NIHR SPHR NEE Network seminar: DAGs for PROMs

Background. Health-related patient-reported outcome measures (PROMs) are commonly used to assess the comparative effectiveness of interventions, explore the determinants of population health, monitor the performance of healthcare providers, and guide clinical decision-making at individual and group levels. The design of such measures depends on the latent concept(s) that the PROM is trying to measure, i.e., the PROM’s intended ‘construct’.

Estimating causal effects of an exposure (e.g., health condition or treatment) on a PROM-represented outcome(s) can have complications depending on the relationship between the PROM’s indicators (e.g., the question items) and the measure’s construct(s). Using directed acyclic graphs (DAGs), we show how to represent a PROM’s potential internal causal relationship between its indicators and latent construct(s), then explain the implications when accounting for external variables to estimate causal effects within natural experiments.

Methods. Measurement theory suggests a PROM’s relationships between its items/indicators and latent construct(s) is reflective (construct causes the indicators) or formative (indicators cause the construct). We present DAGs under reflective and formative models when the PROM is unidimensional (e.g., Patient Health Questionnaire-9 [PHQ-9] representing depression severity) or multidimensional (e.g., EQ-5D representing health-related quality-of-life).

Results. Unidimensional PROMs under a reflective model can be analysed like other unidimensional outcomes (e.g., mortality) to estimate causal effects, thus don’t require additional consideration. In comparison, each indicator of a multidimensional ‘composite’ construct under a formative model needs specific consideration to ensure relevant external variables are appropriately conditioned to estimate causal effects.

Conclusion. We show how multidimensional outcome constructs under a formative model increases the complexity of causal analyses. Despite this, multidimensional measures may particularly aid with a variety of ‘outcome-wide’ natural experiments when assessing exposures that may be beneficial for some outcomes but harmful for others. By showing how PROMs can be incorporated into DAGs to inform such causal analyses, we have taken important steps to supporting such outcome-wide studies using PROMs such as via natural experiments.

Full open-access article online: https://doi.org/10.1007/s11136-025-04007-9