Design strategies for longitudinal spirometry studies: study duration and measurement frequency

Am J Respir Crit Care Med. 2000 Dec;162(6):2134-8. doi: 10.1164/ajrccm.162.6.2003171.


Measuring the longitudinal change in FEV(1) is useful for assessing the adverse effects of respiratory exposures and pulmonary diseases. Investigators seek to estimate the "true" mean FEV(1) slope (mu(beta)) of an infinite population. The difference between the estimated mean FEV(1) slope (mu(beta)) and the true mean slope, resulting from biological variation and measurement errors, can be minimized by increasing the number of subjects (N), years of follow-up (D), or the frequency of measurements (P). We defined maximum error e(max) such that P[|mu(beta) - mu(beta)| < or = e(max)] = 0.95, and thus e(max) is one-half the width of the 95% confidence interval for mu(beta). We computed the values of e(max) on the basis of actual data obtained from 160 coal miners and working nonminers who had completed 11 spirometry measurements, using recommended equipment and procedures, at 6-mo intervals over 5 yr. Individual 5-yr FEV(1) slopes (Delta FEV(1)) were calculated by linear regression. For a range of values of N, D, and P, tables are provided for e(max), the magnitude of detectable differences in Delta FEV(1) between two groups, and the recommended number of subjects needed in each of two groups to reliably detect the anticipated differences in Delta FEV(1). The tables provide unique guidance for investigators in selecting among various study design options.

Publication types

  • Comparative Study
  • Research Support, U.S. Gov't, P.H.S.

MeSH terms

  • Adult
  • Aged
  • Coal Mining
  • Diagnostic Errors / statistics & numerical data
  • Female
  • Forced Expiratory Volume
  • Humans
  • Longitudinal Studies
  • Male
  • Middle Aged
  • Probability
  • Research Design*
  • Spirometry / instrumentation
  • Spirometry / methods*
  • Spirometry / statistics & numerical data
  • Time Factors