OBJECTIVES: To develop methods for prospective registration and analysis of episodes of care (one or more contacts about the same health problem). To compare estimates of point prevalence and incidence of hypertension among a group of elderly patients by means of an elaborate coding of episodes and a new method based on routine coding of contact diagnoses. DESIGN: Prospective longitudinal cohort study. SETTING: General practice in Denmark. SUBJECTS: 31 GPs and 1722 patients 70+ years of age. MAIN OUTCOME MEASURES: During the course one year, data were registered in the doctors' electronic medical record systems. Registration included ICPC diagnoses and codes for diagnostic and therapeutic procedures. Individual contacts were linked into episodes of care by the recording physician. Prevalence and incidence were calculated from the GPs' registration of episodes of care and from contact registration by means of a new method, "the waiting time distribution". RESULTS: Estimates of incidence (2.1 per 100 patient years) and point prevalence (21.2 per 100 patients) of episodes of hypertension care could be calculated when the GPs actively linked contacts into episodes. Based on simple contact registration it was possible to calculate similar estimates. CONCLUSION: A full registration of episodes of care is time-consuming and complicated. Incidence and point prevalence of chronic conditions, however, may be estimated from simple coding of contact diagnoses.
OBJECTIVES: To develop methods for prospective registration and analysis of episodes of care (one or more contacts about the same health problem). To compare estimates of point prevalence and incidence of hypertension among a group of elderly patients by means of an elaborate coding of episodes and a new method based on routine coding of contact diagnoses. DESIGN: Prospective longitudinal cohort study. SETTING: General practice in Denmark. SUBJECTS: 31 GPs and 1722 patients 70+ years of age. MAIN OUTCOME MEASURES: During the course one year, data were registered in the doctors' electronic medical record systems. Registration included ICPC diagnoses and codes for diagnostic and therapeutic procedures. Individual contacts were linked into episodes of care by the recording physician. Prevalence and incidence were calculated from the GPs' registration of episodes of care and from contact registration by means of a new method, "the waiting time distribution". RESULTS: Estimates of incidence (2.1 per 100 patient years) and point prevalence (21.2 per 100 patients) of episodes of hypertension care could be calculated when the GPs actively linked contacts into episodes. Based on simple contact registration it was possible to calculate similar estimates. CONCLUSION: A full registration of episodes of care is time-consuming and complicated. Incidence and point prevalence of chronic conditions, however, may be estimated from simple coding of contact diagnoses.