Search

Background: Prostate cancer is the second leading cause of death among American men. If detected and treated at an early stage, prostate cancer is often curable. However, an advanced stage such as metastatic castration-resistant prostate cancer (mCRPC) has a high risk of mortality. Multiple treatment options exist, the most common included docetaxel, abiraterone, and enzalutamide. Docetaxel is a cytotoxic chemotherapy, whereas abiraterone and enzalutamide are androgen receptor pathway inhibitors (ARPI). ARPIs are preferred over docetaxel due to lower toxicity. No study has used machine learning with patients' demographics, test results, and comorbidities to identify heterogeneous treatment rules that might improve the survival duration of patients with mCRPC., Objective: This study aimed to measure patient-level heterogeneity in the association of medication prescribed with overall survival duration (in the form of follow-up days) and arrive at a set of medication prescription rules using patient demographics, test results, and comorbidities., Methods: We excluded patients with mCRPC who were on docetaxel, cabaxitaxel, mitoxantrone, and sipuleucel-T either before or after the prescription of an ARPI. We included only the African American and white populations. In total, 2886 identified veterans treated for mCRPC who were prescribed either abiraterone or enzalutamide as the first line of treatment from 2014 to 2017, with follow-up until 2020, were analyzed. We used causal survival forests for analysis. The unit level of analysis was the patient. The primary outcome of this study was follow-up days indicating survival duration while on the first-line medication. After estimating the treatment effect, a prescription policy tree was constructed., Results: For 2886 veterans, enzalutamide is associated with an average of 59.94 (95% CI 35.60-84.28) more days of survival than abiraterone. The increase in overall survival duration for the 2 drugs varied across patient demographics, test results, and comorbidities. Two data-driven subgroups of patients were identified by ranking them on their augmented inverse-propensity weighted (AIPW) scores. The average AIPW scores for the 2 subgroups were 19.36 (95% CI -16.93 to 55.65) and 100.68 (95% CI 62.46-138.89). Based on visualization and t test, the AIPW score for low and high subgroups was significant (P=.003), thereby supporting heterogeneity. The analysis resulted in a set of prescription rules for the 2 ARPIs based on a few covariates available to the physicians at the time of prescription., Conclusions: This study of 2886 veterans showed evidence of heterogeneity and that survival days may be improved for certain patients with mCRPC based on the medication prescribed. Findings suggest that prescription rules based on the patient characteristics, laboratory test results, and comorbidities available to the physician at the time of prescription could improve survival by providing personalized treatment decisions., (©Deepika Gopukumar, Nirup Menon, Martin W Schoen. Originally published in JMIR Medical Informatics (https://medinform.jmir.org), 19.11.2024.)

Background: Currently, there is no consensus on how to comprehensively assess comorbidities in lung cancer patients in the clinical setting. Prescription medications may be a preferred comorbidity assessment tool and provide a simple mechanism for predicting postoperative outcomes for lung cancer. We examined the relationship between prescription medications and postoperative outcomes for early-stage non-small cell lung cancer (NSCLC)., Methods: We conducted a retrospective cohort study of patients with clinical stage I NSCLC who underwent surgical resection in the Veterans Health Administration (VHA) between 10/01/2006 and 09/30/2016. Details of all outpatient prescriptions filled by patients within the VHA system from 1-year up to 14 days before surgery were collected. Medications were categorized using the Anatomical Therapeutic Chemical (ATC) Level One classification system. We assessed the association of medications prescribed in the year prior to surgery with postoperative adverse events (composite of death or major complication) at 30 and 90 days following surgery and overall survival (OS)., Results: We included 9,741 veterans in the analysis. The median number of prescription medications filled in the year preceding surgery was 11 (interquartile range: 7-16). In multivariable-adjusted analyses, a higher number of prescription medications was associated with increased risk of 30-day [multivariable-adjusted odds ratio (aOR): 1.016; 95% confidence interval (CI): 1.007-1.026] and 90-day postoperative adverse events (aOR: 1.015; 95% CI: 1.006-1.024) and decreased OS (adjusted hazard ratio: 1.019; 95% CI: 1.014-1.023). Within a subgroup of patients with a high comorbidity burden (Charlson-Deyo Comorbidity Index score of 6-8), a higher number of prescription medications was also associated with reduced OS (P<0.001). Patients prescribed medications from the ATC respiratory system class had elevated risk of postoperative adverse events at 30 days (aOR: 1.255; 95% CI: 1.095-1.439) and 90 days (aOR: 1.254; 95% CI: 1.097-1.434) compared to patients without these prescription medications. Significantly increased odds for 90-day postoperative adverse events were observed with each additional prescription medication from the ATC respiratory (aOR: 1.057; 95% CI: 1.027-1.088) and nervous system (aOR: 1.035; 95% CI: 1.005-1.066) classes., Conclusions: The number of medications prescribed preoperatively is associated with short- and long-term postoperative outcomes for early-stage NSCLC, even when adjusting for several covariates including age and comorbidity burden. Patients prescribed a higher number of medications acting primarily on the respiratory and nervous systems are at elevated risk for postoperative adverse events after curative-intent resection. Prescription medications may be a reliable tool to assess comorbidities and perioperative risk for patients with NSCLC., Competing Interests: Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jtd.amegroups.com/article/view/10.21037/jtd-24-803/coif). D.K. serves as an unpaid editorial board member of Journal of Thoracic Disease from April 2024 to June 2026. S.T. was supported by the Washington University School of Medicine (WUSM) Surgical Oncology Basic Science and Translational Research Training Program grant T32CA009621, from the National Cancer Institute. D.B.E. Jr receives support from VHA 1I01HX002475-01A2 grant. N.E.R. was supported in part by the WUSM StARR Program in Cross-Disciplinary Oncology Clinician-Scientist Training R38CA255575. B.T.H. was funded in part by NIH grant 5T32HL007776-25. B.T.H. is a former consultant at Oncocyte Corporation and was an MBA intern (at Eli Lilly and Company). Y.Y. was funded in part through a VHA 1I01HX002475-01A2 grant. D.K. is supported by NIH grants 1P01AI116501, R01HL094601, R01HL151078, U01163086-01, United States, Veterans Administration Merit Review grant 1I01BX002730, United States, and The Foundation for Barnes-Jewish Hospital, United States. S.H.C. was funded in part through a VHA 1I01HX002475-01A2 grant. V.P. has received the following grants for projects: 1I01HX002475-01A2, R01HL146856, R01CA258681, MATF. V.P. has received payments from PrecisCa (panel discussions) and V.P.’s spouse owns stock in Intuitive Surgical. M.W.S. has funding through the Congressionally Directed Medical Research Program DoD W81XWH-22-1-0602 and has received speaking fees from Pfizer. The other authors have no conflicts of interest to declare., (2024 AME Publishing Company. All rights reserved.)

Background/aim: With new therapies for metastatic prostate cancer, patients are living longer, increasing the need for better understanding of the impact of comorbid disease. Prescription medications may risk-stratify patients independent of established methods, such as the Charlson Comorbidity Index (CCI) and guide treatment selection., Patients and Methods: In a nationwide retrospective study of US Veterans, we used multivariable logistic regression and Cox proportional hazard modeling to evaluate the association between number and class of prescription medications and overall survival (OS) with age, race, body-mass index, prostate specific antigen (PSA), and Charlson comorbidities as covariates in veterans treated for de novo metastatic hormone sensitive prostate cancer (mHSPC) between 2010-2021., Results: Among 8,434 Veterans, a median of nine medications and five medication classes were filled in the year prior to initial treatment with abiraterone or enzalutamide for mHSPC. Veterans on 1-4 medications had an average survival of 38 months compared to 5-9 medicines (33 months), 10-14 medicines (27 months), and 15+ medicines (22 months) (p<0.001). After adjusting for age, race, body mass index (BMI), PSA, CCI, and year of diagnosis, both the number of medications and medication classes were associated with increased mortality. The adjusted hazard ratio (aHR) [95% confidence interval (CI)] was 1.03 (1.02-1.03) for the number of medications and 1.05 (1.04-1.07) for medication classes. Medications within ATC B (blood/blood forming organs), ATC C (cardiovascular), and ATC N (nervous) were associated with worse OS, with aHRs of 1.14 (1.07, 1.21), 1.14 (1.06, 1.22), and 1.12 (1.06, 1.19), respectively., Conclusion: The number and class of medications were independently associated with overall survival in patients undergoing treatment for mHSPC. With new therapies for advanced prostate cancer, patients are living longer, highlighting the need for a better understanding of the impact of comorbid diseases. Simple methods to assess disease burden and prognosticate survival have the potential to guide treatment decisions., (Copyright © 2024 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.)

Background: Lung function is routinely assessed prior to surgical resection for non-small cell lung cancer (NSCLC). Further assessment of chronic obstructive pulmonary disease (COPD) using inhaled COPD medications to determine disease severity, a readily available metric of disease burden, may predict postoperative outcomes and overall survival (OS) in lung cancer patients undergoing surgery., Methods: We retrospectively evaluated clinical stage I NSCLC patients receiving surgical treatment within the Veterans Health Administration from 2006-2016 to determine the relationship between number and type of inhaled COPD medications (short- and long-acting beta2-agonists, muscarinic antagonists, or corticosteroids prescribed within 1 year before surgery) and postoperative outcomes including OS using multivariable models. We also assessed the relationship between inhaled COPD medications, disease severity [measured by forced expiratory volume in 1 second (FEV1)], and diagnosis of COPD., Results: Among 9,741 veterans undergoing surgery for clinical stage I NSCLC, patients with COPD were more likely to be prescribed inhaled medications than those without COPD [odds ratio (OR) =5.367, 95% confidence interval (CI): 4.886-5.896]. Increased severity of COPD was associated with increased number of prescribed inhaled COPD medications (P<0.0001). The number of inhaled COPD medications was associated with prolonged hospital stay [adjusted OR (aOR) =1.119, 95% CI: 1.076-1.165), more major complications (aOR =1.117, 95% CI: 1.074-1.163), increased 90-day mortality (aOR =1.088, 95% CI: 1.013-1.170), and decreased OS [adjusted hazard ratio (aHR) =1.061, 95% CI: 1.042-1.080]. In patients with FEV1 ≥80% predicted, greater number of prescribed inhaled COPD medications was associated with increased 30-day mortality (aOR =1.265, 95% CI: 1.062-1.505), prolonged hospital stay (aOR =1.130, 95% CI: 1.051-1.216), more major complications (aOR =1.147, 95% CI: 1.064-1.235), and decreased OS (aHR =1.058, 95% CI: 1.022-1.095). When adjusting for other drug classes and covariables, short-acting beta2-agonists were associated with increased 90-day mortality (aOR =1.527, 95% CI: 1.120-2.083) and decreased OS (aHR =1.087, 95% CI: 1.005-1.177)., Conclusions: In patients with early-stage NSCLC, inhaled COPD medications prescribed prior to surgery were associated with both short- and long-term outcomes, including in patients with FEV1 ≥80% predicted. Routine assessment of COPD medications may be a simple method to quantify operative risk in early-stage NSCLC patients., Competing Interests: Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jtd.amegroups.com/article/view/10.21037/jtd-23-1273/coif). S.T. was supported in part by the Washington University School of Medicine (WUSM) Surgical Oncology Basic Science and Translational Research Training Program grant T32CA009621 from the NCI. B.T.H. has funding through NIH Grant 5T32HL007776-25, is a former consultant at Oncocyte Corporation, and is an MBA intern (at Eli Lilly and Company). N.E.R. was supported by the WUSM StARR Program in Cross-Disciplinary Oncology Clinician-Scientist Training R38 CA 255575. V.R. was supported through a NHLBI F30 F30HL151136 grant. B.D.K. was supported in part by NCI Grant R01CA258681. V.P. has received the following grants for projects: I01 HX002475, R01HL146856, R01CA258681, MATF and has funding through VHA Grant 1I01HX002475-01A2. V.P. has also received speaking fees from PrecisCa and his spouse has stock in Intuitive Surgical. S.H.C., Y.Y., and D.B.E. have funding through VHA Grant 1I01HX002475-01A2. M.W.S. has funding through the Congressionally Directed Medical Research Program DoD W81XWH-22-1-0602 and received speaking fees from Pfizer. The other authors have no conflicts of interest to declare., (2023 Journal of Thoracic Disease. All rights reserved.)

Background: The Centers for Medicare and Medicaid Services projects that health care costs will continue to grow over the next few years. Rising readmission costs contribute significantly to increasing health care costs. Multiple areas of health care, including readmissions, have benefited from the application of various machine learning algorithms in several ways., Objective: We aimed to identify suitable models for predicting readmission charges billed by hospitals. Our literature review revealed that this application of machine learning is underexplored. We used various predictive methods, ranging from glass-box models (such as regularization techniques) to black-box models (such as deep learning-based models)., Methods: We defined readmissions as readmission with the same major diagnostic category (RSDC) and all-cause readmission category (RADC). For these readmission categories, 576,701 and 1,091,580 individuals, respectively, were identified from the Nationwide Readmission Database of the Healthcare Cost and Utilization Project by the Agency for Healthcare Research and Quality for 2013. Linear regression, lasso regression, elastic net, ridge regression, eXtreme gradient boosting (XGBoost), and a deep learning model based on multilayer perceptron (MLP) were the 6 machine learning algorithms we tested for RSDC and RADC through 10-fold cross-validation., Results: Our preliminary analysis using a data-driven approach revealed that within RADC, the subsequent readmission charge billed per patient was higher than the previous charge for 541,090 individuals, and this number was 319,233 for RSDC. The top 3 major diagnostic categories (MDCs) for such instances were the same for RADC and RSDC. The average readmission charge billed was higher than the previous charge for 21 of the MDCs in the case of RSDC, whereas it was only for 13 of the MDCs in RADC. We recommend XGBoost and the deep learning model based on MLP for predicting readmission charges. The following performance metrics were obtained for XGBoost: (1) RADC (mean absolute percentage error [MAPE]=3.121%; root mean squared error [RMSE]=0.414; mean absolute error [MAE]=0.317; root relative squared error [RRSE]=0.410; relative absolute error [RAE]=0.399; normalized RMSE [NRMSE]=0.040; mean absolute deviation [MAD]=0.031) and (2) RSDC (MAPE=3.171%; RMSE=0.421; MAE=0.321; RRSE=0.407; RAE=0.393; NRMSE=0.041; MAD=0.031). The performance obtained for MLP-based deep neural networks are as follows: (1) RADC (MAPE=3.103%; RMSE=0.413; MAE=0.316; RRSE=0.410; RAE=0.397; NRMSE=0.040; MAD=0.031) and (2) RSDC (MAPE=3.202%; RMSE=0.427; MAE=0.326; RRSE=0.413; RAE=0.399; NRMSE=0.041; MAD=0.032). Repeated measures ANOVA revealed that the mean RMSE differed significantly across models with P<.001. Post hoc tests using the Bonferroni correction method indicated that the mean RMSE of the deep learning/XGBoost models was statistically significantly (P<.001) lower than that of all other models, namely linear regression/elastic net/lasso/ridge regression., Conclusions: Models built using XGBoost and MLP are suitable for predicting readmission charges billed by hospitals. The MDCs allow models to accurately predict hospital readmission charges., (©Deepika Gopukumar, Abhijeet Ghoshal, Huimin Zhao. Originally published in JMIR Medical Informatics (https://medinform.jmir.org), 30.08.2022.)

Big data technologies are increasingly used for biomedical and health-care informatics research. Large amounts of biological and clinical data have been generated and collected at an unprecedented speed and scale. For example, the new generation of sequencing technologies enables the processing of billions of DNA sequence data per day, and the application of electronic health records (EHRs) is documenting large amounts of patient data. The cost of acquiring and analyzing biomedical data is expected to decrease dramatically with the help of technology upgrades, such as the emergence of new sequencing machines, the development of novel hardware and software for parallel computing, and the extensive expansion of EHRs. Big data applications present new opportunities to discover new knowledge and create novel methods to improve the quality of health care. The application of big data in health care is a fast-growing field, with many new discoveries and methodologies published in the last five years. In this paper, we review and discuss big data application in four major biomedical subdisciplines: (1) bioinformatics, (2) clinical informatics, (3) imaging informatics, and (4) public health informatics. Specifically, in bioinformatics, high-throughput experiments facilitate the research of new genome-wide association studies of diseases, and with clinical informatics, the clinical field benefits from the vast amount of collected patient data for making intelligent decisions. Imaging informatics is now more rapidly integrated with cloud platforms to share medical image data and workflows, and public health informatics leverages big data techniques for predicting and monitoring infectious disease outbreaks, such as Ebola. In this paper, we review the recent progress and breakthroughs of big data applications in these health-care domains and summarize the challenges, gaps, and opportunities to improve and advance big data applications in health care.