordingly,
agricultural earniags as a mneasure of social productivity have both an
upward bias--in noC taking account oE social opportunity costs of



67
labor-- and    a downw ard bias because they are often determined
monopsonistically.     In this report we will assume that the true social
benefit of additional output will be equaL 'o half oL the increase in the
annual earnings associated with greater productivity. This value is based
upon two assumptions: the marginal value product. is 50 percent higher than
the monopsony wage and the social opportunity cost of the marginal worker
is one-third of his or her individual productivity.
A study by i-illian Hart (cited in The World Bank 1983: 126) of a
Javanese village in 1975-76 found that adult men had annual earnings of
about $161 and women $58. Assuming eight hour days, the number of days
worked per year was 234 for the inen and 150 for the women. Given
approximately a 200 day a year average and a weighted wage (Rp 30.7), we
can estimate annual earnings at about $118.         Using the 50 percent
adjust:nent factor to obtain the social value of output, this would
translate into $59.     Adjusting it on a per-capita basis requires taking
into account non-earners as well as earners. Although children do perform
some work on both family holdings and in the general labor market, we will
assume that only persons between 15 and 64 comprise the productive
population.    According to The World Bank (1981: 170-1), the population
15-64 years of age constitutes about 55 percent of the total population in
low and middle-income countries. Therefore, the per-capita value of social
output at thle margin is about $32.50 a year in this example.
The estcimated impact of iron intorventions on work output in Table
Five can then be combined with these estimates in a straightforward way to
obtain the vaLue of increases in work output associated with anemia
reduccion.    For examnple, che intermediate value for the change in Hb



68
associated with interventions is 10 percent. Assuming the intermediate
e Las t ic i ty of 1.5, this suggests an increase in work output o f 15 percent.
Multiplying this percentage times $32.50 yields ani increased output per
capita o f about $4. 88.    For supplementationi we can take the internediace
-a lue oF 25 percent For change in H1b -and use an eLasticizy of 1.5,
suggesting an increase in work output of 37.5 percent. Multiplying this
percentage t imes S 32.50 yields an increased output per capita of about
$12. 19.
In the comparisons of benefits with costs, we wilL use this approach
to estimate the beaefits from additional work output. In addition to the
Iadones ian exampLe, we will use the agricultural wage in Kenya and Mexico
as fur ther i l lustrations.  In each case we will assume 200 days a year of
work and a marginal social benefit of work equal to half of the additional
output.   Wie w ill also use an estimate of 55 percent for the working
populat ion.   It should be noted that these techniques will tend to provide
a conservative estimate of the value of additional work output from anemia
reduction by using the relatively low agricultural wage as a criterion and
by assuming that only the population 1.5-64 are doing productive work. This
means that the higher productivity of persons in higher level occupations
is not incLuded and that the work output of very young and very old workers
has not been included in the calculations.
Adjusting for Other Benefits
The methodology set out above was designed to capture the benefits of
itmmediate increases in work output associated with iron interventions. .As
such it igrnores the long term benefits associated with improved health,
vigor, and physical and mental growth of the population. In order to obtain



69
a total estimate of sociaL benefits from anemia red uction, we need to take
account o f these additional benefits. These long-term benefits include
lower morbidity and mortality,       areater physical s tature, higher
productivity outside of the workplace, improved quality of leisure time,
grea ter learning and faster school advancemeat, and improved reelings of
we ll-being. Most of these were discussed above, and each has some vzalue to
society.    For example, lower morbidity and mortality is associated with a
reducction in both human suffering as well as health care needs. Higher
productivity outside of the workpLace means a greater number and variety of
self-produced goods and services as well as an improved caDability of the
family to care for itself and its offspring. Greater learning and more
rapid school advancement impr jve the productivity of school resources
resuLting in lower costs per completion and level of achievement as well as
contributing to reducing scarcities of skiLled labor. Improved feelings of
well-being clearly have value.
But, as difficult as the estimnation of the benefits of short-term
increases in work output may appear, the other benefits are infinitely more
difficult to estimate. For examnple, lack of longitudinal data means that
es t imates of the effects of child nutrition programs on adult productivity
require an even larger number of crucial assumptions and speculation on
relationships than the short-term effects of anemia reduction estimated
above (Selowsky 1981).   Accordingly, we ihave little direct guidance on the
subject.
However, we might s peculate on the pertinence of some parallel
es t imnates of these types of benefits for education. The usual approach to
est;imating che economic value of additional education is to ascertain the



70
additionaL Labor market earnings associated with an additional year of
schooling. In an extremely comprehiensive article, Haveman and Wolfe (1984)
have estimated the magnitude of non-market effects of education on economic
wel l-being. These include improving child quality through home activities,
improvements in lie3lth, iaprovements in labor mark-et searclh and somne 16
other categories of benefits not tied to market productivity and wages.  On
the basis of their analysis and computations they conclude that the
standard estimates of labor market returns to an additional year of
schooling "....may capture only about one half of the total value (p. 401)."
If this f inding were also true with regard to the effects of programs
that provide a major improvement in health, we might expect that the
overall marginal social benefit of anemia reduction would be twice the
value of the additional work output alone.         On the basis of this
assumption, we wilL set the upper Limit of the adjustment at 100 percent,
the lower limit at 25 percent, and the intermediate value at 50 percent.
This suggests that the additional benefits not measured in the work output
sec tion will be assumed to be a mininum of 25 percent and a maximum of 100
percent while we will assume that the intermediate value is the best
estimate (in the absence of more direct informatioa). Applying the 50
percent value would raise the social benefits of the fortification example
set out above fromn $4.88 to ~7.32 and for supplementation from $12aL9 to
$18.29.
'VIL CALCULATING 3ENEFlT-COST RATIOS
Finally, we hlave reached the stage where we can calculate benefit-cost
rat ios for the interventions.      For Indonesia we                   -nave set o'lt an



71
i Llus trac ion of benefits for 1975-76 data.   For Mexico and Kenya we use
wage data from the International Labor Office (L98a3: Table 21).   For 1980
the es c imated w ages of agricultural wor'kers at the exist ing exchange rate
was $ 1150 for Mexico and $689 for 'Kenya on the assumption of a 200 day work
ye ar.
Summary of Benefits
Tab les Six, Seveni and Eight present the estimated per capita benefits
o f the anemia interventions for Indonesia, Kenya, and lMexico respectively.
Each tabLe begins with the annual earnings per. agricultural worker and
divides it by half to obtain the estimated social benefit of an additional
worker for 200 days a year. This amount is multiplied by .55 to obtain a
per capita estimate on the basis that only 55 percent of the populacion is
working.   The per capita social benefit is then applied to the estimates of
gains in work output in Table Five, after adjusting it upward by fifty
percent to take account of other benefits of anemia reduction.     Separate
estimates are made for fortification and supplementation. Thus, under
di fferent assumptions about intervention-induced changes in Rb and the
e ffects of changes in Rb on work output, the estimated social benefits, per
capita, are presented.
In order to understand the construction of the tables, it is tuseful to
review the variouis s teps.    Annual earnings per agricultural worker are
based upon the available sources set out above. In the case of Kenya and
'lexico, they r efer to national estimaates for 1980 reported by the
ln tarnational Labor Office (1983).  In che case of Indonesia, the figure is



72
TABLE SIX
ESTIMATED BENEFITS PER CAPITA OF ANEMIA INTERVENTIONS IN INDONESIA
Annual Earnings per Agricultural Worker                    $118
Social Benefit (divide bv .5)                               59
Per Capita (multiply bv .55)                                 32.5
Benefits per capita for different changes in Hb and Work Elasticities (per
capita social benefit o f additional work output with 50 percent upward
adjuscment for other benefits)
Work Outout Elasticity:
1              1.5                2.0
Fort i fication
i Hb -5                   $ S2.45         $3.66             $4.88
ib = 10                 4.88            7.32              9.75
Hb = 20%                 9.75           14.63             19.50
Supplemencary:
Hb = 10%                 $4.88            $7.32             $9.75
l.Hb = 25%               12.20            18.29             24.3.8
ARb = 50%                 24.38           36.57              48.75



73
TABLE SEVEN
ESTIMATED BENEFITS PER CAPITA OF ANEMIA INTERVENTIONS IN KENYA
Annual Earnings per Agricultural Worker              $689
Social 3enefit (multiply by .5)                       344.5
Per Capita (multiply by .55)                          189.5
Benefits per capita for different changes in Hb and Work Elasticities (per
capita social beaefit of additional work output with 50 percent adjustment
for other benefits)
Work Output Elasticity
1              1.5               2.0
Fortification:
'ib - 5%                  $14.21          $21.32            $28.42
\ Hb = 10%                 28.42           42.64            56.84
Hlb = 20%                  56.84           85.26           113.70
Supplementation:
L\Hb = 10%                $28.42          $42.63            $56.84
QHb = 25%                  71.06          106.59            142.12
A Hb = 50%                142.12          213.18            284.24



74
TABLE EIGHT
ESTIMATED BENEFITS PER CAPITA OF ANEMIA INTERVENTIONS IN 'MEXICO
Annual Earnings per Agricultural Worker              $1,150
Social Benefit (multiply by .5)                         575
?er Person (multiply by .55)                            316
Benefits per capica for different changes in Hb and Work 'lasticities (per
capita social benefit of additional work output with 50 percent upward
adjusrtment for other benefits)
Work Oucput Elasticitvi
1              1.5             2.0
Forti E i.cation
, Hb =   5%               $23.70          $35.55          $47.40
Hb = 10%                  47.40           71.10           94.80
Hb = 20.                   94.80          142.20          139.60
SuDolementation
L Hb = IO                $47.40          $71.10          $94.80
'db = 25;                118.50          177.75          237.00
L Hb = 50%                237.00          355.50         474.00



75
TABLE NINE
ESTIMIATED COSTS OF ANEMIA INTERVE£NTIONS PER CAPITA
A. Costs of Supplements and Delivery Per Capita
Low              Medium          High
Fortification (per capita cost)
Iron                       $0.10           $0.20            $0.30
Ascorbic Acid                                0.60
Supplementation (per capita cost)
Ferrous Sulphate           $1.89           $2.56            $3.17
Ascorbic Acid               3.79             4.46            5.07
B. Social Costs of Additional Energy Requirements Per Caoita
Work Output Elasticitv
1               1.5             2.0
Fortifica tion
R Hb =   5%                 $0.28            $0.41           $0.55
4ib = 10%                 0.55             0.82            1.10
A Rb = 20%                   1.10            1.64             2.20
Supplementation
A Hb = 10%                  $0.55           $0.83            $1.10
LHb = 25%                    1.38            2.06             2.75
Aab = 50%                   2.75             4.13            5.50



76
derived frota a regional s tudy for 1975-76 by Gillian Hart (as cited in
World Bank, 1983: 126).     The adjustmeents to obtain the social value of a
maarginal worker and the per capita values were discussed previously. These
are ad justed upward by 50 percent to incorporate the other benefits of
anemia reduction.     The total social benefits per capita of a marginal
worker are es cimnated to be $48. 75 in Indonesia, $284.25 in Kenya, and
$474.00 in Mexico.
.n order to convert these values into benefits from the intervencions,
we must raultiply them by the expected increase in work output per
agricul tural worker.    Table Five shows the percentage increase in work
output associated with different changes in Hb and different elasticities
relating changes in Hb to changes in work output.    These percentages are
multiplied By the per capita value of total social benefits for an
a gricultural worker to obtain an estimate of the social benefit per capita
of the interventions.      For example, if fortification increases Hb by 10
percent and a work elasticity of 1.5 is assumed, the expected increase in
work output from fortification would be 15 percent according to Table Five.
When this Lncrease in work output is multiplied times the appropriat2
figutres for t'otal so&c I benefits per capita of an additional, agricultural
wor'Ker, che results are an annual expected benefit from the intervention of
$7.32 in Indonesia, S42.64 in Kenya, and $71.10 in Mexico.
Thus, the bottom secciaons of Tables Six, Sevea, and Eight show the
expected social benef.ts attributible to fortification and supplementation
under different assumptions about the effects of the interventions on Hb
and the erfects of increases in Hb on work output.
'hese can be compared directly with the costs of the interventions.



77
Surmnary of Costs
Table 1 ine provides a summary of estimnates of the costs of the anemia
interventions.    The costs of fortification and supplementation are
presented as they were computed in Section V. In the second part of Table
81 ne we have presented the estinated costs o F the additional energy
requi rement. Using the caloric value of rice and the unsubsidized price of
r ice in Indonesia.as well as. estimnates of the additional energy
r2quirrements associated with greater work activity, it was estimated that a
100 percent increase in work output would require almost 1600 additional
zalories a day at an annual cost of almost $30.00 per worker.
But, in Section VI we assumed that only one-third of the additional
output per worker represented a net addition to social output. Presumably.
a h igher level of output for any particular worker will partially di-splace
emp loyment of other workers who could have produced that additional output
in a labor surplus situatiion.      This means that although the energy
requirement for a more productive worker will rise, it will fall for
workers whose work ef forts are displaced by the higher work output of
others.   To be consistent with the assumption that only one-third of the
output o f the marginal agricul tural worker is a cotntribution to social
output (because it is assumed that two-thirds of the output simply
displaces that of other capable workers whose unemployment rises), we umust
also assume that two-thirds of the rise in enecgy input of a more
productive worker is offset by a decline in energy requirements of
displaced workers.     This means that the social costs of additional energy
requirements for higher individual work output will be only about one-third
of the increase in cost for individual workers whose work output increases.



78
In summary, the social costs of additional energy requirements are
based upon taking one-third of the estimated cost of the additional energy
requirements associated with an increase in work output. This provides an
estimate of the social cost per worker of additional energy iaputs.
However, since we have adjusted all benefits to a per-capita measure, chis
cost must also be adjusted to a per-capita reasure by multiplying by .55 as
discussed in seccion VI. Thus, the annual social cost, per capita, of
additional energy intake for a fortification intervention that increases Hb
by 10 percent with   a work elasticity of 1.5 is estimated to be $0.82.  For
a supplementation intervention that increases Rb by 25 percent with a work
elasticity of 1.5, the annual social cost per capita of the additional
erneirgy requirement is $2.06.
Benefit-Cost Ratios
Table Ten provides a summary of benefits and costs of fortification
for Indonesia, Kenya, and Mexico. The costs are taken from Table NIine and
are divided between the costs of fortification and those pertaining to
additional energy requirements under different assumptions regarding change
in Hb and work output elasticities.      These assumptions pertain to three
different conditions in which "low" refers to the lowest work elasticity
and Hb, "medium" refers to an intermediate value for each, and "high"
refers to the upper limit for each in this study. The costs per capita of
fortification and additional energy input are considered to be similar for
the three countries. Benefits are taken from Tables Six, Seven, and Eight
for the appropriate contingencies regarding change in Hlb and work ourput
elasticities.



79.
TABLE TEN
PER CAPITA BENEFITS AND COSTS OF FORTIFICATION
Indonesia       Kenva            Mexico
A. Costs (,medium)
F ortification
Iron                                $0.20            $0.20            $0.20
Ascorbic Acid                         0.60            0.60             0.60
Additional Energy Intake
Low  ( Hb     5%; Elasticity = 1)   $0.28            $0.28            $0.28
Mled. ( Rb   10%; Elasticity = 1.5)   0.82            0.82             0.82
High ( Rb    20%; Elasticity = 2)     2.20            2.20             2.20
B.  Benefits
Low                                 $2.45            $14.21'          $23.70
Med.                                  7.32            42.64            71.10
High                                 19.50           113.70           189.60
C. Benefit-Cost Ratio Iron Fortification (Including energy requirements)
Low                                   5               30               49
Med.                                  7               42               70
High                                  8               47               79



so
TABLE ELEVEN
PER CAPITA BENEFITS AND COSTS OF SUPPLEMENTATION
Indonesia       Kenya           Me:xico
A. Costs (medium)
Ferrous Sulphace (alone)                           S1.10         $1.10           $1.10
Ascorbic Acid (alone)                               3.00          3.00            3.00
Delivery Systems with Ferrous Sulphate              2.56          2.56            2.56
Delivery System with Ascorbic Acid                  4.46          4.46            4.46
Addicional Caloric Incake
Low  (AHb = 10%; Elasticity    1)                0.55          0.55            0.55
Med. t,Hb - 25%; Elasticity     1.5)             2.06          2,06            2.06
-iig.h (,SLb  50%; Elasticity = 2)               5.50          5.50            5.50
B. Benefics
Low                                             $4.88        $23.42           $47.40
MIed.                                           18.29        106.59           177.75
High                                            48.75        284.24           474.00
C. Benefic-Cost Ratios (including energy requirements)
Ferrous Sulphate (alone)
Low                                              3            17               29
Med.                                             6            34               56
High                                             7            43               72
Delivery System wich Ferrous Sulphate
Low                                              1.6           9
Med.                                             4            23
High                                             6            13               59



81
For all three countries and under alL assumptions, the.benefits of
iron a  Dortification exceed the costs by a wide margin. The medium values
represent our "bes t" es timate of the appropriate benefit-cost ratios.
These range fcom 7 in Indonesia and 42. in Kenya to 70 in Mexico.      Even
assuming a rise in Rb of only 5 percent and a rwork elasticity of 1 for {he
country with the lowest agricultural earnings, Iadonesia, the benefit-cost
ratiio i s 5.  Although we have not shown estimates for ascorbic acid, they
can be calculated readily fromn the information provided and are also high.
Under no set of conditions in the table would they be less than about 3,
and under the medium set of conditions they would range from 5 for
Indonest La to 30 for Kenya to 50 for Mexico. Benefit-cost ratios remain
s trong even when the "high" estimate of $0.30 for fortification from Table
Nie;.f .; used. it should a lso be noted that even the $ 0.20 cost of
Eortification used for the benefit-cost estimates is co.nsiderably above the
SO.07-0. 10 cost reported in fortification studies.   Of course, they would
be considerably higher if we assumed a 35 percent increase in Rb as in the
Calcutta example of the Indian salt study.
Tab Le E;leven presents benefits and costs for supplementation.  Costs
are presented separately for the medicinal supplements, the delivery system
wi th the medicinial supplements, and additional energy requirements for the
low, medium, and high assumptions on work output. Again, the "low"
assurmptions represent the lower limit expected for work output elasticity
and change in Rb; the "high" assumptions represent the upper limit; and the
"medium" assumptions represent the "best" estimates. The benefit-cost
ratios are presented for interventions for the medicinal supplements



82
a lone--in this case ferrous sulphate-as well as with the delivery system.
They also include the costs of the increased energy requiremwnens.
The cost of ferrous sulphate alone is relevant only when there already
exists a delivery system so that the marginal cost will be limited to the
cost of the supplement alone. E:xisting employer and community nutritional
or health programs might be pertinent to this assumption.     In all cases,
the benefit-cost ratios exceed I by a good margin, ranging from 6-58 for
the medium benefit category. When the costs of service delivery are added,
the benefit-cost ratios remain high with a range of 4-38 in the medium
range category.     The low benefit assumptions for Indonesia are the one
exception to high ratios with only a 1.6 figure. Even when the high cost
service delivery assumptions from Table Nine are used the ratios remain
s u b s t a n o i a I fo r a l mo s t a ll cas es excep t the Indones ian low bene fi t case
whiich falls to about 1.3. BuC, we should bear in mind that the Indonesian
ratLo is probably understated because of the high value assumed for the
community health worker and the fact that earnings for Indonesia are taken
from a 1975-76 study.     It is likely that agricultural wages were somewhat
higher in 1980 on the basis of productivity increases over that period
(WorLd Bank 1983: 47-51). With a cost for the health auxiliary of about
$600 a year as posited in Section V, the benefit-cost ratio would rise to
over 2.
One area in which costs tSight be understated is the cost of health
persoanel in the Mexican case. Although the high cost figure among three
Asian countries w;as selected and it does appear to be a high estimate for
both Indonesia and Kenya, it is li'kcely to be low among more developed
societies.   Accordingly, one test of the robustness of the results would be



83
to assume that the cost of the community health auxiLiary rises from $1,500
a year to $4,500 a year for Mexico, a rise of $3,000 or $3.00 a person for
the population base of 1000.   Even with this higher cost, raising the cost
of delivery of ferrous sulphate to $5.56 a year per person in Mexico, the
benefit-cost ratios would remaia suibstantial, between 8 and 43.
Under a wide variety of assumptions, the benefit-cost ratios of both
fortification and supplementation exceed unity by a considerable margin.
It should also be borne in mind that most of the assumptions regarding
benefits should impart a conservative bias to our estimates. For example,
we utiLize the relatively low earnings for agricultlural workers; we assume
that a modest portion of the population is economically active; we do not
adjust the elasticities for additional dally work output to take accourt of
lower absenteeism from higher Hb; a substantial downward adjustment is made
for unemployment and underemploymenit; and the value of the benefits not
calculated in short-term labor market effects is set at only an additional
50 percent rather than 100 percent as had been estimated for the case of
education. Further, the benefits estimates do not take account of
fortification and supplementatiort trials in which the results exceeded
substantially the rises in Hb used in this study.     Many of the costs are
also stipulated on the high side, and the costs of additional energy
requirements are included.     Accordingly, we conclude that under a fairly
reasonable set of circumstances, interventions to reduce anemia represent
highly productive investments in LDC's.
Policy Implications
According to our estimates, it appears that both dietary fortification
and supplementation can be highly desirable investments for reducing anemnia



84
in LDC' s. Typical benefit-cost ratios range from 7 to 70 for fortification
and 4 to 38 for supplementation for the three countries. Even when a
variety of higher-cost or lower-benefit assumptions are stipulatad,
benefit-cost ratios remain substantial. Other analysts should feel free to
evaluate the methodology and to use different assumptions and data sets to
see if the ratios are appreciably modified when changes are made to conform
with oCher situations.
It is important to emphasize that fortification and supplementation
should be viewed more as complements than substitutes. Supplementation is a
short-term strategy to raise the iron status of a population to normal
Levels.   Fortification is a long-term strategy to maintain adequate iron
s tatus. 'Liot only are the benefit-cost ratios higher for fortification than
for supplementation, but they rest on fever assumptions about behavior
since they require no special actions on the part of the population.      In
contrast, supplementation requires thac the population cake the complete
regimen of medicinal iron or ascorbic acid as prescribed or that it is.
provided through daily meals on plantations, in factories, and in schools.
To the degree that there is variance in this behavior, it will also affect
benef it-cost ratios.   It is important to note that the benefit-cost ratios
of supplementation are so high that even sub-optimal consumption of the
suppLements is likely to be associated with high returns to investment.
There are a number of issues that ought to be emphasized in the
interpretation    and use of these findings.       First, there may be other
alternatives for reducing anemia thac also ought to be considered. We did
not focus on interventions for reducing anemia associated with blood loss
:rom parasites such as hookworm and schistosomiasis.       Stephenson ec al.



(1983) have shown that at least for children the cost of controLling
roundworms was modest in a four year project in Kenya.  To the degree that
parasites are an important cause of anemia, anti-helminch projects should
be evaluated   for their relatLve cost-effectiveness in improving Rb and
other outcomes.
Second, we did not consider the fact that some members of society may
be hypersusceptible to iron toxicity, so that some screening may have to be
considered (omenn 1982). This is also a consideration in calculating the
variance in consumption of fortified foods among the population. Third is
the issue of f inancing interventions for anemia reduction. Although the
benefit-cost ratios may be high for society and even for individuals, there
may be a reluctance by the populations benefitting from the interventions
to pay for them from incomes that are close to the margin for survival. A
societal mechanism should probably be established that will provide a sound
financial basis for the interventions.
FinalLy, the estimation of benefits and costs in this report assumes
that it is important to consider basing public policy decisions on
available data, even when the data are incomplete. This evaluation
suggests great promise overall in the productivity of investments in anemia
reduction.    However, in any particular situation it is still highly
desirable to carry out field trials which wilL establish more precise
es timates on costs and benefits (WHO 1975) for each situation in which it
is being considered.



86
BIBLIOGRAPHY
Andersen, H. T. and R. Barkve.       "Iron DefEiciency and Muscular Work
Performance," in The Scandanavian Journal of Clinical and Laboraborv
Investigacion, Vol. 25, Suppl. 114, 1970.
Astrand P.   "Physiological Factors Limiting Work Performance and Capacity," in
rron De ficiency and Work Performance L. Hallberg and :4. S. Scrimi-shaw
(eds. )p, Proceedings o f workshop held at the Massachusetts Institute of
Technology, April 28-May 1, 1981.
Astrand, P. 0. and K. RodahL. Textbook of Work Physiology (New York:
McGraw-Hill Book Co. 1977).
Baker, S. J.    "Nutritional Anaemias, Pt. 2: Tropical Asia," in Clinics in
Haemacology, Vol. 10, (October 1981), Dp. 843-871.
Baker. S. H. and E. K. DeMaeyer.    "Nutritional Anemia: Its Understanding and
Control wich Special Reference to the Work of The World Health
Organization," The American Journal of Clinical Nutrition, Vol, 32
(February 1979), pp. 368-417.
Barnum, H. N.', D. Tarantola and I. Setiady.       "Cost Effectiveness of an
Immunization Programme in Indonesia," Bulletin of the WHO Vol. 58 (1980)
pp. 499-503.
Barrett, D. E.    "An Approach to the Conceptua lization and Assessment of
So c ia l-Emo t iona l Functioning in Studying Nutrition--Behavior
Re lationships. " American Journal of Clinical Nutrition, Vol. 35, pp.
12 22-1227.
Basta, Samir Sanad.    "Effect of Anemia," American Journal of Clinical
Nutrition, Vol 31 (February 1978), pp. 197-201.
Basta, Samir Sanad and A. Churchill.       "Iron Deficiency Anemia and the
Productivity of Adulc Mlales in Indonesia," Intarnational Bank for
Reconstruction and Development, Staff Working Paper No. 175 (April 1974).
Basta, Samir Sanad, Soekirman, Darwin Karyadi, and Nevin S. Scrimshaw. " Iron
De ficiency Anemia and the Productivity of Adult Males in Indonesia,"
American Journal of Clinical Nutrition, Vo l. 32 (April 1979), pp.
916-925.
Beard, Jo hn, Clement A. Finch and Bruce Mackler. "Deleterious Effects of Iron
De ficiency, " Progress in Clinical Biological Research, Vol. 77 (1981),
pp. 305-310.
Bender, A. E.    "Food Enrichment," in The ImDortance of Vitamins to Human
Realch, T. G. Taylor (ed.) (Baltimore: University Park Press, 1979),
Chnap. 17.



87
B3jorn-Rasmussen, E. Hallberg and L. Rossander.  "Absorption of 'Fortification'
Iron," in British Journal of Nutrition, VoL. 37 (1977).
Bogert, Jean L. , George M. Briggs, and Doris Howes Calloway.    "NuLtrition and
Physical Work Performance," in Nutrition and Physical Fitness, (1973)
ninth edition, (Philadelphia: W. B. Saunders Company), pp. 481-491.
Bothwell, Thomas H. and Robert W. Charlton. Iron Deficiency            in Women
(Washington, DC: INACG, 1981),
Brise, H. and L. Hallberg.      "Absorbability of Different Iron Compou-ds," in
Acta Med. Scand. Vol. 171 (supplement 376, 1962).
Burman, David.     "Iron Deficiency in infancy and Childhood," in Clinics in
Haemotology, Vol. 11, No. 2 (June 1982), pp. 339-351.
Callender, Sheila T.      "Treatment of Iron Deficiency," in Clinics in
Haematology, Vol. 22, NLo. 2 (June 1982).
Chandra, R. K. and P. M. Newberne.        Nutrition, Immunity and Infection:
Mechanisms of Interactions, (New York: Plenum Press 1977).
Charlton, Robert and Thomas H. Bothwell.         "De fiaition, Prevalence and
Prevention of Iron Deficiency," in Clinics in Haematology, Vol. 11, No.
2 (June 1982), pp. 327-338.
Charlton, Robert, et al.      "Anaemia, Iron Deficiency and Exercise: Extended
Studies in Human Subjects", in Clinical Science and Molecular Medicine,
Vol. 53 (1977), pp. 537-547.
Cook, J. D. and C     Finch.    "Assessing Iron Status of a Population," in
American Journal of Clinical Nutrition, Vol. 32 (October 1979), pp.
2115-2119.
Cook, J. D. and M. E. Reusser.        "Iron Fortification: An Update," in The
American Journal of CliALical. Nutrition, Vol. 38 (October 1983), pp.
648-65 9.
Creese, A. L. and R. Henderson.        "Cost-Benefit Analysis and Immunization
Programmes in Developing Countries, : in Bulletin of the WHO, Vol. 58
(1980), pp. 471-497.
Creese, A. L., N. Sriyabbaya, G. Casabal and D. G. Wiseso. "Cost-Effectiveness
Appraisal of Immunization Programmes,t in Bulletin of the WHO, Vol. 60
(1982) pp. 621-632.
Dallman, Peter.     "Biochemical and Hematologic Indices of Iron Deficiency," in
Iron Deficiency: Brain Biochemistry and Behavior, E. Pollitt and R. L
Leibel (eds.) (New York: Raven Press, 1962), pp. 63-78.



88
Dallman, Peter.  "Tissue Iron Deficiency: General Concepts," in L. Hallberg and
N. Scrimshaw (eds.) Iron Deficiencv and Work Performance, proceedings of
a workshop held at Massachusetts Institute of Technology, April 28-May 1,
1981.
Dallman, Peter R. and LMartti A. Simes.       Iron Deficiency in Infancy and
Childhood, ( Washington, D.C.: INACG, 1979).
Davies, C.T.M.     "Relationship of Maximum Aerobic Power Output to Productivity
and Absenteeism of East African Sugar Cane Workers," in British Journal
of Industrial Medicine, Vol. 30 (1973a), pp. 146-154.
Davies, C. T.M. "Physiological Responses to Exercise in East African Children,
II: The Ef fects of Shistosomiasis, Anaemia and Malnutrition," in Journal
of TroDical Pediatrics, Vol. 19, No. 2, (June 1973b), pp. 115-119.
Davies   C. T. K. , A.C. Chu kweumeka and J.P.M. Van Haaren. "Iron Deficiency
Anemia: Its Ef fect on Maximum Aerobic Power and Responses to Exercise in
African Males aged L7-40 Years," in Clinical Science, Vol. 44, No. 6
(June 1973), pp. 555-562.
Davies, C. T.Y. and J .P.M. Van Haaren. "Effect of Treatment of Physiological
Responses to Exercise in East African Industrial Workers with Iron
Deficiency Anaemia," in British Journal of Industrial Medicine, Vol. 30,
No. 4 (1973), pp. 335-340.
Deinard, Amos, et al.      "Iron Deficiency and Behavioral Deficits," in
Pediatrics, Vol. 68, No. 6 (December 1981).
DeMaeyer, E. M.    "The Role of WHO in the Control of Nutritional Anaemia," in
Nutrition in RealCh and Disease and International Development: Symoosia
From the XXII International Congress of Nutrition, pp. 363-370 (New York:
Alan R. Liss, Inc, 1981).
Derman, D., et al. "Iron Absorption from a Cereal-Based Meal Containing Cane
Sugar Fortified with Ascorbic Acid," in British Journal of Nutrition,
Vol. 38 (1977).
Djukanovic, V. and E. P. Mach (eds.) Alternative Approaches to Meetina Basic
Health Needs in Develooing Countries, joint UNICEF/WHO study, (Geneva:
WHO, 1975).
Edgerton, V. R. and Yoshi Ohira. "Human Studies on the Relationship Between
Iron Deficiency and Work Performance," in Iron DeficiencZ and Work
Performance, L. Hallbera and N. S. Scrimshaw (eds.), Proceedings of
workshop held at the MIassachusetts Institute of Technology, April 28-May
L, 1981.
Edgerton, V. R. et al.      "Effects of Iron Deficiency Anemia on Voluntary
.ALctivities in Rats and Humans," in Iron Deficiency: Brain Biochemistry
and Behavior, E. Pollitt and R. L. Leibel (eds.) (New York: Raven Press
1982), pp. 141-160.



89
Edgerton, V. R., et al.      "Iron-deficiency Anaemia and its Effects on Worker
Productivity and Activity Patterns," in British Medical Journal, Vol. 2
(1979) pp. 1546-1549.
Edgerton, V. R. , et al. "Elevation of Hemoglobin and Work Tolerance in
Iron-De ficient Subjcets," in         Journal of Nutritional Science
Vitaminology, Vol. 27 (1981), pp. 77-86.
Elwood, P.C. and Dafydd Hughes. "Clinical Trial of Iron Therapy on Psychomocor
Function in Anaemic Women," in British Medical Journal, Vol. 3, (1970),
pp. 254-255.
Fleming, Alan F.       "Iron Deficiency in the Tropics," in Clinics in
Haematology, Vol. 11, No. 2 (June 1982), pp. 365-388.
Fleming, Alan F.     "Iron-De ficiency in the Tropics," in. Scandanavian Journal
of Haematology (1977), supDiement 32, pp. 315-321.
Food and Nutrition Board and UNICEF, New Delhi, India. "The Use of Common Salt
Fortified with Iron: A Report on Co llaborative Studies," (August 15,
1981).
Garby, L. and S. Areekul.      "Iron Sup.plementation in Thai Fish Sauce," in
Annals of Tro2ical Medicine and Parasitology, Vol. 68, (1974), pp.
46 7-47 6.
Gardner, Gerald W., et al.       "Cardiorespiratory, Hematological and Physical
Perforraance Responses of Anemic Subjects to Iron Treatment," in American
Journal of Clinical Nutrition, No. 28 (September 1975), pp. 982-988.
Gardner, Gerald W. , et al. "Physical Work Capacity and Metabolic Stress in
Subjects with Iron Deficiency Anemia," in American Journal of Clinical
Nutrition, Vol. 30 (June 1977), pp. 910-917.
Gittinger, J. P.     Economic Analysis of Agricultural Projects, (Baltimore,
IMD.: Johns Hopkins University Press 1982).
Hab icht, J.P. "Delivery of Primary Care by Medical Auxiliaries: Techniques of
Use and Analysis of Benefits Achieved in Some Rural Villages in
Guatemala," in Medical Care Auxiliaries, Scientific Publication No. 278,
(Washington, DC: Pan American Health Organization, 1973), pp. 24-37.
Hallberg, L.    "Effect of Vitamin C on the Bioavailability of iron from Food,"
in J. N. Counsell and D. R. Hornig (eds.) Vicamin C (Englewood, N.J.:
Applied Science Publishers 1981), Chap. 3.
Hal Lberg, L. and N. Scrimshiaw (eds.).   Iron Deficiency and Work Performance,
proceedings of a w -, ..3hop held at the Massachusetts Institute of
Technology, Cambridie,   .issachusetts, April 29-May 1, 1981. Sponsored by
the U. N. University in collaboration with the International Nutritional
Anemia Consultative Group (INAC ).



90
Hraveman, Robert H. and Barbara L. Wolfe. "Schooling and Economic Well-Being:
The Role of Nonmarket Effects,: in The Journal of Human Resources, Vol.
19 (Summer 1984), pp. 377-407.
Heilbroner, Robert L. and Lester C.,Thurow. Understanding Micro-economics,
Tahird Edition. (Englewood Cliffs, INJ: Prentice-Hall Inc., 1975.
Henshall, J.D.     "Ascorbic Acid in Fruit Juices and Beverages," in J. N.
Counsell   and D. H. Horni    (ads. ) Vitamin C (Eng lwood, N.J.: AppLied
Science Publishers 1981), Chap. 9.
H etzel, BasiL S. (ed.)     Basic Health Care in Develooing Countries: An
Epidemological Perspective, (Oxford: Oxford University Press, 1978).
(Handbook sponsored by IEA and WHO)
Immink, Mk. and F. E. Viteri. "Energy Intake and Productivity of Guatemalan
Sugarcane Cutters: Part 1."    In Journal of DeveloDment Economics, Vol. 9
(198La), pp. 251-271.
Immink, M. and F. E. Viteri. "Energy Intake and Productivity of Guatemalan
Sugarcane Cutters: Part 2."    In Journal of Develoomene Economics, Vol. 9
(l981b)., pp. 273-287.
International Labour Office. 1983 Yearbook of Labour Statistics, 43rd Issue,
(Geneva: ILO) pp. 539-629.
INACG (International Nutritional Anemia Consultative Group). The Effects of
Cereals and Legumes on Iron Availability, (Washington, D.C.: The
Nutrition Foundation 1982).
INACG (International Nutritional Anemia Consultative Group). Guidelines for
the Eradication of Iron Deficiency Anemia, (Washington, D.C. 1977).
Jacobs, A. "Non-haematological Effects of Iron Deficiency," in Clinics in
Haematology, Vol. 11, No. 2 (June 1982).
Jamison, Dean.     "Child Malnutrition and School Retardation in China," World
Bank Discussion PaDer #81-27 (supp. paper IV to annex B of World Bank
Report #3391-CHA) (Washington, D.C.: World Bank 1981).
Karyadi, H. D.     "Evaluation of Nutritional Anaemia Intervention Among Anaemic
Female Workers on a Tea Plantation," in Iron Deficiency and Work
Performance, L. Hallberg and N. S. Scrimshaw (eds.) proceedings of
workshop held at the Massachusetts Institute of Technology, April 28-May
1, 1981.
Karyadi, Darwin and Samir Basta.       "Nutrition and Healch of Indonesian
Construction Workers: Endurance and Anemia," International Bank for
Reconstruction and Developoment, International Development Association
Staff Working Paper No. 152, (April 1973).
Latham, Michael C. and Lani S. Stephenson. "Nutrition, Health and Worker
Productivity," an executive summary for the World Bank (1981).



91
Lanzkowsky, Philip.      "Iron Deficiency in Adolescents," in Adolescent
Nutrition Myron Winick (ed.) (New York: John Wiley and Sons 1982).
Layrisse, M. , C. 'Hartinez-Torres, and C. M. Renzl. "Sugar as a Vehicle for
Iron Fortification: Further Studies," in American Journal of Clinical
Nutrition, Vol. 29 (March 1976), pp. 274-279.
Layrisse, M . and C. Martinez-Torres.          "Food Iron Absorption: Iron
Suppelmeatation of Food," in Progress in HemaatoLogy, Vol. 8 (1971), pp.
137-160.
Layr isse, M. , et al .   "Food Iron Absorption: A Comparison of Vegetable and
Animal Foods,: in Blood, Vol. 33 (1969), pp. 430-443.
Leibel, Rudolph L. "Behavioral and Biochemical Correlates of Iron Deficiency,"
in Journal of the American Dietetic Association , Vol. 7., No. 4
(October 1973), pp. 398-404.
Le ibe 1, Rudolph L., Daryl B. Greenfield, and Ernesto Pollit. "Iron Deficiency:
Behavior and Brain Biochemistry," in 4. Winick (ed.) Nutrition: Pre and
Postnatal DeveloDment (New York, Plenum 1979), pp. 383-439.
Leibel, Rudolph L., Ernesto P.ollitt, Insun Kim, and Fernando Viteri. "Studies
Regarding the Impact of Micronu-trient Status on Behavior in Man: Iron
Deficiency as a t{ode*l," in The American Journal of Clinical Nutrition,
No. 35 (May 1982), 1211-1221.
Leibenstein, H.     "Underemployoment in Backward Economnies: Some additional
Notes," in Journal of Political Economy, Vol. 66 (1958), pp. 256-258.
Levin, Henry M.     Cost-Effectiveness: A Primer (Beverly Hills, CA: Sage
Publications 1983).
Levin, Henry M.     "Cost-Effectiveness in Evaluation Research," in M. Guttentag
and E. Struening, Handbook of Evaluation Research, Vol. 2, (Beverly
Hills, CA: Sage Publications 1975), Chapter 5.
Lozo f f, Be tsy, Gary Brittenham, Fernando Viteri, Abraham Wolf, and J. Urrutia.
'"The Effects of Short-Term Oral Iron Therapy on Developmental Deficits in
Iron-Deficient Infants," Journal of Pediatrics, Vo1. 100       (1982a) pp.
35 1-35 7.
Lo zo f f, Be tsy, Gary Brittenaham, Fernando Viteri, and Juan Urrutia.
"Behavioral Abnormalities in Infants with Iron Deficiency Anemia,: in
Iron Deficiency, Brain Biochemistry and Behavior, E. Pollitt and R.
Leibel (eds.) (New York: Raven Press 1982b), pp. 183-194.
Lozo ff, Betsy, Gary Brittenham, Fernando Viteri, Abraham rWJolf  "Developmental
Deficits in Iron-Deficient infants: Effects of Age and Severity of Iron
Lack,: in Journal of Pediatrics Vol. 10, No. 6, (December 1982c).



92
MfacDougall, L.G. et al. "Immune Response in Iron-Deficient Children: Impaired
Cellular Defence Mechanisms with Altered Hormonal Components" in Journal
of Pediatrics, Vol. 86 (1975).
Mackler, Bruce and Clement Finch.    "Iron in Central Nervous System Oxidataive
Metabolism," in E. Pollitt and R. L. Leibel (eds.) Iron Deficiency: Brain
Biochemistry and Behavior (LNew York: Raven Press 1982), pp. 31-38.
Margolis, Harold S. , et al.     "Iron Deficiency in Children: The Relationship
Be tween Pretreatment Laboracory Tests and Subsequent Hemoglobin Response
to Iron Therapy, " in American Journal of Clinical Nutrition, Vol. 34
(Gctober 1981), pp. 2158-2168.
Masawe, Aaron Eben John.      "Nutrtional Anaemias Part 1: Tropical Africa," in
Clinics in Haematolo2y, Vol. 10, No. 3 (October 1981).
Massaro, Thomas F. "Effect of Iron Deficiency on Learning, in E. Pollitt and
R. L. Leibel (eds.)Iron Deficiency: Brain Biochemistry and Behavior (New
York: Raven Press), pp. 125-140.
McDiarmid, 0. J.     Unskilled Labor for Develooment: Its Economic Cost
(Baltimore, MD: Johns Hopkins University Press 1977).
Mears, Leon.    The New Rice Economy of Indonesia, (Gadjah Mada University
Press 1981).
Mirralees, J. A.     "A Pure Theory of Underdeveloped Economics," in L. G.
Reynolds (ed.) Agriculture in Development Theory, (New York: Wiley
1976).
Mlishan, E. J. Cost-Benefic Analysis, (New York: Praeger Publishers 1976).
Moock, Peter R. and Joanne Leslie. "Childhood Malnutrition and Schooling in
the Terai Region of Nepal," World Bank Report (1982).
Nutrition Reviews.     "The Relationship Between Infection and the Iron Status
of an Individual," Vol. 33, No. 4 (April 1975), pp. 103-105.
Ohira, Y. et: al.    "Work Capacity After Iron Treatment as a Function of
Hemoglobin and Iron Deficiency," in Journal of Nutritional Science
Vicaminology, Vol. 27 (1981), pp. 87-96.
Ohira, Y. et al.     "Work Capacity, Heart Rate and Blood Lactate Responses to
Iron Treatment," in British Journal of Haematology, Vol. 41 (1979), pp.
365-3 72.
Cmena, Gilbert S. "Predictive Identification of Hypersusceptible Individuals,"
in Journal of Occupational Medicine, Vol. 24 (May 1982).
Oski, Frank A.    "The Nonhematologic Manifestations of Iron Deficiency," in
American Journal o f Disabled Children, Vol. 133 (March 1979), pp.
315-322.



93
Osk i,  Frank A. and Alice S. Honig.       "The Effects of Therapy on the
Deve lopmental Scores of Iron-Deficient Infants," in Journal of
Pediatrics, Vol. 92 (1978), pp. 21-15
Oski, Frank A.     "Reply to the Editor", in Journal of Pediatrics (January
1976) .
Pan American Health Organization (PAHO). "Medical AuxiLiaries," proceedings of
a symposium heald during thw 12th rmeeting of PAHO Advisory Committee on
Medical Research Sciertific Publications #278 (Washington, D.C.:
PAHO/WHO 1973).
Pollitt, Ernesto and Rudolph L. Leibel (eds. )         Iron Deficiency: Brain
Biochemistry and Behavior (New York: Raven Press 1982).
Pollitt, Ernesto and Rudolph L. Leibel. "Iron Deficiency and Behavior," in
The Journal of Pediatrics, Vol. 88, No. 3 (March 1976), pp. 372-381.
Pollitt, Ernesto and Nita Lewis.   "Nutrition and Educational Achievement: Part
L:  Malnutrtion and Behavioral Test Indicators," in Food and Nutrition
Bulletin, Vol. 2, No. 3.
Po llitt, Ernesto, Fernando Viteri, Carmen Saco-Pollitt, and Rudolph L. Leibel.
"Behavioral Effects of Iron Deficiency Anemia in Children,: in Iron
Deficiency: Brain Biochemistry and Behavior, E. Pollitt and R. L. Leibel
(eds.) (New York: Raven Press 1982), pp. 195-228.
Popkin, Barry M..    "Nutrition and Labor Productivity," in Social Science and
Medicine, Vol. l2C (1978), pp. 117-125.
Popkin, Barry M. and M. Lim-Ybanez. "Nutrition and School Achievement,: in
Social Science and Medicine, Vol. 16 (1982), pp. 53-61.
Read, Merrill S.     "Malnutrition and Behavior" in Applied Research in Mental
Retardation, Vol. 3 (1982), pp. 279-291.
Read, Merrill S.     "Malnutrition and Human Performance," in Mlalnutrition,_
Behavior and Social Organization (New York: Academic Press 1977).
Read, Merrill S. "Anemia and Behavior," in Nutrition, Growth and DeveloDment,
Modern Problems in Pediatrics, Vol. 14 (Basel: Karger 1975) 189-202.
Read, Merrill S.     "Malnutrition, 'Hunger, and Behavior," in Journal of the
American Dietetic Association, Vol. 63, No. 4 (Occober 1973).
Report of the Working Group onFortification of Salt with T.ron. "Use of Common
Salt Fortified with Iron in the Control and Prevention of Anemia--A
Collaborative Stuay," American Journal of Clinical Nutrition, Vol. 35
(June 1982), pp. 1442-1451.
Robertson, Robert.     "Review of Literature on Costs of Health Services in
Developing Countries," for the WorLd Bank (January 31, 1984).



94
Royston, E.    "Z,he Prevalence of Nutritional Anemia in Women in Developing
Countries: A Critical Review of Available Information,: in Rapport
Trimestriel de Statistiques Sanitaires Modiales, Vol. 35 (1982), pp.
5 2-91.
Sayers, M. H. , et al.   "Iron Absorption from Rice Meals Cooked with Fortified
Salt Containing Ferrous Sulphate and Ascorbic Acid," British Journal of
Nutrition, Vol. 31 (1974).
Scrimshaw, Nevin.     "Functional Consequences of Iron Deficiency in Human
Pooulations," in Journal of Nutrition, Science, and Vitaminology, Vol.
30 (1984), pp. 47-63.
Se lowsky, Marcelo. Nutrition, Health, and Education: The Economic Significance
of Complementarities at Early Age", World Bank Reprint Series: Number 218
from Journal of Develooment Economics, Vol. 9, (1981) pp. 331-346.
Sood, S. K. , et aL.    "WHO Sponsored Collaborative Studies on Nutritional
Anaemia    in  India:    1.  The Effects of Supplemental Oral Iron
Administration to Pregnant Women,: in QuarterlY Journal of Medicine,
Vol. 44 (1975).
Sorkin, Alan L.    Health Economics in Developing Countries (Lexingcon, MA:
Lexingcon Books 19 76).
Spurr, G. 3. , M. Barac-Nieto, and M. G. Maksud. "Productivity and Maximal
Oxygen Consumption in Sugar Cane Cutters," in The American Journal of
Clinical Nutrition Vol. 30 (March 1977), pp. 316-321.
Stephenson, L., 0. Crompton, M. Latham, S. Arnold, and A. Jansen. "Evaluation
of a Four Year Project to Control Ascaris Infection in Children in Two
Kenyan Villages,: in Journal of Tropical Pediatrics, Vo. 29 (June 1983).
Stiglitz, J. F.     "The Efficiency Wage Hypothesis, Surplus Labour, and the
Di s tribut ion o f Income in L.D. C. ' s ," in Oxford Economic Pagers (new
series), Vol. 28, (1976), pp. 185-207.
Stott, G. "Anemia in Mauritius," in Bulletin of the WHO, Vol. 23 (1960).
Vellar, Odd D. and Lars Hermansen. "Paysical Performance and Hematological
Parameters,: in Acta Medical Scan. (Supplement) , Vol. 522 (1971) pp.
1-40.
Viteri, Feraando E. and Benjamin Torun. "Anemia and Physical Work Capacity,"
in Clinics in Haematology, Vol. 3, No, 3 (October 1974), pp. 609-626.
Vi teri, Fe rnando E., et al. "Iron Fortification in Developing Countries," in
A. E. Harper and G. K. Davis (eds.) Nutrition in Health and Disease and
International Develonment: Symposia from the XII International Congress of
Nutrition.    Pro ress in Clinical and Biological Research, Vol. 77 (New
York: Alan R. Liss 1981), pp. 345-354.



95
Vi teri, Fernande E., e t a 1.   "De termining Energy Costs of Agricultural
Activities by Respirometer and Energy Balance Techniques," in American
Journal of Clinical Nutrition, Vol. 24 (December 1971), pp. 1418-1430.
Watt, Bernie K. and A. L. Merrill.        Composition of Foods, Agricultural
Handbook #8, U.S. Department of AGriculture (Washington,D.C. 1963).
Webb, Thomas E. and Frank A. Oski. "Iron Deficiency Anemia and Scholastic
Achievement in Young Adolescents," in The Journal of Pediatrics, Vol 82,
No. 5 (May 1973), pp. 827-830.
Wolgamuth, June, et al.      "Worker Productivity and the Nutritional Status of
Kenyan Road Construction Laborers," in American Journal of Clinical
Nutrition Vol. 36 (July 1982), pp. 68-78.
Woodruff, A. W. "Recent Work on Anaemias in the Tropics," in British Medical
Bulletin, Vol. 28, No. 1 (January 1972), pp. 92-95.
Woodruf f, A. W.    "Recent Work Concerning Anemia in the Tropics," in Seminars
in Hematology, Vol. 19, No. 2 (April 1982), pp. 141-147.
World Bank.    World Develooment Report 1981, (New York: Oxford University
Press).
World Bank.    Wages and Employment in Indonesia, report no. 35.86-IND
(Washington, D.C.: The World Bank 1983).
WorLd Health Organization (WHO). "Training and Utilization of AUK. Personnel
for Rural Health Teams in Developing Countries," technical report series
no. 633 (Geneva: WHO 1979).
World Health Organization (WHO). "Control of Nutritional Anaemia with Special
Re ference to Iron Deficiency,: WHO technical report series #580, report of
an IAEAIUSAID/WHO joint meeting (Geneva: WHO 1975).
World Health Organization (WHO).       Nutritional Anemias, series no. 405
(Geneva: WHO 1968).