Abstract
Most studies of survival in sickle cell disease (SCD) include only patients followed at referral centers. We used the public set of the California Patient Discharge Databases (1998–2005) to compare in-hospital mortality and charges by patient characteristics. We included discharges with any diagnostic code for SCD (282.41–2, 282.60–69). We used the Charlson index (adapted for administrative data) to adjust for comorbid conditions. Hospital charges were adjusted to 1998 levels. Statistical methods included ANOVA, Wilcoxon rank-sum, and multivariate linear and logistic regression. We identified 57,887 admissions for SCD and 376 in-hospital deaths (0.65%) from 1998–2005. The mean age of death was 40 years (95% CI 38–42) for sickle cell anemia (HbSS) and 48.3 years (95% CI 44.1–52.5, p<.0005) for all other SCD. Median length of stay was 4 days (IQR 2–7 days) and charges were $10,027 (IQR 5547–18,302) for HbSS and 4 days (IQR 2–6 days) and $8045 (IQR 4375–16,253) for all other SCD (p<.0001). Women with SCD were older at the time of death (43.6 years) than men (41.4 years, p=.29) and age of death increased from 40.6 years in 1998–2000 to 44.2 in 2004–2005 (p=.17) but these differences were not significant. Mortality was increased with a Charlson Index of 1 or 2, older age, private insurance, a diagnosis of HbSS, and transfusion during the admission (Table 1). Women had a lower odds of death than men (OR 0.7, p<.05). Average annual charges for hospitalization in patients with SCD, adjusted to 1998 values, were $117,000,000 and decreased $1,960,000/year (−4.1%, p<.0001) for children and increased $1,150,000/year (1.4% p<0.05) for adults. Adjusted charges for all children were stable and increased 1.5% per year for adults. Charges per admission were $3167 higher for adults than children, $5608 higher per comorbid diagnosis, $6506 higher in transfused patients, and $3595 higher with a diagnosis of crisis (p<.0001 for all). Compared to MediCal, charges were $1186 higher for Medicare (p<.05), $2651 lower for other government insurance (p<.0001), and $3521 lower for self-pay (p<.05). Government (77%) and private insurance (21%) paid for most admissions. We identified a greatly increased risk of in-hospital morality with comorbid diagnoses and older age and moderately increased risk with private insurance, HbSS genotype, and transfusion during the admission. The increased risk with private insurance was surprising, as higher socioeconomic status is often associated with better health outcomes. This may reflect more stringent requirements for admission or financial barriers to outpatient and preventive services for SCD. Total charges decreased dramatically in children, possibly reflecting increased use of hydroxyurea for frequent pain and chronic transfusions for primary stroke prevention.
Table 1: Odds Ratio for In-Hospital Death Based on Multivariate Logistic Regression Variable Odds Ratio (95% CI) P-Value NS indicates not significant Charlson Index=1 3.3 (2.4–4.6) <.0001 Charlson Index=2 12.9 (9.3–18) <.0001 Age (per year) 1.04 (1.03–1.05) <.0001 Private Insurance vs. MediCal 1.6 (1.2–2.2) <.0005 Medicare vs. MediCal 0.9 (0.7–1.3) NS Other SCD vs. HbSS 0.7 (0.5–0.9) <.0005 Transfusion 1.6 (1.2–2.0) <.0001
The molecular basis for the prothrombotic state in sickle cell disease
