Design patterns for clinical guidelines

  • Authors:

  • Mor Peleg;Samson W. Tu

  • Affiliations:

  • Stanford Center for Biomedical Informatics Research, Stanford University, Stanford, CA 94305, USA and Department of Management Information Systems, University of Haifa, Haifa 31905, Israel;Stanford Center for Biomedical Informatics Research, Stanford University, Stanford, CA 94305, USA

  • Venue:
  • Artificial Intelligence in Medicine
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

Objective: Transforming narrative guidelines into a computer-interpretable formalism is still a bottleneck in the development of decision-support systems. Our goal was to support this step by providing computer-interpretable templates for representing guideline knowledge using clinical abstractions that are appropriate for particular guideline sub-domains. Methods and materials: We analyzed guidelines taken from the sub-domains of screening and immunization guidelines to find repeatable clinical abstractions and structured them as design templates to support encoding of these guidelines in a computer-interpretable format. To find guidelines for analysis and validation, we (1) searched the National Guideline Clearinghouse for screening guidelines in internal medicine, that have a clinical algorithm, and which were published during 2002-5 and (2) used adult and childhood immunization guidelines developed by Center of Disease Control and Prevention (CDC) and the Institute for Clinical Systems Improvement. Results: We developed two visual templates that structure screening guidelines as algorithms of guideline steps used for screening and data collection and used them to represent the guidelines collected. We validated the computability of the screening templates by executing a screening guideline in a workflow engine. We validated the computability of immunization templates by writing code that, based on represented knowledge, computes immunization due dates and by creating an algorithm that translates the knowledge into computer-interpretable guidelines. Conclusion: We have demonstrated that our templates could be effectively applied to screening and immunization guidelines to produce computer-interpretable representations using domain-level abstractions.