Menu
Close
.svg)
We claim to prioritize patient experience when we use measures of health-related quality of life (HRQoL) in clinical research, but the "how" matters. Mere use of HRQoL as a proxy or snapshot of a patient's clinical state continues to elevate physicians' priorities—or, at least, what we assume those to be.
This review of the use of predictive modeling in oncology offers a map of where the field could go next:
Far more interesting to me are studies where HRQoL is studied as one of the outcomes to be predicted, valuable in its own right. Yet, such approaches are still a minority (only 21% of the papers reviewed took this approach, while 23% others included it as both outcome and predictor).
Even then, there's a catch: HRQoL studied as an outcome still leans towards physical symptoms and functioning, sidelining the "softer" dimensions of quality of life such as cognitive, emotional, or social functioning.
The skewed focus on physical functioning and symptoms scales as predictors and outcomes reiterates the need to incorporate other HRQoL domains (emotional, role, cognitive functioning) in model development. This is important, as prediction with and of HRQoL could be a tool where patients and clinicians get insight in what can happen during treatment and disease under different treatment choices.
As value-based payment takes root, and shared decision making between patients and physicians becomes the norm, modeling approaches will need to reflect these shifts. And this includes not just coming up with novel patient-reported outcomes or quality of life measures, —but rethinking what it is that we are trying to model in the first place.
van der Heijden, T. G. W., K. M. de Ligt, N. J. Hubel, S. van der Mierden, B. Holzner, L. V. van de Poll-Franse, and B. H. de Rooij. "Exploring the role of health-related quality of life measures in predictive modelling for oncology: a systematic review." Quality of Life Research (2024): 1-19.
