PubMed2001
OBJECTIVE: To calculate the concordance of the computer record and the clinical history (CH) in preventive actions and health problems. DESIGN: Cross-sectional descriptive study. Quality evaluation. SETTING: Urban health centre with 31000 inhabitants. PATIENTS AND OTHER PARTICIPANTS: Randomised batch sample, with 14 cases for each of the 8 attendance base units with computerised records since 1997. N = 112. EXCLUSION CRITERIA: no visit later than January 1997 and absence of records in the CH or computer. MEASUREMENTS AND MAIN RESULTS: Through the checking of the records in the CH and computer, a mean concordance of 73.5 (95% CI, 66.8-80.2) for preventive actions and 93.5 for health problems (95% CI, 90.6-96.4) was found. There was a mean computer under-recording for health problems of 6.5% (95% CI, 3.62-9.32), and for preventive actions of 21% (95% CI, 9.1-33.3) only in those actions based on manual activity. However, in preventive actions based on verbal activity there was 14.3% mean CH under-recording (95% CI, 1.15-27.5). CONCLUSIONS: Concordance is not uniform, with under-recording for some parameters detected. This may affect the reliability and validity of health information in these records. We believe that the way data are collected determines this to a large extent. We suggest as corrective measures improving the training and incentives of health professionals, making computer programmes more appropriate to their purpose and standardising data collection in primary care CR.