Digital Twin Modelling Of Patient Journeys in Healthcare Relationship Management Systems
DOI:
https://doi.org/10.63282/3050-9262.IJAIDSML-V7I3P112Keywords:
Digital Twin, Patient Journey, Healthcare Relationship Management, Journey State Modelling, What-If Simulation, Next Best Action, Patient Engagement, Care-Gap Management, Risk Calibration, Privacy-Preserving Simulation, Clinical Decision SupportAbstract
Healthcare relationship management systems record what has already happened to a patient. They hold appointments, milestones, care gaps, service cases and outreach history, and they act on that record through rules and schedules. What they do not do well is anticipate what happens next, which leaves engagement decisions reactive: contact after a missed appointment rather than before it, channel chosen by predefined rule rather than by predicted responsiveness, and care gaps addressed once already overdue. This paper describes a digital twin of the patient journey built inside a relationship management platform, and reports a research evaluation of it. The twin models the journey as a dynamic personalised process rather than as a copy of the medical record. It holds where the patient sits in the care and engagement lifecycle, how prior interactions produced that position, and the likely next steps under different engagement strategies. It produces three kinds of output: predictions and risk scores, what-if simulations comparing candidate interventions, and an engagement recommendation carrying a next best action, timing, channel, priority and the factors behind it. Validation used a temporal holdout on real reconstructed journeys, with the twin initialised at an earlier point and asked to simulate forward without seeing later events, compared on next-state accuracy, sequence similarity, timing accuracy, risk calibration, subgroup performance, and clinical and operational review of unusual transitions. Messy real-world behaviour was deliberately retained. The evaluation covered about 50,000 patient journeys across five service lines, with an architecture built to support 8 to 10 facilities on a shared platform. Against the existing rule-based engagement process, engagement and response rate improved by about 18 to 25 per cent, appointment adherence by about 12 to 18 per cent, care-gap and follow-up completion by about 15 to 22 per cent, engagement response time for prioritised cases fell by about 25 to 35 per cent, unnecessary outreach fell by about 15 to 20 per cent, and engagement-related operational efficiency improved by about 10 to 18 per cent. Every figure in this paper is a research-evaluation result. None is an audited production outcome, and every range is reported as a range because no point estimate was substantiated. The evaluation also produced findings that changed engagement strategy: timing mattered more than message count, repeated contact on one channel reduced responsiveness, and journey friction was routinely misread as disengagement. Sometimes the best action was no immediate outreach. A three-layer separation of patient identity, clinical data and simulation lets the twin operate on representations rather than on identified records, and simulation output is treated as sensitive in its own right. Throughout, the system is decision support and not autonomous clinical decision-making.
References
[1] Ascarza, E. (2018). Retention Futility: Targeting High-Risk Customers Might be Ineffective. Journal of Marketing Research, 55(1), 80-98. doi:10.1509/jmr.16.0163
[2] Dabbara, B. N. (2026). Ethical AI Governance in Healthcare CRM Systems. International Journal on Science and Technology, 17(2). doi:10.71097/ijsat.v17.i2.11021
[3] Dabbara, B. N. (2026). Integrating Conversational AI into Healthcare CRMs for Patient Engagement Optimization. International Journal of Emerging Trends in Computer Science and Information Technology, 7(1), 16-18. https://doi.org/10.63282/3050-9246.IJETCSIT-V7I1P103
[4] Dabbara, B. N. (2026). Real-Time Care Recommendations in Healthcare Using Agent-to-User Interaction (A2UI) on the Salesforce Ecosystem. International Journal of Innovative Research and Creative Technology, 12(3). doi:10.62970/ijirct.v12.i3.2606012
[5] Dantas, L. F., Fleck, J. L., Cyrino Oliveira, F. L., and Hamacher, S. (2018). No-shows in appointment scheduling – a systematic literature review. Health Policy, 122(4), 412-421. doi:10.1016/j.healthpol.2018.02.002
[6] El Emam, K., Jonker, E., Arbuckle, L., and Malin, B. (2011). A Systematic Review of Re-Identification Attacks on Health Data. PLoS ONE, 6(12), e28071. doi:10.1371/journal.pone.0028071
[7] Finlayson, S. G., Subbaswamy, A., Singh, K., Bowers, J., Kupke, A., Zittrain, J., Kohane, I. S., and Saria, S. (2021). The Clinician and Dataset Shift in Artificial Intelligence. New England Journal of Medicine, 385(3), 283-286. doi:10.1056/NEJMc2104626
[8] Grieves, M., and Vickers, J. (2017). Digital Twin: Mitigating Unpredictable, Undesirable Emergent Behavior in Complex Systems. In Transdisciplinary Perspectives on Complex Systems, pages 85-113. Springer International Publishing. doi:10.1007/978-3-319-38756-7_4
[9] Günal, M. M., and Pidd, M. (2010). Discrete event simulation for performance modelling in health care: a review of the literature. Journal of Simulation, 4(1), 42-51. doi:10.1057/jos.2009.25
[10] Mans, R. S., Schonenberg, M. H., Song, M., van der Aalst, W. M. P., and Bakker, P. J. M. (2008). Application of Process Mining in Healthcare – A Case Study in a Dutch Hospital. In Biomedical Engineering Systems and Technologies, Communications in Computer and Information Science, pages 425-438. Springer Berlin Heidelberg. doi:10.1007/978-3-540-92219-3_32
[11] Nahum-Shani, I., Smith, S. N., Spring, B. J., Collins, L. M., Witkiewitz, K., Tewari, A., and Murphy, S. A. (2018). Just-in-Time Adaptive Interventions (JITAIs) in Mobile Health: Key Components and Design Principles for Ongoing Health Behavior Support. Annals of Behavioral Medicine, 52(6), 446-462. doi:10.1007/s12160-016-9830-8
[12] Niculescu-Mizil, A., and Caruana, R. (2005). Predicting good probabilities with supervised learning. In Proceedings of the 22nd international conference on Machine learning - ICML '05, pages 625-632. ACM Press. doi:10.1145/1102351.1102430
[13] Rajkomar, A., Oren, E., Chen, K., Dai, A. M., Hajaj, N., Hardt, M., Liu, P. J., Liu, X., Marcus, J., Sun, M., et al. (2018). Scalable and accurate deep learning with electronic health records. npj Digital Medicine, 1(1). doi:10.1038/s41746-018-0029-1
[14] Rebuge, Á., and Ferreira, D. R. (2012). Business process analysis in healthcare environments: A methodology based on process mining. Information Systems, 37(2), 99-116. doi:10.1016/j.is.2011.01.003
[15] Shokri, R., Stronati, M., Song, C., and Shmatikov, V. (2017). Membership Inference Attacks Against Machine Learning Models. In 2017 IEEE Symposium on Security and Privacy (SP), pages 3-18. IEEE. doi:10.1109/SP.2017.41
[16] Sonnenberg, F. A., and Beck, J. R. (1993). Markov Models in Medical Decision Making: A Practical Guide. Medical Decision Making, 13(4), 322-338. doi:10.1177/0272989X9301300409
[17] Sutton, R. T., Pincock, D., Baumgart, D. C., Sadowski, D. C., Fedorak, R. N., and Kroeker, K. I. (2020). An overview of clinical decision support systems: benefits, risks, and strategies for success. npj Digital Medicine, 3(1). doi:10.1038/s41746-020-0221-y
[18] Tao, F., Zhang, H., Liu, A., and Nee, A. Y. C. (2019). Digital Twin in Industry: State-of-the-Art. IEEE Transactions on Industrial Informatics, 15(4), 2405-2415. doi:10.1109/TII.2018.2873186
[19] van der Aalst, W. (2016). Process Mining. Springer Berlin Heidelberg. doi:10.1007/978-3-662-49851-4
[20] Venkatesh, K. P., Raza, M. M., and Kvedar, J. C. (2022). Health digital twins as tools for precision medicine: Considerations for computation, implementation, and regulation. npj Digital Medicine, 5(1). doi:10.1038/s41746-022-00694-7.
