A study of underreporting of Down's syndrome on birth certificates in an Ohio County, 1970-78.

In a previous statewide assessment in Ohio of the percentage reporting of Down's syndrome (DS) on birth certificates, it was found that 33.9 percent of the cases chromosomally analyzed were so recorded. The objectives of this study were to gain a greater understanding of the basis of this low reporting percentage by concentrating on Hamilton County births only, to compare these percentages among hospitals in the county, and to determine the commonality of their causes. Since it was anticipated that both a thorough search of hospital records and access to chromosome analysis records would provide essentially complete ascertainment of DS births in the county during 1970-78, the data were also used to test the validity of chromosome analysis as a sampling procedure in estimating the total number of DS births by means of the Lincoln-Peterson capturerecapture method.The percentage of reporting of DS on the birth certificates by hospital ranged from 0.0 to 57.1, with a mean of 26.0 for the 6 hospitals within the county. Although variation among some hospitals was statistically significant, three of the six hospitals had zero percent reporting. The explanation for these low reporting percentages was fundamentally similar for all hospitals-a communications flow problem, produced by poor timing and use of the wrong sources of available information. The only data typically available to the birth certificate clerk were derived from the labor and delivery sheet, the form with the worst reporting percentage of DS of any in the medical record. On the other hand, by the time of discharge, 92.5 percent of the cases had either been definitely diagnosed as DS and were recorded on the face sheet or were entered as "chromosome analysis pending" in the file.The use of the chromosome analysis data gave an estimate of total DS births (N = 201 +/- 34) statistically consistent with the total observed, thereby providing support for this estimation procedure. Annual incidence rates for DS were also calculated for 1970-78, with a total incidence rate for the 9-year period of 1.14 per 1,000 live births.