Placebo-Controlled PrEP Trial in a No-Access District
Ministry of Health officials in a country with 6.1% adult HIV prevalence approve a pre-exposure prophylaxis trial. Daily oral tenofovir is licensed and available in the capital, but the ministry funds it for only about 9,000 of an estimated 210,000 people at substantial risk; in the three trial districts, no one currently receives it. A European university proposes a randomized trial of 1,800 participants comparing a new long-acting injectable to placebo, with all participants receiving condoms, counselling and quarterly testing. The IRB is split 5–4. One member argues placebo is impermissible because a proven effective method exists; another argues that requiring the sponsor to supply tenofovir would make the trial unaffordable and leave the districts with nothing at all, and that the comparison answers a question local clinicians genuinely face. The sponsor has offered free injectable for three years post-trial to participants who want it, but has said nothing about tenofovir.
1. Assess the design under clinical equipoise and the standard-of-care requirement. Is the placebo arm defensible?
Read the worked answer
Clinical equipoise asks whether there is genuine uncertainty among experts about the comparative merits of the arms for the population studied. For tenofovir versus placebo there is none: tenofovir's efficacy is established and consensus. So the trial is not licensed by equipoise about the drug; it is licensed, if at all, only by the gap between the global standard of care and the local one.
The standard-of-care requirement - as in the Declaration of Helsinki's restriction of placebo to cases where no proven intervention exists, or where compelling scientific reasons make placebo necessary and participants are not exposed to serious harm - sets a high bar. The argument for the placebo arm is that the relevant comparator is what the districts actually receive, which is nothing. The argument against is that the districts' lack of access is a distributive failure that the trial would exploit rather than remedy, and that participants are being asked to bear risk that would be unnecessary if the sponsor supplied tenofovir.
The correct judgement is that the placebo arm is defensible only if two conditions hold: the question cannot be answered against an active comparator, and the population receives a fair share of the resulting benefits. The three-year injectable offer to participants addresses the second only partially.
2. Is the design exploitative? Use a transactional account of exploitation and a fair-benefits account, and say where they diverge.
Read the worked answer
A transactional account (Wertheimer) holds that A exploits B when A gains disproportionately from a transaction and B does not receive a fair share of the benefit, even if the transaction is consensual and mutually advantageous. On this account the trial can be exploitative without being coercive: participants gain access to testing and counselling they would not otherwise have, the sponsor gains data and eventual licensing revenue, and the imbalance in the division of benefits is the wrong-making feature. This account permits placebo controls if the overall package is fair.
A fair-benefits account shifts the object of assessment from the individual transaction to the distribution of benefits across the community: the relevant question is whether the community, not merely the enrolled individuals, receives a fair share - capacity building, post-trial access, pricing commitments.
They diverge on cases like this one. The transactional view can clear the trial if each participant's package is net beneficial against the local baseline. The fair-benefits view can condemn it even when every participant benefits, because the districts bear risk and receive no durable access. A power-based account adds a third layer: the ministry's inability to fund tenofovir is itself a background injustice, and the trial's permissibility depends on whether it mitigates or entrenches that dependence.
3. What design decision would satisfy the strongest objections without making the trial unaffordable?
Read the worked answer
Several designs relax the dilemma rather than resolving it by fiat, and each has a cost.
First, an active-comparator non-inferiority design against tenofovir, with the sponsor supplying tenofovir to all arms. This satisfies Helsinki, but the sample must be larger and the follow-up longer, and the result may not generalize to a no-access setting where adherence and background incidence differ.
Second, a randomized rollout or stepped-wedge design in which the comparison is delayed access rather than placebo. Everyone eventually receives an active product, so the ethical objection to placebo disappears, but the design cannot answer long-term efficacy questions that require a true untreated baseline, and it depends on a supply commitment that outlasts the trial.
Third, an adaptive or hybrid design: placebo only in districts where tenofovir is genuinely unavailable, active comparator where it is funded, with a pre-specified plan to supply tenofovir to the placebo districts once the ministry budget allows.
The strongest objection is not "placebo is always wrong" but "the sponsor is externalizing the cost of the host country's distributive failure onto the participants." The remedy is a binding post-trial access and pricing agreement, negotiated and public before enrolment, plus a data-safety monitoring board with community representation. Without that, the three-year injectable offer is a side payment rather than a fair benefit.