Prevalence and Correlates of Respiratory Symptoms and Disease in the Elderly: Statistical Methods

Prevalence and Correlates of Respiratory Symptoms and Disease in the Elderly: Statistical MethodsA standard spirometry system, known to meet or exceed ATS spirometer recommendations, was purchased for all four clinical centers. A water-sealed spirometer (Survey I, Warren E. Collins, Braintree, Mass) was selected so that the results would be comparable with the Atherosclerosis Risk in Communities Study. The spirometer was connected to a personal computer using software (S&M Instruments, Doylestown, Pa) that assists the pulmonary technician with quality control of maneuvers, calculates the pulmonary function (PF) variables, and compresses the results for transmission to the PF reading center. The flow-volume curve and the maneuver duration in seconds were displayed in real-time on the computer monitor. At the end of every FVC maneuver, acceptability and reproducibility checks were applied, and the resulting quality control (QC) messages were displayed. Each participant repeated the FVC maneuver up to eight times or until at least three acceptable and two reproducible FVC maneuvers were obtained, per ATS recommendations. Details of the spirometry methods and resulting normative equations have been previously published.
Variables were grouped into six categories: (1) demographic characteristics (including anthropometric measurements and socioeconomic factors); (2) exposure to cigarette smoking; (3) respiratory symptoms; (4) physician-diagnosed lung diseases; (5) spirometry test results; and (6) coexisting cardiovascular disease. The bivariate relationships among the major outcome variables from groups 3, 4, and 5 were explored by computing a correlation matrix for the continuous variables and by determining the odds ratio for each pair of categorical variables using a computer program (SPSS/PC+, version 4.0) health and care mall.
The continuous spirometry variables of FEVi and FEVi/FVC were transformed into a categorical variable called “obstruction.” Obstruction was set equal to 1 if both FEVi/FVC and FEVi were below the lower limit of the normal range (LLN, fifth percentile cutoffs) as determined by internally generated reference values developed from healthy never smokers; otherwise obstruction was set to zero. This definition of airways obstruction is consistent with the most recent ATS recommendations.

This entry was posted in Pulmonary Function and tagged abnormality rates, elderly, respiratory symptoms.