Feasibility of automatic evaluation of clinical rules in general practice. (April 2017)

• Record Type: Journal Article
• Authors: Opondo, Dedan; Visscher, Stefan; Eslami, Saied; Medlock, Stephanie; Verheij, Robert; Korevaar, Joke C.; Hanna, Ameen Abu
• Is Part Of: International journal of medical informatics. Volume 100(2017)
• Page Start: 90
• Page End: 94
• Publication Date: 2017-04
• Subjects: Assessing Care of Vulnerable Elders (ACOVE) -- Clinical decision support systems -- General practice -- Vulnerable elders -- Clinical rules -- Electronic health records; Medical informatics -- Periodicals; Information science -- Periodicals; Computers -- Periodicals; Medical technology -- Periodicals; Medical Informatics -- Periodicals; Technology, Medical -- Periodicals; Computers; Information science; Medical informatics; Medical technology; Electronic journals; Periodicals; Electronic journals; 610.285
• DOI: 10.1016/j.ijmedinf.2017.01.010
• Languages: English
• ISSNs: 1386-5056

Abstract

Highlights: 40% of ACOVE indicators can be automatically evaluated in Dutch general practice. Sub-optimal statement of indicators impedes automatic evaluation of healthcare. Modification of GP databases is required to optimize evaluation of healthcare. LERM is feasible method for analyzing clinical rules in quality evaluation. Abstract: Purpose: To assess the extent to which clinical rules (CRs) can be implemented for automatic evaluation of quality of care in general practice. Methods: We assessed 81 clinical rules (CRs) adapted from a subset of Assessing Care of Vulnerable Elders (ACOVE) clinical rules, against Dutch College of General Practitioners (NHG) data model. Each CR was analyzed using the Logical Elements Rule Method: (LERM). LERM is a stepwise method of assessing and formalizing clinical rules for decision support. Clinical rules that satisfied the criteria outlined in the LERM method were judged to be implementable in automatic evaluation in general practice. Results: Thirty-three out of 81 (40.7%) Dutch-translated ACOVE clinical rules can be automatically evaluated in electronic medical record systems. Seven out of 7 CRs (100%) in the domain of diabetes can be automatically evaluated, 9/17 (52.9%) in medication use, 5/10 (50%) in depression care, 3/6 (50%) in nutrition care, 6/13 (46.1%) in dementia care, 1/6 (16.6%) in end of life care, 2/13 (15.3%) in continuity of care, and 0/9 (0%) in the fall-related care. Lack of documentation of care activities between primary and secondary health facilities and ambiguous formulation of clinical rules were the main reasons for the inability to automate the clinical rules. Conclusion: Approximately two-fifths of the primary care Dutch ACOVE-based clinical rules can be automatically evaluated. Clear definition of clinical rules, improved GP database design and electronic linkage of primary and secondary healthcare facilities can improve prospects of automatic assessment of quality of care. These findings are relevant especially because the Netherlands has very high automation of primary care.