How can information extraction ease formalizing treatment processes in clinical practice guidelines? A method and its evaluation

Abstract

Objective: Formalizing clinical practice guidelines (CPGs) for a subsequent computer-supported processing is a challenging, but burdensome and time-consuming task. Existing methods and tools to support this task demand detailed medical knowledge, knowledge about the formal representations, and a manual modeling. Furthermore, formalized guideline documents mostly fall far short in terms of readability and understandability for the human domain modeler.

Methods and material: We propose a new multi-step approach using information extraction methods to support the human modeler by both automating parts of the modeling process and making the modeling process traceable and comprehensible. This paper addresses the first steps to obtain a representation containing processes which is independent of the final guideline representation language.

Results: We have developed and evaluated several heuristics without the need to apply natural language understanding and implemented them in a framework to apply them to several guidelines from the medical subject of otolaryngology. Findings in the evaluation indicate that using semi-automatic, step-wise information extraction methods are a valuable instrument to formalize CPGs.

Conclusion: Our evaluation shows that a heuristic-based approach can achieve good results, especially for guidelines with a major portion of semi-structured text. It can be applied to guidelines irrespective to the final guideline representation format.

Figure 1

Guideline transformation process. A multi-step process using intermediate representations to transform CPGs into a formal representation language.

Figure 2

Steps to obtain a semi-formal representation of CPGs. To gain process information from a CPG, the first two steps are accomplished in order to have a representation independent of the final guideline language.

Listing 1

Excerpt of a source listing of the marked-up guideline document “Evidence-based clinical practice guideline for children with acute bacterial sinusitis in children 1–18 years of age”. Relevant sentences are enclosed by HTML-like “a” tags.

Figure 3

Detecting relevant sentences. We split this task into two modules: (1) segmentation and filtering module and (2) template generation module.

Figure 4

Finding processes and extracting required information. We split this task into five modules (i.e., the structure extraction module, the slot extraction module, the merging and grouping module, the process extraction module, and the template generation module).

Listing 2

Action instance of an ActionIR template for the guideline “Evidence-based clinical practice guideline for children with acute bacterial sinusitis in children 1–18 years of age”.