|Home | About | Journals | Submit | Contact Us | Français|
Racial disparities have been described in the utilization of a diverse spectrum of surgical procedures. Our objectives were to determine whether disparities also exist for the use of parenteral nutrition (PN) in IBD.
We analyzed the Nationwide Inpatient Sample from the U. S. between 1998 and 2003 to determine PN utilization among IBD inpatients diagnosed with protein-calorie malnutrition and assessed if utilization patterns differed by race and geographical region.
The proportion of African American IBD admissions with protein-calorie malnutrition who received PN was significantly lower than that in whites (19.9% vs. 28.1%, p=0.001), while there was no difference between Hispanics and non-Hispanic whites. After adjustment for gender, comorbidity, health insurance status, geographic region, and median neighborhood income, African Americans remained less likely than whites to receive PN (OR 0.67; 95% CI: 0.50 – 0.89) while the difference between Hispanics and non-Hispanic whites was marginally significant (OR 0.65; 95% CI: 0.41 – 1.04). PN use varied geographically, with highest rates in the Northeast (44.3%) and lowest in the Midwest (17.3%). Uninsured patients were less than half as likely to receive PN as those with insurance (OR 0.46; 95% CI: 0.31 – 0.69). Compared to whites, Hispanics experienced a longer time interval between admission and initiation of PN (3.5 vs. 4.8 days, p=0.02) and had higher rates of catheter-related complications (5.1% vs. 12.2%, p=0.04).
Among IBD inpatients with clinically diagnosable malnutrition, PN utilization was lower among African Americans compared to whites. The underlying mechanisms of these racial variations merit further investigation.
Patients with inflammatory bowel disease (IBD), including Crohn’s disease (CD) and ulcerative colitis (UC), are susceptible to protein-calorie malnutrition. Nutritional deficiency is secondary to multiple factors including decreased oral intake, malabsorption, increased metabolic expenditure, and drug-nutrient interaction (1). Malnutrition may be present in as many as 40% of hospitalized IBD patients and contributes to increases in both length of stay and hospital costs (2). Supportive interventions to improve the nutritional status of malnourished IBD patients are an important adjunct to medical therapy (1–4). Parenteral nutrition (PN) may be required for hospitalized IBD patients with malnutrition because of impaired bowel function or who are otherwise unable to tolerate oral or enteral feedings.
Initiation of PN requires placement of an indwelling venous catheter. Venous catheterization can be complicated by thrombophlebitis, deep venous thrombosis and pulmonary embolism, and catheter-related infections (1). Racial disparities have been extensively described for utilization of major surgical procedures (5), including colectomy for UC and bowel resection for CD (6;7). Because placement of a venous catheter is a relatively minor and ubiquitously performed procedure, access to colorectal surgeons is not as likely to be a contributing factor to racial variations in utilization as it could be for other major bowel procedures. However, the administration of parenteral nutrition is expensive, may extend length of stay, and usually requires posthospitalization nursing care either at home or in a subacute nursing facility (1). The high costs associated with parenteral nutrition may thus introduce racial and socioeconomic disparities in its utilization. The primary objective of this study was to assess for variations in the utilization of PN among white, African American, and Hispanic hospitalized IBD patients with clinically diagnosed protein-calorie malnutrition.
All data were extracted from the Nationwide Inpatient Sample (NIS) between 1998 and 2003. It is the largest all-payer database of national hospital discharges, maintained as part of the Healthcare Cost and Utilization Project (HCUP) by the Agency for Healthcare Research and Quality (AHRQ). The NIS is a 20% stratified sample of non-federal, acute-care hospitals in the United States. This sample includes community and general hospitals and academic medical centers but excludes long-term care facilities. Hospitals were divided into strata on the basis of geographic region, urban versus rural location, teaching status, ownership, and bedsize. Each record in the NIS represents a single hospital discharge and includes a unique identifier, demographic data (age, gender, and race), admission type (emergent, urgent, or elective), primary and secondary diagnoses (up to 15), primary and secondary procedures (up to 15), expected primary and secondary insurance payers, total hospital charges, length of stay, and hospital characteristics (region, urban versus rural location, bedsize, teaching status). The NIS data concur with the National Hospital Discharge Survey, thus supporting data reliability (8).
We extracted and analyzed all hospital discharges between the years of 1998 and 2003 with the following criteria: 1) the patient was between the ages of 5 and 80 and his or her racial and ethnic data were available; 2) the patient had either a primary diagnosis of inflammatory bowel disease (UC or CD) identified by the Clinical Modification of the International Classification of Diseases, 9th Revision (ICD-9-CM) code of 555.0–556.9 or a primary diagnosis of a complication of UC or CD with a secondary or tertiary diagnosis of UC or CD; 3) and the patient had a diagnosis of protein-calorie malnutrition (260.0–263.9, 799.4, 783.3, 269.8). We excluded children younger than 5 years because of uncertainty of diagnosis in this subpopulation. Adults older than 80 years were excluded because they are more likely to have ischemic colitis misdiagnosed as IBD.
Data on race and ethnicity were derived from hospital administrative data and included the following categories: white, African American, Hispanic, Asian or Pacific Islander, Native American, or other. Because of small numbers, the latter three categories were not included in the analysis. For Hispanic patients, ethnicity took precedence over race and was coded as “Hispanic. “ All other race categories were non-Hispanic. Case-mix adjustment was performed using the validated Deyo modification of the Charlson index (CI) (9;10). The Charlson index is widely used to characterize and adjust for disease burden and case-mix in administrative data. This scale incorporates 17 comorbid conditions that are weighted to yield a summary index. The summarized Charlson index score is a validated predictor of all-cause in-hospital mortality (11). Parenteral nutrition (PN) was identified by the ICD-9-CM procedure code 99.15. The pre-PN interval was defined as the time period between admission and initiation of parenteral nutrition. Catheter-related complications (infection, vessel injury, venous thrombosis) were ascertained from ICD-9-CM codes (900.0–901.9, 903.0–903.9, 996.60, 996.62, 997.2, 999.2, 999.3, 999.79).
Data were analyzed using the Stata 9.0 SE software package (Stata Corporation, College Station, Texas). Analyses took into account the stratified two-stage cluster design using Stata’s SVY (survey data) commands. Two-way χ2 analyses were performed to compare discharge-level categorical variables among different racial groups. Multiple logistic regression was used to calculate the association between race and likelihood of receiving parenteral nutrition, while adjusting for age, gender, primary health insurance carrier, comorbidities, calendar year, and hospital characteristics (region, bedsize, rural versus urban location, and teaching status). The impact of race on the interval between admission and initiation of parenteral nutrition was assessed using multiple linear regression incorporating sociodemographic and clinical covariates. Because the pre-PN interval was skewed, the models were constructed for this outcome after logarithmic transformation. Coefficients from these models were exponentiated to yield a percentage change in the pre-PN interval associated with each predictor.
The research protocol was approved by the Institutional Review Board of the Johns Hopkins Medical Institutions.
Baseline demographic and clinical data for the 3788 non-Hispanic white, 449 African American, and 190 Hispanic admissions for IBD and malnutrition are shown in Table 1. Whites were on average older than African Americans and Hispanics (47.6 years vs. 37.7 years and 40.7 years, respectively, p<0.001). There were modest racial differences in gender distribution, with whites and African Americans exhibiting a slight female predominance (55.2% and 53.8%, respectively) whereas Hispanics had a higher male representation (57.2%). African Americans and Hispanics were less likely to be privately insured and more likely to receive Medicaid than were whites. The proportion of whites who resided in neighborhoods where the median income was greater than the national median was substantially higher than that for African Americans and Hispanics (48.1% vs. 26.4% and 31.5%, respectively, p<0.001). There was also a higher ratio of Crohn’s disease to ulcerative colitis among African Americans compared other races. The geographic distribution of races also differed with African Americans being heavily concentrated in the South and Hispanics residing predominantly in the South and West.
The proportion of IBD admissions with malnutrition who received parenteral nutrition (PN) was lower in African Americans than whites (19.9% vs. 28. 1%, p=0. 001). There was no statistically significant difference between Hispanics and whites (22.9% vs. 28. 1%, p=0. 27). After adjustment for age, gender, health insurance carrier, comorbidity, geographic region, IBD diagnosis, neighborhood income, and hospital characteristics, African Americans remained less likely than whites to receive PN (OR 0.67; 95% CI: 0.50 – 0.89) (Figure 1). A similar difference was seen between Hispanics and non-Hispanic whites but was only marginally significant, probably due to the small sample size (OR 0.65; 95% CI: 0.41 – 1.04). There was significant geographic variation in the implementation of PN for malnutrition. The rate was highest in the Northeast (44.3%) and was 17.3%, 20.3%, and 27.4% in the Midwest, South, and West, respectively (p<0.001). Residing in a zip code area in which the median income was greater than the national median was independently associated with an increased likelihood of receiving parenteral nutrition (OR 1.20; 95% CI: 1.01 – 1.41). Compared to being privately insured, receiving Medicare (OR 0.80; 95% CI: 0.65 – 0.94) and being uninsured (OR 0.46; 95% CI: 0.31 – 0.69) were associated with lower likelihood of PN after adjustment for age, comorbidity, and other confounders.
Among malnourished IBD patients who eventually received PN, Hispanics experienced a longer time interval between admission and initiation of PN than whites (4.8 vs. 3.5 days, p=0.02). African Americans also waited longer prior to initiation of PN than whites, but this difference was not statistically significant (6.2 vs. 3.5 days, p=0.15). After adjustment for age, gender, health insurance, comorbidity, geographic region, and hospital characteristics, African Americans experienced a 42% (95% CI: 11% – 83%) longer pre-PN interval than whites. The pre-PN interval was 78% (95% CI: 26% – 150%) longer in Hispanics than whites. Admission from the South had a 16% (95% CI: 18% – 20%) shorter pre-PN interval than the Northeast.
The overall rate of in-hospital catheter-related complications was 5.5%, but the rate in Hispanics was more than twice that in non-Hispanic whites (12.2% vs. 5.1%, p=0.04). The rates in African Americans and whites were similar (7.6% vs. 5.1%, p=0.35). In-hospital mortality was not different among racial groups.
We found substantial racial differences in the utilization of PN for malnourished IBD patients in the hospital setting. Because nutritional support is an important adjunct to medical therapy, our findings suggest potential differences in quality of care. Complications from PN were also observed at higher rates among Hispanics compared to non-Hispanic whites.
Disparities may arise at one or more of several steps required for the initiation and maintenance of PN. The first challenge is the recognition and evaluation of protein-calorie malnutrition. Racial disparities have been described in nutritional evaluation for other chronic diseases such as diabetes mellitus (12). The longer time interval to initiation of PN among African Americans and Hispanics supports the possibility that recognition of malnutrition may be more delayed in minorities.
The placement of indwelling intravenous catheters for PN is a commonly performed procedure and thus access to surgeons and interventional radiologists is not a rate-limiting factor. However, there may be racial differences in access to hospitals that have ancillary multidisciplinary nutrition management teams that are crucial for both the evaluation and continued management of malnutrition. In addition, patients who are started on PN require supportive care to maintain their venous catheters and nutrition infusions. This posthospitalization long-term care requires an infrastructure for collaboration with home care agencies that may not be widely available. In our study, patients hospitalized in urban settings were more likely to receive parenteral nutrition, though this difference was only marginally significant. Complications from PN, particularly catheter-related infections, are reduced when nurse clinical nutritionists or dedicated nutrition support teams participate in patient care (13;14). Thus, the higher rate of catheter infections among Hispanics receiving parenteral nutrition may also suggest lower access to a nutritional support infrastructure.
Differential access to hospitals that have the infrastructure and multidisciplinary resources for PN may be a common contributing factor to both racial and geographic variations PN utilization. The Northeast was four-fold more likely to use PN than the West. These regional disparities in the management of malnutrition may also reflect differences in practice patterns and physician preferences toward procedural interventions between different geographic areas of the country. The Northeast not only had the highest rate of PN but also the lowest rate of colectomy (6). There are potential financial barriers to receiving nutritional support as evidenced by our findings of less frequent PN among those who resided in lower income neighborhoods and those who were uninsured. In-hospital charges are nearly twice as high for IBD patients receiving PN compared to those who did not, even after adjustment for demographic and clinical factors (15). These findings likely underestimate the true economic consequences associated with PN because they do not include outpatient costs associated with home care.
The analysis of administrative datasets has several limitations. First, the study relies on identification of IBD cases and PN utilization through ICD-9-CM coding which we are unable to validate by chart review. However, we would expect any coding inaccuracies would be non-differential with respect to race. In most cases, this non-differential misclassification leads to attenuation and masking of observed effects, but should not compromise inferences from associations that are actually detected (16). In addition, though malnutrition is a clinical indication for initiation of PN, such decisions may also be influenced by disease severity. Though NIS data allow us to assess comorbidity, we cannot assess disease-specific measures of severity for either Crohn’s disease or ulcerative colitis. It is also possible that if hospitalized African Americans who are malnourished may be better able to tolerate enteral feeds and require PN less often. Unfortunately, enteral feeding is not reliably captured in the NIS dataset because it is infrequently billed as a procedure. Thus, this interesting hypothesis cannot be answered by administrative data and can only be addressed by prospectively collected data.
We have provided exploratory evidence for racial and geographic disparities in the use of parenteral nutrition for hospitalized IBD patients with malnutrition that may have implications for inequities in delivery of IBD healthcare. The impact of these disparities on health outcomes is unknown, especially since the efficacy of parenteral nutrition in IBD remains unproven (17;18). This study likely underestimates the extent of racial disparities since our study population contained individuals who already had a diagnosis of malnutrition. It is possible that there may be greater disparities in the recognition and diagnosis of malnutrition. Future studies must collect prospective data to assess disparities in diagnosis of nutritional deficiencies as well as delineate the mechanisms of racial and geographic variation in PN implementation. The ultimate goal would be to establish multidisciplinary nutritional intervention teams that would eliminate health disparities in treatment of IBD.
This work is supported by an AGA Research Scholar Award by the Foundation for Digestive Health and Nutrition (G. C. N. ).