Towards global model generalizability: independent cross-site feature evaluation for patient-level risk prediction models using the OHDSI network.

BACKGROUND: Predictive models show promise in healthcare, but their successful deployment is challenging due to limited generalizability. Current external validation often focuses on model performance with restricted feature use from the original training data, lacking insights into their suitability at external sites. Our study introduces an innovative methodology for evaluating features during both the development phase and the validation, focusing on creating and validating predictive models for post-surgery patient outcomes with improved generalizability. METHODS: Electronic health records (EHRs) from 4 countries (United States, United Kingdom, Finland, and Korea) were mapped to the OMOP Common Data Model (CDM), 2008-2019. Machine learning (ML) models were developed to predict post-surgery prolonged opioid use (POU) risks using data collected 6 months before surgery. Both local and cross-site feature selection methods were applied in the development and external validation datasets. Models were developed using Observational Health Data Sciences and Informatics (OHDSI) tools and validated on separate patient cohorts. RESULTS: Model development included 41 929 patients, 14.6% with POU. The external validation included 31 932 (UK), 23 100 (US), 7295 (Korea), and 3934 (Finland) patients with POU of 44.2%, 22.0%, 15.8%, and 21.8%, respectively. The top-performing model, Lasso logistic regression, achieved an area under the receiver operating characteristic curve (AUROC) of 0.75 during local validation and 0.69 (SD = 0.02) (averaged) in external validation. Models trained with cross-site feature selection significantly outperformed those using only features from the development site through external validation (P < .05). CONCLUSIONS: Using EHRs across four countries mapped to the OMOP CDM, we developed generalizable predictive models for POU. Our approach demonstrates the significant impact of cross-site feature selection in improving model performance, underscoring the importance of incorporating diverse feature sets from various clinical settings to enhance the generalizability and utility of predictive healthcare models.

Predicting Depression Risk in Patients With Cancer Using Multimodal Data: Algorithm Development Study.

BACKGROUND: Patients with cancer starting systemic treatment programs, such as chemotherapy, often develop depression. A prediction model may assist physicians and health care workers in the early identification of these vulnerable patients. OBJECTIVE: This study aimed to develop a prediction model for depression risk within the first month of cancer treatment. METHODS: We included 16,159 patients diagnosed with cancer starting chemo- or radiotherapy treatment between 2008 and 2021. Machine learning models (eg, least absolute shrinkage and selection operator [LASSO] logistic regression) and natural language processing models (Bidirectional Encoder Representations from Transformers [BERT]) were used to develop multimodal prediction models using both electronic health record data and unstructured text (patient emails and clinician notes). Model performance was assessed in an independent test set (n=5387, 33%) using area under the receiver operating characteristic curve (AUROC), calibration curves, and decision curve analysis to assess initial clinical impact use. RESULTS: Among 16,159 patients, 437 (2.7%) received a depression diagnosis within the first month of treatment. The LASSO logistic regression models based on the structured data (AUROC 0.74, 95% CI 0.71-0.78) and structured data with email classification scores (AUROC 0.74, 95% CI 0.71-0.78) had the best discriminative performance. The BERT models based on clinician notes and structured data with email classification scores had AUROCs around 0.71. The logistic regression model based on email classification scores alone performed poorly (AUROC 0.54, 95% CI 0.52-0.56), and the model based solely on clinician notes had the worst performance (AUROC 0.50, 95% CI 0.49-0.52). Calibration was good for the logistic regression models, whereas the BERT models produced overly extreme risk estimates even after recalibration. There was a small range of decision thresholds for which the best-performing model showed promising clinical effectiveness use. The risks were underestimated for female and Black patients. CONCLUSIONS: The results demonstrated the potential and limitations of machine learning and multimodal models for predicting depression risk in patients with cancer. Future research is needed to further validate these models, refine the outcome label and predictors related to mental health, and address biases across subgroups.

Multimodal data fusion for cancer biomarker discovery with deep learning.

Technological advances now make it possible to study a patient from multiple angles with high-dimensional, high-throughput multi-scale biomedical data. In oncology, massive amounts of data are being generated ranging from molecular, histopathology, radiology to clinical records. The introduction of deep learning has significantly advanced the analysis of biomedical data. However, most approaches focus on single data modalities leading to slow progress in methods to integrate complementary data types. Development of effective multimodal fusion approaches is becoming increasingly important as a single modality might not be consistent and sufficient to capture the heterogeneity of complex diseases to tailor medical care and improve personalised medicine. Many initiatives now focus on integrating these disparate modalities to unravel the biological processes involved in multifactorial diseases such as cancer. However, many obstacles remain, including lack of usable data as well as methods for clinical validation and interpretation. Here, we cover these current challenges and reflect on opportunities through deep learning to tackle data sparsity and scarcity, multimodal interpretability, and standardisation of datasets.

Preoperative Versus Perioperative Risk Factors for Delayed Pain and Opioid Cessation After Total Joint Arthroplasty: A Prospective Cohort Study.

INTRODUCTION: The evolution of pre- versus postoperative risk factors remains unknown in the development of persistent postoperative pain and opioid use. We identified preoperative versus comprehensive perioperative models of delayed pain and opioid cessation after total joint arthroplasty including time-varying postoperative changes in emotional distress. We hypothesized that time-varying longitudinal measures of postoperative psychological distress, as well as pre- and postoperative use of opioids would be the most significant risk factors for both outcomes. METHODS: A prospective cohort of 188 patients undergoing total hip or knee arthroplasty at Stanford Hospital completed baseline pain, opioid use, and emotional distress assessments. After surgery, a modified Brief Pain Inventory was assessed daily for 3 months, weekly thereafter up to 6 months, and monthly thereafter up to 1 year. Emotional distress and pain catastrophizing were assessed weekly to 6 months, then monthly thereafter. Stepwise multivariate time-varying Cox regression modeled preoperative variables alone, followed by all perioperative variables (before and after surgery) with time to postoperative opioid and pain cessation. RESULTS: The median time to opioid and pain cessation was 54 and 152 days, respectively. Preoperative total daily oral morphine equivalent use (hazard ratio-HR 0.97; 95% confidence interval-CI 0.96-0.98) was significantly associated with delayed postoperative opioid cessation in the perioperative model. In contrast, time-varying postoperative factors: elevated PROMIS (Patient-Reported Outcomes Measurement Information System) depression scores (HR 0.92; 95% CI 0.87-0.98), and higher Pain Catastrophizing Scale scores (HR 0.85; 95% CI 0.75-0.97) were independently associated with delayed postoperative pain resolution in the perioperative model. CONCLUSIONS: These findings highlight preoperative opioid use as a key determinant of delayed postoperative opioid cessation, while postoperative elevations in depressive symptoms and pain catastrophizing are associated with persistent pain after total joint arthroplasty providing the rationale for continued risk stratification before and after surgery to identify patients at highest risk for these distinct outcomes. Interventions targeting these perioperative risk factors may prevent prolonged postoperative pain and opioid use.

Prediction of opioid-related outcomes in a medicaid surgical population: Evidence to guide postoperative opiate therapy and monitoring.

BACKGROUND: Treatment of surgical pain is a common reason for opioid prescriptions. Being able to predict which patients are at risk for opioid abuse, dependence, and overdose (opioid-related adverse outcomes [OR-AE]) could help physicians make safer prescription decisions. We aimed to develop a machine-learning algorithm to predict the risk of OR-AE following surgery using Medicaid data with external validation across states. METHODS: Five machine learning models were developed and validated across seven US states (90-10 data split). The model output was the risk of OR-AE 6-months following surgery. The models were evaluated using standard metrics and area under the receiver operating characteristic curve (AUC) was used for model comparison. We assessed calibration for the top performing model and generated bootstrap estimations for standard deviations. Decision curves were generated for the top-performing model and logistic regression. RESULTS: We evaluated 96,974 surgical patients aged 15 and 64. During the 6-month period following surgery, 10,464 (10.8%) patients had an OR-AE. Outcome rates were significantly higher for patients with depression (17.5%), diabetes (13.1%) or obesity (11.1%). The random forest model achieved the best predictive performance (AUC: 0.877; F1-score: 0.57; recall: 0.69; precision:0.48). An opioid disorder diagnosis prior to surgery was the most important feature for the model, which was well calibrated and had good discrimination. CONCLUSIONS: A machine learning models to predict risk of OR-AE following surgery performed well in external validation. This work could be used to assist pain management following surgery for Medicaid beneficiaries and supports a precision medicine approach to opioid prescribing.

Postoperative opioid prescribing patients with diabetes: Opportunities for personalized pain management.

BACKGROUND: Opioids are commonly prescribed for postoperative pain, but may lead to prolonged use and addiction. Diabetes impairs nerve function, complicates pain management, and makes opioid prescribing particularly challenging. METHODS: This retrospective observational study included a cohort of postoperative patients from a multisite academic health system to assess the relationship between diabetes, pain, and prolonged opioid use (POU), 2008-2019. POU was defined as a new opioid prescription 3-6 months after discharge. The odds that a patient had POU was assessed using multivariate logistic regression controlling for patient factors (e.g., demographic and clinical factors, as well as prior pain and opiate use). FINDINGS: A total of 43,654 patients were included, 12.4% with diabetes. Patients with diabetes had higher preoperative pain scores (2.1 vs 1.9, p<0.001) and lower opioid naïve rates (58.7% vs 68.6%, p<0.001). Following surgery, patients with diabetes had higher rates of POU (17.7% vs 12.7%, p<0.001) despite receiving similar opioid prescriptions at discharge. Patients with Type I diabetes were more likely to have POU compared to other patients (Odds Ratio [OR]: 2.22; 95% Confidence Interval [CI]:1.69-2.90 and OR:1.44, CI: 1.33-1.56, respectively). INTERPRETATION: In conclusion, surgical patients with diabetes are at increased risk for POU even after controlling for likely covariates, yet they receive similar postoperative opiate therapy. The results suggest a more tailored approach to diabetic postoperative pain management is warranted.

A Bayesian approach to predictive uncertainty in chemotherapy patients at risk of acute care utilization.

BACKGROUND: Machine learning (ML) predictions are becoming increasingly integrated into medical practice. One commonly used method, ℓ1-penalised logistic regression (LASSO), can estimate patient risk for disease outcomes but is limited by only providing point estimates. Instead, Bayesian logistic LASSO regression (BLLR) models provide distributions for risk predictions, giving clinicians a better understanding of predictive uncertainty, but they are not commonly implemented. METHODS: This study evaluates the predictive performance of different BLLRs compared to standard logistic LASSO regression, using real-world, high-dimensional, structured electronic health record (EHR) data from cancer patients initiating chemotherapy at a comprehensive cancer centre. Multiple BLLR models were compared against a LASSO model using an 80-20 random split using 10-fold cross-validation to predict the risk of acute care utilization (ACU) after starting chemotherapy. FINDINGS: This study included 8439 patients. The LASSO model predicted ACU with an area under the receiver operating characteristic curve (AUROC) of 0.806 (95% CI: 0.775-0.834). BLLR with a Horseshoe+ prior and a posterior approximated by Metropolis-Hastings sampling showed similar performance: 0.807 (95% CI: 0.780-0.834) and offers the advantage of uncertainty estimation for each prediction. In addition, BLLR could identify predictions too uncertain to be automatically classified. BLLR uncertainties were stratified by different patient subgroups, demonstrating that predictive uncertainties significantly differ across race, cancer type, and stage. INTERPRETATION: BLLRs are a promising yet underutilised tool that increases explainability by providing risk estimates while offering a similar level of performance to standard LASSO-based models. Additionally, these models can identify patient subgroups with higher uncertainty, which can augment clinical decision-making. FUNDING: This work was supported in part by the National Library Of Medicine of the National Institutes of Health under Award Number R01LM013362. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Patient-reported distress at a cancer center during the COVID-19 pandemic.

Assessments of health-related quality of life (HRQOL) are conducted by health systems to improve patient-centered care. Studies have shown that the COVID-19 pandemic poses unique stressors for patients with cancer. This study investigates change in self-reported global health scores in patients with cancer before and during the COVID-19 pandemic. In this single-institution retrospective cohort study, patients who completed the Patient-Reported Outcomes Measurement Information System (PROMIS) at a comprehensive cancer center before and during the COVID-19 pandemic were identified. Surveys were analyzed to assess change in the global mental health (GMH) and global physical health (GPH) scores at different time periods (pre-COVID: 3/1/5/2019-3/15/2020, surge1: 6/17/2020-9/7/2020, valley1: 9/8/2020-11/16/2020, surge2: 11/17/2020-3/2/2021, and valley2: 3/3/2021-6/15/2021). A total of 25,192 surveys among 7209 patients were included in the study. Mean GMH score for patients before the COVID-19 pandemic (50.57) was similar to those during various periods during the pandemic: surge1 (48.82), valley1 (48.93), surge2 (48.68), valley2 (49.19). Mean GPH score was significantly higher pre-COVID (42.46) than during surge1 (36.88), valley1 (36.90), surge2 (37.33) and valley2 (37.14). During the pandemic, mean GMH (49.00) and GPH (37.37) scores obtained through in-person were similar to mean GMH (48.53) and GPH (36.94) scores obtained through telehealth. At this comprehensive cancer center, patients with cancer reported stable mental health and deteriorating physical health during the COVID-19 pandemic as indicated by the PROMIS survey. Modality of the survey (in-person versus telehealth) did not affect scores.

Natural Language Processing Methods to Identify Oncology Patients at High Risk for Acute Care with Clinical Notes.

Clinical notes are an essential component of a health record. This paper evaluates how natural language processing (NLP) can be used to identify the risk of acute care use (ACU) in oncology patients, once chemotherapy starts. Risk prediction using structured health data (SHD) is now standard, but predictions using free-text formats are complex. This paper explores the use of free-text notes for the prediction of ACU in leu of SHD. Deep Learning models were compared to manually engineered language features. Results show that SHD models minimally outperform NLP models; an ℓ1-penalised logistic regression with SHD achieved a C-statistic of 0.748 (95%-CI: 0.735, 0.762), while the same model with language features achieved 0.730 (95%-CI: 0.717, 0.745) and a transformer-based model achieved 0.702 (95%-CI: 0.688, 0.717). This paper shows how language models can be used in clinical applications and underlines how risk bias is different for diverse patient groups, even using only free-text data.

Predicting Depression Risk in Patients with Cancer Using Multimodal Data.

When patients with cancer develop depression, it is often left untreated. We developed a prediction model for depression risk within the first month after starting cancer treatment using machine learning and Natural Language Processing (NLP) models. The LASSO logistic regression model based on structured data performed well, whereas the NLP model based on only clinician notes did poorly. After further validation, prediction models for depression risk could lead to earlier identification and treatment of vulnerable patients, ultimately improving cancer care and treatment adherence.

Changes in postoperative opioid prescribing across three diverse healthcare systems, 2010-2020.

Objective: The opioid crisis brought scrutiny to opioid prescribing. Understanding how opioid prescribing patterns and corresponding patient outcomes changed during the epidemic is essential for future targeted policies. Many studies attempt to model trends in opioid prescriptions therefore understanding the temporal shift in opioid prescribing patterns across populations is necessary. This study characterized postoperative opioid prescribing patterns across different populations, 2010-2020. Data Source: Administrative data from Veteran Health Administration (VHA), six Medicaid state programs and an Academic Medical Center (AMC). Data extraction: Surgeries were identified using the Clinical Classifications Software. Study Design: Trends in average daily discharge Morphine Milligram Equivalent (MME), postoperative pain and subsequent opioid prescription were compared using regression and likelihood ratio test statistics. Principal Findings: The cohorts included 595,106 patients, with populations that varied considerably in demographics. Over the study period, MME decreased significantly at VHA (37.5-30.1; p = 0.002) and Medicaid (41.6-31.3; p = 0.019), and increased at AMC (36.9-41.7; p < 0.001). Persistent opioid users decreased after 2015 in VHA (p < 0.001) and Medicaid (p = 0.002) and increase at the AMC (p = 0.003), although a low rate was maintained. Average postoperative pain scores remained constant over the study period. Conclusions: VHA and Medicaid programs decreased opioid prescribing over the past decade, with differing response times and rates. In 2020, these systems achieved comparable opioid prescribing patterns and outcomes despite having very different populations. Acknowledging and incorporating these temporal distribution shifts into data learning models is essential for robust and generalizable models.

Deep Learning Approaches for Predicting Glaucoma Progression Using Electronic Health Records and Natural Language Processing.

Purpose: Advances in artificial intelligence have produced a few predictive models in glaucoma, including a logistic regression model predicting glaucoma progression to surgery. However, uncertainty exists regarding how to integrate the wealth of information in free-text clinical notes. The purpose of this study was to predict glaucoma progression requiring surgery using deep learning (DL) approaches on data from electronic health records (EHRs), including features from structured clinical data and from natural language processing of clinical free-text notes. Design: Development of DL predictive model in an observational cohort. Participants: Adult patients with glaucoma at a single center treated from 2008 through 2020. Methods: Ophthalmology clinical notes of patients with glaucoma were identified from EHRs. Available structured data included patient demographic information, diagnosis codes, prior surgeries, and clinical information including intraocular pressure, visual acuity, and central corneal thickness. In addition, words from patients' first 120 days of notes were mapped to ophthalmology domain-specific neural word embeddings trained on PubMed ophthalmology abstracts. Word embeddings and structured clinical data were used as inputs to DL models to predict subsequent glaucoma surgery. Main Outcome Measures: Evaluation metrics included area under the receiver operating characteristic curve (AUC) and F1 score, the harmonic mean of positive predictive value, and sensitivity on a held-out test set. Results: Seven hundred forty-eight of 4512 patients with glaucoma underwent surgery. The model that incorporated both structured clinical features as well as input features from clinical notes achieved an AUC of 73% and F1 of 40%, compared with only structured clinical features, (AUC, 66%; F1, 34%) and only clinical free-text features (AUC, 70%; F1, 42%). All models outperformed predictions from a glaucoma specialist's review of clinical notes (F1, 29.5%). Conclusions: We can successfully predict which patients with glaucoma will need surgery using DL models on EHRs unstructured text. Models incorporating free-text data outperformed those using only structured inputs. Future predictive models using EHRs should make use of information from within clinical free-text notes to improve predictive performance. Additional research is needed to investigate optimal methods of incorporating imaging data into future predictive models as well.

Exploring approaches for predictive cancer patient digital twins: Opportunities for collaboration and innovation.

We are rapidly approaching a future in which cancer patient digital twins will reach their potential to predict cancer prevention, diagnosis, and treatment in individual patients. This will be realized based on advances in high performance computing, computational modeling, and an expanding repertoire of observational data across multiple scales and modalities. In 2020, the US National Cancer Institute, and the US Department of Energy, through a trans-disciplinary research community at the intersection of advanced computing and cancer research, initiated team science collaborative projects to explore the development and implementation of predictive Cancer Patient Digital Twins. Several diverse pilot projects were launched to provide key insights into important features of this emerging landscape and to determine the requirements for the development and adoption of cancer patient digital twins. Projects included exploring approaches to using a large cohort of digital twins to perform deep phenotyping and plan treatments at the individual level, prototyping self-learning digital twin platforms, using adaptive digital twin approaches to monitor treatment response and resistance, developing methods to integrate and fuse data and observations across multiple scales, and personalizing treatment based on cancer type. Collectively these efforts have yielded increased insights into the opportunities and challenges facing cancer patient digital twin approaches and helped define a path forward. Given the rapidly growing interest in patient digital twins, this manuscript provides a valuable early progress report of several CPDT pilot projects commenced in common, their overall aims, early progress, lessons learned and future directions that will increasingly involve the broader research community.

Leveraging weak supervision to perform named entity recognition in electronic health records progress notes to identify the ophthalmology exam.

OBJECTIVE: To develop deep learning models to recognize ophthalmic examination components from clinical notes in electronic health records (EHR) using a weak supervision approach. METHODS: A corpus of 39,099 ophthalmology notes weakly labeled for 24 examination entities was assembled from the EHR of one academic center. Four pre-trained transformer-based language models (DistilBert, BioBert, BlueBert, and ClinicalBert) were fine-tuned to this named entity recognition task and compared to a baseline regular expression model. Models were evaluated on the weakly labeled test dataset, a human-labeled sample of that set, and a human-labeled independent dataset. RESULTS: On the weakly labeled test set, all transformer-based models had recall > 0.93, with precision varying from 0.815 to 0.843. The baseline model had lower recall (0.769) and precision (0.682). On the human-annotated sample, the baseline model had high recall (0.962, 95 % CI 0.955-0.067) with variable precision across entities (0.081-0.999). Bert models had recall ranging from 0.771 to 0.831, and precision >=0.973. On the independent dataset, precision was 0.926 and recall 0.458 for BlueBert. The baseline model had better recall (0.708, 95 % CI 0.674-0.738) but worse precision (0.399, 95 % CI -0.352-0.451). CONCLUSION: We developed the first deep learning system to recognize eye examination components from clinical notes, leveraging a novel opportunity for weak supervision. Transformer-based models had high precision on human-annotated labels, whereas the baseline model had poor precision but higher recall. This system may be used to improve cohort and feature identification using free-text notes.Our weakly supervised approach may help amass large datasets of domain-specific entities from EHRs in many fields.

Expanding the Secondary Use of Prostate Cancer Real World Data: Automated Classifiers for Clinical and Pathological Stage.

Background: Explicit documentation of stage is an endorsed quality metric by the National Quality Forum. Clinical and pathological cancer staging is inconsistently recorded within clinical narratives but can be derived from text in the Electronic Health Record (EHR). To address this need, we developed a Natural Language Processing (NLP) solution for extraction of clinical and pathological TNM stages from the clinical notes in prostate cancer patients. Methods: Data for patients diagnosed with prostate cancer between 2010 and 2018 were collected from a tertiary care academic healthcare system's EHR records in the United States. This system is linked to the California Cancer Registry, and contains data on diagnosis, histology, cancer stage, treatment and outcomes. A randomly selected sample of patients were manually annotated for stage to establish the ground truth for training and validating the NLP methods. For each patient, a vector representation of clinical text (written in English) was used to train a machine learning model alongside a rule-based model and compared with the ground truth. Results: A total of 5,461 prostate cancer patients were identified in the clinical data warehouse and over 30% were missing stage information. Thirty-three to thirty-six percent of patients were missing a clinical stage and the models accurately imputed the stage in 21-32% of cases. Twenty-one percent had a missing pathological stage and using NLP 71% of missing T stages and 56% of missing N stages were imputed. For both clinical and pathological T and N stages, the rule-based NLP approach out-performed the ML approach with a minimum F1 score of 0.71 and 0.40, respectively. For clinical M stage the ML approach out-performed the rule-based model with a minimum F1 score of 0.79 and 0.88, respectively. Conclusions: We developed an NLP pipeline to successfully extract clinical and pathological staging information from clinical narratives. Our results can serve as a proof of concept for using NLP to augment clinical and pathological stage reporting in cancer registries and EHRs to enhance the secondary use of these data.

A Comparison of Logistic Regression Against Machine Learning Algorithms for Gastric Cancer Risk Prediction Within Real-World Clinical Data Streams.

PURPOSE: Noncardia gastric cancer (NCGC) is a leading cause of global cancer mortality, and is often diagnosed at advanced stages. Development of NCGC risk models within electronic health records (EHR) may allow for improved cancer prevention. There has been much recent interest in use of machine learning (ML) for cancer prediction, but few studies comparing ML with classical statistical models for NCGC risk prediction. METHODS: We trained models using logistic regression (LR) and four commonly used ML algorithms to predict NCGC from age-/sex-matched controls in two EHR systems: Stanford University and the University of Washington (UW). The LR model contained well-established NCGC risk factors (intestinal metaplasia histology, prior Helicobacter pylori infection, race, ethnicity, nativity status, smoking history, anemia), whereas ML models agnostically selected variables from the EHR. Models were developed and internally validated in the Stanford data, and externally validated in the UW data. Hyperparameter tuning of models was achieved using cross-validation. Model performance was compared by accuracy, sensitivity, and specificity. RESULTS: In internal validation, LR performed with comparable accuracy (0.732; 95% CI, 0.698 to 0.764), sensitivity (0.697; 95% CI, 0.647 to 0.744), and specificity (0.767; 95% CI, 0.720 to 0.809) to penalized lasso, support vector machine, K-nearest neighbor, and random forest models. In external validation, LR continued to demonstrate high accuracy, sensitivity, and specificity. Although K-nearest neighbor demonstrated higher accuracy and specificity, this was offset by significantly lower sensitivity. No ML model consistently outperformed LR across evaluation criteria. CONCLUSION: Drawing data from two independent EHRs, we find LR on the basis of established risk factors demonstrated comparable performance to optimized ML algorithms. This study demonstrates that classical models built on robust, hand-chosen predictor variables may not be inferior to data-driven models for NCGC risk prediction.

Outcomes of Primary Trabeculectomy versus Combined Phacoemulsification-Trabeculectomy Using Automated Electronic Health Record Data Extraction.

PURPOSE: Cataract is a known effect of trabeculectomy (TE), but some surgeons are hesitant to perform combined phacoemulsification-TE (PTE) due to a risk of increased TE failure. Herein, we compare intraocular pressure (IOP) lowering between trabeculectomy (TE) and phacoemulsification-TE (PTE) and investigate factors that impact patient outcomes. METHODS: We performed a retrospective study of adults undergoing primary TE or PTE at our institution from 2010 to 2017. We used Kaplan-Meier survival analysis to investigate time to TE failure, and Cox proportional hazards modeling to investigate predictors of TE failure, defined as undergoing a second glaucoma surgery or using more IOP-lowering medications than pre-operatively. RESULTS: 318 surgeries (218 TE; 100 PTE) from 268 patients were included. Median follow-up time was 753 days. Mean baseline IOP was 21.1 mmHg. There were no significant differences in IOP between TE and PTE groups beyond postoperative year 1, with 28.9-46.5% of TE and 35.5-44.4% of PTE groups achieving IOP ≤10. Final IOP was similar in both groups (p = 0.22): 12.41 (SD 4.18) mmHg in the TE group and 14.05 (SD 5.45) in the PTE group. 84 (26.4%) surgeries met failure criteria. After adjusting for surgery type, sex, age, race, surgeon, and glaucoma diagnosis there were no significant differences in TE failure. CONCLUSION: This study suggests there is no significant difference in the risk of TE failure in patients receiving TE versus those receiving PTE.

Cost-Effectiveness Analysis and Microsimulation of Serial Multiparametric Magnetic Resonance Imaging in Active Surveillance of Localized Prostate Cancer.

PURPOSE: Many localized prostate cancers will follow an indolent course. Management has shifted toward active surveillance (AS), yet an optimal regimen remains controversial especially regarding expensive multiparametric magnetic resonance imaging (MRI). We aimed to assess cost-effectiveness of MRI in AS protocols. MATERIALS AND METHODS: A probabilistic microsimulation modeled individual patient trajectories for men diagnosed with low-risk cancer. We assessed no surveillance, up-front treatment (surgery or radiation), and scheduled AS protocols incorporating transrectal ultrasound-guided (TRUS) biopsy or MRI based regimens at serial intervals. Lifetime quality-adjusted life-years and costs adjusted to 2020 US$ were used to calculate expected net monetary benefit at $50,000/quality-adjusted life-year and incremental cost-effectiveness ratios. Uncertainty was assessed with probabilistic sensitivity analysis and linear regression metamodeling. RESULTS: Conservative management with AS outperformed up-front definitive treatment in a modeled cohort reflecting characteristics from a multi-institutional trial. Biopsy decision conditional on positive imaging (MRI triage) at 2-year intervals provided the highest expected net monetary benefit (incremental cost-effectiveness ratio $44,576). Biopsy after both positive and negative imaging (MRI pathway) and TRUS biopsy based regimens were not cost-effective. MRI triage resulted in fewer biopsies while reducing metastatic disease or cancer death. Results were sensitive to test performance and cost. MRI triage was the most likely cost-effective strategy on probabilistic sensitivity analysis. CONCLUSIONS: For men with low-risk prostate cancer, our modeling demonstrated that AS with sequential MRI triage is more cost-effective than biopsy regardless of imaging, TRUS biopsy alone or immediate treatment. AS guidelines should specify the role of imaging, and prospective studies should be encouraged.

Gaps in standardized postoperative pain management quality measures: A systematic review.

BACKGROUND: The goal of this study was an assessment of availability postoperative pain management quality measures and National Quality Forum-endorsed measures. Postoperative pain is an important clinical timepoint because poor pain control can lead to patient suffering, chronic opiate use, and/or chronic pain. Quality measures can guide best practices, but it is unclear whether there are measures for managing pain after surgery. METHODS: The National Quality Forum Quality Positioning System, Agency for Healthcare Research and Quality Indicators, and Centers for Medicare and Medicaid Services Measures Inventory Tool databases were searched in November 2019. We conducted a systematic literature review to further identify quality measures in research publications, clinical practice guidelines, and gray literature for the period between March 11, 2015 and March 11, 2020. RESULTS: Our systematic review yielded 1,328 publications, of which 206 were pertinent. Nineteen pain management quality measures were identified from the quality measure databases, and 5 were endorsed by National Quality Forum. The National Quality Forum measures were not specific to postoperative pain management. Three of the non-endorsed measures were specific to postoperative pain. CONCLUSION: The dearth of published postoperative pain management quality measures, especially National Quality Forum-endorsed measures, highlights the need for more rigorous evidence and widely endorsed postoperative pain quality measures to guide best practices.

Machine Learning Applied to Electronic Health Records: Identification of Chemotherapy Patients at High Risk for Preventable Emergency Department Visits and Hospital Admissions.

PURPOSE: Acute care use (ACU) is a major driver of oncologic costs and is penalized by a Centers for Medicare & Medicaid Services quality measure, OP-35. Targeted interventions reduce preventable ACU; however, identifying which patients might benefit remains challenging. Prior predictive models have made use of a limited subset of the data in the electronic health record (EHR). We aimed to predict risk of preventable ACU after starting chemotherapy using machine learning (ML) algorithms trained on comprehensive EHR data. METHODS: Chemotherapy patients treated at an academic institution and affiliated community care sites between January 2013 and July 2019 who met inclusion criteria for OP-35 were identified. Preventable ACU was defined using OP-35 criteria. Structured EHR data generated before chemotherapy treatment were obtained. ML models were trained to predict risk for ACU after starting chemotherapy using 80% of the cohort. The remaining 20% were used to test model performance by the area under the receiver operator curve. RESULTS: Eight thousand four hundred thirty-nine patients were included, of whom 35% had preventable ACU within 180 days of starting chemotherapy. Our primary model classified patients at risk for preventable ACU with an area under the receiver operator curve of 0.783 (95% CI, 0.761 to 0.806). Performance was better for identifying admissions than emergency department visits. Key variables included prior hospitalizations, cancer stage, race, laboratory values, and a diagnosis of depression. Analyses showed limited benefit from including patient-reported outcome data and indicated inequities in outcomes and risk modeling for Black and Medicaid patients. CONCLUSION: Dense EHR data can identify patients at risk for ACU using ML with promising accuracy. These models have potential to improve cancer care outcomes, patient experience, and costs by allowing for targeted, preventative interventions.