aExclusions were done sequentially. We included participants who were prescribed any of the following medications: (1) statin; (2) metformin; (3) sulfonylureas (including glipizide and gliclazide); (4) meglitinides (including repaglinide and nateglinide); (5) thiazolidinediones (including rosiglitazone and pioglitazone); (6) dipeptidyl peptidase–4 inhibitors (including sitagliptin, saxagliptin, linagliptin, and alogliptin); (7) glucagon-like peptide 1 agonists (including exenatide and liraglutide); and (8) angiotensin-converting enzyme inhibitors, angiotensin II receptor blockers, and direct renin inhibitors; were prescribed a 30-day supply of medication; had a 40% to 80% compliance for relevant medications in the 12 months prior to randomization; were not on the do-not-contact list prior to randomization; had a birthday from January 21 to April 21; and were not living with any other participant in our sample.
bResearchers at the University of Pennsylvania used R, version 3.1.1 (R Core Team) to randomly assign participants to conditions. Participants whose birthdays were between January 21 and March 31 were assigned to 1 of 5 conditions under a 1:1:1:1:1 allocation ratio. Participants who birthdays were between April 1 and April 21 were assigned to the control, new year unframed, or new year framed condition under a 1:1:1 allocation ratio. Participants were enrolled by Humana. All conditions were balanced on available demographic variables (age, sex, race/ethnicity, and income) and proportion of days covered prior to our randomized clinical trial (RCT), suggesting that our randomization was successful.
cWe excluded 3005 participants born in April from our analysis to make conditions comparable. The birthday unframed and birthday framed conditions did not include anyone with April birthdays because our last mailing date was March 25.
Error bars represent standard error. For each participant, our ordinary least squares regression controls for his or her proportion of days covered during the 90 days prior to the mailing date, sex, race/ethnicity, the linear and squared terms of age, log-transformed income, and the type of medication that was listed on the reminder (ie, diabetes, cholesterol, or blood pressure). By design, any differences between the control and treatment conditions reflect a combination of the time of year and the potential treatment effect of receiving a fresh-start reminder.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Dai H, Mao D, Riis J, et al. Effectiveness of Medication Adherence Reminders Tied to “Fresh Start” Dates: A Randomized Clinical Trial. JAMA Cardiol. 2017;2(4):453–455. doi:10.1001/jamacardio.2016.5794
Low medication adherence is problematic.1 Well-designed reminders can increase adherence,2 but when should reminders be sent to maximize their effect? Prior observational studies3 and laboratory experiments4 have shown that engagement in healthy activities increases considerably following fresh-start dates: life and calendar events signaling the beginning of new cycles (eg, birthdays or New Year’s Day).3,4 Extending this insight, we conducted what is, to our knowledge, the first randomized clinical trial examining whether sending medication adherence reminders around fresh-start dates and highlighting these dates as an opportunity for positive changes could boost reminders’ effectiveness.
Between January 21, 2015, and March 25, 2015, we mailed reminders to 13 323 participants (5970 men [45%]; 40%-80% adherence in the past 12 months) with commercial or Medicare Advantage insurance with Humana, encouraging them to regularly take their cholesterol, diabetes, or blood pressure medications. Participants were randomly assigned to 1 of 5 mailing conditions (Figure 1). In the birthday unframed and birthday framed conditions, reminders were sent within 1 week before each participant’s birthday. In the new year unframed and new year framed conditions, reminders were sent 3 weeks after New Year’s Day. Reminders in the birthday framed and new year framed conditions highlighted the participants’ birthday or New Year’s Day, respectively, as an opportunity to make a fresh start and begin taking medications regularly. In the control condition, reminders without any reference to fresh-start dates were sent on a randomly selected day that was at least 1 month after New Year’s Day and 1 week away from the participant’s birthday. The institutional review board of the University of Pennsylvania approved a waiver of informed consent requirements because the research presented no more than minimal risk of harm to participants and involved no procedures for which written consent is normally required outside of the research context. The formal trial protocol can be found in the Supplement.
For each participant, we used pharmacy claims data to calculate the proportion of days covered during our 90-day postmailing observation period, defined as the number of days he/she had any pills in the medication category listed on his/her reminder divided by 90 days. Ordinary least squares regressions to estimate treatment effects in STATA, version 14 (StataCorp) had an 80% power to detect a difference of at least 2 percentage points between the control and each treatment condition with α = .05.
Mean proportion of days covered among participants was 63.3% over 90 days postmailing. Compared with the control condition, proportion of days covered did not significantly differ in the birthday unframed (mean difference, 0.56%; 95% CI, −1.09% to 2.19%), birthday framed (mean difference, 0.55%; 95% CI, −1.08% to 2.20%), new year unframed (mean difference, 1.32%; 95% CI, −0.32% to 2.95%), or new year framed condition (mean difference, 0.38%; 95% CI, −1.26% to 2.03%). The difference in proportion of days covered was also insignificant comparing the birthday unframed and birthday framed conditions (mean difference, −0.02%; 95% CI, −1.66% to 1.63) or the new year unframed and new year framed conditions (mean difference, −0.93%; 95% CI, −2.57% to 0.71%). Figure 2 depicts these results.
Contrary to our expectations, sending reminders following fresh-start dates was not associated with increased medication adherence, and fresh-start–based framing was not associated with increased reminder effectiveness. We encourage further study before concluding that the psychology of fresh starts does not apply to medication adherence. Because fresh-start dates motivate individuals wishing to initiate goal pursuit,3,4 our timing- and framing-based treatments may increase the effectiveness of reminders when reminders involve goal-setting activities. Additionally, there is often a delay between a target fresh-start date and the date when treatment-condition reminders were actually received; in the New Year conditions, reminders often arrived in late January. Reminders received immediately after the target date could be more effective. Further investigation into alternative ways to leverage fresh starts5 and compel patients to attend to public health messaging would be valuable.
This study has several limitations. The insurer sent some customers medication adherence reminders outside of the randomized clinical trial, and we were unable to include a condition without reminders. Also, many participants were involved in another randomized clinical trial comparing reminders that ended shortly before this randomized clinical trial. Furthermore, our participants had lower medication adherence levels than those in other adherence studies (eg, a multisite study published in 20156), possibly because of our selection criteria.
Corresponding Author: Hengchen Dai, PhD, Olin Business School, Organizational Behavior Department, Washington University in St Louis, One Brookings Drive, Campus Box 1156, St Louis, MO 63130 (email@example.com).
Published Online: February 8, 2017. doi:10.1001/jamacardio.2016.5794
Author Contributions: Dr Dai had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Dai, Riis, Volpp, Relish, Milkman.
Acquisition, analysis, or interpretation of data: Dai, Mao, Volpp, Relish, Lawnicki, Milkman.
Drafting of the manuscript: Dai, Mao.
Critical revision of the manuscript for important intellectual content: Riis, Volpp, Relish, Lawnicki, Milkman.
Statistical analysis: Dai, Mao.
Obtained funding: Volpp, Relish, Milkman.
Administrative, technical, or material support: Volpp, Relish, Lawnicki.
Supervision: Dai, Riis, Relish, Milkman.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Dai reports funding in part by grants from Humana during the conduct of the study. Mr Mao reports funding in part by grants from Humana during the conduct of the study. Dr Volpp reports grants from Humana during the conduct of the study; grants from Hawaii Medical Services Agency, Discovery (South Africa), Merck, Weight Watchers, and CVS outside of the submitted work; he has received consulting income from CVS and VALHealth and is a principal in VALHealth, a behavioral economics consulting form. Mr Relish reports employment by Humana, the funding organization, during the course of the study. Dr Lawnicki reports employment by Humana, the funding organization, during the course of the study. Dr Milkman reports grants from Humana during the conduct of the study; personal fees from Health Prize Academic Advisory Board, outside the submitted work. No other disclosures were reported.
Funding/Support: This research was funded by research grants from Humana.
Role of the Funder/Sponsor: The funding organization, in collaboration with the academic investigators, provided input on the design and conduct of the study; oversaw the collection and management of the data; and contributed to the preparation, review, and approval of the manuscript as well as the decision to submit the manuscript for publication. The funding organization had no role in the analysis or interpretation of the data.
Additional Contributions: We thank Bradford Tuckfield, PhD, Berkadia, for his contributions to study conception and design, and Hae Nim (Sunny) Lee, BA, University of Pennsylvania, for her contributions to study conception and design and data analysis; they both received funding in part by grants from Humana during the conduct of the study. We also thank Joelle Friedman, MPA, University of Pennsylvania, for her help with acquisition of data; she was not compensated for her contributions. We also thank Humana Inc, especially Heather E. Pearce, RPh, George J. Spurlock, MBA, Peinie P. Young, PharmD, BCACP, Shannon Clark, MBA, Vicki L. Vogel, BA, and Tova Levin, MBA, for their administrative and technical support. As employees of Humana, they were compensated for their normal work.
Trial Registration: Clinicaltrials.gov Identifier: NCT02411032.
Create a personal account or sign in to: