[Comorbidity in medical guidelines: comparison of the current state, epidemiologic models and expert opinion]

Medical guidelines focusing on monomorbidities can be associated with adverse events in multimorbid patients. This study investigates how comorbidities are actually particularised in a set of German guidelines. In addition, it evaluates whether two epidemiologic approaches (disease combinations or clusters of comorbidities) can be used to systematically integrate multimorbidity in guideline development.

Based on a matrix of 30 comorbidities, mentioning of comorbidities in 8 current German guidelines (diabetes mellitus, hypertension, heart failure, coronary heart disease, chronic obstructive lung disease/asthma, coxarthrosis, low back pain, osteoporosis) was investigated. These so called index diseases were selected on the basis of the hypothetical case of a multimorbid patient published by Cynthia Boyd and colleagues in 2005. Mentioning of comorbidities in the guidelines was compared to the epidemiologic approaches of disease combinations and clusters of comorbidities. In addition, using the comorbidity matrix, 36 physicians involved in everyday care of multimorbid patients assessed whether an explicit recommendation for the listed comorbidities would be helpful.

Mentioning of comorbidities was very heterogeneous across the guidelines investigated, ranging from 0 to more than 10. The proportion of the comorbidities that were considered relevant by the survey participants ranged from 0 % to 62 % with a focus on cardiovascular and metabolic diseases. When using disease combinations, only 0 to 3 of the "relevant" comorbidities were identified. Using the cluster model may be helpful in identifying whether a particular comorbidity is thematically close to the index disease or whether it is associated with an interacting thematic area.

Methodological support is needed for addressing comorbidities in guidelines in a more consistent way. The currently existing epidemiologic approaches should not be used in their current form without being further developed and re-evaluated. Expert opinion of physicians involved in the care of multimorbid patients should be systematically included in methodological refinement studies.