CONTEXT: Although adherence has been identified in practice guidelines for youth with type 1 diabetes to promote optimal glycemic control, there has been no systematic integration of studies investigating the adherence-glycemic control link. This recommendation partly stemmed from the Diabetes Control and Complications Trial (DCCT); however, this trial did not comprehensively measure adherence and had only 195 adolescents.
OBJECTIVE: Our goal was to determine the magnitude of the adherence-glycemic control link in pediatric type 1 diabetes and evaluate its correlates.
METHODS: Our data sources were PubMed (1950–2008), Scopus (1950–2008), and references from reviews in pediatric type 1 diabetes. Studies that included youth under age 19 with type 1 diabetes and a reported association between adherence and glycemic control were eligible for inclusion. Articles were not included if they contained youth with type 2 diabetes, had study samples that overlapped with other studies, or the results came from intervention studies. Of the eligible 26 studies, 21 had sufficient statistical data. Two authors independently extracted information by using a standardized protocol. Agreement between coders was high.
RESULTS: The mean effect size across 21 studies, including 2492 youth with type 1 diabetes, was −0.28 (95% confidence interval: −0.32 to −0.24). As adherence increases, A1c values decrease. No sample or disease characteristics were correlates of the adherence-glycemic control link. Pre-DCCT studies had a mean effect size of −0.32 (8 studies; 1169 participants) compared with −0.25 in post-DCCT studies (13 studies; 1323 participants).
CONCLUSIONS: This meta-analysis supports the adherence-glycemic control link in pediatric type 1 diabetes. The weaker post-DCCT association suggests that the approach to intensive diabetes management has shortcomings. We conclude that this is because of a mismatch between what scientists and clinicians know is the best way to manage pediatric type 1 diabetes and the capabilities of youth and their families.
Adherence to treatment for a chronic disease has been broadly defined as the degree to which a person's behavior corresponds to medical or health advice.1 An individual with type 1 diabetes is prescribed a set of health behaviors on a daily basis that include coordinating the amount and timing of insulin administration with results of frequent blood glucose monitoring, type and amount of dietary intake, and frequency and intensity of physical activity.2 The execution of these behaviors as part of an intensive insulin regimen was shown to promote optimal glycemic control, which in turn was shown to confer risk reduction of the microvascular and macrovascular complications associated with type 1 diabetes.3,4 The Diabetes Control and Complications Trial (DCCT) changed the landscape of type 1 diabetes treatment and solidified the need for optimal adherence to an intensive insulin regimen.
Despite clear scientific findings that achievement of optimal glycemic control has both immediate and long-term health advantages, glycemic control remains suboptimal in subsets of the population of individuals with type 1 diabetes. For example, children and adolescents with type 1 diabetes consistently demonstrate clinic-wide hemoglobin A1c values above target.5–8 Adolescents in the DCCT demonstrated improvements in glycemic control as a result of the intensive insulin regimen; however, both the intensively treated adolescents and those who received standard care had mean A1c values a percentage point higher than the adults.9 This difference is concerning given that a percentage point drop in A1c (eg, 9.0%–8.0%) is associated with a 40% risk reduction of developing retinopathy.10,11
Although suboptimal glycemic control is due in part to growth and puberty,12,13 the imperfect nature of the treatment regimen, and psychosocial variables such as depression,14,15 1 of the gaps in our understanding of the reasons for suboptimal glycemic control is the extent to which adherence is a contributing factor. A primary assumption is that suboptimal adherence leads to suboptimal glycemic control. This is partly based on findings from the DCCT9 and partly on a small number of intervention studies in pediatric type 1 diabetes.16,17 Importantly, it is reflected in practice guidelines that emphasize that promotion of adherence through the integration of information about blood glucose values, dietary intake, and physical activity with the timing and amount of insulin administration contributes to optimal glycemic outcomes.8 However, the DCCT only included 195 adolescents and adherence was not measured in a comprehensive fashion nor was it examined as a correlate of optimal glycemic control.9 In addition, many of the large-scale epidemiologic studies on glycemic control do not report associations with measures of adherence.5–7 A number of smaller scale studies do report on the association between adherence and glycemic control, yet it is not typically the featured aspect of the articles. In addition, these studies have not been systematically reviewed and evaluated with a common metric to determine if individual study findings generalize to expected findings in the population of youth with type 1 diabetes. Thus, to address this gap, a comprehensive review of studies documenting the relationship between adherence and glycemic control in pediatric type 1 diabetes was conducted, followed by a meta-analysis of associations found in these studies. The aims of this work were to (1) determine the magnitude of the association between adherence and glycemic control and (2) elucidate the most robust correlates of this association.
Search Strategy and Study Selection
We searched PubMed (1950 to May 2008) and Scopus (1950 to May 2008) by using combinations of relevant key words associated with pediatric type 1 diabetes, adherence, and glycemic outcomes. The term “type 1 diabetes” was separately paired with each of the terms designating a pediatric population (“children,” “adolescents,” “pediatric”). Then, each resultant search combination was paired with an adherence indicator (“adherence,” “compliance,” “blood glucose monitoring”) and an indicator of glycemic outcomes (“glycemic control,” “A1c”). The results of these searches were then cross-checked and overlapping citations were removed. In addition, the references of 3 recent reviews that included studies on adherence behaviors and glycemic outcomes in pediatric type 1 diabetes18–20 were examined for any articles our searches did not identify. Studies were included if they met the following criteria: (1) individuals in the study sample had type 1 diabetes; (2) the study sample included youth <19 years of age; and (3) the study reported an association between adherence and glycemic control.
Study Coding and Analysis of Reliability
Two authors (Ms Peterson and Ms Rohan) coded studies across 6 characteristics of the sample (year of publication, mean age, percent female, percent ethnic minority, mean duration of type 1 diabetes, and mean A1c), the type of adherence measure (blood glucose meter download, validated self-report measure, or chart review), and the statistics reported for the association (effect size [ES]) between adherence and A1c (eg, correlation coefficient). If more than 1 measure of association between adherence and A1c was reported in the study, we prioritized the adherence-A1c associations in the following way and used the one with the highest “priority score” in the meta-analysis: (1) blood glucose monitoring frequency as indicated by meter download or obtained from chart review documenting actual results of meter downloads; (2) well-validated measure of adherence (child or adolescent report was entered if available; if not, then the caregiver report was used); (3) physician or provider report of adherence or adherence reported in medical chart without report of objective measurement (eg, meter download). This was our standardized protocol for coding adherence and in all cases; adherence indicates performed behaviors around diabetes management as opposed to what was prescribed. The double-entered data were then compared and κ coefficients (for categorical variables) or intra-class correlations (for interval variables) were calculated.
Primary and Secondary Analyses
We were primarily interested in the magnitude of the ES between adherence to the diabetes regimen and glycemic control. Statistical indicators of this ES, in the order of interest, were (1) correlation coefficient, (2) regression or similar multivariate analysis that provided coefficients for individual variables, not just overall R2, and (3) χ2 values (eg, adherence groups by glycemic control groups). Secondary analyses were planned to determine which characteristics of the individual studies and their clinical samples were the most robust predictors of the mean ES. We were interested in 7 study characteristics. The first was an indication of the intensity of the diabetes regimen (eg, basal-bolus versus conventional regimens). Another categorical variable included the age of the sample; whether the sample was primarily made up of children (aged <13 years) or adolescents (aged >13 years). The final categorical variable was the type of adherence measure (eg, objective measure such as the download of blood glucose meter versus validated multidimensional adherence survey). We also examined the association between study ESs and (1) the proportion of female patients, (2) the proportion of ethnic-minority individuals, (3) the mean duration of type 1 diabetes, and (4) the mean indicator of glycemic control, hemoglobin A1c, in the samples.
We used the analytic approach described in Lipsey and Wilson21 and completed analyses in SAS (SAS Institute, Cary, NC) and SPSS (SPSS Inc, Chicago, IL). When calculating the mean ES across studies, we used a weighted least-squares approach in which each ES was weighted by the inverse of its variance.22 This accounts for variability across individual observations (ie, studies) while considering the size of the sample. Specifically, the mean ES, its SE, and 95% confidence intervals were calculated. Homogeneity analyses22 were conducted to determine if the mean ES was accurately predicted by each individual ES. We calculated the Q statistic and determined its statistical significance. Significant Q values indicate rejection of the null hypothesis of homogeneity (ie, there is heterogeneity in study ESs). Finally, on the basis of our a priori plans for secondary analyses, we conducted fixed effects analyses on the 3 categories that potentially influenced the mean ES (intensity of regimen, age group, and type of adherence measure). We calculated between (QB) and within (QW) groups statistics, similar to analysis of variance, to determine if there were differences among the subgroups. A significant QB indicates that a category of the particular variable had a different ES from the other category. A significant QW indicates heterogeneity across the studies in that particular category. Finally, we ran bivariate correlations on individual ESs and sample characteristics. This was done for mean duration of type 1 diabetes, mean A1c, proportion of female patients, and proportion of minority individuals in each sample. These were weighted correlations in that the means or proportions were multiplied by the study's sample size and divided by the total sample size across studies. This was necessary because of the variations in study samples across these characteristics.
Search Results and Coding Reliability
Our searches identified 27 studies that met the inclusion criteria. In addition to meeting these criteria, these 27 studies did not include any type 2 diabetes patients in their samples, none were intervention studies, none used an average of A1c values across multiple time points as their indication of glycemic control, and none were duplicate studies of the same study sample that did meet criteria. Of the 27 studies remaining, 7 required additional information from authors about the ES. More information was needed because the studies reported a nonsignificant ES, but gave no specific values; or the authors reported results of multivariate analysis (eg, R2), but did not provide coefficients or zero-order correlation coefficients. In these 7 cases, all authors were contacted, but only 6 provided information about their findings. Two authors23,24 provided the zero-order correlations and were included in the meta-analysis. Four authors25–28 responded that they no longer had the statistical printouts for the adherence-A1c association so these studies were not included. These 4 authors were not able to provide the exact statistics because of logistic reasons (eg, analyses completed in early 1990s or professional moves since data analysis was complete and results could not be located). The final group of authors29 responded but did not provide the necessary information in time for inclusion in this meta-analysis. Finally, on additional review, we eliminated the only study30 published before 1990 because the measure of adherence in this study, blood glucose monitoring, was a new technique and one that was conducted in a substantially different manner than the other studies employing blood glucose monitoring. Of note, the inclusion of the study did not significantly alter the results of the meta-analysis (results not shown). Thus, the final sample for this meta-analysis was 21 studies.23,24,31–49 Coding of the 8 categories of interest in each study was completed and κ and intra-class correlation coefficients ranged from 0.98 to 1.00. Only 2 discrepancies were found among all the codes and the correct value was validated by the first author and used in analyses.
Table 1 displays the characteristics of the 21 studies. Many studies were conducted in the United States and represent most geographic regions. One study was conducted in the United Kingdom23 and another in Belgium.49 All studies reported on the cross-sectional association between adherence and glycemic control. Study sample sizes ranged from 35 to 360 and the total sample size across studies was 2492. The average age of the sample was 12.1 ± 2.4 years (mean ± SD). All but 1 study reported the proportion of the sample that was female; the average was 50.3%. Eighteen studies reported the race/ethnicity of the sample; the average proportion of nonwhite individuals in those samples was 25.2%. However, there was considerable variability in that the studies ranged from 3% to 74% inclusion of minority participants. Less than half of the studies reported indicators of socioeconomic status (SES) and family structure (eg, proportion of families with 2 caregivers in home), thus we did not collect data on these variables.
Diabetes-specific variables consistently reported in studies were duration of type 1 diabetes and glycemic control. Across the 21 studies, mean duration of type 1 diabetes ranged from 1.7 to 8.3 years; the average across the studies was 4.9 ± 1.4 years. Nineteen studies reported mean values for A1c; these ranged from 6.6% to 11.4%. These studies had considerable variability; SDs indicate A1c values on the participant level ranged from <6% to >14%. The average A1c value reported across these studies was 9.0% ± 1.1%. Two studies reported values as glycosolated hemoglobin; 1 mean value was 8.733 and the other was 12.2.40
The mean ES for these 21 studies, covering 2492 youth with type 1 diabetes, was −0.28. ESs ranged from −0.47 to 0.30; however, only 2 were positive. The mean ES of −0.28 meets conventional standards for a medium effect for a correlation coefficient. This association indicates that as adherence increases, A1c values decrease. On the basis of the SE of 0.02, the 95% confidence interval for the mean ES was −0.32 (lower limit) to −0.24 (upper limit). ESs and confidence intervals for the individual studies are presented in Fig 1.
Factors Associated With Mean ES
Homogeneity analysis revealed significant heterogeneity across the 21 studies (Q = 62.2; P < .0001). We initially intended to examine whether the nature of the insulin regimen (eg, basal-bolus versus conventional) was a factor by determining the proportion of each study sample that used each particular type of regimen. However, most studies did not provide these data. Thus, we classified studies as pre-DCCT or post-DCCT with the assumptions that pre-DCCT studies would include primarily conventional regimens and post-DCCT studies would include a proportion of their samples on basal-bolus regimens. Studies published through 1997 were considered pre-DCCT given the turnaround time in publication and the minimal likelihood that widespread implementation of intensive insulin regimens were done by just 3 years after publication of the DCCT results on adolescents.9 Note that the Kaufman et al study39 was published in 1999, but they reported on a sample from 1995. Thus, it was categorized as pre-DCCT. Eight studies were categorized as pre-DCCT representing 1169 participants. Thirteen studies were classified as post-DCCT and covered 1323 participants. The within-groups comparison for this categorization was statistically significant (QW = 58.9; P < .0001). There was significant heterogeneity among the studies in each category. The between-groups comparison approached statistical significance (QB = 3.3; P < .10); the critical value needed to be above 3.84 to be statistically significant at the P < .05 level. The mean ES for the pre-DCCT studies was −0.32; the mean ES for post-DCCT studies was −0.25. This indicates that the association between adherence and glycemic control was stronger in the pre-DCCT studies, although both effects were still in the medium range.
Next, we examined whether there were differences in the mean ES for children (<13 years old) and adolescents (>13 years old) as well as between studies that used meter downloads or multidimensional adherence surveys. Neither of the between-groups analyses were statistically significant indicating no differences in ESs across either categorization. However, for the studies (n = 4) with meter downloads, there was little variability in their ESs. In contrast, the studies that used adherence surveys (n = 17), there was significant heterogeneity in their ESs. The weighted correlations between each individual ES and proportion of female patients, proportion of ethnic-minority participants, mean duration of diabetes, and mean A1c were all nonsignificant. Results suggest that none of these sociodemographic or diabetes-specific characteristics were associated with the adherence-glycemic control link in these 21 studies.
The results of this meta-analysis show that adherence is linked with glycemic outcomes in pediatric type 1 diabetes. The medium-sized association across studies including 2492 youth with type 1 diabetes was −0.28; as adherence increases, A1c values decrease. Considering that the DCCT clearly showed that adherence to an intensive insulin regimen results in improved glycemic control and subsequently, reduced risk of the long-term complications of the disease, these findings emphasize the benefit to children and adolescents with type 1 diabetes who are adherent. Although this adherence-glycemic control relationship has been assumed and much of pediatric diabetes care is focused on optimizing adherence, this is the first systematic analysis of this association. The results confirm a long-held notion of the importance of measuring adherence50 and provide support for pediatric diabetes care providers to continue to emphasize the importance of adherence.8
A number of study and sample characteristics were examined as factors associated with this relationship between adherence and glycemic control. There were no sociodemographic (age, ethnicity) or disease (duration and A1c levels) characteristics associated with the overall ES or individual study ESs; neither was the type of adherence measure used in the study. This suggests that this relationship between adherence and glycemic control is consistent across the population of children and adolescents with type 1 diabetes and holds up when assessed in different ways. This was surprising given the substantial support for sociodemographic factors such as lower SES, identification as a member of a racial or ethnic minority group, and single caregiver status contributing to both suboptimal adherence and glycemic control.7,51,52 The lack of association found in this meta-analysis among these variables likely reflects the fact that we were specifically examining their association with the adherence-glycemic control link, not each variable considered separately. These findings suggest that all children and adolescents with type 1 diabetes should experience better glycemic outcomes with adherence promotion. Clinicians and researchers are poised to determine the best approaches to tailoring adherence promoting efforts to individual patients and their families while considering the factors that may be barriers to adherence alone.
Of the 21 studies included in this meta-analysis, 90% (n = 19) had a negative association between adherence and glycemic control. Our review of the 2 studies with a positive association23,40 revealed little about why a positive association was found. Specifically, both studies had relatively small sample sizes which may have contributed to different distributions of adherence and glycemic control values from the other studies; however, there were 5 other studies in this meta-analysis with similar sample sizes that showed negative associations. In addition, other studies with similar age ranges (7–1740; 2–823) reported negative associations, suggesting that age is likely not a contributing factor to the study findings. Finally, although these 2 studies did include participants with wide ranges of diabetes duration which could raise the possibility of differing levels of both adherence and glycemic control, we found that no disease characteristics were correlates of the adherence-glycemic control link. Thus, it is unclear to us exactly why these studies were outliers, but they may signal an important consideration about the relationship between adherence and glycemic control. That is, although the preponderance of evidence indicates a negative relationship, there is going to be variability in the magnitude, and possibly direction, of the adherence-glycemic control link among youth with type 1 diabetes. It is important to consider this possibility when treatment planning takes place for the individual patient.
The only characteristic that affected the adherence-glycemic control association was whether the study was conducted before or after the DCCT. Pre-DCCT studies evidenced a larger ES (−0.32) in comparison to the post-DCCT studies (−0.25). Pre-DCCT treatment regimens emphasized carrying out prescribed behaviors such as administering insulin at designated times, checking blood glucose levels, and sticking to a restrictive diet. Post-DCCT treatment regimens are notably more flexible, yet incredibly more complex. The approach is geared toward achieving optimal glycemic control through the achievement of tight glucose control.53 Integrating blood glucose monitoring results with dietary intake and physical activity to determine the amount and timing of insulin administration is the process by which tight glucose control can be achieved. Thus, post-DCCT treatment regimens continue to emphasize carrying out behaviors, but carrying out those behaviors for the purpose of achieving tight glucose control through the integration of multiple pieces of readily available information. The achievement of tight glucose control is also aided by more precision in insulin analogues and devices used to administer insulin. For example, continuous subcutaneous insulin infusion is associated with optimal glycemic outcomes for youth with type 1 diabetes.54–57 Considering the post-DCCT approach and these technologic advances, children and adolescents with type 1 diabetes should be achieving tighter glucose control and better glycemic outcomes. Why then is the association in the post-DCCT era weaker and why are most children and adolescents with type 1 diabetes still demonstrating suboptimal glycemic outcomes?5,6
We believe adherence is harder to achieve in the post-DCCT era because of a mismatch between what scientists and clinicians know is the optimal way to manage type 1 diabetes and the degree to which children and their families are actually capable of managing the disease. Consider the findings from a recent trial investigating continuous glucose monitoring. Children aged 8 to 14 years and adolescents and young adults (aged 15–24 years) who used real-time continuous glucose monitoring did not experience significant improvements in glycemic outcomes compared with the standard care group of similarly aged youth as well as adults above the age of 25.58 It is interesting to note that the sample largely comprised individuals already achieving better glycemic outcomes in comparison to rates found in large-scale epidemiologic studies,5,6 so it would stand to reason that this is a relatively adherent group to start with. The process of intensive insulin management and the ability to integrate all the real-time information may be too complex, demanding, or overwhelming for most youth with type 1 diabetes and their families. It is also a possibility that the measurement of post-DCCT adherence is more difficult to adequately capture given its multidimensional nature. We were unable to examine this statistically; however, measurement of adherence may be a contributing factor to the different findings in the pre-DCCT versus post-DCCT eras.
Considering these points, several adaptations to the current approach to optimizing adherence and glycemic outcomes seem warranted. First, as noted, the adherence-glycemic control link was not associated with sociodemographic and disease characteristics, suggesting that adherence promotion should carry considerable benefit for this population. Adherence promotion should be considered alongside glycemic control improvements as a primary outcome or end point and much more attention should be paid to developing and refining clinic-based interventions to promote adherence. Promoting health behaviors carries substantial benefit in influencing overall health.59 Second, to do this, it seems that getting back to the basics of whether adherence behaviors are conducted is a good first step. Before and after the DCCT, the frequency of blood glucose monitoring has remained a vital predictor of overall adherence and glycemic outcomes.60,61 More complex behaviors and processes cannot be conducted if the most basic behaviors are not executed. Next, conceptualizing and measuring the process of integrating real-time information has to be further developed. This involves understanding the information, responding to the information, and many times coming up with solutions to problems raised by the information. Indicators of diabetes literacy, math skills, and real-time problem solving17 likely make up this notion of present-day adherence.
The results of this meta-analysis and our recommendations should be considered in the context of several limitations. Meta-analytic limitations include the pooling of studies that are made up of differing methodologies and sampling strategies.62 The heterogeneity in individual ESs illustrates that although there is general consensus that poor adherence is associated with suboptimal glycemic control, the magnitude of that relationship may vary from sample to sample and by study design. In addition, we were interested in the most basic relationship between adherence and glycemic control for this meta-analysis. Indeed, glycemic control is multiply determined and future investigations should include an examination of interactions between adherence and other variables when predicting glycemic outcomes. Likewise, a correlation does not imply causation. The studies included in this meta-analysis focused on cross-sectional relationships between adherence and glycemic control. A follow-up meta-analysis on interventions that test whether improving adherence leads to better glycemic outcomes would confirm the direction of this relationship. Finally, other variables that influence glycemic control (eg, growth, puberty, psychosocial factors) were not examined in this study, yet there is compelling evidence for their contribution to suboptimal glycemic outcomes beyond the level of treatment adherence.12–14
The process of conducting this meta-analysis raised several areas of concern about how studies have reported results and consequently, we have identified several recommendations for reporting results. First, sociodemographic variables such as SES and family structure are inconsistently reported. This is surprising given the links between these variables and glycemic outcomes.51,52,63 Future reports should include indicators of SES such as family income, educational achievement, occupational status, or full SES scales. Second, reports rarely show the most basic relationship between adherence, glycemic control, and the other variables of interest in the study. In other words, zero-order correlations or univariate associations should be provided across studies and then followed by the multivariate analyses intended to isolate unique variance associated with predictor variables. Third, means and SDs of adherence measures and their subscales should be reported. As highlighted above, it may be that certain dimensions of adherence are more influential on glycemic outcomes.
In sum, this meta-analysis highlights the link between adherence and glycemic control in pediatric type 1 diabetes as well as areas clinicians and researchers can focus on in future investigations of this link. The results, considered alongside published findings in this area, highlight the importance of ongoing measurement of adherence and adherence promotion. The potential benefits are greater likelihood of achieving optimal glycemic outcomes during childhood and adolescence and reducing the acute and long-term health risks associated with diabetes.
Dr Hood is supported by a career development award from the National Institute of Diabetes and Digestive and Kidney Diseases (K23 DK-073340).
We thank Drs Thomas Frazier and Shoshana Kahana for guidance on developing the study database and conducting data analysis.
- Accepted July 9, 2009.
- Address correspondence to Korey K. Hood, PhD, Cincinnati Children's Hospital, 3333 Burnet Ave, MLC 7039, Cincinnati, OH 45229. E-mail:
Dr Hood had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Financial Disclosure: The authors have indicated they have no financial relationships relevant to this article to disclose.
What's Known on This Subject:
An implicit assumption is that optimal adherence leads to optimal glycemic control. This is reflected in practice guidelines. However, there has been no systematic investigation of this assumption.
What This Study Adds:
This meta-analysis provides a mean ES for the adherence-glycemic control link for nearly 2500 children and adolescents with type 1 diabetes. It will serve as reinforcement for the assumption of this link and should aid in the development of adherence-promoting interventions.
- ↵Haynes RB. Introduction. In: Haynes RB, Taylor DW, Sackett DC, eds. Compliance in Health Care. Baltimore, MD: Johns Hopkins University Press; 1979:1–7
- ↵American Diabetes Association. Standards of medical care in diabetes: 2008. Diabetes Care.2008;31 (suppl 1):S12– S54
- ↵Writing Team for the Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications Research Group. Sustained effect of intensive treatment of type 1 diabetes mellitus on development and progression of diabetic nephropathy: the Epidemiology of Diabetes Interventions and Complications (EDIC) study. JAMA.2003;290 (16):2159– 2167
- ↵Danne T, Mortensen HB, Hougaard P, et al. Persistent differences among centers over 3 years in glycemic control and hypoglycemia in a study of 3805 children and adolescents with type 1 diabetes from the Hvidore Study Group. Diabetes Care.2001;24 (8):1342– 1347
- ↵Mortensen HB, Hougaard P. Comparison of metabolic control in a cross-sectional study of 2873 children and adolescents with IDDM from 18 countries. The Hvidore Study Group on Childhood Diabetes. Diabetes Care.1997;20 (5):714– 720
- ↵Silverstein JH, Klingensmith G, Copeland K, et al. Care of children and adolescents with type 1 diabetes: a statement of the American Diabetes Association. Diabetes Care.2005;28 (1):186– 212
- ↵Group DR. Effect of intensive diabetes treatment on the development and progression of long-term complications in adolescents with insulin-dependent diabetes mellitus: Diabetes Control and Complications Trial. Diabetes Control and Complications Trial Research Group. J Pediatr.1994;125 (2):177– 188
- ↵The relationship of glycemic exposure (HbA1c) to the risk of development and progression of retinopathy in the diabetes control and complications trial. The Diabetes Control and Complications Trial Research Group. Diabetes.1995;44 (8):968– 983
- ↵Lachin JM, Genuth S, Nathan DM, Zinman B, Rutledge BN. Effect of glycemic exposure on the risk of microvascular complications in the Diabetes Control and Complications Trial: revisited. Diabetes.2008;57 (4):995– 1001
- ↵Kovacs M, Obrosky DS, Goldston D, Drash A. Major depressive disorder in youths with IDDM: a controlled prospective study of course and outcome. Diabetes Care.1997;20 (1):45– 51
- ↵Wysocki T, Iannotti R, Weissberg-Benchell J, et al. Diabetes problem solving by youths with type 1 diabetes and their caregivers: measurement, validation, and longitudinal associations with glycemic control. J Pediatr Psychol.2008;33 (8):875– 884
- ↵Winkley K, Ismail K, Landau S, Eisler I. Psychological interventions to improve glycaemic control in patients with type 1 diabetes: systematic review and meta-analysis of randomised controlled trials. BMJ.2006;333 (7558):65
- Hampson SE, Skinner TC, Hart J, et al. Behavioral interventions for adolescents with type 1 diabetes: how effective are they? Diabetes Care.2000;23 (9):1416– 1422
- ↵Lipsey MW, Wilson DB. Practical Meta-analysis. Thousand Oaks, CA: Sage Publications, Inc; 2001
- ↵Hedges LV, Olkin I. Statistical Methods for Meta-analysis. Orlando, FL: Academic Press; 1985
- ↵Wysocki T, Taylor A, Housh BS, Linscheid TR, Yeates KO, Naglieri JA. Deviation from developmentally appropriate self-care autonomy: association with diabetes outcomes. Diabetes Care.1996;19 (2):119– 125
- ↵Anderson B, Ho J, Brackett J, Finkelstein D, Laffel L. Parental involvement in diabetes management tasks: relationships to blood glucose monitoring adherence and metabolic control in young adolescents with insulin-dependent diabetes mellitus. J Pediatr.1997;130 (2):257– 265
- Bougnères PF, Landais P, Mairesse AM, et al. Improvement of diabetic control and acceptability of a three-injection insulin regimen in diabetic adolescents: a multicenter controlled study. Diabetes Care.1993;16 (1):94– 102
- ↵Schafer LC, Glasgow RE, McCaul KD, Dreher M. Adherence to IDDM regimens: relationship to psychosocial variables and metabolic control. Diabetes Care.1983;6 (5):493– 498
- ↵Auslander WF, Thompson S, Dreitzer D, White NH, Santiago JV. Disparity in glycemic control and adherence between African-American and Caucasian youths with diabetes: family and community contexts. Diabetes Care.1997;20 (10):1569– 1575
- ↵Berg CA, Butler JM, Osborn P, et al. Role of parental monitoring in understanding the benefits of parental acceptance on adolescent adherence and metabolic control of type 1 diabetes. Diabetes Care.2008;31 (4):678– 683
- ↵Davis CL, Delamater AM, Shaw KH, et al. Parenting styles, regimen adherence, and glycemic control in 4- to 10-year-old children with diabetes. J Pediatr Psychol.2001;26 (2):123– 129
- ↵Ellis DA, Podolski CL, Frey M, Naar-King S, Wang B, Moltz K. The role of parental monitoring in adolescent health outcomes: impact on regimen adherence in youth with type 1 diabetes. J Pediatr Psychol.2007;32 (8):907– 917
- ↵Greening L, Stoppelbein L, Konishi C, Jordan SS, Moll G. Child routines and youths' adherence to treatment for type 1 diabetes. J Pediatr Psychol.2007;32 (4):437– 447
- ↵Hanson CL, De Guire MJ, Schinkel AM, Kolterman OG, Goodman JP, Buckingham BA. Self-care behaviors in insulin-dependent diabetes: evaluative tools and their associations with glycemic control. J Pediatr Psychol.1996;21 (4):467– 482
- ↵Iannotti RJ, Nansel TR, Schneider S, et al. Assessing regimen adherence of adolescents with type 1 diabetes. Diabetes Care.2006;29 (10):2263– 2267
- ↵Kaufman FR, Halvorson M, Carpenter S. Association between diabetes control and visits to a multidisciplinary pediatric diabetes clinic. Pediatrics.1999;103 (5 pt 1):948– 951
- ↵Lewandowski A, Drotar D. The relationship between parent-reported social support and adherence to medical treatment in families of adolescents with type 1 diabetes. J Pediatr Psychol.2007;32 (4):427– 436
- ↵Lewin AB, Heidgerken AD, Geffken GR, et al. The relation between family factors and metabolic control: the role of diabetes adherence. Journal Pediatr Psychol.2006;31 (2):174– 183
- ↵Miller VA, Drotar D. Discrepancies between mother and adolescent perceptions of diabetes-related decision-making autonomy and their relationship to diabetes-related conflict and adherence to treatment. J Pediatr Psychol.2003;28 (4):265– 274
- ↵Miller VA, Drotar D. Decision-making competence and adherence to treatment in adolescents with diabetes. J Pediatr Psychol.2007;32 (2):178– 188
- ↵Naar-King S, Idalski A, Ellis D, et al. Gender differences in adherence and metabolic control in urban youth with poorly controlled type 1 diabetes: the mediating role of mental health symptoms. J Pediatr Psychol.2006;31 (8):793– 802
- ↵Wiebe DJ, Berg CA, Korbel C, et al. Children's appraisals of maternal involvement in coping with diabetes: enhancing our understanding of adherence, metabolic control, and quality of life across adolescence. J Pediatr Psychol.2005;30 (2):167– 178
- ↵Stewart SM, Emslie GJ, Klein D, Haus S, White P. Self-care and glycemic control in adolescents with type 1 diabetes. Child Health Care.2005;34 (3):239– 244
- ↵Dorchy H, Roggemans MP, Willems D. Glycated hemoglobin and related factors in diabetic children and adolescents under 18 years of age: a Belgian experience. Diabetes Care.1997;20 (1):2– 6
- ↵Phillip M, Battelino T, Rodriguez H, Danne T, Kaufman F. Use of insulin pump therapy in the pediatric age-group: consensus statement from the European Society for Paediatric Endocrinology, the Lawson Wilkins Pediatric Endocrine Society, and the International Society for Pediatric and Adolescent Diabetes, endorsed by the American Diabetes Association and the European Association for the Study of Diabetes. Diabetes Care.2007;30 (6):1653– 1662
- Pankowska E, Blazik M, Dziechciarz P, Szypowska A, Szajewska H. Continuous subcutaneous insulin infusion versus multiple daily injections in children with type 1 diabetes: a systematic review and meta-analysis of randomized control trials. Pediatr Diabetes.2008;10 (1):52– 58
- ↵Weissberg-Benchell J, Antisdel-Lomaglio J, Seshadri R. Insulin pump therapy: a meta-analysis. Diabetes Care.2003;26 (4):1079– 1087
- ↵Johnson SB, Kelly M, Henretta JC, Cunningham W, Tomer A, Silverstein J. A longitudinal analysis of adherence and health status in childhood diabetes. J Pediatr Psychol.1992;17 (5):537– 553
- ↵Gavin L, Wysocki T. Associations of paternal involvement in disease management with maternal and family outcomes in families with children with chronic illness. J Pediatr Psychol.2006;31 (5):481– 489
- ↵Shapiro S. Meta-analysis/Shmeta-analysis. Am J Epidemiol.1994;140 (9):771– 778
- Copyright © 2009 by the American Academy of Pediatrics