NUDGES TOWARD LONGER EXERCISE COMMITMENTS LEAD TO MORE EXERCISE: RESULTS OF A RANDOMIZED TRIAL OF COMMITMENT CONTRACT DESIGN

Wednesday, October 23, 2013
Key Ballroom Foyer (Hilton Baltimore)
Poster Board # P4-5
Applied Health Economics (AHE)

Jeremy D. Goldhaber-Fiebert, PhD1, Alan M. Garber, MD, PhD2 and Jay Bhattacharya, MD, PhD1, (1)Stanford University, Stanford, CA, (2)Office of the President and Provost, Cambridge, MA
Purpose: Growing health burdens and costs are linked to sedentary lifestyles. Many at-risk individuals try repeatedly to exercise more but have difficulty establishing exercise habits. We examined whether exercise commitment contracts and nudges could induce increased exercise and increase post-contract exercise demand.

Methods: We studied StickK.com users considering exercise commitment contracts (October 2010 and April 2012). Users selected a contract duration (weeks); exercise frequency; and financial penalties for each failed week. We randomly assigned the default duration shown to be 8 weeks, 12 weeks, or 20 weeks (nudges) which users could easily change. Analysis 1 (n=8,809) examined how larger nudges increased chosen duration and altered other contract features. Analysis 2 (n=3,397) included users with contracts created through October 2011 to examine whether larger nudges increased the amount of exercise completed within the contract. Analysis 3 (n=4,354) included users with contracts ending by March, 2012 and examined the rate of signing subsequent contracts and their duration, examining whether larger nudges in the first contract induced learning and subsequent demand for exercise commitment.

Results: Analysis 1: larger nudges increased contracted duration (12.5, 14.0, and 18.8 weeks for 8-week, 12-week, and 20-week nudges) but did not change the likelihood of signing the contract or other contract features. Analysis 2: users selecting a duration equal to the nudge were more likely to drop out (OR 1.8 [95% CI: 1.4-2.2]) regardless of the nudge’s value. Despite higher drop-out in this group, overall exercise completed under the 20-week nudge was higher than under the 8-week nudge (8.0 vs. 6.1 weeks, p<0.001), and for those not dropping out, choosing 20 weeks under the 20-week nudge resulted in nearly the same amount of exercise as choosing 20 weeks with an 8-week nudge. Analysis 3: individuals who were randomized to the 20-week nudge but did not select 20 weeks duration evidenced learning, signing subsequent contracts at higher rates than those randomized to 8-week nudges (HR 1.6 [95% CI: 1.1-2.3]), with subsequent contracts of longer duration on average.

Conclusions: Nudges reinforce exercise habit formation. Responses to nudges predict subsequent behavior and drop-out rates, a feature useful for targeting additional interventions and support. Although commitment contracts and nudges support formation of exercise habits, long-term habit maintenance may require ongoing interventions such as repeated commitment contracts.