Alcohol brief intervention and 2-year healthcare costs: An observational study in adult primary care.

Abstract

Aims: To compare healthcare costs over 2 years between those who did and did not receive an alcohol brief intervention (BI) among adult primary care patients screening positive for unhealthy alcohol use.

Design: Population-based observational study, using electronic health record data.

Setting: Kaiser Permanente Northern California, a non-profit, integrated healthcare delivery system of socio-economically and demographically diverse members in California, USA.

Participants: Adult primary care patients, aged 18-85 years, who screened positive for unhealthy alcohol use between 1 January 2014 and 31 December 2017 as part of a systematic alcohol Screening, Brief Intervention and Referral to Treatment (SBIRT) program (n = 287 551). Patients either received a BI for unhealthy alcohol use (BI group) or did not receive a BI (non-BI group).

Measurements: Total emergency department (ED) and inpatient costs summarized in 6-month intervals from index screening through 24 months post-index; multivariable models examined associations between BI receipt and cost, and potential moderation by patient characteristics (age, sex, race/ethnicity, insurance type, clinical characteristics including body mass index, smoking status, physical activity level, the Charlson index of comorbidity, baseline drinking levels, drug and alcohol use disorders and mental health conditions in the prior year and the corresponding cost in the year prior to index date).

Findings: Adjusting for patient characteristics and prior-year cost, the largest declines in cost were found in the 6 months immediately following the index date for both BI and non-BI groups, and patients receiving a BI had greater reductions [estimate = -$209, 95% confidence interval (95% CI) = -$298 to -$119; estimate = -$11, 95% CI = -$14 to -$7, respectively] in total and ED costs, respectively, during this period compared with those who did not. Patients with a Charlson score ≥3 receiving a BI had lower total costs (estimate = -$621, 95% CI = -$1196 to -$46) and lower ED costs (estimate = -$24, 95% CI = -$47 to -$1) over 2 years, and patients with AUD receiving a BI had lower ED costs (estimate = -$33, 95% CI = -$66 to $0, respectively) than those who did not.

Conclusion: Among US adult primary care patients screening positive for unhealthy alcohol use, individuals who receive an alcohol brief intervention at the time of screening appear to have statistically significantly greater declines in healthcare costs in the 6 months following screening than individuals who do not receive an alcohol brief intervention. Moreover, receipt of an alcohol brief intervention appears to be associated with statistically significantly lower costs in two particularly vulnerable (and historically costly) patient groups: patients with alcohol use disorders and those with a Charlson score ≥3 (indicative of significant medical comorbidity).