OBJECTIVES: The goals were (1) to describe rates of diabetes mellitus (DM)-related hospitalizations and retinopathy screening before and after transition to adult care and (2) to test whether different methods of transfer of care were associated with improved outcomes.
METHODS: In a retrospective cohort study, we included 1507 young adults with DM of ≥5-year duration and tracked these patients until 20 years of age.
RESULTS: DM-related hospitalization rates increased from 7.6 to 9.5 cases per 100 patient-years in the 2 years after transition to adult care (P = .03). Previous DM-related hospitalizations, lower income, female gender, and living in areas with low physician supply were associated with higher admission rates. With controlling for all other factors, individuals who were transferred to a new allied health care team with no change in physician were 77% less likely (relative risk: 0.23 [95% confidence interval: 0.05–0.79]) to be hospitalized after the transition than were those transferred to a new physician with either a new or no allied health care team. The rates of eye examinations were stable across the transition to adult care (72% vs 70%; P = .06). Female patients, patients with higher income, and patients with previous eye care were more likely to have an eye care visit after transfer.
CONCLUSIONS: During the transition to adult health care, there is increased risk of DM-related hospitalizations, although this may be attenuated in youths for whom there is physician continuity. Eye care visits were not related to transition; however, rates were below evidence-based guideline recommendations.
Diabetes mellitus (DM) is a common chronic disease in children and adolescents, the incidence of which has been increasing at a rate of 2% to 5% per year.1–3 Short-term complications such as diabetic ketoacidosis are the leading cause of hospitalizations and death among children with type 1 DM.4 Long-term microvascular and macrovascular complications associated with DM are a leading cause of end-stage renal disease5 and acquired blindness.6
Transition to adult care occurs during a critical and vulnerable period for those with childhood-onset chronic diseases, because of the psychological and physiologic changes associated with adolescence. After transfer to adult health care, patients often feel a sense of abandonment and sadness with the loss of the patient-provider relationship, and many find this a negative experience.7–9 During this period, adolescents are at increased risk of leaving medical supervision.7,10,11 Transition care for adolescents with DM is particularly important because regular medical supervision, screening for microvascular complications, and optimizing glycemic control are essential in reducing the onset and progression of DM complications.12,13 However, only 2 small studies of highly selected patients described associations between transfer to adult care and resultant changes in clinical outcomes.14,15 Although 2 cross-sectional studies7,11 and 2 retrospective cohort studies10,14 suggested that adherence to medical supervision deteriorates with transfer to adult DM care, the impact on DM-related hospitalizations and on retinopathy screening has not been studied. In addition, few studies have examined different models of transition and patient health outcomes in DM16,17 or other childhood-onset chronic conditions.18 Our objectives were to compare rates of DM-related hospitalizations and retinopathy screening before and after transfer to adult care and to test whether the method of transfer was associated with these outcomes and processes of care.
We conducted a retrospective cohort study by using population-based, health administrative and survey data from Ontario, which is Canada's largest province, with >12 million residents. Ontario provides universal access to primary and acute health care services for all legal residents, and complete health administrative data are available at the Institute for Clinical Evaluative Sciences, through an agreement with the Ontario Ministry of Health and Long-Term Care. This study was approved by the Hospital for Sick Children and Sunnybrook Health Science Center (Toronto, Canada) research ethics boards.
We used the Ontario Diabetes Database (ODD), a validated,19 population-based, electronic database containing data on patients with DM, to identify all adolescents 16 years of age with prevalent DM for ≥5 years between April 1, 1996, and March 31, 2002. We included patients with DM for ≥5 years because this is the period after which the Canadian Diabetes Association clinical practice guidelines recommend annual retinopathy screening.20 The records for patients from the ODD were linked to health administrative databases by using unique individual identifiers. Each patient's age at diagnosis was identified from the ODD, whereas gender, date of birth, and postal code were obtained from the Registered Person Database, which has demographic information on all residents of Ontario who are registered in the Ontario Health Insurance Plan (OHIP). All legal residents of Ontario are registered in the OHIP and do not have the option of opting out of the plan.
Canadian census data from 1996 and 2001 were used to describe neighborhood-level household income, by using the subjects' residential postal codes. We used hospital admission data from the Canadian Institute for Health Information Hospital Discharge Abstract Database, which are mandatory submissions from all hospitals. The Canadian Institute for Health Information Hospital Discharge Abstract Database includes primary and secondary diagnosis codes from the International Classification of Diseases, Ninth Revision (ICD-9), and International Classification of Diseases, 10th Revision (ICD-10) (as of 2002). We used the fee-for-service claims from OHIP to identify outpatient eye care visits. The Institute for Clinical Evaluative Sciences Physician Database was used to determine pediatric physician supply for each patient's county of residence. For each county, we calculated full-time equivalents (FTEs) of pediatricians and pediatric subspecialists by using standard methods.21 These methods are based on comparisons of billing activity between physicians within their respective specialties and not solely on the number of physicians in each county.21
We conducted a survey of all pediatric DM centers in Ontario regarding their methods of transition of care. In Ontario, all children with DM who are <18 years of age are monitored in 1 of 34 specialized pediatric DM centers (5 tertiary-level and 29 secondary-level programs in the Network of Ontario Pediatric Diabetes Programs). The survey was conducted at the 2007 meeting of the Network of Ontario Pediatric Diabetes Programs, which was attended by representatives of all 34 centers. Each representative completed a self-administered questionnaire consisting of 1 multiple-choice question regarding the center's method of transfer. In general, DM medical care is provided by pediatric endocrinologists in urban centers or by general pediatricians in rural centers. These physicians work in the context of a DM health care team including nurses, dietitians, and mental health professionals. In Ontario, patients with DM are transferred to adult care at 18 years of age.
The main outcomes were DM-related hospitalizations and retinopathy screening visits from 18 to 20 years of age. We defined hospitalizations for acute complications as those with diagnoses of hyperglycemia (ICD-9 code 250.1; ICD-10 codes E10.0–14.0 and E10.1–14.1), which includes diabetic ketoacidosis (ICD-9 code 250.2; ICD-10 codes 10.1–14.1) and hyperosmolar hyperglycemic coma (ICD-9 code 250.3; ICD-10 codes 10.0–14.0), or hypoglycemia (ICD-9 code 251; ICD-10 codes E10.63–E14.63). We defined eye examinations as any visit to an ophthalmologist or optometrist for a periodic ocular examination with a diagnosis code for DM or any visit for a comprehensive ocular examination.
We categorized the method of transition of each pediatric DM center in Ontario as (1) transfer to a new physician and allied health care team, (2) transfer to a new physician but remaining with the same allied health care team, (3) no change in physician or allied health care team, (4) transfer to a new allied health care team but remaining with the same physician, or (5) transfer to a new physician with no allied health care team follow-up care. We combined the transition groups on the basis of having no continuity with the DM team, representing a major change in care (categories 1 and 5). Because patients with DM are monitored by the DM center closest to their home, we assigned patients to the DM center closest to their residential postal code at 16 years of age. To measure access to eye care professionals, we used the patient's postal code at 16 years of age and calculated the shortest distance to the nearest eye care provider. We used the census-based, mean household income quintile of the subject's enumeration (1996) or dissemination (2001) area (average population: 400–700 individuals) as a proxy for socioeconomic status. The income quintiles were adjusted for household and community size.22 Income quintiles for patients residing in areas with unstable populations were suppressed; these were assigned to a “missing income data” category. Pediatric physician supply was determined for each subject's county of residence at 16 years of age, on the basis of FTEs. We divided counties into 4 categories on the basis of FTE levels, that is, high supply (academic health centers and pediatric subspecialists), no supply (no pediatricians), medium supply (above the median pediatrician FTE value in the remaining counties), or low supply (below the median pediatrician FTE value in the remaining counties).
Statistical analyses of the data were performed by using SAS 9.1 (SAS Institute, Cary, NC).
We used a paired t test to compare hospitalization rates in the 2 years before and after transfer. Rates were also calculated with the omission of outliers, because the inclusion of outliers resulted in hospitalizations rates that were much higher than those reported previously12,23 and were likely not representative of patients' clinical outcomes during the transition period. We identified outliers as observations with DM-related hospitalizations above the 99th percentile of the distribution (n = 23; 1.5% of the total cohort); this corresponded to having >5 DM-related hospital admissions.
We tested the association between differing methods of transition and subsequent DM-related hospitalizations by using multivariate negative binomial regression analysis for event counts. We used other regression analysis techniques, including Poisson regression, 0-inflated Poisson regression, and 0-inflated negative binomial regression.24 The negative binomial regression model was the best-fit model on the basis of the Vuong statistic and the χ2 likelihood ratio test.25 Adjusted relative risk (RR) estimates were calculated with the final regression model, and the corresponding 95% confidence intervals (CIs) were calculated by using the likelihood ratio method.
Eye Care Visits
We determined whether the proportions of patients who had ≥1 eye examination were significantly different in the 2-year periods before and after transition by using McNemar's χ2 test. We used multivariate logistic regression to examine the association of different transition methods with eye care use after transfer to adult care. Adjusted odds ratio (OR) estimates and the corresponding 95% CIs were calculated with the final logistic regression model.
Methods of Transition
The total cohort consisted of 1507 patients, with equal gender distribution (Table 1). We had a 100% response rate for our survey of DM centers. Most patients (63%) were associated with DM centers that transferred patients to a new physician and allied health care team for adult DM care (Table 2).
Overall, 10% of the cohort experienced DM-related hospitalizations during the 4 years of the transition period. Rates of DM-related hospitalizations did not increase significantly after transfer to adult care (Table 3). However, after omission of outliers, rates of hospitalization because of acute hyperglycemia increased from 7.2 to 9.1 cases per 100 patient-years in the 2 years after transition (P = .02). Total DM-related hospitalizations also increased after transition (from 7.6 to 9.5 cases per 100 patient-years; P = .03). Most of the patients in the outlier group were in the lowest income quintiles (39% vs 4% in the highest income quintile; data not shown) but did not differ with respect to other baseline characteristics.
Table 4 summarizes the adjusted RRs among the entire cohort for DM-related hospitalizations after transition. With controlling for all other factors, individuals who were transferred to a new allied health care team with no change in physician were 77% less likely (RR: 0.23 [95% CI: 0.05–0.79]) to be hospitalized after transition, compared with those who were transferred to a new physician with either a new or no allied health care team. Furthermore, individuals with no change in care after transition demonstrated a decreased risk of DM-related hospitalizations, compared with those with a complete change in care (RR: 0.09 [95% CI: 0.003–0.71]). To delineate whether the effect on DM-related hospitalizations was attributable to the change of physician versus allied health care team, we combined the transition groups on the basis of whether there was a change in physician. With controlling for all other factors, individuals who were transferred to a new physician were 4 times (RR: 4.39 [95% CI: 1.62–14.4]; data not shown) more likely to be hospitalized after transition than were those who remained with the same physician after transition.
Regression analysis revealed an increased risk of hospitalization with decreasing neighborhood income, with hospitalization rates for the lowest quintile being double those for the highest quintile (RR: 2.25 [95% CI: 1.27–4.02]). Other significant independent predictors of DM-related hospitalizations after transfer included previous DM-related hospitalizations, female gender, and living in areas with low physician supply.
The proportions of individuals who had ≥1 eye care examination were stable across transition of care (72% vs 70%; P = .06). Female patients, patients from higher-income neighborhoods, and patients with previous eye care visits were more likely to have had an eye care visit after transition of care (Table 5). Distance to an eye care professional was not associated with visits.
This is the largest study to date to document increased rates of admissions because of acute complications of DM in youths in the period after transfer to adult care. Transfer that included physician continuity was associated with lower DM-related hospitalization rates. Consistent with previous studies, the majority of our cohort did not have any hospitalizations during the 4 years of observation.23 A retrospective cohort study of 229 young adults demonstrated similar associations with respect to medical supervision.16 Clinic attendance was compared within 4 health districts in the Oxford, England, region with differing transition models. The greatest decreases in clinic attendance were among subjects who either were transferred to a young-adult clinic in a different hospital or were transferred directly to adult care. In contrast, clinic attendance was better when subjects were transferred to a young-adult clinic within the same hospital as the pediatric clinic or to an adolescent clinic run jointly by pediatric and adult physicians. Consistent with our results, these findings suggest that patients who had some degree of familiarity with the adult health care providers experienced improved outcomes.
Our results support the qualitative evidence regarding challenges youths face in establishing new relationships with their adult health care providers. For young adults with childhood-onset chronic diseases, studies describe the difficulties patients face in establishing a trusting relationship with their adult care providers,9,26,27 in part because they lack confidence in the providers' knowledge and experience regarding childhood-onset chronic conditions.9,26,27 Patients' lack of trust and satisfaction with their health care providers can lead to poor glycemic control28,29 and decreased treatment compliance.30 Our findings of an increased risk of hospitalizations for patients with a change in their physician may be explained in part by this relationship loss and lack of therapeutic consistency occurring during the transfer to adult care.
Our findings that female patients were more likely to be hospitalized after transfer to adult care is new. However, the risk for DM-related hospitalizations has been shown to be higher among adolescent and young-adult female patients.23,31–33 These results may reflect, at least in part, the increased risk of disordered eating behaviors and insulin misuse among teen and young-adult female patients with DM.34–36
Although Canadian patients enjoy universal access to health care, we documented income gradients in DM-related hospitalizations. Another Ontario study demonstrated a clear inverse relationship between hospital admissions attributable to acute complications of DM and neighborhood-level incomes, although with a more-attenuated effect in children than in this study.37,38 However, the previous study included children of all ages and patients whose hospitalizations were the first presentation of disease. Although we cannot explain the disparities in admissions according to income in our cohort, they may be driven by factors such as compliance with medical care, including decreased blood glucose monitoring,39 access to transportation, and irregular physician visits.13 Our results also suggest that, irrespective of socioeconomic factors, problems with access to specialist care because of geographic maldistribution may be associated with worse outcomes for youths with DM. Although we used pediatric physician supply to describe this maldistribution, it is strongly correlated with adult physician supply, with the majority of pediatric and adult endocrinologists working in urban centers.40,41 Other studies support the association between hospitalizations attributable to ambulatory care-sensitive conditions such as DM and low physician supply.42
Unlike admission rates, we report no significant relationship between eye care use and transition to adult care. However, despite the use of an inclusive definition of provider visits for retinopathy screening, rates were well below current practice guidelines.20 This suggests that many young adults with DM are not benefiting from preventive eye care and are at risk for vision-threatening diabetic retinopathy. Although similar suboptimal rates have been found among older populations with health care insurance in both the United States and Canada,43–45 this is the first study to describe eye care among youths with DM. Our results suggest that, despite universal funding for these visits, low-income patients are not accessing services. This is particularly important in the context of more-poorly controlled disease, as measured by admission rates, among low-income youths with DM.
To our knowledge, this is the first population-based study describing acute complications of DM and appropriate use of eye care during the transfer to adult care and the largest study evaluating different methods of transfer. However, our study has several limitations. First, the administrative data used did not allow control for the potential, individual-level, confounding factors that can contribute to the development of acute DM-related complications, including glycemic control,23 level of education,46 and comorbid psychiatric conditions.23 We did not assess ambulatory care use,13 because some pediatric subspecialists are salaried and we have not yet validated the completeness of their “shadow” billing data. Second, measurement of DM-related hospitalizations as a proxy measure of glycemic control may underestimate the effect of transition on clinical outcomes. Patients with poor glycemic control may not develop acute complications of DM; however, they remain at risk for increased DM-related deaths and morbidity.47 Third, we could not distinguish between individuals with type 1 versus type 2 DM. However, the vast majority of our cohort likely had type 1 DM, because our inclusion criteria required patients to have been diagnosed as having DM by 11 years of age and the peak age at onset of type 2 DM is 14 to 15 years.48–50 Type 2 DM among youths <10 years of age is extraordinarily rare,48–50 and most youths <20 years of age have type 1 DM.48,51 Finally, we could not verify that patients were being cared for at the center closest to their residence, although patients in Ontario generally are monitored at the center closest to home. A few exceptions exist, and patients with complicated DM courses may receive some of their care at tertiary care centers. Finally, although all patients remained in Ontario during the observation period, we were not able to capture patients who might have moved within Ontario during the transition period.
Our results suggest that having continuity with the pediatric physician during the transition to adult care decreases admissions attributable to acute DM complications. Current programs, such as those in children's hospitals, that include a complete transfer of care should consider implementing transition care programs that involve some continuity with the pediatric physician or earlier integration of the adult team into care. Although transition care seems to be unrelated to eye care use, important gaps in care exist. DM care providers need to focus attention on optimizing the use of eye examinations to ensure that youths with DM are benefiting from preventive eye care.
This study was performed at the Institute for Clinical Evaluative Sciences, which is funded by an annual grant from the Ontario Ministry of Health and Long-Term Care. Funds for data collection were from the R.S. McLaughlin Foundation Chair in Pediatrics awarded to Dr Daneman. Dr Guttmann holds a Canadian Institutes for Health Research Career Award. Dr Paradis holds a Canadian Institutes for Health Research Public Health Research Chair in Chronic Disease Prevention. Dr Nakhla was supported by a Hospital for Sick Children Research Training Award.
- Accepted June 23, 2009.
- Address correspondence to Meranda Nakhla, MD, MSc, Division of Endocrinology and Metabolism, Children's Hospital of Eastern Ontario, 401 Smyth Rd, Ottawa, ON K1H 8L1, Canada. E-mail:
The opinions, results, and conclusions reported in this article are those of the authors and are independent from all funding sources. No endorsement by the Institute for Clinical Evaluative Sciences or the Ontario Ministry of Health and Long-Term Care is intended or should be inferred.
Financial Disclosure: The authors have indicated they have no financial relationships relevant to this article to disclose.
What's Known on This Subject:
Adherence to medical supervision around the time of transition to adult health care services for youths with DM is not optimal. Little is known about the impact of this nonadherence on DM-related hospitalizations and retinopathy screening.
What This Study Adds:
Youths with DM with some continuity with their physician are less likely to be hospitalized with acute complications of DM after the transition to adult care, compared with those with a complete transition. Overall retinopathy screening rates are low.
- Vehik K, Hamman RF, Lezotte D, et al. Increasing incidence of type 1 diabetes in 0- to 17-year-old Colorado youth. Diabetes Care.2007;30 (3):503– 509
- ↵Guttman A, Nakhla M, Henderson M, et al. Validation of a health administrative data algorithm for assessing the epidemiology of diabetes in Canadian children. Pediatr Diabetes. 2009;Jun 5 (Epub ahead of print)
- ↵Edge JA, Ford-Adams ME, Dunger DB. Causes of death in children with insulin dependent diabetes 1990–96. Arch Dis Child.1999;81 (4):318– 323
- ↵US Renal Data System. USRDS 2004 Annual Data Report: Atlas of End-Stage Renal Disease in the United States. Bethesda, MD: National Institutes of Health; 2004
- ↵Pacaud D, McConnell B, Huot C, Aebi C, Yale J. Transition from pediatric care to adult care for insulin-dependent diabetes patients. Can J Diabetes Care.1996;20 (4):14– 20
- Scott L, Vallia M, Charette M, Murray A, Latta R. Transition of care: researching the needs of young adults with type 1 diabetes. Can J Diabetes.2005;29 (3):203– 210
- ↵Frank M. Factors associated with non-compliance with a medical follow-up regimen after discharge from a pediatric diabetes clinic. Can J Diabetes Care.1996;20 (3):13– 20
- ↵Pacaud D, Yale J, Stephure D, Trussell R, Davies H. Problems in transition from pediatric care to adult care for individuals with diabetes. Can J Diabetes.2005;29 (1):13– 18
- ↵Hux JE, Ivis F, Flintoft V, Bica A. Diabetes in Ontario: determination of prevalence and incidence using a validated administrative data algorithm. Diabetes Care.2002;25 (3):512– 516
- ↵Canadian Diabetes Association. Canadian Diabetes Association 2008 clinical practice guidelines for the prevention and management of diabetes in Canada. Can J Diabetes.2008;32 (suppl 1):S1– S201
- ↵Chan B, Schultz SE. Supply and Utilization of General Practitioner and Family Physician Services in Ontario: An Institute for Clinical Evaluative Sciences (ICES) Integrative Report. Toronto, Canada: Institute for Clinical Evaluative Sciences; 2005
- ↵Wilkins R. PCCF+ Version 3G User's Guide (Geocodes/PCCF): Automated Geographic Coding Based on Statistics Canada Postal Code Conversion Files, Including Postal Codes to June 2001. Ottawa, Canada: Statistics Canada; 2001
- ↵Long JS. Regression Models for Categorical and Limited Dependent Variables. Thousand Oaks, CA: Sage Publications; 1997
- ↵Reiss JG, Gibson RW, Walker LR. Health care transition: youth, family, and provider perspectives. Pediatrics.2005;115 (1):112– 120
- ↵Bryden KS, Peveler RC, Stein A, Neil A, Mayou RA, Dunger DB. Clinical and psychological course of diabetes from adolescence to young adulthood: a longitudinal cohort study. Diabetes Care.2001;24 (9):1536– 1540
- ↵Ciechanowski PS, Hirsch IB, Katon WJ. Interpersonal predictors of HbA1c in patients with type 1 diabetes. Diabetes Care.2002;25 (4):731– 736
- ↵Lee JM, Okumura MJ, Freed GL, Menon RK, Davis MM. Trends in hospitalizations for diabetes among children and young adults: United States, 1993–2004. Diabetes Care.2007;30 (12):3035– 3039
- ↵Peveler RC, Bryden KS, Neil HA, et al. The relationship of disordered eating habits and attitudes to clinical outcomes in young adult females with type 1 diabetes. Diabetes Care.2005;28 (1):84– 88
- Jones JM, Lawson ML, Daneman D, Olmsted MP, Rodin G. Eating disorders in adolescent females with and without type 1 diabetes: cross sectional study. BMJ.2000;320 (7249):1563– 1566
- ↵Neumark-Sztainer D, Patterson J, Mellin A, et al. Weight control practices and disordered eating behaviors among adolescent females and males with type 1 diabetes: associations with sociodemographics, weight concerns, familial factors, and metabolic outcomes. Diabetes Care.2002;25 (8):1289– 1296
- ↵Jaakkimainen L, Shah B, Kopp A. Sources of physician care for people with diabetes. In: Hux JE, Booth GL, Slaughter PM, Laupacis A, eds. Practice Atlas: Diabetes in Ontario. Toronto, Canada: Institute for Clinical Evaluative Sciences; 1998:181–192
- ↵Chan B, Harju M. Supply and utilization of health care services for diabetes. In: Hux JE, Booth GL, Slaughter PM, Laupacis A, eds. Diabetes in Ontario: An ICES Practice Atlas. Toronto, Canada: Institute for Clinical Evaluative Sciences; 2003:14.249–14.268
- ↵Buhrmann R, Assaad D, Hux JE, Tang M, Sykora K. Diabetes and the eye. In: Hux JE, Booth GL, Slaughter PM, Laupacis A, eds. Practice Atlas: Diabetes in Ontario. Toronto, Canada: Institute for Clinical Evaluative Sciences; 1998:193–208
- ↵Icks A, Rosenbauer J, Strassburger K, Grabert M, Giani G, Holl RW. Persistent social disparities in the risk of hospital admission of paediatric diabetic patients in Germany: prospective data from 1277 diabetic children and adolescents. Diabet Med.2007;24 (4):440– 442
- Bell RA, Mayer-Davis EJ, Beyer JW, et al. Diabetes in non-Hispanic white youth: prevalence, incidence, and clinical characteristics: the SEARCH for Diabetes in Youth Study. Diabetes Care.2009;32 (suppl 2):S102– S111
- ↵Lawrence JM, Mayer-Davis EJ, Reynolds K, et al. Diabetes in Hispanic American youth: prevalence, incidence, demographics, and clinical characteristics: the SEARCH for Diabetes in Youth Study. Diabetes Care.2009;32 (suppl 2):S123– S132
- ↵Amed S, Dean H, Hamilton J. National surveillance for non-type 1 diabetes (NT1DM) in Canadian children. Presented at the Canadian Diabetes Association/Canadian Society of Endocrinology and Metabolism Professional Conference and Annual Meetings; October 15–18, 2008; Montreal, Canada
- Copyright © 2009 by the American Academy of Pediatrics