How to translate therapeutic recommendations in clinical practice guidelines into rules for critiquing physician prescriptions? Methods and application to five guidelines

Table 3 Characteristics of the knowledge bases.

		Hyp.	Diab.	Dys.	Tob.	Atr.	Thr.
Modelled	Number of pharmaco-therapeutic drug classes	7	7	10	5	11	3
CPG	Number of recommended treatments	30	39	15	12	14	3
recommendations	Number of "one should prescribe..." recommendations	30	14	9	7	6	6
	Number of "one should not prescribe..." recommendations	19	10	6	11	4	1
	Number of "treatments of increasing power..." recommendations	1	3	2	1	0	0
	Total number of recommendations	50	27	17	19	10	7
Generated rules	Number of "if ... then criticism" rules	119	102	73	42	51	16
	Number of treatment patterns	826	1121	708	267	591	82

Hyp.: hypertension, Diab.: type 2 diabetes, Dys.: dyslipaemia, Tob.: tobacco addiction, Atr.: atrial fibrillation, Thr.: thrombo-embolic risk. User-defined knowledge bases refer to the recommendations as entered into the system, and generated knowledge bases to the rule bases automatically generated by the system, including generic rules and after applying macros to transform all recommendations into "if conditions then criticism" rules.

ISSN: 1472-6947