Abstract
Background
Prior studies suggest that disease activity alone does not reliably predict hospital readmission among patients with inflammatory bowel diseases (IBD). Using a national database, we aimed to further describe the burden of readmissions for IBD and identify modifiable risk factors.
Methods
We performed a retrospective cohort study using 2013 data from the Nationwide Readmission Database (NRD). Using International Classification of Diseases-9th Revision, Clinical Modification (ICD-9-CM) codes, we identified adult patients with discharge diagnoses of ulcerative colitis (UC) or Crohn's disease (CD) and ascertained diagnoses of anxiety, depression, chronic pain, tobacco use, and other comorbidites during index admission. Logistic regression was used to estimate factors associated with hospital readmission.
Results
Among 52,498 hospitalizations of IBD patients (63% CD and 37% UC), 12,407 (24%) were readmitted within 90 days of the index hospitalization resulting in roughly $576 million in excess charges. In multivariable analysis of patients with CD, anxiety (OR 1.31, 95%CI:1.21–1.43), depression (OR 1.27, 95%CI:1.07–1.50), chronic pain (OR 1.31, 95%CI:1.18–1.46), and tobacco abuse (OR 1.13, 95%CI:1.06–1.22) were associated with a significant increase in odds of readmission. Among patients with UC, anxiety (OR 1.28, 95%CI:1.14–1.45), depression (OR 1.35, 95%CI:1.07–1.70), and chronic pain (OR 1.44, 95%CI:1.21–1.73) were associated with a significant increase in odds of readmission.
Conclusions
Readmission occurs frequently in patients with IBD, and is costly. Anxiety, depression, and chronic pain may represent targets for interventions to prevent 90 day hospital readmission in this population.
Keywords: anxiety, depression, Crohn's disease, ulcerative colitis, readmission
Introduction
Hospital readmissions are a key metric associated with quality of health care in the United States (US).1 Furthermore, hospital readmissions place a great burden on the health care system and patient.1 In the US, an estimated 1.2-1.4 million patients have chronic inflammatory bowel disease (IBD); including both ulcerative colitis (UC) and Crohn's disease (CD).2,3 These disorders are characterized by recurring gastrointestinal symptoms often requiring gastrointestinal procedures, intravenous fluid and medication support, hospitalization and/or surgery. Thus, those with IBD represent a significant portion of patients with gastrointestinal disorders who are hospitalized.4
Many studies evaluating readmissions in IBD have been focused on the identification of factors associated with readmission in the post-operative period, including the need for emergent surgery during the index admission and pre-operative comorbidities.5–8 The existing literature on IBD readmissions after non-surgical hospitalizations in the US includes single center experiences, where generalizability may be limited. However, these studies have suggested that poorly-controlled pain and depression,9 and lack of compliance with outpatient follow-up4 were associated with a higher likelihood of readmission. Prior work has also demonstrated that disease activity was not related to readmission within 90 days,9 suggesting the role of other, potentially modifiable factors, such as psychiatric disease, tobacco abuse, and chronic pain.
To date, there has not been an assessment of national data to assess the factors associated with readmission in patients with IBD. Our objective was to use the Nationwide Readmissions Database (NRD) to describe the overall disease burden of readmission in patients with IBD and analyze the factors associated with hospital readmission within 90 days, with a specific interest in the role that modifiable risk factors play in readmission.
Materials and Methods
Data source
We conducted a retrospective cohort study using 2013 data from the NRD, a resource of the Healthcare Cost and Utilization Project (HCUP) State Inpatient Databases. The NRD is a publicly available all-payer inpatient database from the US that includes data from 14,325,172 hospitalizations. Utilizing data from 21 states, the NRD is designed to be nationally-representative of readmission rates for all payers and the uninsured. The NRD contains more than 100 clinical and non-clinical variables,10 up to 15 diagnostic and procedure codes based on International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) coding, hospital characteristics, and outcomes including hospital length of stay (LOS), readmission, and total charges. Due to the de-identified nature of the data, this study was determined exempt from review by the University of North Carolina Institutional Review Board.
Study sample
We examined discharge data from patients age 18 and older. Any patient with an ICD-9-CM diagnosis code indicating CD (555.xx) or UC (556.xx) was eligible for inclusion in this study. To minimize misclassification of subgroups of IBD (CD or UC), we excluded patients with discharge codes for both CD and UC (n=557). The index admission was defined as the first admission where an ICD-9-CM diagnosis code for CD or UC was recorded as one of the first 3 diagnosis codes. Given our interest in evaluating 90 day readmissions, only those patients with a full 90 days between the date of discharge from their index admission and December 31, 2013 were included in these analyses.
Outcomes of Interest
The primary outcome of interest, the first readmission within 90 days of the index hospitalization, was determined by linking discharge data of individual patients using a unique identifier specified within the NRD. We also evaluated the designated reason for readmission, as defined by the primary ICD-9-CM diagnosis code and the total charges and LOS associated with readmission.
Covariates
During the index admission, multiple covariates associated with a patient's underlying IBD or perceived risk of readmission were evaluated. Patient demographics including age, sex, and primary insurance/payer for the index hospitalization. ICD-9-CM diagnosis codes were used to identify additional clinical factors suggested to be independent predictors or confounders of the relationship between IBD and hospital readmission. These included the components of the Deyo-modification of the Charlson comorbidity index, which has been validated for use as a marker of comorbid illness in studies utilizing large administrative databases.11 Given our specific interest in the role that modifiable risk factors such as depression (ICD-9-CM 296.2, 296.3, 298,0 300.4, 311), chronic pain (338.2 or 338.4), tobacco abuse (305.1), and anxiety (300, 300.2) play in readmission we made the a priori decision to include these covariates in our analyses. For the definition of anxiety and depression, we used ICD-9-CM definitions that have previously been validated in IBD using administrative data.12 Similarly, we used diagnosis coding for Clostridium difficile infection (CDI) that has previously been utilized in HCUP databases.13 All ICD-9-CM codes used in our analyses are detailed in the Supplementary Table.
We also evaluated hospital specific factors including hospital location/teaching status, as well as IBD-related hospital volume. For the evaluation of IBD-related hospital volume, we utilized previously identified cutoffs from the Nationwide Inpatient Sample (NIS) of low, medium and high centers (≤50 admissions annually, 51-150, and >150 admissions respectively).14 Due to high correlation between location/teaching status and hospital volume, only hospital location/teaching status was included in our final models. IBD-related surgery was included in the models for CD but not UC, given that patients with UC undergoing colectomy could have a scheduled readmission for repeat operation within the 90 day follow-up for readmission, with the ultimate plan for ileal pouch-anal anastomosis or other pouch operation.
Statistical Analysis
All analyses were performed using survey procedures in SAS version 9.4 (SAS Institute, Cary, NC, USA) to account for the complex survey design. Data were weighted to reflect estimates of the national population, including the means and proportions presented. Chi-Square test and Student's t-test were used to compare proportions and continuous variables respectively. We used logistic regression modeling to estimate the unadjusted odds of a readmission within 90 days of the index hospitalization for the covariates of interest. Based on clinical knowledge and previously reported risk factors for readmission, multivariable logistic regression models were then constructed to evaluate the association between risk factors and readmission within 90 days, adjusting for potential confounders. All factors were included in the model due to their clinical relevance, including payer and hospital specific factors, and all factors included in the final models were identified a priori. A two tailed p-value of 0.05 was chosen as the threshold for statistical significance for all tests. Odds ratios (OR) and 95% confidence intervals (CI) are presented.
Results
Among 52,498 patients with a diagnosis of IBD in the NRD (63% CD, 37% CD), 12,407 (24%) required readmission within 90 days of their index hospitalization. In univariate analysis, patients requiring readmission were younger, had higher Charlson comorbidity index scores, were more likely to be admitted to metropolitan teaching hospitals, and were more likely to have Medicare or Medicaid insurance (Tables 1 and 2). Additionally, patients requiring hospital readmission were more likely to have a diagnosis of anxiety, depression, chronic pain, tobacco abuse or CDI. Among patients with CD, those requiring readmission were more likely to have a diagnosis of a fistula (Table 1).
Table 1. Baseline clinical and demographic characteristics of patients with Crohn's disease in the 2013 Nationwide Readmission Database.
| No readmission n=24,076 (%) |
Readmission n=8,267 (%) |
p-value | |
|---|---|---|---|
| Age | <0.001 | ||
| 18-39 | 10,798 (45) | 4,040 (49) | |
| 40 - 59 | 7,968 (33) | 2,634 (32) | |
| ≥60 | 5,002 (21) | 1,516 (18) | |
| Female Sex | 13,370 (56) | 4,441 (54) | 0.005 |
| Anxiety | 2,678 (11) | 1,268 (15) | <0.001 |
| Depression | 592 (2) | 285 (3) | <0.001 |
| Tobacco Abuse | 4,246 (17) | 1,805 (22) | <0.001 |
| Chronic Pain | 1,434 (6) | 811 (10) | <0.001 |
| Charlson Comorbidity Score | |||
| ≤ 1 | 22,050 (92) | 7,397 (89) | <0.001 |
| 2 | 1,180 (5) | 492 (6) | |
| >2 | 846 (4) | 378 (5) | |
| Location and Teaching Status of Hospital | <0.001 | ||
| Metropolitan Non-Teaching | 9,439 (39) | 3,075 (37) | |
| Metropolitan Teaching | 12,830 (51) | 4,673 (57) | |
| Rural | 1,807 (8) | 519 (6) | |
| Payer/Insurance | <0.001 | ||
| Medicare | 5,818 (24) | 2,392 (29) | |
| Medicaid | 3,383 (14) | 1,807 (22) | |
| Private Insurance | 11,657 (48) | 2,993 (36) | |
| Self-Pay | 1,833 (8) | 618 (5) | |
| No Charge | 292 (1) | 95 (1) | |
| Other | 1,072 (4) | 346 (4) | |
| Clostridium difficile Infection | 479 (2) | 249 (3) | <0.001 |
| Need for IBD related surgery | 3,834 (16) | 941 (11) | <0.001 |
| Length of Stay for index hospitalization (days) | <0.001 | ||
| ≤3 | 12,370 (51) | 3,599 (44) | |
| 4-7 | 8,050 (33) | 2,970 (36) | |
| >7 | 3,656 (15) | 1,698 (21) | |
| Presence of a fistula (CD only) | 1,464 (6) | 643 (8) | <0.001 |
| Small bowel obstruction (CD only) | 10,608 (18) | 3,572 (17) | 0.534 |
Table 2. Baseline clinical and demographic characteristics of patients with ulcerative colitis in the 2013 Nationwide Readmission Database.
| No readmission n=16,015 (%) |
Readmission n=4,140 (%) |
p-value | |
|---|---|---|---|
| Age | 0.002 | ||
| 18-39 | 5,787 (36) | 1,584 (38) | |
| 40 - 59 | 5,044 (31) | 1,199 (29) | |
| ≥60 | 4,990 (31) | 1,297 (31) | |
| Female Sex | 8,441 (53) | 2,069 (50) | 0.004 |
| Anxiety | 1,508 (9) | 501 (12) | <0.001 |
| Depression | 368 (2) | 110 (3) | 0.023 |
| Tobacco Abuse | 1,388 (9) | 330 (8) | 0.178 |
| Chronic Pain | 606 (4) | 229 (6) | <0.001 |
| Charlson Comorbidity Score | |||
| ≤ 1 | 13,847 (86) | 3,489 (84) | <0.001 |
| 2 | 1,116 (7) | 347 (7) | |
| >2 | 1,052 (7) | 344 (8) | |
| Location and Teaching Status of Hospital | <0.001 | ||
| Metropolitan Non-Teaching | 6,409 (40) | 1,499 (36) | |
| Metropolitan Teaching | 8,583 (54) | 2,432 (59) | |
| Rural | 1,023 (6) | 209 (5) | |
| Payer/Insurance | <0.001 | ||
| Medicare | 4,402 (27) | 1,224 (30) | |
| Medicaid | 1,899 (12) | 678 (16) | |
| Private Insurance | 7,614 (48) | 1,771 (43) | |
| Self-Pay | 1,081 (7) | 203 (5) | |
| No Charge | 168 (1) | 41 (1) | |
| Other | 827 (5) | 223 (5) | |
| Clostridium difficile Infection | 772 (5) | 227 (5) | 0.056 |
| Length of Stay for index hospitalization (days) | <0.001 | ||
| ≤3 | 7,108 (44) | 1,425 (34) | |
| 4-7 | 5,830 (36) | 1,665 (40) | |
| >7 | 3,077 (19) | 1,050 (25) |
Readmission resulted in an additional 74,059 hospital days for those affected. The median LOS for the first readmission for patients with CD was 3.45 days [Interquartile range (IQR): 1.93–6.07] compared to a median LOS of 4.14 days (IQR: 2.40–7.03) for patients with UC (p=0.181). The 5 most frequent primary diagnoses for readmissions among patients with CD were CD (47%), intestinal obstruction (560.81, 560.89, 560.9, 1.7%) septicemia (ICD-9-CM 038.9, 1.2%), anal and rectal abscess (566, 0.4%), and acute kidney failure (584.9 0.3%). The 5 most frequent primary diagnoses for readmissions among patients with UC were UC (60%), septicemia (3%), intestinal obstruction (1.4%), CDI (1.2%), and acute kidney failure (0.6%).
Among patients with IBD who were readmitted within 90 days, the median total charges for the first readmission were $27,162 (IQR: $15,159–$50,606). The median total charges for the first readmission for patients with UC was significantly greater than patients with CD ($29,881 vs. $25,862, p=0.019). When aggregated among all IBD patients in the NRD, the total charges attributed to the first readmission within 90 days were $576,702,174.
Factors associated with readmission in Crohn's disease
Several factors were associated with readmission within 90 days among patients with CD, in both unadjusted and adjusted analysis. In multivariable analysis, anxiety (OR 1.18, 95% CI: 1.11–1.26), depression (OR 1.27, 95% CI: 1.07–1.50), chronic pain (OR 1.31, 95% CI: 1.18–1.46), and tobacco abuse (OR 1.13, 95% CI: 1.06–1.22) were associated with a significant increase in odds of readmission within 90 days of index hospitalization (Table 3). Additionally, younger age, diagnosis of a fistula, higher Charlson comorbidity scores, CDI, Medicare and Medicaid insurance status or self-pay status, and a LOS greater than 3 days on the index admission all were associated with increased odds of readmission within 90 days (Table 3). An IBD related surgery during the index admission was associated with decreased odds of readmission within 90 days (OR 0.56, 95% CI: 0.51-0.61). When evaluating hospital related factors, patients admitted to a rural (OR 0.75, 95% CI: 0.67–0.84) or a metropolitan non-teaching hospital (0.88, 95% CI: 0.83–0.93) were less likely to be readmitted when compared to patients admitted to metropolitan teaching hospitals.
Table 3. Unadjusted and Adjusted Analyses, odds of readmission among patients with Crohn's Disease in the 2013 Nationwide Readmission Database.
| Unadjusted OR (95% CI) |
Adjusted OR (95% CI) |
|
|---|---|---|
| Age | ||
| 18-39 | 1.14 (1.07 – 1.22) | 1.18 (1.11 – 1.26) |
| 40 - 59 | Reference | Reference |
| >59 | 0.94 (0.86 – 1.01) | 0.78 (0.71 – 0.86) |
| Female Sex | 0.93 (0.88 – 0.98) | 0.87 (0.82 – 0.92) |
| Anxiety | 1.51 (1.39 – 1.63) | 1.31 (1.21 – 1.43) |
| Depression | 1.53 (1.30 – 1.79) | 1.27 (1.07 – 1.50) |
| Tobacco Abuse | 1.30 (1.21 – 1.39) | 1.13 (1.06 – 1.22) |
| Chronic Pain | 1.69 (1.53 – 1.87) | 1.31 (1.18 – 1.46) |
| Charlson Comorbidity Score | ||
| ≤ 1 | Reference | Reference |
| 2 | 1.25 (1.11 – 1.41) | 1.17 (1.03 – 1.33) |
| >2 | 1.38 (1.21 – 1.58) | 1.30 (1.12 – 1.51) |
| Location and Teaching Status of Hospital | ||
| Metropolitan Non-Teaching | 0.87 (0.83 – 0.92) | 0.88 (0.83 – 0.93) |
| Metropolitan Teaching | Reference | Reference |
| Rural | 0.77 (0.69 – 0.85) | 0.75 (0.67 – 0.84) |
| Payer/Insurance | ||
| Medicare | 1.57 (1.47 – 1.68) | 1.64 (1.50 – 1.78) |
| Medicaid | 2.05 (1.90 – 2.21) | 1.79 (1.66 – 1.94) |
| Private Insurance | Reference | Reference |
| Self-Pay | 1.25 (1.12 – 1.39) | 1.13 (1.02 – 1.27) |
| No Charge | 1.36 (1.03 – 1.78) | 1.17 (0.89 – 1.55) |
| Other | 1.12 (0.98 – 1.29) | 1.07 (0.93 – 1.23) |
| Clostridium difficile Infection | 1.70 (1.44 – 2.02) | 1.47 (1.23 – 1.75) |
| Need for IBD related surgery | 0.67 (0.61 – 0.73) | 0.56 (0.51 – 0.61) |
| Length of Stay for index hospitalization (days) | ||
| ≤3 | Reference | Reference |
| 4-7 | 1.27 (1.19 – 1.35) | 1.31 (1.23 – 1.40) |
| >7 | 1.56 (1.44 – 1.68) | 1.62 (1.49 – 1.77) |
| Presence of a fistula | 1.32 (1.18 – 1.46) | 1.42 (1.26 – 1.59) |
All variables included in the final multivariable analysis are listed above.
Factors associated with readmission in ulcerative colitis
When patients with UC were evaluated in multivariable analysis, anxiety (OR 1.28, 95% CI: 1.14–1.45), depression (1.35, 95% CI:1.07–1.70), and chronic pain (OR 1.44, 95% CI: 1.21–1.73) were also associated with a significant increase in odds of readmission within 90 days of index hospitalization (Table 4). Younger age and lengths of stay of greater than 3 days were associated with increased odds of readmission, as were Medicare and Medicaid insurance status, and higher Charlson comorbidity scores (Table 4). Additionally, patients admitted to a rural or metropolitan non-teaching hospital were less likely to be readmitted when compared to patients admitted to metropolitan teaching hospitals (Table 4).
Table 4. Unadjusted and Adjusted Analyses, odds of readmission among patients with ulcerative colitis in the 2013 Nationwide Readmission Database.
| Unadjusted OR (95% CI) |
Adjusted OR (95% CI) |
|
|---|---|---|
| Age | ||
| 18-39 | 1.18 (1.08 – 1.30) | 1.24 (1.12 – 1.36) |
| 39 - 59 | Reference | Reference |
| >59 | 1.10 (1.00 – 1.22) | 0.98 (0.87 – 1.11) |
| Female Sex | 0.89 (0.82 – 0.96) | 0.86 (0.79 – 0.93) |
| Anxiety | 1.33 (1.18 – 1.50) | 1.28 (1.14 – 1.45) |
| Depression | 1.46 (1.17 – 1.84) | 1.35 (1.07 – 1.70) |
| Tobacco Abuse | 0.82 (0.80 – 1.05) | 0.92 (0.80 – 1.06) |
| Chronic Pain | 1.54 (1.0 – 1.83) | 1.44 (1.21 – 1.73) |
| Charlson Comorbidity Score | ||
| ≤ 1 | Reference | Reference |
| 2 | 1.18 (1.02 – 1.36) | 1.12 (0.96 – 1.31) |
| >2 | 1.39 (1.21 – 1.60) | 1.28 (1.10 – 1.48) |
| Location and Teaching Status of Hospital | ||
| Metropolitan Non-Teaching | 0.79 (0.73 – 0.86) | 0.82 (0.76 – 0.89) |
| Metropolitan Teaching | Reference | Reference |
| Rural | 0.66 (0.56 – 0.77) | 0.69 (0.59 – 0.81) |
| Payer/Insurance | ||
| Medicare | 1.14 (1.04 – 1.25) | 1.19 (1.09 – 1.35) |
| Medicaid | 1.42 (1.27 – 1.59) | 1.40 (1.25 – 1.57) |
| Private Insurance | Reference | Reference |
| Self-Pay | 0.72 (0.61 – 0.85) | 0.75 (0.63 – 0.90) |
| No Charge | 0.83 (0.58 – 1.20) | 0.83 (0.57 – 1.20) |
| Other | 1.07 (0.90 – 1.27) | 1.04 (0.88 – 1.24) |
| Clostridium difficile Infection | 1.17 (0.99 – 1.38) | 1.03 (0.87 – 1.22) |
| Length of Stay for index hospitalization (days) | ||
| ≤3 | Reference | Reference |
| 4-7 | 1.41 (1.29 – 1.54) | 1.39 (1.27 – 1.52) |
| >7 | 1.68 (1.52 – 1.86) | 1.57 (1.42 – 1.74) |
All variables included in the final multivariable analysis are listed above.
Discussion
In a nationally representative database we identified several potentially modifiable risk factors for readmission among patients with UC and CD, including concomitant diagnoses of anxiety, depression, or chronic pain conditions. Almost one fourth of patients admitted with a diagnosis of IBD in the 2013 NRD were readmitted within 90 days of their index hospitalization. Additional risk factors for readmission include markers of more complex disease such as fistulizing disease among patients with CD and an increased LOS on the index admission. Insurance or payer status was also significantly associated with increased odds of readmission, with Medicare or Medicaid insurance being significantly associated with readmission among patients with both UC and CD.
Prior single center studies have associated an increased risk of hospital readmission with factors such as chronic pain and depression.9 Abdominal pain is a common symptom among patients with IBD,15 and may be associated with IBD-irritable bowel syndrome.16,17 Although there were no available data on medication use within the NRD, a prior study evaluating narcotic pain use among inpatients with IBD demonstrated that use of narcotic pain medications was not associated with disease severity.18 Among patients with IBD, chronic pain has been associated with significant effects on quality of life, as well as functional and social outcomes.19
Chronic abdominal pain involves dysregulation of the brain-gut axis, which may lead to increased rates of anxiety and depression.20 Both anxiety and depression are increased among patients with IBD as compared to other disease states and the general population.21–23 Additionally, anxiety and depression have been associated with increased risk of clinical recurrence of disease in patients with IBD24 and higher anxiety has been associated with increased resource utilization among patients with IBD.25,26 Although the impact of these psychological comorbidities has been identified in multiple studies of patients with CD and UC, a majority of patients may not be receiving adequate psychosocial support and services as part of their overall care.27
There appears to be a significant opportunity for interventions aimed at decreasing readmissions among patients with psychosocial and chronic pain risk factors. In particular, early recognition of patients with co-existing psychiatric disease and/or chronic pain syndromes during the index admission is critical. After recognition of these comorbidities, multidisciplinary approaches including the use of inpatient case management teams, behavioral health consultations, and the integration of pain management specialists into inpatient care models may lead to better discharge planning. Additionally, early identification of responsible stakeholders after discharge including the patient's outpatient gastroenterologist, outpatient pain management specialists, and primary care provider may ultimately decrease readmission rates as well.
With 24% of patients requiring readmission within 90 days of the index hospitalization, readmission represents a significant burden to patients with IBD and the healthcare system overall. Prior single center experiences have reported 90 day readmission rates between 21% and 31% for patients with IBD.9,28 While the demonstrated rate of readmission in this study was similar, our use of a nationally representative sample is perhaps a more accurate representation of the proportion of patients with UC and CD requiring readmission within 90 days. Additionally, our findings are similar to the rate of readmission within one year reported from an analysis utilizing a Canadian national health database.29
The significant cost burden of hospital readmission has previously been demonstrated by Hazratjee, et al.4 who found that hospital readmissions within 30 days of index hospitalization increased cost of care by 111%. We also demonstrated the substantial economic impacts of readmission. When aggregated among all IBD patients in the NRD, the total charges associated with readmissions within 90 days were over $576 million. Additionally, those patients requiring readmission represented over 74,000 excess hospital days as compared to those patients who did not require readmission.
When the reason for readmission has been evaluated in earlier studies, recurrent UC or CD has been identified as the driver of readmission in 27.8 to 31.4% of cases.4,9 Small bowel obstruction has been commonly associated with readmission among patients with CD in multiple studies,4,9,30 and represented 1.7% of readmissions among patients with CD in our population. The association between infectious complications and the need for readmission among patients with IBD is significant, particularly the association between septicemia and both UC and CD readmissions. In a prior study, patients with UC and sepsis demonstrated a significantly higher mortality rate during hospitalization when compared to a control population.31 In a separate study using NIS, patients with IBD admitted to centers with a high annual volume demonstrated lower in-hospital mortality.14 In a sensitivity analysis where hospital volume was included in our multivariable models, our final OR were similar to those demonstrated in the original models.
In our analysis, younger patients demonstrated increased odds of readmission while patients over 59 years of age demonstrated the lowest odds of readmission. An increased rate of hospitalizations and readmissions among younger patients with IBD has been demonstrated in other administrative databases as well.29,32 The increase in odds of readmission among younger patients is likely multifactorial, as patients may have a more severe disease course potentially associated with an earlier onset of IBD.32 Additionally, older patients are likely to have other comorbidities, which might lead to increased follow up with primary care or other providers after discharge from and index admission. These factors would not be captured in the NRD, however, and thus we were not able to explore further.
Our study is the one of the first nationally representative study of hospital readmissions among patients with IBD. However, our study has inherent limitations that should be noted. The NRD does not include information regarding an individual patient's medication profile prior to admission or during the hospitalization. While infections such as CDI and pneumonia represented significant etiologies of readmission, we were unable to evaluate the potential relationship between steroid use or other immunosuppressive therapy and risk of subsequent infection requiring hospitalization. Additionally, the NRD does not contain laboratory data or clinical data regarding the phenotype or severity of an individual patient's IBD at the time of admission. Although we demonstrated increased odds of readmission among patients with CD and a fistula, and increased odds of readmission with longer LOS on index hospitalization, our ability to evaluate the relationship between the severity of disease and odds of readmission is limited.
With a large database such as NRD, there is potential for misclassification of those factors associated with readmission, particularly tobacco use which was associated with increased odds of readmission among patients with CD.33 There is also potential for misclassification of UC and CDI given the, given the nature of ICD-9-CM coding and the common term colitis. This misclassification is a potential explanation for the significant relationship between CDI and readmission among patients with CD but not UC. The potential exists that comorbid conditions such as anxiety, depression, and chronic pain may be less likely to be coded, due to the high likelihood that active CD or UC or related complications are driving the index admission. If this were the case, we would expect to under-estimate the burden of these comorbid diseases, even when using validated algorithms for case-finding. As such, despite demonstrating a significant increase in odds of readmission associated with these diagnoses, we may still be under-estimating their effect on readmission. In large databases, the potential also exists for statistically significant differences due to large sample sizes that may not be clinically meaningful. Multivariable analysis with logistic regression is particularly helpful in the assessment of such findings, as both the clinical significance and the magnitude of the associations demonstrated after adjusting for potential confounders must be considered. In our study, modifiable risk factors such as anxiety, depression, and chronic pain demonstrated increases in odds of readmission that are larger and likely more clinically meaningful than smaller, statistically significant associations such as the decreased odds of readmission among female patients.
The NRD also does not include information on race, which is a significant limitation as black race has been associated with shorter time to readmission among pediatric patients34 and an increased risk for readmission after colorectal surgery.35 Given the structure of the NRD, we were also unable to evaluate provider factors that may have contributed to readmission, including early discharge from the index hospitalization or treatment plans that may have led to worse outcomes including readmission. While future approaches to decreasing readmission will likely require multidisciplinary approaches, a continued focus on provider decision making during the hospitalization and discharge planning will remain a key point in improving outcomes among hospitalized patients with IBD.
In conclusion, within a nationally representative database we identified multiple risk factors for readmission among patients with UC and CD that are associated with a significant increase in charges. These include several potentially modifiable risk factors including psychiatric disease and chronic pain conditions. Earlier recognition of comorbid diagnoses of chronic pain, anxiety, and depression, along with increased use of multidisciplinary teams and care transition models, may lead to decreased hospital readmissions and improved outcomes among patients with UC and CD.
Supplementary Material
Acknowledgments
Sources of Funding: Dr. Kochar is supported by the National Institutes of Health [T32DK07634].
Dr. Crockett's effort was supported, in part, by the National Institutes of Health [KL2-RR025746].
Footnotes
Conflicts of Interest: MDL: consultant for Abbvie, Takeda, Pfizer, Theravance
ELB, BK, MDK, CFM, JRK, SDC: Each of these authors report no disclosures or conflicts of interest relevant to this study.
References
- 1.Jencks SF, Williams MV, Coleman EA. Rehospitalizations among Patients in the Medicare Fee-for-Service Program. N Engl J Med. 2009;360(14):1418–28. doi: 10.1056/NEJMsa0803563. [DOI] [PubMed] [Google Scholar]
- 2.Kappelman MD, Moore KR, Allen JK, Cook SF. Recent trends in the prevalence of Crohn's disease and ulcerative colitis in a commercially insured US population. Dig Dis Sci. 2013;58(2):519–25. doi: 10.1007/s10620-012-2371-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Long MD, Hutfless S, Kappelman MD, et al. Challenges in designing a national surveillance program for inflammatory bowel disease in the United States. Inflamm Bowel Dis. 2014;20(2):398–415. doi: 10.1097/01.MIB.0000435441.30107.8b. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Hazratjee N, Agito M, Lopez R, et al. Hospital readmissions in patients with inflammatory bowel disease. Am J Gastroenterol. 2013;108(7):1024–32. doi: 10.1038/ajg.2012.343. [DOI] [PubMed] [Google Scholar]
- 5.Frolkis A, Kaplan GG, Patel AB, et al. Postoperative complications and emergent readmission in children and adults with inflammatory bowel disease who undergo intestinal resection: a population-based study. Inflamm Bowel Dis. 2014;20(8):1316–23. doi: 10.1097/MIB.0000000000000099. [DOI] [PubMed] [Google Scholar]
- 6.Ozturk E, Kiran RP, Remzi F, Fazio VW. Early Readmission After Ileoanal Pouch Surgery. Dis Colon Rectum. 2009;52(11):1848–53. doi: 10.1007/DCR.0b013e3181b15610. [DOI] [PubMed] [Google Scholar]
- 7.Medress Z, Fleshner PR. Can we predict unplanned hospital readmission after colectomy for ulcerative colitis and indeterminate colitis? Am Surg. 2007;73(10):998–1001. [PubMed] [Google Scholar]
- 8.Sutton JM, Wima K, Wilson GC, et al. Factors Associated With 30-Day Readmission After Restorative Proctocolectomy With IPAA. Dis Colon Rectum. 2014;57(12):1371–8. doi: 10.1097/DCR.0000000000000227. [DOI] [PubMed] [Google Scholar]
- 9.Allegretti JR, Borges L, Lucci M, et al. Risk Factors for Rehospitalization Within 90 Days in Patients with Inflammatory Bowel Disease. Inflamm Bowel Dis. 2015;21(11):2583–9. doi: 10.1097/MIB.0000000000000537. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Agency for the Healthcare Cost and Resource Utilization Project. Introduction to the HCUP Nationwide Readmission Database 2013. Available at: https://www.hcup-us.ahrq.gov/db/nation/nrd/NRD_Introduction_2013.jsp.
- 11.Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol. 1992;45(6):613–9. doi: 10.1016/0895-4356(92)90133-8. [DOI] [PubMed] [Google Scholar]
- 12.Marrie RA, Walker JR, Graff LA, et al. Performance of administrative case definitions for depression and anxiety in inflammatory bowel disease. J Psychosom Res. 2016;89:107–13. doi: 10.1016/j.jpsychores.2016.08.014. [DOI] [PubMed] [Google Scholar]
- 13.Kassam Z, Cribb Fabersunne C, Smith MB, et al. Clostridium difficile associated risk of death score (CARDS): a novel severity score to predict mortality among hospitalised patients with C. difficile infection. Aliment Pharmacol Ther. 2016;43(6):725–33. doi: 10.1111/apt.13546. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Ananthakrishnan AN, McGinley EL, Binion DG. Does it matter where you are hospitalized for inflammatory bowel disease? A nationwide analysis of hospital volume. Am J Gastroenterol. 2008;103(11):2789–98. doi: 10.1111/j.1572-0241.2008.02054.x. [DOI] [PubMed] [Google Scholar]
- 15.Long MD, Drossman DA. Inflammatory bowel disease, irritable bowel syndrome, or what?: A challenge to the functional-organic dichotomy. Am J Gastroenterol. 2010;105(8):1796–8. doi: 10.1038/ajg.2010.162. [DOI] [PubMed] [Google Scholar]
- 16.Grover M, Herfarth H, Drossman DA. The Functional–Organic Dichotomy: Postinfectious Irritable Bowel Syndrome and Inflammatory Bowel Disease–Irritable Bowel Syndrome. Clin Gastroenterol Hepatol. 2009;7(1):48–53. doi: 10.1016/j.cgh.2008.08.032. [DOI] [PubMed] [Google Scholar]
- 17.Halpin SJ, Ford AC. Prevalence of symptoms meeting criteria for irritable bowel syndrome in inflammatory bowel disease: systematic review and meta-analysis. Am J Gastroenterol. 2012;107(10):1474–82. doi: 10.1038/ajg.2012.260. [DOI] [PubMed] [Google Scholar]
- 18.Long MD, Barnes EL, Herfarth HH, Drossman DA. Narcotic use for inflammatory bowel disease and risk factors during hospitalization. Inflamm Bowel Dis. 2012;18(5):869–76. doi: 10.1002/ibd.21806. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Morrison G, Van Langenberg DR, Gibson SJ, Gibson PR. Chronic Pain in Inflammatory Bowel Disease. Inflamm Bowel Dis. 2013;19(6):1210–7. doi: 10.1097/MIB.0b013e318280e729. [DOI] [PubMed] [Google Scholar]
- 20.Regueiro M, Greer JB, Szigethy E. Etiology and Treatment of Pain and Psychosocial Issues in Patients with Inflammatory Bowel Diseases. Gastroenterology. 2017;152(2):430–9. doi: 10.1053/j.gastro.2016.10.036. [DOI] [PubMed] [Google Scholar]
- 21.Kappelman MD, Long MD, Martin C, et al. Evaluation of the Patient-Reported Outcomes Measurement Information System in a Large Cohort of Patients With Inflammatory Bowel Diseases. Clin Gastroenterol Hepatol. 2014;12(8):1315–23.e2. doi: 10.1016/j.cgh.2013.10.019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Loftus EV, Guérin A, Yu AP, et al. Increased Risks of Developing Anxiety and Depression in Young Patients With Crohn's Disease. Am J Gastroenterol. 2011;106(9):1670–7. doi: 10.1038/ajg.2011.142. [DOI] [PubMed] [Google Scholar]
- 23.Mikocka-Walus A, Knowles SR, Keefer L, Graff L. Controversies Revisited. Inflamm Bowel Dis. 2016;22(3):752–62. doi: 10.1097/MIB.0000000000000620. [DOI] [PubMed] [Google Scholar]
- 24.Mikocka-Walus a, Pittet V, Rossel J, von Känel R. Symptoms of Depression and Anxiety Are Independently Associated With Clinical Recurrence of Inflammatory Bowel Disease. Clin Gastroenterol Hepatol. 2016 Jan;(25):1542–3565. doi: 10.1016/j.cgh.2015.12.045. [DOI] [PubMed] [Google Scholar]
- 25.Ramos-Rivers C, Regueiro M, Vargas EJ, et al. Association between telephone activity and features of patients with inflammatory bowel disease. Clin Gastroenterol Hepatol. 2014;12(6):986–94.e1. doi: 10.1016/j.cgh.2013.11.015. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Click B, Ramos Rivers C, Koutroubakis IE, et al. Demographic and Clinical Predictors of High Healthcare Use in Patients with Inflammatory Bowel Disease. Inflamm Bowel Dis. 2016;22(6):1442–9. doi: 10.1097/MIB.0000000000000763. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Bennebroek Evertsz' F, Thijssens NAM, Stokkers PCF, et al. Do Inflammatory Bowel Disease patients with anxiety and depressive symptoms receive the care they need? J Crohn's Colitis. 2012;6(1):68–76. doi: 10.1016/j.crohns.2011.07.006. [DOI] [PubMed] [Google Scholar]
- 28.Tinsley A, Naymagon S, Mathers B, Kingsley M, Sands BE, Ullman TA. Early readmission in patients hospitalized for ulcerative colitis: incidence and risk factors. Scand J Gastroenterol. 2015;50(9):1103–9. doi: 10.3109/00365521.2015.1020862. [DOI] [PubMed] [Google Scholar]
- 29.Bernstein CN, Nabalamba A. Hospitalization, surgery, and readmission rates of IBD in Canada: A population-based study. Am J Gastroenterol. 2006;101(1):110–8. doi: 10.1111/j.1572-0241.2006.00330.x. [DOI] [PubMed] [Google Scholar]
- 30.White E, Melmed GY, Vasiliauskas EA, et al. A prospective analysis of clinical variables, serologic factors, and outcome of ileal pouch-anal anastomosis in patients with backwash ileitis. Dis Colon Rectum. 2010;53(7):987–94. doi: 10.1007/DCR.0b013e3181dcb3f2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Colbert JF, Schmidt EP, Faubel S, Ginde AA. Severe Sepsis Outcomes Among Hospitalizations with Inflammatory Bowel Disease. SHOCK. 2016;1 doi: 10.1097/SHK.0000000000000742. [DOI] [PubMed] [Google Scholar]
- 32.Nguyen GC, Bollegala N, Chong CA. Factors Associated With Readmissions and Outcomes of Patients Hospitalized for Inflammatory Bowel Disease. Clin Gastroenterol Hepatol. 2014;12(11):1897–904.e1. doi: 10.1016/j.cgh.2014.02.042. [DOI] [PubMed] [Google Scholar]
- 33.Al Kazzi ES, Lau B, Li T, Schneider EB, Makary MA, Hutfless S. Differences in the prevalence of obesity, smoking and alcohol in the United States Nationwide Inpatient Sample and the Behavioral Risk Factor Surveillance System. PLoS One. 2015;10(11):1–11. doi: 10.1371/journal.pone.0140165. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Dotson JL, Kappelman MD, Chisolm DJ, Crandall WV. Racial disparities in readmission, complications, and procedures in children with Crohn's disease. Inflamm Bowel Dis. 2015;21(4):801–8. doi: 10.1097/MIB.0000000000000325. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Gunnells DJ, Morris MS, DeRussy A, et al. Racial Disparities in Readmissions for Patients with Inflammatory Bowel Disease (IBD) After Colorectal Surgery. J Gastrointest Surg. 2016;20:985–93. doi: 10.1007/s11605-015-3068-9. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.