Housing Vouchers and Health Spending

The new rage in health care is attending to social determinants of health.  Now hopefully we are beginning to see some research on whether addressing social needs does impact health care outcomes.  A study reported in the Journal of the American Medical Association looked at whether providing housing vouchers affected health care.   (JAMA Article)   From 1994 to 1998 about 4600 families who lived in public housing projects in Baltimore, Boston, Chicago, New York and Los Angeles were randomized to get a voucher for housing anywhere, housing in a low-poverty area or no voucher.  This study attempted to track differences between the two voucher groups and the non-voucher group in health outcomes, but really only looked at hospitalizations, including the number of hospitalizations, number of days hospitalized and annual hospital spending.  All-payer databases were used for some states and/or Medicaid claims for others.  The study is obviously small, only about 4070 families were used in the final analysis, and it isn’t clear why such a limited set of outcomes was used.  The groups with vouchers did appear to use them to move to neighborhoods with less poverty.  Among adults at the time the study started, the rate of of hospitalizations over the followup period was slightly lower, 14 per 100 person years, than for adults in the non-vouchered group, at 14.7 hospitalizations per 100 person years, but after adjustment, there was no statistically significant difference in the hospitalization rate, hospital days or spending between the groups.  Among those family members who were children at the start of the study, the voucher group had a hospitalization rate during the followup period of 6.3 per 100 person years, compared to 7.3 in the non-voucher group.  Hospital spending was also lower for this group but the number of inpatient days was not different.  But these differences applied only to younger children, those under 13 years of age at the start of of the study.

Hard to understand what conclusions to draw.  For adults and older children, moving to a lower-poverty neighborhood did not appear to affect hospital use and spending.  For younger children, while the number of hospitalizations and spending was lower, the total number of hospital days wasn’t.  And more important measures, like the total utilization pattern, total spending, changes in health behaviors and health status, weren’t even analysed, although some data must have been available.  It seems unlikely to me that just moving to a lower-poverty neighborhood alone is going to do much in regard to access to health care or changes in health behavior.  The last item is the important one.  Low-income people tend to have behaviors that only worsen their plight:  poor attendance and effort at school, poor diet, poor exercise habits, more smoking and drug use, and less use of available health care resources.  Addressing those behaviors has been repeatedly shown to take more than just throwing money at low-income people.