Distribution and predictive accuracy of the Index of Caring Complexity in a semi-intensive care unit

To analyze nursing complexity level and its relationship with mortality and length of stay.

Observational study carried out between 1/06/2014 to 21/06/2015 at High Dependency Unit of an Italian University Hospital. The nursing needs and the Index of Caring Complexity (Indice di Complessità Assistenziale, ICA) were collected following the "Modello delle Prestazioni Infermieristiche".

Data from 506 patients were collected for a maximum of 4 days of hospital stay (1,436 data collection). Nursing needs with the higher score were mostly related to cardio-circulatory function. Average ICA was 38.4. Only 17.2% of patients were classified as "high complexity" at admission. During the hospital stay, the ICA showed a progressive decrease in survivors while increased in those who died. Overall, the ICA showed high predictive accuracy regarding the risk of death (90.3%; p=0.006); the best ICA cut-off point was 41.5.

The risk of death increased with the increasing complexity of nursing care. Exceeding the ICA cut-off value, as well as an increasing trend in the ICA daily monitoring, may be adopted as criteria to identify patients at higher risk, for which increasing, for example, the nurse-patients ratio. Identification of the patient's complexity level may enable the delivery of more personalized nursing activities or interventions based on the intensity of care needed, as well as the proper allocation of patients during hospitalization. This strategy may allow the rationalization of available resources and better planning and implementation of nursing care.