Overarching Foci

Longitudinal Data Image
Health Utilities Image
Knowledge Translation Image

Individual Projects

Methodology for causal inference is rapidly expanding, however most methods have adopted a frequentist approach. We are investigating Bayesian approaches, via importance sampling and latent classes.

We are expanding existing inverse-intensity weighting and semi-parametric joint model approaches to handle an additional layer of clustering, e.g. due to individuals coming from the same family, school or hospital.

Longitudinal data is useful for describing disease trajectories. To date, focus has been on capturing population trajectories. This project is working on capturing subject specific trajectories through using an initial portion of the follow-up to predict long-term follow-up, and also selecting covariates (from a potentially high-dimensional set) to use in making predictions.

Using prospective longitudinal data gathered through the TargetKids! study, we are studying the relationships among wheezing, air quality and outdoor free play among children aged 2 to 9 years living in the Greater Toronto Area. Particular areas of methodological interest are irregular follow-up, causal inference, and clustering by Toronto neighbourhood.

Discrete choice experiments (DCEs) have been of interest for some time, and have been used to create value sets for some of the EuroQol instruments. Given the resource-intensive nature of the time trade-off (TTO), efficient approaches to valuation using DCEs are appealing. However, current approaches to anchoring the latent utilities derived from a DCE to the cardinal scale may be suboptimal. When DCE is implemented without duration, anchoring is typically achieved by including some TTO tasks. In this project we consider the optimal number of health states to be valued using TTO, whether these health states should be clustered at the severe end of the utility scale or spread evenly, and whether, given a fixed number of respondents and number of tasks per respondent, it is better to have more health states valued by fewer respondents or fewer health states valued by more respondents. Furthermore, we are exploring analytic strategies to improve precision by (a) extending the hybrid model to incorporate non-linear associations between latent and TTO utilities, and (b) incorporating spatial correlation among health states in the hybrid model.

Valuation of the EuroQol instruments is currently underway. The recommended sample size of 1000 has proven infeasible for some studies, and so efficient approaches to analysis that offer the same predictive precision but with a smaller sample size would be helpful. Three non-parametric approaches have been proposed: one frequentist approach two Bayesian. These have led to 20-65% reductions in mean squared error. Sample size calculations using the US EQ-5D-3L suggest that the sample size could have been reduced three-fold when coupled with the non-parametric approach while still achieving better predictive accuracy than the full sample coupled with the traditional analysis.

These three approaches have never been compared to one another on the same dataset. Moreover, evidence that they lead to such substantial improvements in predictive accuracy in a range of datasets would provide more compelling evidence for their widespread use. The purpose of this project is to use EQ-5D-5L data from 9 countries to assess the reductions in sample size that could be achieved by using these alternative modelling strategies, and identify whether one method generally outperforms the others.

Most approaches to longitudinal data subject to irregular observation have been semi-parametric. Joint modeling of the longitudinal outcomes and gap times has been considered, but the models are difficult to specify correctly. However, recent work has suggested that parametric models are more robust to mis-specification than semi-parametric models. Patient charts often contain information about the recommended time to the next visit, and we will build on our previous work to incorporate this into the joint models.