Sickle cell disease (SCD) affects about 100,000 people in the US. The disease causes significant mortality, morbidity and frequent hospitalizations secondary to primary manifestations of the disease and its complications. We conducted a retrospective, single institution study in adult patients (pts), ≥ 18 years with SCD to assess factors associated with health care utilization (HCU: Hospital admissions and/or Emergency Department (ED) visits). For pts admitted to the hospital, we studied the association of demographic, clinical, laboratory and treatment factors with the length of stay.
Methods:
After IRB approval, we conducted a chart review of adult pts with SCD between January 1st 2014 and December 31st 2015. For all pts, we recorded demographic data including age, gender, if living alone, insurance type and availability of primary care provider, co-morbidities, type of hemoglobin, treatment with hydroxyurea and chelation. We studied the differences in sociodemographic and clinical characteristics of pts with HCU and those without. In pts who were admitted, we collected additional data including admission vitals, labs and diagnosis, treatment with patient controlled analgesia, blood transfusions, antibiotics and the total length of stay. Computations were performed using SPSS statistical software. Poisson regression analysis with an over dispersion parameter was applied to identify and evaluate the risk factors associated with the number of admissions and ED visits. The Akaike information criterion was used for model evaluation and selection.
Results:
We found a total of 92 unique pts in the given time period, of which 67 had HCU while 25 did not. Pts with private insurance were less likely to have HCU compared to pts with government-issued insurance (25% vs 75%, p= 0.002). Private insured pts had a median of 1 episode of HCU vs 4 for others (p 0.02). This persisted on regression analysis.
There were a total of 341 admissions during the period of study. In the univariate analysis for the total length of stay, we found statistical significance with the following factors (p< 0.05): Chronic kidney disease, thromboembolic disease, home oxygen, high WBC, high reticulocyte count, elevated creatinine, need for blood transfusions and antibiotics use. However, on multivariate analysis, only blood transfusions and antibiotic use were found to be statistically significant (Table 1).
Conclusions:
In our study of adult pts with SCD , patients with private insurances had lesser likelihood of HCU. For pts admitted to the hospital, need for blood transfusions and antibiotic therapy were found to be associated with increased length of stay. A larger prospective multicenter study is needed to confirm these findings.