Abstract
Background: Computerized Provider Order Entry (CPOE) systems are critical for medication safety, but their effectiveness relies heavily on the completeness of entered data. Incomplete clinical and anthropometric information can disable Clinical Decision Support Systems (CDSSs), compromising patient safety. Objective: This study aimed to assess the longitudinal evolution of CPOE data completeness, specifically focusing on "Breadth Completeness" (the presence of essential clinical variables), and to identify factors predicting data integrity in a tertiary care setting. Methods: A retrospective cross-sectional study was conducted at a 500-bed tertiary referral center in Riyadh. Data were extracted from the Cerner Millennium CPOE system for three "steady-state" years (2015, 2017, and 2019); years involving major system overhauls (2016 and 2018) were excluded to avoid structural bias. A total of 600 unique patient encounters (200 per year) were selected using systematic random sampling from a chronologically ordered sampling frame to minimize temporal bias. The primary outcome was "Breadth Completeness," defined as the presence of eight key variables: age, gender, marital status, weight, height, diagnosis, vital signs, and allergies. Secondary outcomes included documentation consistency (daily notes). Multivariable logistic regression, adjusted for potential confounders, was used to determine predictors of completeness. Results: The rate of primary data completeness (Breadth) improved significantly over the study period, rising from 5.5% in 2015 to 26% in 2017 and 49.5% in 2019. In the multivariable analysis, the year of documentation (OR = 17.47 for 2019 vs. 2015, p < 0.0001) and length of hospitalization (OR = 1.04, p = 0.045) were significant predictors of completeness. Pharmacist-led medication reconciliation was associated with a 2.5-fold increase in data completeness in bivariate analysis (p < 0.0001). Conclusions: While system maturity has driven substantial improvements in CPOE documentation, critical gaps persist, particularly in anthropometric data required for safety alerts. The study underscores the necessity of mandating "hard stops" for core variables and formalizing pharmacist involvement in data reconciliation to ensure patient safety.