Matrika Household Survey in India - User Guide

Data Description:

Data produced as part of a study to evaluate the impact of the Matrika social franchising model – a multi-faceted intervention that established a network of private providers and strengthened the skills of both public and private sector clinicians – and determine whether it has improved the quality and coverage of health services along the continuum of care for maternal, newborn and reproductive health in Uttar Pradesh, India.  The datasets cover two rounds of a household survey, performed in January 2015 and May 2016, of women who had recently given birth.

Data Collection Methods:

Data were collected during household interviews using CAPI (computer assisted personal interviewing) by staff of Sambodhi Research and Communications. The household survey was administered to women as a repeated cross-section in January 2015 (round 1) and May 2016 (round 2). Women were selected from the same clusters in each round. Eligible respondents included all women who gave birth in the previous 24 months (round 1) or 18 months (round 2), including those who had a stillbirth or whose child died since birth. Eligible women were identified through a census of households, conducted one month before the household survey round. Every member of the household was listed and then, for women aged 15 to 49 years, a series of questions probed whether she gave birth to a baby that was born alive, born dead or lost before birth. Using this sampling frame, eligible women in each cluster were randomly selected for interview. The household survey tool included the following modules: (1) household listing, (2) general healthcare interactions, (3) household characteristics, (4) wellbeing of husband, (5) pregnancy history, (6) family planning and antenatal care, (7) delivery and postnatal care, (8) child health, (9) interactions with community health workers, (10) information and perceptions of healthcare, and (11) wellbeing, mental health and physical health.

Sampling strategy of clusters:

The study involved the selection of three types of clusters: 1) ‘intervention’ clusters with a Sky provider; 2) ‘internal comparison’ clusters with no social franchisee in the three intervention districts; 3) ‘external comparison’ clusters in three neighbouring districts where the social franchise model was not operating. Study clusters were selected one year after the first social franchisee was contracted using the following procedures. First, every Sky health provider was linked to its census area. At the time of selection there were 393 private providers in the network. This process identified 216 possible intervention clusters from which 60 clusters were selected at random. Second, internal comparison clusters were selected by matching without replacement the intervention clusters to 60 comparison areas within the same three districts. We performed exact matching on district and urban residences, and then within each strata, selected pairs of clusters with the smallest distance based on a Mahalanobis metric that was computed using census data on village characteristics (total population, % under 6 years, % females under 6 years, % female literate females, % scheduled tribe, % scheduled caste, % cultivator, and % “other” workers). To limit problems of contamination, comparison clusters adjacent to intervention areas could not be selected. Finally, the same matching procedure was performed to select 60 external comparison clusters from neighbouring districts.

Data Analysis and Preparation:

Personally identifiable information removed. Data analysed for research purposes.

Geographic regions:

Data were collected from eligible women in 180 clusters (defined as a ward (urban) or a village (rural) according to the most recent census) across 6 districts (Kannuaj, Kanpur Nagar, Kanpur Dehat, Auraiya, Etawah, and Fatehpur) of Uttar Pradesh, India.

Key dates:

• Round 1 collection - January 2015
• Round 2 collection - May 2016.

Species:

Human population

Privacy:

Personally identifiable data has been removed, however there remains a risk of indirect identification through the large number of variables contained.

Ethics:

Ethics approval for study obtained from LSHTM Ethics Committee in London, and from the Public Healthcare Society (PHS) Ethics Review Board and the Indian Council of Medical Research in India.

Keywords:

Maternal health, Newborn health, Social franchising

Language of written material:

English, Hindi

Project title:

Evaluation of the Matrika Social Franchising Programme in Uttar Pradesh, India

Funder/Sponsor:

Data collection supported by funding from MSD, through its MSD for Mothers programme (called Merck for Mothers in the US and Canada). MSD had no role in the design, collection, analysis or interpretation of data. MSD for Mothers is an initiative of Merck & Co., Inc., Kenilworth, N.J., U.S.A.

Additional Information:

Project collaboration between LSHTM and Sambodhi Research and Communications

Data Creators

Researchers are part of the Maternal healthcare markets Evaluation Team (MET) research group at LSHTM (http://met-lshtm.com/).

File Description