Large language models could change the future of behavioral healthcare: a proposal for responsible development and evaluation.

Large language models (LLMs) such as Open AI's GPT-4 (which power ChatGPT) and Google's Gemini, built on artificial intelligence, hold immense potential to support, augment, or even eventually automate psychotherapy. Enthusiasm about such applications is mounting in the field as well as industry. These developments promise to address insufficient mental healthcare system capacity and scale individual access to personalized treatments. However, clinical psychology is an uncommonly high stakes application domain for AI systems, as responsible and evidence-based therapy requires nuanced expertise. This paper provides a roadmap for the ambitious yet responsible application of clinical LLMs in psychotherapy. First, a technical overview of clinical LLMs is presented. Second, the stages of integration of LLMs into psychotherapy are discussed while highlighting parallels to the development of autonomous vehicle technology. Third, potential applications of LLMs in clinical care, training, and research are discussed, highlighting areas of risk given the complex nature of psychotherapy. Fourth, recommendations for the responsible development and evaluation of clinical LLMs are provided, which include centering clinical science, involving robust interdisciplinary collaboration, and attending to issues like assessment, risk detection, transparency, and bias. Lastly, a vision is outlined for how LLMs might enable a new generation of studies of evidence-based interventions at scale, and how these studies may challenge assumptions about psychotherapy.

Marital status and perceived stress in men with advanced prostate cancer: A randomized-controlled trial of cognitive behavioral stress management.

OBJECTIVE: Relationship status predicts numerous outcomes among medical populations. Few interventions evaluate the role of marital status on response to psychosocial treatment, and no such studies exist within advanced prostate cancer (APC). This study examined whether marital status modified the effect of a cognitive behavioral stress management (CBSM) intervention on perceived stress. METHODS: Men with APC (N = 190) were randomized to 10-week CBSM or a health promotion (HP) intervention (#NCT03149185). The Perceived Stress Scale assessed perceived stress at baseline and 12-month follow-up. Medical status and sociodemographics were captured at enrollment. RESULTS: Participants were mostly White (59.5%), non-Hispanic (97.4%), heterosexual (97.4%) men, 66.8% of whom were partnered. Neither condition nor marital status predicted perceived stress change at follow-up. However, a significant interaction was found between condition and marital status (p = 0.014; Cohen's f = 0.07), such that partnered men who received CBSM and unpartnered men who received HP reported greater reductions in perceived stress. CONCLUSION: This is the first study to assess the impact of marital status on psychosocial intervention effects among men with APC. Partnered men derived greater benefit from a cognitive-behavioral intervention and unpartnered men equally benefitted from a HP intervention. Further research is necessary to understand the mechanisms underlying these relationships.

Technology-Based Psychosocial Intervention to Improve Quality of Life and Reduce Symptom Burden in Men with Advanced Prostate Cancer: Results from a Randomized Controlled Trial.

BACKGROUND: Men with advanced prostate cancer (APC) face multiple challenges including poor prognosis, poor health-related quality of life (HRQOL), and elevated symptom burden. This study sought to establish the efficacy of a tablet-delivered, group-based psychosocial intervention for improving HRQOL and reducing symptom burden in men with APC. We hypothesized that men randomized to cognitive-behavioral stress management (CBSM) would report improved HRQOL and reduced symptom burden relative to men randomized to an active control health promotion (HP) condition. Condition effects on intervention targets and moderators of these effects were explored. METHODS: Men with APC (N = 192) were randomized (1:1) to 10-week tablet-delivered CBSM or HP, and followed for 1 year. Multilevel modeling was used to evaluate condition effects over time. RESULTS: Changes in HRQOL and symptom burden did not differ between groups. Men in both groups improved across several intervention targets; men in the CBSM condition reported greater increases in self-reported ability to relax, and both conditions showed improvements in cancer-related anxiety, cancer-related distress, and feelings of cohesiveness with other patients over time. Moderating factors included baseline interpersonal disruption, fatigue, and sexual functioning. CONCLUSIONS: Tablet-delivered CBSM and HP were well received by men with APC. The hypothesized effects of CBSM on HRQOL and symptom burden were not supported, though improvements in intervention targets were observed across conditions. Participants reported high-baseline HRQOL relative to cancer and general population norms, possibly limiting intervention effects. The identified moderating factors should be considered in the development and implementation of interventions targeting HRQOL and symptom burden. TRIAL REGISTRATION: ClinicalTrials.gov Identifier: NCT03149185.