Related Experiment Video
Updated: Dec 10, 2025

Implementation of a Real-Time Psychosis Risk Detection and Alerting System Based on Electronic Health Records using CogStack
Published on: May 15, 2020
Implementation and Impact of a Risk-Stratified Prostate Cancer Screening Algorithm as a Clinical Decision Support
Anand Shah1, Thomas J Polascik1, Daniel J George1
1Duke University, Durham, NC, USA.
Background:
Implementation methods of risk-stratified cancer screening guidance throughout a health care system remains understudied.
Objective:
Conduct a preliminary analysis of the implementation of a risk-stratified prostate cancer screening algorithm in a single health care system.
Design:
Comparison of men seen pre-implementation (2/1/2016-2/1/2017) vs. post-implementation (2/2/2017-2/21/2018).
Participants:
Men, aged 40-75 years, without a history of prostate cancer, who were seen by a primary care provider.
Interventions:
The algorithm was integrated into two components in the electronic health record (EHR): in Health Maintenance as a personalized screening reminder and in tailored messages to providers that accompanied prostate-specific antigen (PSA) results.
Main Measures:
Primary outcomes: percent of men who met screening algorithm criteria; percent of men with a PSA result. Logistic repeated measures mixed models were used to test for differences in the proportion of individuals that met screening criteria in the pre- and post-implementation periods with age, race, family history, and PSA level included as covariates.
Key Results:
During the pre- and post-implementation periods, 49,053 and 49,980 men, respectively, were seen across 26 clinics (20.6% African American). The proportion of men who met screening algorithm criteria increased from 49.3% (pre-implementation) to 68.0% (post-implementation) (p < 0.001); this increase was observed across all races, age groups, and primary care clinics. Importantly, the percent of men who had a PSA did not change: 55.3% pre-implementation, 55.0% post-implementation. The adjusted odds of meeting algorithm-based screening was 6.5-times higher in the post-implementation period than in the pre-implementation period (95% confidence interval, 5.97 to 7.05).
Conclusions:
In this preliminary analysis, following implementation of an EHR-based algorithm, we observed a rapid change in practice with an increase in screening in higher-risk groups balanced with a decrease in screening in low-risk groups. Future efforts will evaluate costs and downstream outcomes of this strategy.
Related Concept Videos
Cancer Survival Analysis
Imaging Studies for Cardiovascular System VI: Calcium -Scoring CT
Preventive Healthcare Services

