The relationship between physicians' malpractice claims history and later claims. Does the past predict the future?

OBJECTIVE: To investigate whether an association exists between physicians' past and subsequent claims of medical malpractice, particularly whether a history of even unpaid claims ($0) or small claims (< $30,000) predicts subsequently higher rates of claims, especially large paid claims (> or = $30,000) (all in 1990 dollars). DATA: All medical malpractice claims closed in the state of Florida from January 1975 through August 1988 (N = 20,016, 92% involving physicians), matched with the American Medical Association's Physician Masterfile on all practicing physicians in the state of Florida during that period. Claims history was automated into physician-year claims files, then partitioned into a baseline period (1975 through 1980) and a subsequent period (1981 through 1983). Inconsequential claims were excluded, ie, cases closed without a named claimant and without expense for investigation (30.4% of raw claims).
METHODS: Descriptive analysis of all physician claims; odds ratio analysis of physicians in practice throughout both periods (N = 8247), comparing claims experience in baseline vs subsequent period, adjusted for specialty of practice.
RESULTS: For all consequential physician claims, 60% were unpaid claims, 17% were small paid claims, and 23% were large paid claims. The 8247 continuously practicing physicians had a total of 6614 claims, averaging 0.9 per year, but 59.2% of physicians had no claims in 9 years, only 13.4% had any paid claims, and 7.2% had multiple paid claims. Less than 8% of physicians had any large paid claims during the baseline period, and less than 7% had any in the subsequent period. Physicians with any baseline claims (whether paid or unpaid, small or large, single or multiple) had elevated odds of subsequent claims (whether defined as any claims, any paid claims, any large claims, or multiple claims) relative to physicians with no baseline claims. With a baseline of all small claims, the adjusted odds ratio for any subsequent claim was 2.84 (95% confidence interval [CI], 2.32 to 3.49), for any subsequent paid claim was 2.97 (95% CI, 2.34 to 3.77), for all large subsequent claims was 2.42 (95% CI, 1.76 to 3.33), and for subsequent multiple claims was 2.83 (95% CI, 2.08 to 3.86). Even having a single unpaid baseline claim approximately doubled the odds.
CONCLUSIONS: Claims history had predictive value, even with only unpaid claims. Small paid claims were better predictors than unpaid claims, large paid claims were better predictors than small paid claims, and multiple paid claims were better predictors than single paid claims. Claims history of all kinds is a reasonable statistical measure, eg, for the screening purposes of the National Practitioner Data Bank.