OBJECTIVE: To investigate whether patients with a $1,000 annual prescription drug benefit cap reduced their drug consumption prior to exceeding the cap threshold. Previously, we found that the drug cap reduced overall drug consumption during the year. METHODS: All 183,640 subjects were 65+ years with Medicare insurance, had tiered copayments ($10 for generic & $15-35 for brand drugs), and were members of an integrated, prepaid delivery system: 146,050 subjects had an annual $1,000 drug benefit limit; and 37,590 subjects had no benefit limit (because of supplementary insurance from former employers rather than individual choice). To compare drug consumption (measured in dollars) below the cap amount, we examined the risk of cap and non-cap subjects consuming $250, $500, $750, and $1000 in 2003 using proportional hazard models for each of these thresholds. We adjusted for age, gender, race/ethnicity, brand copayment amount, prior visits, socioeconomic status, comorbidity, and having a regular primary care provider. RESULTS: Among the 183,640 subjects, 16,657 (11%) of subjects with a cap and 7,888 (21%) of subjects without caps exceeded the $1,000 cap threshold during 2003. After adjustment for covariates, subjects with a cap were significantly less likely to exceed the $1,000 cap threshold (HR=0.61, 95% CI: 0.56 – 0.66), compared with subjects without a cap. Similarly, subjects with a cap were significantly less likely to exceed lower drug consumption thresholds during the year, compared with subjects without a cap: HR=0.75 for $750 (95% CI: 0.70 – 0.80); HR=0.80 for $500 (95% CI: 0.75 – 0.85); HR=0.81 for $250 (95% CI: 0.77 – 0.85). CONCLUSIONS: These preliminary analyses indicate that patients with annual drug benefit caps appear to anticipate exceeding the benefit limit, and reduce their drug consumption prior to reaching the cap amount. Further research is needed to assess how patients reduce their drug consumption and the clinical and economic impact.