Authors: Luu NP, Hussain T, Chang HY, Pfoh E, Pollack CE, Luu NP, Hussain T, Chang HY, Pfoh E, Pollack CE
Title: ReCAP: Readmissions After Colon Cancer Surgery: Does It Matter Where Patients Are Readmitted?
Journal: J Oncol Pract :-
Date: 2016 Apr 05
Abstract: QUESTION ASKED: Do mortality and cost differ for patients with cancer who get readmitted to the hospital within 30 days after their initial cancer surgery on the basis of whether they are admitted to the same hospital where they had their surgery or to a different hospital? SUMMARY ANSWER: Patients with colon cancer who had a 30-day readmission to a different hospital than the one where they received their cancer surgery did not experience higher all-cause mortality, colon cancer-specific mortality, or costs compared with patients who were readmitted to the hospital where they underwent their cancer surgery. However, patients readmitted to a different hospital did have a higher risk of short-term (90-day) mortality. METHODS: We used SEER-Medicare linked data of patients with colon cancer diagnosed between 2000 and 2009. Our cohort consisted of patients with stage I to III disease who received colon cancer surgery and were subsequently readmitted within 30 days of their initial discharge. We compared patients who were readmitted to the same hospital as where their surgery was performed and those admitted to a different hospital. We performed propensity score-weighted doubly robust models to examine the association with all-cause mortality, cancer-specific mortality, 90-day all-cause mortality, 90-day cancer-specific mortality, and costs (Table 2). Our models were adjusted for patient, physician, and hospital characteristics. We used Cox proportional hazard models for all-cause mortality, the method of Fine and Gray for competing risk regression for colon cancer-specific mortality, and log transformed generalized linear models with gamma distribution to model costs. BIAS/CONFOUNDING FACTORS, DRAWBACKS: The choice of both the initial hospital where patients received surgery and the subsequent hospital where they were readmitted is unlikely to be random. Though we used propensity score adjusted models, we were unable to account for the reasons why patients selected one hospital versus another and were unable to assess factors such as travel distance. In addition, we were unable to determine whether hospitals were part of an integrated network, which may bias our findings toward the null. Finally, the generalizability of our sample is limited to Medicare fee-for-service beneficiaries and to patients who were not transferred during their initial hospital admission. REAL-LIFE IMPLICATIONS: Readmission after cancer surgery is a costly problem, and a large number of clinical initiatives and policy efforts have centered on reducing rates of readmission. Less attention has been focused on whether it matters where patients get readmitted. We were concerned that patients who are readmitted to a different hospital from where they received their cancer surgery may be more likely to experience poorly coordinated and costly care and be subject to poor outcomes. We focus on readmissions after colon cancer, the third most common cancer and one for which surgery is the cornerstone of treatment of patients with stage I to III disease. Our findings indicate that readmission to the same hospital does not necessarily confer long-term mortality or costs savings after initial discharge. Further investigations should focus on the reasons for the short-term differences in mortality and determine mechanisms that may underlie the lack of long-term benefits or cost savings.jop;JOP.2015.007757v1/T02T1T02Table 2.Short- and Long-Term All-Cause and Colon Cancer-Specific Mortality, and Cost Associated With Readmission in 30 Days, to Different Hospital Versus Same Hospital From Propensity Score-Matched Doubly Robust ModelsHazardHazard Ratio (95% CI) for All-Cause Mortality (n = 3,399)Subhazard Ratio (95% CI) for Colon Cancer-Specific Mortality (n = 534)90-day Hazard Ratio (95% CI) for All-Cause Mortality (n = 3,399)90-day Subhazard Ratio (95% CI) for Colon Cancer-Specific Mortality (n = 534)Dollars Saved (95% CI) From Generalized Linear Model for Patients After Discharge (n = 3,399)*Readmitted to same hospitalRefRefRefRefRefReadmitted to different hospital1.04 (0.87 to 1.26)1.09 (0.79 to 1.51)1.18 (1.02 to 1.38)1.58 (0.38 to 6.65)$8,405 (-$4,202 to $23,114)NOTE. Estimates are fully adjusted for all patient, provider, and hospital characteristics listed in Table 1 and also for year of diagnosis, SEER site, and length of stay at index hospital. Long-term mortality is from date of diagnosis. Short-term mortality is from date of index hospital discharge.Abbreviation: Ref, reference.*Costs were calculated from date of discharge from the initial hospital. All dollar values were inflated to 2009 using the annual Gross Domestic Product price index.
Last Updated: 02 Mar 2015