OBJECTIVE

We sought to develop and pilot a new measure, the Diabetes-Specific Risk-Taking Inventory (DSRI), to assess unhealthy risk-taking behaviors among adolescents with type 1 diabetes.

Methods

Thirteen diabetes health care providers, 30 adolescents with type 1 diabetes (aged 15–19 years, 60% female, mean A1C 8.7% [72 mmol/mol], and 33% on insulin pumps), and the adolescents’ caregivers rated the perceived riskiness of each item on the DSRI. Adolescents completed the DSRI, for which they reported how often they engaged in 34 behaviors that could place them at risk for acute complications of type 1 diabetes or out-of-range blood glucose levels. Adolescents also completed the risk-taking subscale from the Risk-Taking and Self-Harm Inventory for Adolescents, and parents completed the parent-proxy Diabetes Management Questionnaire. Mean A1C during the previous year was obtained via medical chart review.

Results

Results indicated good content validity and feasibility for using the DSRI in a research context, as 100% of adolescents who consented to the study completed the measure. The DSRI was positively correlated with general risk-taking and negatively correlated with diabetes management, indicating preliminary evidence of convergent validity. The DSRI also correlated with A1C.

Conclusion

This pilot study extends our previous work in developing a conceptual model for illness-specific risk-taking. The DSRI is a promising new measure to assess illness-specific risk-taking, including unhealthy risk-taking behaviors, for adolescents with type 1 diabetes.

Across pediatric chronic diseases, self-management (previously known as “adherence”) of medical regimens (i.e., the extent to which a person’s behavior coincides with medical advice [1]) is typically at its worst during adolescence (2). It is estimated that 50–75% of medical treatments are not followed as prescribed by youth with chronic conditions and that the self-management behavior patterns established in adolescence carry forward into adulthood (3). Thus, the period of adolescence is a crucial time for behavioral interventions that aim to improve disease self-management behaviors.

Recent estimates suggest that >18,000 youths <20 years of age are newly diagnosed with type 1 diabetes every year in the United States (4), making it one of the most common childhood chronic illnesses. People with type 1 diabetes must follow a complex daily treatment regimen in which insulin dosing and blood glucose monitoring must be coordinated with food intake and physical activity to keep glucose levels within a target range. There can be severe health consequences as glucose levels become further from the desired range (5). Short-term consequences include seizures resulting from severe hypoglycemia and life-threatening diabetic ketoacidosis resulting from a lack of insulin (6). Long-term consequences of years of chronic hyperglycemia include retinopathy, nephropathy, neuropathy, cardiovascular disease, and neurocognitive impairment (7,8).

Despite extensive research on self-management behaviors in adolescents with type 1 diabetes (9), adolescents in the United States continue to have higher A1C values than any other age-group and are least likely to meet American Diabetes Association–recommended glycemic targets (10). A1C, an estimate of average glucose levels during the past 8–12 weeks, is routinely assessed at quarterly diabetes clinic appointments (11). Psychosocial factors (e.g., depression, anxiety, diabetes distress, and family conflict) as well as sociodemographic factors (e.g., income, family structure, parental educational attainment, insurance status, and access to health care) have been documented to be associated with self-management (12). Still, our understanding of modifiable factors underlying self-management difficulties is incomplete. Interventions targeting these known factors demonstrate only modest effect sizes on glycemic outcomes, and participants’ A1C levels often remain in a suboptimal range (13).

There are few new conceptualizations for why adolescents are particularly vulnerable to self-management difficulties (9). Prior researchers have noted that poor self-management behaviors such as forgoing insulin or not taking a pill could be a form of risk-taking (14). The current authors took this assertion a step further by creating a model for and formally defining “illness-specific risk-taking” as a type of behavior in which youths make decisions about disease management that put them at risk for acute complications of their health condition or adverse events resulting from potential consequences of their health condition (15). For adolescents, risk-taking behaviors can either precede poor self-management choices (e.g., excessive alcohol use can lead to an inability to participate in type 1 diabetes self-care tasks) or take the form of poor self-management choices (e.g., engaging in unhealthy weight loss strategies by restricting insulin). On the basis of our extensive clinical and research experience with adolescents with type 1 diabetes, our team wanted to study this new construct, “diabetes-specific risk-taking,” as an overlooked factor that potentially contributes to poor adolescent self-management of the complex daily type 1 diabetes regimen.

As a first step to explore illness-specific risk-taking in youth with type 1 diabetes, the authors developed a new measure, the Diabetes-Specific Risk-Taking Inventory (DSRI), and piloted it in the current study. The primary objectives in this exploratory pilot study were to determine whether the DSRI was feasible to administer to our target population, whether it captured the construct of perceived diabetes risk, and whether it was related to established measures of type 1 diabetes self-management, general risk-taking, and A1C. Specifically, the study had the following exploratory aims: 1) to determine whether the DSRI demonstrates adequate content validity among adolescents with type 1 diabetes, their parents, and diabetes health care professionals such that all groups would rate each behavior as at least “a little risky”; 2) to determine whether the DSRI would demonstrate adequate feasibility in a research context such that ≥95% of adolescents who entered the study would complete the DSRI; and 3) to explore whether scores on the DSRI are related to other established measures (i.e., self-management and general risk-taking behaviors) and glycemic level (average A1C) over the past year.

Participants and Measure Development Procedures

To develop the DSRI, the authors followed guidelines for creating patient-reported outcome measures (16) and psychological measures. The steps followed in this process are described below and summarized in Figure 1 (17).

FIGURE 1

Process for developing the DSRI. PROMs, patient-reported outcome measures.

FIGURE 1

Process for developing the DSRI. PROMs, patient-reported outcome measures.

Close modal

Step 1: Review of established measures. First, the authors reviewed established measures of frequency of general adolescent risk-taking behaviors. The DSRI structure was modeled on other adolescent risk-taking inventories such as the Youth Risk Behavior Surveillance System (YRBSS) (18) and the Risk-Taking Inventory for Adolescents (RTI-A) (19). For example, we modeled instructions for the DSRI on those of the RTI-A, which aim to normalize risk-taking behaviors (i.e., “this questionnaire asks about a number of things that young people with diabetes sometimes do”), ensure the adolescent is in a private setting (i.e., “please complete this questionnaire on your own”), encourage truthfulness (i.e., “please try to answer as truthfully as possible”), and assure confidentiality of participants’ answers (i.e., “your answers are kept strictly private and confidential; they will not be shared with your parents or your medical team”). We also modeled the recall period (i.e., “have you ever…?”), the response options (i.e., “never,” “once,” “more than once,” or “many times”), and the scoring from the RTI-A (19). Finally, we used language from the YRBSS that described alcohol and drug use to define these behaviors in our measure (e.g., “Used drugs [for example, pills that were not prescribed to you, marijuana, etc.”]).

Step 2: Generation of initial items. To generate items for the measure, the authors consulted with other psychologists who specialize in diabetes to generate an initial list of example type 1 diabetes risk-taking behaviors.

Step 3: Focus groups/interviews. Once a full draft of the measure was completed, the authors conducted one-on-one interviews and group meetings with diabetes health care providers to explore new ideas and items that should be added to the measure and obtain feedback on current items. Five endocrinologists, two endocrinology nurse practitioners, and four psychologists with experience in diabetes participated in these focus groups and interviews.

Step 4: Development of initial measure. On the basis of this input, items were refined, dropped, or added until consensus was reached among the investigators.

Step 5: Cognitive debriefings with target population. Next, four adolescents with type 1 diabetes (aged 17–19 years) completed the first draft of the DSRI, either before or after their regular diabetes clinic visit, and provided feedback on the measure through a semi-structured cognitive debriefing interview. These cognitive debriefings (as recommended by the U.S. Department of Health and Human Services [16]) were conducted to ensure participants’ understanding of the meaning and intent of questions. Participants completed the questionnaire in a private space separate from the clinic area, and cognitive debriefings were conducted immediately after they finished the survey. Each cognitive debriefing interview was audio-recorded and transcribed.

Step 6: Incorporation of feedback into the DSRI. The research team reviewed the cognitive debriefings to come to a consensus regarding what aspects to change on the DSRI. Suggested changes included adding a response option to the response scale to allow for an answer in between “once a year” and “once a month,” deleting one item from the scale, and adding a “not applicable” response option.

Step 7: Expert review. After implementing changes from the cognitive debriefings, the research team sent a survey to diabetes health care providers at the same institution, for which providers anonymously reported on the perceived level of “riskiness” for each DSRI item (1 = “not risky,” 2 = “a little risky,” 3 = “moderately risky,” 4 = “very risky,” and 5 = “extremely risky”). An administrative assistant who was not involved with the research team distributed the online survey link to the diabetes health care providers at our institution, and the survey did not ask for any personal or demographic information of the health care providers.

Step 8: Piloting of measure with target population. For the pilot stage, research staff recruited adolescents with type 1 diabetes and their caregivers from the pediatric diabetes clinic of a major metropolitan children’s hospital in the south central region of the United States. Research staff reviewed electronic medical records (EMRs) to screen for adolescents with upcoming appointments who met the following inclusion criteria: had no major health problems other than type 1 diabetes, had been diagnosed for at least 1 year, had a recent A1C level between 6.0% (42 mmol/mol) and 13% (119 mmol/mol), were between the ages of 15 and 20 years, both adolescent and caregiver were able to speak and read English, and were not currently enrolled in another behavioral study. This age range was chosen because older adolescents typically spend more time away from their parents, manage their diabetes care more independently, and would be more likely to engage in risk-taking behaviors.

This pilot study was approved by our institutional review board. Research staff sent 61 eligible families a letter describing the study and called to inquire about their interest in participating. Six families were determined via follow-up to be ineligible based on inclusion/exclusion criteria, 16 were unable to participate because of scheduling conflicts, and 9 families declined participation because of time limitations (n = 7) or lack of interest in research participation (n = 2). If a family could be reached ahead of their clinic appointment, a research staff member scheduled a research appointment to occur immediately before or after the adolescent’s diabetes clinic appointment. If a family could not be reached before the clinic appointment, a research staff member left a message on their home phone to inform them that they would be approached in clinic and would have the opportunity to participate in a survey study after their appointment.

At the research appointment, the research staff member explained the study to the adolescent and parent and obtained their consent (or assent, as appropriate) before proceeding. The study appointment lasted ∼20–30 minutes. It included parent and adolescent questionnaires and a brief diabetes-related interview, which asked about the youth’s diabetes regimen and use of technology (e.g., insulin pump and continuous glucose monitoring [CGM] system). After completing the DSRI frequency items, participants also rated each item on the DSRI as to “How likely is this behavior to result in a serious health problem for an older teen/young adult with type 1 diabetes?” on a 5-point Likert scale (1 = “not risky,” 2 = “a little risky,” 3 = “moderately risky,” 4 = “very risky,” and 5 = “extremely risky”). Parents also rated their own perceived level of riskiness for each DSRI item. Adolescents and their caregivers were compensated for their time ($12 for parking, $10 for adolescent participation, and $10 for caregiver participation).

Measures

Demographics

Demographic information was obtained from a questionnaire completed by the caregiver that included questions about race/ethnicity, insurance status, and parental level of education. A research study staff member obtained adolescents’ date of birth and sex from their EMR.

Diabetes-Specific Risk-Taking

Adolescents completed the newly developed DSRI measure. The DSRI piloted in this study was a 34-item measure that assesses the frequency with which adolescents (aged 15–20 years) engage in diabetes-specific risk-taking behaviors (Table 1). Respondents reported the frequency of each behavior using a 6-point Likert scale (5 = “daily,” 4 = “weekly,” 3 = “monthly,” 2 = “every few months,” 1 = “yearly,” or 0 = “never”) or “not applicable (N/A).” The N/A option was provided for items for which a participant may not have had the opportunity to engage in that risk-taking behavior (e.g., a person might answer N/A to “recorded a made-up blood sugar number in your logbook” because he or she did not have a logbook). The current study used the total mean score, with a higher score indicating more frequent diabetes-specific risk-taking behavior. Any item that was answered “N/A” was dropped from the mean score calculation. The measure demonstrated excellent internal reliability (α = 0.92) in the current study.

TABLE 1

Items From the DSRI

How often have you …
 Guessed the number of carbs in a snack/meal you were eating when the nutrition information was right there?
 Decided not to wear a diabetes ID?
 Engaged in physical activity without wearing a diabetes ID?
 Engaged in physical activity without first checking your blood glucose?
 Participated in an organized sport or activity without telling the coach you have diabetes?
 Ate without first checking your blood glucose?
 Drunk anything besides water/diet drinks without first checking your blood glucose?
 Felt your blood glucose might be low and did not check?
 Felt your blood glucose might be low and did not treat it?
 Had high blood sugars and did not check ketones?
 Gone without checking your blood glucose for at least 24 hours?
 Told someone you checked your blood glucose when you really had not?
 Entered made-up blood glucose numbers in your logbook?
 Reported a made-up blood glucose number to someone?
 Taken insulin without checking your blood glucose first?
 Ate without taking short-acting insulin to cover carbs (except when blood glucose is low)?
 Drunk anything besides water without taking short-acting insulin to cover carbs?
 Taken less insulin than you knew you needed, based on carbs, blood glucose, and exercise?
 Taken more insulin than you knew you needed, based on carbs, blood glucose, and exercise?
 Waited until you were out of insulin before telling your parents or getting more from the pharmacy?
 Waited until your glucagon had expired before telling parents or getting more from the pharmacy?
 Gone without taking insulin for at least 24 hours?
 Told your parents you had taken insulin when you really had not?
 Told your doctor you had taken insulin when you really had not?
 Driven a car without first checking your blood glucose?
 Driven a car without fast-acting carbs within reach?
 Drunk alcohol without eating extra carbs?
 Drunk alcohol when no one around knew you had diabetes?
 Drunk alcohol without wearing a diabetes ID?
 Gone to sleep after drinking alcohol with no plan for checking blood sugars during the night?
 Gotten drunk to the point where you could not take care of your diabetes?
 Used drugs (for example, pills that were not prescribed to you, marijuana, etc.) when no one around knew you had diabetes?
 Had sex without first checking your blood glucose?
 Gone to sleep after sex without checking blood glucose? 
How often have you …
 Guessed the number of carbs in a snack/meal you were eating when the nutrition information was right there?
 Decided not to wear a diabetes ID?
 Engaged in physical activity without wearing a diabetes ID?
 Engaged in physical activity without first checking your blood glucose?
 Participated in an organized sport or activity without telling the coach you have diabetes?
 Ate without first checking your blood glucose?
 Drunk anything besides water/diet drinks without first checking your blood glucose?
 Felt your blood glucose might be low and did not check?
 Felt your blood glucose might be low and did not treat it?
 Had high blood sugars and did not check ketones?
 Gone without checking your blood glucose for at least 24 hours?
 Told someone you checked your blood glucose when you really had not?
 Entered made-up blood glucose numbers in your logbook?
 Reported a made-up blood glucose number to someone?
 Taken insulin without checking your blood glucose first?
 Ate without taking short-acting insulin to cover carbs (except when blood glucose is low)?
 Drunk anything besides water without taking short-acting insulin to cover carbs?
 Taken less insulin than you knew you needed, based on carbs, blood glucose, and exercise?
 Taken more insulin than you knew you needed, based on carbs, blood glucose, and exercise?
 Waited until you were out of insulin before telling your parents or getting more from the pharmacy?
 Waited until your glucagon had expired before telling parents or getting more from the pharmacy?
 Gone without taking insulin for at least 24 hours?
 Told your parents you had taken insulin when you really had not?
 Told your doctor you had taken insulin when you really had not?
 Driven a car without first checking your blood glucose?
 Driven a car without fast-acting carbs within reach?
 Drunk alcohol without eating extra carbs?
 Drunk alcohol when no one around knew you had diabetes?
 Drunk alcohol without wearing a diabetes ID?
 Gone to sleep after drinking alcohol with no plan for checking blood sugars during the night?
 Gotten drunk to the point where you could not take care of your diabetes?
 Used drugs (for example, pills that were not prescribed to you, marijuana, etc.) when no one around knew you had diabetes?
 Had sex without first checking your blood glucose?
 Gone to sleep after sex without checking blood glucose? 

General Risk-Taking

The Risk-Taking and Self-Harm Inventory for Adolescents (RTSHI-A) (15) was used to measure adolescent self-reported frequency (0 = “never,” 1 = “once,” 2 = “more than once,” or 3 = “many times”) with which they have engaged in general risk-taking behaviors such as risky driving and substance use. The RTSHI-A includes two independently validated subscales—one for risk-taking and one for self-harm. Only the nine items comprising the risk-taking subscale were used in the current study. To score the risk-taking subscale, a total mean score was calculated, with a higher score indicating more frequent risk-taking behavior. The risk-taking subscale had good internal consistency in a large sample of youth aged 11–18 years (α = 0.85) (15) as well as in the current study (α = 0.85).

Adolescent Self-Management

Parents completed the Diabetes Management Questionnaire (DMQ) (20) as a parent proxy report of adolescents’ self-management behaviors with respect to their prescribed diabetes medical regimen. The DMQ assesses diabetes self-management behaviors (insulin management, physical activity, diet, hyperglycemia, hypoglycemia, and blood glucose monitoring) over the past month. Parents were asked to report on the frequency of their child’s self-management behaviors on a 5-point Likert scale ranging from “almost never” to “almost always.” A total mean score was calculated for the DMQ, with a higher score indicating more frequent self-management behaviors. The DMQ parent report has demonstrated validity for pediatric populations aged 8–18 years (20) and had acceptable reliability in the current study (α = 0.79). Although there is a self-report version of the DMQ, we decided to include only a parent proxy of self-management to reduce the amount of time required for the study and limit the burden on the family, as parents and adolescents completed their respective questionnaires at the same time.

Average Glycemic Level

The adolescents’ average glycemic levels were assessed with measurement of A1C. A1C is routinely collected at diabetes clinic visits via fingerstick capillary blood sample and analyzed using the DCA 2000 (Siemens-Bayer, Munich, Germany). From EMR review, research staff members obtained A1C values over the previous year, including the value obtained at the clinic visit on the day of the research appointment. Participants had anywhere from two to six A1C values, depending on the frequency with which they came to the clinic. Average A1C values over the previous year were calculated to obtain an estimated measure of average glycemic level over the past year. An A1C goal of <7.5% (58 mmol/mol) is recommended for school-aged children and adolescents (8).

Data Analysis Plan

Analyses were performed using the SPSS statistics package, v. 24 (IBM Corp., Armonk, NY). Descriptive analyses were conducted to determine the frequency and distribution of each variable. To assess content validity, the research team examined ratings of how “risky” each group of reporters (adolescents, parents, and health care providers) evaluated each item on the DSRI. To assess feasibility, the research team examined study engagement data to determine how many participants completed the DSRI. Finally, the research team explored whether there was preliminary evidence for concurrent and criterion validity of our questionnaire by conducting bivariate correlations to determine associations between diabetes-specific risk-taking and general risk-taking, diabetes-specific risk-taking and self-management, and diabetes-specific risk-taking and average glycemic level.

Descriptive Statistics

Descriptive statistics (i.e., mean, SD, and percentages) of the current sample were calculated for demographic and health status variables and scores on the DSRI, RTSHI-A, and DMQ (Table 2). Average A1C was 8.7 ± 1.4% (72 mmol/mol), which was above the recommended target range for this age-group, although it is consistent with other samples of similarly aged youths with type 1 diabetes (e.g., Petitti et al. [21]).

TABLE 2

Adolescents’ Demographic and Clinical Characteristics (n = 30)

Youth age, years 17.2 ± 1.2 (15–19) 
Youth sex
 Female
 Male 

18 (60)
12 (40) 
Race/ethnicity
 White
 Black
 Hispanic
 Asian
 Other/biracial 

15 (50)
3 (10)
8 (27)
1 (3)
3 (10) 
Duration of diabetes, years 6.8 ± 3.7 (1.2–16.5) 
Treatment type
 Insulin pump
 Injections 

10 (33)
20 (67) 
CGM use within the past month
 CGM
 No CGM 

6 (20)
24 (80) 
Parental education
 Less than high school
 High school/general education diploma
 Partial college
 College education
 Graduate degree 

3 (10)
7 (23)
9 (30)
8 (27)
3 (10) 
Type of insurance
 Private
 Public 

18 (60)
12 (40) 
DSRI score 2.4 ± 0.8 (1.0–3.7) 
RTSHI-A score 0.5 ± 0.6 (0–2.6) 
DMQ score 3.5 ± 0.7 (2.0–4.7) 
Average A1C over past year, % 8.7 ± 1.4 (6.5–14) 
Average A1C over past year, mmol/mol 72 (48–130) 
Youth age, years 17.2 ± 1.2 (15–19) 
Youth sex
 Female
 Male 

18 (60)
12 (40) 
Race/ethnicity
 White
 Black
 Hispanic
 Asian
 Other/biracial 

15 (50)
3 (10)
8 (27)
1 (3)
3 (10) 
Duration of diabetes, years 6.8 ± 3.7 (1.2–16.5) 
Treatment type
 Insulin pump
 Injections 

10 (33)
20 (67) 
CGM use within the past month
 CGM
 No CGM 

6 (20)
24 (80) 
Parental education
 Less than high school
 High school/general education diploma
 Partial college
 College education
 Graduate degree 

3 (10)
7 (23)
9 (30)
8 (27)
3 (10) 
Type of insurance
 Private
 Public 

18 (60)
12 (40) 
DSRI score 2.4 ± 0.8 (1.0–3.7) 
RTSHI-A score 0.5 ± 0.6 (0–2.6) 
DMQ score 3.5 ± 0.7 (2.0–4.7) 
Average A1C over past year, % 8.7 ± 1.4 (6.5–14) 
Average A1C over past year, mmol/mol 72 (48–130) 

Data are mean ± SD (range) or n (%).

Content Validity

To examine content validity of the DSRI measure, the research team analyzed the level of perceived “riskiness” for each item among three groups of reporters: diabetes health care providers (n = 13, including 9 pediatric endocrinologists, 2 nurse practitioners, and 1 certified diabetes educator), adolescents with type 1 diabetes (n = 30), and parents of adolescents with type 1 diabetes (n = 28). Among each group of reporters, the mean perceived level of risk for each DSRI item ranged from 3.0 ± 1.0 (“moderately risky”) to 4.8 ± 0.4 (“very risky”) for adolescents, 3.3 ± 1.2 (“moderately risky”) to 4.8 ± 0.8 (“very risky”) for parents, and 2.5 ± 0.8 (“a little risky”) to 4.9 ± 0.3 (“very risky”) for diabetes health care providers.

Feasibility

We examined feasibility of using the DSRI in a research setting with our target population by examining the rate of decline among participants who consented. Of the 30 adolescents who consented to the study, 100% completed the entire DSRI. No participants opted out of completing the DSRI. Also, despite specific instructions that they could skip any item they did not feel comfortable answering, all participants answered every item. It took participants ∼1–3 minutes to complete the questionnaire, and none asked questions about any of the items for clarification.

Concurrent and Construct Validity

First, associations between the DSRI and other adolescent measures were examined. As shown in Table 3, adolescents who engaged more frequently in diabetes-specific risk-taking also engaged in less frequent diabetes self-management behaviors, as reported by parents (r = –0.56, P <0.01). These adolescents also reported more frequent general risk-taking (r = 0.41, P <0.05), despite diabetes self-management and general risk-taking being unrelated (r = –0.22, P >0.05). Second, associations between diabetes-specific risk-taking and A1C were examined. As shown in Table 3, adolescents who engaged more frequently in diabetes-specific risk-taking behaviors had higher average A1C levels over the previous year (r = 0.57, P <0.01).

TABLE 3

Correlations Between DSRI and Established Measures of Self-Management and General Risk-Taking Behavior and A1C

1234
1. DSRI     
2. DMQ −0.560** —   
3. RTSHI 0.419* −0.22 —  
4. Average A1C 0.575** −0.243 0.292 — 
1234
1. DSRI     
2. DMQ −0.560** —   
3. RTSHI 0.419* −0.22 —  
4. Average A1C 0.575** −0.243 0.292 — 

Bold type indicates significance.

**

Correlation is significant at the 0.01 level (two-tailed).

*

Correlation is significant at the 0.05 level (two-tailed).

This pilot study is the first to test an innovative behavioral construct—illness-specific risk-taking—using the DSRI, a new measure targeting youths with type 1 diabetes. The DSRI is a carefully constructed measure of specific types of adolescent risk-taking behaviors that could exacerbate risks of type 1 diabetes. This pilot study provides strong initial psychometrics for the DSRI, including concurrent validity with other established measures of similar constructs and with a biological variable (A1C).

While correlated with a measure of self-management, the DSRI also provides distinct information about behaviors that overlap with general risk-taking behaviors, as evident by its association with a measure of general risk-taking. This risk-taking information is not assessed in the DMQ, as is evident by the lack of association in this study between the DMQ and the RTSHI-A. The authors of the current study will evaluate other forms of validity and psychometric properties of the DSRI in an upcoming larger, longitudinal study. Given the lack of research examining type 1 diabetes–specific risk-taking, the DSRI has the potential to launch a new area of pediatric health psychology research and intervention development for adolescents struggling with disease self-management.

This pilot study was the first step in establishing content validity of the DSRI and providing support for our conceptualization of the construct “diabetes-specific risk-taking” (12). It demonstrated strong content validity of the DSRI by including many types of content experts (i.e., health care providers, adolescents with type 1 diabetes, and their parents) in the development, cognitive debriefings, and content validity testing of the DSRI. Another strong indicator of concurrent validity was our finding that the DSRI was significantly related to other established measures of similar constructs, including the RTSHI-A and the DMQ, as well as average glycemic level over the last year (mean A1C). These associations are especially promising given that this pilot study had a small sample and thus was underpowered to detect significance in such associations. Another strength is that the associations were significant across reporters and methods of measurement (i.e., self-report data were correlated with both parent report data and objective glycemic data).

Additionally, this study establishes the feasibility of using this measure with adolescents in a research setting. One potential barrier to the feasibility of the DSRI was that adolescents might have felt uncomfortable with the content and refuse to complete the questionnaire. On the contrary, 100% of our participants completed the DSRI without any expressed concern about the content. The research team also checked the responses on the DSRI for any participant who might have responded with N/A or “never” to all of the items, but this did not happen. Finally, participants were able to complete the questionnaire in a reasonable amount of time, in a paper-and-pencil format, without needing to clarify instructions or the meaning of specific questions.

Because of the small sample size of this pilot study, it is important to note that these findings may not be generalizable; thus, a follow-up study with a larger sample is necessary to replicate our findings. Additionally, this study included a convenience sample of adolescents who attended their regular diabetes clinic visit during study enrollment. Ideally, a follow-up study would include a larger, more diverse sample. Although the DSRI demonstrated excellent preliminary psychometrics (internal reliability and validity), a larger study is needed to confirm and expand on these positive indications.

We propose that diabetes-specific risk-taking is an important new concept that has the potential to extend our understanding of behaviors that can contribute to the poor health outcomes often seen in adolescents with type 1 diabetes. Further research is needed to support the findings from this initial pilot study. A larger, longitudinal study is planned to allow for a more complete assessment of psychometric properties of the DSRI, which will provide a greater understanding of which diabetes-specific risk-taking behaviors are most common and most closely associated with poor health outcomes (e.g., acute complications such as severe hypoglycemia and diabetic ketoacidosis) and a thorough examination of moderating factors such as age, sex, family environment, and social influences. This future study could inform the development of screening and intervention programs to reduce diabetes-specific risk-taking behaviors in a clinical setting. Additionally, given that the measure is rather lengthy (34 items), future research is planned to identify ways to condense the measure for clinical use so that it could be used to more quickly screen for the most relevant diabetes-specific risk-taking behaviors. To accomplish this, we may use factor analysis and/or sensitivity/specificity analyses. Finally, future studies using the DSRI may wish to consider also using a self-report measure of adolescent adherence. We chose to use only a parent proxy measure to reduce response burden and the overall time required for families to participate in the study. However, adolescents may be better reporters of their own self-management behaviors, given that they are likely to be more independent with diabetes care tasks.

Although the psychometric properties of this measure require further investigation, this study has important clinical implications that could be immediately considered and addressed in routine care. For example, the International Society of Pediatric and Adolescent Diabetes (ISPAD) recommends that adolescents with diabetes receive education around “driving, alcohol, drugs, sexual health, and contraception” (22), but these topics are often skipped during busy clinical appointments (23). The findings from this study highlight the importance of assessing risk-taking in youths with type 1 diabetes and thus support ISPAD’s recommendations for diabetes health care providers to discuss these topics in their clinic appointments. Eventually, the DSRI could be used as the basis for in-depth discussions around risk-taking between adolescents and diabetes clinicians (24). The DSRI has significant potential to be a useful clinical tool to screen for high-risk behaviors in adolescents with type 1 diabetes and identify those who might benefit the most from targeted education and behavioral intervention.

Acknowledgments

This work was supported by the National Institute of Diabetes and Digestive and Kidney Diseases under grant 1K12 DK097696 (principal investigator B.J.A.). The authors appreciate the participation of the families of youths with type 1 diabetes who dedicated their time to this study. They also thank the health care providers in the Division of Endocrinology at Texas Children’s Hospital who contributed to the content validation of this questionnaire.

Duality of Interest

No potential conflicts of interest relevant to this article were reported.

Author Contributions

R.M.W. contributed to the conception of the research project, measure development, research strategy, data collection, data analyses, and writing of the manuscript. D.D.S. contributed to editing of the measure, research strategy, and editing of the manuscript. B.J.A. contributed to measure development, research strategy, and editing of the manuscript. R.M.W. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

1.
Modi
AC
,
Pai
AL
,
Hommel
KA
, et al
.
Pediatric self-management: a framework for research, practice, and policy
.
Pediatrics
2012
;
129
:
e473
e485
2.
Taddeo
D
,
Egedy
M
,
Frappier
JY
.
Adherence to treatment in adolescents
.
Paediatr Child Health
2008
;
13
:
19
24
3.
Rapoff
MA
.
Consequences of nonadherence and correlates of adherence
. In
Adherence to Pediatric Medical Regimens: Issues in Clinical Child Psychology
. 2nd ed.
Rapoff
MA
, Ed.
Berlin, Germany
,
Springer
,
2010
, p.
33
45
4.
Centers for Disease Control and Prevention
.
National Diabetes Statistics Report, 2020
.
5.
Brink
SJ
.
Complications of pediatric and adolescent type 1 diabetes mellitus
.
Curr Diab Rep
2001
;
1
:
47
55
6.
Rewers
A
,
Chase
HP
,
Mackenzie
T
, et al
.
Predictors of acute complications in children with type 1 diabetes
.
JAMA
2002
;
287
:
2511
2518
7.
Bragd
J
,
Adamson
U
,
Bäcklund
LB
,
Lins
PE
,
Moberg
E
,
Oskarsson
P
.
Can glycaemic variability, as calculated from blood glucose self-monitoring, predict the development of complications in type 1 diabetes over a decade?
Diabetes Metab
2008
;
34
:
612
616
8.
Lin
A
,
Northam
EA
,
Werther
GA
,
Cameron
FJ
.
Risk factors for decline in IQ in youth with type 1 diabetes over the 12 years from diagnosis/illness onset
.
Diabetes Care
2015
;
38
:
236
242
9.
Martinez
K
,
Frazer
SF
,
Dempster
M
,
Hamill
A
,
Fleming
H
,
McCorry
NK
.
Psychological factors associated with diabetes self-management among adolescents with type 1 diabetes: a systematic review
.
J Health Psychol
2018
;
23
:
1749
1765
10.
Foster
NC
,
Beck
RW
,
Miller
KM
, et al
.
State of type 1 diabetes management and outcomes from the T1D Exchange in 2016–2018
.
Diabetes Technol Ther
2019
;
21
:
66
72
11.
American Diabetes Association
.
12. Children and adolescents: Standards of Medical Care in Diabetes—2018
.
Diabetes Care
2018
;
41
(
Suppl. 1
):
S126
S136
12.
Gonzalez
JS
,
Tanenbaum
ML
,
Commissariat
PV
.
Psychosocial factors in medication adherence and diabetes self-management: implications for research and practice
.
Am Psychol
2016
;
71
:
539
551
13.
Hood
KK
,
Rohan
JM
,
Peterson
CM
,
Drotar
D
.
Interventions with adherence-promoting components in pediatric type 1 diabetes: meta-analysis of their impact on glycemic control
.
Diabetes Care
2010
;
33
:
1658
1664
14.
Schwartz
DD
,
Axelrad
ME
.
Healthcare Partnerships for Pediatric Adherence: Promoting Collaborative Management for Pediatric Chronic Illness Care
.
Berlin, Germany
,
Springer International Publishing
,
2015
15.
Wasserman
R
,
Anderson
BJ
,
Schwartz
DD
.
Illness-specific risk-taking in adolescence: a missing piece of the nonadherence puzzle for youth with type 1 diabetes?
Diabetes Spectr
2017
;
30
:
3
10
16.
U.S. Department of Health and Human Services U.S. FDA Center for Drug Evaluation and Research
;
U.S. Department of Health and Human Services FDA Center for Biologics Evaluation and Research
.
U.S. Department of Health and Human Services FDA Center for Devices and Radiological Health
.
Guidance for industry: patient-reported outcome measures: use in medical product development to support labeling claims: draft guidance
.
Health Qual Life Outcomes
2006
;
4
:
79
17.
Holmbeck
GN
,
Devine
KA
.
Editorial: an author’s checklist for measure development and validation manuscripts
.
J Pediatr Psychol
2009
;
34
:
691
696
18.
Centers for Disease Control and Prevention
.
YRBS questionnaire content, 1991–2017
.
19.
Vrouva
I
,
Fonagy
P
,
Fearon
PR
,
Roussow
T
.
The risk-taking and self-harm inventory for adolescents: development and psychometric evaluation
.
Psychol Assess
2010
;
22
:
852
865
20.
Mehta
SN
,
Nansel
TR
,
Volkening
LK
,
Butler
DA
,
Haynie
DL
,
Laffel
LM
.
Validation of a contemporary adherence measure for children with type 1 diabetes: the Diabetes Management Questionnaire
.
Diabet Med
2015
;
32
:
1232
1238
21.
Petitti
DB
,
Klingensmith
GJ
,
Bell
RA
, et al
.
Glycemic control in youth with diabetes: the SEARCH for Diabetes in Youth study
.
J Pediatr
2009
;
155
:
668
672.e1-3
22.
Cameron
FJ
,
Garvey
K
,
Hood
KK
,
Acerini
CL
,
Codner
E
.
ISPAD clinical practice consensus guidelines 2018: diabetes in adolescence
.
Pediatr Diabetes
2018
;
19
(
Suppl. 27
):
250
261
23.
Bento
SP
,
Campbell
MS
,
Soutullo
O
,
Cogen
FR
,
Monaghan
M
.
Substance use among adolescents and young adults with type 1 diabetes: discussions in routine diabetes care
.
Clin Pediatr (Phila)
2020
;
59
:
388
395
24.
de Wit
M
,
Pulgaron
ER
,
Pattino-Fernandez
AM
,
Delamater
AM
.
Psychological support for children with diabetes: are the guidelines being met?
J Clin Psychol Med Settings
2014
;
21
:
190
199
Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. More information is available at https://www.diabetesjournals.org/content/license.