[Show abstract][Hide abstract] ABSTRACT: Determination of comparative effectiveness in a randomized controlled trial requires consideration of an intervention’s comparative uptake (or acceptance) among randomized participants and the intervention’s comparative efficacy among participants who use their assigned intervention. If acceptance differs across interventions, then simple randomization of participants can result in post-randomization losses that introduce bias and limit statistical power.
We develop a novel preference-adaptive randomization procedure in which the allocation probabilities are updated based on the inverse of the relative acceptance rates among randomized participants in each arm. In simulation studies, we determine the optimal frequency with which to update the allocation probabilities based on the number of participants randomized. We illustrate the development and application of preference-adaptive randomization using a randomized controlled trial comparing the effectiveness of different financial incentive structures on prolonged smoking cessation.
Simulation studies indicated that preference-adaptive randomization performed best with frequent updating, accommodated differences in acceptance across arms, and performed well even if the initial values for the allocation probabilities were not equal to their true values. Updating the allocation probabilities after randomizing each participant minimized imbalances in the number of accepting participants across arms over time. In the smoking cessation trial, unexpectedly large differences in acceptance among arms required us to limit the allocation of participants to less acceptable interventions. Nonetheless, the procedure achieved equal numbers of accepting participants in the more acceptable arms, and balanced the characteristics of participants across assigned interventions.
Preference-adaptive randomization, coupled with analysis methods based on instrumental variables, can enhance the validity and generalizability of comparative effectiveness studies. In particular, preference-adaptive randomization augments statistical power by maintaining balanced sample sizes in efficacy analyses, while retaining the ability of randomization to balance covariates across arms in effectiveness analyses.
ClinicalTrials.gov, NCT01526265; https://clinicaltrials.gov/ct2/show/NCT01526265 31 January 2012
[Show abstract][Hide abstract] ABSTRACT: Advertisers can learn a lot from our digital footprints, including about our health. TIMOTHY LIBERT AND COLLEAGUES : argue that we need to tighten restrictions on such data for marketing purposes while looking at their potential to improve health.
[Show abstract][Hide abstract] ABSTRACT: Importance:
Financial incentives to physicians or patients are increasingly used, but their effectiveness is not well established.
To determine whether physician financial incentives, patient incentives, or shared physician and patient incentives are more effective than control in reducing levels of low-density lipoprotein cholesterol (LDL-C) among patients with high cardiovascular risk.
Design, setting, and participants:
Four-group, multicenter, cluster randomized clinical trial with a 12-month intervention conducted from 2011 to 2014 in 3 primary care practices in the northeastern United States. Three hundred forty eligible primary care physicians (PCPs) were enrolled from a pool of 421. Of 25 627 potentially eligible patients of those PCPs, 1503 enrolled. Patients aged 18 to 80 years were eligible if they had a 10-year Framingham Risk Score (FRS) of 20% or greater, had coronary artery disease equivalents with LDL-C levels of 120 mg/dL or greater, or had an FRS of 10% to 20% with LDL-C levels of 140 mg/dL or greater. Investigators were blinded to study group, but participants were not.
Primary care physicians were randomly assigned to control, physician incentives, patient incentives, or shared physician-patient incentives. Physicians in the physician incentives group were eligible to receive up to $1024 per enrolled patient meeting LDL-C goals. Patients in the patient incentives group were eligible for the same amount, distributed through daily lotteries tied to medication adherence. Physicians and patients in the shared incentives group shared these incentives. Physicians and patients in the control group received no incentives tied to outcomes, but all patient participants received up to $355 each for trial participation.
Main outcomes and measures:
Change in LDL-C level at 12 months.
Patients in the shared physician-patient incentives group achieved a mean reduction in LDL-C of 33.6 mg/dL (95% CI, 30.1-37.1; baseline, 160.1 mg/dL; 12 months, 126.4 mg/dL); those in physician incentives achieved a mean reduction of 27.9 mg/dL (95% CI, 24.9-31.0; baseline, 159.9 mg/dL; 12 months, 132.0 mg/dL); those in patient incentives achieved a mean reduction of 25.1 mg/dL (95% CI, 21.6-28.5; baseline, 160.6 mg/dL; 12 months, 135.5 mg/dL); and those in the control group achieved a mean reduction of 25.1 mg/dL (95% CI, 21.7-28.5; baseline, 161.5 mg/dL; 12 months, 136.4 mg/dL; P < .001 for comparison of all 4 groups). Only patients in the shared physician-patient incentives group achieved reductions in LDL-C levels statistically different from those in the control group (8.5 mg/dL; 95% CI, 3.8-13.3; P = .002).
Conclusions and relevance:
In primary care practices, shared financial incentives for physicians and patients, but not incentives to physicians or patients alone, resulted in a statistically significant difference in reduction of LDL-C levels at 12 months. This reduction was modest, however, and further information is needed to understand whether this approach represents good value.
JAMA The Journal of the American Medical Association 11/2015; 314(18):1926-1935. DOI:10.1001/jama.2015.14850 · 35.29 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: Background:
Social media may offer insight into the relationship between an individual's health and their everyday life, as well as attitudes towards health and the perceived quality of healthcare services.
To determine the acceptability to patients and potential utility to researchers of a database linking patients' social media content with their electronic medical record (EMR) data.
Adult Facebook/Twitter users who presented to an emergency department were queried about their willingness to share their social media data and EMR data with health researchers for the purpose of building a databank for research purposes. Shared posts were searched for select terms about health and healthcare.
Of the 5256 patients approached, 2717 (52%) were Facebook and/or Twitter users. 1432 (53%) of those patients agreed to participate in the study. Of these participants, 1008 (71%) consented to share their social media data for the purposes of comparing it with their EMR. Social media data consisted of 1 395 720 posts/tweets to Facebook and Twitter. Participants sharing social media data were slightly younger (29.1±9.8 vs 31.9±10.4 years old; p<0.001), more likely to post at least once a day (42% vs 29%; p=0.003) and more likely to present to the emergency room via self-arrival mode and have private insurance. Of Facebook posts, 7.5% (95% CI 4.8% to 10.2%) were related to health. Individuals with a given diagnosis in their EMR were significantly more likely to use terms related to that diagnosis on Facebook than patients without that diagnosis in their EMR (p<0.0008).
Many patients are willing to share and link their social media data with EMR data. Sharing patients have several demographic and clinical differences compared with non-sharers. A database that merges social media with EMR data has the potential to provide insights about individuals' health and health outcomes.
[Show abstract][Hide abstract] ABSTRACT: Adverse and suboptimal health behaviors and habits are responsible for approximately 40 % of preventable deaths, in addition to their unfavorable effects on quality of life and economics. Our current understanding of human behavior is largely based on static “snapshots” of human behavior, rather than ongoing, dynamic feedback loops of behavior in response to ever-changing biological, social, personal, and environmental states. This paper first discusses how new technologies (i.e., mobile sensors, smartphones, ubiquitous computing, and cloud-enabled processing/computing) and emerging systems modeling techniques enable the development of new, dynamic, and empirical models of human behavior that could facilitate just-in-time adaptive, scalable interventions. The paper then describes concrete steps to the creation of robust dynamic mathematical models of behavior including: (1) establishing “gold standard” measures, (2) the creation of a behavioral ontology for shared language and understanding tools that both enable dynamic theorizing across disciplines, (3) the development of data sharing resources, and (4) facilitating improved sharing of mathematical models and tools to support rapid aggregation of the models. We conclude with the discussion of what might be incorporated into a “knowledge commons,” which could help to bring together these disparate activities into a unified system and structure for organizing knowledge about behavior.
Translational Behavioral Medicine 09/2015; 5(3). DOI:10.1007/s13142-015-0324-1
[Show abstract][Hide abstract] ABSTRACT: The experiences and behaviors revealed in our everyday lives provide as much insight into health and disease as any analysis of our genome could ever produce. These characteristics are not found in the genome, but may be revealed in our online activities, which make up our social mediome.
Published by Elsevier Ltd.
Trends in Molecular Medicine 07/2015; 21(9). DOI:10.1016/j.molmed.2015.06.004 · 9.45 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: Primary care provider (PCP) turnover is common and can disrupt patient continuity of care. Little is known about the effect of PCP turnover on patient care experience and quality of care.
To measure the effect of PCP turnover on patient experiences of care and ambulatory care quality.
Observational, retrospective cohort study of a nationwide sample of primary care patients in the Veterans Health Administration (VHA). We included all patients enrolled in primary care at the VHA between 2010 and 2012 included in 1 of 2 national data sets used to measure our outcome variables: 326 374 patients in the Survey of Healthcare Experiences of Patients (SHEP; used to measure patient experience of care) associated with 8441 PCPs and 184 501 patients in the External Peer Review Program (EPRP; used to measure ambulatory care quality) associated with 6973 PCPs.
Whether a patient experienced PCP turnover, defined as a patient whose provider (physician, nurse practitioner, or physician assistant) had left the VHA (ie, had no patient encounters for 12 months).
Five patient care experience measures (from SHEP) and 11 measures of quality of ambulatory care (from EPRP).
Nine percent of patients experienced a PCP turnover in our study sample. Primary care provider turnover was associated with a worse rating in each domain of patient care experience. Turnover was associated with a reduced likelihood of having a positive rating of their personal physician of 68.2% vs 74.6% (adjusted percentage point difference, -5.3; 95% CI, -6.0 to -4.7) and a reduced likelihood of getting care quickly of 36.5% vs 38.5% (adjusted percentage point difference, -1.1; 95% CI, -2.1 to -0.1). In contrast, PCP turnover was not associated with lower quality of ambulatory care except for a lower likelihood of controlling blood pressure of 78.7% vs 80.4% (adjusted percentage point difference, -1.44; 95% CI, -2.2 to -0.7). In 9 measures of ambulatory care quality, the difference between patients who experienced no PCP turnover and those who had a PCP turnover was less than 1 percentage point. These effects were moderated by the patients' continuity with their PCP prior to turnover, with a larger detrimental effect of PCP turnover among those with higher continuity prior to the turnover.
Primary care provider turnover was associated with worse patient experiences of care but did not have a major effect on ambulatory care quality.
JAMA Internal Medicine 05/2015; 175(7). DOI:10.1001/jamainternmed.2015.1853 · 13.12 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: Background:
Financial incentives promote many health behaviors, but effective ways to deliver health incentives remain uncertain.
We randomly assigned CVS Caremark employees and their relatives and friends to one of four incentive programs or to usual care for smoking cessation. Two of the incentive programs targeted individuals, and two targeted groups of six participants. One of the individual-oriented programs and one of the group-oriented programs entailed rewards of approximately $800 for smoking cessation; the others entailed refundable deposits of $150 plus $650 in reward payments for successful participants. Usual care included informational resources and free smoking-cessation aids.
Overall, 2538 participants were enrolled. Of those assigned to reward-based programs, 90.0% accepted the assignment, as compared with 13.7% of those assigned to deposit-based programs (P<0.001). In intention-to-treat analyses, rates of sustained abstinence from smoking through 6 months were higher with each of the four incentive programs (range, 9.4 to 16.0%) than with usual care (6.0%) (P<0.05 for all comparisons); the superiority of reward-based programs was sustained through 12 months. Group-oriented and individual-oriented programs were associated with similar 6-month abstinence rates (13.7% and 12.1%, respectively; P=0.29). Reward-based programs were associated with higher abstinence rates than deposit-based programs (15.7% vs. 10.2%, P<0.001). However, in instrumental-variable analyses that accounted for differential acceptance, the rate of abstinence at 6 months was 13.2 percentage points (95% confidence interval, 3.1 to 22.8) higher in the deposit-based programs than in the reward-based programs among the estimated 13.7% of the participants who would accept participation in either type of program.
Reward-based programs were much more commonly accepted than deposit-based programs, leading to higher rates of sustained abstinence from smoking. Group-oriented incentive programs were no more effective than individual-oriented programs. (Funded by the National Institutes of Health and CVS Caremark; ClinicalTrials.gov number, NCT01526265.).
New England Journal of Medicine 05/2015; 372(22). DOI:10.1056/NEJMoa1414293 · 55.87 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: Objective . To evaluate the use of behavioral economics to design financial incentives to promote health behavior change and to explore associations with demographic characteristics. Data Source . Studies performed by the Center for Health Incentives and Behavioral Economics at the University of Pennsylvania published between January 2006 and March 2014. Study Inclusion and Exclusion Criteria . Randomized, controlled trials with available participant-level data. Studies that did not use financial incentives to promote health behavior change were excluded. Data Extraction . Participant-level data from seven studies were pooled. Data Synthesis . Meta-analysis on the pooled sample using a random-effects model with interaction terms to examine treatment effects and whether they varied by incentive structure or demographic characteristics. Results . The pooled study sample comprised 1403 participants, of whom 35% were female, 70% were white, 24% were black, and the mean age was 48 years (standard deviation 11.2 years). In the fully adjusted model, participants offered financial incentives had higher odds of behavior change (odds ratio [OR]: 3.96; p < .01) when compared to control. There were no significant interactions between financial incentives and gender, age, race, income, or education. When further adjusting for incentive structure, blacks had higher odds than whites of achieving behavior change (OR: 1.67; p < .05) with a conditional payment. Compared to lower-income participants, higher-income participants had lower odds of behavior change (OR: 0.46; p = .01) with a regret lottery. Conclusion . Financial incentives designed using concepts from behavioral economics were effective for promoting health behavior change. There were no large and consistent relationships between the effectiveness of financial incentives and observable demographic characteristics. Second-order examinations of incentive structure suggest potential relationships among the effectiveness of financial incentives, incentive structure, and the demographic characteristics of race and income.
American journal of health promotion: AJHP 05/2015; 29(5):314-23. DOI:10.4278/ajhp.140714-LIT-333 · 2.37 Impact Factor
[Show abstract][Hide abstract] ABSTRACT: As hard as it may be for clinicians to adopt new practices, it is often harder for them to "de-innovate," or give up old practices, even when new evidence reveals that those practices offer little value. In this article we explore recent controversies over screening for breast and prostate cancer and testing for sleep disorders. We show that these controversies are not caused solely by a lack of clinical data on the harms and benefits of these tests but are also influenced by several psychological biases that make it difficult for clinicians to de-innovate. De-innovation could be fostered by making sure that advisory panels and guideline committees include experts who have competing biases; emphasizing evidence over clinical judgment; resisting "indication creep," or the premature extension of innovations into unproven areas; and encouraging clinicians to explicitly consider how their experiences bias their interpretations of clinical evidence.
Project HOPE—The People-to-People Health Foundation, Inc.
Health Affairs 02/2015; 34(2):239-44. DOI:10.1377/hlthaff.2014.0983 · 4.97 Impact Factor