Associate Professor King Khalid University Abha, Asir, Saudi Arabia
Background: Pharmacoepidemiological studies are crucial for understanding prescribing trends and treatment outcomes in Parkinson’s disease (PD). However, biases within these studies can undermine the reliability of evidence and impact clinical decision-making.
Objectives: This systematic review aims to identify and evaluate reported biases in pharmacoepidemiological studies on nonrandomized prescribing patterns, drug utilization, or treatment trends in PD.
Methods: A systematic search was conducted in PubMed, Embase, and Web of Science for pharmacoepidemiological studies published from the outset until November 30, 2024. The search focused on studies investigating nonrandomized prescribing patterns, drug utilization, or treatment trends in PD. Inclusion criteria required studies that explicitly reported or demonstrated biases. Types of bias analyzed included confounding, selection bias, and information bias.
Results: Of 63 studies identified, 17 met the inclusion criteria. Common biases included confounding due to disease severity or comorbidities (present in 58.8% of studies) and selection bias from non-representative patient populations (41.1%). Recall bias, as a form of information bias, was reported in 17.6% of studies. Geographic disparities in prescribing practices were also observed, potentially reflecting systemic biases. Few studies (23.5%) adequately addressed methodological limitations, such as appropriate adjustment for confounders or stratification by disease progression. Only one study (5.8%) used the ROBINS-I tool to assess and report biases.
Conclusions: Although 63 studies were identified, only about 27% explicitly reported some form of methodological bias. Considering that these types of studies are inherently prone to biases, this suggests a lack of awareness among researchers regarding the identification and reporting of biases in pharmacoepidemiological studies on PD. Improved study designs and heightened researcher awareness are necessary to enhance the validity and reliability of evidence in this field.
