Implementing high-dimensional propensity score principles to improve confounder adjustment in UK electronic health records

Evaluating the Role of High-Dimensional Proxy Data in Confounding Adjustment in Multiple Sclerosis Research: A Case Study

PURPOSE

Given the historical use of limited confounders in multiple sclerosis (MS) studies utilizing administrative health data, this brief report evaluates the impact of incorporating high-dimensional proxy information on confounder adjustment in MS research. We have implemented high-dimensional propensity score (hdPS) and high-dimensional disease risk score (hdDRS) methods to assess changes in effect estimates for the association between disease-modifying drugs (DMDs) and all-cause mortality in an MS cohort from British Columbia (BC), Canada.

METHODS

We conducted a population-based retrospective study using linked administrative databases from BC, including health insurance registries, demographics, physician visits, hospitalizations, prescriptions, and vital statistics. The cohort comprised 19 360 individuals with MS, followed from January 1, 1996, to December 31, 2017. DMD exposure was defined as at least 180 days of use for beta-interferon or glatiramer acetate, or at least 90 days for other DMDs. The outcome was time to all-cause mortality. We compared Cox proportional hazards models adjusting for investigator-specified covariates with those incorporating additional empirical covariates using hdPS and hdDRS methods.

RESULTS

In the unadjusted analysis, DMD exposure was associated with a 69% lower risk of mortality (HR 0.31; 95% CI: 0.27-0.36). Adjusting for investigator-specified covariates, the adjusted hazard ratio (aHR) was 0.76 (95% CI: 0.65-0.89). HdPS analyses showed a 20%-23% lower mortality risk (aHRs: 0.77 to 0.80), while hdDRS analyses indicated a 19%-21% reduction (aHRs: 0.79 to 0.81).

CONCLUSIONS

Incorporating high-dimensional proxy information resulted in minor variations in effect estimates compared to traditional covariate adjustment. These findings suggest that the impact of residual confounding in the question under consideration may be modest. Further research should explore additional data dimensions and replicate these findings across different datasets.

Risk of mortality between warfarin and direct oral anticoagulants: population-based cohort studies

BACKGROUND

Direct oral anticoagulants (DOACs) have been reported to be associated with a higher risk of mortality compared with an older alternative, warfarin using primary care data in the United Kingdom (UK). However, other studies observed contradictory findings. We therefore aimed to investigate the association between mortality and warfarin, compared with DOACs.

METHODS

We conducted cohort studies using UK Clinical Practice Research Datalink (CPRD) Aurum and Hong Kong Clinical Data Analysis and Reporting System (CDARS) to identify the association between warfarin and hazard of mortality, compared to DOACs. Individuals with non-valvular atrial fibrillation aged ≥ 18 years who had first anticoagulant therapy (warfarin or DOAC) during 1/1/2011-31/12/2019 were included.

RESULTS

Compared with DOAC use, a lower hazard of all-cause mortality was found in warfarin users (hazard ratio (HR) = 0.81, 95% confidence interval (CI) = 0.77-0.86) in CPRD; while a higher hazard was observed in warfarin users (HR = 1.31, 95% CI = 1.24-1.39) in CDARS, versus DOAC users. In our exploratory analysis, consistent results were seen in both databases when stratified warfarin users by time in therapeutic range (TTR) using post-baseline measurements: a lower hazard of all-cause mortality in warfarin users with TTR ≥ 65% (CPRD: HR = 0.68, 95% CI = 0.65-0.72; CDARS: HR = 0.86, 95% CI = 0.77-0.96) and increased hazard in warfarin users with TTR < 65% (CPRD: HR = 1.14, 95% CI = 1.05-1.23; CDARS: HR = 1.59, 95% CI = 1.50-1.69), versus DOAC users.

CONCLUSIONS

The differences in hazard of all-cause mortality associated with warfarin compared with DOAC, in part may depend on anticoagulation control in warfarin users. Notably, this study is unable to establish a causal relationship between warfarin and mortality stratified by TTR, versus DOACs, requiring future studies for further investigation.

Application of High-Dimensional Propensity Score Methods to the National Health and Aging Trends Study

BACKGROUND

High-dimensional propensity scoring (HDPS) is a method for empirically identifying potential confounders within large healthcare databases such as administrative claims data. However, this method has not yet been applied to large national health surveys such as the National Health and Aging Trends Study (NHATS), an ongoing nationally representative survey of older adults in the United States and important resource in gerontology research.

METHODS

In this Research Practice article, we present an overview of HDPS and describe the specific data transformation steps and analytic considerations needed to apply it to national health surveys. We applied HDPS within NHATS to investigate the association between self-reported visual difficulty and incident dementia, comparing HDPS to conventional confounder selection methods.

RESULTS

Among 7 207 dementia-free NHATS Wave 1 respondents, 528 (7.3%) had self-reported visual difficulty. In an unadjusted discrete time proportional hazards model accounting for the complex survey design of NHATS, self-reported visual difficulty was strongly associated with incident dementia (odds ratio [OR] 2.34, 95% confidence interval [CI]: 1.95-2.81). After adjustment for standard investigator-selected covariates via inverse probability weighting, the magnitude of this association decreased, but evidence of an association remained (OR 1.44, 95% CI: 1.11-1.85). Adding 75 HDPS-prioritized variables to the investigator-selected propensity score model resulted in further attenuation of the association between visual impairment and dementia (OR 0.94, 95% CI: 0.70-1.23).

CONCLUSIONS

HDPS can be successfully applied to national health surveys such as NHATS and may improve confounder adjustment. We hope developing this framework will encourage future consideration of HDPS in this setting.

A methodological review of the high-dimensional propensity score in comparative-effectiveness and safety-of-interventions research finds incomplete reporting relative to algorithm development and robustness

OBJECTIVES

The use of secondary databases has become popular for evaluating the effectiveness and safety of interventions in real-life settings. However, the absence of important confounders in these databases is challenging. To address this issue, the high-dimensional propensity score (hdPS) algorithm was developed in 2009. This algorithm uses proxy variables for mitigating confounding by combining information available across several healthcare dimensions. This study assessed the methodology and reporting of the hdPS in comparative effectiveness and safety research.

STUDY DESIGN AND SETTING

In this methodological review, we searched PubMed and Google Scholar from July 2009 to May 2022 for studies that used the hdPS for evaluating the effectiveness or safety of healthcare interventions. Two reviewers independently extracted study characteristics and assessed how the hdPS was applied and reported. Risk of bias was evaluated with the Risk Of Bias In Non-randomised Studies - of Interventions (ROBINS-I) tool.

RESULTS

In total, 136 studies met the inclusion criteria; the median publication year was 2018 (Q1-Q3 2016-2020). The studies included 192 datasets, mostly North American databases (n = 132, 69%). The hdPS was used in primary analysis in 120 studies (88%). Dimensions were defined in 101 studies (74%), with a median of 5 (Q1-Q3 4-6) dimensions included. A median of 500 (Q1-Q3 200-500) empirically identified covariates were selected. Regarding hdPS reporting, only 11 studies (8%) reported all recommended items. Most studies (n = 81, 60%) had a moderate overall risk of bias.

CONCLUSION

There is room for improvement in the reporting of hdPS studies, especially regarding the transparency of methodological choices that underpin the construction of the hdPS.

Evaluating the Performance of High-Dimensional Propensity Scores Compared with Standard Propensity Scores for Comparing Antihypertensive Therapies in the CPRD GOLD Database

INTRODUCTION

Propensity score (PS) matching is widely used in medical record studies to create balanced treatment groups, but relies on prior knowledge of confounding factors. High-dimensional PS (hdPS) is a semi-automated algorithm that selects variables with the highest potential for confounding from medical databases. The objective of this study was to evaluate performance of hdPS and PS when used to compare antihypertensive therapies in the UK clinical practice research datalink (CPRD) GOLD database.

METHODS

Patients initiating antihypertensive treatment with either monotherapy or bitherapy were extracted from the CPRD GOLD database. Simulated datasets were generated using plasmode simulations with a marginal hazard ratio (HRm) of 1.29 for bitherapy versus monotherapy for reaching blood pressure control at 3 months. Either 16 or 36 known covariates were forced into the PS and hdPS models, and 200 additional variables were automatically selected for hdPS. Sensitivity analyses were conducted to assess the impact of removing known confounders from the database on hdPS performance.

RESULTS

With 36 known covariates, the estimated HRm (RMSE) was 1.31 (0.05) for hdPS and 1.30 (0.04) for PS matching; the crude HR was 0.68 (0.61). Using 16 known covariates, the estimated HRm (RMSE) was 1.23 (0.10) and 1.09 (0.20) for hdPS and PS, respectively. Performance of hdPS was not compromised when known confounders were removed from the database.

RESULTS ON REAL DATA

With 49 investigator-selected covariates, the HR was 1.18 (95% CI 1.10; 1.26) for PS and 1.33 (95% CI 1.22; 1.46) for hdPS. Both methods yielded the same conclusion, suggesting superiority of bitherapy over monotherapy for time to blood pressure control.

CONCLUSION

HdPS can identify proxies for missing confounders, thereby having an advantage over PS in case of unobserved covariates. Both PS and hdPS showed superiority of bitherapy over monotherapy for reaching blood pressure control.

High-dimensional propensity scores for empirical covariate selection in secondary database studies: Planning, implementation, and reporting

Real-world evidence used for regulatory, payer, and clinical decision-making requires principled epidemiology in design and analysis, applying methods to minimize confounding given the lack of randomization. One technique to deal with potential confounding is propensity score (PS) analysis, which allows for the adjustment for measured preexposure covariates. Since its first publication in 2009, the high-dimensional propensity score (hdPS) method has emerged as an approach that extends traditional PS covariate selection to include large numbers of covariates that may reduce confounding bias in the analysis of healthcare databases. hdPS is an automated, data-driven analytic approach for covariate selection that empirically identifies preexposure variables and proxies to include in the PS model. This article provides an overview of the hdPS approach and recommendations on the planning, implementation, and reporting of hdPS used for causal treatment-effect estimations in longitudinal healthcare databases. We supply a checklist with key considerations as a supportive decision tool to aid investigators in the implementation and transparent reporting of hdPS techniques, and to aid decision-makers unfamiliar with hdPS in the understanding and interpretation of studies employing this approach. This article is endorsed by the International Society for Pharmacoepidemiology.

Transparency of high-dimensional propensity score analyses: guidance for diagnostics and reporting

Purpose

The high‐dimensional propensity score (HDPS) is a semi‐automated procedure for confounder identification, prioritisation and adjustment in large healthcare databases that requires investigators to specify data dimensions, prioritisation strategy and tuning parameters. In practice, reporting of these decisions is inconsistent and this can undermine the transparency, and reproducibility of results obtained. We illustrate reporting tools, graphical displays and sensitivity analyses to increase transparency and facilitate evaluation of the robustness of analyses involving HDPS.

Methods

Using a study from the UK Clinical Practice Research Datalink that implemented HDPS we demonstrate the application of the proposed recommendations.

Results

We identify seven considerations surrounding the implementation of HDPS, such as the identification of data dimensions, method for code prioritisation and number of variables selected. Graphical diagnostic tools include assessing the balance of key confounders before and after adjusting for empirically selected HDPS covariates and the identification of potentially influential covariates. Sensitivity analyses include varying the number of covariates selected and assessing the impact of covariates behaving empirically as instrumental variables. In our example, results were robust to both the number of covariates selected and the inclusion of potentially influential covariates. Furthermore, our HDPS models achieved good balance in key confounders.

Conclusions

The data‐adaptive approach of HDPS and the resulting benefits have led to its popularity as a method for confounder adjustment in pharmacoepidemiological studies. Reporting of HDPS analyses in practice may be improved by the considerations and tools proposed here to increase the transparency and reproducibility of study results.