Search

Text-based computational approaches for assessing the quality of psychotherapy are being developed to support quality assurance and clinical training. However, due to the long durations of typical conversation based therapy sessions, and due to limited annotated modeling resources, computational methods largely rely on frequency-based lexical features or dialogue acts to assess the overall session level characteristics. In this work, we propose a hierarchical framework to automatically evaluate the quality of transcribed Cognitive Behavioral Therapy (CBT) interactions. Given the richly dynamic nature of the spoken dialog within a talk therapy session, to evaluate the overall session level quality, we propose to consider modeling it as a function of local variations across the interaction. To implement that empirically, we divide each psychotherapy session into conversation segments and initialize the segment-level qualities with the session-level scores. First, we produce segment embeddings by fine-tuning a BERT-based model, and predict segment-level (local) quality scores. These embeddings are used as the lower-level input to a Bidirectional LSTM-based neural network to predict the session-level (global) quality estimates. In particular, we model the global quality as a linear function of the local quality scores, which allows us to update the segment-level quality estimates based on the session-level quality prediction. These newly estimated segment-level scores benefit the BERT fine-tuning process, which in turn results in better segment embeddings. We evaluate the proposed framework on automatically derived transcriptions from real-world CBT clinical recordings to predict session-level behavior codes. The results indicate that our approach leads to improved evaluation accuracy for most codes when used for both regression and classification tasks., Accepted by Computer Speech & Language

Aerobic exercise is important in the rehabilitation of individuals with prior burn injuries, but no studies have examined whether adult burn survivors demonstrate cardiac remodeling to long-term aerobic exercise training. In this study, we tested the hypothesis that 6 months of progressive exercise training would improve cardiac magnetic resonance imaging-based measures of cardiac structure and function in well-healed burn survivors. Our findings highlight the ability of exercise training to modify cardiac structure and function in well-healed burn survivors and nonburned sedentary controls alike.

Advances in artificial intelligence (AI) are enabling systems that augment and collaborate with humans to perform simple, mechanistic tasks like scheduling meetings and grammar-checking text. However, such Human-AI collaboration poses challenges for more complex, creative tasks, such as carrying out empathic conversations, due to difficulties of AI systems in understanding complex human emotions and the open-ended nature of these tasks. Here, we focus on peer-to-peer mental health support, a setting in which empathy is critical for success, and examine how AI can collaborate with humans to facilitate peer empathy during textual, online supportive conversations. We develop Hailey, an AI-in-the-loop agent that provides just-in-time feedback to help participants who provide support (peer supporters) respond more empathically to those seeking help (support seekers). We evaluate Hailey in a non-clinical randomized controlled trial with real-world peer supporters on TalkLife (N=300), a large online peer-to-peer support platform. We show that our Human-AI collaboration approach leads to a 19.60% increase in conversational empathy between peers overall. Furthermore, we find a larger 38.88% increase in empathy within the subsample of peer supporters who self-identify as experiencing difficulty providing support. We systematically analyze the Human-AI collaboration patterns and find that peer supporters are able to use the AI feedback both directly and indirectly without becoming overly reliant on AI while reporting improved self-efficacy post-feedback. Our findings demonstrate the potential of feedback-driven, AI-in-the-loop writing systems to empower humans in open-ended, social, creative tasks such as empathic conversations.

High dietary phosphate intake may lead to adverse outcomes including cardiovascular disease (CVD). Urinary phosphate excretion, a marker of intestinal phosphate absorption, may be a more reliable marker of phosphate homeostasis in steady state than serum phosphate. Studies report good agreement between urine phosphate-to-creatinine ratio (uPiCr) and 24-hour urinary phosphate; however, whether uPiCr is associated with increased risk of CVD or mortality remains uncertain. This study aimed to assess the relationship between uPiCr and all-cause and CVD mortality.This is an observational longitudinal cohort study using data from the population-based national Australian Diabetes, Obesity and Lifestyle study (n = 10,014 participants). Non-linear association between uPiCr and all-cause and CVD mortality was assessed using fractional polynomial transformations. Cox proportional hazards regression models were used to estimate adjusted hazard ratios for all-cause and CVD mortality.Median age [interquartile range] was 50 [41-62] years, and 46% were male. Median uPiCr was 1.38 [1.02-1.79] mmol/mmol. Median follow-up time was 16.9 years with 1,735 deaths. uPiCr was associated with all-cause and CVD mortality in univariate models and when adjusted for age and gender. However, associations were not significant in multivariate models. Sensitivity analyses excluding participants with chronic kidney disease (CKD) revealed a significant J-shaped association between uPiCr and all-cause mortality. Urine phosphate alone showed an association with increased all-cause mortality in a similar J-shape relationship.Although no association between uPiCr and all-cause and CVD mortality was observed in multivariate analyses in the whole cohort, a significant relationship between uPiCr and mortality in those without CKD suggests that uPiCr may have predictive validity for future adverse outcomes in people with no CKD.

The purpose of this study was to evaluate whether burn survivors have lower adherence compared to non-burned control individuals during a 6-month community-based exercise program. In burn survivors we sought to answer if there was a relation between the size of the burn injury and dropout frequency. Fifty-two burn survivors and 15 non-burned controls (n=67) were recruited for a 6-month community-based (i.e., non-supervised), progressive, exercise training program. During the exercise program, 27% (i.e., 4 of 15 enrolled) of the non-burned individuals dropped out of the study, while 37% (i.e., 19 of 52) of the burn survivors dropped out from the study. There was no difference in the percentage of individuals who dropped out between groups (p = 0.552). There was no difference in size of the burn injury, expressed as percent body surface area burned (%BSA) between the burn-survivors that dropped out versus those who completed the exercise regimen (p = 0.951). We did not observe a relation between %BSA burned and dropouts (log odds = -0.15-0.01(%BSA), B= -0.01, SE = 0.015, p = 0.541). There was no effect of %BSA burned on the probability of dropout [Exp (B) = 0.991, 95% CI (0.961, 1.020)] and there were no differences in the percentage of individuals who dropped out of the study based on %BSA burned (χ2(1) = 0.44, p = 0.51). These data demonstrate that burn survivors have similar exercise adherence relative to a non-burned group and the extent of a burn injury does not affect exercise program adherence.