Democratic and Republican physicians provide different care on politicized health issues

Abstract

Physicians frequently interact with patients about politically salient health issues, such as drug use, firearm safety, and sexual behavior. We investigate whether physicians' own political views affect their treatment decisions on these issues. We linked the records of over 20,000 primary care physicians in 29 US states to a voter registration database, obtaining the physicians' political party affiliations. We then surveyed a sample of Democratic and Republican primary care physicians. Respondents evaluated nine patient vignettes, three of which addressed especially politicized health issues (marijuana, abortion, and firearm storage). Physicians rated the seriousness of the issue presented in each vignette and their likelihood of engaging in specific management options. On the politicized health issues-and only on such issues-Democratic and Republican physicians differed substantially in their expressed concern and their recommended treatment plan. We control for physician demographics (like age, gender, and religiosity), patient population, and geography. Physician partisan bias can lead to unwarranted variation in patient care. Awareness of how a physician's political attitudes might affect patient care is important to physicians and patients alike.

Keywords: health care; partisanship; physicians; politics; primary care.

Fig. 1.

Perceptions of seriousness, by party affiliation. (A) Histograms for each vignette by party affiliation. Red represents Republicans; blue represents Democrats. (B) Circles represent coefficients from regressions, with 95% CIs.

Fig. 2.

Perceptions of seriousness, by subgroup. Difference of means with 95% confidence intervals are shown. The lower plot uses quartiles of a principal-components factor analysis based on a physicians' perceptions of their patient population.

Fig. 3.

Partisan differences in treatment plans on politicized issues. Points represent coefficients from regression, with 95% CIs. Positive values indicate that Republicans are more likely to include a particular item in a treatment plan; negative values indicate that Democrats are more likely to do so.

Fig. 4.

Partisan differences are especially pronounced on politicized issues. Note: Points represent coefficients from regression, with 95% CIs.

Fig. S1.

Specialty of physicians in NPI file vs. physicians in NPI-Catalist–matched sample.

Fig. S2.

Work state of physicians in NPI file vs. physicians in NPI-Catalist–matched sample.

Fig. S3.

Distribution of NPI physicians vs. NPI-Catalist–matched physicians on zip code population.