Essential Work: Psychiatric Clinical Investigation in the Midst of a Pandemic

For the past decade, the tools have been available to enable clinical trials to be conducted remotely—not in part, but in full. Electronic health records facilitate recruitment, teleconferencing enables consent and monitoring, and mobile applications allow self-report scales and collection of passive measures (2). If clinical trials of new interventions, not just large-scale effectiveness bake-offs, are to continue, this will change. Truly remote trials can dramatically increase accessibility to particularly vulnerable populations—including those who cannot safely travel to a study site during business hours. If schools in resource-limited areas can provide low-cost laptops to children, investigators can certainly do no less for clinical trial participants.

As COVID-19 exploded around the world, new treatment trials were launched hastily, in many cases with insufficient sample sizes, duplicative efforts, and incompatible designs. This problem sounds all too familiar to psychiatrists accustomed to underpowered studies that yield ambiguous results. Recruitment challenges may finally force investigators to pay more attention to the nontraditional designs common in oncology and other areas of medicine (3). Availability of shared trial platforms and data sources developed to expedite clinical investigation for COVID-19 (4) should also inform future psychiatric studies. Virtual trial networks need not be sinecures granted to well-connected sites but, rather, efficient means of collaboration. At minimum, the embrace of preprint servers and accelerated online data sharing is a trend psychiatry should follow enthusiastically.

COVID-19 and its social consequences represent profound secular trends (5). Individuals enrolled in a trial 3 months before pandemic onset, compared with 6 months later in a changed environment, are likely to look profoundly different clinically. Analytic strategies for study outcomes will need to take this discontinuity into account. In reality, any psychiatrist will recognize the phenomenon of the patient who reports feeling far worse from week to week because of some new adversity, effects rarely modeled systematically in trials. This is a chance to imbue the phrase “computational psychiatry” with meaning at last, leveraging more sophisticated approaches to analyzing time-series data and incorporating individual-level variation.

The new experimental condition imposed by COVID-19 may provide badly needed data about how to prevent or treat adverse outcomes associated with trauma. Trialists should consider, at minimum, adding analyses and measures to explicitly examine the emergence of psychiatric sequelae, even in studies not focused on stress per se. Extending follow-up or adding a 6- or 12-month outcome to acute trials may increase the ability to detect the effect of short-term interventions on longer-term outcomes even if the primary treatment target was not depression or anxiety, generating important hypotheses for future studies. COVID-19 also provides a reminder about the importance of how health disparities affect disease risk and outcomes, effects rarely considered explicitly in clinical trials.

Based on experience with the 2003 outbreak of SARS (severe acute respiratory syndrome) and with the terrorist attacks of September 11, 2001, among others, the psychiatric community quickly recognized the potentially devastating psychiatric impact of the COVID-19 pandemic—both in the near term and in the years to come (6). As a result, a thousand surveys have bloomed, with opportunities to expand psychiatric screening in community settings. These assessments provide a unique opportunity to engage a new group of individuals in clinical trials, provided that investigators can do more to make intervention studies appealing. Given the widespread availability of effective treatments for psychiatric disorders, placebos are very difficult to justify even if they offer advantages in signal detection. Trialists can and should seek to employ active comparators and noninferiority designs: clinical trials should, at minimum, provide standard-of-care treatments.

Clinical investigators are eager to find a new normal. But the constraints imposed by COVID-19 allow an opportunity for a much-needed reconsideration of how new interventions in psychiatry are studied. In a pandemic that began with squandered opportunities to save lives, the psychiatric research community must not miss this next opportunity to save lives.