RT Journal Article SR Electronic T1 Assessing the quality of care of multiple conditions in general practice: practical and methodological problems JF Quality and Safety in Health Care JO Qual Saf Health Care FD BMJ Publishing Group Ltd SP 421 OP 427 DO 10.1136/qhc.12.6.421 VO 12 IS 6 A1 Kirk, S A A1 Campbell, S M A1 Kennell-Webb, S A1 Reeves, D A1 Roland, M O A1 Marshall, M N YR 2003 UL http://qualitysafety.bmj.com/content/12/6/421.abstract AB Objective: To investigate practical and methodological problems in assessing the quality of care of multiple conditions in general practice. Setting: Sixteen general practices from two socioeconomically diverse regions in the UK. Method: Quality of care was assessed in 100 randomly selected patient records in each practice using an established set of quality indicators covering 23 conditions commonly seen in primary care. Inter-rater reliability assessment was carried out for five of the conditions. Results: Conducting simultaneous quality assessment across multiple conditions is highly resource intensive. Poor data quality and the low prevalence of some items of care defined by the indicators are significant problems. Scores for individual indicators require very large samples for reliable assessment. Quality scores are more reliable when reported at a higher unit of analysis. This is particularly true for indicators and conditions with low prevalence where data may need to be aggregated to the level of groups of conditions or organisational providers. There is no single ideal way of aggregating quality scores. Conclusion: The study identified some of the practical and methodological difficulties in assessing quality of care across multiple conditions. For improved quality assessment, advances in information technology and improvements in data quality are required for more efficient and reliable data extraction from medical records, together with the development of methods for combining scores across indicators, conditions, and practices. However, electronic data extraction methods will still be based on the assumption that the care recorded reflects the care provided.