The use of primary care databases: case-control and case-only designs

Area of Science:

• Epidemiology
• Primary Care Research
• Biostatistics

Background:

• Epidemiological studies often rely on case identification, traditionally from hospital records.
• Utilizing computerized clinical data from primary care presents unique opportunities and challenges for case identification.
• Limited information on confounding variables in primary care databases is a significant issue.

Purpose of the Study:

• To discuss case identification study designs applicable to primary care research data.
• To explore the application and principles of case-only designs (case-crossover, within-person case-series) in primary care research.
• To address the challenge of confounding variables when using primary care data.

Main Methods:

• Review and discussion of traditional case-control study designs.
• Detailed explanation of case-only designs: case-crossover and within-person case-series.
• Focus on the identification of cases and selection of controls within primary care contexts.

Main Results:

• Case-only designs are presented as alternatives to overcome confounding in primary care research.
• Principles and examples of case-crossover and within-person case-series designs are illustrated.
• Comparison of advantages and limitations of different case identification designs is provided.

Conclusions:

• Case-only designs offer valuable alternatives for epidemiological research using primary care data, particularly when confounding is a concern.
• Understanding the nuances of case identification and control selection is crucial for robust primary care research.
• These designs enhance the utility of computerized clinical data for epidemiological investigations.