Development of Machine Learning Algorithms Incorporating Electronic Health Record Data, Patient-Reported Outcomes, or Both to Predict Mortality for Outpatients With Cancer.

PURPOSE: Machine learning (ML) algorithms that incorporate routinely collected patient-reported outcomes (PROs) alongside electronic health record (EHR) variables may improve prediction of short-term mortality and facilitate earlier supportive and palliative care for patients with cancer. METHODS: We trained and validated two-phase ML algorithms that incorporated standard PRO assessments alongside approximately 200 routinely collected EHR variables, among patients with medical oncology encounters at a tertiary academic oncology and a community oncology practice. RESULTS: Among 12,350 patients, 5,870 (47.5%) completed PRO assessments. Compared with EHR- and PRO-only algorithms, the EHR + PRO model improved predictive performance in both tertiary oncology (EHR + PRO v EHR v PRO: area under the curve [AUC] 0.86 [0.85-0.87] v 0.82 [0.81-0.83] v 0.74 [0.74-0.74]) and community oncology (area under the curve 0.89 [0.88-0.90] v 0.86 [0.85-0.88] v 0.77 [0.76-0.79]) practices. CONCLUSION: Routinely collected PROs contain added prognostic information not captured by an EHR-based ML mortality risk algorithm. Augmenting an EHR-based algorithm with PROs resulted in a more accurate and clinically relevant model, which can facilitate earlier and targeted supportive care for patients with cancer.

Performance Drift in a Mortality Prediction Algorithm during the SARS-CoV-2 Pandemic.

Research Objective: Health systems use clinical predictive algorithms to allocate resources to high-risk patients. Such algorithms are trained using historical data and are later implemented in clinical settings. During this implementation period, predictive algorithms are prone to performance changes ("drift") due to exogenous shocks in utilization or shifts in patient characteristics. Our objective was to examine the impact of sudden utilization shifts during the SARS-CoV-2 pandemic on the performance of an electronic health record (EHR)-based prognostic algorithm. Study Design: We studied changes in the performance of Conversation Connect, a validated machine learning algorithm that predicts 180-day mortality among outpatients with cancer receiving care at medical oncology practices within a large academic cancer center. Conversation Connect generates mortality risk predictions before each encounter using data from 159 EHR variables collected in the six months before the encounter. Since January 2019, Conversation Connect has been used as part of a behavioral intervention to prompt clinicians to consider early advance care planning conversations among patients with ≥10% mortality risk. First, we descriptively compared encounter-level characteristics in the following periods: January 2019-February 2020 ("pre-pandemic"), March-May 2020 ("early-pandemic"), and June-December 2020 ("later-pandemic"). Second, we quantified changes in high-risk patient encounters using interrupted time series analyses that controlled for pre-pandemic trends and demographic, clinical, and practice covariates. Our primary metric of performance drift was false negative rate (FNR). Third, we assessed contributors to performance drift by comparing distributions of key EHR inputs across periods and predicting later pandemic utilization using pre-pandemic inputs. Population Studied: 237,336 in-person and telemedicine medical oncology encounters. Principal Findings: Age, race, average patient encounters per month, insurance type, comorbidity counts, laboratory values, and overall mortality were similar among encounters in the pre-, early-, and later-pandemic periods. Relative to the pre-pandemic period, the later-pandemic period was characterized by a 6.5-percentage-point decrease (28.2% vs. 34.7%) in high-risk encounters (p<0.001). FNR increased from 41.0% (95% CI 38.0-44.1%) in the pre-pandemic period to 57.5% (95% CI 51.9-63.0%) in the later pandemic period. Compared to the pre-pandemic period, the early and later pandemic periods had higher proportions of telemedicine encounters (0.01% pre-pandemic vs. 20.0% early-pandemic vs. 26.4% later-pandemic) and encounters with no preceding laboratory draws (17.7% pre-pandemic vs. 19.8% early-pandemic vs. 24.1% later-pandemic). In the later pandemic period, observed laboratory utilization was lower than predicted (76.0% vs 81.2%, p<0.001). In the later-pandemic period, mean 180-day mortality risk scores were lower for telemedicine encounters vs. in-person encounters (10.3% vs 11.2%, p<0.001) and encounters with no vs. any preceding laboratory draws (1.5% vs. 14.0%, p<0.001). Conclusions: During the SARS-CoV-2 pandemic period, the performance of a machine learning prognostic algorithm used to prompt advance care planning declined substantially. Increases in telemedicine and declines in laboratory utilization contributed to lower performance. Implications for Policy or Practice: This is the first study to show algorithm performance drift due to SARS-CoV-2 pandemic-related shifts in telemedicine and laboratory utilization. These mechanisms of performance drift could apply to other EHR clinical predictive algorithms. Pandemic-related decreases in care utilization may negatively impact the performance of clinical predictive algorithms and warrant assessment and possible retraining of such algorithms.

Impact of Behavioral Nudges on the Quality of Serious Illness Conversations Among Patients With Cancer: Secondary Analysis of a Randomized Controlled Trial.

PURPOSE: Serious Illness Conversations (SICs) are structured conversations between clinicians and patients about prognosis, treatment goals, and end-of-life preferences. Although behavioral interventions may prompt earlier or more frequent SICs, their impact on the quality of SICs is unclear. METHODS: This was a secondary analysis of a randomized clinical trial (NCT03984773) among 78 clinicians and 14,607 patients with cancer testing the impact of an automated mortality prediction with behavioral nudges to clinicians to prompt more SICs. We analyzed 318 randomly selected SICs matched 1:1 by clinicians (159 control and 159 intervention) to compare the quality of intervention vs. control conversations using a validated codebook. Comprehensiveness of SIC documentation was used as a measure of quality, with higher integer numbers of documented conversation domains corresponding to higher quality conversations. A conversation was classified as high-quality if its score was ≥ 8 of a maximum of 10. Using a noninferiority design, mixed effects regression models with clinician-level random effects were used to assess SIC quality in intervention vs. control groups, concluding noninferiority if the adjusted odds ratio (aOR) was not significantly < 0.9. RESULTS: Baseline characteristics of the control and intervention groups were similar. Intervention SICs were noninferior to control conversations (aOR 0.99; 95% CI, 0.91 to 1.09). The intervention increased the likelihood of addressing patient-clinician relationship (aOR = 1.99; 95% CI, 1.23 to 3.27; P < .01) and decreased the likelihood of addressing family involvement (aOR = 0.56; 95% CI, 0.34 to 0.90; P < .05). CONCLUSION: A behavioral intervention that increased SIC frequency did not decrease their quality. Behavioral prompts may increase SIC frequency without sacrificing quality.

Two-Stage Approaches to Accounting for Patient Heterogeneity in Machine Learning Risk Prediction Models in Oncology.

PURPOSE: Machine learning models developed from electronic health records data have been increasingly used to predict risk of mortality for general oncology patients. But these models may have suboptimal performance because of patient heterogeneity. The objective of this work is to develop a new modeling approach to predicting short-term mortality that accounts for heterogeneity across multiple subgroups in the presence of a large number of electronic health record predictors. METHODS: We proposed a two-stage approach to addressing heterogeneity among oncology patients of different cancer types for predicting their risk of mortality. Structured data were extracted from the University of Pennsylvania Health System for 20,723 patients of 11 cancer types, where 1,340 (6.5%) patients were deceased. We first modeled the overall risk for all patients without differentiating cancer types, as is done in the current practice. We then developed cancer type-specific models using the overall risk score as a predictor along with preselected type-specific predictors. The overall and type-specific models were compared with respect to discrimination using the area under the precision-recall curve (AUPRC) and calibration using the calibration slope. We also proposed metrics that characterize the degree of risk heterogeneity by comparing risk predictors in the overall and type-specific models. RESULTS: The two-stage modeling resulted in improved calibration and discrimination across all 11 cancer types. The improvement in AUPRC was significant for hematologic malignancies including leukemia, lymphoma, and myeloma. For instance, the AUPRC increased from 0.358 to 0.519 (∆ = 0.161; 95% CI, 0.102 to 0.224) and from 0.299 to 0.354 (∆ = 0.055; 95% CI, 0.009 to 0.107) for leukemia and lymphoma, respectively. For all 11 cancer types, the two-stage approach generated well-calibrated risks. A high degree of heterogeneity between type-specific and overall risk predictors was observed for most cancer types. CONCLUSION: Our two-stage modeling approach that accounts for cancer type-specific risk heterogeneity has improved calibration and discrimination than a model agnostic to cancer types.

Patient, disease, and treatment factors associated with overall survival in esthesioneuroblastoma.

BACKGROUND: Esthesioneuroblastomas (ENB) are uncommon and data regarding outcomes are often limited to single-institution series. The National Cancer Database (NCDB), which contains outcomes information from treatment centers across the United States, represents an opportunity to evaluate outcomes for rare diseases such as ENB across multiple institutions. METHODS: The NCDB was queried for location codes corresponding to the nasal cavity and paranasal sinuses and the histology code for ENB. Multivariate analyses were performed to evaluate for contributing factors to overall survival. RESULTS: A total of 1225 patients with ENB met the inclusion criteria. The 5-year overall survival was 76.2% (95% confidence interval [CI], 73.4-79.0%). Overall survival was associated with Kadish stage, grade, treatment sequence, margin status, Charlson/Deyo score, age, and gender (p < 0.05). Multivariate analysis demonstrated that, compared with surgery alone, surgery followed by radiation without chemotherapy had improved all-cause mortality (odds ratio [OR], 0.61; 95% CI, 0.40-0.95). Surgery with chemotherapy alone was associated with increased odds of all-cause mortality (OR, 4.86; 95% CI, 2.31-10.25). Multivariate subanalysis for Kadish stages A and B demonstrated no difference in survival between surgery and surgery followed by radiation, but surgery followed by chemoradiation had worse overall survival (OR, 3.03; 95% CI, 1.07-8.56). For Kadish stage C, surgery followed by radiation had improved overall survival compared with surgery alone (OR, 0.44; 95% CI, 0.24-0.81). CONCLUSION: The most common treatment for ENB is surgery followed by radiation, which is associated with the highest overall survival. The role of adjunctive chemotherapy needs to be re-evaluated in further studies.