期刊名称：International Journal of Population Data Science
电子版ISSN：2399-4908
出版年度：2017
卷号：1
期号：1
页码：1-1
DOI：10.23889/ijpds.v1i1.98
出版社：Swansea University
摘要：ABSTRACTIntroductionThe Hospital Information System of the Brazilian Unified Health System (SIH-SUS) can be an important tool in evaluating care through comparisons of mortality rates among hospitals. However, the SIH-SUS has limited availability of data on comorbidities, which are needed in order to measure the severity of patients' condition for risk adjustment when comparing mortality among hospitals. ObjectivesTo analyze previous hospital admissions as an indicator of severity of patients' condition using as an example hip fracture admissions. ApproachWe analyzed 1984 patients with 62 years of age or older with hip fracture who were admitted to the public health system in Rio de Janeiro City between 2010 and 2011. The causes of previous hospital admissions were obtained through probabilistic record linkage with the SIH-SUS for all causes of admissions between 2008 and 2010. We analyzed the association between the patient having had a any previous admission, adjusted by sex and age, with hospital death, through a logistic regression model. We also analyzed the association between groups of causes of admission with hospital death. ResultsHaving had at least one previous hospital admission due to any cause before the admission for hip fracture was associated with a 77% increase in the odds of hospital death (CI 95% = 1.06 – 2.94). Among comorbidities, the most relevant were severe or moderate kidney disease with OR = 20.196 (CI 95% = 3.042 – 134.092), ischemic diseases with OR = 9.099 (CI 95% = 0.973 – 85.106), pneumonia with OR = 3.619 (CI 95% = 0.977 – 13.401) and diverse fractures and lesions with OR = 2.041 (CI 95% = 0.900 – 4.627). We were not able to analyze the association for some comorbidities due to an absence in outcome variability in some groups. ConclusionResults point to a promising use of previous hospital admissions in the Hospital Information System as a proxy for patient's risk adjustment.
