An automated method for analyzing adherence to therapeutic guidelines: application in diabetes

Stud Health Technol Inform. 2008:136:339-44.

Abstract

Background: Physicians' adherence to guidelines can be used for measuring prescribing appropriateness. We present a simple approach allowing the automation of this process.

Design: The drug therapy is described in terms of treatment type, pharmacotherapeutic classes, international non proprietary names (INN) and doses. A rule-based engine implementing the guideline generates recommendations for each patient record. These are automatically compared with prescriptions of the same patient in three levels of detail.

Participants: Ambulatory patients admitted for the follow-up of their type 2 diabetes between June 2003 and September 2004 in a university hospital in France.

Results: For 574 patient records included in the study, physicians agreed with the guideline recommendations over the choice of type of treatment in 473 cases (82%). When agreement over pharmacotherapeutic class of drugs was also taken into account, the adherence ratio decreased to 448 cases (78%). Finally, when the dosage of each drug was taken into account, the adherence ratio dropped to 396 cases (69%). Adherence ratios were also dependent on the type of treatment at admission: low for patients on oral tritherapy, and on diet and exercise. The results also highlighted inertia of physicians for beginning drug therapy and the underuse of biguanides.

Conclusions: The proposed method provides an automatable way of measuring the appropriateness of treatment choice, which can be used for chronic diseases.

MeSH terms

  • Combined Modality Therapy
  • Decision Support Systems, Clinical*
  • Diabetes Mellitus, Type 2 / drug therapy*
  • Dose-Response Relationship, Drug
  • Drug Therapy, Combination
  • France
  • Guideline Adherence / statistics & numerical data*
  • Humans
  • Hypoglycemic Agents / administration & dosage*
  • Medical Records Systems, Computerized / statistics & numerical data*

Substances

  • Hypoglycemic Agents