Related Experiment Video
Updated: May 31, 2026

The Participant-Reported Implementation Update and Score (PRIUS): A Novel Method for Capturing Implementation-Related Data Over Time
Published on: February 19, 2021
Temporal and within practice variability in the health improvement network.
Kevin Haynes1, Warren B Bilker, Tom R Tenhave
1Department of Biostatistics and Epidemiology, University of Pennsylvania School of Medicine, Philadelphia, PA, USA. khaynes@mail.med.upenn.edu
The Health Improvement Network (THIN) database shows yearly trends and practice variations in mortality, cancer incidence, and prescribing. These factors are crucial for designing pharmacoepidemiologic studies using electronic health records.
Area of Science:
- Epidemiology
- Health Informatics
- Pharmacoepidemiology
Background:
- The Health Improvement Network (THIN) is a UK primary care electronic medical record database used for pharmacoepidemiologic research.
- Standard methods like matching on practice and calendar year are used to control for temporal and practice-specific differences.
- Limited understanding exists regarding the consistency of data within practices over time and among different practices.
Purpose of the Study:
- To assess the consistency of THIN data within practices across years and among practices within a year.
- To quantify secular trends and practice variations in key health metrics.
- To inform the design of pharmacoepidemiologic studies utilizing large electronic health record databases.
Main Methods:
- Analysis of mortality, cancer incidence, prescribing, and encounter rates from 2000-2007 using 415 UK practices.
- Longitudinal Poisson regression models (random and fixed effects) with practice-year as the unit of observation.
- Adjusted models incorporated practice-level characteristics like smoking, obesity, age, and software experience.
Main Results:
- Significant linear trends (p < 0.001) were observed: decreasing mortality, increasing cancer reporting, increasing prescriptions per patient, and decreasing encounters per patient over calendar years.
- In 2007, the interquartile range (75th to 25th percentile ratio) for crude rates was 1.63 for cancer and mortality, 1.45 for prescriptions, and 1.42 for encounters.
- Accounting for practice characteristics reduced the variation among practices by approximately 40%.
Conclusions:
- THIN data exhibit significant secular trends and substantial variation among practices.
- These temporal and practice-specific variations must be considered when designing pharmacoepidemiologic studies.
- The study could not definitively distinguish whether observed trends reflect data quality changes or true secular health trends.
Related Concept Videos
Health Information Technology and Healthcare Information System
Health Information Technology, commonly called HIT, integrates advanced information systems and technology in healthcare settings. Its primary functions include:
Current Trends in Nursing II
Variability: Analysis
The range is a simple measure of variability, indicating the difference between the highest and...
Regression Toward the Mean
Healthcare Associated Infections II: Preventive Measures
The best practices for preventing healthcare-associated infections include hand hygiene, patient risk...
Dosage Regimen: Individualization