The Data-Driven Definition of "Complication" in Medicine

Area of Science:

• Medical Informatics
• Natural Language Processing
• Patient Safety

Background:

• The term "complication" lacks a standardized definition in medicine.
• This ambiguity hinders systematic tracking and improvement of medical care safety.
• A clear definition is crucial for both individual clinics and the broader healthcare system.

Purpose of the Study:

• To establish a precise, generally accepted definition of "complication" in the medical context.
• To identify the generic concept and key distinguishing features of complications.
• To leverage natural language processing for concept definition.

Main Methods:

• Linguistic and statistical analysis of a large corpus (90,688 neurosurgical cases, 2000-2017).
• Tokenization, normalization, and lexicon development (Lexicon of Adverse Medical Events - LAME).
• N-gram vector representations and cosine similarity to identify core terms related to complications.

Main Results:

• "Pathology" identified as the generic concept for complication.
• Key features: novelty and emergence during observation of a primary phenomenon.
• Proposed definition: A complication (in medicine) is an intercurrent pathology detected during observation of an underlying disease, physiological process, or the result of intervention.

Conclusions:

• The study provides a scientifically grounded definition for "complication" to enhance medical care safety monitoring.
• The proposed definition clarifies the concept, facilitating systematic tracking and analysis.
• The methodology can be applied to define other poorly understood medical concepts.