Development and initial validation of the barriers to diabetes adherence measure for adolescents

Abstract

Aims

The purpose of this study was to develop a measure of psychosocial barriers to adherence in adolescents with type 1 diabetes (T1D) and examine relationships to patient characteristics, adherence, and hemoglobin A1C (A1C).

Methods

Barriers to diabetes adherence (BDA) items were generated by researchers, clinicians, and patients. Adolescents aged 12–17 with T1D completed the BDA and an adherence measure. Hemoglobin A1C was obtained through medical chart review.

Results

Factor analysis from 123 adolescents resulted in a 21-item, five-component solution that accounted for 64.5% of the variance. The components were stress and burnout, time pressure and planning, social support, parental autonomy support, and stigma. The BDA total and subscales were internally consistent. The BDA total and some components were associated with adherence and A1C. The BDA was the only predictor of A1C compared to demographic, clinical, and adherence variables (F 6.17, p < .05). Subjects with higher A1C (>8.5) showed a higher level of barriers (F 15.20, p < .001) and a differential profile of barriers (F 5.75, p < .05).

Conclusions

The BDA may be useful in research and clinical settings as a compliment to adherence measures and to tailor educational programs. Additional research is necessary to establish test–retest reliability and discriminant validity.

Introduction

Adolescents with T1D are at risk for poor adherence and glycemic control. Adherence in T1D involves performing many tasks over the course of each day, such as carrying supplies, monitoring blood glucose levels, dosing insulin according to glucose meter readings, managing carbohydrate intake, and taking insulin or using an insulin pump. Many self-care tasks need to be carried out around mealtimes, and in contexts such as restaurants, social gatherings, school, and sports. Given the frequency and nature of these tasks, it is not surprising that adherence and glycemic control in adolescents are often suboptimal [1], [2], [3], [4], [5].

Assessment of adherence frequency is an important part of diabetes research and clinical care. However, self-report questionnaires that assess the frequency of performing self-care tasks do not provide insight regarding reasons for inadequate adherence. Measurement of barriers to adherence compliments adherence frequency, as it provides unique information regarding why adherence frequencies may be suboptimal. The examination of barriers to self-care has provided useful information for educators, and predicted glycemic control [6], [7], [8], [9], [10], [11]. Educational programs may use barriers assessments to tailor intervention content.

Assessment of barriers to adherence within research or clinical practice may also help address social desirability and restriction of range observed on adherence frequency measures. Adherence assessment may be subject to socially desirable responding [12], [13]. Socially desirable responding results in an overestimate of adherence levels. However, unlike adherence, there may be less pressure for socially desirable responses in the assessment of barriers to adherence. There are no “gold standards” or expectations from adults regarding the level of barriers adolescents might experience. Adolescents may more freely admit to factors that act as barriers to adherence, and thus measures of barriers to adherence may be less vulnerable to social desirability effects. This may be particularly true when adolescents report frequent adherence in the context of poor glycemic control. Further investigation using a barriers instrument may provide insights for education and communication about adherence with young patients.

A large number of psychosocial factors have been identified that negatively influence adherence in adolescents with T1D. For example, many studies have identified stress as negatively associated with adherence and glycemic control [14], [15], [16], [17]. Interventions that reduce stress have resulted in improved levels of adherence and glycemic control [18], [19], [20]. Social support has been consistently positively related to adolescent diabetes outcomes [21], [22], [23] as has autonomy support, a key motivation during adolescent development [24], [25], [26]. An area of self-care barriers not well-studied in this population, but very relevant, are those related to the awareness of time, rushing, and planning. Several studies have found that time management, feeling pressured, or an unwillingness to take the time for self-care create barriers to adherence [27], [28], [29].

Finally, stigma and embarrassment associated with chronic illness have been well-documented as a negative influence on quality of life, self-care, and medical outcomes. In pediatric chronic illness, the impact of stigma has been documented in various conditions such as cancer, HIV/AIDS, epilepsy, and cystic fibrosis [30], [31], [32], [33], [34]. Embarrassment and perceived stigma may impact the ability of the individual to communicate needs, seek help, or perform self-care tasks around others. In adolescent diabetes, limited research has documented levels of stigma or the extent to which it impacts self-care [28], [35], [36].

Measures of psychosocial barriers to adherence have been developed for pediatric asthma, cystic fibrosis, and other chronic illnesses [37], [38] as well as adult type 1 and type 2 diabetes [7], [11]. However, available measures focus on and have been validated with adults, focus on one type of self-care task, on one domain of barriers or measure the extent to which a barrier causes negative feelings [6], [39], [40], [41], [42]. Given a lack of appropriate barriers to self-care adherence measures in T1D, the purpose of the present study was to develop and provide initial validation for a measure of barriers to adherence for use in clinical and research settings. We limited development and the review here to psychosocial barriers that have been related to diabetes adherence or glycemic control, are modifiable, and did not already have clinically feasible or brief measures associated with them. For example, we did not attempt to include barriers to adherence that may be better assessed separately, such as depression. A measure of barriers to diabetes adherence will allow clinicians and researchers to identify patient- and population-based barriers to adherence and tailor educational resources to overcome them.