Healthcare-associated infections including central-line associated bacteremias and surgical site infections increase patient morbidity and mortality. Most states across the United States are requiring hospitals to publicly report healthcare-associated infection rates. Proper risk adjustment adjusts for patient factors without controlling for differences in the effectiveness of care enabling consumers and healthcare maintenance organizations to be informed purchasers of healthcare. Numerous prominent organizations and authors agree that the current method of risk adjustment for healthcare-associated infections is sub-optimal. The current knowledge as to which patient comorbid conditions can be used for risk-adjustment is not known. Our long- term goal is to use electronic data to identify risk factor for healthcare-associated infections that can be used for accurate risk adjustment. The overall objective of this application is to determine which comorbid conditions can be used for risk adjustment for central-line associated bacteremia and surgical site infection. Our central hypothesis is that comorbid conditions identified by ICD-9 codes or admission medications that are easily electronically-available by hospital across the United States can be used to better risk-adjust healthcare- associated infection rates. We plan to test our central hypothesis and, thereby, accomplish the objective of this application by pursuing the following two specific aims:
Aim 1. Perform a cohort study of adult patients with central venous catheters across greater than 20 hospitals to determine which electronically-obtained comorbidities are risk factors for central-line associated bacteremia (CLABSI).
Aim 2. Perform a cohort study of adult patients who underwent surgery across greater than 20 hospitals to determine which electronically- obtained comorbidities are risk factors for surgical site infection. Sites will be recruited as par of an arrangement between Premier Inc. and the University of Maryland. The advantage of this agreement is that it will allow us to recruit a large number of sites and perform the study in a short time period and in a cost- effective fashion. The proposed work will be significant because it will determine whether easily-obtainable electronic data on patient comorbid conditions can be used to risk adjust nationally-reported surgical site infection rates and central-line associated infection rates. The proposed research is innovative, in our opinion, because: a) it explores the use of patient comorbid conditions as risk factors and risk adjustment variables;and b) it explores variables that are easily obtained by electronic sources, making them potentially very easy to use for risk adjustment in national reporting c) it uses a large number of hospitals.
The proposed research is relevant to public health because the work will improve future risk adjustment by controlling for comorbid conditions. This will lead to the ability to compare healthcare-associated infection rates more appropriately and realize the full benefits of public reporting. It will also hopefully lead to states, CDC and CMS considering the use of ICD codes and admission medications to adjust nationally-reported infection rates as hospitals advance in their use of electronic medical records.
|Rock, Clare; Thom, Kerri A; Harris, Anthony D et al. (2016) A Multicenter Longitudinal Study of Hospital-Onset Bacteremia: Time for a New Quality Outcome Measure? Infect Control Hosp Epidemiol 37:143-8|
|Harris, Anthony D; Fleming, Brandon; Bromberg, Jonathan S et al. (2015) Surgical site infection after renal transplantation. Infect Control Hosp Epidemiol 36:417-23|
|Pepin, Christopher S; Thom, Kerri A; Sorkin, John D et al. (2015) Risk factors for central-line-associated bloodstream infections: a focus on comorbid conditions. Infect Control Hosp Epidemiol 36:479-81|
|Rock, Clare; Harris, Anthony D; Johnson, J Kristie et al. (2015) Infrequent air contamination with Acinetobacter baumannii of air surrounding known colonized or infected patients. Infect Control Hosp Epidemiol 36:830-2|