Routinely-collected general practice data are complex, but with systematic processing can be used for quality improvement and research

From Clinfowiki
Jump to: navigation, search

The primary care setting is an information rich environment. Focusing on the use of primary care data for improving chronic disease management, health services research, and health system planning while encouraging greater transparency among informaticians and epidemiologists, the authors describe their method for using primary care data which incorporates a system for ensuring traceability, a design which facilitates consistent approaches to planning and extraction, and a processing system with quality controls at each step. Methods for ensuring traceability focus on identifying the source of origin of each data element including storage of queries to libraries, look up tables for coding and decoding data, and unique identifiers for each practice, patient, and primary care organization. Design methods prioritize identification of an appropriate data set, development of robust audit criteria, and development of an analysis plan that precedes data collection. Key aspects of their processing system include appraising data entry issues so that bias is minimized, extracting data using software that ideally is compatible with the “national Spine”, migrating and integrating data using a valid metadata system that ensures consistent meaning across similarly sets of aggregate data from different sources, cleaning data for correction of inappropriate values or inconsistent units of measurement, processing data using the software package SPSS, and analyzing the data with an initial step focused on ensuring an appropriate denominator. This process culminates in feed back to the end users (caregivers, researchers, and health system planners) to guide their efforts toward improving the health status of the population. The authors note challenges related to frequent patient transitions between primary care organizations obscuring the denominator and contributing to data loss, limited information on ethnicity and social class, and inaccuracies associated with data entry.

Atrial fibrillation, a condition diagnosed objectively (EKG) and treated based on longstanding evidence (anticoagulation) is less vulnerable to the latter challenge than bronchitis with a less robust evidence base and less consistently obtained objective supporting data. Lack of a ubiquitous, unchanging evidence base and variable use of coding creating inaccurate diagnoses are both cited by the authors as challenges and well illustrated by the examples provided.

Comment: Written in a setting where electronic health records are ubiquitous, this article underscores the need for a consistent approach to data analysis that promotes accuracy and validity of data by creating transparency at all levels. Approaching data collection and analysis in this manner is likely to contribute to unanticipated benefits and insights into critical elements of quality improvement. A major limitation to effectiveness is the lack of collaboration between different primary care organizations and an apparent limitation in interoperability between systems. The extent to which the health care system can overcome these inter-organizational barriers to data sharing will probably dictate the value that can be derived from the richness of primary care data.

--Elmer Washington