THE RELIABILITY AND VALIDITY OF DOCTOR-RECORDED MORBIDITY DATA IN ACTIVE DATA-COLLECTION SYSTEMS

Objective - To test the reliability and validity of morbidity data rec orded by general practitioners (family physicians) on structured recor ding forms in active data collection systems. Design - The consulting general practitioner recorded the problems managed at 347 video-taped doctor-patient contacts. Two observers independently viewed the video- tapes and completed a recording form for each. Problems were centrally coded, using the International Classification of Primary Care (ICPC). Setting - Primary health care. Main outcome measures - Overall distri bution of the morbidity; positive agreement regarding the morbidity ma naged at matched contacts at three levels of specificity (chapter; cha pter-component; specific rubric); agreement taking negative agreement into account using Cohen's Kappa. Results-Overall distribution of morb idity did not differ between observers. Positive mean agreement was 78 .8% at chapter level, 69.6% at chapter-component, and 55.9% at rubric level. Kappa statistics showed agreement better than chance in all cha pters, but failed to reich significance at rubric level In three chapt ers. Conclusion - Morbidity data actively collected by GPs provide a r eliable overview of morbidity managed in general practice. At a specif ic contact the morbidity recorded is reliable and valid at chapter lev el, and in most cases, at chapter-component level. At rubric level var iance between practitioners in labelling the problem calls into questi on the validity and reliability of the data.