Variational Bayes machine learning for risk adjustment of general outcome indicators with examples in urology.
Journal
NPJ digital medicine
ISSN: 2398-6352
Titre abrégé: NPJ Digit Med
Pays: England
ID NLM: 101731738
Informations de publication
Date de publication:
14 Sep 2024
14 Sep 2024
Historique:
received:
03
11
2023
accepted:
01
09
2024
medline:
15
9
2024
pubmed:
15
9
2024
entrez:
14
9
2024
Statut:
epublish
Résumé
Risk adjustment is often necessary for outcome quality indicators (QIs) to provide fair and accurate feedback to healthcare professionals. However, traditional risk adjustment models are generally oversimplified and not equipped to disentangle complex factors influencing outcomes that are out of a healthcare professional's control. We present VIRGO, a novel variational Bayes model trained on routinely collected, large administrative datasets to risk-adjust outcome QIs. VIRGO uses detailed demographics, diagnosis, and procedure codes to provide individualized risk adjustment and explanations on patient factors affecting outcomes. VIRGO achieves state-of-the-art on external datasets and features capabilities of uncertainty expression, explainable features, and counterfactual analysis capabilities. VIRGO facilitates risk adjustment by explaining how patient factors led to adverse outcomes and expresses the uncertainty of each prediction, allowing healthcare professionals to not only explore patient factors with unexplained variance that are associated with worse outcomes but also reflect on the quality of their clinical practice.
Identifiants
pubmed: 39277683
doi: 10.1038/s41746-024-01244-z
pii: 10.1038/s41746-024-01244-z
doi:
Types de publication
Journal Article
Langues
eng
Pagination
249Informations de copyright
© 2024. The Author(s).
Références
Evans, S. M., Bohensky, M., Cameron, P. A. & McNeil, J. A survey of Australian clinical registries: can quality of care be measured? Intern. Med. J. 41, 42–48 (2011).
doi: 10.1111/j.1445-5994.2009.02068.x
pubmed: 19811553
McNeil, J. J., Evans, S. M., Johnson, N. P. & Cameron, P. A. Clinical-quality registries: their role in quality improvement. Med. J. Aust. 192, 244–245 (2010).
doi: 10.5694/j.1326-5377.2010.tb03499.x
pubmed: 20201755
Ivers, N. et al. Audit and feedback: effects on professional practice and healthcare outcomes. Cochrane. Database. Syst. Rev. https://doi.org/10.1002/14651858.CD000259.pub3 (2012).
doi: 10.1002/14651858.CD000259.pub3
pubmed: 22696318
pmcid: 11338587
The Australian Commission on Safety and Quality in Health Care. Economic Evaluation of Clinical Quality Registries: Final Report. https://www.safetyandquality.gov.au/wp-content/uploads/2016/12/Economic-evaluation-of-clinical-quality-registries-Final-report-Nov-2016.pdf (2016).
Copnell, B. et al. Measuring the quality of hospital care: an inventory of indicators. Intern. Med. J. 39, 352–360 (2009).
doi: 10.1111/j.1445-5994.2009.01961.x
pubmed: 19323697
Koh, H. J. W. et al. Quality indicators in the clinical specialty of urology: a systematic review. Eur. Urol. Focus.9, 435–446 (2022).
Lane-Fall, M. B. & Neuman, M. D. Outcomes measures and risk adjustment. Int. Anesthesiol. Clin. 51, 10–21 (2013).
doi: 10.1097/AIA.0b013e3182a70a52
pubmed: 24088885
pmcid: 3816010
Ossai, C. I., Rankin, D. & Wickramasinghe, N. Preadmission assessment of extended length of hospital stay with RFECV-ETC and hospital-specific data. Eur. J. Med. Res. 27, 128 (2022).
doi: 10.1186/s40001-022-00754-4
pubmed: 35879803
pmcid: 9310419
Irvin, J. A. et al. Incorporating machine learning and social determinants of health indicators into prospective risk adjustment for health plan payments. BMC Public Health 20, 608 (2020).
doi: 10.1186/s12889-020-08735-0
pubmed: 32357871
pmcid: 7195714
Kan, H. J. et al. Exploring the use of machine learning for risk adjustment: a comparison of standard and penalized linear regression models in predicting health care costs in older adults. PLOS ONE 14, e0213258 (2019).
doi: 10.1371/journal.pone.0213258
pubmed: 30840682
pmcid: 6402678
Assareh, H., Achat, H. M., Stubbs, J. M., Guevarra, V. M. & Hill, K. Incidence and variation of discrepancies in recording chronic conditions in Australian hospital administrative data. PloS ONE 11, e0147087 (2016).
doi: 10.1371/journal.pone.0147087
pubmed: 26808428
pmcid: 4726608
Janssen, A., Talic, S., Gasevic, D., Kay, J. & Shaw, T. Exploring the Intersection Between Health Professionals’ Learning and eHealth Data: Protocol for a Comprehensive Research Program in Practice Analytics in Health Care. JMIR Res. Protoc. 10, e27984 (2021).
doi: 10.2196/27984
pubmed: 34889768
pmcid: 8727065
Herm, L.-V., Heinrich, K., Wanner, J. & Janiesch, C. Stop ordering machine learning algorithms by their explainability! a user-centered investigation of performance and explainability. Int. J. Inf. Manag. 69, 102538 (2023).
doi: 10.1016/j.ijinfomgt.2022.102538
Kruschke, J. K. Bayesian analysis reporting guidelines. Nat. Hum. Behav. 5, 1282–1291 (2021).
doi: 10.1038/s41562-021-01177-7
pubmed: 34400814
pmcid: 8526359
Department of Health. Victoria, Australia. Victorian Admitted Episodes Dataset. https://www.health.vic.gov.au/data-reporting/victorian-admitted-episodes-dataset (2024).
Shaaban, A. N., Peleteiro, B. & Martins, M. R. O. Statistical models for analyzing count data: predictors of length of stay among HIV patients in Portugal using a multilevel model. BMC Health Serv. Res. 21, 372 (2021).
doi: 10.1186/s12913-021-06389-1
pubmed: 33882911
pmcid: 8061202
Diamantopoulos, L. N. et al. Patterns and timing of perioperative blood transfusion and association with outcomes after radical cystectomy. Urol. Oncol. Semin. Orig. Investig. 39, 496.e1–496.e8 (2021).
Abu-Ghanem, Y. & Ramon, J. Impact of perioperative blood transfusions on clinical outcomes in patients undergoing surgery for major urologic malignancies. Ther. Adv. Urol. 11, 1756287219868054 (2019).
doi: 10.1177/1756287219868054
pubmed: 31447936
pmcid: 6691668
Wang, Y.-L. et al. Perioperative blood transfusion promotes worse outcomes of bladder cancer after radical cystectomy: a systematic review and meta-analysis. PLoS ONE 10, e0130122 (2015).
doi: 10.1371/journal.pone.0130122
pubmed: 26080092
pmcid: 4469696
Callum, J. & Siemens, D. R. We should redouble efforts to minimize transfusions in urological surgery. J. Urol. 209, 471–473 (2023).
doi: 10.1097/JU.0000000000003146
pubmed: 36629474
Sarkies, M. N., White, J., Henderson, K., Haas, R. & Bowles, J. Additional weekend allied health services reduce length of stay in subacute rehabilitation wards but their effectiveness and cost-effectiveness are unclear in acute general medical and surgical hospital wards: a systematic review. J. Physiother. 64, 142–158 (2018).
doi: 10.1016/j.jphys.2018.05.004
pubmed: 29929739
Barr, M. L. et al. Understanding the use and impact of allied health services for people with chronic health conditions in Central and Eastern Sydney, Australia: a five-year longitudinal analysis. Prim. Health Care Res. Dev. 20, e141 (2019).
doi: 10.1017/S146342361900077X
pubmed: 31640837
pmcid: 6842649
Juhnke, C., Bethge, S. & Mühlbacher, A. C. A review on methods of risk adjustment and their use in integrated healthcare systems. Int. J. Integr. Care 16, 4 (2026).
Pfohl, S. R., Duan, T., Ding, D. Y. & Shah, N. H. Counterfactual reasoning for fair clinical risk prediction. In Proc. 4th machine learning for healthcare conference 325–358 (PMLR, 2019).
Seoni, S. et al. Application of uncertainty quantification to artificial intelligence in healthcare: a review of last decade (2013–2023). Comput. Biol. Med. 165, 107441 (2023).
doi: 10.1016/j.compbiomed.2023.107441
pubmed: 37683529
Jain, R., Singh, M., Rao, A. R. & Garg, R. Machine learning models to predict length of stay in hospitals. In Proc. 10th International Conference on Healthcare Informatics (ICHI) 545–546 https://doi.org/10.1109/ICHI54592.2022.00105 (IEEE, Rochester, MN, USA, 2022)
Dietterich, T. G. An experimental comparison of three methods for constructing ensembles of decision trees: bagging, boosting, and randomization. Mach. Learn. 40, 139–157 (2000).
Szakiel, J. Hospital casemix protocol-medibank private perspective. Health Inf. Manag. J. 39, 47–49 (2010).
Institute of Medicine. Crossing the Quality Chasm: A New Health System for the 21st Century. https://doi.org/10.17226/10027 (2001).
van Mourik, M. S. M., van Duijn, P. J., Moons, K. G. M., Bonten, M. J. M. & Lee, G. M. Accuracy of administrative data for surveillance of healthcare-associated infections: a systematic review. BMJ Open 5, e008424 (2015).
doi: 10.1136/bmjopen-2015-008424
pubmed: 26316651
pmcid: 4554897
Agrawal, R., Trippe, B., Huggins, J. & Broderick, T. The Kernel Interaction Trick: Fast Bayesian Discovery of Pairwise Interactions in High Dimensions. in Proceedings of the 36th International Conference on Machine Learning 141–150 (PMLR, 2019).
Leckie, G. Multiple membership multilevel models. https://doi.org/10.48550/arXiv.1907.04148 (2016).
Blei, D. M., Kucukelbir, A. & McAuliffe, J. D. Variational inference: a review for statisticians. J. Am. Stat. Assoc. 112, 859–877 (2017).
doi: 10.1080/01621459.2017.1285773
Lingsma, H. F. et al. Evaluation of hospital outcomes: the relation between length-of-stay, readmission, and mortality in a large international administrative database. BMC Health Serv. Res. 18, 116 (2018).
doi: 10.1186/s12913-018-2916-1
pubmed: 29444713
pmcid: 5813333
Australian Commission on Safety and Quality in Health Care. Hospital-Acquired Complications (HACs) List. https://www.safetyandquality.gov.au/publications-and-resources/resource-library/hospital-acquired-complications-hacs-list-specifications-version-31-12th-edn (2019).
Versace, V. L., Skinner, T. C., Bourke, L., Harvey, P. & Barnett, T. National analysis of the modified Monash model, population distribution and a socio-economic index to inform rural health workforce planning. Aust. J. Rural Health 29, 801–810 (2021).
doi: 10.1111/ajr.12805
pubmed: 34672057
World Health Organization. International Classification of Diseases (ICD). https://www.who.int/standards/classifications/classification-of-diseases (2019).
Quan, H. et al. Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Med. Care 43, 1130–1139 (2005).
doi: 10.1097/01.mlr.0000182534.19832.83
pubmed: 16224307
Gelman, A., Simpson, D. & Betancourt, M. The prior can often only be understood in the context of the likelihood. Entropy 19, 555 (2017).
doi: 10.3390/e19100555
Preisser, J. S., Stamm, J. W., Long, D. L. & Kincade, M. E. Review and recommendations for zero-inflated count regression modeling of dental caries indices in epidemiological studies. Caries Res. 46, 413–423 (2012).
doi: 10.1159/000338992
pubmed: 22710271
Piironen, J. & Vehtari, A. Sparsity information and regularization in the horseshoe and other shrinkage priors. Electron. J. Stat. 11, 5018–5051 (2017).
doi: 10.1214/17-EJS1337SI
Vehtari, A., Gelman, A. & Gabry, J. Practical Bayesian model evaluation using leave-one-out cross-validation and WAIC. Stat. Comput. 27, 1413–1432 (2017).
doi: 10.1007/s11222-016-9696-4
Florek, P. & Zagdański, A. Benchmarking state-of-the-art gradient boosting algorithms for classification. Preprint at https://doi.org/10.48550/arXiv.2305.17094 (2023).
Bergstra, J., Bardenet, R., Bengio, Y. & Kégl, B. Algorithms for Hyper-Parameter Optimization. In Proc. Advances in Neural Information Processing Systems 24 (Curran Associates, Inc., 2011).
Phan, D., Pradhan, N. & Jankowiak, M. Composable Effects for Flexible and Accelerated Probabilistic Programming in NumPyro. Preprint at https://doi.org/10.48550/arXiv.1912.11554 (2019).