All-Patient-Refined Diagnosis Related Groups’ Severity of Illness (SOI) and Risk of Mortality (ROM) as prediction tools of in-hospital mortality: an example from burns

João Vasco Santos1, Júlio Souza1, João Viana1, Alberto Freitas1

1MEDCIDS – Department of Community Medicine, Informatics and Decision in Health; CINTESIS – Center for Health Technology and Services Research

Aim:

Our aim was to study the All-Patient-Refined Diagnosis Related Groups’ Severity of Illness (SOI) and Risk of Mortality (ROM) as prediction tools for in-hospital mortality among inpatients with burns and to compare with burn-specific mortality prediction scores.

Methods:

We performed a retrospective nationwide study, including Portuguese mainland hospitalizations of patients of all ages, from 2011 to 2014, with a major diagnosis category (MDC) of 22 – Burns (i.e. DRG 841-844) and percent of total body surface area (%TBSA) burned available, provided by the Portuguese Ministry of Health’s Authority for Health Services (ACSS). Receiver Operating Characteristic (ROC) curves for in-hospital death were performed, using SOI, ROM, ABSI, Baux and modified Baux score, and area under the curve (AUC) for each curve was calculated with 95% confidence intervals.

Results:

A total of 4656 hospitalizations took place with the MDC 22 (Burns) and had a %TBSA, between 2011 and 2014 in Portugal. In-hospital mortality rate was of 4.1 % (n=189). Only the AUC of ROM (95%CI 0.890-0.929) was significantly inferior to that from modified Baux score (95%CI 0.930-0.957). There were no more significant differences between AUCs for each other scores in predicting in-hospital mortality rate; in a decreasing order: modified Baux (AUC=0.943), ABSI (0.936), Baux (0.933), SOI (0.910) and ROM (0.909).

Discussion:

SOI and ROM were good prediction tools of in-hospital mortality in burns and comparable to burn-specific mortality prediction scores. This can be a tool to be used in the future for several diseases and in the quality evaluation or benchmarking.

keywords: prognosis, administrative databases, burns

Presentation: All-Patient-Refined Diagnosis Related Groups’ Severity of Illness and Risk of Mortality as prediction tools of in-hospital mortality