Abstract
Background:
Management of inflammatory bowel diseases (IBD) is complex and variation in care has been well-documented. However, the drivers of practice variation remain unexplored. We examined variation based on the treating gastroenterologist’s IBD focus (proportion of outpatient visits for IBD).
Methods:
We conducted a retrospective cohort of newly diagnosed patients with IBD using data from Optum’s de-identified Clinformatics® Data Mart Database (2000–2020). The exposure variable was whether the treating gastroenterologist had an IBD-focus (>90th percentile of IBD visits/total outpatient visits). We used adjusted regression models to evaluate associations between provider IBD focus and process measures (use of meslamine, corticosteroid, biologic and narcotic medications and endoscopic or radiographic imaging) and clinical outcomes (time to IBD-related hospitalization and bowel resection surgery). We tested for change in treatment patterns over time by including an interaction term for study era (2004–2012 versus 2013–2020).
Results:
The study included 772 children treated by 493 providers and 2864 adults treated by 2076 providers. In children, none of the associations between provider focus and process or outcome measures were significant. In adults, care from an IBD-focused provider was associated with more use of biologics, combination therapy, and imaging and endoscopy, and less mesalamine use for Crohn’s disease (P<.05 for all comparisons) but not with other process measures. Biologics were prescribed more frequently and narcotics less frequently during the later era (P<.05 for both). Hospitalization and surgery rates were not associated with IBD-focus or era.
Conclusions:
IBD care for adults varies by provider specialization. Given the evolving complexity, novel methods may be needed to standardize care.
Keywords: practice volume, Crohn’s disease, ulcerative colitis, biologic therapy, steroids, narcotics
Graphical Abstract
Introduction
The management of Crohn’s disease (CD) and ulcerative colitis (UC), collectively referred to as inflammatory bowel diseases (IBD), has become more complex over the last several decades, with the approval of multiple new medications, development of new surgical techniques, and evolution of treatment guidelines for patients of all ages.1–6 Not surprisingly, variation in IBD care has been well-documented.7–11
Greater provider and hospital experience has been associated with adherence to guidelines and improved outcomes across many areas of medical care, including surgical and endoscopic outcomes.12–14 It is more challenging to assess the impact of a provider’s experience with non-surgical components of managing specific diseases. Evidence of improved outcomes with greater volume exists in a few medical subspecialties such as cardiology, oncology and infectious diseases.15–17 Little is known about how a gastroenterologist’s experience caring for patients with IBD is related to patterns of IBD care or health outcomes.
In this research, we sought to evaluate IBD treatment patterns and outcomes according to the experience of the treating gastroenterologist. We hypothesized that IBD-focused providers would use biologic therapies earlier in the course of disease and that this may result in less use of corticosteroids, narcotics and possibly hospitalizations and surgery, particularly in patients with CD.18
Methods
We conducted a retrospective cohort study using data from Optum’s de-identified Clinformatics® Data Mart Database (2000–2020) (CDM), a collection of anonymized patient-level insurance claims data. CDM contains data on demographics, diagnoses, procedures and dispensed medications. Data on providers includes their specialty and subspecialty and whether the service was provided by an individual provider or a facility.
Inclusion criteria
The study included patients newly diagnosed with IBD between April 30, 2004 and September 10, 2020, identified using our validated algorithm (PPV=91%).19 An incident diagnosis required the following 5 criteria be met: 1) a minimum of 4 years of follow-up prior to the first diagnosis of IBD in the claims data by any provider, 2) no prior therapy with a medication used to treat IBD unless there was another indication, such as psoriasis in a patient treated with an anti-TNF medication, 3) underwent a colonoscopy, sigmoidoscopy, capsule endoscopy or bowel resection surgery within 6 weeks prior to the first recorded diagnosis code, 4) had a second diagnosis of IBD within 12 months of the first diagnosis, and 5) the first and follow-up diagnoses were required to be from a gastroenterologist or surgeon. Exclusion criteria are described in the Supplemental Materials.
Exposure variables
The first visit with a gastroenterologist at least 90 days and not more than 365 days after the incident IBD diagnosis date was the index date. This was not necessarily the first encounter after the IBD diagnosis date. Rather, the 90-day minimum provided time for referral to the provider who would be directing the majority of the patient’s IBD care, such as the scenario where one provider made the diagnosis at colonoscopy but was not the patient’s primary gastroenterologist. The provider who saw the patient on the index date was considered the primary gastroenterologist for this study and the provider of interest even if the patient later switched to a provider with a different level of IBD focus.
The exposure variable of interest was the IBD focus of the primary gastroenterologist, defined as the proportion of the provider’s outpatient office visit encounters in the year prior to the index date that included a diagnosis of IBD. To avoid lack of precision in estimating IBD focus, we excluded patients whose gastroenterologist had fewer than 30 visits with a beneficiary of the health plans in CDM in the year prior to the index date. Thirty visits per year was the 10th percentile of annual visit volume among providers in this cohort.
Providers were divided into those with the top 10% of the IBD focus distribution (“high IBD focus”) and the remainder (“low IBD focus”). Among IBD-focused providers, the minimum proportion of office visits with an IBD diagnosis was 26% and the median was 38% (Supplemental Table 1).
Outcome measures
We examined seven process measures and two outcome measures during the follow-up period (See Supplemental Materials for details). The process measures were 1) corticosteroid use in the last 60 days of the first year of follow-up, 2) the average daily corticosteroid dose during the first year of follow-up, 3) time to first prescription biologic therapy which included antibodies to tumor necrosis factor alpha (anti-TNF), vedolizumab, ustekinumab, and tofacitinib, 4) time to first narcotic medication, 5) use of combination anti-TNF and immunomodulator medications (thiopurine or methotrexate) in the first year, 6) use of mesalamine (5ASA) for Crohn’s disease in the first year, and 7) time to first diagnostic imaging (CT or MRI of the abdomen and pelvis, colonoscopy or flexible sigmoidoscopy). The outcome measures were time to first IBD-related hospitalization and bowel resection surgery.
Covariates
We measured covariates that may be related to either the exposure or outcome variables. These included demographics, insurance type, markers of healthcare utilization, calendar year, comorbidities and markers of disease severity. See Supplemental Materials for additional details.
Statistical analysis
Separate analyses were conducted for children (age<20 on index date) and adults. We used descriptive statistics, chi-squared test and Wilcoxon rank sum test to compare the characteristics of patients cared for by high versus low IBD focus providers. We used logistic regression with forward and backwards stepwise variable selection to identify factors independently associated with receiving care from IBD-focused providers. See Supplemental Materials for additional details.
Time-to-event outcomes were assessed using Cox regression with robust variance estimates clustering on the provider. Patients were censored at the outcome of interest, death, disenrollment, or end of the study period (December 31, 2020). Steroid use in the last 60 days of the first year was analyzed with logistic regression using generalized estimating equations in which provider was considered a random effect. To compare the mean daily steroid dose in the first year of follow-up we used Poisson regression accounting for those with incomplete follow-up. All models were adjusted for the factors that were independently associated with receiving care from a IBD-focused provider. To assess each of the associations separately for CD and UC, an interaction term between provider IBD focus and IBD type was added to each of the models used in the primary analysis. Similarly, we assessed the impact of time by including an interaction term for era, dividing the study period into two eras, 2004–2012 and 2013–2020. A sensitivity analysis was performed classifying high IBD-focused providers as those with >50% of office visits with an IBD diagnosis. A type 1 error rate of 5% was used for the primary analyses and 10% to test for interaction by disease type and time.
Results
There were 772 children included in the cohort who were treated by 493 providers and 2864 adults under the care of 2076 providers (Figure 1). The providers caring for children had a median (interquartile range, IQR) of 93 (56–166) encounters in the prior year recorded within CDM. Among providers caring for adults, the median (IQR) was 137 (78–237) encounters in the prior year. Lower volume providers were more likely to have a high IBD focus (Supplemental Figure 1). On the index date, 65% of children and 71% of adults were under the care of the same provider who performed the diagnostic lower endoscopy as part of the diagnosis of IBD. Among children, 58% were diagnosed with CD (Table 1); among adults, 32% were diagnosed with CD (Table 2). The median time from the diagnostic colonoscopy to the index date was approximately 130 days in children and adults and was not significantly different between the groups, however 91% of included patients had at least one visit with a gastroenterologist between the date of the diagnostic endoscopy and the index date (Tables 1 and 2). Median (IQR) post-index follow-up time was 871 (378–1806) days among children and 815.5 (344–1707) days among adults. Among the 77% of patients with at least one additional gastroenterology office visit between the index date and the end of the first year of follow-up, 89% had at least one office visit with the index provider during the first year. See Supplemental Materials for additional details.
Figure 1.
Derivation of the study cohort.
Table 1.
Characteristics of the pediatric cohorta
| Cared for by high case mix provider (n=110) | Cared for by low case mix provider (n=662) | P value | |
|---|---|---|---|
| Age (years) | 0.38 | ||
| < 11 | 18 (16.4) | 83 (12.5) | |
| 11–15 | 44 (40.0) | 249 (37.6) | |
| 16–19 | 48 (43.6) | 330 (49.8) | |
| Female sex | 40 (36.4) | 297 (44.9) | 0.10 |
| Race/ethnicity | 0.52 | ||
| White | 87 (79.1) | 500 (75.5) | |
| Black | 6 (5.5) | 38 (5.7) | |
| Asian | -- | 20 (3.0) | |
| Hispanic | -- | 43 (6.5) | |
| Other | 9 (8.2) | 61 (9.2) | |
| Region | <.0001 | ||
| Midwest | 49 (44.5) | 176 (26.6) | |
| Northeast | 21 (19.1) | 74 (11.2) | |
| South | 33 (30.0) | 334 (50.5) | |
| West | 7 (6.4) | 78 (11.8) | |
| Household income | 0.20 | ||
| <$50,000 | 6 (5.5) | 48 (7.3) | |
| $50,000-$99,999 | 18 (16.4) | 134 (20.2) | |
| >$100,000 | 73 (66.4) | 368 (55.6) | |
| Unknown | 13 (11.8) | 112 (16.9) | |
| IBD diagnosis | 0.04 | ||
| Crohn’s disease | 78 (70.9) | 401 (60.6) | |
| Ulcerative colitis | 32 (29.1) | 261 (39.4) | |
| Year of diagnosis | 0.69 | ||
| 2004–2010 | 30 (27.3) | 197 (29.8) | |
| 2011–2015 | 38 (34.5) | 240 (36.3) | |
| 2016–2020 | 42 (38.2) | 225 (34.0) | |
| Time from diagnosis to index date (days)g | 132.5 (111.0–163.0) | 125.0 (105.0–160.0) | 0.12 |
| Follow-up time (days) | 816.0 (312.0–1632) | 877.5 (392.0–1824) | 0.30 |
| Comorbid conditions | |||
| Asthma | 24 (21.8) | 82 (12.4) | 0.01 |
| Diabetes mellitus | -- | -- | 0.003 |
| Depression | -- | 38 (5.7) | 0.61 |
| Anxiety disorder | 11 (10.0) | 48 (7.3) | 0.31 |
| Anemia | 38 (34.5) | 175 (26.4) | 0.08 |
| Malnutrition | 35 (31.8) | 201 (30.4) | 0.76 |
| Number of Elixhauser comorbidities | 0.26 | ||
| 0 | 79 (71.8) | 489 (73.9) | |
| 1 | 13 (11.8) | 90 (13.6) | |
| 2 | -- | 39 (5.9) | |
| 3+ | 13 (11.8) | 44 (6.6) | |
| Parenteral nutritionb | 8 (7.3) | 22 (3.3) | 0.05 |
| Intravenous ironb | 6 (5.5) | 12 (1.8) | 0.02 |
| Antibiotic usec | 13 (11.8) | 61 (9.2) | 0.39 |
| Antifungal usec | -- | -- | 0.68 |
| Antiviral usec | -- | -- | 0.56 |
| Allopurinol | -- | -- | 0.41 |
| CT scan, MRI or SBFTd,g | 12 (10.9) | 78 (11.8) | 0.79 |
| Lower endoscopye,g | -- | 14 (2.1) | 0.69 |
| Capsule endoscopye,g | -- | -- | 0.10 |
| Hospitalizationf | 36 (32.7) | 227 (34.3) | 0.75 |
| ED visitf | 16 (14.5) | 101 (15.3) | 0.85 |
| Visits with provider other than gastroenterologistsf | 6.0 (3.0–9.0) | 5.0 (3.0–8.0) | 0.03 |
| Non-IBD medicationsb | 3.0 (1.0–5.0) | 3.0 (2.0–5.0) | 0.52 |
| Narcotic medicationd | 9 (8.2) | 49 (7.4) | 0.77 |
| Oral corticosteroidsd | 39 (35.5) | 291 (44.0) | 0.10 |
| Rectal corticosteroidsd | -- | 12 (1.8) | 0.52 |
| Average prednisone equivalents per day dispensed from diagnosis to index date (mg) | 6.5 (0.0–19.0) | 9.1 (0.0–19.3) | 0.26 |
| Influenza vaccinef | 48 (43.6) | 249 (37.6) | 0.23 |
Pediatric grouping includes patients aged under 20 years on index date. Categorical variables reported as n (%); continuous variables reported as median and interquartile range unless otherwise noted
In 6 months prior to index date
In 1 month prior to index date
In 3 months prior to index date
In 6 weeks prior to index date
In 1 year prior to index date
Variable not included in analysis of factors independently associated with receiving care from a high-case mix provider
CT – computed tomography; ED – emergency department; IBD – inflammatory bowel disease; MRI – magnetic resonance imaging, SBFT – small bowel follow-through study Cell sizes less than 6 are suppressed to preserve confidentiality
Table 2.
Characteristics of the adult cohorta
| Cared for by high case mix provider (n=254) | Cared for by low case mix provider (n=2610) | P value | |
|---|---|---|---|
| Age (years) | <.0001 | ||
| 20–34 | 105 (41.3) | 682 (26.1) | |
| 35–49 | 79 (31.1) | 797 (30.5) | |
| 50–64 | 39 (15.4) | 651 (24.9) | |
| 65+ | 31 (12.2) | 480 (18.4) | |
| Female sex | 126 (49.6) | 1,287 (49.3) | 0.93 |
| Race/ethnicity | 0.64 | ||
| White | 206 (81.1) | 2,038 (78.1) | |
| Black | 15 (5.9) | 206 (7.9) | |
| Asian | 11 (4.3) | 95 (3.6) | |
| Hispanic | 12 (4.7) | 159 (6.1) | |
| Other | 10 (3.9) | 112 (4.3) | |
| Region | <.0001 | ||
| Midwest | 138 (54.3) | 657 (25.2) | |
| Northeast | 25 (9.8) | 278 (10.7) | |
| South | 66 (26.0) | 1,252 (48.0) | |
| West | 25 (9.8) | 423 (16.2) | |
| Insurance | 0.007 | ||
| Traditional commercial | 234 (92.1) | 2,248 (86.1) | |
| Medicare Advantage | 20 (7.9) | 362 (13.9) | |
| Household income | 0.14 | ||
| <$50,000 | 40 (15.7) | 446 (17.1) | |
| $50,000-$99,999 | 67 (26.4) | 757 (29.0) | |
| >$100,000 | 123 (48.4) | 1,079 (41.3) | |
| Unknown | 24 (9.4) | 328 (12.6) | |
| IBD diagnosis | 0.07 | ||
| Crohn’s disease | 98 (38.6) | 822 (31.5) | |
| Ulcerative colitis | 141 (55.5) | 1,627 (62.3) | |
| IBD not otherwise specified | 15 (5.9) | 161 (6.2) | |
| Year of diagnosis | 0.12 | ||
| 2004–2010 | 80 (31.5) | 961 (36.8) | |
| 2011–2015 | 90 (35.4) | 929 (35.6) | |
| 2016–2020 | 84 (33.1) | 720 (27.6) | |
| Time from diagnosis to index date (days)g | 132.5 (109.0–182.0) | 135.0 (109.0–188.0) | 0.98 |
| Follow-up time (days) | 820.5 (366.0–1651) | 815.5 (340.0–1717) | 0.99 |
| Comorbid conditions | |||
| Asthma | 20 (7.9) | 230 (8.8) | 0.61 |
| Diabetes mellitus | 15 (5.9) | 246 (9.4) | 0.06 |
| Congestive heart failure | 2 (0.8) | 75 (2.9) | 0.05 |
| Depression | 30 (11.8) | 277 (10.6) | 0.56 |
| Anxiety disorder | 25 (9.8) | 256 (9.8) | 0.99 |
| Anemia | 63 (24.8) | 470 (18.0) | 0.01 |
| Malnutrition | 31 (12.2) | 227 (8.7) | 0.06 |
| Number of Elixhauser comorbidities | 0.33 | ||
| 0 | 180 (70.9) | 1,973 (75.6) | |
| 1 | 18 (7.1) | 180 (6.9) | |
| 2 | 19 (7.5) | 150 (5.7) | |
| 3+ | 37 (14.6) | 307 (11.8) | |
| Parenteral nutritionb | -- | 22 (0.8) | 0.08 |
| Intravenous ironb | 8 (3.1) | 34 (1.3) | 0.02 |
| Antibiotic usec | 13 (5.1) | 216 (8.3) | 0.08 |
| Antifungal usec | -- | 25 (1.0) | 0.37 |
| Antiviral usec | -- | 22 (0.8) | 0.44 |
| Allopurinol | -- | 13 (0.5) | 0.26 |
| CT scan, MRI or SBFTd,g | 23 (9.1) | 212 (8.1) | 0.61 |
| Lower endoscopye,g | 6 (2.4) | 75 (2.9) | 0.64 |
| Capsule endoscopye,g | -- | 8 (0.3) | 0.81 |
| Hospitalizationf | 82 (32.3) | 657 (25.2) | 0.01 |
| ED visitf | 68 (26.8) | 538 (20.6) | 0.02 |
| Visits with provider other than gastroenterologistsf | 4.0 (2.0–7.0) | 4.0 (2.0–7.0) | 0.72 |
| Non-IBD medicationsb | 4.0 (2.0–7.0) | 5.0 (2.0–8.0) | 0.003 |
| Narcotic medicationd | 28 (11.0) | 342 (13.1) | 0.35 |
| Oral corticosteroidsd | 88 (34.6) | 793 (30.4) | 0.16 |
| Rectal corticosteroidsd | 7 (2.8) | 51 (2.0) | 0.39 |
| Average prednisone equivalents per day dispensed from diagnosis to index date (mg) | 2.0 (0.0–15.5) | 0.0 (0.0–11.0) | 0.01 |
| Influenza vaccinef | 97 (38.2) | 640 (24.5) | <.0001 |
Includes patients aged 20 years and older on index date. Categorical variables reported as n (%); continuous variables reported as median and interquartile range unless otherwise noted
In 6 months prior to index date
In 1 month prior to index date
In 3 months prior to index date
In 6 weeks prior to index date
In 1 year prior to index date
Variable not included in analysis of factors independently associated with receiving care from a high-case mix provider
CT – computed tomography; ED – emergency department; IBD – inflammatory bowel disease; MRI – magnetic resonance imaging, SBFT – small bowel follow-through study
Cell sizes less than 6 are suppressed to preserve confidentiality
Variables independently associated with receipt of care from a IBD-focused provider differed for children and adults. Geography, comorbidities, age, markers of healthcare utilization and insurance plans were associated with provider type (Supplemental Table 2).
Medication use
A trend in which children who were under the care of a IBD-focused provider received less corticosteroids during the first year of follow-up when measured as the average daily steroid dose was not statistically significant (adjusted IRR 0.66, 95% CI 0.44–1.04) The proportion with a steroid prescription in the last 60 days of the first year of follow-up did not differ (IBD-focused 13.0%, non IBD-focused 12.6%, OR 1.03, 95% CI 0.50–2.13). Receipt of care from an IBD-focused provider was not associated with steroid use in adults (last 60 days of first year IBD-focused 13.6%, non IBD-focused 13.2%, OR 0.93, 95% CI 0.59–1.49; average daily dose in first year adjusted (IRR 0.95, 95% CI 0.73–1.23) (Table 3).
Table 3.
Comparison of treatment patterns and outcomes among children and adults treated by high and low case mix providersa
| Outcome | Unadjusted HR, IRR or OR (95% CI)b | Adjusted HR, IRR or OR (95% CI)b, c,d |
|---|---|---|
| Children | ||
| Steroid prescription in last 60 days of first year | 1.04 (0.51–2.10) | 1.03 (0.50–2.13) |
| Average daily steroid dose in first year | 0.58 (0.37–0.91) | 0.68 (0.44–1.04) |
| Time to first biologic | 1.06 (0.81–1.39) | 0.97 (0.75–1.27) |
| Treatment with combination therapy in first yeare | 1.09 (0.56–2.15) | 0.99 (0.48–2.08) |
| Time to first narcotic prescription | 0.85 (0.59–1.25) | 0.94 (0.64–1.39) |
| Treatment with 5ASA for Crohn’s disease in first year | 0.52 (0.25–1.09) | 0.64 (0.30–1.37) |
| Time to first IBD hospitalization | 1.07 (0.69–1.66) | 1.02 (0.65–1.60) |
| Time to first surgery | 1.00 (0.44–2.29) | 0.97 (0.41–2.33) |
| Time to first diagnostic imagingf | 0.99 (0.74–1.33) | 0.99 (0.73–1.34) |
| Adults | ||
| Steroid prescription in last 60 days of first year | 1.04 (0.67–1.61) | 0.93 (0.59–1.49) |
| Average daily steroid dose in first year | 1.06 (0.82–1.36) | 0.95 (0.73–1.23) |
| Time to first biologic | 2.06 (1.68–2.51) | 1.52 (1.23–1.88) |
| Treatment with combination therapy in first yeare | 2.56 (1.68–3.89) | 1.84 (1.16–2.90) |
| Time to first narcotic | 0.82 (0.67–1.01) | 0.83 (0.68–1.02) |
| Treatment with 5ASA for Crohn’s disease in first year | 0.39 (0.23–0.67) | 0.43 (0.24–0.76) |
| Time to first IBD hospitalization) | 1.07 (0.77–1.50) | 0.96 (0.67–1.36) |
| Time to first surgery | 1.31 (0.80–2.13) | 1.13 (0.68–1.88) |
| Time to first diagnostic imagingf | 1.43 (1.20–1.70) | 1.34 (1.13–1.60) |
Low case mix providers are the reference group
Reported values are OR for steroid prescription in last 60 days of the first year, 5ASA use in the first year and combination therapy use in the first year, IRR for average steroid dose per day in the first year, and HR for time to event outcomes. The exposure variable is treatment by an IBD-focused provider.
Age < 20 years - Adjusted for region, UC versus CD, sex, asthma, parenteral nutrition in the 6 months prior to index date, oral steroids in the 3 months prior to index date and number of non-gastroenterologist office visits
Age 20 years and older - Adjusted for age, region, insurance type, anemia, antibiotic use in the month prior to index date, hospitalization in year prior to index date, emergency department visits in year prior to index date, influenza vaccine
Anti-TNF therapy plus azathioprine, 6 mercaptopurine or methotrexate
CT or MRI of abdomen and pelvis, colonoscopy or flexible sigmoidoscopy
HR – hazard ratio; IRR – incidence rate ratio; OR – odds ratio
The initial biologic agent was anti-TNF in 96% of children and 88% of adults (Supplemental Table 3). Overall, the 1-year cumulative incidence of biologic initiation was 45% in children and 22% in adults. Provider IBD focus was not associated with use of biologics in children (OR 0.97, 95% CI 0.75–1.27) but was positively associated in adults (OR 1.52, 95% CI 1.23–1.88) (Figure 2). Combination therapy with anti-TNF and immunomodulator was more common in adult patients care for by high IBD-focused providers (OR 1.84, 95% C 1.16–2.90) but not in children (Table 3). The time to first narcotic prescription did not differ between patients cared for by high and low IBD focus providers among children or adults (Table 3, Supplemental Figure 2). Treatment of Crohn’s disease with 5ASA in the first year was less common in children and adults cared for by high IBD-focused providers, although only statistically significant for adults (OR 0.43, 95% CI 0.24–0.76) (Table 3).
Figure 2. Time to first dispensing of a biologic therapy in (A) children and (B) adults.
Patients who received a prescription on or before the index date were categorized has having received the prescription on the first day of follow-up.
Imaging
High IBD focus providers ordered diagnostic imaging sooner for adults (HR 1.34, 95% CI 1.13–1.60) but not children (HR 0.99, 95% CI 0.73–1.34) (Table 3, Supplemental Figure 3).
Hospitalizations and surgery
The time to first hospitalization for IBD or first bowel resection surgery did not differ based on the provider IBD focus for children or adults (Table 3, Supplemental Figures 4 and 5).
Interaction between IBD focus and disease type
Inclusion of an interaction term between disease type and provider IBD focus identified differences between CD and UC in the association of provider IBD focus and hospitalization for children (interaction P values <.05, Supplemental Table 4). Among children with CD, a positive trend between provider IBD focus and hospitalization was not statistically significant (OR 1.45, 95% CI 0.88–2.40); an inverse trend for children with UC was also not statistically significant (OR 0.57, 95% CI 0.25–1.29). No other associations with IBD focus differed based on disease type.
Interaction between IBD focus and era
None of the tests of interaction of IBD focus by era were statistically significant (Supplemental Table 4). However, era was strongly associated with process measures (Supplemental Table 5). For children, rates of steroid prescribing were lower in the later period, particularly among non IBD-focused providers. For both adults and children, rates of prescribing biologics were higher and narcotics and mesalamine for CD were lower in the later period for both high and low IBD focus providers (Figure 3 and Supplemental Table 5). During the later period, biologic prescribing in the first year was approximately 60% for IBD-focused adult and all pediatric gastroenterologists compared to 29% among low case-mix adult gastroenterologists (Supplemental Table 6). Hospitalization and surgery rates did not differ by era.
Figure 3. Time to first dispensing of a biologic therapy in (A) children and (B) adults and narcotics in (C) children and (D) adults stratified by time period.
Patients who received a prescription on or before the index date were categorized has having received the prescription on the first day of follow-up.
Sensitivity analysis based on definition of high IBD focus
Using a stricter definition of high IBD focus, there were 32 (4.1%) children and 96 (3.4%) adults treated by a high IBD-focused provider. The magnitude of the significant associations observed among adults using the primary definition were stronger using the stricter definition. No new associations reached statistical significance although some associations for children were qualitatively stronger (Supplemental Table 7).
Discussion
Caring for patients with IBD is complex and evolving. We hypothesized that patients treated by IBD-focused gastroenterologists would receive different care and have better outcomes. In this study, we observed significant change in prescribing of biologics and narcotics over time and these changes were comparable among IBD-focused and non IBD-focused providers. We confirmed our hypothesis among adults but not children with IBD. IBD-focused providers treating adults used more and earlier biologic therapies, combination anti-TNF and immunomodulator therapy, diagnostic imaging and less mesalamine in patients with CD. Although there was less variability in care of children with IBD, rates of biologic and steroid prescribing were similar for children and adults cared for by IBD-focused gastroenterologists. Biologic prescribing for these patients was substantially higher than among adult patients cared for by non IBD-focused gastroenterologists in both treatment eras. Despite these associations, we did not observe significant differences in IBD-related hospitalization or surgery rates by era or provider IBD focus, perhaps due to the short follow-up period.
Strong data support better surgical outcomes with greater experience and specialization12–14, and there is some evidence that this may apply to medical decision-making.15–17 In this study, we observed differences in process measures, including prescribing of biologics, combination therapy, mesalamine for Crohn’s disease and use of diagnostic imaging, between IBD-focused and non IBD-focused providers caring for adults. Earlier use of biologic therapy has been associated with several favorable outcomes, particularly in CD. Patients with shorter duration of disease were more likely to achieve remission with biologic therapy.20 Among children, earlier initiation of anti-TNF therapy has been shown to prevent penetrating complications.21 In contrast, prolonged steroid use and narcotic use are associated with worse outcomes.22–24 Thus, greater emphasis on steroid and narcotic sparing strategies, such as with earlier prescribing of biologics, may be considered a measure of better quality of care.
We studied use of mesalamine in CD as an example of overuse, given little evidence of efficacy.25 Difference in rates of use of mesalamine for CD was among the strongest association we observed. Although not statistically significant among children, the OR was 0.64, suggesting a similar association as that seen among adults. These results also suggest greater use of evidence-based treatment algorithms by IBD-focused providers.
Biologic use was more common in children than adults, particularly in the first half of the study period (2004–2012). By the latter half (2013–2020), IBD-focused gastroenterologists caring for adults had prescribing patterns similar to pediatricians. We hypothesize that this may reflect increased adherence to a model of early intervention with the hopes of preventing disease progression and permanent bowel damage among all pediatric gastroenterologists and IBD-focused gastroenterologists that treat adults.5,6 Supporting this hypothesis was the observation that the rates of biologic prescribing in the second half of the study were 2- to 3-fold higher than in the first half (Figure 3 and Supplemental Table 5). Although we did not observe differences in IBD-related hospitalization or surgery rates based on the IBD focus of the treating provider or era despite greater biologic use, with longer follow-up, it is possible that differences in these outcomes would emerge.
Interestingly, we observed less practice variability among treatment of children than adults. We suspect this may be due to the smaller size of the pediatric gastroenterology community, the general acceptance of anti-TNF monotherapy and reluctance to use concomitant immunomodulators, and the widespread acknowledgment of the negative impact of steroids on growth. Additionally, many pediatric gastroenterologists and centers in the US are engaged in the ImproveCareNow, which may have further accelerated standardization of care.26
This study included only newly diagnosed patients, used a validated algorithm to identify these patients, and categorized IBD focus based on the provider caring for the patient on the index date, which was at least 90 days after their first diagnosis. This design, along with our adjustment for confounders, minimized the risk of bias from channeling sicker patients to more experienced providers. However, it does not completely exclude early channeling of patients who appeared particularly ill at the time of diagnosis. Indeed, the mean proportion of outpatient visits for IBD was 18% when the index provider was not the provider who performed the diagnostic colonoscopy and 11% when the index provider was the same.
While a strength of commercial claims data is complete capture of dispensed outpatient medications, hospitalizations, and surgeries, lack of information on the reason for treatment decisions could result in residual confounding. We were not able to assess how long the provider had been in practice or the IBD focus of their practice in other insurance plans. However, we observed that the provider’s IBD focus in the prior year was strongly correlated with their IBD focus using all available data (r=0.90, P<0.0001, data not shown), suggesting that IBD focus is relatively stable over time. To the extent that our definition of IBD focus was imperfect, particularly if this misclassification was non-differential, we may have obscured meaningful differences. Although this study leveraged a large administrative claims database, the power to detect effect modification by disease type and era was low. Thus, statistical interactions may have been missed. We conducted separate analyses for children and adults, but we could not determine whether providers caring for children were trained as pediatric gastroenterologists. Our specific results may not be generalizable to other countries or payers, but the general principles likely apply. Similarly, these data do not allow for identification of patients with secondary insurance, such as Medicaid insurance for children with IBD. We limited our study to patients seen by gastroenterologists; future studies will be necessary to understand care patterns and outcomes of patients cared for by advanced practice providers. Finally, our design did not account for patients switching between providers over time and may not generalize to patients with longstanding disease.
In conclusion, we demonstrated marked changes in prescribing patterns for patients with newly diagnosed IBD over the last two decades. Biologic use has increased at the same time as narcotic prescribing has decreased. Moreover, adult gastroenterologists with a high IBD focus demonstrated greater use of proven therapies and less use of unproven therapies, suggesting more evidence-based prescribing. Given the high prevalence27 and evolving complexity of IBD management, novel methods are needed to standardize evidence-based, quality care for patients with IBD. Similar studies focused on other gastrointestinal conditions are warranted to assess the impact of gastroenterologists’ clinical specialization and identify opportunities for reducing variation in care.
Supplementary Material
WHAT YOU NEED TO KNOW.
Background:
Management of inflammatory bowel disease is complex. Provider experience and case volume are important predictors of practice patterns and outcomes in many health conditions.
Findings:
Differences in practice patterns according to the extent of concentration of a physician’s practice on inflammatory bowel disease is evident for adult patients but not pediatric patients.
Implication for patient care:
Care of adult patients with inflammatory bowel disease differs by provider experience treating inflammatory bowel disease. Strategies are needed to reduce practice variation.
Financial support:
Center for Disease Control and Prevention U01-DP006369
NIH: P30-DK050306
NIH: UL1TR002489
Potential competing interests:
JDL consulted or served on an advisory board for Eli Lilly and Company, Samsung Bioepis, UCB, Bristol-Myers Squibb, Nestle Health Science, Merck, Celgene, Janssen Pharmaceuticals, Bridge Biotherapeutics, Entasis Therapeutics, AbbVie, Pfizer, Gilead, Galapagos, Sanofi, Arena Pharmaceuticals, Protagonist Therapeutics, Amgen, and Scipher Medicine. He has had research funding from Nestle Health Science, Takeda, Janssen Pharmaceuticals, and AbbVie. He has had educational grants from Takeda and Janssen. He has performed legal work on behalf of generic manufacturers of ranitidine, including L. Perrigo Company, Glenmark Pharmaceuticals Inc., Amneal Pharmaceuticals LLC, Aurobindo Pharma USA, Inc., Dr. Reddy’s Laboratories, Inc., Novitium Pharma, Ranbaxy Inc. and Sun Pharmaceutical Industries, Inc., Strides Pharma, Inc., and Wockhardt USA LLC. He owns stock in Dark Canyon Labs.
LEP has conducted studies on behalf of Sanofi and Pfizer, but they are not related to research presented in this manuscript.
MDK has consulted for Abbvie, Pfizer, Takeda and Lilly, is a shareholder in Johnson & Johnson, and has received research support from Janssen and Abbvie.
All other authors report no potential conflict of interest.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Lichtenstein GR, Loftus EV, Isaacs KL, Regueiro MD, Gerson LB, Sands BE. ACG Clinical Guideline: Management of Crohn’s Disease in Adults. Am J Gastroenterol. 2018;113(4):481–517. [DOI] [PubMed] [Google Scholar]
- 2.Rubin DT, Ananthakrishnan AN, Siegel CA, Sauer BG, Long MD. ACG Clinical Guideline: Ulcerative Colitis in Adults. Am J Gastroenterol. 2019;114(3):384–413. [DOI] [PubMed] [Google Scholar]
- 3.Feuerstein JD, Isaacs KL, Schneider Y, et al. AGA Clinical Practice Guidelines on the Management of Moderate to Severe Ulcerative Colitis. Gastroenterology. 2020;158(5):1450–1461. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Ko CW, Singh S, Feuerstein JD, et al. AGA Clinical Practice Guidelines on the Management of Mild-to-Moderate Ulcerative Colitis. Gastroenterology. 2019;156(3):748–764. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Peyrin-Biroulet L, Sandborn W, Sands BE, et al. Selecting Therapeutic Targets in Inflammatory Bowel Disease (STRIDE): Determining Therapeutic Goals for Treat-to-Target. Am J Gastroenterol. 2015;110(9):1324–1338. [DOI] [PubMed] [Google Scholar]
- 6.Turner D, Ricciuto A, Lewis A, et al. STRIDE-II: An Update on the Selecting Therapeutic Targets in Inflammatory Bowel Disease (STRIDE) Initiative of the International Organization for the Study of IBD (IOIBD): Determining Therapeutic Goals for Treat-to-Target strategies in IBD. Gastroenterology. 2020. [DOI] [PubMed] [Google Scholar]
- 7.Assa A, Avni I, Ben-Bassat O, Niv Y, Shamir R. Practice Variations in the Management of Inflammatory Bowel Disease Between Pediatric and Adult Gastroenterologists. J Pediatr Gastroenterol Nutr. 2016;62(3):372–377. [DOI] [PubMed] [Google Scholar]
- 8.Krishnakumar C, Ballengee CR, Liu C, et al. Variation in Care in the Management of Children With Crohn’s Disease: Data From a Multicenter Inception Cohort Study. Inflamm Bowel Dis. 2019;25(7):1208–1217. [DOI] [PubMed] [Google Scholar]
- 9.Lewis JD, Cross RK, Long M, et al. High Degree of Practice Variability in Colonic Dysplasia Surveillance for Patients With Inflammatory Bowel Disease. Inflamm Bowel Dis. 2022;28(8):1289–1292. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Limketkai BN, Singh S, Jairath V, Sandborn WJ, Dulai PS. US Practice Patterns and Impact of Monitoring for Mucosal Inflammation After Biologic Initiation in Inflammatory Bowel Disease. Inflamm Bowel Dis. 2019;25(11):1828–1837. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Weaver KN, Kappelman MD, Sandler RS, et al. Variation in Care of Inflammatory Bowel Diseases Patients in Crohn’s and Colitis Foundation of America Partners: Role of Gastroenterologist Practice Setting in Disease Outcomes and Quality Process Measures. Inflamm Bowel Dis. 2016;22(11):2672–2677. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Gooiker GA, van Gijn W, Post PN, van de Velde CJ, Tollenaar RA, Wouters MW. A systematic review and meta-analysis of the volume-outcome relationship in the surgical treatment of breast cancer. Are breast cancer patients better of with a high volume provider? Eur J Surg Oncol. 2010;36 Suppl 1:S27–35. [DOI] [PubMed] [Google Scholar]
- 13.Gruen RL, Pitt V, Green S, Parkhill A, Campbell D, Jolley D. The effect of provider case volume on cancer mortality: systematic review and meta-analysis. CA Cancer J Clin. 2009;59(3):192–211. [DOI] [PubMed] [Google Scholar]
- 14.Campos S, Papaefthymiou A, Florou T, et al. The impact of center and endoscopist ERCP volume on ERCP outcomes: a systematic review and meta-analysis. Gastrointest Endosc. 2023. [DOI] [PubMed] [Google Scholar]
- 15.Fleming LM, Jones P, Chan PS, Andrei AC, Maddox TM, Farmer SA. Relationship of Provider and Practice Volume to Performance Measure Adherence for Coronary Artery Disease, Heart Failure, and Atrial Fibrillation: Results From the National Cardiovascular Data Registry. Circ Cardiovasc Qual Outcomes. 2016;9(1):48–54. [DOI] [PubMed] [Google Scholar]
- 16.Landon BE, Wilson IB, Cohn SE, et al. Physician specialization and antiretroviral therapy for HIV. J Gen Intern Med. 2003;18(4):233–241. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Huntington SF, Hoag JR, Zhu W, et al. Oncologist volume and outcomes in older adults diagnosed with diffuse large B cell lymphoma. Cancer. 2018;124(21):4211–4220. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Law CCY, Tkachuk B, Lieto S, et al. Early Biologic Treatment Decreases Risk of Surgery in Crohn’s Disease but not in Ulcerative Colitis: Systematic Review and Meta-Analysis. Inflamm Bowel Dis. 2023. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Dawwas GK, Weiss A, Constant BD, et al. Development and Validation of Claims-Based Definitions to Identify Incident and Prevalent Inflammatory Bowel Disease in Administrative Healthcare Databases. Inflamm Bowel Dis. 2023;29(12):1993–1996. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Ben-Horin S, Novack L, Mao R, et al. Efficacy of Biologic Drugs in Short-Duration Versus Long-Duration Inflammatory Bowel Disease: A Systematic Review and an Individual-Patient Data Meta-Analysis of Randomized Controlled Trials. Gastroenterology. 2022;162(2):482–494. [DOI] [PubMed] [Google Scholar]
- 21.Kugathasan S, Denson LA, Walters TD, et al. Prediction of complicated disease course for children newly diagnosed with Crohn’s disease: a multicentre inception cohort study. Lancet. 2017;389(10080):1710–1718. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Lewis JD, Gelfand JM, Troxel AB, et al. Immunosuppressant medications and mortality in inflammatory bowel disease. Am J Gastroenterol. 2008;103(6):1428–1435; quiz 1436. [DOI] [PubMed] [Google Scholar]
- 23.Lewis JD, Scott FI, Brensinger CM, et al. Increased Mortality Rates With Prolonged Corticosteroid Therapy When Compared With Antitumor Necrosis Factor-alpha-Directed Therapy for Inflammatory Bowel Disease. Am J Gastroenterol. 2018;113(3):405–417. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Lichtenstein GR, Feagan BG, Cohen RD, et al. Serious Infection and Mortality in Patients With Crohn’s Disease: More Than 5 Years of Follow-Up in the TREAT Registry. Am J Gastroenterol. 2012;107(9):1409–1422. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Lim WC, Hanauer S. Aminosalicylates for induction of remission or response in Crohn’s disease. Cochrane Database Syst Rev. 2010(12):Cd008870. [DOI] [PubMed] [Google Scholar]
- 26.Pratt J, Jeffers D, King EC, et al. Implementing a Novel Quality Improvement-Based Approach to Data Quality Monitoring and Enhancement in a Multipurpose Clinical Registry. EGEMS (Wash DC). 2019;7(1):51. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Lewis JD, Parlett LE, Jonsson Funk ML, et al. Incidence, Prevalence, and Racial and Ethnic Distribution of Inflammatory Bowel Disease in the United States. Gastroenterology. 2023;165(5):1197–1205 e1192. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.