World Bank / CODEX Productions COVID-19 Vaccine Inequities and Hesitancy in Iraq Results from June, July, and August 2021 Rounds of Iraq High Frequency Phone Survey (IHFPS) 1 Table of Contents List of Tables Acronyms 04 Table 1: Respondents and household characteristics (by survey round). 14 Acknowledgements 05 Executive summary 07 List of Figures Background 09 Results I: Vaccination and vaccine hesitancy 15 Results 2: Disparities in vaccination and vaccine hesitancy 18 Discussion and Conclusions 28 Figure 1: Daily new cases and deaths attributed to Covid-19 (seven day rolling average). 10 Works Cited 30 Annex I: Sampling Design 33 Figure 2: Cumulative cases and deaths per 1 million population as of January 19, 2022. 11 Annex II: Reweighting for the High Frequency Phone Survey in Iraq (IHFPS 2021) 34 Figure 3: Average stringency index as of January 19, 2022. 11 Figure 4: Daily new Covid-19 vaccine doses administered (seven day rolling average). 13 Figure 5: Share of population vaccinated against Covid-19 as of January 19, 2022. 14 Figure 6: vaccination and vaccine hesitancy among Iraqi adults aged 18 or over. 16 Figure 7: Awareness and knowledge regarding Covid-19 vaccine availability among 17 unvaccinated Figure 8: Awareness and knowledge of vaccine in Iraq by hesitancy category (among 18 unvaccinated adult Iraqis) Figure 9: Knowledge of where to get vaccine by vaccine hesitancy category (among 18 unvaccinated adult Iraqis) Figure 10: Reasons why unvaccinated adult Iraqis are not sure or not planning to get 19 vaccinated Figure 11: Vaccination by education status 21 Figure 12: Vaccine hesitancy by education status 21 Figure 13: Vaccination by employment status 22 Figure 14: Vaccine hesitancy by employment status 22 Figure 15: Vaccination by age group 24 Figure 16: Vaccine hesitancy by age group 24 Figure 17: Vaccination by gender, region, and environment type 24 Figure 18: Vaccine hesitancy by gender, region, environment type 24 Figure 19: Vaccination and respondent’s household food security and expenditure 25 Figure 20: Vaccine hesitancy and respondent’s household food security and expenditure 25 Figure 21: Vaccination by asset ownership 26 Figure 22: Vaccine hesitancy by asset ownership 26 Figure 23: Vaccination by respondent’s household wealth index 27 Figure 24: Vaccine hesitancy by respondent’s household wealth index 27 2 3 Acronyms Acknowledgements CEPI - Coalition for Epidemic Preparedness Innovations This report was prepared by Kwabena Krah and Lokendra Phadera from the World Bank’s Poverty and Equity Global Practice. The project was led by Lokendra Phadera, and Matthew Wai-Poi co-led CFSVA - Comprehensive Food Security and Vulnerability Analysis the project in its early stage. The data collection was a collaborative effort between the World Food FCS - Food Consumption Score Programme (WFP) Iraq and the World Bank. We are grateful for the guidance of the peer reviewers Utz Johann Pape and Nandini Krishnan. Helpful comments and suggestions were also received from IDPs - Internal Displaced Peoples Johannes G. Hoogeveen (World Bank), Fawad Raza, and Tiwonge Machiwenyika (both WFP Iraq). IHFPS - Iraq High Frequency Phone Survey We thank Johannes G. Hoogeveen, Practice Manger for the Poverty and Equity Global Practice for the Middle East and North Africa region, and Ramzi Afif Neman, Head of the Iraq Country Office, for MENA - Middle East and North Africa their overall guidance and support. We are also thankful to Suhair Murad Al-Zubairi for excellent MICS - Multi Indicator Cluster Survey administrative support throughout the project. MNOs - Mobile Network Operators We acknowledge the financial support from the Iraq Reform, Recovery, and Reconstruction Fund (I3RF) trust fund. We are also grateful to Armine Juergenliemk and Timothy D Robertson from the Agriculture mVAM - mobile Vulnerability Analysis and Mapping and Food Global Practice for their collaboration on the data collection of the last three rounds of the PSM - Propensity Score Matching Iraq High Frequency Phone Survey (IHFPS). Finally, we gratefully acknowledge the support of WFP in translation, and thanks to Sandra Estefane (WFP) for designing the report. WFP - World Food Programme 4 5 Executive Summary While Iraq continues to experience a surge in the number of coronavirus cases, the proportion of the country’s population that is vaccinated remains very low. According to data from Our World in Data, as of mid-January 2022, only 15 percent of Iraqis were fully vaccinated, and another 7 percent were partially vaccinated against Covid-19. These numbers are among the lowest in the Middle East and North Africa (MENA) region and are well below the global vaccination rates of 51 and 10 percent fully and partially vaccinated. For a more in-depth understanding of the state of vaccination in Iraq, this note presents findings on Covid-19 vaccination disparities and hesitancy from the last three rounds of the Iraq High Frequency Phone Survey (IHFPS) conducted between June and August 2021. Starting in August 2020, the World Bank collaborated with the World Food Programme (WFP) and implemented nine (9) rounds of the IHFPS as part of the WFP’s monthly mVAM survey. More than 1,600 adults from across Iraq (nationally representative) were interviewed in each round using mobile phones. The survey included several socioeconomic modules in addition to the WFP’s food security module to assess the evolving situation during the pandemic. The last three rounds of the survey also included a module on Covid-19 vaccination to monitor vaccine distribution and acceptance in the country. Findings from the survey suggest a low but increasing vaccination trend among adult Iraqis, however, resistance to the vaccine remain high. With the Delta variant surge in the country, share of Iraqis 18 years and above with at least the first dose of the Covid-19 vaccine doubled between June and August. While another one third were planning to receive the vaccine, the majority (51.9 and 43.6 percent in June and August) were either undecided or were not planning to receive the vaccine. Lack of knowledge about availability did not appear to be a problem as almost all the unvaccinated adults were aware and had knowledge of where to get the vaccine. Fear and mistrust towards the Covid-19 vaccine risks to undermine the efficacy of the country’s vaccination campaign. The majority of those who are reluctant to the vaccine are worried about its side effects. More than half (51 percent) of those who were either unsure or not planning to receive the vaccine in June 2021 indicated concern of possible side effects as the main reason for their reluctancy. Despite a spike in the number of infections, the figure increased significantly in the subsequent two months to 61.2 and 73.4 percent. Low vaccination rate and the important disparities that have emerged in vaccination and vaccine hesitancy highlight the need for a robust and more inclusive vaccination campaign. Iraqis with higher levels of education, with formal public-sector jobs, men, and those in urban areas are more likely to have been vaccinated than those with lower levels of education, with informal private-sector jobs and self-employment, women, and in rural areas. Households’ food consumption and financial security are other distinct features for being vaccinated. Iraqis from poorer households are significantly less likely to have been vaccinated than those from relatively wealthier households. In contrast, these trends reverse when it comes to vaccine hesitancy. The low vaccination among the elderly, and those with higher risk of exposure- poorer households and informal workers that are less likely to work from home and more likely to live in large households in cramped conditions- is of particular concern. Iraq should prioritize the elderly and other vulnerable groups and increase overall access to covid-19 vaccines. Public messaging should be tailored and targeted to different segments of the population. World Bank / CODEX Productions 6 7 ‫موجز تنفيذي‬ ‫رغم ان العراق مازال يشهد ارتفاعا في عدد حاالت اإلصابة بفيروس كورونا ‪ ،‬فإن نسبة السكان الذين‬ ‫ً‬ ‫وفقا لبيانات موقع «عالمنا في البيانات» ‪ ،‬و‬ ‫تم تطعيمهم ضد الفايروس ما تزال منخفضة جدا‪ .‬و‬ ‫ر ا من منتصف كانون الثاني (يناير) ‪ ، 2202‬فان ‪ 51‬في المائة فقط من العراقيين جرى تطعيمهم‬ ‫اعتبا ً‬ ‫بالكامل ‪ ،‬و ‪ 7‬في المائة أخرين تم تطعيمهم جزئيًا ضد كوفيد ‪ .91-‬وتعد هذه األرقام من بين أدنى المعدالت‬ ‫في منطقة الشرق األوسط وشمال إفريقيا (‪ ، )MENA‬وهي أقل بكثير من معدالت التطعيم العالمية‬ ‫البالغة ‪ 15‬و ‪ 01‬في المائة من التطعيمات الكاملة والجزئية على التوالي‪.‬‬ ‫تعمقا لحالة التطعيم في العراق ‪ ،‬تقدم هذه المذكرة نتائج حول‬ ‫ً‬ ‫وللحصول على صورة أوضح و أكثر‬ ‫التفاوتات والتردد تجاه تلقي التطعيم (اللقاح) ضد كوفيد ‪ 91 -‬من خالل الجوالت الثالث األخيرة من‬ ‫المسح الهاتفي عالي التردد في العراق (‪ )IHFPS‬الذي تم إجراؤه بين حزيران ‪ /‬يونيو و آب ‪ /‬أغسطس‬ ‫ابتداء من شهر آب ‪ /‬أغسطس ‪ ، 0202‬تعاون البنك الدولي مع برنامج األغذية العالمي ونفذ تسع‬ ‫ً‬ ‫‪ .1202‬و‬ ‫(‪ )9‬جوالت من المسح الهاتفي عالي التردد في العراق ‪ IHFPS‬كجزء من مسح برنامج األغذية العالمي‬ ‫الشهري عن طريق تحليل نقاط الضعف المتنقلة ورسم الخرائط ‪ . mVAM‬وقد تم اجراء مقابالت مع أكثر‬ ‫من ‪ 0061‬شخص بالغ من جميع أنحاء العراق في كل جولة باستخدام الهواتف المحمولة‪ .‬اشتمل المسح‬ ‫على العديد من النماذج االجتماعية واالقتصادية باإلضافة إلى نماذج األمن الغذائي لبرنامج األغذية‬ ‫العالمي لتقييم حالة األمن الغذائي المتطورة خالل فترة الوباء‪ .‬وقد تضمنت الجوالت الثالث األخيرة من‬ ‫أيضا برنامجا حول التلقيح ضد كوفيد ‪ 91 -‬لرصد توزيع اللقاح وقبوله في البلد‪.‬‬‫المسح ً‬ ‫تشير نتائج المسح إلى وجود اتجاه منخفض ولكنه متزايد ألخذ اللقاح بين العراقيين البالغين ‪ ،‬ومع‬ ‫ذلك ‪ ،‬ال تزال مقاومة اخذ اللقاح عالية‪ .‬و مع انتشار المتغير (المتحور) دلتا في البالد ‪ ،‬تضاعفت حصة‬ ‫العراقيين الذين تبلغ اعمارهم ‪ 81‬عامً ا فما فوق مع الجرعة األولى على األقل من لقاح كوفيد ‪ 91 -‬بين‬ ‫حزيران ‪ /‬يونيو و آب ‪ /‬أغسطس‪ .‬بينما كان ثلث آخر من السكان يخططون لتلقي اللقاح ‪ ،‬كانت الغالبية‬ ‫(‪ 9.15‬و ‪ 6.34‬في المائة في حزيران ‪ /‬يونيو وآب ‪ /‬أغسطس على التوالي) إما مترددة في اخذ اللقاح أو لم تكن‬ ‫‪Background‬‬ ‫تخطط لتلقي اللقاح‪ .‬و ال يبدو أن االفتقار الى المعرفة بشأن التوافر يمثل مشكلة ألن جميع البالغين غير‬ ‫الملقحين تقريبًا كانوا على دراية ولديهم معرفة بمكان الحصول على اللقاح‪.‬‬ ‫إن الخوف وانعدام الثقة تجاه لقاح كوفيد ‪ 91 -‬من المخاطر التي تقوض فعالية حملة التطعيم في البلد‪.‬‬ ‫ان غالبية الذين يترددون في اخذ اللقاح لديهم شعور بالقلق بشأن آثاره الجانبية‪ .‬و أشار أكثر من نصف (‪15‬‬ ‫بالمائة) الذين كانوا غير متأكدين أو ال يخططون لتلقي اللقاح في حزيران ‪ /‬يونيو من عام ‪ 1202‬إلى القلق‬ ‫والمخاوف من اآلثار الجانبية المحتملة باعتباره السبب الرئيسي لترددهم‪ .‬على الرغم من االرتفاع الكبير‬ ‫لعدد اإلصابات ‪ ،‬فقد ارتفع الرقم بشكل ملحوظ في الشهرين التاليين إلى ‪ 2.16‬و ‪ 4.37‬في المائة‪.‬‬ ‫إن انخفاض معدالت التطعيم والتفاوتات المهمة التي ظهرت في التطعيم والتردد في اخذ اللقاح‬ ‫ً‬ ‫شموال‪ .‬إن العراقيين الحاصلين على مستويات‬ ‫يسلط الضوء على الحاجة إلى حملة تطعيم قوية وأكثر‬ ‫تعليمية عالية ‪ ،‬والذين لديهم وظائف رسمية في القطاع العام ‪ ،‬والرجال ‪ ،‬وأولئك الذين يعيشون في‬ ‫المناطق الحضرية من المرجح ان يكونوا قد تلقوا التطعيم بنسبة اكثر من أولئك الذين لديهم مستويات‬ ‫تعليم اقل ‪ ،‬وأولئك الذين لديهم وظائف في القطاع الخاص غير الرسمي والعمل الحر ‪ ،‬والنساء ‪ ،‬و في‬ ‫المناطق الريفية‪ .‬يعد استهالك الطعام واألمن المالي لألسرة من السمات المميزة األخرى للتلقيح‪.‬‬ ‫فالعراقيون الذين ينتمون إلى أسر فقيرة هم أقل احتماال بشكل ملحوظ لتلقي التطعيم من أولئك الذين‬ ‫ثراء نسبيًا‪ .‬في المقابل ‪ ،‬تنعكس هذه االتجاهات عندما يتعلق األمر بالتردد في تلقي‬ ‫ً‬ ‫ينتمون إلى أسر أكثر‬ ‫اللقاح‪ .‬ومما يثير القلق بوجه خاص انخفاض مستوى التطعيم بين كبار السن ‪ ،‬و بين هؤالء الذين لديهم‬ ‫مخاطر أعلى للتعرض – مثل األسر الفقيرة والعمال غير الرسميين (غير النظاميين) الذين هم أقل احتماال‬ ‫للعمل من المنزل واألكثر احتماال للعيش في أسر كبيرة في ظروف ضيقة ‪ .‬ينبغي على العراق إعطاء‬ ‫األولوية لكبار السن والفئات الضعيفة األخرى (مثل الفقراء والعمال غير الرسميين) وزيادة فرص‬ ‫الحصول بشكل عام على لقاحات كوفيد ‪ .91 -‬ينبغي أن تكون الرسائل العامة مصممة وموجهة إلى شرائح‬ ‫مختلفة من السكان‪.‬‬ ‫‪8‬‬ ‫‪9‬‬ Figure 2: Cumulative cases and deaths per 1 Figure 3: Average stringency index as of January Background million population as of January 19, 2022. 19, 2022. 200 100 180 90 Iraq, like many countries around the world, has experienced a surge in the number of Covid-19 160 80 cases with the spread of Omicron variant. After a peak of 12,000 daily new cases during the height 140 120 70 Strigency index of the third surge caused by the Delta variant in the summer of 2021, the numbers receded gradually 100 60 and dropped below 300 by the end of 2021 (Figure 1). However, with the spread of the highly 80 50 contagious Omicron variant, the cases in Iraq increased drastically. The country reported more than 60 40 4,000 new cases on January 19, 2022, and the trend is rising. Number of daily deaths attributed to the 40 30 20 virus followed a similar trend. After a sharp rise during the first wave in July 2020, the number of new 0 20 deaths dropped and remained below 100 throughout. While it increased during another surge in the 10 en ld AE ti n ia n e t on ar n q ria yp ai da in ou I ra I ra ab or at U m Sy n hr st Eg r jib ba Q W Ar Ye Jo le Ba 0 summer of 2021, the daily mortality remained below 100 and gradually dropped to less than 10 per D Le Pa i ud Sa en ld ti ia an e t n ar n q ria AE n day by the end of 2021. Yet, the spread of the new variant into the new year elevated daily mortality yp no in ou ai I ra I ra ab or Cases (000) Deaths (00) at rd m Sy U st Eg hr jib ba Q W Ar Ye Jo le Ba D Le Pa i figures once again. ud Sa Source: Our World in Data https://ourworldindata.org/covid-vaccinations Figure 1: Daily new cases and deaths attributed to Covid-19 (seven day rolling average). Iraq has made only modest progress in inoculating its population against the virus and has a vaccination rate that is well below the global average and one of the lowest in the region. Iraq 13 350 signed up to the COVAX facility2 in the last quarter of 2020 and agreed to pay for 16 million doses. 12 However, it was unable to procure any vaccines until March 2021, when the first 50,000 doses arrived Daily new case s 11 300 in the country from Sinopharm, a Chinese state-owned company. This was followed by 336,000 doses Daily new deaths 10 of the AstraZeneca vaccine through the COVAX facility. The country rolled out its vaccination program 250 prioritizing its elderly population and health workers. As seen in Figure 4, vaccination remained very 9 low until June 2021 with less than 10,000 doses administered per day. Even during the most severe New cases (000) 8 period of the pandemic in July and August, daily inoculation remained below 50,000 doses. After a New deaths 200 7 significant spike for short few days in early September when the country administered almost half 6 a million doses daily, low inoculation rates continue to persist till now with only between 50,000 to 150 5 100,000 doses administered daily. Not surprisingly, the share of population vaccinated against the virus is low in Iraq (Figure 5). As of January 19, 2022, only 21.8 percent of Iraqis had received any dose 4 100 of vaccines; while 14.9 percent were fully vaccinated, 6.9 percent were only partially vaccinated.3 3 This is significantly lower than the global vaccination rate of 60.2 percent - 50.7 fully vaccinated and 2 50 an additional 9.5 partially vaccinated. Albeit most of MENA countries lag behind the global rate and 1 significant disparity within the region, only Syria, Djibouti and Yemen in the region fare worse than 0 0 Iraq in terms of share of population vaccinated against the virus (Figure 5). In addition to the economic and logistical challenges, wide-spread public fear, and mistrust 20 21 0 1 20 0 21 1 0 20 1 21 20 0 21 1 20 21 21 20 21 0 1 -2 -2 -2 -2 -2 -2 l-2 l-2 -2 -2 p- p- r- r- - - g- g- n- n- n- b- b- ec ec ay ay ar ar ov ov ct ct Ju Ju towards the vaccine risks to undermine the efficacy of the country’s vaccination campaign. Ap Ap Au Au Fe Se Fe Se Ju Ju Ja O O M M M M D D N N As has been the case in other countries, vaccine hesitancy, and disinformation regarding the effects Source: Our World in Data https://ourworldindata.org/covid-vaccinations of vaccination remains a hinderance in Iraq. Evidence suggests a clear lack of enthusiasm among some Iraqis even before the vaccination program was rolled out. An online survey conducted in February 2021, also highlighted vaccine hesitancy. Out of a total of 9,431 online respondents, 16 Despite employing swift containments measures, Iraq remains among the most Covid-19 percent indicated that they would not accept COVID-19 vaccines and another 45 percent said they impacted countries in the Middle East and North Africa (MENA) region in terms of cumulative were unsure about getting the vaccines when they become available in the country. Hesitancy against cases and deaths. After Iran, Iraq has registered the highest number of Coronavirus cases the vaccine continued once the vaccine became available. Although not a nationally representative and deaths in the region. As of January 19, 2022, Iraq had registered 2.1 million cases and 24,267 sample, a survey conducted by the CARE Iraq in the governorates of Ninewa and Duhok in July 2021 deaths attributed to the virus. These figures, however, are lower than in Bahrain, Lebanon, Jordan, showed that 68 percent of the people interviewed were not willing to receive the COVID-19 vaccines. Palestine, and Iran when accounting for the size of country’s population (Figure 2). Iraq’s 51,761 cases and 598 deaths per million people are also comparable to the global rates of 43,062 and 707 respectively.1 Nonetheless, Iraq has been swift to implement various containment measures 2 COVAX is a global alliance co-led by the Coalition for Epidemic Preparedness Innovations (CEPI), Gavi - The Vaccine Al- throughout the pandemic, which may have prevented the further spread and deaths from the virus. liance, and the World Health Organization (WHO), with the UNICEF as a key delivery partner. COVAX seeks to provide equitable distribution of vaccines primarily to lower-income countries. Since the rollout began in February 2021, COVAX has delivered Both the federal and the Kurdistan regional governments enforced curfews, lockdowns, school the Covid-19 vaccines to 144 primarily low- and middle-income countries. closures and other measures during various waves of the virus. Based on the stringency index 3 The vaccination rate as reported by the Our World in Data and the IHFPS differ due to various undelaying factors. compiled by Our World in Data, Iraq either topped the chart or remained near the top among the First, while the Our World in Data repots the rate for the entire population, the respondents for the IHFPS are Iraqis 18 years MENA countries throughout the pandemic (Figure 3). or above, and, hence, it reports vaccination rate among adult Iraqis only. Second, the biases related to data sources are dif- ferent. The survey results are estimates and concerns such as the lack of national representation, participation, and response biases, etc. that are related to phone surveys could be of factors. The figures from the Our World in Data, on the other hand, 1 Given limited testing, these figures, however, may be lower than the true number of infections; 416 tests per are calculated using the COVID-19 Data Repository by the Johns Hopkins University and their accuracy depends on how well thousand people is among the lowest in the region and below the global average (Ritchie, et al., 2020). the repository compiles the vaccination information. 10 11 Figure 4: Daily new Covid-19 vaccine doses administered (seven day rolling average). 500 450 400 350 New doses (000) 300 250 200 150 100 50 0 Jan-21 Feb-21 Mar-21 Apr-21 May-21 Jun-21 Jul-21 Aug-21 Sep-21 Oct-21 Nov-21 Dec-21 Source: Our World in Data https://ourworldindata.org/covid-vaccinations For a more in-depth understanding of the state of vaccination, this note presents findings on Covid-19 vaccination disparities and hesitancy in Iraq from the last three rounds of the Iraq High Frequency Phone Survey (IHFPS) conducted between June and August 2021. The World Bank collaborated with the World Food Programme (WFP) and implemented nine (9) rounds of the IHFPS as part of the WFP’s monthly mVAM survey. To monitor the evolving situation and provide policy relevant recommendations, the survey included several socioeconomic modules in addition to the WFP’s food security module. For each round, more than 1,600 adult respondents from across Iraq (nationally representative) were interviewed using mobile phones. However, we note that although the phone survey was designed to be a panel, individuals who cannot be tracked are replaced to meet the target sample size in each round. Since the survey is part of the WFP’s long ongoing monthly mVAM survey, the attrition is substantial. For the last three rounds of the 1,627 individuals interviewed in June 2021, 1,404 could be tracked through July to August 2021. Thus, an attrition rate of 24.3 percent.4 While the first 6 rounds of the survey were implemented between August 2020 and January 2021, the last three rounds were conducted between June and August 2021. These last three rounds also included a module on Covid-19 vaccination to monitor vaccine distribution and acceptance in Iraq. World Bank / CODEX Productions 4 The data collection methodology consists of a countrywide survey covering the 18 governorates in Iraq. The sample size is disaggregated by 18 governorates and the survey firm applied a random sampling approach to reach participants from different governorates to reach the given geographical quotas. The governorate population and details of quota are provided in Annex I. All major Mobile Network Operators (MNOs) active in the country were included within the sampling frame to ensure a representative sample. The sample size is designed to detect changes in the prevalence of food insecurity (mainly people with inadequate food consumption) at governorate level as reported in the 2016 Comprehensive Food Security and Vulnerability Analysis (CFSVA) survey in Iraq. Average number of attempts per phone number was below 1.5 calls for all three rounds. Krah, Phadera, & Wai-Poi (2021) provide details on sampling and survey design. 12 13 Figure 5: Share of population vaccinated against Covid-19 as of January 19, 2022. UAE 92 7 Saudi Arabia 67 5 Iran 63 8 Bahrain 68 1 World 51 10 Share of people fully vaccinated Jordan 40 4 Palestine 29 11 Share of people only partly Egypt vaccinated 25 11 Lebanon 29 7 Iraq 15 7 Syria 5 7 Djibouti 8 4 Yeme n 1 1 0.0 20.0 40.0 60.0 80.0 100.0 Source: Our World in Data https://ourworldindata.org/covid-vaccinations Table 1 reports summary statistics of the respondents and their household characteristics Results I: for each of the three last rounds of the survey. As in the earlier rounds, most of the survey respondents (about 65 percent, unweighted) in the June to August 2021 rounds were males. About 60 percent had at least a secondary education and most of the respondents (over 60 percent) were their households’ main breadwinners. The average respondent was about 36 years old and lived Vaccination and in a household with about 6 members including herself. Across all three rounds, above 54 percent reported living in their own dwelling.5 Vaccine Hesitancy Table 1: Respondents and household characteristics (by survey round). Characteristic June July August Percent of respondents who are male 64.3 64.6 65.3 Percent of respondents with secondary or higher level of education 60.0 58.8 59.0 Percent of respondents who are breadwinners 61.6 64.4 62.5 Average age of respondents 35.8 35.9 35.9 Average household size 6.4 6.5 6.5 Percent of respondents living in a dwelling that their household owned 54.2 54.1 54.5 Sample size 1627 1635 1628 Source: Authors’ calculation using IHFPS 2021. The results in the note are presented in two parts. The first part provides overall trends of vaccination and vaccine hesitancy among the adult population, awareness of vaccine availability and knowledge of where to get vaccines, and concerns/reasons why some may be hesitant to receive one. The second part of the results focuses on the latest data (August 2021 round) to delve deeper into vaccination inequities and heterogeneity in vaccine hesitancy. 5 To ensure representativeness at national level, we construct cross-sectional survey weights for each round. Although, cost-effective, flexible and can be implemented rapidly, lack of national representativeness of phone surveys is of concern. Therefore, using the nationally representative Multi Indicator Cluster Survey (MICS) 2018 as a reference survey, we reweight the initial sampling weights through propensity score matching (PSM) and post-stratification procedures. Three set of weights - household, population, and adult – are created to make the phone survey resemble the distribution of the specific population in the MICS survey. Weight calculation and reweighting procedure are detailed in Annex II. All the results presented in the report are calculated using adult weights unless specified otherwise. 14 15 Results I: Vaccination and Vaccine Hesitancy The survey suggests a low but an increasing trend of Covid-19 vaccination among adult Iraqis (i.e., those 18 years old and above) between June and August 2021. As of June 2021, only 11.9 percent of the survey respondents had received at least a dose of COVID-19 vaccine. This figure increased slightly to 12.4 percent in July (Figure 6). With the increased availability of vaccines in the country, the figure increased by 7.3 percentage points to 19.7 percent by the end of August 2021. World Bank / CODEX Productions Figure 6: vaccination and vaccine hesitancy among Iraqi adults aged 18 or over. 45 40 38.4 36.2 36.8 35 33.9 Percent of adult population 30 30.6 25 24.1 20 18.7 19.7 19.5 Lack of knowledge regarding the availability of vaccine does not appear to be a problem as 17.9 15 most unvaccinated adults are aware and have knowledge about where to get the vaccine. The 11.9 12.4 survey showed that most of the unvaccinated adults are aware of the country’s vaccination program. 10 By June 2021, 90.7 percent of adult Iraqis who were yet to receive a single dose of the vaccine were aware of the availability of vaccine in the country. This figure increased to 93.4 and 95.7 percent in July 5 and August respectively (Figure 7). Moreover, most of the unvaccinated adults are also aware of where 0 to get the vaccine. While 83.0 and 89.8 percent of the unvaccinated adults in June and July indicated Jun-21 Jul-21 Aug-21 that they knew where to get the vaccine, the share rose to 95 percent by end of August. Received at least a dose of the vaccine Planning to be vaccinated Not planning to get vaccinated Not sure Source: Authors’ calculation using IHFPS 2020/21. The figures are weighted using adult survey weights. Figure 7: Awareness and knowledge regarding Covid-19 vaccine availability among unvaccinated 98 While it decreased slightly, vaccine reluctancy remained high even during the period when 96 95.7 daily infections were highest in the country. The share of adults planning to receive the vaccine, 94.9 94 remained constant during the three months (Figure 6). The figure increased slightly from 36.2 to 38.4 93.4 percent from June to July, but fell back to 36.8 percent in August, which, to a degree, may have been 92 driven by some of those who were planning to receive vaccine in the earlier rounds already getting 90.7 vaccinated. More worrying, on the other hand, is high share of adults who still indicated either being 90 Percent 89.8 unsure or not planning to be vaccinated against the virus. One third (33.9 percent) of the adults in June reported that they did not plan to receive the vaccine. When the Covid-19 cases started to spike 88 in July and August, the figure dropped significantly by 6.5 percentage points from 30.6 and 24.1 respectively. Those still unsure of getting vaccinated against the virus increased from 17.9 percent in 86 June to 19.5 percent in August. Almost half of the adults in Iraq (51.9 percent in June and 43.6 percent in August) still being unsure or not planning to be vaccinated is concerning given the importance of 84 the vaccine’s efficacy against severe symptoms and hospitalization. This may undermine the efforts to 83.0 82 reduce the spread of the virus and returning of social and economic normalcy in the country.6 80 June July August Awareness of COVID-19 vaccine availabilty in Iraq Knowledge of where to get vaccine 6 Results from the January 2021 round of the Iraq High Frequency Phone Survey (IHFPS) showed that over 32 percent of Iraqi adults were uncertain about getting the shots - 16.3 percent of the respondents expressed “no opinion” to the state- Source: Authors’ calculation using IHFPS 2020/21. The figures are weighted using adult survey weights. ment “I would accept a COVID-19 vaccine which is proven to be safe and effective”, while 15.9 percent disagreed either strongly or to some degree with the statement (Krah K. , Phadera, Tanner, & Mugera, 2021). 16 17 Furthermore, awareness as regards to vaccine accessibility remain similar between the group Concerns regarding possible side effects of the vaccine is the single most cited reason why a hesitant to be vaccinated and the one planning to receive the vaccine. Among both hesitant significant number of adults in Iraq remain hesitant to the COVID-19 vaccine. A majority of groups (unsure and not intended to be vaccinated) in June to August 2021, above 87 percent were those who are reluctant to the vaccine are worried about its side effects. More than half (51 percent) aware of vaccine availability in the country. This trend was very similar for the group that was planning of those who were either unsure or not planning to receive vaccine in June 2021 indicated concern to receive the vaccine (Figure 8). Share of people with knowledge of where to get vaccine was similarly of possible side effects as the main reason for their reluctancy. Despite spikes in the number of high and increased over time for all the three groups (Figure 9). In fact, the share was consistently infections driven by the Delta variant, the figure increased significantly in the subsequent two months greater among the undecided group, even more so than among those who were planning to receive to 61.2 and 73.4 percent (Figure 10). This trend has a potential to impede the progress of vaccination. the vaccine. These finding suggest that while awareness of vaccine availability in the country is important, it might not be enough in getting people to be vaccinated. Although significantly less, other more cited reasons such as believing in having lower risks of contracting the virus and the vaccine not being effective decreased with rising infections. At the beginning of the third wave driven by the Delta variant in June, about 10 percent cited “I don’t Figure 8: Awareness and knowledge of vaccine in Iraq by hesitancy category (among think the vaccine will work” and another 13 percent cited “I am not at risk of contracting Covid-19” as unvaccinated adult Iraqis) reasons for not wanting to take vaccine (Figure 10). By the end of August when daily infections were at the peak, only 4.4 and 4.3 cited such reasons. 100 100.0 96.4 95 Figure 10: Reasons why unvaccinated adult Iraqis are not sure or not planning to get vaccinated 94.7 93.8 93.6 80 91.2 90.8 90 73.4 89.7 Percent 87.6 70 85 61.2 60 50.7 80 50 Percent of unvaccinated adult Iraqis 75 40 June July August 30 Planning to receive vaccine Not planning to receive vaccine Not sure Figure 9: Knowledge of where to get vaccine by vaccine hesitancy category (among 20 14.915.6 unvaccinated adult Iraqis) 13.1 9.6 8.3 10 7.2 7.3 7.3 4.5 4.8 5.8 4.3 4.8 100 4.4 2.9 97.6 0 I don't think it will I am worried I already had I am not at risk of In general , I don't Other 95 work about the side covid-19 contracting trust vaccines 94.2 93.2 93.5 effects COVID-19 June (N= 762) July (N = 682) August (N = 554) 90 89.0 Percent 88.8 88.8 Source: Authors’ calculation using IHFPS 2020/21. The figures are weighted using adult survey weights. 85 83.1 80 79.9 75 June July August Planning to receive vaccine Not planning to receive vaccine Not sure Source: Authors’ calculation using IHFPS 2020/21. The figures are weighted using adult survey weights. 18 19 Results 2: Disparities in vaccination and vaccine hesitancy Vaccination and vaccine reluctancy differ by level of education. Based on the latest available phone survey data (August 2021), individuals with higher education are significantly more likely to receive the Covid-19 vaccine, whereas vaccine hesitancy is more prevalent among those with lower levels of education. As of August 2021, only 4.8 percent of adult Iraqis with no formal education had received any vaccine compared to 29.1 and 45.6 percent of those with secondary and tertiary levels of education (Figure 11). In contrast, more than two third (68.9 percent) of adults with no formal education were reluctant/hesitant (either not sure or were not planning) to receive the vaccine while only 25.7 of people with tertiary education indicated so (Figure 12). Efforts need to be made to reach the less educated people in the country with information about the safety of Covid-19 vaccines. Figure 11: Vaccination by respondents’ education status 100% Share of adults with at least the first Covid-19 Results II: 90% 80% 70% Disparities in 60% vaccine dose 50% 45.6% 40% Vaccination and 29.1% 30% 20% 11.8% 10% Vaccine Hesitancy 4.8% 0% No education Primary Secondary Tertiary Figure 12: Vaccine hesitancy by respondents’ education status 100% 90% Share who are not sure / not planning to get 80% 68.9% 70% 60% Covid-19 vaccine 48.9% 50% 40% 29.0% 30% 25.7% 20% 10% 0% No education Primary Secondary Tertiary Source: Authors’ calculation using IHFPS August 2021 round. The figures are weighted using adult survey weights.. 20 21 Iraqis employed in the formal public sector are significantly more likely to have received the Covid-19 vaccine and are also less hesitant to receiving it compared to those in private sector, self-employed and those without jobs. By August 2021, more than 42 percent of adults working in the public sector had received at least the first dose of Covid-19 vaccine, while only 17.8 percent of those employed in the private sector had received any doses of the vaccine (Figure 13). The rate was even lower for those self-employed (6.8 percent) and unemployed (13.9 percent). Prevalence of the vaccine hesitancy, on the other hand, was highest among self-employed (Figure 14). Unemployed adults and the private sector workers were also significantly more likely to be resistant against the vaccine take-up. Figure 13: Vaccination by respondents’ employment status 100% 90% 80% Share of adults with at least the first 70% Covid-19 vaccine dose 60% 50% 42.1% 40% World Bank / CODEX Productions 30% 17.8% 20% 13.9% 10% 6.8% 0% Wage Public Wage Private Self employed Unemployed Figure 14: Vaccine hesitancy by respondents’ employment status 100% Older Iraqis are slightly less likely to be vaccinated and more resistant to the vaccine. Aside 90% from people between 40 to 49 years, other age groups had vaccination rates that is relatively close to that of the overall rate of 19.7 percent in August (Figure 15). However, it is notable that planning to get Covid-19 vaccine 80% Share who are not sure / not only 17.7 percent of the elderly (60 and more) population was vaccinated, which was lower than 70% 64.7% among some of the younger age groups with less covid risks. Furthermore, compared to the young 60% adults between 18 to 29, older Iraqis are more hesitant to receive the COVID-19 vaccine. Among those 47.0% in age group 18-29, 38 percent indicated that they were not sure or not planning to get the vaccine but 50% 40.0% more than half of those in the age category 40-49 expressed such view in August (Figure 16). Vaccine 40% resistance was high (45.5 percent) even among elderly, 60 years and above. While the aged and those 30% 26.8% with pre-existing conditions are known to have higher risks of developing severe symptoms, younger Iraqis tend to be more receptive to the vaccine. 20% Vaccination rates are relatively higher for men, in the Central region, and in urban areas but 10% the prevalence of vaccine hesitancy remain consistent across gender, geographical regions 0% and environment. The results suggest that, as of August 2021, while 21.0 percent of adult male Wage Public Wage Private Self employed Unemployed Iraqis had received at least a dose of the vaccine, only 18.3 percent of the adult female population had done so (Figure 17). A similar percent of male and female adults (44.0 vs 43.2 percent) was either not sure or were not planning to receive the vaccine as of August 2021 (Figure 18). Similar findings emerge among rural and urban populations. While 21.4 percent of urban dwellers had received at Source: Authors’ calculation using IHFPS August 2021 round. The figures are weighted using least one dose of COVID-19 vaccine as of August 2021, only 15.3 percent of rural dwellers had done adult survey weights.. so during this period. In terms of vaccine hesitancy, 43.2 percent, and 44.6 percent of adult Iraqis in urban and rural areas indicated they’re not sure / not planning to receive the vaccine. 22 23 Figure 15: Vaccination by respondents’ age group Figure 16: Vaccine hesitancy by respondents’ Both vaccination and vaccine hesitancy are strongly correlated with the respondents’ aage group household food consumption and financial security. Individuals from households that consume adequate diets are more likely to be vaccinated and so are those whose household have enough 100% resources to sustain current expenditures. Based on the WFP’s food consumption score (FCS), which Share of adults with at lest the first Covid- 90% 100% measures the dietary diversity and frequency of food consumed, we define households with “poor” planning to get Covid-19 vaccine 90% and “borderline” food consumption, that is FCS less than 35, as those with inadequate diet. While Share who are not sure / not 80% 80% 70% 19.8 percent of respondents from households that consumed adequate diets indicated they had 70% 19 vaccine dose 60% 60% received at least a dose of the vaccine as of August 2021, 16.9 percent of those from households 52.9% 50% 50% 47.7% 45.5% that consumed inadequate diets indicated so (Figure 19). The survey also asked if the households 40% 40% 38.0% 37.4% had enough resources to sustain usual expenditures for the current month. Adult vaccination rate 30% 21.0% 20.9% 21.7% 30% was even more contrasting between households with and without the resources to meet the regular 17.7% 20% 14.9% 20% expenditures, 26.4 and 9.0 percent respectively (Figure 19). Prevalence of vaccine resistance, on 10% 10% the other hand, are reversed. Adults from households with inadequate diets and economically less 0% 18 - 29 30 - 39 40 - 49 50-59 60+ 0% secured (not having sufficient resources to meet current expenditures), were significantly more likely 18 - 29 30 - 39 40 - 49 50-59 60+ Age group of respondents Age group of respondents to be unsure or not planning to receive the vaccine (Figure 20). Source: Authors’ calculation using IHFPS August 2021 round. The figures are weighted using adult survey weights.. Figure 19: Vaccination and respondent’s household food consumption and expenditure Figure 17: Vaccination by gender, region, and environment type 80% Share of adults with at least the first 70% 100% Covid-19 vaccine dose 60% 90% Share of adults with at least the first 50% 80% Covid-19 vaccine dose 70% 40% 60% 30% 26.4% 50% 19.8% 20% 16.9% 40% 9.0% 30% 26.2% 10% 21.0% 21.4% 18.3% 16.3% 17.4% 20% 15.9% 15.3% 0% 10% Yes No Yes No/Not sure 0% Household consume adequate diet Can meet current month's Male Female KRI North Center South Urban Rural household expenditures Gender Region Environment Figure 20: Vaccine hesitancy and respondent’s household food consumption and expenditure 100% Vaccine hesitancy by by gender, region, environment type Figure 18: Share who are not sure / not planning to get 90% 80% 80% 67.6% 70% planning to get Covid-19 vaccine Share who are not sure / not 70% 60% 56.5% Covid-19 vaccine 60% 50.0% 50% 50% 44.0% 43.2% 46.3% 43.2% 44.5% 42.0% 39.7% 40% 37.4% 40% 35.4% 30% 30% 20% 20% 10% 10% 0% Male Female KRI North Center South Urban Rural 0% Gender Region Environment Yes No Yes No/Not sure Household consume adequate diet Can meet current month's household expenditures Source: Authors’ calculation using IHFPS August 2021 round. The figures are weighted using adult survey weights.. Source: Authors’ calculation using IHFPS August 2021 round. The figures are weighted using adult survey weights.. 24 25 Figure 23: Vaccination by respondent’s household wealth index World Bank / CODEX Productions 100% 90% 80% Share of adults with at least the first 70% COvid-19 vaccine dose 60% 50% 40.7% 40% 30% 19.6% 20% 9.0% 9.3% Household’s ownership of some key assets is also highly correlated with vaccination status 10% and vaccine acceptance. Household owning a dwelling (23.2 precent), computer (29.2 percent) 0% and having internet access at home (22.0 percent) are associated with greater adult vaccination rate Bottom Second Third Top (Figure 21). Adults from households that own these key assets are also less resistant to the vaccine Wealth quartile index compared to those without the assets, but the rate of hesitancy remains high across the board. In August 2021, about 40 percent of adults from the households with these assets were not sure or not planning to receive the vaccine (Figure 22). In comparison, 67.6 percent 48.8 and 52.5 percent of the Figure 24: Vaccine hesitancy by respondent’s household wealth index adults from the households that did not own dwelling, computer and did not have internet at home were resistant to the vaccine. 100% 90% Figure 21: Vaccination and asset ownership 80% Share who are not sure / not planning to 70% 60% 54.3% at Home Interne t No 9.2% 51.6% get COVID-19 vaccine 50% Yes 21.9% 40% 34.2% 34.4% Household computer owns a 30% No 10.0% 20% Yes 29.2% 10% owns dwelling Household 0% No 12.4% Bottom Second Third Top Yes 23.2% Wealth quartile index 0.0% 10.0% 20.0% 30.0% 40.0% 50.0% 60.0% 70.0% 80.0% Share of adults with at least the first COvid-19 vaccine dose Source: Authors’ calculation using IHFPS August 2021 round. The figures are weighted using adult survey weights.. Figure 22: Vaccine hesitancy and asset ownership at Home Interne t No 52.5% Overall, adult vaccination and vaccine acceptance is closely associated with household Yes 41.6% economic status. Based on the food consumption, expenditure, and asset ownership variables discussed above, we create household wealth index using the principal component analysis (PCA). Household computer Consistent with the earlier findings, adults from relatively wealthier households are more likely owns a No 48.8% Yes to have been vaccinated but skepticism against the vaccine is more prevalent among the poorer 38.5% households. Below 10 percent of the adults from the bottom two wealth quartiles were vaccinated as owns dwelling of August 2021 compared to 19.6 and 40.7 from the top two quartiles (Figure 23). While 34.2 and 34.4 Household No 49.4% percent of adults from the third and fourth wealth quartiles were either undecided or had already Yes 40.7% decided on not getting the vaccine, more than a half of the adults from the bottom and second quartiles (54.2 vs 51.8 percent) indicated their reservation against the vaccine (Figure 24). The low 0.0% 10.0% 20.0% 30.0% 40.0% 50.0% 60.0% 70.0% 80.0% Share who are not sure / not planning to get COVID-19 vaccine overall vaccination rate in Iraq is of concern, however, emerging disparities in vaccination and vaccine rejections are equally concerning. These diverging trends appear to be born out of the discrepancies in both the access to vaccine and the critical awareness regarding Covid-19 diseases and the vaccine. Source: Authors’ calculation using IHFPS August 2021 round. The figures are weighted using adult survey weights.. 26 27 Discussion and Conclusions While vaccines remain the most effective methods to control the Covid-19 pandemic and critical for gradual re-opening of the country, the share of Iraqis receiving Covid-19 vaccine remains very low. Early on when the Covid-19 vaccines became available, countries across the world grappled with procuring enough vaccine for their populations. Poorer countries struggled to afford and faced logistical challenges with procurement and remained behind in the queue. But with the increased availability of vaccine, especially, after the wealthier nations’ inoculation programs, vaccination rates have increased even in the poorer countries. Vaccination in Iraq, on the other hand, are still sluggish and the vaccination rate lags well behind the global rate. Besides a bump for few days in September 2021 (after the country received over 500,000 COVID-19 vaccine doses from the US government), the daily doses of vaccine administered remained low throughout (Figure 4). According to data from Our World in Data, only 15 percent of the Iraqi population is fully vaccinated (Figure 5). This number is significantly lower than the global rate of 51 percent and among the lowest in the region; only lowered by Syria, Yemen, and Djibouti (Figure 5). Fear and misinformation regarding the Covid-19 vaccine remain widespread. Even during the height of the Delta wave in August 2021, 45 percent of adult Iraqis were either undecided or had decided against receiving the COVID-19 vaccine. More than 70 percent cited fear of the vaccine’s side effects as the main reasons for resisting to be inoculated against the virus. While there is no other nationally representative survey to compare against, these findings are in line with other studies that have documented the high prevalence of public fear and hesitancy of the vaccine in Iraqi Kurdistan (Tahir, et al., 2021), in Ninewa and Duhok Governorates (CARE Iraq, 2021), and even among healthcare workers in Kurdistan (Luma, Haveen, Faiq, Stefania, & Leonardo, 2022). This is driven by Discussion and misinformation regarding the Covid-19 vaccine, which remains widespread among Iraqi communities (Al-Rubaye, Abdulwahid, Ejbary, Al-Rubaye, & Albadran, 2022). While the overall low vaccination rate in Iraq is of concern, important trends have emerged Conclusions regarding vaccination and vaccine acceptance pointing to disparities in access and awareness about the vaccine. Iraqis with higher levels of education, with formal public-sector jobs, men, and those in urban areas are more likely to have been vaccinated than that with lower levels of education, with informal private-sector jobs and self-employment, women, and those in rural areas. Most importantly, the vaccination among the elderly, who are most vulnerable, remain very low; even lower than some of the young age groups with less Covid risks. Households’ food and financial security are other distinct features for being vaccinated. Poorer Iraqis are significantly less likely to have been vaccinated than wealthier people even though poor people tend to be more at risk of infection as they are more likely to live in large households in cramped conditions and often do informal jobs that require direct interaction with people and less likely to work from home. In contrast, these trends reverse when it comes to vaccine hesitancy. These vulnerable groups also tend to have lesser access to correct knowledge (CARE Iraq, 2021) and high misinformation (Al-Rubaye, Abdulwahid, Ejbary, Al- Rubaye, & Albadran, 2022) regarding the Covid-19 vaccine. Iraq should focus on increasing access to the Covid-19 vaccine for all its population through a robust and more equitable vaccination strategy. Health system in Iraq was already weakened by the decades of war and conflict and mismanagement of public services and was not prepared to handle the pandemic (Al-Saiedi, 2021). This is reflected by the number of cases and deaths, which remain among the highest in the region. While the government has made some efforts, it should speed up the inoculation process including targeting the elderly, marginalized groups such as IDPs, refugees, informal workers, women, and those from the lower socio-economic groups. It is equally vital that the government intensifies public education campaigns to counter the widespread misinformation regarding the Covid-19 vaccine. Infections in Iraq, so far, have spiked with the emergence of new COVID-19 variants like Delta and Omicron. The country should aggressively counter the misinformation through effective information campaign and readied itself for the future waves. High prevalence of vaccine hesitancy and misinformation among the most marginalized and vulnerable groups is of concern. Public messaging should be tailored and target these vulnerable groups. 28 29 Works Cited • Al-Rubaye, A., Abdulwahid, D., Ejbary, A., Al-Rubaye, L., & Albadran, A. (2022). COVID-19 Vaccine– Related Misinformation in The Iraqi Community. Iraqi National Journal of Medicine, 4(1), 129–137. doi:https://doi.org/10.37319/IQNJM.4.1.13 • Al-Saiedi, A. (2021, 8 12). Latest COVID-19 Surge Pushes More Iraqis to Get Vaccinated, But Hesitancy Still Remains. Retrieved 2 16, 2022, from https://phr.org/our-work/resources/latest- covid-19-surge-push-more-iraqis-to-get-vaccinated-but-hesitancy-still-remains/ • CARE Iraq. (2021). COVID-19 Vaccination Uptake: A study of Knowledge, Attitudes and Practices of Marginalized Communities in Iraq. • Ceyhun, H. E. (2019). Arab Barometer - Wave V: Social Capital in the Middle East and North Africa. Arab Barometer. • Kolenikov, S. (2014). Calibrating survey data using iterative proportional ftting. The Stata Journal, 14 (1): 22 - 59. • Kolenikov, S. (2019). Updates to the ipfraking ecosystem. The Stata Journal, 19 (1): 143-184. • Krah, K., Phadera, L., & Wai-Poi, M. (2021). Iraq high frequency phone survey (IHFPS): Resutls from August, September, and October Rounds. Washington, DC: Wold Bank. • Krah, K., Phadera, L., Tanner, J., & Mugera, H. (2021). Iraq High Frequency Phone Survey (IHFPS) To Monitor Socioeconomic Trends During COVID-19: Results from October, November, December 2020, and January 2021 Rounds. Washington DC: World Bank. Works Cited • Luma, A. H., Haveen, A. H., Faiq, B. B., Stefania, M., & Leonardo, E. G. (2022). Hesitancy towards Covid-19 vaccination among the healthcare workers in Iraqi Kurdistan. Public Health in Practice, 3, 100222. • Ritchie, H., Mathieu, E., Rodés-Guirao, L., Appel, C., Giattino, C., Ortiz-Ospina, E., . . . Roser, M. (2020). Coronavirus Pandemic (COVID-19). Retrieved 1 19, 2022, from Our World in Data: https:// ourworldindata.org/coronavirus • Tahir, A. I., Ramadhan, D. S., Taha, A. A., Abdullah, R. Y., Karim, S. K., Ahmed, A. K., & Ahmed, S. F. (2021). Public fear of COVID-19 vaccines in Iraqi Kurdistan region: a cross-sectional study. Middle East Current Psychiatry, 28(46). Retrieved from https://doi.org/10.1186/s43045-021-00126-4 • UNICEF. (2020, 5 31). Spread awareness, not fear— UNICEF reaches 14 million people in Iraq through SMS with critical information related to Covid-19. Baghdad: UNICEF. Retrieved 2 16, 2022, from https://www.unicef.org/iraq/press-releases/spread-awareness-not-fear-unicef-reaches-14- million-people-iraq-through-sms-critical • WHO; UNICEF. (2021). Iraq received almost 3 million doses of the COVID-19 vacccine through COVAX. Baghdad. Retrieved from https://reliefweb.int/report/iraq/unhcr-press-release-9-june- 2021-enar/iraq-receives-almost-3-million-doses-covid-19-%09vaccine-%09through-covax-enar • World Bank. (2021, 9 28). US$100 Million to Support COVID-19 Vaccination in Iraq. Baghdad. Retrieved from https://www.worldbank.org/en/news/press-release/2021/09/24/us-100-million-to- support-covid-19-vaccination-in-iraq 30 31 Annex I: Sampling Design To ensure representativeness, the sample size for the survey was disaggregated by the 18 governorates with a quota set for each governorate. Given a 4% margin of error and 95% confidence level, the sample size of 1620 was the minimum required to ensure a representative sample across governorates. This was calculated by considering the population proportion, margin of error and the confidence level as follows: Unlimited population: Finite population : Where score, margin of error, population size, population proportion A minimum of 1620 households/individuals were interviewed monthly. With approximately 405 interviews carried out every week. The call center applied a random sampling approach to reach the given geographical quotas. Table A presents detailed information about Governorate Population, Target Sample Size and Quotas. Table A: Governorate Population, Target Sample Size and Quotas Monthly Weekly Bi-weekly tar- ADM1 target per target per Monthly Population get per gover- Name governor- governor- Adjusted norate ate ate Target Anbar 2,069,768 81 21 42 84 Basrah 3,383,447 126 32 64 128 Annexes Muthanna Qadissiya Najaf 990,453 1,873,089 1,653,244 34 66 66 9 17 17 18 34 34 36 68 68 Erbil 2,681,017 96 24 48 96 Kirkuk 1,792,045 66 17 34 68 Babil 2,330,682 81 21 42 84 Kerbala 1,470,412 50 13 26 52 Missan 1,338,393 50 13 26 52 Ninewa 4,215,084 154 39 78 156 Salah al-Din 1,628,457 66 17 34 68 Sulaymani- 2,282,730 81 21 42 84 yah Baghdad 8,242,789 262 66 132 264 Wassit 1,548,814 50 13 26 52 Thi-Qar 2,500,447 96 24 48 96 Dahuk 2,703,872 96 24 48 96 Diyala 1,657,588 66 17 34 68 Total 44,362,331 1587 405 810 1620 Note: Population information is from: https://sedac.ciesin.columbia.edu/data/collection/gpw-v4 32 33 Annex II: Reweighting for the High In Iraq, we use the following time-invariant as the target to be matched across surveys: • household size Frequency Phone Survey in Iraq • household size squared (IHFPS 2020) • dependent share • elderly share The spread of COVID-19 and government-imposed social distancing practices across the globe has • adult category (2 or less, between 3 to 5, 6 or more) severely limited the use of traditional, face-to-face interviews in population-based surveys to address • urban/rural the data needs. Recently, a more commonly adopted strategy for collecting household survey data is through phone surveys, which do not require face-to-face interactions and can elicit information from • accommodation (dwelling ownership) individuals and households rapidly and at low cost. Furthermore, these platforms offer flexibility to • residence of region (KRI, North, Center, or South) alter sampling and/or questionnaire design in response to evolving information needs. Initial weights, from the reference and phone surveys serve as a starting point.8 Once, these variables The biggest concern with the phone surveys, however, is the lack of national representativeness. are created in both the surveys, the two dataset are then appended generating a variable named Presumably, people who could be more easily reached by phone should have very different “append”, which takes the value of 1 if an observation is from the phone survey and takes the value of characteristics from people with no phone. For example, it is likely that households who own a phone 0 if it comes from the reference survey. are wealthier than those without. Additionally, households with a phone installed are more likely to reside in urban areas with better infrastructure, whereas households with no phone are more likely to Using “append” as the dependent variable, we implement a logit regression with the above listed be in remote/rural areas. Therefore, phone surveys only represent a certain group of households with variables as regressors (variables that are correlated with the respondent’s likelihood of being reached characteristics, thereby failing to be nationally representative. by phone). Based on the predicted probability, the combined appended data set is then divided into five quintiles. The rest of the procedure is as follow: To address such concern in the Iraq’s High Frequency Phone Surveys (IHFPS), we follow the reweighting procedure developed by the World Bank’s Poverty and Equity Global Practice.7 It I. Compute the quintile-level sum of predicted probability for the reference and phone calibrates the phone surveys against a nationally representative reference household survey and surveys, respectively. readjusts the phone survey to make it nationally representative. II. Compute the sum of predicted probability for both the reference and phone surveys, The 2018 Multiple Indicator Cluster Survey (MICS) was selected as the reference survey. It is a respectively. nationally representative survey with a representative sample at the national and governorate levels of more than 20 thousand families throughout Iraq. The readjustment of the phone survey is done as III. Divide the quintile-level sum by the survey-level sum of predicted probability for both surveys. described below. IV. Divide the quintile-to-total ratio from the reference survey by the quintile-to-total ratio from the phone survey and obtain a new ratio which we name as “coefficient”. 1. Household and population weights: V. Generate a new household weight by multiplying the initial household weights from the phone survey by the coefficient: Step 1: Using the population and the monthly sample size (as reported in Annex I), starting or initial population weights are calculated as: VI. Generate a new population weight by multiplying the initial population weight from the phone survey by the coefficient: where, is the starting/initial population weight of household in governorate . While is the governorate population, is the number of complete phone interviews in a month from governorate . Using the average household size in a governorate, , (from the MICS 2018 survey) we calculate the initial household weights as below: and all the subscripts have same meaning as in equation 1. Step 2: The calculated “initial weights” are then adjusted using the propensity score weighting procedure. The goal is to make the phone survey resemble the distribution of the nationally representative survey as much as possible. To achieve this goal, we need to compare variables that are time-invariant between the two surveys. If these variables are close enough across the two 8 MICS 2018 survey reports only the normalized weighs (the sum of the weights match to total sample size in- surveys, we can safely conclude that the phone survey has resembled the reference survey quite well, stead of population). In order to create the full weights, we multiply the normalize household weights by a constant or, the reweighting has been implemented successfully. - 6,056,089/20,214 (total households in Iraq in 2018/MICS 2018 sample size). Total households in Iraq was identified by dividing the population (as reported in World Development Indicator (WDI)) by average household size. We reproduce Table 7 See Annex 2 of “High Frequency Mobile Phone Surveys of Households to Assess the Impacts of COVID-19 Guidelines SR 4.1 “Age distribution of household population by sex” in the MICS 2018 report using the full weights to access its accura- on Sampling Design”, Version: April 29, 2020, which provides various ways to implement re-weighting procedures. cy. The age distribution matched perfectly with the distribution produced by the normalized weights. 34 35 Step 3: PSM adjusted weights are then calibrated to match governorate and other population sums. To create adult weights, we follow the similar procedure as described in section 1 for household and While the propensity-score-matching-based procedure makes the phone survey closer to “being population weights and make the following adjustment: nationally representative” by overweighing the group of people that were hard to be reached by phone, the population distribution in the phone survey may still differ from the reference survey. Step 1: Since the sampling design did not use the adult population distribution to select the sample At this stage, we implement a procedure named post-stratification or raking to exactly match the and there is no auxiliary data, all the observations are assigned a starting/initial weight of 1 i.e. governorate-level and other population sums between the reference and phone surveys. To be exact, using the population sums (number of households and individuals for households and population weights, respectively) from the reference survey by (i) governorate; (ii) household size category – between 1 to 3, between 4 to 5, between 6 to 7, between 8 to 9, and 10 or more; (iii) dependency ratio Step 2: Here we follow all the steps as described in section 1, except that the MICS 2018 sample category (number of children/household size) – 0-0.1, 0.1-0.3, 0.3-0.5, 0.5-0.6, and 0.6-1; and (iv) urban is limited to adults only when appending the two data sets. In addition to the 8 regressors used in population, we use Kolenikov’s (2014; 2019) ipfraking procedure in Stata and calculate post-stratified section 1, respondents’ age, age squared, gender and education levels are added as extra regressors weights.9 in calculating one’s propensity to be part of the phone surveys. Finally, to correct for outliers, we re-run the above ipfraking procedure by controlling for smallest Step 3: The PSM adjusted adult weights then are calibrated to match adult population sums by and highest raked weights allowed. While the value of 1st percentile of the post-stratified weight from (i) governorate; (ii) household size category; (iii) dependency ratio category; (iv) urban population; the previous ipfraking procedure is used as the lower bound, the 99th percentile is used as the upper (v) region by gender adult population sums; (vi) region by education level (primary, secondary, bound. Weights that are below (above) this lower (upper) bound are increased (trimmed down). bachelors or more), and (viii) region by age category 18-24, 25-29, 30-39, 40-49, 50-59 and 60 or more) population sums. Again, the outliers are corrected as described in section 1. 2. Adult weights: All three full (raw) sample weights are then standardized (or normalized) to make the weighted sum of the interviewed sample units equal to the total sample size. Normalization is done by multiplying Unlike the regular face-to-face household surveys, the Iraq’s High Frequency Phone Surveys do not the full sample weights by a constant factor equal to the unweighted number of total completed collect information on all household members. They elicit household and individual level information interviews (sample size) divided by the weighted total number of completed interviews i.e., total from the survey respondents only. Moreover, by design only adult, 18 years or older, are selected as households, individuals, or adults. survey respondents. Therefore, to assess individual level indicator such as one’s employment status or labor force participation for adults, the calculated household and population weights are not adequate and need to adjust such that the adult population in the phone survey resemble the adult distribution in the nationally representative survey. 9 See Kolenikov, S. 2014, “Calibrating survey data using iterative proportional fitting (raking).” Stata Journal 14: 22- 59 and Kolenikov, S. 2019, “ (Kolenikov, Updates to the ipfraking ecosystem, 2019).” Stata Journal 19: 143-184. World Bank / CODEX Productions 36 37