Small Incentives and Nudges for HIV Testing

Background: The proportion of HIV-infected people who are undiagnosed has held constant at twenty percent; many who are newly diagnosed are late diagnoses.  In 2006, the Centers for Disease Control and Prevention (CDC) issued new HIV screening guidelines, endorsing universal routine opt-out testing, including in emergency rooms.  Several trials have shown that such testing programs are feasible and have yielded a wide range of test acceptance rates; how to best implement the CDC guidelines remains unclear. This study examines several HIV testing policies.  It is the first study to measure differences in take-up rates across a variety of approaches within a unified randomized trial using a factorial design.  In this field experiment, patients are offered HIV tests and a questionnaire by research assistants in parallel with their routine care in the ED.

Methods: This study tests three defaults: informed opt-in (test patients only if they request testing), opt-out (test patients unless they decline), and active-choice testing (patients are required to state whether they want to be tested). The study also tests the effect of small monetary incentives (none, $1, $5, $10) on test take-up.  All subjects offered an HIV test are offered a questionnaire designed to elicit risk of HIV infection.  Testing default and questionnaire timing (before or after the offer of the HIV test) are randomized at the patient level; monetary incentives are randomized by ED zone (four zones) on a daily basis.

Results:  8,571 of 10,630 approached patients consented to study inclusion; 4,852 (56.6%) of participants accepted an HIV test.  Opt-in testing yielded a 43.7% acceptance rate.  In comparison, active choice yielded a 55.2% acceptance rate and opt out yielded a 67.7% acceptance rate. Monetary incentives of $1, $5, and $10 increased test acceptance rates by 1, 10, and 15 percentage points, respectively, compared to a rate of 51.6% with no incentive.

Using self-reported risk of infection, we categorized patients into low, intermediate, and high risk groups.  Test acceptance rates increased with increasing risk (49.2%, 61.6%, and 69.3%), but defaults and incentives affected risk groups similarly.  

Conclusion:  Small incentives and changes in test offer wording can have large effects, highlighting the importance of carefully approaching patient-centered decision-making.  Opt-out regimens can substantially raise HIV testing rates, even after controlling for information about test availability, and does so equally for both low-risk and high-risk subjects.  Active choice presents a compromise approach in contexts where coercion is a concern.  A $1 monetary incentive is not enough to change test acceptance rates.  $5 and $10 incentives increase test acceptance rates, but not as much as changing the offer wording from opt in to opt out, assuaging concerns that small cash offers are coercive rather than encouraging.  Our findings suggest that defaults and monetary incentives can be used to increase test acceptance rates across all risk groups to help identify occult infection.