The validity of survey data on utilization of health and social services among the very old.

Survey responses to questions regarding utilization of health and social services among very old persons were compared to independently registered utilization records. Survey data came from in-person interviews with a sample of noninstitutionalized 75-84-year-olds and all noninstitutionalized individuals over the age of 84 in a Swedish community. Agreement regarding utilization vs no utilization during the past three months was found to be very high for hospitalizations and home help, somewhat lower for home visits by nurses, and lower yet for visits to physicians and to nurses. Except for visits to nurses, aggregate survey estimates of proportion with utilization were not significantly different from those based on records. Reporting patterns were generally in agreement with predictions made on the basis of frequency and saliency of service utilization. Reasons for over- and underreporting were investigated on the basis of register data, and some evidence for telescoping was observed.