Integrated prediction and decision models are valuable in informing personalized decision making
Objectives: To show how prediction models can be incorporated into decision models, to allow for personalized decisions, and to assess the value of this approach using the management of the neck in early-stage oral cavity squamous cell carcinoma as an example. Study Design and Setting: In a decision model, three approaches were compared: a “population-based” approach in which patients undergo the strategy that is optimal for the population; a “perfectly predicted” approach, in which each patient receives the optimal strategy for that specific patient; and a “prediction model” approach in which each patient receives the strategy that is optimal based on prediction models. The average differences in costs and quality-adjusted life years (QALYs) for the population between these approaches were studied. Results: The population-based approach resulted on average in 4.9158 QALYs with €8,675 in costs, per patient. The perfectly predicted approach yielded 0.21 more QALYs and saved €1,024 per patient. The prediction model approach yielded 0.0014 more QALYs and saved €152 per patient compared with the population-based approach. Conclusion: The perfectly predicted approach shows that personalized care is worthwhile. However, current prediction models in the field of oral cavity squamous cell carcinoma have limited value. Incorporating prediction models into decision models appears to be a valuable method to assess the value of personalized decision making.
|Keywords||Cost-effectiveness, Decision modeling, Head and neck neoplasms, Individualized medicine, Lymph nodes, Prediction rule|
|Persistent URL||dx.doi.org/10.1016/j.jclinepi.2018.08.016, hdl.handle.net/1765/110532|
|Journal||Journal of Clinical Epidemiology|
Govers, T.M. (Tim M.), Rovers, M.M, Brands, M.T. (Marieke T.), Dronkers, E.A.C, Baatenburg de Jong, R.J, Merkx, M.A.W, … Grutters, J.P.C. (2018). Integrated prediction and decision models are valuable in informing personalized decision making. Journal of Clinical Epidemiology, 104, 73–83. doi:10.1016/j.jclinepi.2018.08.016