Using Machine Learning to Predict Operating Room Case Duration: A Case Study in Otolaryngology.

OBJECTIVE: Optimizing operating room (OR) efficiency depends on accurate case duration estimates. Machine learning (ML) methods have been used to predict OR case durations in other subspecialties. We hypothesize that ML methods improve projected case lengths over existing non-ML techniques for otolaryngology-head and neck surgery cases. METHODS: Deidentified patient information from otolaryngology surgical cases at 1 academic institution were reviewed from 2016 to 2020. Variables collected included patient, surgeon, procedure, and facility data known preoperatively so as to capture all realistic contributors. Available case data were divided into a training and testing data set. Several ML algorithms were evaluated based on best performance of predicted case duration when compared to actual case duration. Performance of all models was compared by the average root mean squared error and mean absolute error (MAE). RESULTS: In total, 50,888 otolaryngology surgical cases were evaluated with an average case duration of 98.3 ± 86.9 minutes. Most cases were general otolaryngology (n = 16,620). Case features closely associated with OR duration included procedure performed, surgeon, subspecialty of case, and postoperative destination of the patient. The best-performing ML models were CatBoost and XGBoost, which reduced operative time MAE by 9.6 minutes and 8.5 minutes compared to current methods, respectively. DISCUSSION: The incorporation of other easily identifiable features beyond procedure performed and surgeon meaningfully improved our operative duration prediction accuracy. CatBoost provided the best-performing ML model. IMPLICATIONS FOR PRACTICE: ML algorithms to predict OR case time duration in otolaryngology can improve case duration accuracy and result in financial benefit.

Data-driven segmentation of audiometric phenotypes across a large clinical cohort.

Pure tone audiograms are used to assess the degree and underlying source of hearing loss. Audiograms are typically categorized into a few canonical types, each thought to reflect distinct pathologies of the ear. Here, we analyzed 116,400 patient records from our clinic collected over a 24-year period and found that standard categorization left 46% of patient records unclassified. To better account for the full spectrum of hearing loss profiles, we used a Gaussian Mixture Model (GMM) to segment audiograms without any assumptions about frequency relationships, interaural symmetry or etiology. The GMM converged on ten types, featuring varying degrees of high-frequency hearing loss, flat loss, mixed loss, and notched profiles, with predictable relationships to patient age and sex. A separate GMM clustering of 15,380 audiograms from the National Health and Nutrition Examination Survey (NHANES) identified six similar types, that only lacked the more extreme hearing loss configurations observed in our patient cohort. Whereas traditional approaches distill hearing loss configurations down to a few canonical types by disregarding much of the underlying variability, an objective probabilistic model that accounted for all of the data identified an organized, but more heterogenous set of audiogram types that was consistent across two large clinical databases.

Audiometric Predictors of Bothersome Tinnitus in a Large Clinical Cohort of Adults With Sensorineural Hearing Loss.

OBJECTIVE: To identify demographic and audiometric predictors of bothersome tinnitus within a large clinical cohort. STUDY DESIGN: Retrospective chart review. SETTING: Tertiary care hospital. PATIENTS: 51,989 English-speaking patients between 18 and 80 years of age that received initial audiometric evaluations at the Massachusetts Eye and Ear Infirmary between the years 2000 and 2016. MAIN OUTCOME MEASURES: Patients were categorized according to whether or not tinnitus was the primary reason for their visit. The likelihood of tinnitus as a primary complaint (TPC) was evaluated as a function of age, sex, and audiometric configuration. Patient-reported tinnitus percepts were qualitatively assessed in relation to audiometric configuration. RESULTS: Approximately 20% of adults who presented for an initial hearing evaluation reported TPC. The prevalence of TPC increased with advancing age until approximately 50 to 54 years, and then declined thereafter. In general, men were significantly more likely to report TPC than women. TPC was statistically associated with specific audiogram configurations. In particular, TPC was most prevalent for notched and steeply sloping hearing losses, but was relatively uncommon in adults with flat losses. Patients with frequency-restricted threshold shifts often reported tonal tinnitus percepts, while patients with asymmetric configurations tended to report broadband percepts. CONCLUSIONS: The probability of seeking audiological evaluation for bothersome tinnitus is highest for males, middle-aged patients, and those with notched or high-frequency hearing losses. These findings support the theory that tinnitus arises from sharp discontinuities in peripheral afferent innervation and cochlear amplification, which may induce topographically restricted changes in the central auditory pathway.

Leveraging Existing Tools in Electronic Health Record Systems to Automate Clinical Registry Compilation.

Clinical registries have proven beneficial by providing a resource to address research questions, monitor care, and identify suitable subjects for clinical studies. Despite a well-organized registry, population is often low because of the human capital required. The increasing prevalence of electronic medical health records provides the opportunity to integrate registry compilation into routine patient encounters. Here we describe how one tool existing within the Epic Medical Record software suite, Smart Phrases, can be adapted to automate population of a hearing loss patient registry. The usage rate of Smart Phrases was high and resulted in a significant reduction in the time burden associated with registry population. Use of Smart Phrases could become an important factor in the design of future registries that allow broad uptake and convenient data input.