Clostridium Difficile Hospital Discharges: Differences in Patients Admitted With or Without the Infection and the Role of the Hospital in Hospital-Acquired Clostridium Difficile Obioma Nwaiwu, MBBS

Co-authors Darcy K. McMaughan, PhD Rachel Edwards, BA Szu-hsuan Lin, MPH Charles Phillips, PhD, MPH

Disclosures No funding source supported this work None of the authors have any conflicts of interest to disclose

Outline Introduction Research Objectives Method Results Discussion Limitation

Introduction Clostridium difficile – potentially life threatening healthcare-associated infection (HAI) – Death occurs in 9% of cases (compared to 2% of all other inpatients) – Mean cost per hospital stay ~$24,400 – Rate continues to rise despite reduction in rates of other HAIs

– Newer cases - more severe and more resistant to treatment – Previous studies have focused on association between antibiotics use and hospital stay on the likelihood of acquiring the infection – Our study differs by its focus on the differences in individual characteristics of those who acquired the infection in the hospital

Research Objectives To investigate: – the incidence and demographics of hospital acquired C. difficile infection in Texas – the individual characteristics that increase the likelihood a patient admitted to hospital without a diagnosis of C. difficile infection will acquire the infection before discharge

Method Data – 2011 Texas inpatient discharge data – 2,620,000 discharges Analysis – Using ICD-9-CM diagnostic code, we identified all hospital discharges with C.difficile infection – SAS 9.3 – Logistic regression

Exclusion Criteria All cases from facilities exempt from reporting Diagnosis present on admission (POA) codes All cases with a diagnosis of C. difficile prior to being admitted to a hospital After exclusions were applied, we retained over 2,300,000 hospital discharges for our analysis

Baseline Characteristics Hospital acquired C.difficile (N=4,595) No C.difficile on discharge (N=2,329,367) P value < Female <.0001 Black <.0001 Other White Hispanic Origin <.0001 More than one comorbidity <.0001

Hospital acquired C.difficile (N=4,595) No C.difficile on discharge (N=2,329,367) P value At least one surgical procedure performed < days - 10 days in Hosp <.0001 Greater than 10 days in Hosp Less than 3 days in Hosp Extreme Severity <.0001 Major severity minor or moderate severity MSA

Findings VariableModel (3.401,4.102)** (4.667, 5.639)*** (5.813, 6.997)*** (5.327, 6.614)*** Female 0.950(0.896, 1.008) Black 1.073(0.981, 1.174) Other 1.087(0.996, 1.186) Hispanic Origin0.807(0.740, 0.880)*** More than one comorbidity At least one surgical procedure performed 3 days - 10 days in Hosp Greater than 10 days in Hosp

Findings VariableModel 1Model (3.401,4.102)**1.035(0.940,1.140) (4.667, 5.639)***1.159(1.051,1.279)** (5.813, 6.997)***1.304(1.184,1.437)*** (5.327, 6.614)***1.198(1.070,1.341)** Female 0.950(0.896, 1.008)1.199(1.129,1.272)*** Black 1.073(0.981, 1.174)0.780(0.712,0.854)*** Other 1.087(0.996, 1.186)1.007(0.922,1.099) Hispanic Origin0.807(0.740, 0.880)***0.798(0.732,0.870)*** More than one comorbidity 1.014(0.940,1.094) At least one surgical procedure performed 0.899(0.831,0.974)** 3 days - 10 days in Hosp5.307(4.347,6.480)** Greater than 10 days in Hosp (19.891,29.759)***

Findings VariableModel 1Model 2Model (3.401,4.102)**1.035(0.940,1.140)1.023(0.928,1.129) (4.667, 5.639)***1.159(1.051,1.279)**1.142(1.033,1.263)** (5.813, 6.997)***1.304(1.184,1.437)***1.284(1.162,1.418)*** (5.327, 6.614)***1.198(1.070,1.341)**1.171(1.043,1.314)** Female 0.950(0.896, 1.008)1.199(1.129,1.272)***1.198(1.128,1.273)*** Black 1.073(0.981, 1.174)0.780(0.712,0.854)***0.833(0.757,0.915)** Other 1.087(0.996, 1.186)1.007(0.922,1.099)0.943(0.850,1.047) Hispanic Origin0.807(0.740, 0.880)***0.798(0.732,0.870)***0.941(0.845,1.049) More than one comorbidity 1.014(0.940,1.094)0.988(0.914,1.069) At least one surgical procedure performed 0.899(0.831,0.974)**0.959(0.884,1.042) 3 days - 10 days in Hosp5.307(4.347,6.480)**5.320(4.356,6.495)** Greater than 10 days in Hosp (19.891,29.759)***23.163(18.911,28.369)***

Discussion These results indicate that the driving factors of hospital acquired C. difficile may be more attributable to individual characteristics than hospital characteristics This finding may be of considerable importance as payers introduce performance- based reimbursement measures that often include non-payment for hospital-acquired infections

Within such systems, the use of measures largely driven by individual characteristics is far from ideal Other measures may merit consideration in the reimbursement for the care of patients with hospital-acquired C. difficile

Limitations Conclusions drawn from the data may be subject to errors caused by the inability of the hospital to communicate complete data due to form constraints, subjectivity in the assignment of codes, and normal clerical error If ICD-9-CM coding of subjects are inaccurate, patients may have been misidentified as having C. difficile.

Limitations (Cntd) The generalizability of the study remains unknown since all hospitals contributing to the study are within Texas

Questions?