Evaluation of behavioral economic strategies to raise influenza vaccination rates across a health system: Results from a randomized clinical trial.

Influenza vaccination rates are low. Working with a large US health system, we evaluated three health system-wide interventions using the electronic health record's patient portal to improve influenza vaccination rates. We performed a two-arm RCT with a nested factorial design within the treatment arm, randomizing patients to usual-care control (no portal interventions) or to one or more portal interventions. We included all patients within this health system during the 2020-2021 influenza vaccination season, which overlapped with the COVID-19 pandemic. Through the patient portal, we simultaneously tested: pre-commitment messages (sent September 2020, asking patients to commit to a vaccination); monthly portal reminders (October - December 2020), direct appointment scheduling (patients could self-schedule influenza vaccination at multiple sites); and pre-appointment reminder messages (sent before scheduled primary care appointments, reminding patients about influenza vaccination). The main outcome measure was receipt of influenza vaccine (10/01/2020-03/31/2021). We randomized 213,773 patients (196,070 adults ≥18 years, 17,703 children). Influenza vaccination rates overall were low (39.0%). Vaccination rates for study arms did not differ: Control (38.9%), pre-commitment vs no pre-commitment (39.2%/38.9%), direct appointment scheduling yes/no (39.1%/39.1%), pre-appointment reminders yes/no (39.1%/39.1%); p > 0.017 for all comparisons (p value cut-off adjusted for multiple comparisons). After adjusting for age, gender, insurance, race, ethnicity, and prior influenza vaccination, none of the interventions increased vaccination rates. We conclude that patient portal interventions to remind patients to receive influenza vaccine during the COVID-19 pandemic did not raise influenza immunization rates. More intensive or tailored interventions are needed beyond portal innovations to increase influenza vaccination.

Impact of perceived contraceptive policy changes on long acting reversible contraception dispensing trends.

OBJECTIVE: We observed the long-term trend of Long Acting Reversible Contraception (LARC) usage before and after the 2016 presidential election. METHODS: We observed the rate of LARC dispensed at a university student health centre in the 18 months preceding and 27 months following the 2016 U.S. presidential election which posed threat to contraception access. We applied a segmented regression model using two linear regression line segments to evaluate whether there is a time point where the trend of LARC dispensing changed. We fit the regression models with a breakpoint at month 0 (election month) and 3 months with a Locally Estimated Scatterplot Smoothing (LOESS) estimate with parameters obtained by estimating simple linear regression models separately below and above the breakpoint '0'. RESULTS: There were a total of 2067 LARC methods dispensed from May 2015 to February 2019. The average number of LARC methods dispensed before November 2016 was 38/month and increased to 51/month post-presidential election. The LARC dispense rate significantly increased each month (0.38, 1.74; 95% confidence level, p < 0.05) until a breakpoint at 6 months (standard error 4.11) post-election followed by slower decrease (-0.59/month, 95% confidence level: -1.37, 0.20; p=not significant). CONCLUSION: Our study is the first to report long-term trends (27 months post-election) in LARC uptake amidst the public discourse that suggested the end of a policy that provided LARC insurance coverage. Although this observational study cannot suggest causality, the findings could reflect actions taken to prevent unintended pregnancy in response to the event of uncertain national policy.