What Is Known/What Is New
What Is Known
- Paediatric-onset inflammatory bowel disease is different from adult-onset inflammatory bowel disease with respect to disease extent and severity.
- Compared with adult inflammatory bowel disease patients, adolescents have increased rates of nonadherence, hospital admissions, anxiety, depression, inadequate self-efficacy, insufficient knowledge of disease, and developmental immaturity.
What Is New
- This study shows that a significant proportion of adolescent inflammatory bowel disease patients is diagnosed and treated by adult care providers.
- In this study, paediatric care setting is associated with significantly lower steroid and biological use, and inflammatory bowel disease-related hospital admissions.
Childhood inflammatory bowel disease (IBD) differs from adult disease in disease extent and long-term treatment outcomes (1). Due to increased severity in childhood, treatment choices between immunosuppressive therapy or biologicals may have greater impact on the disease course than in adult IBD (2,3). Furthermore, compared with adult IBD patients, adolescents have increased rates of nonadherence, hospital admissions, anxiety, depression, inadequate self-efficacy, insufficient knowledge of disease, and developmental immaturity (4–7).
In chronic diseases that occur in the adolescent age group, care setting (paediatric vs adult-oriented) affects health outcomes. In acute lymphoblastic leukemia, choice of care setting is even associated with a difference in survival (8). Care for paediatric IBD patients in several European countries, such as the Netherlands, is dispersed among paediatric and adult-oriented care providers. Choice of setting appears to be based on availability rather than on needs of the developing teenagers. With respect to needs of adolescent IBD patients, paediatricians claim to have better knowledge and tools to handle the developmental aspects, whereas adult gastroenterologists claim that adolescents are better off in adult care as they generally have a significantly greater degree of exposure to IBD patients. No proof for these concepts exist.
The objective of this study was to estimate impact of paediatric versus adult care setting on healthcare utilization in adolescents with IBD. We studied differences in treatment of adolescent IBD patients by paediatric versus adult care providers, including steroid and biological use, and treatment outcomes, including IBD-related hospital admissions and IBD-related surgery.
In this study, we used anonymized data from Achmea, a Dutch health insurance firm. Enrolment was between January 1, 2007 and January 1, 2014. Achmea covered about 4.2 million persons in this period. With a Dutch population of more than 16 million persons, this amounts to approximately 25% of the Dutch population. Patients were included in this study if they had an IBD-related hospital visit or admission between 16.0 and 17.9 years of age, were continuously insured with Achmea for at least 12 months before study inclusion, and had at least 1 IBD-related pharmaceutical prescription in the 12 months following study inclusion. Patient characteristics, such as birth date and sex, could be derived from the insurance files. Insurance files were without missing data. The study was performed according to the guidelines of the ethical committee of the University Medical Center Utrecht.
Each eligible patient was included in the study only once. Patients were followed from study inclusion until switch to different health insurer; death; end of observation period (December 31, 2014); 19th birthday or transfer to adult-oriented care, whichever came first.
IBD-related hospital visits and admissions were identified from insurance claims data, which in the Netherlands are filed according to a national diagnosis-treatment classification (DTC) system. This DTC system is based on the hospital registration of diagnoses [International Classification of Diseases, 9th revision, clinical modification (ICD-9-CM) codes]. Age at study inclusion and insurance coverage preceding and following study inclusion were derived from insurance administrative records.
Identification of IBD-related pharmaceutical prescriptions was based on Anatomical Therapeutic Chemical (ATC) codes. In the ATC classification system, active substances are divided into groups according to the organ or system on which they act, and on their therapeutic, pharmacological, and chemical properties (9). We included corticosteroids, azathioprine, methotrexate, mesalazine, and biologicals (see Supplemental Digital Content, http://links.lww.com/MPG/B656, for extensive list).
Determination of care setting (paediatric vs adult-oriented) was based on the medical specialty (Paediatrics vs Internal Medicine or Gastroenterology) reported at study inclusion, which could be derived from the insurance claims. For the years 2006 to 2011, the first 4 digits of the DTC code identified the medical specialty. By the end of 2011, the DTC system underwent a major revision. For the period 2012 to 2014, we therefore used publicly available cross-walk tables provided by the Dutch Healthcare Authority (Nederlandse Zorgautoriteit) to translate the original DTC codes to the corresponding DTC codes under the revised system.
Study outcomes included time-to-first steroid prescription, time-to-first IBD-related hospital admission, time-to-first IBD-related surgery, and time-to-first use of a biological agent (anti-tumor necrosis factor [TNF] agents Infliximab or Adalimumab), as measured from the date of study inclusion. Data on these outcomes were available either for the full study period (steroid prescriptions), or limited to the years 2007 to 2011 (IBD-related hospital admissions and surgery), or to the years 2012 to 2014 (use of biological agents). Data on IBD-related hospital admissions and surgery were not available after 2011, because of the earlier mentioned major revisions of the DTC classification. Data on use of biological agents were not available before 2012, because only from 2012 onwards these were filed and reimbursed on a per-patient basis by Dutch health insurers (10).
We adjusted study outcomes for potential confounders, including age and sex at study inclusion, disease stage (newly diagnosed vs known IBD), and hospital type (general vs academic). Disease phenotype (Crohn disease [CD] or ulcerative colitis [UC]) may also act as a confounder. However, the distinction between CD or UC was unobserved for patients in the paediatric care setting (the Dutch DTC classification for this setting classifies both diseases as “IBD”). In a subgroup of these patients, the patients transferred to the adult setting before the end of the study period and disease phenotype could be inferred from the DTC codes filed in the adult setting. We so could compare the proportion of disease phenotypes between the paediatric and adult treatment setting.
We estimated (crude) utilization rates for each of the outcomes, separately for patients in paediatric care and those in adult-oriented care. We calculated utilization rates as the number of patients with at least 1 treatment, and because of different follow-up time for each patient, this was divided by the cumulative follow-up time, that is, the number of days until treatment or end of follow-up.
We used Cox proportional hazards regression (11) to estimate adjusted hazard ratios for the effect of paediatric versus adult-oriented care on utilization. Specifically, we regressed time-to-first event (separately for each outcome) on treatment setting (paediatric care = 1; adult care = 0), and the confounders described in the previous section. All analyses were carried out with R (12) version 3.3.3, using the R packages survival (13) and epiR (14).
Using the estimated Cox model, we derived the estimated cumulative proportion of patients with any steroid prescriptions, any IBD-related hospitalization, any IBD-related surgery, and any use of biologicals over time since study inclusion. Specifically, we used the method of direct adjusted survival curves, outlined in Gail and Byar (15). We defined “survival” as the time from study inclusion until the first occurrence of the outcome of interest. The predicted individual curves for all subjects in the data set were retrieved, assuming both care settings (paediatric vs adult-oriented) one by one, after which the average for each setting was taken.
We identified 897 unique, 16-to-17-year-old patients with IBD between the years 2007 and 2014, who were continuously insured for at least a 12-month period immediately preceding study inclusion. We excluded 271 patients as we could not confirm the use of IBD medication in the 12 months following the DTC code for IBD. The final study population included 626 patients of which 380 (61%) were treated in paediatric care setting, and 246 (39%) were treated in adult-oriented care (Fig. 1).
Patients treated in the adult-oriented care setting were significantly more often newly diagnosed, 0.7 years older (median age 17.3 vs 16.6 years, Wilcoxon rank-sum test P < 0.001), and more often treated in general hospitals than patients treated in the pediatric care setting (Table 1). The follow-up time was at least 1 year for 410 patients (65%); the median follow-up time was 1.13 years (IQR 0.52–1.65) in the paediatric care setting, and 1.37 years (IQR = 1.08–1.88) in the adult-oriented care setting. In the paediatric care setting, the follow-up mostly ended by transfer to adult care; in the adult-oriented care setting, the end of follow-up was mostly by definition (turned 19 years). A comparable small proportion of patients in both treatment settings left the insurance company before the end of the study (Table 1).
In the paediatric care setting, disease phenotype (CD vs UC) could be inferred from DTC codes filed in the adult setting for 65% of the patients. When comparing this subset of the paediatric care group with the adult care group, we found similar proportions of patients with CD versus UC in both treatment settings. Specifically, in the paediatric care group, 63% of patients had CD (vs UC 37%), compared with 65% patients with CD in the adult care group (vs UC 35%).
Treatment Outcomes Per Care Setting
Table 2 shows medication use, IBD-related surgical interventions and IBD-related hospital admissions per care setting. Crude utilization rates in the paediatric care setting were lower for steroid prescriptions (paediatric setting: 33.4 per 100 person-years; adult-oriented setting: 56.7 per 100 person-years), IBD-related hospitalizations (paediatric setting: 27.3 per 100 person-years; adult setting: 51.9 per 100 person-years) and biological use (paediatric setting: 28.2 per 100 person years; adult setting: 36.8 per 100 person-years). The adjusted rates of steroid prescription, IBD-related hospitalizations and biological use were significantly lower in the paediatric care setting (hazard ratio [HR] 0.72; 95% confidence interval [CI] 0.52–0.99); HR 0.58; 95% CI 0.37–0.92; and HR 0.57; 95% CI 0.34–0.97, respectively). The HR for IBD-related surgery was also lower in the paediatric care setting (HR 0.51, 95% CI 0.18–1.39), but the absolute numbers were low and no statistical significance was shown (Fig. 2).
Figure 2 shows lower cumulative proportions of steroid prescriptions, IBD-related hospitalizations, IBD-related surgery, and biological use in patients treated in paediatric care versus adult-oriented care over time since study inclusion.
This study shows that 61% versus 39% of Dutch adolescents with IBD between 16 and 18 years old is diagnosed and treated in paediatric versus adult-oriented care, respectively. The paediatric-oriented care group had a lower incidence of steroid prescriptions, and was almost half as likely to be treated with anti-TNF agents (Infliximab or Adalimumab). The lower use of steroids and anti-TNF agents in the paediatric care setting was not accompanied by an increase in IBD-related hospitalizations and IBD-related surgery rates. On the contrary, we observed significantly lower IBD-related hospitalization rates in paediatric care.
Strengths and Limitations
To our knowledge, this is the first study, which evaluates care setting-dependent differences in health care utilization of adolescents with IBD. The strength of this study is its large cohort size, as compared with other studies on adolescent IBD patients. Secondly, this observational study has good external validity because the study population and setting are highly representative of the real world. Achmea is an insurance company for all Dutch inhabitants and covers approximately 25% of the Dutch population. There is no selection of patients because insurance companies in the Netherlands are prohibited to refuse patients based on their health situation. Thus, the cohort is a good reflection of the Dutch population.
A limitation of our observational study is that the distinction between UC and CD in the pediatric group was inferred from data obtained when the patients were adult, and were also incomplete. Yet the proportion of patients with UC versus CD in the adult and pediatric group was almost identical, suggesting the 2 age groups do not differ widely for this aspect.
An additional limitation is that we could not control for confounders not observable in our study data, such as disease severity and disease activity. However, we expect that adolescents with severe and active IBD would be transferred to academic paediatric care more often. As this group with severe disease is likely to use more steroids, anti-TNF agents, and is more frequently hospitalized, correction for disease severity would probably only strengthen the results of our study. Furthermore, we corrected for disease stage (newly diagnosed vs known disease), which is expected to reflect most of the difference in disease activity.
Another potential limitation is that the data were collected for administrative (reimbursement) purposes, rather than scientific research, and therefore, lack detail in terms of secondary diagnoses, and laboratory test results. Medication use in this study is based on medication that patients actually collected at the pharmacy and may differ from the doctor's prescriptions. However, registration of the data is complete, accurate, subject to extensive control, and comprehensive auditing because of the economic function of the data (16).
As this is a study based on a health insurance database, clinical data, such as growth, use of exclusive enteral nutrition (EEN), and disease activity were unfortunately not available. We think that even though these data are not known, the large and significant differences between treatment settings is remarkable. Furthermore, we do not think that the use of EEN (in the paediatric setting often used for remission induction of CD) acts as a confounder, as the groups had comparable percentages of patients with CD. Additionally, we corrected for disease stage (newly diagnosed vs known disease), which may reflect differences in disease activity. Apart from differences in disease stage, we do not expect disease activity to be different between the treatment groups.
In this study, we show a remarkable difference in treatment outcomes between the paediatric and adult care setting. Due to the observational nature of the study, we cannot draw conclusions on the cause of differences in study outcomes. However, we speculate that these may be the effect of—inherent—differences between paediatric and adult-oriented care. The previously reported increased rates of nonadherence (17), decline of self-management and increase in hospital admissions during transition to adult care (18), may support this hypothesis. Literature that specifically focuses on the effect of treatment setting in adolescents with chronic disease is scarce. A literature review on adolescents with cancer indicates that adolescents with cancer may have a survival advantage when treatment followed paediatric protocols compared with adult protocols (19). Other effects of treatment setting on adolescent care can be derived from studies on transition of patients from paediatric to adult care. Young people frequently drop out of healthcare systems after they leave paediatric care (20). A systematic review exploring the experiences of adolescents with chronic kidney disease regarding the transition process found that in adult care, adolescents felt out of place among the older patients, overwhelmed by the environment and the lack of attention from health professionals (21). Several studies on the effect of transition to adult care setting show unfavourable outcomes after transition, such as increased disease activity in rheumatic disease (22,23), or significantly increased risk of graft failure in paediatric kidney transplant recipients (24–26). Interestingly, 1 study showed that the effect of transfer depended on age at transfer: individuals transferred at <21 years old had a 57% higher risk of graft failure than individuals of the same age who had been transferred at ≥21 years old (27). Follow-up strategies in paediatric and adult-oriented care typically differ on several aspects, generally reflecting the particular needs of paediatric and adult populations. Paediatric care is often family focused, multidisciplinary, and reliant on parental involvement for consent and guidance. In contrast, typical adult-oriented care is focused on the individual patient, carried out by single providers, and advocates patient independence (28).
In the adult care group, more newly diagnosed patients were seen as compared with the paediatric care group, which could explain some of the higher incidence of corticosteroid prescriptions, as this is used in induction therapy. However, the adult care setting also showed a higher biological use, which cannot be explained by this way.
This study shows that a significant proportion of adolescent IBD patients is diagnosed and treated by adult care providers. In the paediatric-oriented care setting we observed a lower use of corticosteroids and biologicals as well as a lower incidence of IBD-related hospitalizations. This indicates that adolescent IBD patients of similar age get different treatment and possibly also different treatment outcomes according to their treatment setting. These research findings should stimulate more extensive collaboration between paediatric and adult professionals to search for the best treatment approach to reach optimal outcomes for adolescents with IBD.
We thank Onno Dijt for providing research assistance, and Achmea Insurance Company for providing the data used in this study.
1. Ruemmele FM, Turner D. Differences in the management of pediatric and adult onset ulcerative colitis-lessons from the Joint ECCO and ESPGHAN Consensus Guidelines for the Management of Pediatric Ulcerative Colitis. J Crohns Colitis
2. Goodhand J, Dawson R, Hefferon M, et al. Inflammatory bowel disease in young people: the case for transitional clinics. Inflamm Bowel Dis
3. Van Limbergen J, Russell RK, Drummond HE, et al. Definition of phenotypic characteristics of childhood-onset inflammatory bowel disease. Gastroenterology
4. Bollegala N, Brill H, Marshall JK. Resource utilization during pediatric to adult transfer of care in IBD. J Crohns Colitis
5. Goodhand J, Hedin CR, Croft NM, et al. Adolescents with IBD: the importance of structured transition care. J Crohns Colitis
6. Paine CW, Stollon NB, Lucas MS, et al. Barriers and facilitators to successful transition from pediatric to adult inflammatory bowel disease care from the perspectives of providers. Inflamm Bowel Dis
7. Gray WN, Denson LA, Baldassano RN, et al. Treatment adherence in adolescents with inflammatory bowel disease: the collective impact of barriers to adherence and anxiety/depressive symptoms. J Pediatr Psychol
8. de Bont JM, van der Holt B, Dekker AW, et al. Significant difference in outcome for adolescents with acute lymphoblastic leukemia treated on pediatric vs adult protocols in the Netherlands. Leukemia
9. WHO Collaborating Centre for Drug Statistics Methodology. ATC structure and principles. Available at: https://www.whocc.no/atc/structure_and_principles/
. Accessed November 1, 2017.
10. Evaluatie overheveling geneesmiddelen. Available at: https://zoek.officielebekendmakingen.nl/blg-719581.pdf
. Accessed November 1, 2017.
11. Cox DR. Regression models and life-tables. J Royal Stat Soc B
12. R Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Available at: https://www.R-project.org/
. Accessed November 1, 2017.
13. Therneau T. A Package for Survival Analysis in S. version 2.38. Available at: https://cran.r-project.org/web/packages/survival/citation.html
. Accessed November 1, 2017.
14. Stevenson M, Nunes T, Heuer C, et al. EpiR: Tools for the Analysis of Epidemiological Data. R package version 0.9-87. Available at: https://CRAN.R-project.org/package=epiR
. Accessed November 1, 2017.
15. Gail MH, Byar DP. Variance calculations for direct adjusted survival curves, with applications to testing for no treatment effect. Biom J
16. Smeets HM, de Wit NJ, en Hoes AW. Routine health insurance data for scientific research: potential and limitations of the Agis Health Database. J Clin Epidemiol
17. Jackson CA, Clatworthy J, Robinson A, et al. Factors associated with non-adherence to oral medication for inflammatory bowel disease: a systematic review. Am J Gastroenterol
18. Abraham BP, Kahn SA. Transition of care in inflammatory bowel disease. Gastroenterol Hepatol (NY)
19. Tai E, Buchanan N, Westervelt L, et al. Treatment setting, clinical trial enrollment, and subsequent outcomes among adolescents with cancer: a literature review. Pediatrics
20. Viner R. Transition from paediatric to adult care. Bridging the gaps or passing the buck? Arch Dis Child
21. Crawford K, Wilson C, Low JK, et al. Transitioning adolescents to adult nephrology care: a systematic review of the experiences of adolescents, parents, and health professionals. Pediatr Nephrol
2019; doi: 10.1007/s00467-019-04223-9. [Epub ahead of print].
22. Hersh AO. Growing up and moving on—transition of care for patients with childhood onset rheumatic disease. J Rheumatol
23. Son MB, Sergeyenko Y, Guan H, et al. Disease activity and transition outcomes in a childhood-onset systemic lupus erythematosus cohort. Lupus
24. Samuel SM, Nettel-Aguirre A, Hemmelgarn BR, et al. Graft failure and adaptation period to AdultHealthcare Centers in pediatric renal transplant patients. Transplantation
25. Prestidge C, Romann A, Djurdjev O, et al. Utility and cost of a renal transplant transition clinic. Pediatr Nephrol
26. Kreuzer M, Prüfe J, Oldhafer M, et al. Transitional care and adherence of adolescents and young adults after kidney transplantation in Germany and Austria. A Binational Observatory Census Within the TRANSNephro Trial. Medicine (Baltimore)
27. Foster BJ, Platt RW, Dahhou M, et al. The impact of age at transfer from pediatric to adult oriented care on renal allograft survival. Pediatr Transplant
28. Bollegala N, Nguyen GC. Transitioning the adolescent with IBD from pediatric to adult care: a review of the literature. Gastroenterol Res Pract