Soft and hard commitment devices to increase HIV testing

We will investigate whether a soft commitment device (a scheduled appointment) and a hard commitment device (small money transfer conditional on HIV testing) can increase HIV testing among men in Zomba Town, Malawi.

We believe that procrastination (a tendency to delay) is a barrier to HIV testing for many people. We will assess the impact of two interventions that aim at reducing procrastination and thereby overcome the problem of continuously putting off HIV testing.

(1) A soft commitment device: selected participants will be offered scheduled appointments to test for HIV. Appointments are devices for mental and social commitment.

(2) A hard commitment device: selected participants will be invited to invest in a mechanism that conditions a small transfer of money on testing for HIV.

The scripts of the interventions are described in the questionnaire.

2019-06-21

2019-09-30

Our primary outcome variable is a testing dummy collected at clinics when participants redeem their voucher. The testing dummy will be equal to 1 for participants who redeemed their voucher and then agreed to get tested, and equal to 0 otherwise.

We will consider two secondary outcomes collected at clinics:

(1) A "tested HIV+ dummy": a dummy variable equal to 1 for participants who redeemed their voucher and tested HIV+, and equal to 0 otherwise.

(2) An "ART dummy": a dummy variable equal to 1 for participants who redeemed their voucher and agreed to initiate ARV treatment after testing HIV+, and equal to 0 otherwise.

These outcomes are considered as secondary not because they are uninteresting (quite the contrary) but rather because power will most likely be insufficient to detect any effect.

We use a factorial design.

Participants are split into four treatment arms. They are randomized at the individual level without stratification.

All participants receive a 1000K gift of airtime at the end of the survey, and a 500K airtime voucher for redemption at any HIV testing site in Zomba Town. They do not have to agree to an HIV test to redeem this voucher.

Prior to applying the results of the randomization, we will elicit preference for a hard commitment device from each participant.

0: 25% are assigned to the control group.

1: 25% of participants will then be offered the hard commitment device. The original preference we elicited from all participants will be honoured. If their preference was to refuse the commitment device, they receive an immediate payment of 1000K. If their preference was to accept, they forego the immediate payment of 1000K, but can receive a 1000K transfer if they show up at one of the HIV testing sites. They do not have to agree to an HIV test.

2: 25% of participants will be offered a scheduled appointment to test for HIV. They are allowed to refuse the appointment.

3: 25% of participants will be offered both the hard commitment device and the appointment.

The effect of the soft commitment device (being offered an appointment) can measured by comparing those randomly selected to get an appointment (groups 2 and 3) to those who were not (groups 0 and 1).

The intent-to-treat effect of offering the hard commitment device can measured by comparing those who were assigned to groups 1 and 3 to those assigned to groups 0 and 2. We can also measure the effect on compliers using the same comparison, but restricting our sample to those who preferred the hard commitment device.

The randomization process will be predefined in a .csv downloaded on survey tablets, and hence reproducible. Randomization will be at the individual level without stratification. Each study participant will be automatically allocated to a treatment arm based on their Study ID which is composed of their enumerator ID, day of the study, and number of the survey. For example, the first enumerator’s first respondent on the first day of the study will have Study ID 010101.

Individual

No

No cluster

1200 individuals

For the soft commitment device intervention (appointments), we will have 600 participants in the treatment group and 600 participants in the control group.

For the hard commitment device intervention (conditional money transfer) , we will have 600 participants in the treatment group and 600 participants in the control group for the intent-to-treat analysis. When calculating the average treatment effect on the treated, the sample size will depend on the take-up rate of the hard commitment device: denoting this take-up rate r, 600*r participants will be in the treatment group and 600*r participants will be in the control group for this analysis.

PAP

We show that ordinary appointments can act as effective substitutes for hard commitment devices and increase demand for a critical healthcare service, particularly among those with self-control problems. We show this using an experiment that randomly offered HIV testing appointments and hard commitment devices to high-risk men in Malawi. Appointments more than double testing rates, with effects concentrated among those who demand commitment. In contrast, most men who take up hard commitments lose their investments. Appointments overcome commitment problems without the potential drawback of commitment failure, and have the potential to increase demand for healthcare in the developing world.

Derksen, Laura, Jason Kerwin, Natalia Ordaz Reynoso, and Olivier Sterck. (2023). Healthcare Appointments as Commitment Devices. Working Paper.

https://www.jasonkerwin.com/Papers/Appointments/DKOS_Appointments_Latest.pdf