|Home | About | Journals | Submit | Contact Us | Français|
Patients with sickle cell disease (SCD) who undergo surgical procedures experience greater risk for preoperative and postoperative complications than patients without SCD; however, the impact of SCD on inpatient resource use and costs had not been reported [1–7]. We recently examined inpatient length of stay and total costs for patients with and without SCD who underwent cholecystectomy or hip replacement and found that patients with SCD had longer lengths of stay and incurred higher costs . In this study, we extend our previous work to surgical procedures that are less commonly performed in patients with SCD: hysterectomy, appendectomy, and knee replacement. Using a large national database of inpatient stays, we found that patients with SCD had significantly longer lengths of stay and incurred higher costs than patients without SCD who underwent the same procedures, after adjustment for patient, hospital, and procedural characteristics (P < 0.001). Higher inpatient costs were largely attributable to longer hospital stays. Future work should investigate the extent to which preoperative and postoperative complications and other factors contribute to longer stays among patients with SCD and how these factors might be addressed.
SCD is the most commonly inherited hemoglobinopathy, characterized by anemia and unpredictable, acute complications that can rapidly become life-threatening [9–14]. SCD affects more than 70,000 people in the United States and is predominant in African Americans, occurring in approximately 1 out of every 500 African American births [9,15]. Compared to patients without SCD, patients with SCD who undergo surgical procedures are at greater risk for preoperative and postoperative complications, with SCD-related postoperative complications occurring in over 30% of patients, even after preoperative blood transfusion [5–7, 16]. In a health care environment increasingly moving toward differentiated payments for lower- and higher- quality care, it is important to understand patient and procedural factors that contribute to higher complication rates, longer stays, and higher costs. We recently performed an analysis comparing inpatient outcomes between patients with and without SCD who underwent the 2 most commonly performed nonobstetric surgical procedures among patients with SCD in the United States: cholecystectomy and hip replacement . The study revealed that patients with SCD who underwent either procedure incurred longer inpatient stays and higher inpatient costs than patients without SCD.
In the present study, we expanded our analysis to other less commonly performed procedures (ie, hysterectomy, appendectomy, and knee replacement) in patients with SCD and found similar results. Using the 2002–2006 Nationwide Inpatient Sample (NIS) from the Healthcare Cost and Utilization Project (HCUP), we identified 118 SCD and 85,073 non-SCD discharges for hysterectomy, 69 SCD and 24,802 non-SCD discharges for appendectomy, and 62 SCD and 14,517 non-SCD discharges for knee replacement. For each procedure, patients with SCD were younger than patients without SCD (Table; P < 0.001 for each comparison). Hysterectomy and knee replacement were less likely to be recorded as elective surgery for patients with SCD than for patients without SCD (79.7% vs 86.0% [P = 0.05] for hysterectomy; 77.4% vs 93.1% [P < 0.001] for knee replacement). Compared to patients without SCD, patients with SCD were significantly more likely to receive blood transfusions during the inpatient stay (28.0% vs 4.3% [P < 0.001] for hysterectomy; 18.8% vs 2.5% [P < 0.001] for appendectomy; 30.6% vs 17.5% [P = 0.007] for knee replacement). Unadjusted comparisons of length of stay revealed that patients with SCD had mean hospital stays that were approximately 2 times longer than for patients without SCD who underwent the same procedure (P < 0.001 for each comparison). Mean inpatient costs followed a similar pattern, ranging from 32% higher for knee replacement to 91% higher for appendectomy (P < 0.001 for each comparison). The average cost among patients with SCD who underwent appendectomy was nearly $10,000 more than for patients without SCD. Percentages of patients with SCD who died after appendectomy or knee replacement were also significantly higher than for patients without SCD who underwent the same procedure (P = 0.02 for appendectomy; P = 0.001 for knee replacement), but the absolute numbers of patients with SCD who died were small (n ≤ 10 for each procedure).
After adjustment for patient, procedural, and hospital characteristics, length of stay and inpatient cost estimates remained significantly higher for patients with SCD than for patients without SCD (P < 0.001 for each comparison). Compared to patients without SCD, the adjusted length of stay for patients with SCD was 35% longer for hysterectomy (length of stay ratio, 1.35; 95% confidence interval [CI], 1.23–1.49), 85% longer for appendectomy (length of stay ratio, 1.85; 95% CI, 1.56–2.20), and 81% longer for knee replacement (length of stay ratio, 1.81; 95% CI, 1.62–2.02). Compared to patients without SCD, the adjusted cost estimates for patients with SCD were 28% higher for hysterectomy (cost ratio, 1.28; 95% CI, 1.18–1.39), 61% higher for appendectomy (cost ratio, 1.61; 95% CI, 1.40–1.85), and 19% higher for knee replacement (cost ratio, 1.19; 95% CI, 1.08–1.32).
After including inpatient length of stay and discharge disposition in the regression model for inpatient costs, we found that higher inpatient costs were largely attributable to longer hospital stays for all 3 surgical procedures. Inpatient costs for patients with SCD who underwent hysterectomy or appendectomy remained 8% and 22% higher, respectively, than for patients without SCD, after adjustment for inpatient length of stay and discharge disposition (cost ratio for hysterectomy, 1.08; 95% CI, 1.01–1.15; cost ratio for appendectomy, 1.22; 95% CI, 1.11–1.33).
Among patients with SCD, sickle cell crisis occurred in 31.9% (22/69) of patients who underwent appendectomy. In patients who underwent hysterectomy or knee replacement, ≤ 10 patients in each group experienced sickle cell crisis. Due to the small number of patients with sickle cell crisis, we did not evaluate the impact of this complication on length of stay, costs, or mortality.
In summary, after adjustment for patient, procedural, and hospital characteristics, patients with SCD incurred significantly longer inpatient stays and higher total costs for hysterectomy, appendectomy, and knee replacement than patients without SCD. Patients with SCD who underwent hysterectomy, appendectomy, or knee replacement were also significantly more likely to have a procedure code for blood transfusion during the inpatient stay, though the overall rates may be somewhat low due to undercoding of this procedure . After accounting for inpatient length of stay and discharge disposition, the impact of SCD on inpatient costs was attenuated for all 3 surgeries, but costs remained significantly higher for appendectomy and hysterectomy. The higher costs for appendectomy and hysterectomy may be associated with higher rates of sickle cell crisis and nonelective admissions for these 2 procedures. The study results were consistent with previous findings that focused on cholecystectomy and hip replacement , indicating that the study results may be generalizable to other procedures.
This study used the 2002–2006 NIS, part of the Healthcare Cost and Utilization Project sponsored by the Agency for Healthcare Research and Quality (AHRQ). The 2006 NIS consisted of complete discharge information for 1045 hospitals in 38 states in the United States . The NIS includes patient demographic characteristics, primary and secondary diagnoses, procedures performed, discharge status, admission and discharge dates, and inpatient charges. Diagnoses and procedures are coded using both International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) and Clinical Classification Software (CCS) codes. Inpatient charges were converted to costs using inpatient cost-to-charge ratios for individual NIS hospitals provided by AHRQ [19, 20].
We selected discharges of adults aged 18 to 64 years who had undergone hysterectomy (CCS code 124), appendectomy (CCS code 80), or knee replacement (CCS code 152). To limit the possibility that differences in inpatient outcomes between patients with and without SCD were attributable to hospital-level factors, we limited the sample of patients without SCD for each procedure to hospitals where at least 1 patient with SCD was discharged for the same procedure during the study period. Discharges of patients with SCD were identified using CCS code 61 as the primary or any secondary diagnosis. Sickle cell crisis was identified using ICD-9-CM diagnosis code 282.42, 282.62, or 282.64. Laparoscopic hysterectomy was identified using ICD-9-CM procedure code 68.31, 68.41, 68.51, 68.61, or 68.71. Laparoscopic appendectomy was identified using ICD-9-CM procedure code 47.01 or 47.11. We used the presence of at least one of ICD-9-CM procedure codes 99.01 to 99.04 to identify patients who received at least 1 blood transfusion during the hospitalization. In addition, we used the Elixhauser comorbidity algorithm, developed by AHRQ, to identify relevant comorbid conditions [21, 22].
Analyzing each procedure separately, we used descriptive statistics to summarize patient, procedural, and hospital characteristics and patient outcomes. Comparisons of categorical variables were based on chi-square tests or Fisher exact tests; comparisons of continuous variables were based on t tests. A generalized linear regression model was used to evaluate the independent association of SCD with length of stay and costs, after adjustment for potentially confounding variables. For length of stay, we used a negative binomial distribution with a log link. For costs, we specified the model using a gamma distribution with a log link. There were too few inpatient deaths to model mortality. Covariates included in each model for inpatient length of stay and costs were age (as a continuous variable); sex (for appendectomy and knee replacement); primary expected payer (categorized as private, Medicare, Medicaid, self-pay or no charge, or other); Elixhauser comorbid conditions [21, 22] (including congestive heart failure, valvular disease, pulmonary circulation disease, peripheral vascular disease, paralysis, other neurological disorders, chronic pulmonary disease, diabetes with or without chronic complications, hypothyroidism, renal failure, liver disease, peptic ulcer disease, acquired immune deficiency syndrome, lymphoma, metastatic cancer, solid tumor without metastasis, rheumatoid arthritis, coagulopathy, obesity, weight loss, fluid and electrolyte disorders, chronic blood loss anemia, deficiency anemias, alcohol abuse, drug abuse, psychoses, and depression); elective vs nonelective surgery; surgery type (ie, laparoscopic vs open surgery for hysterectomy and appendectomy); urban vs rural hospital location; hospital teaching status; hospital bed size; hospital ownership (public, private, not specified); and US geographic location.
To evaluate the impact of SCD on inpatient costs after adjustment for differences in length of stay and discharge disposition, we added length of stay as a continuous variable and discharge disposition as a categorical variable (ie, routine vs non-routine status, with non-routine status including transfer to short-term hospital, transfer to other, home health care, left against medical advice, died, and alive but destination unknown). In addition, we sought to determine the effect of sickle cell crisis on inpatient length of stay and costs. However we could not perform multivariable regression analysis, because of the small number of patients with SCD who had a diagnosis of sickle cell crisis.
We used SAS software 9.1.3 (SAS Institute Inc, Cary, North Carolina) for all statistical analyses. The institutional review board of the Duke University Health System approved the study.
Funding/Support: This study was supported by grant U10HL083698 from the National Heart, Lung, and Blood Institute as part of the Sickle Cell Disease Clinical Research Network.
Additional Contributions: The authors thank Damon Seils of Duke University for assistance with manuscript preparation. Mr Seils did not receive compensation for his assistance apart from his employment at the institution where the study was conducted.