Impact evaluation of an intervention to reduce neonatal mortality in Kerala, India

Publication Details

Author
Shannon Maloney, Rohini Ghosh
Country
India
Region
South Asia
Sector
Health Nutrition and Population
Subsector
Health Sector Reform, Mortality, Preventive Health and Health Behavior
Gender analysis
No
Equity Focus
None specified
Evaluation design
Difference-in Difference (DID)
Status
Ongoing 3ie Funded Studies
3ie Funding Window
Policy Window Round 3

Synopsis

The project assesses the impact of new health standards introduced by the National Rural Health Mission (NRHM). These new standards concern neonatal health and care practices implemented across 91 government hospitals in 14 districts of Kerala, a state in south India.

Context

In Kerala, neonatal complications are extremely high leading to high infant mortality.  Structural changes within government hospitals and local delivery rooms, such as those imposed by NRHM, have the potential to reduce neonatal mortality. NRHM standards include provision of improved medical equipment, training of staff and access to medicines. The results from the study will help inform decision making to improve the quality of care in hospitals in Kerala.

Research questions

  1. Do the quality standards improve quality of care?
  2. Do the standards reduce neonatal complications?
  3. Do the standards reduce probability of neonatal death within first 48 hours of life?
  4. Do hospitals have adequate facilities to support the implementation of NRHM quality standards to bring down Infant Mortality Rate (IMR)?
  5. What barriers do hospitals face in implementing the intervention and how can they be overcome?

Methodology

Intervention design

This study evaluates the effectiveness of NRHM standards to improve neonatal outcomes and care practices in Kerala. NRHM conducts training sessions for hospital staff where standards are introduced, followed by distribution of poster-sized flow charts elaborating on the standards. A pilot intervention is implemented in a small set of government hospitals and a full-scale rollout across Kerala follows. The study also assesses the quality standards of the intervention hospitals; looking primarily at availability of required medical supplies and neonatal care equipment, and the extent to which the NRHM guidelines were being adhered. A pilot intervention is implemented in a small set of government hospitals, followed by a full-scale rollout across Kerala.

Theory of change

Current delivery and neonatal facilities found in hospitals in Kerala do not provide adequate care and support for pregnant mothers and new-borns. NRHM is implementing standards of care to reduce neonatal mortality in Kerala hospitals. The thirteen IMR quality standards address common risk factors for neonatal death, including respiratory disorders, hypothermia, hypoglycaemia and preterm labour and protective factors, including first and second trimester scanning, clean birthing instruments, breastfeeding within one hour of birth, and continuous monitoring during the first and second stages of labour. The IMR criteria are expected to improve standards of neonatal health, leading to better care and reduced complications after birth.

Evaluation design

This study uses a quasi-experimental Difference-in-Difference (DID) approach across districts and hospitals. Researchers plan to collect at least 12 months of pre-intervention data and 12 months of post-intervention data for all 9 intervention hospitals. The primary data used in this study will be patient medical records accompanied by qualitative interviews, and labour room registers. The study identifies three key measurable outcomes: presence of key hospital equipment and medicines, use of NRHM practices within the hospital, and neonatal survival 48 hours post-delivery. To assess impacts on process and health outcomes, researchers will estimate weighted least squares regression models where the weights consist of the number of observations in each hospital-month cell (cells with more observations have smaller variance and should be accorded higher weight). Authors will assess impacts across districts and within hospitals. Standard errors are adjusted for clustering and serial correlation, with wild bootstrap corrections utilised in both models due to the small number of districts in the sample.

Scroll to Top