Variational Disentanglement for Rare Event Modeling.

Combining the increasing availability and abundance of healthcare data and the current advances in machine learning methods have created renewed opportunities to improve clinical decision support systems. However, in healthcare risk prediction applications, the proportion of cases with the condition (label) of interest is often very low relative to the available sample size. Though very prevalent in healthcare, such imbalanced classification settings are also common and challenging in many other scenarios. So motivated, we propose a variational disentanglement approach to semi-parametrically learn from rare events in heavily imbalanced classification problems. Specifically, we leverage the imposed extreme-distribution behavior on a latent space to extract information from low-prevalence events, and develop a robust prediction arm that joins the merits of the generalized additive model and isotonic neural nets. Results on synthetic studies and diverse real-world datasets, including mortality prediction on a COVID-19 cohort, demonstrate that the proposed approach outperforms existing alternatives.

Accelerometry Data Delineate Phases of Recovery and Supplement Patient-Reported Outcome Measures Following Lumbar Laminectomy.

BACKGROUND: Patient-reported outcome measures (PROMs) are traditionally used to track recovery of patients after spine surgery. Wearable accelerometers have adjunctive value because of the continuous, granular, and objective data they provide. We conducted a prospective study of lumbar laminectomy patients to determine if time-series data from wearable accelerometers could delineate phases of recovery and compare accelerometry data to PROMs during recovery tracking. METHODS: Patients with lumbar stenosis for whom lumbar laminectomy was indicated were prospectively recruited. Subjects wore accelerometers that recorded their daily step counts from at least 1 week preoperatively to 6 months postoperatively. Subjects completed the Oswestry Disability Index and the 12-Item Short Form Health Survey preoperatively and at 2 weeks, 1 month, 3 months, and 6 months postoperatively. Daily aggregate median steps and individual visit-specific median steps were calculated. The Pruned Linear Exact Time method was used to segment aggregate median steps into distinct phases. Associations between visit-specific median steps and PROMs were identified using Spearman rank correlation. RESULTS: Segmentation analysis revealed 3 distinct postoperative phases: step counts rapidly increased for the first 40 days postoperatively (acute healing), then gained more slowly for the next 90 days (recovery), and finally plateaued at preoperative levels (stabilization). Visit-specific median steps were significantly correlated with PROMs throughout the postoperative period. PROMs significantly exceeded baseline at 6 months postoperatively, while step counts did not (all P < 0.05). CONCLUSIONS: Continuous data from accelerometers allowed for identification of 3 distinct stages of postoperative recovery after lumbar laminectomy. PROMs remain necessary to capture subjective elements of recovery.