Health Care Transition in Young Adults With Type 1 Diabetes
Health Care Transition in Young Adults With Type 1 Diabetes
We assessed transition barriers as part of a survey of emerging adults with type 1 diabetes. The survey also examined transition preparation and satisfaction, gaps in care, current self care, and demographics.
Survey development, recruitment, and fielding are described in detail elsewhere. Briefly, survey development was informed by literature review, focus groups consisting of posttransition young adults with diabetes, expert review, and cognitive testing with young adult patients with type 1 diabetes.
Eligible subjects included young adults with type 1 diabetes between the ages of 22 and 30 years who were under the care of an adult diabetes specialist at a single tertiary diabetes center. A review of electronic medical records identified patients with the following criteria: encounters classified as International Classification of Disease-Ninth Revision (ICD-9) codes 250.X1 or 250.X3, diabetes diagnosed before the age of 18 years, previous pediatric diabetes care, and a medical visit with hemoglobin A1c (A1c) measured in the adult clinic within 2 years of the study.
We mailed a paper survey to all 484 eligible subjects in three waves. We reminded subjects by telephone and offered a secure electronic Internet option using the Research Data Electronic Capture Survey (REDCap). The institutional committee on human subjects at Joslin Diabetes Center approved this study.
For assessment of barriers, we asked respondents about six specific barriers, which reflected major themes identified in focus groups and provider interviews: (1) "I didn't have a name for an adult provider;" (2) "I didn't know how to contact the new adult provider;" (3) "I couldn't get an appointment with the new adult provider;" (4) I had other priorities;" (5) "I felt upset about leaving my child/adolescent providers;" and (6) "I didn't have health insurance." For each barrier, the four response options included: not at all a barrier; small barrier; moderate barrier; or major barrier. Patients reported the gap in time between their last pediatric diabetes visit and first adult diabetes visit as ≤3 months, 4 to 6 months, 7 to 12 months, 13 to 24 months, or >24 months.
All variables were self-reported, except for the most recent adult clinic A1c level (Tosoh, San Francisco, CA), which was measured and obtained from chart review. Self-reported pretransition A1c was assessed with seven survey response options: ≤7.0%, 7.1 to 8.0%, 8.1 to 9.0%, 9.1 to 10.0%, 10.1 to 12%, >12%, or "don't know." The 24 subjects who responded "don't know" for this item were eliminated from multivariate analyses that included pretransition A1c as a covariate. The other six categories were ordered as continuous variables using the mean for each range and an imputed value of 13% for the >12% option. To validate the self-reported pretransition A1c values, electronically stored measured results from 69 respondents who had previously received their pediatric diabetes care at the same tertiary center were compared with the self-reports. The pretransition A1c values matched the measured values for 72% of these respondents; of the 28% of respondents for whom the values did not agree, two-thirds of the self-reported values were higher than the measured values and one-third were lower.
Analyses were conducted using SAS 9.2 software (SAS Institute, Cary, NC). Descriptive statistics were calculated as means and standard deviations or proportions. For all analyses, P<.05 was considered significant.
In multivariate analyses, barriers were dichotomized as not at all/small versus moderate/major in order to sharpen the distinction between those subjects who perceived significant barriers to the establishment of adult care and those who did not. Similarly, we dichotomized gaps at ≤6 months or >6 months for comparison with other reports, given the American Diabetes Association recommendation that insulin-treated patients >18 years of age have diabetes visits at least every 6 months.
We explored bivariate relationships between specific barriers and gaps in care using chi-square tests. We used logistic regression to assess the odds of a gap >6 months between pediatric and adult care for subjects reporting each of the six moderate/major barriers, adjusting for pretransition A1c, sex, education, and transition age.
Finally, the presence or absence of at least one moderate/major barrier was entered into a logistic regression model with a gap >6 months between pediatric and adult care as the dependent variable, adjusting for the same covariates.
Methods
Data Collection
We assessed transition barriers as part of a survey of emerging adults with type 1 diabetes. The survey also examined transition preparation and satisfaction, gaps in care, current self care, and demographics.
Survey development, recruitment, and fielding are described in detail elsewhere. Briefly, survey development was informed by literature review, focus groups consisting of posttransition young adults with diabetes, expert review, and cognitive testing with young adult patients with type 1 diabetes.
Eligible subjects included young adults with type 1 diabetes between the ages of 22 and 30 years who were under the care of an adult diabetes specialist at a single tertiary diabetes center. A review of electronic medical records identified patients with the following criteria: encounters classified as International Classification of Disease-Ninth Revision (ICD-9) codes 250.X1 or 250.X3, diabetes diagnosed before the age of 18 years, previous pediatric diabetes care, and a medical visit with hemoglobin A1c (A1c) measured in the adult clinic within 2 years of the study.
We mailed a paper survey to all 484 eligible subjects in three waves. We reminded subjects by telephone and offered a secure electronic Internet option using the Research Data Electronic Capture Survey (REDCap). The institutional committee on human subjects at Joslin Diabetes Center approved this study.
Key Variables
For assessment of barriers, we asked respondents about six specific barriers, which reflected major themes identified in focus groups and provider interviews: (1) "I didn't have a name for an adult provider;" (2) "I didn't know how to contact the new adult provider;" (3) "I couldn't get an appointment with the new adult provider;" (4) I had other priorities;" (5) "I felt upset about leaving my child/adolescent providers;" and (6) "I didn't have health insurance." For each barrier, the four response options included: not at all a barrier; small barrier; moderate barrier; or major barrier. Patients reported the gap in time between their last pediatric diabetes visit and first adult diabetes visit as ≤3 months, 4 to 6 months, 7 to 12 months, 13 to 24 months, or >24 months.
All variables were self-reported, except for the most recent adult clinic A1c level (Tosoh, San Francisco, CA), which was measured and obtained from chart review. Self-reported pretransition A1c was assessed with seven survey response options: ≤7.0%, 7.1 to 8.0%, 8.1 to 9.0%, 9.1 to 10.0%, 10.1 to 12%, >12%, or "don't know." The 24 subjects who responded "don't know" for this item were eliminated from multivariate analyses that included pretransition A1c as a covariate. The other six categories were ordered as continuous variables using the mean for each range and an imputed value of 13% for the >12% option. To validate the self-reported pretransition A1c values, electronically stored measured results from 69 respondents who had previously received their pediatric diabetes care at the same tertiary center were compared with the self-reports. The pretransition A1c values matched the measured values for 72% of these respondents; of the 28% of respondents for whom the values did not agree, two-thirds of the self-reported values were higher than the measured values and one-third were lower.
Statistical Analyses
Analyses were conducted using SAS 9.2 software (SAS Institute, Cary, NC). Descriptive statistics were calculated as means and standard deviations or proportions. For all analyses, P<.05 was considered significant.
In multivariate analyses, barriers were dichotomized as not at all/small versus moderate/major in order to sharpen the distinction between those subjects who perceived significant barriers to the establishment of adult care and those who did not. Similarly, we dichotomized gaps at ≤6 months or >6 months for comparison with other reports, given the American Diabetes Association recommendation that insulin-treated patients >18 years of age have diabetes visits at least every 6 months.
We explored bivariate relationships between specific barriers and gaps in care using chi-square tests. We used logistic regression to assess the odds of a gap >6 months between pediatric and adult care for subjects reporting each of the six moderate/major barriers, adjusting for pretransition A1c, sex, education, and transition age.
Finally, the presence or absence of at least one moderate/major barrier was entered into a logistic regression model with a gap >6 months between pediatric and adult care as the dependent variable, adjusting for the same covariates.