Policy Research Working Paper 8879 The 2019 Update of the Health Equity and Financial Protection Indicators Database An Overview Adam Wagstaff Patrick Eozenou Sven Neelsen Marc Smitz Development Economics Development Research Group & Health, Nutrition and Population Global Practice June 2019 Policy Research Working Paper 8879 Abstract This paper outlines changes that have been made in the thereof ); and refinements to the estimation of out-of-pocket 2019 version of the Health Equity and Financial Protec- expenditures. On the health equity side, the 2019 database tion Indicators database. On the financial protection side, includes 198 more data points than the 9,733 in the 2018 the changes include an increase in the number of indi- database, reflecting the addition of 535 new datapoints, and cators from five to 14; revisions to several previous data the dropping of 337 previously included data points now points, reflecting the analysis of new surveys (or adaptations considered to be substandard. This paper is a product of the Development Research Group, Development Economics and the Health, Nutrition and Population Global Practice. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://www.worldbank.org/prwp. The authors may be contacted at awagstaff@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team The 2019 Update of the Health Equity and Financial Protection Indicators Database: An Overview Adam Wagstaffa*, Patrick Eozenoub, Sven Neelsenb, and Marc Smitzb a Development Research Group, The World Bank, 1818 H Street, NW, Washington DC 20433, USA b Health, Nutrition and Population Global Practice, The World Bank, 1818 H Street, NW, Washington DC 20433, USA Keywords: Health indicators; health equity; health and inequality; out-of-pocket health expenditures; financial protection; health and poverty; millennium development goals; sustainable development goals; universal health coverage; non-communicable diseases JEL codes: I1, I3, J13 1 Acknowledgments We are grateful to Maxime Émile Armand Roche and Benoît Simon who assisted in the processing of the microdata in this phase of the project, to Gabriel Leung and Irene Wong for help with the Hong Kong data, and to Rachel Lu for help with the Taiwan data. The findings, interpretations and conclusions expressed in this paper are entirely those of the authors, and do not necessarily represent the views of the World Bank, its Executive Directors, or the governments of the countries they represent. Accessing the database A portal version of the 2019 database with visualization functionality can be accessed at http://datatopics.worldbank.org/health-equity-and-financial-protection/. The data set itself can be accessed and downloaded, indicator by indicator, or in its entirety, from https://datacatalog.worldbank.org/node/142861 from which model Stata ‘do’ files can be downloaded to replicate the datapoints in the HEFPI data set. Citing the database The reference citation for the data is: Wagstaff, Adam, Eozenou, Patrick, Neelsen, Sven and Smitz, Marc. 2019. The Health Equity and Financial Protection Indicators Database 2019. World Bank: Washington, DC. * Corresponding author: Adam Wagstaff. Development Research Group, The World Bank, 1818 H Street, NW, Washington DC 20433, USA. Tel: +1 202 473 0566, awagstaff@worldbank.org 2 1. Introduction In a recent paper (Wagstaff et al. 2018), we provided an overview of an international database on Health Equity and Financial Protection Indicators (HEFPI). The data set provides data on the delivery of health service interventions, health outcomes, and ‘financial protection’ in health, at both the population level, and for subpopulations (defined by household living standards) along with a summary measure of inequality known as the concentration index (Wagstaff et al. 1991; Kakwani et al. 1997). The data are computed from well-known household surveys that have been conducted by, or in partnership with, national governments, such as the Demographic and Health Survey (DHS) and the Living Standards Measurement Study (LSMS). This paper outlines changes that have been made in the 2019 version of the HEFPI database. On the financial protection side, the number of indicators has been expanded in the 2019 database from five to 14. Moreover, several estimates have changed from the 2018 database, reflecting in part our analysis of new surveys (or adaptations thereof) since releasing the 2018 database, but also refinements we have made to out-of-pocket expenditure estimates for some surveys included in the 2018 database. On the health equity side, the 2019 HEFPI database includes 197 more datapoints than the 9,733 in the 2018 database. This increase is the result of two major counteracting enhancements of the database: the addition of 535 new datapoints, of which 493 come from recent Multiple Indicator Cluster Surveys (MICS) and several new DHS; and the removal of 338 previously included datapoints for eight health equity indicators which, after a thorough quality review, were concluded to be substandard. These changes are detailed in the sections below. 2. Financial protection indicators On the financial protection side, for the 2019 HEFPI database, 1,846 surveys were analyzed and 650 were retained covering 149 countries. The number of indicators has also been expanded from 5 to 14. In addition, in many cases, estimates have changed from the 2018 database, reflecting 3 in part our analysis of new surveys (or adaptations thereof) since releasing the 2018 database, but also refinements we have made to out-of-pocket expenditure estimates for some surveys included in the 2018 database. 2.1. New indicators The HEFPI 2018 database included five financial protection indicators: the incidence of (i.e. fraction of households experiencing) ‘catastrophic’ health expenditures, defined alternately as expenditures exceeding 10% of consumption or income, or as expenditures exceeding 25% of consumption or income; and the incidence of (i.e. fraction of households experiencing) impoverishing expenditures, defined as expenditures without which the household would have been above the poverty line, but because of the expenditures is below the poverty line, with two absolute poverty lines ($1.90-a-day and $3.20-a-day in 2011 purchasing power parity (PPP) dollars) and one relative poverty line (50% of median consumption or income). The HEFPI 2019 database includes these five indicators plus nine new financial protection indicators: 1. The incidence of impoverishment for two other international poverty lines proposed by Joliffe and Prydz (2016), namely $5.50-a-day (relevant to upper middle-income countries) and $21.70-a-day (relevant to high-income countries). 2. The incidence of impoverishment using the societal poverty line (SPL) proposed by Joliffe and Prydz (2017), equal to the higher of two poverty lines: the $1.90-a-day line or 50% of median consumption. 3. For all four $-a-day poverty lines, the HEFPI 2019 database also includes not just the incidence of impoverishing expenditures but also the addition to the per capita poverty gap (in $ terms) attributable to out-of-pocket expenditures.1 1 This is computed (for the $1.90-a-day poverty line) as 1.9 times the difference between the per capita poverty gap for consumption (or income) net of out-of-pocket expenditures and the per capita poverty gap for consumption (or income) gross of out-of-pocket expenditures. 4 4. The HEFPI 2019 database also includes mean annual household per capita out-of-pocket expenditure in 2011 PPP $ terms. 5. In addition, the 2019 database includes the out-of-pocket expenditure “budget share” – the fraction of household consumption or income absorbed by out-of-pocket expenditures. This turns out to be highly correlated with the incidence of catastrophic expenditures using the 10% threshold (Wagstaff et al. 2019). Figure 1 shows the new structure of the financial protection side of the 2019 HEFPI database, updating Figure 2 in Wagstaff et al. (2018). The yellow-edged boxes are the new indicators in the 2019 database. Figure 1: New structure of the financial protection side of the 2019 HEFPI database Financial  Protection Catastrophic  Impoverishing  Out‐of‐pocket  expenditures (CATA) expenditures (IMPOV) expenditures (OOP) IMPOV poverty  gap CATA 10% IMPOV absolute IMPOV relative Household p.c. OOP (in $ due to OOP) For all abs. poverty lines  60% of median p.c.  Household OOP budget  CATA25% $1.90 and $3.20 per day ($1.90, $3.20, $5.50,  consumption share $21.70 per day) Societal Poverty Line (SPL):  $5.50 and $21.70 per day $1.90 per day or 50% of  median p.c. consumption 2.2. Refinements to out-of-pocket expenditure estimates In preparing the 2019 HEFPI database, out-of-pocket expenditure estimates at the household level have been revised for many of the datapoints included the 2018 database. 5 Estimates from the Luxembourg Income Study (LIS) previously corresponded to total health expenditures (“hcmed” in LIS terminology). These, it turns out, include nonmonetary expenditures, such as the imputed value of publicly financed health insurance in some countries, e.g. the US. In the 2019 HEFPI database, expenditures are – wherever possible – monetary expenditures only (“hmcmed” in LIS terminology).2 This process led to some datapoints being no longer available from LIS. For example, in the case of the US, out-of-pocket expenditures before 2012 reflect only the imputed value of Medicare and Medicaid; actual out-of-pocket expenditures are not available. Of the 54 LIS points that were included in the 2018 HEFPI database, 15 were affected by this revision. Aside from the US, the other dramatic change was Taiwan whose catastrophic health expenditure rate felt from 30 or so percent to under 7% and is now in line with values reported elsewhere (van Doorslaer et al. 2007). Over 30 of the surveys we use ask about out-of-pocket expenditures in both the consumption and health modules. Very often there are differences across them in the number of items and the recall period. Lu et al. (2009) find that use of the single-item question leads to a smaller estimate of out-of-pocket expenditure than the survey’s multi-item question and a 4-week recall period leads to a larger (annualized) estimate of out-of-pocket spending than a 12-month recall period. Our sense is that a single item is unrealistic and that – at least up to a point – asking more items is better. Ideally, the list should explicitly separate outpatient and inpatient care. It also seems clear that a while a 4-week recall period might be appropriate for frequently consumed items, such as over-the- counter medicines, it is too short for most items of health expenditure, and far too short for inpatient care, where a 12-month window seems more appropriate. We ended up switching modules in several surveys we used in the HEFPI 2018 database, mostly to the consumption module, which typically did 2 The total consumption variable was also altered to match and now covers only monetary consumption. 6 better in terms of the number of items, separating outpatient and inpatient care, and the appropriateness of the recall period. 2.3. Survey changes For the 2018 HEFPI database, 1,707 surveys (or adaptations thereof) were analyzed and 575 were retained in the data set covering 142 countries. For the 2019 database, 1,846 surveys were analyzed and 650 were retained covering 149 countries. Of the 650 surveys retained in the 2019 database, 120 were not in the 2018 database. Not all the 575 surveys (or adaptations thereof) retained in the HEFPI 2018 database were retained in 2019 database: 46 were replaced by different ones. This was based in part on comparisons between the old and new surveys of the quality-check indicators outlined in Wagstaff et al. (2018). Sometimes, however, the decision was based on other factors: for example, we learnt that the public release of the US Consumer Expenditure Survey is top-coded which is an issue given we are interested in especially large out-of-pocket expenditures. We also realized that our estimates had not exploited the panel nature of the data set. We ended up favoring instead our (new) estimates from the US Current Population Survey. 3. Health Equity Indicators On the health equity side, the 2019 HEFPI database includes 9,930 data points – 197 points more than the 9,733 points in the 2018 version. This increase is the result of two major counteracting enhancements of the database: on the one hand, we added 535 entirely new points, of which 493 come from the addition of all available fifth and sixth wave surveys of the MICS and several new DHS; on the other hand, we conducted a major quality review of all data points for eight health equity indicators for which we examined the country time-series and dropped 338 previously included outlying points for which we could not find a policy or epidemiological explanation. 7 This trend-check exercise is detailed further in section 3.1 on dropped data points. Section 3.2 discussed data points which are included in both the 2018 and 2019 HEFPI versions but change in value due to changes in their underlying source or our indicator definitions, or because of coding corrections. Section 3.3 introduces the data points we newly added to the HEFPI database in its 2019 version. Finally, section 3.4 gives an overview of the subnational surveys included on the health equity side of HEFPI. 3.1. Dropped data points For the 2019 HEFPI database, we quality-checked all points from the 2018 database and all newly obtained points for the following eight health equity indicators: Inpatient care use, 4+ antenatal care visits, skilled birth attendance, full child immunization, formal health care provider visits for children with acute respiratory infections, receipt of oral rehydration salts among children with diarrhea, adult inpatient care use, mammograms, and pap smears. The quality check consisted in reviewing each point in a country’s time-series for the respective indicator via visual inspection of trend charts. For outlying points that we computed from microdata we reviewed, and if necessary corrected our code.3 If the data came from publications or if we found our coding to be correct, we re-checked if the survey’s raw data indeed allowed us to construct the indicator in line with or sufficiently similarly to the HEFPI indicator definitions4 – and dropped the point if we found this to not be possible. If definitions were sufficiently similar, we 3 Changes in indicator values between the 2018 and 2019 HEFPI versions due to such coding corrections or for other reasons are discussed further in section 3.2.3 below. 4 See Tables 1-3 in Wagstaff et al. (2018) for detailed HEFPI indicator definitions. 8 further examined if there was a policy or epidemiological explanation for the outlying indicator value. If no such explanation could be found, we dropped the data point.5 For example, we found no coding errors or definitory inconsistencies in the 2011 DHS Benin oral rehydration salts receipt rates for children with diarrhea, but nevertheless dropped the point as it was far above the rate of prior and subsequent DHS and MICS surveys (see Figure 2), and no policy or epidemiological explanation for this divergence could be found. Figure 2: Trend-chart check for rate of children under five with diarrhea receiving ORS in Benin Kept Dropped 60% 50% 50.7% 40% DHS 30% 20% 25.9% 23.5% 25.4% 23.7% 22.3% DHS DHS MICS DHS DHS 10% 0% 96 97 98 99 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 Year The trend checks also led us to exclude all insecticide-treated bed net use data points from the second round of the MICS that was carried out between 1998 and 2003, as the discrepancy in 5 The trend-checks underlying our exclusion decisions for the pap smear data are based on new indicator values that use a three-year reference period and women age 20-69 as our preferred age-group instead of a two-year period and women age 30-49 as in the 2018 HEFPI version. Section 3.2.2 discusses this change in detail. 9 definitions6 with the other HEFPI bed net use data points rendered meaningful comparisons impossible. The trend checks, dropping of bed net use data points from the second MICS round and additional ad hoc discovery of surveys which did not fulfill HEPFI requirements7 led us to exclude a total of 338 data points from the 2019 HEFPI database which were included in the 2018 version. 3.2. Changes in data points 3.2.1. Changes in indicator values due to changes in sources The HEFPI 2019 database changed the underlying data sources of 161 data points that were already included in the 2018 version. From multiple to single sources For the 2018 database, we allowed multiple data sources for data points of a given year and indicator. In these cases, the HEFPI data point’s value was computed as the unweighted mean of the indicator values from each source. For instance, the HEFPI 2018 stunting rate for Rwanda in 2000 – 46.2% – was computed as the mean over the data points from the 2000 DHS and MICS surveys. In the 2019 database we no longer use such averaging and instead rely on a single source per data point. When data is available from multiple sources, we select the point to include based on four criteria: (1) Consistency in data sources over time (e.g. prefer the Rwanda 2000 DHS stunting point 6 The other HEPFI insecticide-treated bed net points consider a bed net treated if it is a) a long-lasting treated net, b) a pre-treated net that was purchased or soaked in insecticides less than 12 months ago, or c) a non-pre- treated net which was soaked in insecticides less than 12 months ago. MICS 2 data points by contrast consider bed nets treated if they were ever treated. 7 For example, we dropped two STEPS surveys as their samples did not represent populations meaningful for HEFPI’s purposes, namely the 2015 STEPS in Turkey which focused on Syrian refugees and the 2004 Bhutan STEPS which was carried out among patients of one hospital’s outpatient ward. 10 over that from the MICS as all other points in the stunting time series for Rwanda come from DHS); (2) survey sample size with a preference for larger surveys in order to minimize sampling error; (3) fit with the overall time series where points are dropped if they form outliers without a credible explanation; (4) preference for data points we computed from the micro-data ourselves over data points from publications.8,9 Complete replacement In some cases, we replaced points’ previous source(s) with an entirely new source. Most of the time, this occurred when we managed to obtain the microdata of a point that we had previously sourced from a publication – such as Kosovo’s 2013 skilled birth attendance rate that came from the MICS survey report in the 2018 HEFPI database and that we now computed from the microdata ourselves. Encouragingly, in most cases, any changes in indicator values resulting from the changes in sources are small – the median change relative to the 2018 HEFPI indicator value is 3.9 percent. But sometimes, the changes are meaningful, such as for the United Kingdom’s 2003 inpatient care rate which drops by five percentage points to 8.3 percent when we exclude Eurobarometer data and exclusively rely on the General Household Survey (GHS). 8 For some of the points in the 2018 database which were computed as averages over multiple points, the underlying points came from the same survey. For instance, the 2018 database’s 2013 mammography rate for Belgium was computed as the mean over two rates from the European Health Interview Survey – our own microdata-based rate and the rate published by the OECD. The 2019 database uses the microdata-based rate only. 9 In some cases, sources changed because a source had been erroneously entered: For example, for the 2008 mammography rate for Canada, we mistakenly computed the 2018 HEFPI version’s data point as an average over a point from the 2002 Joint Canada/United States Survey of Health survey report and the 2008 point reported by OECD. This error has been corrected in the 2019 database. 11 Table 1 lists the 39 data points for which changes in data sources from the 2018 to the 2019 HEFPI version led to changes in indicator values of more than 10 percent. 10,11 The table reports the 2018 and 2019 survey identifiers, the 2019 population mean indicator value ̅ and the difference between the 2019 and 2018 indicator values ̅ ̅ . 10 Where we also corrected indicator computations for sources used both in the 2018 and 2019 versions of HEFPI, the difference in indicator values between the two versions is of course also affected by these corrections. 11 For Table 2 we did not consider differences in 2018 and 2019 HEFPI indicator values for pap smear points with changing sources as the change in the pap smear reference period between the database’s two versions confounds any change in indicator values from the change in sources. 12 Table 1: Data points which changed by more than 20 percent due to changes in underlying sources ̅ Indicator 2018 survey identifier 2019 survey identifier ̅ ̅ a_bp_treat AUT_2006_EUBM_v01_M; AUT_2006_EHIS_v01_M AUT_2006_EUBM_v01_M 0.07 -0.06 a_inpatient_1yr BEL_2000_MCSS_v01_M; BEL_2000_ECHP_v01_M BEL_2000_ECHP_v01_M 0.10 -0.01 a_inpatient_1yr BEL_2003_EUBM_v01_M; BEL_2003_WHS_v01_M BEL_2003_WHS_v01_M 0.14 -0.02 CHL_2011_National Socioeconomic Characterization a_inpatient_1yr CHL_2011_ISSP_v01_M 0.10 0.01 Survey_v01_M; CHL_2011_ISSP_v01_M a_inpatient_1yr DNK_2003_WHS_v01_M; DNK_2003_EUBM_v01_M DNK_2003_WHS_v01_M 0.07 -0.03 a_inpatient_1yr FRA_2003_EUBM_v01_M; FRA_2003_WHS_v01_M FRA_2003_WHS_v01_M 0.10 -0.01 a_inpatient_1yr IRL_2000_MCSS_v01_M; IRL_2000_ECHP_v01_M IRL_2000_ECHP_v01_M 0.09 -0.02 a_inpatient_1yr IRL_2003_EUBM_v01_M; IRL_2003_WHS_v01_M IRL_2003_WHS_v01_M 0.11 -0.02 a_inpatient_1yr ITA_2003_WHS_v01_M; ITA_2003_EUBM_v01_M ITA_2003_WHS_v01_M 0.10 -0.02 a_inpatient_1yr JPN_2011_ISSP_v01_M; JPN_2011_KHPS_v01_M JPN_2011_KHPS_v01_M 0.05 -0.02 a_inpatient_1yr KAZ_2002_HBS_v01_M; KAZ_2002_WHS_v01_M KAZ_2002_HBS_v01_M 0.05 -0.02 a_inpatient_1yr LUX_2003_WHS_v01_M; LUX_2003_EUBM_v01_M LUX_2003_WHS_v01_M 0.15 -0.03 a_inpatient_1yr NIC_2001_DHS_v01_M; NIC_2001_EMNV_v01_M NIC_2001_DHS_v01_M 0.06 -0.06 a_inpatient_1yr PRT_2000_MCSS_v01_M; PRT_2000_ECHP_v01_M PRT_2000_ECHP_v01_M 0.05 -0.02 a_inpatient_1yr PRT_2003_WHS_v01_M; PRT_2003_EUBM_v01_M PRT_2003_WHS_v01_M 0.09 -0.01 RUS_2011_WBHHS_v01_M; a_inpatient_1yr RUS_2011_HBS_v01_M; RUS_2011_ISSP_v01_M RUS_2011_HBS_v01_M 0.14 0.02 a_inpatient_1yr ESP_2000_MCSS_v01_M; ESP_2000_ECHP_v01_M ESP_2000_ECHP_v01_M 0.07 -0.01 GBR_2001_GHS 72-04_v01_UKDAP; GBR_2001_GHS 72- a_inpatient_1yr GBR_2001_ECHP_v01_M; GBR_2001_OECD_HS 04_v01_UKDAP 0.09 -0.01 GBR_2003_GHS 72-04_v01_UKDAP; GBR_2003_GHS 72- a_inpatient_1yr GBR_2003_EUBM_v01_M 04_v01_UKDAP 0.08 -0.05 GBR_2004_WHS_v01_M; GBR_2004_GHS 72- GBR_2004_GHS 72- a_inpatient_1yr 04_v01_UKDAP 04_v01_UKDAP 0.09 -0.02 USA_2012_NHIS_v01_IPUMS; USA_2012_NHIS_v01_IPUM a_inpatient_1yr USA_2012_ISSP_v01_M S 0.09 -0.03 c_anc GHA_2003_WHS_v01_M; GHA_2003_DHS_v01_M GHA_2003_DHS_v01_M 0.69 0.09 c_anc MAR_2003_WHS_v01_M; MAR_2003_DHS_v01_M MAR_2003_DHS_v01_M 0.30 0.05 c_anc PHL_2003_DHS_v01_M; PHL_2003_WHS_v01_M PHL_2003_DHS_v01_M 0.68 0.08 c_anc VNM_2002_WHS_v01_M; VNM_2002_DHS_v01_M VNM_2002_DHS_v01_M 0.31 0.04 c_treatARI RWA_2000_MICS_v01_M; RWA_2000_DHS_v01_M RWA_2000_DHS_v01_M 0.15 -0.02 c_treatdiarrhea RWA_2000_MICS_v01_M; RWA_2000_DHS_v01_M RWA_2000_DHS_v01_M 0.14 0.02 m_bmi NPL_2003_STEPS_v01_FS; NPL_2003_WHS_v01_M NPL_2003_WHS_v01_M 21.11 -2.85 m_obese RUS_2003_RLMS_v01_M; RUS_2003_WHS_v01_M RUS_2003_RLMS_v01_M 0.13 0.02 RUS_2011_ISSP_v01_M; RUS_2011_WBHHS_v01_M; m_obese RUS_2011_RLMS_v01_M RUS_2011_RLMS_v01_M 0.16 0.02 w_mam_2y GRC_2003_EUBM_v01_M; GRC_2003_WHS_v01_M GRC_2003_EUBM_v01_M 0.38 0.06 w_mam_2y HKG_2014_PHS_v01_M_pub HKG_2014_PHS_v01_M 0.38 0.20 w_mam_2y HUN_2003_OECD_HS; HUN_2003_WHS_v01_M HUN_2003_OECD_HS 0.60 0.06 MEX_2006_National Health and Nutrition w_mam_2y Survey_v01_M; MEX_2006_OECD_HS MEX_2006_NHNS_v01_M 0.38 0.08 w_mam_2y PRT_2003_WHS_v01_M; PRT_2003_EUBM_v01_M PRT_2003_EUBM_v01_M 0.69 0.09 RUS_2011_WBHHS_v01_M; w_mam_2y RUS_2011_RHS_v01_M_pub RUS_2011_WBHHS_v01_M 0.13 -0.16 w_obese NPL_2003_STEPS_v01_FS; NPL_2003_WHS_v01_M NPL_2003_WHS_v01_M 0.02 -0.04 w_obese RUS_2003_RLMS_v01_M; RUS_2003_WHS_v01_M RUS_2003_RLMS_v01_M 0.30 0.05 RUS_2011_ISSP_v01_M; RUS_2011_WBHHS_v01_M; w_obese RUS_2011_RLMS_v01_M RUS_2011_RLMS_v01_M 0.30 0.06 13 Changes in survey identifiers For a number of data points in the 2019 HEFPI database which we list in Table 2, we abbreviated survey identifiers used in the 2018 version or corrected them where we previously had used an incorrect survey year. Table 2: Changes in survey identifiers 2018 survey identifier 2019 survey identifier ARG_2005_National Risk Factor Survey_v01_M ARG_2005_NRFS _v01_M ARG_2009_National Risk Factor Survey_v01_M ARG_2009_NRFS _v01_M BEL_2009_EHIS_v01_M BEL_2008_EHIS_v01_M BGR_2009_EHIS_v01_M BGR_2008_EHIS_v01_M BRA_2006_National Demographic and Maternal and Child Health BRA_2006_NDMCHS_v01_M Survey_v01_M BRA_2008_National Household Sampling Survey_v01_M BRA_2008_NHSS_v01_M CHL_2009_National Health Survey_v01_M CHL_2009_NHS_v01_M CHL_2011_National Socioeconomic Characterization Survey_v01_M CHL_2011_NSCS_v01_M CRI_1999_Sexual and Reproductive Health Survey_v01_M CRI_1999_SRHS_v01_M CRI_2006_National Health Survey_v01_M CRI_2006_NHS_v01_M ESP_-1_EHIS2_v01_M ESP_2014_EHIS2_v01_M EST_2007_EHIS_v01_M EST_2006_EHIS_v01_M GBR_2014_EHIS2_v01_M GBR_2013_EHIS2_v01_M LVA_2015_EHIS2_v01_M LVA_2014_EHIS2_v01_M MEX_2000_National Health Survey_v01_M MEX_2000_NHNS_v01_M MEX_2006_National Health and Nutrition Survey_v01_M MEX_2006_NHNS_v01_M MEX_2012_National Health and Nutrition Survey_v01_M MEX_2012_NHNS_v01_M MLT_2015_EHIS2_v01_M MLT_2014_EHIS2_v01_M 3.2.2. Changes in indicator values due to changes in definitions For the 2019 HEFPI version, we substantively revised the definition of one of the 46 health equity indicators: pap smears. Following WHO guidelines (World Health Organization 2013), the 2018 version of the indicator was defined as the rate of women – preferably in the 30-49 age-group – who received a pap smear in the five years prior to the survey interview. In formulating its recommendations, WHO considers not only clinical evidence but also “financial, infrastructural, and other resources”, i.e. the recommendations reflect both need and feasibility. HEFPI’s main aim is to facilitate measurement of progress towards UHC, the definition of which – all people having access to the health services they need without suffering financial hardship – considers need only. For the 2019 HEFPI version, we therefore switch to a definition of the pap smear indicator which is based on 14 high-income country guidelines that arguably are more aligned with pure need. Specifically, we now use the rate of women in the 20-69 age group who had a pap smear in the three years prior to the survey interview. The three-year interval is used in the United States as well as in most countries in the European Union (International Agency for Research on Cancer 2017). There is more heterogeneity in the age-groups recommended for screening across high-income countries, but the guidelines typically recommend screening from a woman’s 20s to her 60s.12 Whenever the raw survey data permit, we therefore use the 20-69 age-range, which is also used by OECD (OECD 2018). We continue to apply the same method to standardize pap smear rates with different reference periods in the original survey data as for the 2018 HEFPI database.13 Nevertheless, the shortening of the reference period from five to three years and the extension of our preferred age- group from 30-49 to 20-69 years changes the values of all previously presented pap smear data points in the 2019 HEFPI version. The 2019 HEFPI database makes a more minor change to the definition of two additional indicators: skilled birth attendance and having 4 or more antenatal care checks. Here, we change the age-range14 of women for whom we compute the indicators from 18-49 to 15-49. Most surveys with maternal and child health modules like the DHS, MICS and Reproductive Health Surveys (RHS) use the 15-49 age-range to collect data on pregnancy outcomes. We nevertheless used the 18-49 age- 12 In the United States the recommended screening age-range is 21-64 and the commonest range in the European Union is 25-64, although only 28% of countries use this range (no European Union country guideline recommends ending screening at age 49) (International Agency for Research on Cancer 2017). 13 When a survey used a pap-smear utilization question with a reference period other than 3 years, we transform the survey utilization rates to a 3 year reference period using the formula for the probability of an event over multiple trials, 1 1 where is the percentage of women obtaining pap smears over the survey’s reference period (in years), and is our target reference period of 3 years. For surveys where the reference period is unspecified (“Have you ever had a pap smear/mammogram?”), we assume where and are the upper and lower bounds, respectively, of the age-group for which the survey question is asked (e.g. 65 and 21). 14 Age-ranges always refer to the age at the time of the survey interview (i.e. not the age during which the pregnancy the woman is asked about occurred). 15 range in the 2018 HEFPI version to achieve definitory consistency between these survey families and the World Health Survey (WHS) that only collected pregnancy outcome data for women 18 and older. However, as the number of surveys in HEFPI that have maternal and child health data available from age 15 increases (e.g. by adding MICS 5 and 6 surveys) and as the WHS that was conducted in 2002-2004 becomes increasingly dated, the 2019 HEFPI database switches to the 15-49 age-group whenever the raw survey data permit it. Users of the 2019 HEFPI version’s skilled birth attendance and antenatal care use data should be mindful of the new age-group inconsistency in the WHS data points that continue to come from women age 18-49. All other indicator definitions of the 2018 HEFPI database remain the same.15 Table 3 shows the newly changed definitions of the pap smear, skilled birth attendance, and antenatal care indicators. Table 3: HEFPI indicators with changed definitions Indicator Definition Main Data Source16 Pregnancies with 4 or more Percentage of most recent births in last two years with at least 4 DHS, MICS, antenatal care visits (% of total) antenatal care visits (women age 15-49 at the time of the survey) WHS Births attended by skilled health Percentage of most recent births in last 2 years attended by any DHS, MICS, staff (% of total) skilled health personnel (women age 15-49 at the time of the WHS survey). Definition of skilled varies by country and survey but always includes doctor, nurse, midwife and auxiliary midwife). Pap smear in last 3 years Percentage of women who received a pap smear in the last 3 years DHS, EHIS, (preferably age 20-69 but age groups may vary) STEPS, WHS 3.2.3. Changes in indicator values due to coding corrections Coding corrections led to changes in indicator values for a number of data points, even if their underlying data source remained unchanged. 15 They are presented in Tables 1-3 of Wagstaff et al. (2018). 16 Main data sources are survey families with 15 or more data points for an indicator. 16 A prominent example are the ANC data points from the WHS where a coding error caused the 2018 HEFPI version’s points to substantively understate the true rate of pregnancies with 4 or more ANC visits – in the case of Burkina Faso by over 14 percentage points.17 In some cases, we discovered that the omission of a questionnaire’s filter questions or errors in the coding of missing values in raw survey data sets caused mistakes when we computed indicator values from micro-data. These errors were sometimes also committed by the survey reports – for instance, due to an erroneous coding of missings in the Côte d’Ivoire 2006 MICS immunization variables, the survey report and the 2018 HEFPI data point overstate the true full immunization rate by more than 15 percentage points. For the 2019 version of HEFPI, we have corrected such errors whenever we were confident that we could recover the true indicator values from the raw data. Moreover, for the 2019 HEFPI database, we reviewed all MICS data points. We corrected coding errors, and, among other things, better aligned the definitions of “skilled birth attendants”, “modern contraceptives”, and “formal health care providers” (to treat child acute respiratory infections) with those used for points from DHS and other maternal and child health surveys. In total, coding corrections led to changes of 10 percent or more relative to 2018 HEFPI indicator value levels for 152 data points for which sources remained unchanged. These data points 17Another survey-family-wide correction in the 2019 HEFPI version is that we now follow the official DHS (StatCompiler) method to compute rates of stunting and underweight from DHS data by (a) using all children sampled in the anthropometry module and (b) applying anthropometry module sampling weights. For the 2018 database, we had limited our analysis to children whose mothers were present in the household and applied the ordinary DHS child sample weights. The changes make only small differences in practice, as all changes in rates of stunting and underweight are smaller than one percentage point. 17 are listed in Appendix Table A1. The median change among points with coding corrections was 20 percent.18 3.3. New data points The 2019 version of the HEFPI database includes 535 newly added data points. Most new points come from surveys which were not included in the 2018 HEFPI database, but there are also a number of new points from previously included surveys. For instance, we added mammography data points from the previously included Jordan 2002 and 2007 DHS, and the child stunting and underweight data points from the previously included Myanmar 2000 and Albania 2005 MICS. Most newly added surveys come from the MICS family, and there mainly from the addition of all currently available fifth and sixth round surveys. We also added a number MICS surveys from earlier rounds that were missing from the 2018 HEFPI version, namely the round two surveys from Bosnia and Herzegovina (2000), Equatorial Guinea (2000), Indonesia (2000), Iraq (2000), Madagascar (2000), South Sudan (2000), Trinidad and Tobago (2000), Uzbekistan (2000) and Zambia (1999); the round three survey from Jamaica (2005), and the round four surveys from Bangladesh (2012), Saint Lucia (2012), South Sudan (2010), and Trinidad and Tobago (2011). With these additions, we have included all MICS surveys from the second to the fifth round which are nationally representative and where the micro-data are publicly available.19 We also added points from newly available surveys of the seventh round of the DHS, namely Albania (2017), Benin (2017), Burundi (2016), Haiti (2016), Jordan (2017), Malawi (2016), Maldives 18 Appendix Table A1 and our calculation of the median change exclude the pap smear indicator, as all values of the indicator changed between the 2018 and 2019 database versions due the change in the indicator’s reference period from three to two years. 19 Due to problems with the structure and labeling of the raw micro-data, we could not include the second-round Botswana (2000) and the fourth-round Guinea-Bissau (2014) surveys. 18 (2016), Philippines (2017), Senegal (2016, 2017), South Africa (2016), Tajikistan (2017), and Timor- Leste (2016).20 Newly added country-specific surveys on the health equity side of HEFPI include the 2015 Kenya Integrated Household Budget Survey (KIHBS) and the fourth round of the 2016 Malawi Integrated Household Survey (IHS). 3.4. Subnational data points Generally, HEFPI data points come from nationally representative surveys. There are, however, a number of exceptions: because data on the prevalence and treatment of non- communicable diseases from low- and middle-income countries are extremely scarce, we include data points from subnational STEPS surveys for indicators where no data point (from any year) is available from a nationally representative survey. Most of these subnational surveys sample capital cities where epidemiological and healthcare use patterns likely differ from those of the general population – a caveat HEFPI users should bear in mind. Table 4 lists the subnational surveys on the health equity side of HEFPI, together with the number of data points we source from them, and the subnational area they are representative of. 20So far, we have only included the following indicators from these new DHS surveys: 4+ ANC visits, skilled birth attendance, full child immunization, receipt of oral rehydration salts for children under five with diarrhea, and formal health care provider visits for children under five with acute respiratory infections. 19 Table 4: Subnational surveys on the health equity side of HEFPI # of HEFPI survey identifier Country Year Subnational area sampled data points CAF_2010_STEPS_v01_FS Central African Rep. 2010 Bangui 12 COD_2005_STEPS_v01_FS Congo, Dem. Rep. 2005 Ville-Province de Kinshasa 12 COL_2010_STEPS_v01_FS Colombia 2010 Department of Santander 15 GAB_2009_STEPS_v01_FS Gabon 2009 Libreville and Owendo 12 GHA_2006_STEPS_v01_FS Ghana 2006 Greater Accra Region 3 GIN_2009_STEPS_v01_FS Guinea 2009 Conakry and Basse Guinée 13 IDN_2003_STEPS_v01_FS Indonesia 2003 Abadijaya, Depok municipality 6 IDN_2006_STEPS_v01_FS Indonesia 2006 Depok 5 KNA_2007_STEPS_v01_FS St. Kitts and Nevis 2007 Saint Kitts 9 LAO_2008_STEPS_v01_FS Lao PDR 2008 Vientiane 3 MDG_2005_STEPS_v01_FS Madagascar 2005 Antananarivo and Toliara 12 MDV_2004_STEPS_v01_FS Maldives 2004 Malé 8 MDV_2011_STEPS_v01_FS Maldives 2011 Malé 12 District of Bamako, Commune MLI_2007_STEPS_v01_FS Mali 2007 of Kati Central, and Commune 3 of Ouéléssebougou MRT_2006_STEPS_v01_FS Mauritania 2006 Nouakchott 4 OMN_2006_STEPS_v01_FS Oman 2006 Sur City 7 SDN_2005_STEPS_v01_FS Sudan 2005 Khartoum state 10 TCD_2008_STEPS_v01_FS Chad 2008 N’Djaména 4 ZMB_2008_STEPS_v01_FS Zambia 2008 Lusaka 5 4. Conclusion The HEFPI database will continue to evolve. New datapoints will continue to be added, through a mix of adding new indicators and additional datapoints to already included indicators. The process of improving the reliability of estimates will also continue, meaning that some existing datapoints may be revised or dropped in future versions. Documentation (like this working paper) will be released to accompany new versions of the HEFPI database. 20 References International Agency for Research on Cancer (2017). Cancer Screening in the European Union. Report on the implementation of the Council Recommendation on Cancer Screening. Lyon, France, International Agency for Research on Cancer. Jolliffe, D. and E. B. Prydz (2016). "Estimating International Poverty Lines from Comparable National Thresholds." Journal of Economic Inequality 14 2: 185-198. Jolliffe, D. M. and E. B. Prydz (2017). Societal poverty: a relative and relevant measure. The World Bank, Policy Research Working Paper Series, 8073. Kakwani, N., A. Wagstaff and E. Van Doorslaer (1997). "Socioeconomic inequalities in health: Measurement, computation and statistical inference." Journal of Econometrics 77(1): 87-104. Lu, C., B. Chin, G. Li and C. J. Murray (2009). "Limitations of methods for measuring out-of-pocket and catastrophic private health expenditures." Bulletin of the World Health Organization 87(3): 238-244, 244A-244D. OECD (2018). Health at a Glance: Europe 2018: State of Health in the EU Cycle. Paris. van Doorslaer, E., O. O'Donnell, R. P. Rannan-Eliya, A. Somanathan, S. R. Adhikari, C. C. Garg, D. Harbianto, A. N. Herrin, M. N. Huq, S. Ibragimova, A. Karan, T. J. Lee, G. M. Leung, J. F. Lu, C. W. Ng, B. R. Pande, R. Racelis, S. Tao, K. Tin, K. Tisayaticom, L. Trisnantoro, C. Vasavid and Y. Zhao (2007). "Catastrophic payments for health care in Asia." Health Econ 16(11): 1159-1184. Wagstaff, A., P. Paci and E. Van Doorslaer (1991). "On the Measurement of Inequalities in Health." Social Science & Medicine 33(5): 545-557. Wagstaff, A., P. Eozenou, S. Neelsen and M. Smitz (2018). The 2018 Health Equity and Financial Protection Indicators Database: Overview and Insights. World Bank Policy Research Working Paper 8577, The World Bank. Wagstaff, A., P. H.-V. Eozenou and M.-F. Smitz (2019). Out-of-Pocket Expenditures on Health : A Global Stocktake. Policy Research Working Paper 8808. Washington, D.C. World Health Organization (2013). WHO guidelines for screening and treatment of precancerous lesions for cervical cancer prevention, World Health Organization. 21 Appendix Table A1: Data points with changes of more than 10 percent due to coding corrections ̅ Indicator Country Year 2019 survey identifier ̅ ̅ a_inpatient_1yr Philippines 2008 PHL_2008_DHS_v01_M 0.05 -0.01 a_inpatient_1yr Philippines 2013 PHL_2013_DHS_v01_M 0.05 0.01 a_overweight Tonga 2012 TON_2012_STEPS_v01_FS 0.91 0.68 Trinidad and a_overweight Tobago 2011 TTO_2011_STEPS_v01_FS 0.56 0.19 c_ITN Mozambique 2008 MOZ_2008_MICS_v01_M 0.19 -0.05 c_ITN Suriname 2006 SUR_2006_MICS_v01_M 0.09 -0.41 c_anc Bangladesh 2003 BGD_2003_WHS_v01_M 0.16 0.06 c_anc Burkina Faso 2002 BFA_2002_WHS_v01_M 0.22 0.14 c_anc Chad 2003 TCD_2003_WHS_v01_M 0.23 0.08 c_anc Guatemala 2003 GTM_2003_WHS_v01_M 0.66 0.12 c_anc India 2003 IND_2003_WHS_v01_M 0.27 0.10 c_anc Lao PDR 2003 LAO_2003_WHS_v01_M 0.20 0.06 c_anc Mauritania 2003 MRT_2003_WHS_v01_M 0.28 0.11 c_anc Myanmar 2003 MMR_2003_WHS_v01_M 0.56 0.08 c_anc Nepal 2003 NPL_2003_WHS_v01_M 0.24 0.07 c_anc Pakistan 2003 PAK_2003_WHS_v01_M 0.20 0.07 c_anc Paraguay 2002 PRY_2002_WHS_v01_M 0.71 0.08 c_anc Senegal 2003 SEN_2003_WHS_v01_M 0.42 0.23 Swaziland c_anc (Eswatini) 2003 SWZ_2003_WHS_v01_M 0.66 0.18 c_anc Tunisia 2003 TUN_2003_WHS_v01_M 0.71 0.25 c_anc Zambia 2003 ZMB_2003_WHS_v01_M 0.74 0.18 c_anc Zimbabwe 2003 ZWE_2003_WHS_v01_M 0.73 0.17 Bosnia and c_fullimm Herzegovina 2006 BIH_2006_MICS_v01_M 0.72 -0.16 c_fullimm Burundi 2005 BDI_2005_MICS_v01_M 0.56 0.08 c_fullimm Cameroon 2006 CMR_2006_MICS_v01_M 0.59 -0.11 Central African c_fullimm Republic 2006 CAF_2006_MICS_v01_M 0.26 -0.03 c_fullimm Congo, Dem. Rep. 2010 COD_2010_MICS_v01_M 0.49 -0.05 c_fullimm Côte d'Ivoire 2006 CIV_2006_MICS_v01_M 0.58 -0.16 c_fullimm Georgia 2005 GEO_2005_MICS_v01_M 0.39 0.14 c_fullimm Guyana 2006 GUY_2006_MICS_v01_M 0.43 -0.31 c_fullimm Kazakhstan 2006 KAZ_2006_MICS_v01_M 0.77 -0.18 c_fullimm Kenya 2000 KEN_2000_MICS_v01_M 0.55 -0.07 c_fullimm Lao PDR 2000 LAO_2000_MICS_v01_M 0.26 0.07 c_fullimm Mauritania 2007 MRT_2007_MICS_v01_M 0.21 -0.07 c_fullimm Mauritania 2011 MRT_2011_MICS_v01_M 0.23 -0.04 c_fullimm Moldova 2000 MDA_2000_MICS_v01_M 0.67 -0.22 c_fullimm Myanmar 2000 MMR_2000_MICS_v01_M 0.51 -0.31 c_fullimm Sierra Leone 2000 SLE_2000_MICS_v01_M 0.36 -0.06 c_fullimm Somalia 2006 SOM_2006_MICS_v01_M 0.08 -0.02 c_fullimm Tajikistan 2000 TJK_2000_MICS_v01_M 0.58 -0.15 c_measles_vacc Afghanistan 2010 AFG_2010_MICS_v01_M 0.39 -0.08 Bosnia and c_measles_vacc Herzegovina 2006 BIH_2006_MICS_v01_M 0.75 -0.12 c_measles_vacc Cameroon 2006 CMR_2006_MICS_v01_M 0.81 -0.08 c_measles_vacc Chad 2003 TCD_2003_WHS_v01_M 0.39 0.12 c_measles_vacc Côte d'Ivoire 2006 CIV_2006_MICS_v01_M 0.72 -0.17 c_measles_vacc Ecuador 2003 ECU_2003_WHS_v01_M 0.55 0.09 c_measles_vacc Guatemala 2003 GTM_2003_WHS_v01_M 0.53 0.06 c_measles_vacc Guyana 2006 GUY_2006_MICS_v01_M 0.86 -0.10 c_measles_vacc India 2003 IND_2003_WHS_v01_M 0.51 0.07 22 ̅ Indicator Country Year 2019 survey identifier ̅ ̅ c_measles_vacc Kazakhstan 2006 KAZ_2006_MICS_v01_M 0.83 -0.17 c_measles_vacc Lao PDR 2000 LAO_2000_MICS_v01_M 0.44 0.13 c_measles_vacc Mali 2003 MLI_2003_WHS_v01_M 0.34 -0.04 c_measles_vacc Mauritania 2007 MRT_2007_MICS_v01_M 0.71 -0.09 c_measles_vacc Myanmar 2000 MMR_2000_MICS_v01_M 0.53 -0.35 c_measles_vacc Pakistan 2003 PAK_2003_WHS_v01_M 0.37 0.04 c_measles_vacc Sri Lanka 2003 LKA_2003_WHS_v01_M 0.87 0.15 Swaziland c_measles_vacc (Eswatini) 2003 SWZ_2003_WHS_v01_M 0.58 -0.06 c_measles_vacc Tunisia 2003 TUN_2003_WHS_v01_M 0.86 0.19 c_measles_vacc Zimbabwe 2009 ZWE_2009_MICS_v01_M 0.80 -0.11 c_sba Burkina Faso 2003 BFA_2003_DHS_v01_M 0.56 0.18 c_sba Guatemala 2014 GTM_2014_DHS_v01_M 0.69 -0.28 c_sba Mali 2009 MLI_2009_MICS_v01_M 0.58 0.28 c_sba Niger 2000 NER_2000_MICS_v01_M 0.15 -0.09 c_sba Togo 2006 TGO_2006_MICS_v01_M 0.52 -0.17 Bosnia and c_stunted Herzegovina 2006 BIH_2006_MICS_v01_M 0.12 0.11 c_stunted Burkina Faso 2006 BFA_2006_MICS_v01_M 0.43 0.05 c_stunted Congo, Dem. Rep. 2010 COD_2010_MICS_v01_M 0.43 0.05 c_stunted Guinea-Bissau 2000 GNB_2000_MICS_v01_M 0.36 0.04 c_stunted Mauritania 2007 MRT_2007_MICS_v01_M 0.27 -0.04 c_stunted Nigeria 2007 NGA_2007_MICS_v01_M 0.42 0.05 c_stunted Senegal 2000 SEN_2000_MICS_v01_M 0.30 0.05 c_stunted Sierra Leone 2000 SLE_2000_MICS_v01_M 0.38 0.06 c_stunted Sierra Leone 2005 SLE_2005_MICS_v01_M 0.47 0.05 c_stunted Sudan 2010 SDN_2010_MICS_v01_M 0.34 0.07 c_stunted Togo 2006 TGO_2006_MICS_v01_M 0.31 0.04 c_treatARI Bolivia 2000 BOL_2000_MICS_v01_M 0.54 0.07 c_treatARI Kyrgyz Republic 2005 KGZ_2005_MICS_v01_M 0.58 -0.07 c_treatARI Myanmar 2000 MMR_2000_MICS_v01_M 0.49 0.14 c_treatARI Nigeria 2011 NGA_2011_MICS_v01_M 0.35 -0.05 c_treatARI Sudan 2010 SDN_2010_MICS_v01_M 0.55 0.08 c_treatARI Yemen, Rep. 2006 YEM_2006_MICS_v01_M 0.68 0.27 c_treatdiarrhea Algeria 2012 DZA_2012_MICS_v01_M 0.25 -0.09 c_treatdiarrhea Belize 2011 BLZ_2011_MICS_v01_M 0.56 0.32 c_treatdiarrhea Cuba 2010 CUB_2010_MICS_v01_M 0.53 0.07 c_treatdiarrhea El Salvador 2008 SLV_2008_RHS_v01_M 0.53 0.45 c_treatdiarrhea Iraq 2006 IRQ_2006_MICS_v01_M 0.31 -0.21 c_treatdiarrhea Mali 2009 MLI_2009_MICS_v01_M 0.11 -0.10 c_treatdiarrhea Mauritania 2011 MRT_2011_MICS_v01_M 0.25 -0.14 c_treatdiarrhea Serbia 2005 SRB_2005_MICS_v01_M 0.17 -0.63 c_treatdiarrhea Sudan 2010 SDN_2010_MICS_v01_M 0.22 -0.18 c_treatdiarrhea Thailand 2012 THA_2012_MICS_v01_M 0.58 -0.06 c_treatdiarrhea Venezuela, RB 2000 VEN_2000_MICS_v01_M 0.38 0.34 c_treatdiarrhea Vietnam 2006 VNM_2006_MICS_v01_M 0.27 -0.35 c_treatdiarrhea Zimbabwe 2009 ZWE_2009_MICS_v01_M 0.09 -0.29 c_underweight Belarus 2005 BLR_2005_MICS_v01_M 0.02 0.00 c_underweight Burkina Faso 2006 BFA_2006_MICS_v01_M 0.38 0.05 Central African c_underweight Republic 2006 CAF_2006_MICS_v01_M 0.27 0.03 c_underweight Chad 2010 TCD_2010_MICS_v01_M 0.30 0.03 c_underweight Dominican Republic 2000 DOM_2000_MICS_v01_M 0.04 0.00 c_underweight Guinea-Bissau 2000 GNB_2000_MICS_v01_M 0.22 0.03 c_underweight Kyrgyz Republic 2005 KGZ_2005_MICS_v01_M 0.03 0.00 c_underweight Mauritania 2007 MRT_2007_MICS_v01_M 0.22 -0.03 c_underweight Nigeria 2007 NGA_2007_MICS_v01_M 0.26 0.03 c_underweight Senegal 2000 SEN_2000_MICS_v01_M 0.21 0.04 c_underweight Sierra Leone 2000 SLE_2000_MICS_v01_M 0.24 0.03 c_underweight Sierra Leone 2005 SLE_2005_MICS_v01_M 0.28 0.03 c_underweight Sudan 2010 SDN_2010_MICS_v01_M 0.28 0.06 23 ̅ Indicator Country Year 2019 survey identifier ̅ ̅ c_underweight Togo 2006 TGO_2006_MICS_v01_M 0.25 0.03 m_overweight Tonga 2012 TON_2012_STEPS_v01_FS 0.87 0.57 w_CPR Afghanistan 2010 AFG_2010_MICS_v01_M 0.26 0.06 w_CPR Algeria 2012 DZA_2012_MICS_v01_M 0.56 0.07 w_CPR Burkina Faso 2006 BFA_2006_MICS_v01_M 0.16 0.02 w_CPR Burundi 2005 BDI_2005_MICS_v01_M 0.09 0.02 w_CPR Cameroon 2000 CMR_2000_MICS_v01_M 0.13 -0.03 w_CPR Cameroon 2006 CMR_2006_MICS_v01_M 0.17 0.03 Central African w_CPR Republic 2006 CAF_2006_MICS_v01_M 0.13 0.04 Central African w_CPR Republic 2010 CAF_2010_MICS_v01_M 0.15 0.04 w_CPR Chad 2010 TCD_2010_MICS_v01_M 0.05 0.01 w_CPR Congo, Dem. Rep. 2001 COD_2001_MICS_v01_M 0.12 0.07 w_CPR Congo, Dem. Rep. 2010 COD_2010_MICS_v01_M 0.09 0.01 w_CPR Djibouti 2006 DJI_2006_MICS_v01_M 0.20 0.02 w_CPR Gambia, The 2000 GMB_2000_MICS_v01_M 0.00 0.00 w_CPR Ghana 2006 GHA_2006_MICS_v01_M 0.16 0.02 w_CPR Ghana 2011 GHA_2011_MICS_v01_M 0.28 0.04 w_CPR Guinea-Bissau 2000 GNB_2000_MICS_v01_M 0.05 0.00 w_CPR Guinea-Bissau 2006 GNB_2006_MICS_v01_M 0.11 0.02 w_CPR Guyana 2006 GUY_2006_MICS_v01_M 0.36 0.04 w_CPR Iraq 2006 IRQ_2006_MICS_v01_M 0.47 0.07 w_CPR Iraq 2011 IRQ_2011_MICS_v01_M 0.42 0.06 w_CPR Lesotho 2000 LSO_2000_MICS_v01_M 0.32 0.20 w_CPR Malawi 2006 MWI_2006_MICS_v01_M 0.45 0.07 w_CPR Mali 2009 MLI_2009_MICS_v01_M 0.11 0.02 w_CPR Mauritania 2011 MRT_2011_MICS_v01_M 0.12 0.07 w_CPR Mozambique 2008 MOZ_2008_MICS_v01_M 0.17 0.04 w_CPR Nigeria 2007 NGA_2007_MICS_v01_M 0.12 0.02 w_CPR Nigeria 2011 NGA_2011_MICS_v01_M 0.14 0.03 w_CPR Sierra Leone 2000 SLE_2000_MICS_v01_M 0.05 -0.02 w_CPR Sierra Leone 2005 SLE_2005_MICS_v01_M 0.06 0.01 w_CPR Sierra Leone 2010 SLE_2010_MICS_v01_M 0.12 0.02 w_CPR Sudan 2010 SDN_2010_MICS_v01_M 0.10 0.01 w_CPR Suriname 1999 SUR_1999_MICS_v01_M 0.44 -0.09 Syrian Arab w_CPR Republic 2006 SYR_2006_MICS_v01_M 0.53 0.06 w_CPR Togo 2006 TGO_2006_MICS_v01_M 0.13 0.02 w_CPR Togo 2010 TGO_2010_MICS_v01_M 0.15 0.02 West Bank and w_CPR Gaza 2010 PSE_2010_MICS_v01_M 0.51 0.09 w_CPR Yemen, Rep. 2006 YEM_2006_MICS_v01_M 0.30 0.05 w_CPR Zimbabwe 2009 ZWE_2009_MICS_v01_M 0.72 0.08 w_mam_2y Dominican Republic 1996 DOM_1996_DHS_v01_M 0.10 -0.06 w_mam_2y Dominican Republic 2013 DOM_2013_DHS_v01_M 0.35 -0.23 w_mam_2y Nicaragua 1997 NIC_1997_DHS_v01_M 0.06 -0.06 w_mam_2y Nicaragua 2001 NIC_2001_DHS_v01_M 0.09 -0.57 w_overweight Tonga 2012 TON_2012_STEPS_v01_FS 0.94 0.78