Abstract

The 2019 Pakistan Maternal Mortality Survey (2019 PMMS) was the first stand-alone maternal mortality survey conducted in Pakistan. A nationally representative sample of 1,396 primary sampling units were randomly selected. The survey was expected to result in about 14,000 interviews with ever-married women age 15-49.

The primary objective of the 2019 PMMS is to provide up-to-date estimates of basic demographic and health indicators. Specifically, the survey was designed and carried out with the purpose of assessing where Pakistan stands on maternal health indicators and how well the country is moving toward these targets. Overall aims of the 2019 PMMS were as follows: - To estimate national and regional levels of maternal mortality for the 3 years preceding the survey and determine whether the MMR has declined substantially since 2006-07 - To identify medical causes of maternal deaths and the biological and sociodemographic risk factors associated with maternal mortality - To assess the impact of maternal and newborn health services, including antenatal and postnatal care and skilled birth attendance, on prevention of maternal mortality and morbidity - To estimate the prevalence and determinants of common obstetric complications and morbidities among women of reproductive age during the 3 years preceding the survey

Geographic coverage

National coverage

Analysis unit

• Household
• Individual
• Woman age 15-49
• Community

Kind of data

Sample survey data [ssd]

Sampling procedure

The 2019 PMMS used a multistage and multiphase cluster sampling methodology based on updated sampling frames derived from the 6th Population and Housing Census, which was conducted in 2017 by the Pakistan Bureau of Statistics (PBS). The sampling universe consisted of urban and rural areas of the four provinces of Pakistan (Punjab, Sindh, Khyber Pakhtunkhwa, and Balochistan), Azad Jammu and Kashmir (AJK), Gilgit Baltistan (GB), Federally Administered Tribal Areas (FATA), and the Islamabad Capital Territory (ICT). A total of 153,560 households (81,400 rural and 72,160 urban) were selected using a two-stage and two-phase stratified systematic sampling approach. The survey was designed to provide representative results for most of the survey indicators in 11 domains: four provinces (by urban and rural areas with Islamabad combined with Punjab and FATA combined with Khyber Pakhtunkhwa), Azad Jammu and Kashmir (urban and rural), and Gilgit Baltistan. Restricted military and protected areas were excluded from the sample.

The sampled households were randomly selected from 1,396 primary sampling units (PSUs) (740 rural and 656 urban) after a complete household listing. In each PSU, 110 randomly selected households were administered the various questionnaires included in the survey. All 110 households in each PSU were asked about births and deaths during the previous 3 years, including deaths among women of reproductive age (15-49 years). Households that reported at least one death of a woman of reproductive age were then visited, and detailed verbal autopsies were conducted to determine the causes and circumstances of these deaths to help identify maternal deaths. In the second phase, 10 households in each PSU were randomly selected from the 110 households selected in the first phase to gather detailed information on women of reproductive age. All eligible ever-married women age 15-49 residing in these 10 households were interviewed to gather detailed information, including a complete pregnancy history.

Note: A detailed description of the sample design is provided in Appendix A of the final report.

Mode of data collection

Face-to-face [f2f]

Research instrument

Six questionnaires were used in the 2019 PMMS: the Short Household Questionnaire, the Long Household Questionnaire, the Woman’s Questionnaire, the Verbal Autopsy Questionnaire, the Community Questionnaire, and the Fieldworker Questionnaire. A Technical Advisory Committee was established to solicit comments on the questionnaires from various stakeholders, including representatives of government ministries and agencies, nongovernmental organisations, and international donors. The survey protocol was reviewed and approved by the National Bioethics Committee, the Pakistan Health Research Council, and the ICF Institutional Review Board. After being finalised in English, the questionnaires were translated into Urdu and Sindhi. The 2019 PMMS used paper-based questionnaires for data collection, while computer-assisted field editing (CAFE) was used to edit questionnaires in the field.

Cleaning operations

The processing of the 2019 PMMS data began simultaneously with the fieldwork. As soon as data collection was completed in each cluster, all electronic data files were transferred via the Internet File Streaming System (IFSS) to the NIPS central office in Islamabad. These data files were registered and checked for inconsistencies, incompleteness, and outliers. A double entry procedure was adopted by NIPS to ensure data accuracy. The field teams were alerted about any inconsistencies and errors. Secondary editing of completed questionnaires, which involved resolving inconsistencies and coding open-ended questions, was carried out in the central office. The survey core team members assisted with secondary editing, and the NIPS data processing manager coordinated the work at the central office. Data entry and editing were carried out using the CSPro software package. The concurrent processing of the data offered a distinct advantage because it maximised the likelihood of the data being error-free and accurate.

Response rate

In the four provinces, the sample contained a total of 116,169 households. All households were visited by the field teams, and 110,483 households were found to be occupied. Of these households, 108,766 were successfully interviewed, yielding a household response rate of 98%. The subsample selected for the Long Household Questionnaire comprised 11,080 households, and interviews were carried out in 10,479 of these households. A total of 12,217 ever-married women age 15-49 were eligible to be interviewed based on the Long Household Questionnaire, and 11,859 of these women were successfully interviewed (a response rate of 97%).

In Azad Jammu and Kashmir, 16,755 households were occupied, and interviews were successfully carried out in 16,588 of these households (99%). A total of 1,707 ever-married women were eligible for individual interviews, of whom 1,666 were successfully interviewed (98%). In Gilgit Baltistan, 11,005 households were occupied, and interviews were conducted in 10,872 households (99%). A total of 1,219 ever-married women were eligible for interviews, of whom 1,178 were successfully interviewed (97%).

A total of 944 verbal autopsy interviews were conducted in Pakistan overall, 150 in Azad Jammu and Kashmir, and 88 in Gilgit Baltistan. The Verbal Autopsy Questionnaire was used in almost all of the interviews, and response rates were nearly 100%.

Sampling error estimates

The estimates from a sample survey are affected by two types of errors: nonsampling errors and sampling errors. Nonsampling errors are the results of mistakes made in implementing data collection and data processing, such as failure to locate and interview the correct household, misunderstanding of the questions on the part of either the interviewer or the respondent, and data entry errors. Although numerous efforts were made during the implementation of the 2019 Pakistan Maternal Mortality Survey (2019 PMMS) to minimise this type of error, nonsampling errors are impossible to avoid and difficult to evaluate statistically.

Sampling errors, on the other hand, can be evaluated statistically. The sample of respondents selected in the 2019 PMMS is only one of many samples that could have been selected from the same population, using the same design and sample size. Each of these samples would yield results that differ somewhat from the results of the actual sample selected. Sampling errors are a measure of the variability among all possible samples. Although the degree of variability is not known exactly, it can be estimated from the survey results.

Sampling error is usually measured in terms of the standard error for a particular statistic (mean, percentage, etc.), which is the square root of the variance. The standard error can be used to calculate confidence intervals within which the true value for the population can reasonably be assumed to fall. For example, for any given statistic calculated from a sample survey, the value of that statistic will fall within a range of plus or minus two times the standard error of that statistic in 95% of all possible samples of identical size and design.

If the sample of respondents had been selected by simple random sampling, it would have been possible to use straightforward formulas for calculating sampling errors. However, the 2019 PMMS sample was the result of a multi-stage stratified design, and, consequently, it was necessary to use more complex formulas. Sampling errors are computed using SAS programmes developed by ICF. These programmes use the Taylor linearisation method to estimate variances for survey estimates that are means, proportions, or ratios and use the Jackknife repeated replication method for variance estimation of more complex statistics such as fertility and mortality rates.

A more detailed description of estimates of sampling errors are presented in Appendix B of the survey report.

Data appraisal

Data Quality Tables

- Household age distribution