Qualitative Exploration of Barriers to Statin Adherence and Lipid Control: A Secondary Analysis of a Randomized Clinical Trial | Cardiology | JAMA Network Open | JAMA Network
[Skip to Navigation]
Sign In
Figure.  Recruitment Flowchart
Recruitment Flowchart

Staff contacted participants following a rank-order approach until all stratification cells were complete. Incentive groups were stratified equally across the 3 incentive groups by interview group. LDLC indicates low-density lipoprotein cholesterol.

Table 1.  Demographic and Clinical Characteristics of Study Participants
Demographic and Clinical Characteristics of Study Participants
Table 2.  Demographic and Clinical Characteristics of Participants in the Randomized Clinical Trial With the Highest and Lowest Deciles of LDLC Level Change During the Trial
Demographic and Clinical Characteristics of Participants in the Randomized Clinical Trial With the Highest and Lowest Deciles of LDLC Level Change During the Trial
Table 3.  Understanding of Intervention Design
Understanding of Intervention Design
Table 4.  Dietary and Medication Habits Entering the Trial
Dietary and Medication Habits Entering the Trial
1.
Heron  M. Deaths: leading causes for 2016. National Vital Statistics Reports. Published July 26, 2018. Accessed May 7, 2020. https://www.cdc.gov/nchs/data/nvsr/nvsr67/nvsr67_06.pdf
2.
 Randomised trial of cholesterol lowering in 4444 patients with coronary heart disease: the Scandinavian Simvastatin Survival Study (4S).   Lancet. 1994;344(8934):1383-1389.PubMedGoogle Scholar
3.
Maningat  P, Gordon  BR, Breslow  JL.  How do we improve patient compliance and adherence to long-term statin therapy?   Curr Atheroscler Rep. 2013;15(1):291. doi:10.1007/s11883-012-0291-7 PubMedGoogle ScholarCrossref
4.
Kearney  PM, Blackwell  L, Collins  R,  et al; Cholesterol Treatment Trialists’ (CTT) Collaborators.  Efficacy of cholesterol-lowering therapy in 18,686 people with diabetes in 14 randomised trials of statins: a meta-analysis.   Lancet. 2008;371(9607):117-125. doi:10.1016/S0140-6736(08)60104-X PubMedGoogle ScholarCrossref
5.
Mihaylova  B, Emberson  J, Blackwell  L,  et al; Cholesterol Treatment Trialists’ (CTT) Collaborators.  The effects of lowering LDL cholesterol with statin therapy in people at low risk of vascular disease: meta-analysis of individual data from 27 randomised trials.   Lancet. 2012;380(9841):581-590. doi:10.1016/S0140-6736(12)60367-5 PubMedGoogle ScholarCrossref
6.
Jackevicius  CA, Mamdani  M, Tu  JV.  Adherence with statin therapy in elderly patients with and without acute coronary syndromes.   JAMA. 2002;288(4):462-467. doi:10.1001/jama.288.4.462 PubMedGoogle ScholarCrossref
7.
Benner  JS, Glynn  RJ, Mogun  H, Neumann  PJ, Weinstein  MC, Avorn  J.  Long-term persistence in use of statin therapy in elderly patients.   JAMA. 2002;288(4):455-461. doi:10.1001/jama.288.4.455 PubMedGoogle ScholarCrossref
8.
Colantonio  LD, Huang  L, Monda  KL,  et al.  Adherence to high-intensity statins following a myocardial infarction hospitalization among Medicare beneficiaries.   JAMA Cardiol. 2017;2(8):890-895. doi:10.1001/jamacardio.2017.0911 PubMedGoogle ScholarCrossref
9.
Shah  ND, Dunlay  SM, Ting  HH,  et al.  Long-term medication adherence after myocardial infarction: experience of a community.   Am J Med. 2009;122(10):961.e7-961.e13. doi:10.1016/j.amjmed.2008.12.021PubMedGoogle ScholarCrossref
10.
Laibson  D.  Golden eggs and hyperbolic discounting.   Q J Econ. 1997;112(2):443-478. doi:10.1162/003355397555253 Google ScholarCrossref
11.
Augenblick  N, Niederle  M, Sprenger  C.  Working over time: dynamic inconsistency in real effort tasks.   The Quarterly Journal of Economics. 2015;130(3):1067-1115. doi:10.1093/qje/qjv020 Google ScholarCrossref
12.
Putt  ME, Reese  PP, Volpp  KG,  et al.  The Habit Formation trial of behavioral economic interventions to improve statin use and reduce the risk of cardiovascular disease: Rationale, design and methodologies.   Clin Trials. 2019;16(4):399-409. doi:10.1177/1740774519846852PubMedGoogle ScholarCrossref
13.
Barankay  I, Reese  PP, Putt  ME,  et al.  Effect of patient financial incentives on statin adherence and lipid control: a randomized clinical trial.   JAMA Netw Open. 2020;3(10):e2019429. doi:10.1001/jamanetworkopen.2020.19429 PubMedGoogle Scholar
14.
Bargh  JA, Chartrand  TL.  The unbearable automaticity of being.   Am Psychol. 1999;54(7):462-479. doi:10.1037/0003-066X.54.7.462 Google ScholarCrossref
15.
Schumpeter  J, Backhaus  U. The theory of economic development. In: Backhaus  J, ed.  Joseph Alois Schumpeter: Entrepreneurship, Style and Vision. Springer US; 2003:61-116. doi:10.1007/0-306-48082-4_3
16.
Nelson  RR, Winter SG.  An Evolutionary Theory of Economic Change. Belknap Press of Harvard University Press; 1982.
17.
Chetty  R, Szeidl  A.  Consumption commitments and habit formation.   Econometrica. 2016;84(2):855-890. doi:10.3982/ECTA9390 Google ScholarCrossref
18.
Ryder  HE, Heal  GM.  Optimal growth with intertemporally dependent preferences.   Rev Econ Studies. 1973;40(1):1-31. doi:10.2307/2296736 Google ScholarCrossref
19.
O’Cathain  A, Thomas  KJ, Drabble  SJ, Rudolph  A, Hewison  J.  What can qualitative research do for randomised controlled trials? a systematic mapping review.   BMJ Open. 2013;3(6):e002889. doi:10.1136/bmjopen-2013-002889 PubMedGoogle Scholar
20.
Ivankova  NV, Creswell  JW, Stick  SL.  Using mixed-methods sequential explanatory design: from theory to practice.   Field Meth. 2006;18(1):3-20. doi:10.1177/1525822X05282260 Google ScholarCrossref
21.
Penn Medicine. Facts and figures 2021 edition. Published 2020. Accessed March 1, 2021. https://www.pennmedicine.org/news/publications-and-special-projects/facts-and-figures
22.
Stone  NJ, Robinson  JG, Lichtenstein  AH,  et al; American College of Cardiology/American Heart Association Task Force on Practice Guidelines.  2013 ACC/AHA guideline on the treatment of blood cholesterol to reduce atherosclerotic cardiovascular risk in adults: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines.   Circulation. 2014;129(25)(suppl 2):S1-S45. doi:10.1161/01.cir.0000437738.63853.7a PubMedGoogle ScholarCrossref
23.
Creswell  J, Poth  C.  Qualitative Inquiry and Research Design: Choosing Among Five Approaches. 4th ed. SAGE Publications Inc.; 2018.
24.
Yeo  A, Legard  R, Keegan  J, Ward  K, Nicholls  CM, Lewis  J. In-depth interviews. In: Ritchie J, Lewis J, Nicholls CM, Ormtron R, eds.  Qualitative Research Practice: A Guide for Social Science Students and Researchers. 2nd ed: SAGE Publications; 2014:177-210.
25.
Hennink  MM, Kaiser  BN, Marconi  VC.  Code saturation versus meaning saturation: how many interviews are enough?   Qual Health Res. 2017;27(4):591-608. doi:10.1177/1049732316665344PubMedGoogle ScholarCrossref
26.
Tavory  I, Timmermans  S.  Abductive Analysis: Theorizing Qualitative Research. University of Chicago Press; 2014. doi:10.7208/chicago/9780226180458.001.0001
27.
Alvesson  M, Karreman  D.  Constructing mystery: empirical matters in theory development.   Acad Manage Rev. 2007;32(4):1265-1281. doi:10.5465/amr.2007.26586822 Google ScholarCrossref
28.
Osterberg  L, Blaschke  T.  Adherence to medication.   N Engl J Med. 2005;353(5):487-497. doi:10.1056/NEJMra050100 PubMedGoogle ScholarCrossref
29.
Conn  VS, Ruppar  TM.  Medication adherence outcomes of 771 intervention trials: systematic review and meta-analysis.   Prev Med. 2017;99:269-276. doi:10.1016/j.ypmed.2017.03.008 PubMedGoogle ScholarCrossref
30.
O’Donoghue  T, Rabin  M.  Present bias: lessons learned and to be learned.   Am Econ Rev. 2015;105(5):273-279. doi:10.1257/aer.p20151085 Google ScholarCrossref
31.
O'Donoghue  T, Rabin  M.  Doing it now or later.   Am Econ Rev. 1999;89(1):103-124. doi:10.1257/aer.89.1.103 Google ScholarCrossref
32.
Asch  DA, Troxel  AB, Stewart  WF,  et al.  Effect of financial incentives to physicians, patients, or both on lipid levels: a randomized clinical trial.   JAMA. 2015;314(18):1926-1935. doi:10.1001/jama.2015.14850 PubMedGoogle ScholarCrossref
33.
Gneezy  U, Meier  S, Rey-Biel  P.  When and why incentives (don’t) work to modify behavior.   J Econ Perspectives. 2011;25(4):191-210. doi:10.1257/jep.25.4.191 Google ScholarCrossref
34.
Tøttenborg  SS, Lange  P, Johnsen  SP, Nielsen  H, Ingebrigtsen  TS, Thomsen  RW.  Socioeconomic inequalities in adherence to inhaled maintenance medications and clinical prognosis of COPD.   Respir Med. 2016;119:160-167. doi:10.1016/j.rmed.2016.09.007 PubMedGoogle ScholarCrossref
35.
Yang  C, Hao  Z, Yu  D, Xu  Q, Zhang  L.  The prevalence rates of medication adherence and factors influencing adherence to antiepileptic drugs in children with epilepsy: a systematic review and meta analysis.   Epilepsy Res. 2018;142:88-99. doi:10.1016/j.eplepsyres.2018.03.018 PubMedGoogle ScholarCrossref
36.
Hamilton  AJ, Caskey  FJ, Casula  A, Inward  CD, Ben-Shlomo  Y.  Associations with wellbeing and medication adherence in young adults receiving kidney replacement therapy.   Clin J Am Soc Nephrol. 2018;13(11):1669-1679. doi:10.2215/CJN.02450218 PubMedGoogle ScholarCrossref
37.
Gast  A, Mathes  T.  Medication adherence influencing factors-an (updated) overview of systematic reviews.   Syst Rev. 2019;8(1):112. doi:10.1186/s13643-019-1014-8 PubMedGoogle ScholarCrossref
38.
Gaalema  DE, Elliott  RJ, Morford  ZH, Higgins  ST, Ades  PA.  Effect of socioeconomic status on propensity to change risk behaviors following myocardial infarction: implications for healthy lifestyle medicine.   Prog Cardiovasc Dis. 2017;60(1):159-168. doi:10.1016/j.pcad.2017.01.001 PubMedGoogle ScholarCrossref
39.
Nelson  LA, Ackerman  MT, Greevy  RA  Jr, Wallston  KA, Mayberry  LS.  Beyond race disparities: accounting for socioeconomic status in diabetes self-care.   Am J Prev Med. 2019;57(1):111-116. doi:10.1016/j.amepre.2019.02.013 PubMedGoogle ScholarCrossref
40.
Kim  J, Lee  E, Park  B-J, Bang  JH, Lee  JY.  Adherence to antiretroviral therapy and factors affecting low medication adherence among incident HIV-infected individuals during 2009-2016: a nationwide study.   Sci Rep. 2018;8(1):3133. doi:10.1038/s41598-018-21081-x PubMedGoogle ScholarCrossref
41.
Osborn  CY, Kripalani  S, Goggins  KM, Wallston  KA.  Financial strain is associated with medication nonadherence and worse self-rated health among cardiovascular patients.   J Health Care Poor Underserved. 2017;28(1):499-513. doi:10.1353/hpu.2017.0036 PubMedGoogle ScholarCrossref
42.
Patel  MS, Asch  DA, Volpp  KG.  Wearable devices as facilitators, not drivers, of health behavior change.   JAMA. 2015;313(5):459-460. doi:10.1001/jama.2014.14781 PubMedGoogle ScholarCrossref
Limit 200 characters
Limit 25 characters
Conflicts of Interest Disclosure

Identify all potential conflicts of interest that might be relevant to your comment.

Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.

Err on the side of full disclosure.

If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.

Not all submitted comments are published. Please see our commenting policy for details.

Limit 140 characters
Limit 3600 characters or approximately 600 words
    Original Investigation
    Pharmacy and Clinical Pharmacology
    May 4, 2021

    Qualitative Exploration of Barriers to Statin Adherence and Lipid Control: A Secondary Analysis of a Randomized Clinical Trial

    Author Affiliations
    • 1Department of Management, The Wharton School, University of Pennsylvania, Philadelphia
    • 2Department of Business Economics and Public Policy, The Wharton School, University of Pennsylvania, Philadelphia
    • 3Center for Health Incentives and Behavioral Economics, University of Pennsylvania, Philadelphia
    • 4Department of Biostatistics, Epidemiology and Informatics, Perelman School of Medicine, University of Pennsylvania, Philadelphia
    • 5Division of Renal Electrolyte and Hypertension, Department of Medicine, Perelman School of Medicine, University of Pennsylvania, Philadelphia
    • 6Department of Medical Ethics and Health Policy, Perelman School of Medicine, University of Pennsylvania, Philadelphia
    • 7Leonard Davis Institute of Health Economics, University of Pennsylvania, Philadelphia
    • 8currently a medical student at Rutgers New Jersey Medical School, Newark
    • 9Department of Emergency Medicine, Perelman School of Medicine, University of Pennsylvania, Philadelphia
    • 10Division of General Internal Medicine, Department of Medicine, Perelman School of Medicine, University of Pennsylvania, Philadelphia
    • 11Center for Health Equity Research and Promotion, Cresencz Veterans Affairs Medical Center, Philadelphia, Pennsylvania
    • 12Department of Health Care Management, The Wharton School, University of Pennsylvania, Philadelphia
    • 13Department of Anesthesiology and Critical Care, Perelman School of Medicine, University of Pennsylvania, Philadelphia
    JAMA Netw Open. 2021;4(5):e219211. doi:10.1001/jamanetworkopen.2021.9211
    Key Points

    Question  What barriers to statin therapy adherence and control of cholesterol levels are revealed through qualitative interviews with participants in a randomized trial of financial incentives for adherence?

    Findings  In this qualitative study of 54 participants, individuals whose cholesterol levels did not improve described a greater burden of chronic illness, were less frequently employed, were less focused on the risks of high cholesterol levels, appeared to have lower health literacy, were less engaged in their cholesterol level management, made fewer specific nutritional choices for optimizing health, and had greater difficulty obtaining healthy food compared with participants with marked improvement of cholesterol levels.

    Meaning  These findings suggest that future interventions should consider addressing socioeconomic circumstances in combination with adherence interventions among patients needing to reduce cholesterol levels.

    Abstract

    Importance  Financial incentives may improve health by rewarding patients for focusing on present actions—such as medication regimen adherence—that provide longer-term health benefits.

    Objective  To identify barriers to improving statin therapy adherence and control of cholesterol levels with financial incentives and insights for the design of future interventions.

    Design, Setting, and Participants  This qualitative study involved retrospective interviews with participants in a preplanned secondary analysis of a randomized clinical trial of financial incentives for statin therapy adherence. A total of 636 trial participants from several US insurer or employer populations and an academic health system were rank ordered by change in low-density lipoprotein cholesterol (LDLC) levels. Participants with the most LDLC level improvement (high-improvement group) and those with LDLC levels that did not improve (nonimprovement group) were purposively targeted, stratified across all trial groups, for semistructured telephone interviews that were performed from April 1 to June 30, 2018. Interviews were coded using a team-based, iterative approach. Data were analyzed from July 1, 2018, to October 31, 2020.

    Main Outcomes and Measures  The primary outcome was mean change in LDLC level from baseline to 12 months; the secondary outcome, statin therapy adherence during the first 6 months.

    Results  A total of 54 patients were interviewed, divided equally between high-improvement and nonimprovement groups, with a mean (SD) age of 43.5 (10.3) years; 36 (66.7%) were women, 28 (51.9%) had diabetes, and 18 (33.3%) had cardiovascular disease. Compared with the high-improvement group, the nonimprovement group had fewer interviewees with an annual income of greater than $50 000 (11 [40.7%] vs 22 [81.5%]), worse self-reported health (fair to poor, 13 [48.1%] vs 3 [11.1%]), more Black interviewees (16 [59.3%] vs 4 [14.8%]), and lower baseline LDLC levels (>160 mg/dL, 2 [7.4%] vs 25 [92.6%]). Participants in the nonimprovement group had a greater burden of chronic illness (≥2 chronic conditions, 13 [48.1%] vs 6 [22.2%]) and were less frequently employed (full-time, 6 [22.2%] vs 12 [44.4%]). In interviews, the nonimprovement group was less focused on risks of high LDLC levels, described less engagement in LDLC level management, articulated fewer specific nutritional choices for optimizing health, and recounted greater difficulty obtaining healthy food. Participants in both groups had difficulty describing the structure of the financial incentives but did recall features of the electronic pill containers used to track adherence and how those containers affected medication routines.

    Conclusions and Relevance  Participants in a statin adherence trial whose LDLC levels did not improve found it more difficult to create medication routines and respond to financial incentives in the context of complex living conditions and a high burden of chronic illness. These findings suggest that future studies should be more attentive to socioeconomic circumstances of trial participants.

    Trial Registration  ClinicalTrials.gov Identifier: NCT01798784

    Introduction

    Heart disease is the leading cause of death in North America.1 Individuals with atherosclerotic cardiovascular disease (ASCVD) can improve their survival by taking statins (hydroxymethyl glutaryl coenzyme A reductase inhibitors).2 Among individuals at elevated risk for ASCVD but without prior cardiac events, such as adults with diabetes, the US Preventive Services Task Force3 concluded that statins reduce the risk of death and ASCVD. Statins involve manageable adverse effects and are available as low-cost generics.4,5 Unfortunately, nonadherence to statin therapy is surprisingly common.6,7 Successful and durable strategies are needed to remove barriers to adherence and healthy routines.3,8,9 Present bias, whereby individuals undervalue the effect of healthy actions on their future well-being, is a fundamental behavioral barrier to adherence that may be addressed with small targeted incentives rewarding present actions.10,11

    The Habit Formation trial, a randomized clinical trial described elsewhere,12,13 tested the ability of 6 months of financial rewards for daily statin therapy adherence to establish longer-term adherence habits. The primary outcome was change in low-density lipoprotein cholesterol (LDLC) levels at 12 months. The incentives were designed using behavior modification approaches based on insights from economics, management, and psychology.12,14-18 Every participant received daily reminders and was asked to use an electronic pill container. Container opening times were transmitted via the cellphone network to track adherence. The trial had a control group as well as 3 incentive groups with equal expected payouts: the first group received daily rewards for adherence; the second group’s incentives were halved if they took their statins after receipt of a daily reminder; and the third group received half their rewards daily and the other half as a monthly deposit that decreased according to the number of days of nonadherence to statin therapy.

    Financial incentives make immediate the importance of adherence, which is particularly relevant because statins carry no immediately perceptible health benefits. In addition, incentives may enhance a patient’s ability to focus on establishing lasting daily routines for adherence, which is important because patients at high risk of ASCVD typically need to take statins for life.

    It is unknown, however, how best to design and communicate about incentives for medication adherence. In our clinical trial, the mean change in LDLC levels at 12 months was very similar between controls and those receiving daily financial incentives, despite improvements in measured statin therapy adherence recorded through electronic pill-bottle openings among intervention arm participants. In this qualitative study, we used semistructured interviews to compare the experiences of participants with large improvements vs no improvement in LDLC levels during the 12-month study period. This work was consistent with the trial’s secondary aim of exploring socioeconomic backgrounds as a factor in treatment response. We aimed to better understand the variation in effectiveness of incentives between individuals and general barriers to control of cholesterol levels and thereby inform the design of future adherence interventions.19

    Methods
    Overview

    This study follows the Standards for Reporting Qualitative Research (SRQR) guideline. We applied a mixed-methods sequential explanatory design.20 The randomized clinical trial was conducted from August 1, 2013, to July 30, 2018 (the first phase; the trial protocol is found in Supplement 1), with a qualitative interview study among a subset of trial participants from April 1 to June 30, 2018 (the second phase), to illuminate aspects of the quantitative findings from the trial. Investigators are increasingly implementing sequential explanatory designs to help explain the results of randomized clinical trials.19 The 4-group randomized clinical trial assessed the efficacy of financial incentives to promote sustained adherence to statin therapy and reduce LDLC levels in individuals at high risk for ASCVD.12,13 The qualitative study reported herein enrolled trial participants from Penn Medicine, the largest health care organization serving the Philadelphia region,21 and was approved by the University of Pennsylvania institutional review board. All participants provided written informed consent.

    As described elsewhere, participants were eligible for the trial if they had a daily statin prescription, self-reported incomplete adherence to the statin therapy, and 1 or more of the following: diabetes, ASCVD with elevated LDLC levels, 10-year ASCVD risk score of greater than 7.5%, or high LDLC levels (>190 mg/dL [to convert to millimoles per liter, multiply by 0.0258]).12,13,22 All participants received electronic pill containers, which transmitted a signal to the study team at each opening as a measure of adherence. Participants also received daily reminders to take their statin. Participants had LDLC levels assessed at baseline and 6 and 12 months. The control group received no further intervention. Those in the treatment groups were offered daily financial incentives for adherence, as measured by the pill container, and received daily feedback about their earnings. Participants were paid monthly. Incentives were implemented for 6 months, followed by a 6-month follow-up period without incentives or reminders.

    The primary outcome was mean change in LDLC level from baseline to 12 months. The secondary outcome was medication adherence during the first 6 months. On average, all groups saw substantial improvement in LDLC levels. Although financial incentives raised measured adherence in the treatment vs control groups, no significant differences were found between intervention and control groups in LDLC level at 12 months.

    Sampling and Data Collection for the Qualitative Study

    We used extreme case sampling (a type of purposive sampling) to select interviewees. As used in the qualitative phase of an explanatory sequential design, extreme sampling targets cases at the outer ends of the distribution along a quantitative parameter of interest (in this case, change in LDLC level from baseline to 12 months). Such cases are assumed to present particularly rich information about that parameter. This sampling strategy facilitates comparison between cases to explore potential causes of stark quantitative differences. We used this sampling approach to examine how the experiences of patients with high LDLC level improvement in the trial differed from those whose LDLC levels improved least.23 First, of 805 total trial participants, we rank ordered the 636 participants with LDLC level measurements at both baseline and 12 months according to their change in LDLC levels. We then recruited from 2 groups: (1) the high-improvement group, starting with the participant with the largest decline in LDLC levels and working down the rank-ordered list; and (2) the nonimprovement group, starting with the participant with the largest increase of LDLC levels and working up the rank-ordered list, eventually including some participants with small LDLC level declines. We stratified both groups across the trial’s 4 groups: as we neared code saturation, we began capping the number of interviewees recruited from specific groups, skipping over eligible participants in the LDLC level rank orders if they belonged to a capped group, until we achieved an equal stratification. The overall response rate was 41.9% (54 of 129 participants).

    Semistructured interviews were conducted by telephone by 3 team members (C.P., D.P., and S.C., with all research staff who had worked on the randomized clinical trial) and supervised by a faculty anthropologist with substantial experience in qualitative methods (J.T.C.). Recruited participants engaged in audio-recorded, one-on-one interviews (see eMethods in Supplement 2 for interview guide).24 Interviews addressed participants’ overall health status, their risk perceptions of high cholesterol levels and any behaviors they had implemented to address it, circumstances of their daily lives, their statin-taking routines, their experiences with financial incentives during the trial, their understandings of how they performed in the trial, and perceived effects of the trial on their long-term behavior. As described below, data were collected until code saturation was achieved, that is, until additional data did not necessitate any alterations to the codebook.25

    Qualitative Analysis

    A professional service transcribed the recordings. Coding of transcripts was managed with NVivo, version 12 (QSR International) and used an iterative process of codebook development and rounds of double coding to assess reliability and refine the codebook. eMethods in Supplement 2 provides details of this process.

    Once coding was complete, 2 investigators (C.P. and D.P.) independently compared the responses of participants in the high-improvement vs nonimprovement groups and of participants in the control group vs those in the combined financial incentives groups. After independently identifying thematic differences along these 2 dimensions, they met to reach consensus on a core set of differences. Finally, a third investigator (J.T.C.) reviewed and verified the differences agreed on by the previous 2 investigators. After coding and comparison were complete, the observed differences were explored via regular group discussion using an abductive approach, which is designed to facilitate explanation of findings that are not well explained by extant literature on a topic.26,27 We identified potential explanations for unexpected thematic trends, further evaluating them against the data to assess their level of empirical support until we arrived at an account that best reflected the data.

    Statistical Analysis

    Data were analyzed from July 1, 2018, to October 31, 2020. Characteristics of interviewees and trial participants were described at baseline and 12 months using percentages and 95% CIs.

    Results

    Table 1 displays the 54 interviewees’ characteristics (36 women [66.7%] and 18 men [33.3%]; mean [SD] age, 43.5 [10.3] years) by improvement group. Twenty-eight interviewees (51.9%) had diabetes, and 18 (33.3%) had cardiovascular disease. At baseline, 25 of 27 interviewees (92.6%) in the high-improvement group had highly elevated LDLC levels (>160 mg/dL) vs 2 of 27 (7.4%) in the nonimprovement group. By 12 months, the distribution had changed notably: 16 of the high-improvement interviewees (59.3%) had LDLC levels of less than 100 mg/dL, whereas 11 of the nonimprovement group interviewees (40.7%) had LDLC levels of greater than 160 mg/dL. Interviewees in the high-improvement group achieved a mean LDLC level decrease of 122 mg/dL compared with a mean increase of 33 mg/dL in the nonimprovement group and had higher statin therapy adherence during 6 months (standardized mean difference, 0.313).

    The Figure depicts recruitment for this qualitative study. We interviewed 54 individuals equally divided between high-improvement and nonimprovement groups. Each group included 12 control and 15 treatment participants, stratified equally across the 3 incentive groups.

    There were important socioeconomic and health differences between the groups. Ten interviewees in the high-improvement group (37.0%) were 60 years or older, 4 (14.8%) were Black, and 22 (81.5%) had incomes of $50 000 or more, compared with 15 (55.6%) who were 60 years or older, 16 (59.3%) who were Black, and 11 (40.7%) with incomes of $50 000 or more in the nonimprovement group. Three interviewees in the high-improvement group (11.1%) reported fair or poor health compared with 13 in the nonimprovement group (48.1%). Last, although some individuals in the high-improvement group had cardiovascular disease (5 [18.5%]) or diabetes (7 [25.9%]), almost half of the nonimprovement group (13 [48.1%]) had cardiovascular disease, and more than three-quarters (21 [77.8%]) had diabetes. These differences were consistent with those of the eligible pool of interviewees in the top and bottom deciles of LDLC level change, suggesting that our sample is representative based on observable characteristics (Table 2).

    Overall Health, Function, and Employment Status Reported in Interviews

    Six interviewees in the high-improvement group (22.2%) reported at least 2 chronic conditions compared with 13 in the nonimprovement group (48.1%). This greater burden of chronic illness accords with the lower self-reported health status in the nonimprovement group (Table 1). Despite these differences, interviewees from both groups described similar levels of functional limitation and dependency. However, individuals in the high-improvement group were more likely to work: 12 of 27 (44.4%) were full-time employees, 5 (18.5%) were part-time employees, 2 (7.4%) were retirees, and none were disability beneficiaries. In comparison, only 6 interviewees in the nonimprovement group (22.2%) reported working full-time, 1 (3.7%) worked part time, 2 (7.4%) were retired, and 3 (11.1%) received disability benefits.

    Recollection of Financial Incentives

    During the trial’s 6-month intervention period, intervention group participants received incentives for adherence and daily feedback about earnings. However, few interviewees who were enrolled in the study’s incentives groups, irrespective of improvement status, could describe the structure of rewards (see Table 3, section 1). Some participants did not distinguish between incentives for adherence and payments for study participation; as 1 interviewee in the nonimprovement group stated, “All I know is they paid me some for every so many months or something.”

    In contrast, those in the incentive groups in particular offered detailed recollections of the features of the electronic pill bottles. Recalled 1 trial participant, “[the pill bottle] wasn’t round. But it looked round, but yet it was square and it had a round top and it had a light that went around the bottom of it, so when you opened the top, the bottom of it lit up” (Table 3, section 2). Of note, no interviewees expressed concern about their adherence being monitored through the pill bottle.

    Interviewees were asked about the effect of the study earnings on their financial situations and pill-taking habits. Twenty-five of the 27 interviewees in the high-improvement group (92.6%) said that their study earnings did not change their regular spending habits, compared 15 of the 27 in the nonimprovement group (55.6%).

    Most respondents in the high-improvement group did not find study earnings a substantial motivator to take their statin (Table 3, section 3), explaining that they joined the study to get their cholesterol levels under control, which was a reward in itself. “I don’t think that there would have been a monetary amount that would have made a difference,” said a high-improvement participant. “Seeing my test results and feeling better made a difference.” However, interviewees in the nonimprovement group were more likely to characterize study earnings as a motivator (Table 3, section 4).

    Understanding and Prioritization of LDLC Level Management

    When articulating their concerns about having high LDLC levels, the language and viewpoints of the groups differed. These differences may point to discrepancies in health literacy and patient engagement between groups. Interviewees in the high-improvement group universally described their concerns in terms of pathology, often using medicalized language in noting that they worried about “stroke,” “heart disease,” “heart failure,” “plaque buildup,” “cardiovascular” problems, or exacerbation of chronic medical conditions, including some unrelated to LDLC level.

    Many interviewees in the nonimprovement group also used medical terms to articulate disease-centered concerns, but their responses were otherwise more varied. Unlike those in the high-improvement group, several interviewees said they currently were not concerned about their LDLC level or had only become concerned after being warned by a health professional that their cholesterol levels were worrisome. Use of colloquial language (eg, “clogged” arteries) to describe pathology was more frequent among interviewees in the nonimprovement group.

    When asked about family history of cholesterol issues, 14 interviewees in the nonimprovement group (51.9%) and 15 in the high-improvement group (55.6%) responded affirmatively. When further asked how their family history affected their approach to their own LDLC levels, interviewees in the high-improvement group more frequently stated that it inspired them to alter their health behaviors, with several specifically mentioning medication adherence.

    Dietary and Medication Habits

    Respondents in the high-improvement group more frequently mentioned specific nutritional changes they made to improve their LDLC levels than did those from the nonimprovement group (Table 4, section 1). Six of 7 employed interviewees in the nonimprovement group cited a lack of access to healthy foods in their work environments or an inability to find time to eat healthy foods due to hectic work schedules (Table 4, section 2). Conveyed 1 interviewee in the nonimprovement group: “I didn’t exactly live the greatest of lifestyle[s]. I worked shift work almost my entire life. When you’re working at night the only place open to eat at is WaWa, a gas station, or there was a diner that was in town.” In contrast, just 3 of the 17 employed interviewees in the high-improvement group described such difficulties.

    When asked about pill-taking routines before the study, interviewees in the high-improvement group largely characterized their pill-taking routine as well established, whereas those in the nonimprovement group more often described themselves as forgetful and unable to establish medication routines (Table 4, section 3). Discussing pill-taking behavior before the trial, a participant in the nonimprovement group recalled, “I can’t even remember. I really wasn’t doing anything, just sitting around getting fat.… I would forget to take it if it’s not on my mind. If I took it, I took it. Not really, because I forgot.” This difference carried through interviewees’ responses about which aspects of the trial were unhelpful. Eight interviewees in the high-improvement group cited the specialized pill bottle used in the trial, which they said interfered with their preestablished routines. An interviewee in the high-improvement group stated: “I’ll tell you that [using the study pill bottle] was harder to remember to take than the way I’ve always taken it.… I would have to remember to go to 2 places to take all my meds.… [I]t was even harder for me to remember with that bottle because I just have all of my evening meds in one place, and I take them every night without a problem.” In contrast, few interviewees in the nonimprovement group reported any aspect of the trial as particularly troublesome.

    Discussion

    Medication nonadherence is a major obstacle to achieving gains in health from safe and effective therapies such as statins. The high prevalence of medication nonadherence among patients with ASCVD and other diseases has motivated research to identify adherence barriers and strategies to overcome them.28,29 Behavioral economics points to the role of present bias, because the benefits of adherence are often delayed and may seem too far away and abstract to motivate patients.30,31 Present bias provides the rationale for trials of small financial incentives that provide immediate benefits for adherence and, if successful, prevent future deterioration in health and associated regret.12,32

    This qualitative study of our trial of financial incentives for statin adherence provides important insights into the context and living conditions of study participants that may have prevented the incentives from enabling a more substantial improvement in adherence routines.33 The differences found between high-improvement and nonimprovement groups suggest the influence of socioeconomic status on participants’ efforts to lower LDLC levels, consistent with prior research on medication adherence.34-41 Socioeconomic status is a complex collection of factors, including income, wealth, educational attainment, occupation, location, and social capital. Participants in the high-improvement group were more likely to be employed, evinced greater health literacy and engagement, had better access to healthy food at work and work schedules more amenable to healthy eating, were more able to make specific dietary changes, and had less difficulty establishing medication routines. They were also more likely to have an annual income of $50 000 or greater, to identify as White, to have lower rates of cardiovascular disease and diabetes, and to have self-reported better overall health yet with higher baseline LDLC levels. The same differences appeared when we compared the top and bottom deciles of LDLC level improvement among all participants in the underlying clinical trial, suggesting that the interviewees were representative of these groups.

    Our results indicate that individuals of lower socioeconomic status might face distinct challenges in responding to incentives and feedback. Specifically, the physical, logistical, cognitive, and emotional burdens of low socioeconomic status may have made it harder for these individuals to initiate new routines, manage their health, and engage with the study intervention. Even simple changes, such as taking a statin daily, appear to be difficult to initiate in this population, although they were well informed about their medication because they had existing prescriptions and access to routine care. More generous incentives might have been more salient for behavior modification and should be explored in future research, although higher incentives might not be cost-effective.

    Our interviews also revealed important aspects of patients’ pill-taking routines and the challenges that remote monitoring via electronic pill bottles may present. In this and a growing number of other trials, remote monitoring of medication adherence is the platform for delivering financial incentives, feedback, and other interventions.42 Many individuals in the high-improvement group reported that the pill bottles disrupted their medication routines, whereas the nonimprovement group did not report this problem. Instead, those in the nonimprovement group reported that the device and incentives were not sufficient by themselves to change their routines. Future studies should pay close attention to how electronic pill containers interact with participants’ established routines and provide coaching and health education to those who have clinical inertia and struggle to engage with interventions such as incentives and feedback.

    Although study participants were able to provide details about the electronic pill containers, most participants from the incentive groups had little memory of how and which behavior was financially rewarded. This finding stands in contrast to the common assumption that with proper incentives, it is in the interest of participants to understand rewards and how to qualify for them. It emphasizes the potential need to accompany these interventions with multipronged approaches addressing structural barriers to cholesterol level control (such as better access to healthy food and community engagement), as well as customized, motivational interviewing and health education, especially when participants do not show improvement. As noted, compared with a control condition, participants with incentives had significantly higher measured adherence but no difference in LDLC level change. Taken together, these findings and the qualitative data suggest that financial rewards for adherence led some participants to focus attention on the pill container but did not cause participants to remember the incentive structure itself. Follow-up studies need to explore what the barriers to comprehension are and how to overcome them to improve the design of interventions.

    Limitations

    This study has limitations. Interviewees’ ability to recall the details of the trial and their experiences participating in it may have been limited because the interviews were conducted more than 6 months after the incentives ended. The response rate was 41.9%; however, qualitative participants were comparable to eligible participants based on observable characteristics. Furthermore, in comparing the experiences of interviewees in the high-improvement and nonimprovement groups, we used 12-month change in LDLC level as a proxy for how successful participants were in the trial. Because the high-improvement group had significantly higher baseline LDLC levels, their much larger decrease in LDLC levels may have in part been owing to their simply having greater room for improvement. However, we affirm that the contrast between groups was valid, because the LDLC levels of most participants in the nonimprovement group actually increased (in stark contrast to the mean decline of >32 mg/dL in the overall trial). Furthermore, in light of the substantial experiential and demographic differences between the groups, we do not believe that the difference in baseline LDLC levels can be posited as the primary explanation of why the groups achieved such divergent outcomes.

    Conclusions

    This qualitative study suggests the need for a new approach to the provision of financial incentives. Whereas incentives theories assume that it is in the interest of individuals to manage changes autonomously to qualify for the highest financial rewards, this study highlights that participants with socioeconomic disadvantages and clinical burdens may need a more customized approach along with additional support to address, where possible, structural barriers to improving health.

    Back to top
    Article Information

    Accepted for Publication: March 17, 2021.

    Published: May 4, 2021. doi:10.1001/jamanetworkopen.2021.9211

    Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2021 Barankay I et al. JAMA Network Open.

    Corresponding Author: Iwan Barankay, MSc, PhD, Department of Management, The Wharton School, University of Pennsylvania, 3620 Locust Walk, Steinberg-Dietrich Hall, Ste 2000, Philadelphia, PA 19104 (barankay@wharton.upenn.edu).

    Author Contributions: Drs Barankay and Reese contributed as co–first authors. Drs Barankay and Clapp had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

    Concept and design: Barankay, Reese, Putt, Pagnotti, Chadha, Volpp, Clapp.

    Acquisition, analysis, or interpretation of data: Barankay, Reese, Putt, Russell, Phillips, Chadha, Oyekanmi, Yan, Zhu, Clapp.

    Drafting of the manuscript: Barankay, Reese, Putt, Russell, Phillips, Chadha, Clapp.

    Critical revision of the manuscript for important intellectual content: Barankay, Reese, Putt, Russell, Pagnotti, Oyekanmi, Yan, Zhu, Volpp, Clapp.

    Statistical analysis: Barankay, Putt, Chadha, Yan.

    Obtained funding: Barankay, Volpp.

    Administrative, technical, or material support: Pagnotti, Chadha, Oyekanmi, Volpp.

    Supervision: Barankay, Reese, Clapp.

    Conflict of Interest Disclosures: Dr Barankay reported receiving research support from Humana Inc outside the scope of the submitted work. Dr Reese reported receiving investigator-initiated grants from Merck & Co, Inc, to the University of Pennsylvania to support studies of medication adherence, grants from Merck & Co, Inc, and AbbVie to support clinical trials in transplantation, consulting fees from VAL Health, LLC (to identify patients with kidney disease), and personal fees from the American Journal of Kidney Diseases outside the submitted work. Dr Putt reported receiving grants from the National Institutes of Health (NIH) for financial support provided to Drs Barankay, Reese, and Volpp during the conduct of the study. Dr Russell reported receiving grants from the National Heart, Lung, and Blood Institute for financial support provided to Drs. Reese, Volpp, and Barankay during the conduct of the study. Dr Zhu reported receiving grants from the NIH during the conduct of the study. Dr Volpp reported receiving grants from National Institute of Aging and CVS Health during the conduct of the study and receiving personal fees and being a part-owner of VAL Health, LLC; receiving grants from Hawaii Medical Services Association, Vitality/Discovery, Humana Inc, and WW; and receiving personal fees from Center for Corporate Innovation, Lehigh Valley Medical Center, Vizient Inc, Greater Philadelphia Business Coalition on Health, American Gastroenterological Association Tech Conference, Bridge to Population Health Meeting, and Irish Medtech Summit outside the submitted work. No other disclosures were reported.

    Funding/Support: This work was supported by CVS Health and grant R01HL118195 from the NIH.

    Role of the Funder/Sponsor: The sponsors had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

    References
    1.
    Heron  M. Deaths: leading causes for 2016. National Vital Statistics Reports. Published July 26, 2018. Accessed May 7, 2020. https://www.cdc.gov/nchs/data/nvsr/nvsr67/nvsr67_06.pdf
    2.
     Randomised trial of cholesterol lowering in 4444 patients with coronary heart disease: the Scandinavian Simvastatin Survival Study (4S).   Lancet. 1994;344(8934):1383-1389.PubMedGoogle Scholar
    3.
    Maningat  P, Gordon  BR, Breslow  JL.  How do we improve patient compliance and adherence to long-term statin therapy?   Curr Atheroscler Rep. 2013;15(1):291. doi:10.1007/s11883-012-0291-7 PubMedGoogle ScholarCrossref
    4.
    Kearney  PM, Blackwell  L, Collins  R,  et al; Cholesterol Treatment Trialists’ (CTT) Collaborators.  Efficacy of cholesterol-lowering therapy in 18,686 people with diabetes in 14 randomised trials of statins: a meta-analysis.   Lancet. 2008;371(9607):117-125. doi:10.1016/S0140-6736(08)60104-X PubMedGoogle ScholarCrossref
    5.
    Mihaylova  B, Emberson  J, Blackwell  L,  et al; Cholesterol Treatment Trialists’ (CTT) Collaborators.  The effects of lowering LDL cholesterol with statin therapy in people at low risk of vascular disease: meta-analysis of individual data from 27 randomised trials.   Lancet. 2012;380(9841):581-590. doi:10.1016/S0140-6736(12)60367-5 PubMedGoogle ScholarCrossref
    6.
    Jackevicius  CA, Mamdani  M, Tu  JV.  Adherence with statin therapy in elderly patients with and without acute coronary syndromes.   JAMA. 2002;288(4):462-467. doi:10.1001/jama.288.4.462 PubMedGoogle ScholarCrossref
    7.
    Benner  JS, Glynn  RJ, Mogun  H, Neumann  PJ, Weinstein  MC, Avorn  J.  Long-term persistence in use of statin therapy in elderly patients.   JAMA. 2002;288(4):455-461. doi:10.1001/jama.288.4.455 PubMedGoogle ScholarCrossref
    8.
    Colantonio  LD, Huang  L, Monda  KL,  et al.  Adherence to high-intensity statins following a myocardial infarction hospitalization among Medicare beneficiaries.   JAMA Cardiol. 2017;2(8):890-895. doi:10.1001/jamacardio.2017.0911 PubMedGoogle ScholarCrossref
    9.
    Shah  ND, Dunlay  SM, Ting  HH,  et al.  Long-term medication adherence after myocardial infarction: experience of a community.   Am J Med. 2009;122(10):961.e7-961.e13. doi:10.1016/j.amjmed.2008.12.021PubMedGoogle ScholarCrossref
    10.
    Laibson  D.  Golden eggs and hyperbolic discounting.   Q J Econ. 1997;112(2):443-478. doi:10.1162/003355397555253 Google ScholarCrossref
    11.
    Augenblick  N, Niederle  M, Sprenger  C.  Working over time: dynamic inconsistency in real effort tasks.   The Quarterly Journal of Economics. 2015;130(3):1067-1115. doi:10.1093/qje/qjv020 Google ScholarCrossref
    12.
    Putt  ME, Reese  PP, Volpp  KG,  et al.  The Habit Formation trial of behavioral economic interventions to improve statin use and reduce the risk of cardiovascular disease: Rationale, design and methodologies.   Clin Trials. 2019;16(4):399-409. doi:10.1177/1740774519846852PubMedGoogle ScholarCrossref
    13.
    Barankay  I, Reese  PP, Putt  ME,  et al.  Effect of patient financial incentives on statin adherence and lipid control: a randomized clinical trial.   JAMA Netw Open. 2020;3(10):e2019429. doi:10.1001/jamanetworkopen.2020.19429 PubMedGoogle Scholar
    14.
    Bargh  JA, Chartrand  TL.  The unbearable automaticity of being.   Am Psychol. 1999;54(7):462-479. doi:10.1037/0003-066X.54.7.462 Google ScholarCrossref
    15.
    Schumpeter  J, Backhaus  U. The theory of economic development. In: Backhaus  J, ed.  Joseph Alois Schumpeter: Entrepreneurship, Style and Vision. Springer US; 2003:61-116. doi:10.1007/0-306-48082-4_3
    16.
    Nelson  RR, Winter SG.  An Evolutionary Theory of Economic Change. Belknap Press of Harvard University Press; 1982.
    17.
    Chetty  R, Szeidl  A.  Consumption commitments and habit formation.   Econometrica. 2016;84(2):855-890. doi:10.3982/ECTA9390 Google ScholarCrossref
    18.
    Ryder  HE, Heal  GM.  Optimal growth with intertemporally dependent preferences.   Rev Econ Studies. 1973;40(1):1-31. doi:10.2307/2296736 Google ScholarCrossref
    19.
    O’Cathain  A, Thomas  KJ, Drabble  SJ, Rudolph  A, Hewison  J.  What can qualitative research do for randomised controlled trials? a systematic mapping review.   BMJ Open. 2013;3(6):e002889. doi:10.1136/bmjopen-2013-002889 PubMedGoogle Scholar
    20.
    Ivankova  NV, Creswell  JW, Stick  SL.  Using mixed-methods sequential explanatory design: from theory to practice.   Field Meth. 2006;18(1):3-20. doi:10.1177/1525822X05282260 Google ScholarCrossref
    21.
    Penn Medicine. Facts and figures 2021 edition. Published 2020. Accessed March 1, 2021. https://www.pennmedicine.org/news/publications-and-special-projects/facts-and-figures
    22.
    Stone  NJ, Robinson  JG, Lichtenstein  AH,  et al; American College of Cardiology/American Heart Association Task Force on Practice Guidelines.  2013 ACC/AHA guideline on the treatment of blood cholesterol to reduce atherosclerotic cardiovascular risk in adults: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines.   Circulation. 2014;129(25)(suppl 2):S1-S45. doi:10.1161/01.cir.0000437738.63853.7a PubMedGoogle ScholarCrossref
    23.
    Creswell  J, Poth  C.  Qualitative Inquiry and Research Design: Choosing Among Five Approaches. 4th ed. SAGE Publications Inc.; 2018.
    24.
    Yeo  A, Legard  R, Keegan  J, Ward  K, Nicholls  CM, Lewis  J. In-depth interviews. In: Ritchie J, Lewis J, Nicholls CM, Ormtron R, eds.  Qualitative Research Practice: A Guide for Social Science Students and Researchers. 2nd ed: SAGE Publications; 2014:177-210.
    25.
    Hennink  MM, Kaiser  BN, Marconi  VC.  Code saturation versus meaning saturation: how many interviews are enough?   Qual Health Res. 2017;27(4):591-608. doi:10.1177/1049732316665344PubMedGoogle ScholarCrossref
    26.
    Tavory  I, Timmermans  S.  Abductive Analysis: Theorizing Qualitative Research. University of Chicago Press; 2014. doi:10.7208/chicago/9780226180458.001.0001
    27.
    Alvesson  M, Karreman  D.  Constructing mystery: empirical matters in theory development.   Acad Manage Rev. 2007;32(4):1265-1281. doi:10.5465/amr.2007.26586822 Google ScholarCrossref
    28.
    Osterberg  L, Blaschke  T.  Adherence to medication.   N Engl J Med. 2005;353(5):487-497. doi:10.1056/NEJMra050100 PubMedGoogle ScholarCrossref
    29.
    Conn  VS, Ruppar  TM.  Medication adherence outcomes of 771 intervention trials: systematic review and meta-analysis.   Prev Med. 2017;99:269-276. doi:10.1016/j.ypmed.2017.03.008 PubMedGoogle ScholarCrossref
    30.
    O’Donoghue  T, Rabin  M.  Present bias: lessons learned and to be learned.   Am Econ Rev. 2015;105(5):273-279. doi:10.1257/aer.p20151085 Google ScholarCrossref
    31.
    O'Donoghue  T, Rabin  M.  Doing it now or later.   Am Econ Rev. 1999;89(1):103-124. doi:10.1257/aer.89.1.103 Google ScholarCrossref
    32.
    Asch  DA, Troxel  AB, Stewart  WF,  et al.  Effect of financial incentives to physicians, patients, or both on lipid levels: a randomized clinical trial.   JAMA. 2015;314(18):1926-1935. doi:10.1001/jama.2015.14850 PubMedGoogle ScholarCrossref
    33.
    Gneezy  U, Meier  S, Rey-Biel  P.  When and why incentives (don’t) work to modify behavior.   J Econ Perspectives. 2011;25(4):191-210. doi:10.1257/jep.25.4.191 Google ScholarCrossref
    34.
    Tøttenborg  SS, Lange  P, Johnsen  SP, Nielsen  H, Ingebrigtsen  TS, Thomsen  RW.  Socioeconomic inequalities in adherence to inhaled maintenance medications and clinical prognosis of COPD.   Respir Med. 2016;119:160-167. doi:10.1016/j.rmed.2016.09.007 PubMedGoogle ScholarCrossref
    35.
    Yang  C, Hao  Z, Yu  D, Xu  Q, Zhang  L.  The prevalence rates of medication adherence and factors influencing adherence to antiepileptic drugs in children with epilepsy: a systematic review and meta analysis.   Epilepsy Res. 2018;142:88-99. doi:10.1016/j.eplepsyres.2018.03.018 PubMedGoogle ScholarCrossref
    36.
    Hamilton  AJ, Caskey  FJ, Casula  A, Inward  CD, Ben-Shlomo  Y.  Associations with wellbeing and medication adherence in young adults receiving kidney replacement therapy.   Clin J Am Soc Nephrol. 2018;13(11):1669-1679. doi:10.2215/CJN.02450218 PubMedGoogle ScholarCrossref
    37.
    Gast  A, Mathes  T.  Medication adherence influencing factors-an (updated) overview of systematic reviews.   Syst Rev. 2019;8(1):112. doi:10.1186/s13643-019-1014-8 PubMedGoogle ScholarCrossref
    38.
    Gaalema  DE, Elliott  RJ, Morford  ZH, Higgins  ST, Ades  PA.  Effect of socioeconomic status on propensity to change risk behaviors following myocardial infarction: implications for healthy lifestyle medicine.   Prog Cardiovasc Dis. 2017;60(1):159-168. doi:10.1016/j.pcad.2017.01.001 PubMedGoogle ScholarCrossref
    39.
    Nelson  LA, Ackerman  MT, Greevy  RA  Jr, Wallston  KA, Mayberry  LS.  Beyond race disparities: accounting for socioeconomic status in diabetes self-care.   Am J Prev Med. 2019;57(1):111-116. doi:10.1016/j.amepre.2019.02.013 PubMedGoogle ScholarCrossref
    40.
    Kim  J, Lee  E, Park  B-J, Bang  JH, Lee  JY.  Adherence to antiretroviral therapy and factors affecting low medication adherence among incident HIV-infected individuals during 2009-2016: a nationwide study.   Sci Rep. 2018;8(1):3133. doi:10.1038/s41598-018-21081-x PubMedGoogle ScholarCrossref
    41.
    Osborn  CY, Kripalani  S, Goggins  KM, Wallston  KA.  Financial strain is associated with medication nonadherence and worse self-rated health among cardiovascular patients.   J Health Care Poor Underserved. 2017;28(1):499-513. doi:10.1353/hpu.2017.0036 PubMedGoogle ScholarCrossref
    42.
    Patel  MS, Asch  DA, Volpp  KG.  Wearable devices as facilitators, not drivers, of health behavior change.   JAMA. 2015;313(5):459-460. doi:10.1001/jama.2014.14781 PubMedGoogle ScholarCrossref
    ×