Search

Randomized controlled trials are the gold standard for measuring the causal effects of treatments on clinical outcomes. However, randomized trials are not always feasible, and causal treatment effects must, therefore, often be inferred from observational data. Observational study designs do not allow conclusions about causal relationships to be drawn unless statistical techniques are used to account for the imbalance of confounders across groups while key assumptions hold. Propensity score (PS) and balance weighting are two useful techniques that aim to reduce the imbalances between treatment groups by weighting the groups to look alike on the observed confounders. There are many methods available to estimate PS and balancing weights. However, it is unclear a priori which will achieve the best trade-off between covariate balance and effective sample size. Weighted analyses are further complicated by small studies with limited sample sizes, which is common when studying rare diseases. To address these issues, we present a step-by- step guide to covariate balancing strategies, including how to evaluate overlap, obtain estimates of PS and balancing weights, check for covariate balance, and assess sensitivity to unobserved confounding. We compare the performance of a number of commonly used estimation methods on a synthetic data set based on the Physical Activity and Exercise Outcomes in Huntington Disease (PACE-HD) study, which explored whether enhanced physical activity affects the progression and severity of the disease. We provide general guidelines for the choice of method for estimation of PS and balancing weights, interpretation, and sensitivity analysis of results.

Background The impact of modifiable lifestyle factors on Huntington’s Disease (HD) is an emerging area of interest in the HD community. DOMINO-HD is the first study of its kind to integrate prospective longitudinal physical activity, sleep and nutrition metrics to explore the interplay between multi-domain environmental factors and HD outcomes. Aims Inspired by the Lifestyle for Brain Health (LIBRA) index aimed at quantifying a person’s risk of developing dementia based on modifiable risk factors, we wanted to develop a similar Lifestyle Index (LifeHD) which aggregates information from a range of lifestyle factors (including physical activity, sleep, diet, and body mass index) alongside non-modifiable characteristics such as genetic information (CAG repeat length), age, and HD stage. Methods Continuous variables are first transformed to categorical variables using recommended thresholds. Regression analysis was used to identify the range of values for each variable less likely to be linked to disease progression, with these ranges contributing zero points to the LifeHD score. These categorical variables are all included in a regression model, and the coefficients are used as characteristic-specific scores. An individual’s LifeHD score is the sum of the coefficients that are relevant to its characteristics. Higher scores indicate more favourable lifestyle habits and more favourable non-modifiable characteristics. This methodology could be extended to include more characteristics. Outcome The LifeHD association with cUHDRS progression could become a useful tool for studying the progression and severity of HD. The relationship of LifeHD score with cUHDRS (or other severity measures) would require validation on independent sample.

Randomised controlled trials (RCTs) are considered the gold standard to estimate the effect of interventions on the progression of diseases. Observational studies are gaining traction to study the effect of modifiable factors on the progression and severity of rare diseases, as they are often easier and cheaper to implement than RCTs. Observational studies do not require tailored interventions; rather, they involve observing the behaviour of the participants and monitoring their progression. However, making inferences using data from observational studies can be challenging for several reasons, including imbalances in confounders between the control and treatment groups. In this thesis, I use observational data in an attempt to identify risk factors for Huntington’s disease (HD) severity and progression. HD is an inherited disorder that results in the death of brain cells and typically leads to death 15-20 years after clinical diagnosis. It is a rare disease (prevalence 4-8 per 100,000). I began by proposing a set of steps to make inferences using data from observational studies, controlling for confounding bias using Propensity Score and Balancing Weights (PSBW), and developing a web application (CoBWeb), which implements these steps in a user-friendly environment. Next, I used a well established simulation study to understand how the limited sample size (typical for rare diseases), the choice of baseline covariates for balancing could affect the ability to control for confounding bias. Finally, I applied these methods to the Enroll-HD dataset to investigate the effect of several modifiable lifestyle factors (e.g., use of antidepressant medication) on the progression of HD. Analysis of Enroll-HD data, using the proposed methodology, identified a potentially harmful effect of antidepressant medication on the progression of HD among those in the early stage of the disease.

Randomised controlled trials (RCTs) are considered the gold standard to estimate the effect of interventions on the progression of diseases. Observational studies are gaining traction to study the effect of modifiable factors on the progression and severity of rare diseases, as they are often easier and cheaper to implement than RCTs. Observational studies do not require tailored interventions; rather, they involve observing the behaviour of the participants and monitoring their progression. However, making inferences using data from observational studies can be challenging for several reasons, including imbalances in confounders between the control and treatment groups. In this thesis, I use observational data in an attempt to identify risk factors for Huntington’s disease (HD) severity and progression. HD is an inherited disorder that results in the death of brain cells and typically leads to death 15-20 years after clinical diagnosis. It is a rare disease (prevalence 4-8 per 100,000). I began by proposing a set of steps to make inferences using data from observational studies, controlling for confounding bias using Propensity Score and Balancing Weights (PSBW), and developing a web application (CoBWeb), which implements these steps in a user-friendly environment. Next, I used a well established simulation study to understand how the limited sample size (typical for rare diseases), the choice of baseline covariates for balancing could affect the ability to control for confounding bias. Finally, I applied these methods to the Enroll-HD dataset to investigate the effect of several modifiable lifestyle factors (e.g., use of antidepressant medication) on the progression of HD. Analysis of Enroll-HD data, using the proposed methodology, identified a potentially harmful effect of antidepressant medication on the progression of HD among those in the early stage of the disease.