The Matlab Health and Socioeconomic Survey (Bangladesh)

The Matlab Health and Socioeconomic Survey, or MHSS, was carried out in the Matlab region of rural Bangladesh in 1996. Matlab is notable for its ongoing prospective Demographic Surveillance System.[1] The ultimate objective of the MHSS effort was to enter into the public domain a new and unique microlevel data set for research on aging. In particular, these new data will support in-depth analyses--not possible with existing survey data--on interrelated topics pertaining to life-cycle investments in the physical, economic, and social well-being of adults and the elderly.

The MHSS addresses the following broad areas of concern to rural adults and the elderly: the effect of socioeconomic and behavioral factors on adult and elderly health status and on health care utilization; the linkages between adult/elderly well-being, social and kin network characteristics, and resource flows; and the impact of community services and infrastructure on adult/elderly health and schooling and training. The National Institute on Aging funded this survey under Grant P01AG11952. The principal investigator is Dr. Omar Rahman, M.D., D.Sc., Associate Professor of Epidemiology and Demography at the Harvard School of Public Health. The MHSS is a collaborative effort of RAND, the Harvard School of Public Health, the University of Pennsylvania, the University of Colorado at Boulder, Brown University, Mitra and Associates, and the International Centre for Diarrhoeal Disease Research, Bangladesh (ICDDR,B). The other co-investigators are Jane Menken, Andrew Foster, Christine Peterson, Mohammad Nizam Khan, Paul Gertler, and Randall Kuhn.

The MHSS actually comprises four separate and distinct surveys that have different samples and serve different analytic objectives:

• The Main survey, consisting of household- and individual-level information on 4,364 households clustered in 2,687 baris, or residential compounds, a random sample of approximately one-third of the total number of baris in the Surveillance area. This is likely to be the most useful component for the widest group of analysts. The Main sample data contain an additional 174 households clustered in 94 other baris. These households fell outside the prescribed sampling scheme and should be dropped from analyses. Thus, the total number of households in the Main sample data is 4,538.

• The Determinants of Natural Fertility Survey (DNFS), a specialized survey consisting of household- and individual-level information on a particular follow-up group of 1,790 households and 2,441 of the women who were originally interviewed about their health and pregnancy status in the mid-1970s. The available data on 1,806 DNFS women in these 1,790 households represent a sample of approximately 80 percent of the surviving 2,273 DNFS women.

• * The Outmigrant survey, another specialized survey consisting of household- and individual-specific information on 552 outmigrants who had left the households of the primary sample between 1982 and the date of the MHSS and who had not returned to their original households or baris. This represents an approximately 8-percent sample of the outmigrants since 1982.

• The Community/Provider survey, which gathered information on community infrastructure and services on the 141 constituent villages of the primary sample respondents and detailed data on 254 health/family planning providers, as well as data on 100 educational facilities potentially serving (in the opportunity set of) the primary sample households in the MHSS. This constitutes an almost complete census of schools and health and family planning clinics serving the study population, and a sample of individual health/family planning providers.

MHSS Household Questionnaire

---

The MHSS Household questionnaire consists of six sections. The contents of the survey are summarized in Table 3. Most of the contents are self-explanatory.

NOTE: Book IV is an additional set of questions given only to ever-married female Book III respondents (includes all ages). The child questionnaire of Book V is administered by proxy. Physical and cognitive tests were given to all Book III respondents; cognitive tests were also given to Book V respondents aged 6-14 years. Height and weight measurements were done for all Book III and Book V respondents.

The MHSS Provider Questionnaire

---

The MHSS Provider questionnaire had separate instruments for each of the nine types of providers (seven health providers and two schools). The health providers were Thana Health Complexes, MBBS physicians, Family Welfare Centers, traditional practitioners, village doctors, pharmacists, and trained/traditional birth attendants. The education providers were primary and secondary schools.

MHSS Health Provider Instruments

Instruments were administered to 254 health providers serving the 141 villages of the Main sample survey area. Although separate instruments were prepared in MHSS for each of the seven types of health providers, there are strong similarities across the instruments for three of the strata: (1) Thana Health Complexes and Family Welfare Centers, (2) MBBS doctors, village doctors, and pharmacists, and (3) the practices of traditional healers and trained/traditional birth attendants. The maintaining of distinct instruments recognizes that different types of providers offer different types of services. Of the Health Provider instruments, the Thana Health Complexes instrument was the most comprehensive. The individual practitioner modules were similar in content, although some modules were scaled back to reflect the fact that most individual practices offer a less elaborate array of services than do the Thana Health Complexes. With the exception of traditional healers and trained/traditional birth attendants, all providers were asked to answer the hypothetical patient vignette modules. Table 4 summarizes the modules used for health providers.

Education Provider Instruments

MHSS interviewers administered the Education Provider instruments to two levels of schools, primary and secondary (including public, private, and religious schools), serving the 141 villages of the Main sample survey area. The instruments were quite similar across school levels. Table 5 summarizes the Education Provider modules.

Community Questionnaire

---

Information on community characteristics for the 141 villages in the MHSS area for the Main sample was obtained from village leaders and other knowledgeable sources. Table 6 summarizes this questionnaire's content.

MHSS Data and Documentation

---

Each sample described above (Main, DNFS, Outmigrant, Provider, and Community) has its own set of data subfiles, as does each module. The Community and Provider samples have separate sets of subfiles for each provider type and for the Community Characteristics survey. Documentation on the MHSS is available in five separate volumes. All volumes must be referred to when working with the MHSS data:

DRU-2018/1-NIA, The 1996 Matlab Health and Socioeconomic Survey: Overview and User's Guide, O. Rahman, J. Menken, A. Foster, C. E. Peterson, M. N. Khan, R. Kuhn, P. Gertler, ICDDR,B, and Mitra and Associates, 1999.

Provides a detailed description of the MHSS, including response rates and data quality, as well as guidance on how to work with the subfile data structure to construct analysis files.

DRU-2018/2-NIA, The 1996 Matlab Health and Socioeconomic Survey: Questionnaires for Household Survey, O. Rahman, J. Menken, A. Foster, M. N. Khan, C. E. Peterson, R. Kuhn, P. Gertler, ICDDR,B, and Mitra and Associates, 1999.

Contains the MHSS Household survey instruments in English.

DRU-2018/3-NIA, The 1996 Matlab Health and Socioeconomic Survey: Questionnaires for Community/Provider Survey, O. Rahman, J. Menken, A. Foster, M. N. Khan, C. E. Peterson, R. Kuhn, P. Gertler, ICDDR,B, and Mitra and Associates, 1999.

Contains the MHSS Community/Provider survey instruments in English.

DRU-2018/4-NIA, The 1996 Matlab Health and Socioeconomic Survey: Codebook for Household Survey, C. E. Peterson, D. Wesley, N. Murray, 1999.

Contains the formal codebook for the MHSS Household survey data subfiles. The codebook describes data structure, data conventions, and all variables in the Household database.

DRU-2018/5-NIA, The 1996 Matlab Health and Socioeconomic Survey: Codebook for Community/Provider Survey, C. E. Peterson, D. Wesley, N. Murray, 1999.

Contains the formal codebook for the MHSS Community/Provider survey data subfiles. The codebook describes data structure, data conventions, and all variables in the Community/Provider database.

The MHSS data and documentation are available from RAND and will also be available from ICPSR through its NACDA public holdings. For details on how to get the data and documentation, please see the "Obtaining FLS Data" section of this newsletter.

Questions about the MHSS data may be directed to the following email address: mhss-supp@rand.org. Please acknowledge RAND and the other institutions involved in any publications based on the MHSS data. A citation to the following document suffices as a proper acknowledgment:

Rahman et al., The Matlab Health and Socioeconomic Survey: Overview and User's Guide, Santa Monica, CA: RAND, DRU-2018/1-NIA, 1999.

The MHSS project maintains a bibliography that will be available to interested users. Please send copies of any papers based on the MHSS data to:

Dr. Jane Menken
Department of Sociology
University of Colorado, Boulder
Institute of Behavioral Science
Campus Box 484
Boulder, CO 80309-0484