Morbidity Reporting in Local Areas: II. The Problem of Measuring the Completeness of Reporting

Abstract

A first step in improving morbidity data is the measurement of the completeness of reporting. Only on the basis of such measurements can the present status of reporting be determined and the effectiveness of measures taken to improve reporting be evaluated. Since the source of most morbidity reports is the local health department, there is particular need for a method which can be used to measure the level of reporting in local health jurisdictions. Two general approaches have been made to the problem in the past: (1) comparison of disease incidence as reported to the health department with that found by sampling and questioning the general population; (2) the use of indices derived from death data and case reports, either as case fatality rates or as the proportion of fatal cases reported before death (1, 2). The first method was used in 1929 and 1930 in connection with attempts to set up a morbidity reporting area for the United States (3). Estimates of the completeness of reporting were also made as by-products of the Hagerstown Morbidity Studies (4, 5) and the National Health Survey (2, 6). Almost 20 years ago, as part of a proposal for setting up a morbidity reporting area, there was some experimentation (in at least three States) with house-to-house surveys of 1 percent of the population to find cases of diphtheria, poliomyelitis, scarlet fever, smallpox, typhoid fever, and tuberculosis. The cases found in these surveys were checked against health department reports to measure the level of reporting. In one State a study covering a sample of 1.7 percent of a total population of 7,000,000 was completed; in a second, a houseto-house survey representing a 1.3 percent population sample was made by public health nurses.