Cut off values for Oays hourly plots
|
We are using the OASYS program and want to use the reported areas between the graphs for the time of day plots and the time from waking graphs in our interpration of the record. What are the cut-off values that you have established for evidence of a work-related pattern. Do you have a reference I can use for this?
|
This is more difficult than it seems.
There is a diurnal variation in PEF which usually has its nadir on waking. The peak (acrophase) is very variable, anything from 1-16 hours later. The acrophase is affected by the timing of work exposure related to waking time and changes in sleep pattern. The development versions of Oasys calculate the mean value of PEF over each 2 hour period separately for each shift pattern and restdays. The program also calculates the mean values in hours from waking as well as clock time.
At present we believe that the best method of interpretation is by expert visual interpretation. To validate a cut off value we need to have workers whose diagnosis is validated independently of the peak flow record, and to have workers providing quality records for morning, afternoon and night shifts. We are seeking funding to do this.
We have done some preliminary analysis of day shift workers, comparing day shifts with workdays, in workers where there is little difference in waking times on days at and away from work. We have only done the first stage of validation by constructing receiver/operator curves and selecting a cut off (about 170)(ROC Curves for area under the curve analysis of hourly peak flows. V.J.Drought, V.Huggins, S.Burge. Eur Respir J 2004;24 Suppl 48:368s). This value however depends on the length of the record, unless the overnight period is to be included. Since there are no overnight readings, and since the greatest difference between the "days" is often overnight, this is not calculated reliably from the available readings.
We also do not know the minimum data quality required for calculating the difference between the hourly plots. The data quality required for the analysis of maximum, mean and minimum PEF is more dependant on the length of the record and the number of consecutive workdays than on the number of readings/day (Effect of peak expiratory flow data quality on diagnostic sensitivity and specificity in occupational asthma. W.Anees, P.F.Gannon, V.Huggins, C.F.A.Pantin, P.S.Burge. Eur Respir J 23:730-734). These quality standards are unlikely to apply to the hourly plot, where we suppose that 2-houly readings are going to be important, and the length of the record less important. (We are also looking for money to do this, we have the data from which to do the testing).
All this is probably not much help. We would however be delighted to look at your records and give an independant opinion (for nothing), and work with you if you have data from workers with independently validated diagnoses (or would like to fund the work with us)
|
Please sign in or register to add your thoughts.