The TESTsmART Trial consists of two main aims. The overall goal of the two aims is to investigate the impact of malaria rapid diagnostic test (mRDT) subsidies and conditional artemisinin combination therapy (ACT) subsidies on the testing and treatment behavior of participants seeking care for their febrile illness in the private retail sector. Conditional ACT subsidies are discounts on quality-assured ACTs which are linked to the results of a malaria rapid diagnostic test administered at the retail outlet; only participants with a positive test will have access to an additional discount on a quality-assured ACT.

The main objective of Aim 1 of this study is to identify a combination of conditional ACT and RDT subsidies that maximizes the proportion of participants that choose to have a malaria diagnostic test before taking a drug. The investigators will test two levels of conditional ACT subsidy (100% subsidy versus ~67% subsidy) and two levels of RDT subsidy (0% subsidy and 50% subsidy) in a factorial designed experiment. Because dose size and therefore the price of an ACT course are dependent upon patient age, the ACT subsidy amount will also be scaled with patient age. These subsidy levels were chosen to keep the estimated program cost of the combined subsidy within $0.30-0.60 USD per person (assuming 100% testing uptake and between 20-40% of participants having a positive RDT). These estimates represent an upper bound since testing is unlikely to reach 100%. Current subsidy levels for ACT costs the program between 1.30-2.50 USD per treatment, with more than a third of that investment spent on individuals without malaria.

Individuals presenting to a retail outlet for a treatment of a fever or suspected malaria illness will be randomized to one of the four groups in equal proportions. A total of 840 participants will be enrolled (210 per arm). Their choices concerning uptake of testing and drug purchase will be recorded. The main outcome will be the proportion of participants that choose to take a test. Secondary outcomes include the proportion of participants who adhered to the results of the RDT among those who were tested (used ACT when positive and did not use an ACT when negative or without a test). The results of this study will be used to inform the subsidy levels in the intervention for Aim 2 of this trial.

RATIONALE:

From previous studies and implementation experiences, we know that consumption of ACTs increases as the price declines. Declining prices of ACTs create a trade-off between access and targeting; lower prices improve uptake of effective therapies by those with malaria but also increase inappropriate use by those without malaria. Curbing inappropriate use and targeting ACTs to malaria cases requires parasitological diagnosis which is virtually absent in the retail sector. It has also been shown that RDTs can be safely deployed in the retail sector. Most clients will agree to have an RDT if it is free. However, uptake of RDTs is sensitive to the price of the RDT. Previous work has not specifically evaluated the relationship between the price of the drug and the price of the test, but available evidence suggests that uptake of the RDT is sensitive to the price of the ACT as well.

In this work, we will link these two commodities through a diagnosis-dependent ACT subsidy. Access to the additional ACT subsidy depends on having a positive RDT. What we need to understand before scaling up a conditional subsidy is how the price of these two commodities should be related. Should the conditionally-subsidized ACT be less expensive than the RDT? Must the RDT be considerably less expensive than the retail price of ACT in order to motivate people to purchase a test?

OBJECTIVES:

In this study (Aim 1) we will use an individually-randomized experiment conducted among customers at medicine retail outlets to identify the combination of RDT subsidy-level and conditional ACT subsidy-level that maximizes uptake of diagnostic testing. We choose to focus on diagnostic testing because this is the first step to achieving the downstream goals of ACT targeting and rational use.

STUDY DESIGN:

We will use a factorial design to test two ACT subsidy levels and two RDT subsidy levels. The unit of randomization will be the individual customer.

STUDY POPULATION:

This study will be carried out in a sample of retail shops that carry quality-assured ACTs in our study area in western Kenya. Ten shops will be randomly selected to participate in the study. The study population will be any individual presenting to the shop with a malaria-like illness. Children older than 1 year of age are eligible to be enrolled provided they are physically present and accompanied by a parent or legal guardian. Customers with a prescription from a health facility, who have already received a malaria diagnostic test or who have already taken antimalarials prior to coming to the outlet will be excluded. Individuals who have signs of severe disease will be excluded and referred immediately to a health facility for care.

STUDY PROCEDURES:

A research assistant (RA) will be stationed at participating outlets on random days in order to avoid influencing treatment seeking behavior - in other words, to avoid attracting customers because of the study team's presence. The RA will obtain consent and offer participants a scratch card with a secret subsidy offer that will be revealed after the participant is enrolled. Using the scratch card, the participants will be randomized, in a 1:1:1:1 ratio, to one of four study arms: 1) No subsidy for RDT (price to consumer=$0.40); 100% ACT subsidy (price to consumer=0) // 2) No subsidy for RDT (price to consumer=$0.40); 67% ACT subsidy (price to consumer=$0.40) // 3) 50% subsidy for RDT (price to consumer=$0.20); 100% ACT subsidy (price to consumer=0) // 4) 50% subsidy for RDT (price to consumer=$0.20); 67% ACT subsidy (price to consumer=$0.10-$0.40, dependent on patient age).

These four arms represent a 2x2 factorial experiment.

If the participant chooses to purchase a test at their assigned price (subsidy level), the outlet will collect the money and the RA will perform the test. (RAs have been trained in RDTs and blood safety and have conducted thousands of RDTs. If the test is positive, the participant is entitled to an additional discount on their ACT purchase according to their group identified on the scratch card. If the test is negative, the participant may purchase any medicine they choose, including a regularly-priced ACT. Those who opt not to purchase an RDT may continue with their transaction as they choose, including purchasing a regularly-priced ACT.

The outlet attendant will sell the medicines to the customer, including a discounted ACT, if eligible. The study team will reimburse the outlet the difference between the retail price and the discounted price.

The RDTs selected for the study will be a World Health Organization (WHO) approved product that exceed 95% sensitivity and 95% specificity for Plasmodium falciparum [Malaria Rapid Diagnostic Test Performance, Round 1-5, WHO 2014].

STUDY OUTCOME MEASURES:

The primary outcome for Aim 1 is the customer's decision to purchase an RDT (yes/no). Using the 2x2 factorial design we will separately evaluate the effect of RDT price (2 levels) and of conditional ACT subsidies (2 levels) on the primary outcome.

The main secondary outcome is the proportion of tested participants who are adherent to the test result among those tested. Adherence is defined as taking a quality-assured ACT if the RDT is positive or taking another drug (or no drug) if the test is negative. We will also measure the proportion of people who purchase a full-price ACT among those who do not use an RDT.

SAMPLE SIZE:

We estimated the sample size required in each of the four study arms for a design with equal numbers of individuals allocated to each arm and where we wish to detect a 15-percentage point increase in RDT testing between an unsubsidized RDT compared to a subsidized RDT and to detect a 10 percentage-point difference in testing uptake between a partially and a fully subsidized ACT (conditional on a positive RDT) with at least 90% power and 5% chance of a two-tailed Type I error for each of the two comparisons. To do so, we estimate that we will need 210 subjects per arm (total=840).

ENROLLMENT AND FOLLOW UP:

All participants will be screened and enrolled on the day they visit the outlet. Data collection will be brief and will be completed on the same day. Participants will be screened as they arrive, enrolled if eligible and willing, tested if they choose and then allowed to proceed with their transaction at the outlet. Upon completion, they will be briefly interviewed again before leaving. Because the interaction with the participant is short and completed in one encounter, we expect minimal loss to follow-up (for example, participants leaving before the final questions).

Inclusion Criteria:

• Participants with fever or history of fever or malaria like illness
• Individual with malaria-like illness must be present at recruitment
• 1 year of age or older

Exclusion Criteria:

• Any individual with signs of severe illness requiring immediate referral
• Individuals who have taken an antimalarial in the last seven days, including for the current illness
• Individuals who already have a prescription from a facility or medical provider
• Pregnant women will be enrolled and offered an mRDT, but will be advised to seek treatment through a health care provider.