Risk aversion and costs: A comparison of family physicians and general internists

Kevin Fiscella, Peter Franks, Jack Zwanziger, Cathleen Mooney, Melony Sorbero, Geoffrey C. Williams

Research output: Contribution to journalArticlepeer-review

53 Scopus citations


BACKGROUND - The authors of previous studies have suggested that family physicians generate lower health care expenditures than internists. Explanations for this difference have not been explored. METHODS - We surveyed 61 family physicians and 112 internists within a managed care organization regarding their demographic, practice, and psychological characteristics. We derived physician costs per enrollee and case-mix adjustment using claims data. RESULTS - In a multivariate analysis, we found that family physicians were significantly less risk averse than general internists. After adjustment for case mix, family physicians generated 5% lower costs (95% confidence interval [CI], 2% - 9%). After adjustment for case mix, risk-averse physicians generated higher expenditures; a one standard deviation increase in risk-aversion was associated with a 3% increase in expenditures (95% CI, 1% - 5%). After adjustment for case mix and risk aversion, family physicians' costs were no longer significantly lower (3%; 95% CI, -1% to 7%). None of the other physician demographic, practice, or psychological characteristics were significantly associated with case-mix- adjusted expenditures. CONCLUSIONS - The lower costs per patient generated by family physicians compared with internists may reflect psychological differences in risk aversion.

Original languageEnglish (US)
Pages (from-to)12-17+18
JournalJournal of Family Practice
Issue number1
StatePublished - Jan 2000
Externally publishedYes


  • Attitude
  • Physician's practice

ASJC Scopus subject areas

  • Public Health, Environmental and Occupational Health


Dive into the research topics of 'Risk aversion and costs: A comparison of family physicians and general internists'. Together they form a unique fingerprint.

Cite this