In recent decades, research has generated an enormous growth of medical science, technology, and therapeutics. Knowledge from basic research, translational research, randomized clinical trials, and outcomes research has enabled experts in many fields to develop and disseminate evidence-based clinical practice guidelines with recommendations for medical practitioners. Yet health services research suggests that health care could perform a great deal better than it does today (1). For example, an audit of medical records of 4000 adults in 12 cities in the United States showed that only 55% of recommended preventive, acute, and chronic care was being received (2). A study of 3000 hospitals found that recommended care was provided to a large majority of adult patients for only 5 of 10 measures (3). Deficits in the quality of ambulatory care provided to children appear to be similar in magnitude to those reported for adults (4). The National Scorecard on US Health System Performance, an assessment of health care outcomes, quality, access, equity, and efficiency, found that the United States achieves an average score of only 66%. If the United States improved performance in key areas, it could save an estimated 100,000 to 150,000 lives and $50 to $100 billion annually (5).
Despite advances in technology and medications, there is evidence that Crohn disease outcomes have not improved. A systematic review evaluating mortality, cancer, disease recurrence, extraintestinal manifestations, and medication use failed to show hard evidence of improvement in disease outcome in Crohn disease during the past 4 decades (6). Furthermore, despite advances in therapy, inflammatory bowel disease (IBD) hospitalization and surgery rates in the United States between 1990 and 2003 did not decrease, and there was a significant increase in hospitalizations for Crohn disease (7).
One of the barriers to improved outcomes is unnecessary variation in care. Wide variation in the use of a diagnostic test or therapy may indicate the test or therapy is being underused, overused, or misused (8). In Crohn disease, failing to perform a colonoscopy with biopsy to establish the diagnosis or neglecting to identify and treat weight loss are potential examples of underuse. An example of overuse is prolonged treatment with prednisone. An example of misuse is prescribing infliximab to a patient with tuberculosis. Knowledge about the types of variation in disease management is the basis for initial efforts to improve care and outcomes (9–11).
The primary aims of this study were to characterize and measure the variation in care as currently provided by a broad sample of pediatric gastroenterologists and to use variation-in-care data to identify diagnostic tests and therapies that may be underused or overused. The primary hypothesis was that there is wide variation in the use of diagnostic tests and therapies owing to underuse or overuse.
We conducted cross-sectional analyses at the time patients were being enrolled in a study of children and adolescents with Crohn disease who were initiating treatment for the first time with either thiopurine or infliximab. The patients were divided into 2 groups. One group of patients was initiating therapy with a thiopurine (6-mercaptopurine or azathioprine), and the other group was initiating therapy with infliximab.
The inclusion criteria were patients 1 to 17 years old with Crohn disease, diagnosed by a pediatric gastroenterologist based on clinical, radiographic, endoscopic, and pathological findings, who were about to begin treatment with either thiopurine or infliximab. Exclusion criteria for the thiopurine group were previous treatment with a thiopurine or current treatment with infliximab. The exclusion criterion for the infliximab group was previous treatment with infliximab (patients could be treated concurrently with a thiopurine). Approval of the protocol by the local institutional review board, parental consent, and patient assent were obtained.
All pediatric gastroenterologists who were members of the North American Society for Pediatric Gastroenterology, Hepatology and Nutrition (NASPGHAN) were invited to participate in the study. Each network physician was asked to recruit and enroll up to 4 consecutive patients with Crohn disease, 2 or 3 initiating treatment with thiopurine and 1 or 2 initiating treatment with infliximab, providing care as usual.
We examined variation in care in 4 areas: (1) diagnostic tests that had been performed at the time of diagnosis to establish the diagnosis of Crohn disease and to assess the phenotype, extent, and severity of disease: a complete blood count and an erythrocyte sedimentation rate, a stool test for pathogens, small bowel imaging by either an upper gastrointestinal and small bowel series radiograph or a computerized tomography scan with contrast, and endoscopy and colonoscopy with biopsy. (2) Thiopurine (6-mercaptopurine and azathiopurine) treatment: measurement of the thiopurine methyltransferase (TPMT) phenotype or genotype before initiating thiopurine treatment to determine the dose of thiopurine at the onset of treatment and the initial dose of thiopurine prescribed. (3) Infliximab treatment: a skin test and/or chest radiograph to screen for tuberculosis before initiating treatment with infliximab and the initial dose of infliximab administered. (4) Nutritional therapy: the administration of supplements of vitamins, minerals, and formula and of nasogastric tube feedings.
The data reported here were obtained at the time patients were enrolled and included demographic information about the patient, characterization of the nature of the disease, disease activity, prior diagnostic evaluation, current treatment, and the indication for initiating treatment with the drug. Data were collected using an Internet-enabled secure system. Before enrollment of patients, the Web-based data form was tested for completeness, correctness, functionality, and ease of use. At the time of enrollment, data were entered for each patient including age, sex, race, ethnicity, type of medical insurance, date of diagnosis, location of disease, tests that had been performed at any time before enrollment, symptoms, physical findings, body weight and height, current medications, tests performed at the time of enrollment, the Physician Global Assessment, indications for treatment, and the starting doses of thiopurine and infliximab. Disease activity was assessed by the Physician Global Assessment and the Harvey Bradshaw Index (12). The Physician Global Assessment is a subjective measure in which the physician classifies the disease status as 1 of 4 categories: inactive, mild, moderate, or severe. Body weight was defined as satisfactory, underweight (mild subpar), or severely underweight (weight loss >10%) compared with the previous highest weight, excluding intentional or prednisone-related weight loss.
To characterize variations in current care delivery approaches, we computed point estimates and standard deviation for patient characteristics (age, sex, ethnicity, race, and medical insurance), disease characteristics (extent and phenotype of disease, disease duration and severity, nutritional and growth status, and clinical course before initiating therapy), and care delivery approaches (diagnostic testing, current medications, and indications for treatment). The least variation is present when essentially all or none of the patients have an intervention; the most variation is present when 50% of patients have an intervention. Characteristics of the 2 groups were compared using the Mann-Whitney test for continuous variables, such as duration, and ordinal variables, such as the Harvey Bradshaw Index. Variables expressed in percentages were compared using the χ2 test.
There were 246 patients enrolled by 93 pediatric gastroenterologists at 48 practice sites (44 in the United States, 3 in Canada, and 1 in Australia). The sites were both large and small and included both university and private practices. Patients were enrolled between October 2004 and July 2006. There were 146 patients in the group initiating treatment with thiopurine and 100 patients in the group initiating treatment with infliximab. There were no significant differences between the patients in the 2 groups in demographic information or in the extent or the phenotype of disease: 58% were male, 85% were white, the age at enrollment was 13.3 ± 2.9 (mean ± SD, median 13.7, range 3.6–17.8 years), 93% had health insurance, including 12% with Medicaid.
The distribution of disease according to the Montreal classification (13) was as follows: 12% involved the ileum but not the colon (L1); 23% involved the colon but not the ileum (L2); 62% involved both the ileum and colon (L3); and 3% involved only the upper gastrointestinal tract (L4). By the Montreal classification, 81% were nonstricturing, nonpenetrating (B1); 17% were stricturing (B2); and 2% were penetrating (B3). There was perianal disease in 30% and extraintestinal disease in 5%.
The body weight was satisfactory in 44%, underweight (mild subpar) in 36%, and severely underweight (weight loss >10%) in 19% of patients. Assessment of growth status was not performed. The duration of Crohn disease at the time of enrollment was 6.8 ± 17.6 (median 1, range 0–154) months in the thiopurine group and 22.2 ± 23.0 (median 14, range 0–136) months in the infliximab group (P < 0.001). The indications for initiating treatment with a thiopurine or infliximab are shown in Table 1 and disease severity at the time of enrollment is shown in Table 2. Patients in the infliximab group had a longer duration of disease and more severe disease.
Variation in Care
Variation in diagnostic interventions is shown in Table 3. Diagnostic interventions in which there was little or no variation were a complete blood count, erythrocyte sedimentation rate, colonoscopy with biopsy, and esophagogastroduodenoscopy with biopsy. In contrast, diagnostic interventions in which there was substantial variation in care were stool test for pathogens, upper gastrointestinal and small bowel series radiography, and TPMT. Small bowel imaging (with either radiography or CT scan) was not performed in 19% of patients.
When patients were treated with a thiopurine, there was variation in the selection of thiopurine: 70% were treated with 6-mercaptopurine and 30% were treated with azathioprine. Thiopurine methyltransferase was measured in 65% of patients before initiating thiopurine treatment in the thiopurine group and had been measured in 55% of patients in the infliximab group who were receiving thiopurine at the time infliximab treatment was initiated (Table 4). The TPMT was normal in 120 (85%) of those patients tested. Of the 120 patients in whom TPMT was normal, the dose of 6-mercaptopurine was between 0.55 and 2.78 mg/kg/day (31% of patients had a dose <1.0 mg/kg/day); the dose of azathioprine was between 0.85 and 3.45 mg/kg/day (66% had a dose <2.0 mg/kg/day). When the TPMT enzyme activity is normal, the recommended initial dose of 6-mercaptopurine is 1 to 1.5 mg/kg/day, and the recommended initial dose of azathioprine is 2 to 3 mg/kg/day (14,15). Thus, 40% of patients (48/120) were receiving doses of thiopurine that were lower than recommended. In the infliximab group, some of the patients on a maintenance dose may have had a decrease or increase in dose in response to blood concentrations of metabolites or to toxicity.
Screening for tuberculosis with either a skin test or a chest radiograph before starting treatment with infliximab was not performed in 30% of patients. There also appeared to be variation in the initial dose of infliximab. The recommended initial dose of infliximab is 5 mg/kg (14–17). The mean dose was 5.4 mg/kg, the median dose 5.1 mg/kg, and the range was 4.4 to 11.0 mg/kg. The dose was 4.4 to 5.0 mg/kg in 49% of patients, 5.1 to 6.0 mg/kg in 37% of patients, 6.1 to 7.0 mg/kg in 11%, and 10.1 to 11.0 mg/kg in 3%. Infliximab is dispensed in vials of 100 mg, and pediatric gastroenterologists commonly administer the entire contents of any vial used to reach a dose of 5 mg/kg, a practice that would result in a dose greater than 5 mg/kg. Such a practice could account for almost all of the doses above 5.0 mg/kg. Thus, only 3% of patients had an unexplained high dose of infliximab.
However, there was considerable variation in nutritional interventions: 37% of patients in both groups received a multivitamin supplement, 14% a calcium supplement, 4% a vitamin D supplement, 7% a formula supplement, and 4% tube feeding. Forty-seven patients (19%) were severely underweight (weight decreased >10%). Nutritional interventions performed in these patients are shown in Table 5; of note, 36% of patients had not received any of these interventions.
We also measured variation in other diagnostic tests. Liver tests had been performed in 95% of patients in both groups, C-reactive protein in 64%, serological tests for IBD in 36%, bone densitometry in 18%, and ophthalmologic examination in 15%.
The efficacy of 5-aminosalicylic acid medications, although well established in the treatment of ulcerative colitis, is controversial in Crohn disease, and the optimal dosage in pediatric patients is not known. In our study, 60% of the patients in both groups were treated with 5-aminosalicylic acid. The doses of sulfasalazine varied between 26 and 79 mg/kg/day; the doses of a pH 7-dependent mesalamine varied between 21 and 104 mg/kg/day; and the doses of a controlled-release mesalamine varied between 13 and 145 mg/kg/day.
The results of this study indicate there is significant variation in diagnostic and therapeutic care of children and adolescents with Crohn disease. There appears to be variation of certain diagnostic tests, including a stool test for pathogens, imaging of the small intestine, measurement of TPMT before starting treatment with a thiopurine, and testing for tuberculosis before starting treatment with infliximab. Nearly half of the patients starting treatment with a thiopurine (in the thiopurine group) and more than one quarter of patients on maintenance thiopurine (in the infliximab group) were treated with a dose of thiopurine that was less than recommended for initial treatment (14,15). In addition, more than one third of severely underweight patients were not receiving nutritional interventions.
When a test is not performed for a stool pathogen, it increases the risk of an undetected infection or misdiagnosis at onset (18,19). When small bowel imaging is not performed, the extent of disease cannot be delineated and the chance of detecting a stricture or penetrating disease is reduced, potentially leading to incorrect classification of the disease phenotype and suboptimal treatment (18–21). When patients initiating treatment with infliximab are not tested for tuberculosis, it increases the risk of the development of active tuberculous infection (15).
There is widespread recognition of the value of measuring TPMT before initiating treatment with 6-mercaptopurine or azathioprine, including by the Food and Drug Administration (22), by analyses of cost-effectiveness (23,24), by many insurance companies (25), and as part of clinical practice guidelines (14,15). When the TPMT is normal, treatment with 6-mercaptopurine can be initiated with a dose of 1 to 1.5 mg/kg/day, or azathioprine with a dose of 2 to 3 mg/kg/day. Yet only 61% of patients in the 2 groups had measurement of TPMT before thiopurine treatment. Some patients do not have TPMT testing because their medical insurance company will not pay for this test despite its importance. Nonetheless, even when the TPMT was normal, 40% of the patients were treated with a dose of thiopurine that was less than is recommended (14,15). Without receiving adequate doses, patients are significantly less likely to attain or maintain remission (26,27).
In a previous report of variation in the initial management of children with Crohn disease, there was significant variation in the frequency of use of medications, including immunomodulators, antibiotics, 5-aminosalicylates, and infliximab at the 10 centers studied (28). Another study found that adults with IBD referred for a second opinion often were not receiving optimal medical therapy (29). There was suboptimal dosing of immunomodulatory medications, prolonged use of corticosteroids, failure to use steroid-sparing agents, inadequate measures to prevent metabolic bone disease and inadequate screening for colorectal cancer.
This is the first study to report variation of both diagnostic and therapeutic interventions in pediatric IBD care and potential gaps between recommended and actual care. Strengths of this study include the participation of a large number of practice sites, providing a broad representation, and Web-based data entry that facilitated the process of data collection. Limitations of the study include the lack of verification that consecutive patients were recruited; auditing of site records was not performed. Investigation is necessary to determine to what extent a reduction in the variation in care reported here would result in better outcomes. In addition, the paucity of recommendations for standard diagnostic testing and initial evaluation limits the interpretation of the variation in testing for stool pathogens, small bowel imaging, and less commonly performed tests.
The causes of variation in the care provided are likely multifactorial. The absence of an evidence-based pediatric clinical practice guideline from a North American authority could contribute to variation in care. It is likely that the lack of a single, promulgated, widely accepted guideline with recommendations for standard diagnostic testing and initial evaluation, as well as for indications for medications and their proper dosages, is an important factor. The habits and preferences of physicians could be an important factor. Patient preferences and adherence to medication regimens, particularly when influenced by financial concerns, are also important. The refusal of medical insurance companies to pay for necessary testing is a factor. Not all diagnostic tests are necessary in some clinical situations; for example, when Crohn disease is diagnosed by gross and histopathological examination of a surgical specimen, it may not be necessary to perform a colonoscopy or to test for stool pathogens. In addition, many physicians practice in a delivery system that does not provide the structure, including the information systems and decision-making support, necessary to deliver consistent, reliable care.
Randomized trials have demonstrated that quality improvement can lead to a reduction in unnecessary variation in care and significant improvements in care and outcomes of adults with chronic illness (30). A prospective controlled cohort study of 65 adults with IBD demonstrated that practice guidelines for IBD used by physicians can reduce practice variation (31). The findings of our study indicate that there is variation in the care of children and adolescents with IBD, and gaps exist between recommended and actual care. These results suggest there is an opportunity to apply quality improvement methods at multiple institutions to evaluate the effectiveness of a pediatric IBD guideline to reduce variation and improve clinical outcomes.
All of the authors have indicated they do not have any affiliations with or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or material discussed in the manuscript, with the following exceptions: Richard B. Colletti (investigator, Abbott Laboratories; investigator and consultant, Centocor; Consultant, Accordant Health Services); Athos Bousvaros (speakers' bureau, Abbott Laboratories); Wallace Crandall (investigator and consultant, Centocor; investigator, Abbott Laboratories; speakers' bureau, TAP Pharmaceuticals); Marla Dubinsky (consultant, Prometheus).
All people named in the Acknowledgment have provided written permission to be named.
Other PIBDNet Network Physicians who contributed patient data to this study include C. Allan Pratt (Anchorage, AK), Shehzad Saeed (Birmingham, AL), Sesi Ogunbi (Montgomery, AL), Timothy Bohane (Sydney, Australia), Saied Dallalzadeh (Encino, CA), Elizabeth Gleghorn (Oakland, CA), Philip McDonald (Sacramento, CA), Melvin Heyman (San Francisco, CA), Jeffrey Hyams (Hartford, CT), Jonathan Evans (Jacksonville, FL), Alan Sacks (Pensacola, FL), Jay Hochman (Atlanta, Georgia), Claire Wilson (Honolulu, HI), Richard Sandler (Chicago, IL), Jean Molleston, Joseph Fitzgerald, Steven Steiner, Mark Corkins (Indianapolis, IN), Adrienne Scheich, Alan Leichtner, Anne Wolf, Daniel Kamin, Menno Verhave, Laurie Fishman (Boston, MA), Maria Oliva-Hemker (Baltimore, MD), Mark Integlia (Portland, ME), Pamela Brown (Ann Arbor, MI), Harold Conrad (Grand Rapids, MI), Robert Rothbaum (St. Louis, MO), Jon Vanderhoof (Omaha, NE), Jeffrey Critch (St. John's, Newfoundland), Jonathan Teitelbaum (Long Branch, NJ), Iona Monteiro (Newark, NJ), Anthony Otley (Halifax, Nova Scotia), Howard Baron (Las Vegas, NV), Abraham Jellin (Brooklyn, NY), Keith Benkov (New York, NY), John Bucuvalas (Cincinnati, OH), Carlo DiLorenzo, Hayat Mousa, Jane Balint, John Barnard, John Russo (Columbus, OH), John Grunow (Oklahoma City, OK), Sulaiman Bharwani (London, Ontario), David Mack (Ottawa, ON), David Piccoli, Edisio Semeao, Janice Kelly, Jonathan Markowitz, Joshua Friedman, Kathleen Loomes, Mei Lun Wang, Randolph Matthews, Ritu Verma (Philadelphia, PA), Manisha Dave, Alka Goyal (Pittsburgh, PA), Corey Strobel (Knoxville, TN), Abiodun Johnson (Amadillo, TX), George Ferry (Houston, TX), John Pohl (Temple, TX), Lynn Duffy, Otto Louis-Jacques, Peter Lee, Catherine Chao (Fairfax, VA), Michael Stephens, Subra Kugathasan (Milwaukee, WI).
1. Berwick DM, Nolan TW. Physicians as leaders in improving health care. Ann Intern Med 1998; 128:289–292.
2. McGlynn EA, Asch SM, Adams J, et al. The quality of health care delivered to adults in the United States. N Engl J Med 2003; 348:2635–2645.
3. Jha AK, Li Z, Orav EJ, et al. Care in U.S. hospitals—the Hospital Quality Alliance Program. N Engl J Med 2005; 353:265–274.
4. Mangione-Smith R, DeCristofaro AH, Setodji CM, et al. The quality of ambulatory care delivered to children in the United States. N Engl J Med 2007; 357:1515–1523.
6. Wolters FL, Russel MGVM, Stockbrugger RW. Systematic review: has disease outcome in Crohn's disease changed during the last four decades? Aliment Pharmacol Ther 2004; 20:483–496.
7. Bewtra M, Su C, Lewis JD. Trends in hospitalization rates for inflammatory bowel disease in the United States. Clin Gastroenterol Hepatol 2007; 5:597–601.
8. Chassin MR, Galvin RW, the National Roundtable on Health Care Quality. The urgent need to improve health care quality. JAMA 1998; 280:1000–1005.
9. Berwick DM. Controlling variation in health care. Med Care 1991; 29:1212–1225.
10. Institute of Medicine. Crossing the Quality Chasm: A New Health System for the Twenty-first Century. Washington, DC: National Academy Press; 2001.
11. Schechter MS, Margolis P. Improving subspecialty healthcare: lessons from cystic fibrosis. J Pediatr 2005; 147:295–301.
12. Harvey RF, Bradshaw JM. A simple index of Crohn's disease activity. Lancet 1980; 1:514.
13. Satsangi J, Silverberg MS, Vermeire S, et al. The Montreal classification of inflammatory bowel disease: controversies, consensus, and implications. Gut 2006; 55:749–753.
14. Lichtenstein GR, Abreu MT, Cohen R, et al. American Gastroenterological Association Institute medical position statement on corticosteroids, immunomodulators, and infliximab in inflammatory bowel disease. Gastroenterology 2006; 130:935–939.
16. Clark M, Colombel JF, Feagan BC, et al. American Gastroenterological Association Consensus Development Conference on the use of biologics in the treatment of inflammatory bowel disease, June 21–23, 2006. Gastroenterology 2007; 133:312–339.
17. Panaccione R, Fedorak RN, Aumais G, et al. Canadian Association of Gastroenterology clinical practice guidelines: the use of infliximab in Crohn's disease. Can J Gastroenterol 2004; 18:503–508.
18. IBD Working Group of the European Society for Pediatric Gastroenterology, Hepatology and Nutrition (ESPGHAN). Inflammatory bowel disease in children and adolescents: recommendations for diagnosis—the Porto criteria. J Pediatr Gastroenterol Nutr 2005;41:1–7.
19. Nikolaus S, Schreiber S. Diagnostics of inflammatory bowel disease. Gastroenterology 2007; 133:1670–1689.
20. Bousvaros A, Antonioli D, Colletti R, et al. Differentiating ulcerative colitis from Crohn's disease in children and young adults: a report of a working group of the North American Society of Pediatric Gastroenterology Hepatology and Nutrition, and the Crohn's and Colitis Foundation of America. J Pediatr Gastroenterol Nutr 2007; 44:653–674.
21. Walfish A, Sachar D. Phenotype classification in IBD: is there an impact on therapy? Inflamm Bowel Dis 2007; 13:1573–1575.
22. Package insert for azathioprine/6-mercaptopurine. Imuran®, Prometheus Laboratories, Inc., San Diego, California; 2005.
23. Winter J, Walker A, Shapiro D, et al. Cost-effectiveness of thiopurine methyltransferase genotype screening in patients about to commence azathioprine therapy for treatment of inflammatory bowel disease. Aliment Pharmacol Ther 2004; 20:593–599.
24. Dubinsky MC, Reyes E, Ofman J, et al. A cost-effectiveness analysis of alternative disease management strategies in patients with Crohn's disease treated with azathioprine or 6-mercaptopurine. Am J Gastroenterol 2005; 100:2239–2247.
26. Pearson DC, May GR, Fick G, et al. Azathioprine for maintaining remission of Crohn's disease. Cochrane Database Syst Rev 2000:CD000067.
27. Colletti RB. Next steps on the TPMT 6-TGN pathway. J Pediatr Gastroenterol Nutr 2006; 43:282–283.
28. Kappelman MD, Bousvaros A, Hyams J, et al. Intercenter variation in initial management of children with Crohn's disease. Inflamm Bowel Dis 2007; 13:890–895.
29. Reddy SI, Friedman S, Telford JJ, et al. Are patients with inflammatory bowel disease receiving optimal care? Am J Gastroenterol 2005; 100:1357–1361.
30. Tsai AC, Morton SC, Mangione CM, et al. A meta-analysis of interventions to improve care for chronic illnesses. Am J Manag Care 2005; 11:478–488.
31. Tremaine WJ, Sandborn WJ, Loftus EV, et al. A prospective cohort study of practice guidelines in inflammatory bowel disease. Am J Gastroenterol 2001; 96:2401–2406.
© 2009 Lippincott Williams & Wilkins, Inc.