EHR data as RWD sources: common problems & solutions
Electronic health records (EHR) are one of the most significant sources for Real World Data (RWD) in healthcare. EHR data have several advantages including that there are already collected, thus not extra workload for healthcare professionals, the large sample size, while also they reflect in many ways the real practice and clinical routine. However, it could be easily understood that there are common challenges and problems that may hinder the full utilization of these data.
In a recent Viewpoint article by Sauer et al., published in the Lancet Digital Health recently, the authors discuss the key pitfalls that should be taken into consideration when working with EHR data. Six common pitfalls that should be avoided are discussed: sample selection bias, imprecise variable definitions, limitations to deployment, variable measurement frequency, subjective treatment allocation, and model overfitting. Further, the authors propose a solution to further improve the EHR analysis methodology.
The full article is available here.
In a recent Viewpoint article by Sauer et al., published in the Lancet Digital Health recently, the authors discuss the key pitfalls that should be taken into consideration when working with EHR data. Six common pitfalls that should be avoided are discussed: sample selection bias, imprecise variable definitions, limitations to deployment, variable measurement frequency, subjective treatment allocation, and model overfitting. Further, the authors propose a solution to further improve the EHR analysis methodology.
The full article is available here.