Supply of Medical Services

Social Networks and Access to Care

Jason Shafrin

On the news yesterday that Microsoft purchased a 1.5% stake in Facebook for $240 million, social networking appears to  highly valued commodity.

“…[T]he strength of social networks among Mexican-Americans is positively related to access to care.”

Is this true? That is hypothesis that a recent NBER working paper by Roan Gresenz, Escarce and Rogowski attempt to test. The authors uses data from 1996-2002 MEPS, along with data from the 2000 Census, Area Resource File, the CPS and the Bureau of Primary Health Care Uniform Data System.

Methods

The econometric specification is relatively simple. A probit regression is used where the dependent variables are either (1) whether the person has a “usual source of care,â€? (2) whether or not the person has an office-based physician or non-physician visit during a year; (3) whether the individual has any prescription drug expenditures during the year; (4) whether the individual has any medical expenditures during the year.

Social networks are measured three ways. The first is the percentage of Hispanics in an individuals ZIP code of residence, the second is the percentage of the population that is foreign born and Spanish speaking, and third is the percentage of Spanish speakers in a ZIP code. Each social network measure is interacted “…with an individual-level variable indicating whether the individual was born in the U.S., is foreign-born but has lived in the U.S. for more than five years, or is foreign-born and a recent immigrant.”

Results

The authors find that Mexican-Americans who only speak Spanish have less access to care than bilingual Mexican-Americans. Insured Mexican-Americans who live in areas of dense Hispanic populations have more access to care than insured Mexican-Americans living in areas with a weaker Hispanic presence. The results for uninsured Mexican-Americans are mixed. It seems that uninsured individuals living in an area with more Hispanics or more Spanish speakers leads to more medical expenditures and office visits, but seems to decrease the percentage of the population who spend money on pharmaceuticals or have a usual source of care.

Healthcare Economist commentary

One worry is that areas with larger Hispanic populations are areas with higher economic growth or are somehow different than areas with fewer Hispanics or Spanish speakers. Since the authors use data on a ZIP code level, this may be less of a problem. Further the authors state the following:

“We find no effects of these characteristics of the local population on access to care for U.S. born Mexican-Americans, suggesting that similarities in race and language may contribute more to the formation of social ties among individuals who are less acculturated to the U.S. “

While the data seem to prove that Mexican-Americans–especially the insured– have more access to care in areas with high levels of Spanish-speakers, the mechanism of how this manifests itself is unclear. The authors claim that social networks is the cause. Social networks provide information concerning which doctors are of high quality and are Spanish-speaking. On the other hand, a supply-side story could explain this phenomenon as well. In areas with many Hispanics, there will be more Spanish-speaking doctors and Mexican-Americans will gain more utility from a physician office visit. This is likely true regardless of whether a single individual has a large or small social network.

Even though this paper uses a large data set and has an intuitive result, since their is no measure of an individual’s social network the results can not prove that social networks–rather than other features of heavily Hispanic areas–are causing changes in access to care.