Patient classification systems can have negative consequences, especially under scarcity

It is well known that both scarcity of resources and expectations of efficacy influence preferences and decisions in healthcare. But how do these two factors (scarcity and expectations) affect when both are combined? Why, for example, were specific groups of patients (mainly elderly and disabled) so obviously discriminated against during the COVID-19 pandemic, when medical resources were particularly scarce? Aranzazu Vinas, Fernando Blanco, and Helena Matute wanted to answer this question and thus gain a better understanding of the mechanisms involved in health decisions under scarcity. The results of their research have been published in the journal Scientific Reports. 

In the two experiments described in the article, volunteers had to imagine that they were doctors and had to treat a series of fictitious patients. Participants were randomly assigned to two groups, one with scarce resources (although sufficient to treat all their patients) and one with abundant resources. Participants were then presented a series of patients sequentially, and for each of them, they decided whether to administer the treatment or not. These patients were labelled either as very sensitive or weakly sensitive to the treatment, which led to high or low expectations of efficacy of the treatment for each patient type. Immediately after each decision, the participants were informed whether the patient had healed. 

Importantly, the labelling was incorrect, and both types of patients healed with the same probability. In the first experiment, the treatment was effective, meaning that administering it always increased the patients' probability of cure. In the second experiment, not only was the labelling of the patients incorrect, but the treatment was also totally ineffective in all cases: it did not increase the probability of cure.

At the end of both experiments, participants answered whether they believed the treatment was effective and whether it was more effective in a group of patients. The aim was to test whether, despite the prior expectations induced by the erroneous labelling, participants were able to learn from the evidence they received throughout the experiments. That is, they had the opportunity to prove the classification wrong and to determine the true effectiveness of the treatment, just by observing the patients' cures.

The results showed that volunteers administered less treatment to patients classified (incorrectly) as not very sensitive, especially when resources were scarce. Moreover, participants judged the treatment as less effective with these patients, even in experiment 2, when it was ineffective for all! That is, their prior expectation, induced by mislabelling, caused participants to allocate resources unfairly and further interfered with their ability to learn that the treatment was equally effective (in experiment 1) or equally ineffective (in experiment 2).

In short, the scarcity of resources together with the initial expectations induced by patient classification systems that are sometimes erroneous can result in an unfair distribution of resources, in the incorrect use of treatments that are not really effective, and worse, in people being less able to learn from the evidence and thus correct their mistakes.

Reference:

Vinas, A., Blanco, F., & Matute, H. (2025). The combined effect of patient classification and the availability of resources can bias the judgments of treatment effectiveness. Scientific Reports, 15, 15915. https://doi.org/10.1038/s41598-025-01043-w