Assessing the precision of a time-sampling-based study among GPs: balancing sample size and measurement frequency

Area of Science:

• Health workforce research
• Occupational health
• Epidemiology

Background:

• A novel time-sampling SMS method was used to measure Dutch general practitioners' (GPs) working hours.
• Standard power analyses are insufficient for time-sampling studies due to measurement fluctuation.
• Understanding sample size requirements is crucial for health workforce planning, especially for hard-to-reach groups.

Purpose of the Study:

• To investigate the impact of participant numbers and measurement frequency on confidence intervals (CIs) for GP working hours.
• To provide guidance on sample size calculation for time-sampling research.

Main Methods:

• Statistical analysis of time-use data collected via SMS from 1051 GPs.
• Calculation of 95% confidence intervals (CIs) using equations and simulations.
• Evaluation of various participant numbers and measurement frequencies.

Main Results:

• Confidence intervals for weekly working hours decreased significantly with increased numbers of GPs and measurement frequency.
• Precision gains diminished with larger numbers of GPs.
• Reducing measurement intervals (e.g., hourly vs. every 3 hours) substantially decreased the required number of GPs (e.g., 100 vs. 300 GPs for a 1-hour CI).

Conclusions:

• The optimal sample size for time-sampling studies depends on study design, objectives, target population, and budget.
• Both the number of participants and the frequency of measurements per participant critically influence the precision of working hour estimates.
• Balancing these factors is key to efficient and accurate health workforce research.