Combining administrative data from multiple agencies: Housing, Justice, Education, and Health

The use of big data to evaluate interventions and policies directed at public health and the social determinants of health is aided by the linking of different data elements at a population level.  As part of the Pathways to Health and Social (PATHS) Equity for Children research program, the data repository housed at the Manitoba Centre for Health Policy is able to bring together complete population level data (encrypted and anonymized) to evaluate policies or programs and their effects on health outcomes, educational outcomes, and the social determinants of health.  Some examples of the data elements include: a complete population registry with basic demographics; all ambulatory physician visits; hospitalizations; prescription drug dispensations; school enrollment and grades; receipt of income assistance; involvement with child and family services; criminal justice data.  Clinical data and social program data have also been merged with the data held in the repository to facilitate specific research aims.  Two completed projects are briefly described outlining how disparate data elements were merged and utilized.  1) To determine whether the location of social housing units in rich or poor neighborhoods influences vaccinations at age two, school readiness, teenage births, and high school completion.  2)  To determine whether receipt of a modest unconditional income supplement during the second and third trimester of pregnancy affected birth outcomes, breastfeeding initiation and initial hospital outcomes.  The success of both of these studies illustrates the power of big data to facilitate important policy evaluation work.