Abstract
OBJECTIVES: The aim of this study is to develop a new predictive model to measure complexity of patients in medical wards. SETTING: 29 Internal Medicine departments in Italy. MATERIALS AND METHODS: The study cohort was made of 541 consecutive patients hospitalized for any cause, aged more than 40 years and with at least two chronic diseases. First, we applied a hierarchical cluster analysis and the principal component analysis (PCA) to a panel of questionnaires [comorbidity (Charlson, CIRS), clinical stability (MEWS), social frailty (Flugelman), cognitive dysfunction (SPSMQ), depression (5-item GDS), functional dependence (ADL, IADL, Barthel), risk of sore threats (Exton-Smith scale), nutrition (MNA), pain (NRPS), adherence to therapy (Morisky scale)], in order to select domains informative for the definition of complexity. The following step was to create the score(s) needed to quantify it. RESULTS: Two main clusters were identified: the first includes 7 questionnaires whose common denominator is dependence and frailty, the second consists of 3 questionnaires representative of comorbidity. Globally, they account for about 70% of the total variance (55.2% and 13.8%, respectively). The first principal component was simplified in "Complimed Score 1" (CS1) as a recalibrated average between the Barthel Index and the Exton Smith score, whereas the second cluster was approximated to "Complimed Score 2" (CS2), by using the Charlson score only. CONCLUSIONS: Complexity is a two-dimensional clinical phenomenon. The FADOI-Complimed Score(s) is a new tool useful for the routine evaluation of complexity in medical patients, simple to use and taking around 10 minutes to complete.