An Assessment of the Validity of the Comprehensive Severity Index (CSI) as a Measure of Severity of Influenza Infection in Children. (4th October 2017)

• Record Type: Journal Article
• Authors: Tran, Dat; Richardson, Susan E; Ipp, Moshe; Schuh, Suzanne; Bitnun, Ari; Paterson, Andrew
• Is Part Of: Open forum infectious diseases. Volume 4(2017)Supplement 1
• Page Start: S693
• Page End: S693
• Publication Date: 2017-10-04
• Subjects: Communicable diseases -- Periodicals; Medical microbiology -- Periodicals; Infection -- Periodicals; 616.9
• DOI: 10.1093/ofid/ofx163.1859
• Languages: English
• ISSNs: 2328-8957

Abstract

Abstract: Background: A standardized quantitative severity score that reflects the breadth of influenza-related complications would prove valuable in epidemiologic analyses. The maximum CSI score (maxCSI) is a composite, continuous measure of illness severity, based on the degree of abnormality of individual signs and symptoms of a patient's disease or diseases. Importantly, the index contains criteria for influenza as well as related complications. Methods: We evaluated the spectrum of influenza illness as measured by maxCSI and assessed its discriminatory power on 321 influenza-infected, otherwise healthy children (0–17 years) enrolled into a prospective study from the emergency department and inpatient units of a pediatric tertiary care hospital and an urban community pediatric clinic. The area under ROC curve (AUC) was computed for univariate (maxCSI as a sole predictor variable) and multivariable logistic regression models of two outcome measures: (1) influenza-related respiratory and extra-respiratory complications based on physician diagnosis and (2) hospitalization. Multivariable models incorporated maxCSI, age, household crowding, influenza type/subtype and antiviral therapy. For each outcome, the Hanley-McNeil method was used to compare AUCs of univariate and multivariable models. Results: Of the 321 children enrolled, 200 (62.3%) were male and the median age was 5.25 years (range 0.07–17.96). 73 (22.7%) had complications while 61 (19.0%) were hospitalized; the median maxCSI was 25 (range 0–140). In univariate and multivariable modeling, maxCSI was significantly associated with both influenza-related complications and hospitalization (all P < 0.0001). The univariate models discriminated well between children with and without complications [AUC 0.88 (95% CI 0.83–0.93)] and between those who were and were not hospitalized [AUC 0.94 (95% CI 0.91–0.97)]. The AUCs for the corresponding multivariable models were not statistically significantly different: 0.90 (95% CI 0.86–0.94; P = 0.13) for complications and 0.94 (95% CI 0.91–0.97; P = 0.34) for hospitalization. Conclusion: The maxCSI represents a valid continuous outcome measure that can be leveraged to increase statistical power in epidemiologic studies aimed at identifying factors associated with severe influenza. Disclosures: All authors: No reported disclosures.