When sharing is the point

Meena is 64 and lives two hours from Pondicherry. She has glaucoma – a condition that causes no pain, gives no warning, and, if left untreated, leads to irreversible blindness. The specialist who manages her care sees roughly 180 patients a day. The average consultation lasts six minutes. One afternoon, the hospital offers her a different arrangement: she and four other glaucoma patients will be seen together. Each will receive an individual examination and personalized recommendations – but in the presence of the others.

Her instinct, we suspect, was discomfort. A sense that something private was being traded away. That instinct is widely shared, and it is the reason shared medical appointments have struggled to gain adoption in most health systems despite evidence of their benefits.

That instinct is understandable. But in this setting, it turns out to be incomplete. Running the first randomized controlled trial of shared medical appointments for any disease in India forced us to rethink assumptions we had not known we were making.

Testing a radical idea

Between 2016 and 2018, in a study published in PLOS Global Public Health, together with colleagues at Aravind Eye Hospital, Harvard Business School, and London Business School, we randomly assigned 1,000 glaucoma patients to one of two conditions across four consecutive appointments: standard one-on-one consultations, or shared medical appointments (SMAs) in groups of five. Because assignment was random, any differences in outcomes can be attributed to the format itself – not to self-selection.

One detail about our setting matters. Aravind is not a struggling clinic. It is one of the most operationally sophisticated health systems in the world, already running baseline consultations of just over three minutes per patient. We were not testing shared appointments against a slow, wasteful baseline. We were testing them against a model that was already highly optimized.

That rules out a simple efficiency story. What we found was a behavioral one. It also requires an honest caveat: physician time per patient was slightly higher in the SMA condition – 3.44 minutes against 3.04 minutes for one-on-one appointments – because patients asked more questions. Shared appointments, in this trial, were not faster at the individual level. The engagement gains did not come from saving time. They came from changing how that time was used.

What surprised us

Patients in shared appointments reported higher satisfaction on multiple dimensions – including how well their doubts were addressed and how much they learned. On objective knowledge tests, they scored significantly higher. Their medication compliance rate was 97.0 percent, against 94.9 percent for one-on-one patients.

That gap deserves unpacking. Two percentage points sound modest. But it represents a 39 percent reduction in non-compliance – a 39 percent reduction in the proportion of patients failing to follow their prescribed regimen. For a disease where daily adherence is the primary mechanism of disease control, that behavioral shift matters. Clinical outcomes, meanwhile, were statistically indistinguishable between groups. SMAs improved satisfaction, learning, and medication compliance, without compromising follow-up rates or measured clinical outcomes. What surprised us most was not the compliance finding. It was what we observed in how patients engaged with the appointment itself. Patients did not simply receive care; they interacted with it differently, drawing on shared questions and experiences.

Exit surveys reinforced this. Patients in SMAs reported that the physician had been more caring – not less – and that they had received more individual attention, despite the shared format. The group had not eroded the therapeutic relationship. In some respects, it had strengthened it.

Patients as co-producers

One finding captures the arc of the experience precisely. Early in the trial, interruptions by peers – a fellow patient asking a question during someone else’s examination – were negatively correlated with satisfaction. Patients were still adjusting. By later appointments, the relationship had completely reversed: interruptions were positively correlated with satisfaction. What had felt intrusive had become community.

For service designers, that trajectory has a direct implication. Onboarding shapes whether participants develop the relationship with a shared format that makes its benefits possible. The early friction is not a failure of the model – it is part of the transition into it. The gender finding deserves a note of empirical caution. Female patients experienced a significantly larger compliance improvement than male patients; younger patients showed stronger gains than older ones. The data show differential effects – they do not explain them definitively. What we can say is that the populations most expected to find the privacy trade-off costly were not, in fact, worse off. Some benefited more. That is worth noting, without overclaiming why.

Three questions for service designers

The implications extend beyond healthcare. Many high-value services – consulting, financial advice, executive coaching, professional education – default to one-on-one delivery on the assumption that individualization is quality. The evidence here should prompt careful reconsideration.

When does visibility improve engagement rather than compromise it? Seeing others navigate the same situation – asking the same questions, expressing the same uncertainties – can be enabling rather than embarrassing. The peer becomes a resource. That dynamic applies wherever customers share similar needs and where the barrier to engagement is partly social rather than informational.

What does the one-on-one format conceal? Conventional appointments obscure the physician’s time constraint; shared appointments make it visible and collective. In our trial, that transparency increased trust rather than undermining it. Managed visibility of constraints and trade-offs may do the same in other service contexts.

What is being designed away in the name of personalization? The peer dynamic in SMAs generated normalization of uncertainty, social motivation to comply, and collective learning. Hyper-individualized services eliminate those dynamics by design. Sometimes that is right. But it should be a deliberate choice, not an unexamined default.

The paradox, reconsidered

The resistance to shared medical appointments – in wealthy health systems and in Meena’s initial hesitation – rests on an equation we rarely examine: privacy equals dignity equals quality. Our research does not dismiss that equation. It shows that it is contingent.

Psychological safety, it turns out, can be created through the presence of others facing the same situation – not just through the privacy of a closed door. In a context where the alternative to shared care is no qualified care at all, that realization is practically urgent. But even where the alternative is a private appointment, the assumption that individualization is always better deserves scrutiny.

Meena left her shared appointment more satisfied and better informed than patients who saw the doctor alone. Her compliance with her medication was higher. Whether the shared format directly protected her vision the research cannot say – the trial was not designed to show that. But it produced the behaviors most likely to do so.

The closed door, it turns out, may not be the only architecture for trust.