Is unemployment associated with inefficient sleep habits? A cohort study using objective sleep measurements

Measuring algorithmic bias to analyze the reliability of AI tools that predict depression risk using smartphone sensed-behavioral data

AI tools intend to transform mental healthcare by providing remote estimates of depression risk using behavioral data collected by sensors embedded in smartphones. While these tools accurately predict elevated depression symptoms in small, homogenous populations, recent studies show that these tools are less accurate in larger, more diverse populations. In this work, we show that accuracy is reduced because sensed-behaviors are unreliable predictors of depression across individuals: sensed-behaviors that predict depression risk are inconsistent across demographic and socioeconomic subgroups. We first identified subgroups where a developed AI tool underperformed by measuring algorithmic bias, where subgroups with depression were incorrectly predicted to be at lower risk than healthier subgroups. We then found inconsistencies between sensed-behaviors predictive of depression across these subgroups. Our findings suggest that researchers developing AI tools predicting mental health from sensed-behaviors should think critically about the generalizability of these tools, and consider tailored solutions for targeted populations.

Measuring algorithmic bias to analyze the reliability of AI tools that predict depression risk using smartphone sensed-behavioral data

AI tools intend to transform mental healthcare by providing remote estimates of depression risk using behavioral data collected by sensors embedded in smartphones. While these tools accurately predict elevated symptoms in small, homogenous populations, recent studies show that these tools are less accurate in larger, more diverse populations. In this work, we show that accuracy is reduced because sensed-behaviors are unreliable predictors of depression across individuals; specifically the sensed-behaviors that predict depression risk are inconsistent across demographic and socioeconomic subgroups. We first identified subgroups where a developed AI tool underperformed by measuring algorithmic bias, where subgroups with depression were incorrectly predicted to be at lower risk than healthier subgroups. We then found inconsistencies between sensed-behaviors predictive of depression across these subgroups. Our findings suggest that researchers developing AI tools predicting mental health from behavior should think critically about the generalizability of these tools, and consider tailored solutions for targeted populations.

Trajectories of sleep disturbance in cancer survivors during the first 2 years post-treatment

STUDY OBJECTIVES

To examine the trajectories of sleep disturbance in cancer survivors during the first 2 years post-treatment and to investigate whether psychological, cognitive, and physical factors differentiate trajectories.

METHODS

A total of 623 Chinese cancer survivors of diverse cancer types participated in a 2-year-long prospective study after the completion of cancer treatment. Sleep disturbance was measured using Pittsburgh Sleep Quality Index at 3 (T2), 6 (T3), 12 (T4), 18 (T5), and 24 (T6) months after baseline (within 6-months post-treatment; T1). Latent growth mixture modeling identified distinctive sleep disturbance trajectories and tested if these longitudinal patterns were predicted by baseline psychological distress, attentional control, attentional bias and physical symptom distress and T2 cancer-related distress. Fully adjusted multinomial logistic regression then identified whether these factors differentiated trajectories.

RESULTS

Two distinct sleep disturbance trajectories were identified, namely stable good sleepers (69.7%) and persistent high sleep disturbance (30.3%). Compared to those in the stable good sleep group, patients in the persistent high sleep disturbance group were less likely to report avoidant (OR=0.49, 95% CI = 0.26-0.90), while more likely to report intrusive thoughts (OR = 1.76, 95% CI = 1.06-2.92) and cancer-related hyperarousal (OR = 3.37, 95% CI = 1.78-6.38). Higher depression scores also predicted persistent high sleep disturbance group membership (OR = 1.13, 95% CI = 1.03-1.25). Attentional bias, attentional control, anxiety, and physical symptom distress did not predict sleep trajectory membership.

CONCLUSIONS

One in three cancer survivors experienced persistent high sleep disturbance. Screening and managing depressive symptoms and cancer-related distress in early cancer rehabilitation may reduce risk of persistent sleep disturbance among cancer survivors.