**4.1 Sample caracteristics**

The sample consisted of 85 adolescents, 51% males and 49% females. Their age ranged from 12 to 19 with an average of 15.13 (SD=1.97), 15.12 for males (SD=2.00) and 15.14 for females (SD=1.96). Glycated hemoglobin in the sample was, in average, 9.06 (SD=1.58) specifically 9.00 (SD=1.72) for boys and 9.13 (SD=1.44) for girls. Therefore, girls had a poor metabolic control than boys but they were all at high risk. Average of duration of diabetes was 6.61 years (SD=3.68) with boys being diagnosed longer (M=7.05 years; SD=4.10) than girls (M=6.17 years; SD=3.19). In our sample, girls reported better adherence to self-care, less social support, higher school support and family social support when compared to boys but differences were non-significant. Girls showed less quality of life than boys and this difference was significant (t(83)=-2.004; p=.048) (table 1).


Statistics: M (mean), SD (standard deviation)

Table 1. Characteristics of the Adolescents' Sample by Clinical, Socio-demographic and Psychosocial variables

Predictors of Adherence, Metabolic Control and Quality of Life in Adolescents with Type 1 Diabetes 129

Table 2. Predictors of Adherence, Metabolic Control and Quality of Life in Adolescents on Gender, Duration of Disease, Glycated Hemoglobin, Family Support, School Support and

Parental Coping (N=85 adolescents; N= 85 family members)

74% of adolescents lived with their nuclear families, 15% belonged to monoparental families, 9.4% to stepfamilies and, only, 1.2% lived in an extended family. 20% of family members, who participated in the study, were fathers and 80% mothers. Average age for fathers was 46 years (SD=4.55) and for mothers was 44 years (SD=6.19).

#### **4.2 Predictors of adherence, metabolic control and quality of life in adolescents on gender, duration of disease, glycated hemoglobin, family support, school support and parental coping**

When all variables were included in the model, adherence was predicted by gender of adolescent (p<.05), glycated hemoglobin (p<.05) and family support (p<.001), explaining 30% of the total variance. None of the family variables predicted adherence. Taking in consideration what a high score means, in each instrument, results showed that low perception of family support, gender (being male) and high glycated hemoglobin (bad metabolic control) predicted lower adherence to diabetes self-care.

Metabolic Control was predicted by family support (total) (p<.05), adherence (total) (p<.05), quality of life (total) (p<.05) and parental coping (understanding the medical situation) (p<.05), explaining 15.9% of total variance. As a result, higher adherence of adolescent to self-care and parental understanding of the medical situation predicted lower levels of glycated hemoglobin (better metabolic control). On the other hand, low quality of life and low perception of family support predicted high values of glycated hemoglobin (poor metabolic control).

Quality of life was predicted by gender (p<.05), glycated hemoglobin (p<.05) and school support (total) (p<.01) explaining 26.5% of the total variance. Higher values of glycated hemoglobin (poor metabolic control) predicted lower quality of life. On the other hand, higher adherence and a higher school support predicted better quality of life. Like in adherence, none of the family variables predicted quality of life, in adolescents. Table 2 shows the results.

#### **4.3 Predictors of adherence, metabolic control and quality of life in adolescents on glycated hemoglobin and illness representations**

Overall, adherence was predicted by personal control of adolescent's illness representations (p<.001) and family's representation of timeline (p<.05) explaining 20.3% of the total variance. Thus, lower adolescents' perception of personal control predicted lower adherence to self care and higher family perception of diabetes duration (timeline) predicted higher adherence to self care, in adolescents.

Metabolic control, in adolescents, was predicted by emotional representation of adolescents' illness perceptions (p<.001) and by family's perceptions of illness coherence (p<.05), explaining 16.6% of the total variance. Therefore, higher adolescents' perception of emotional representation (diabetes seen as a threatening disease) predicted higher values of glycated hemoglobin (poor metabolic control) and lower family's comprehension of diabetes predicted higher values of glycated hemoglobin.

Quality of life was predicted by glycated hemoglobin (p<.05), adolescent's perception of consequences (p<.05) and emotional representation (p<.05) explaining 31.6% of the total variance. Higher perception of the consequences of diabetes by adolescents and higher perception of emotional representation (diabetes seen as a threatening disease) predicted lower quality of life. None of the family variables predicted adolescent's quality of life. Table 3 shows the results.


Table 2. Predictors of Adherence, Metabolic Control and Quality of Life in Adolescents on Gender, Duration of Disease, Glycated Hemoglobin, Family Support, School Support and Parental Coping (N=85 adolescents; N= 85 family members)

74% of adolescents lived with their nuclear families, 15% belonged to monoparental families, 9.4% to stepfamilies and, only, 1.2% lived in an extended family. 20% of family members, who participated in the study, were fathers and 80% mothers. Average age for

**4.2 Predictors of adherence, metabolic control and quality of life in adolescents on gender, duration of disease, glycated hemoglobin, family support, school support and** 

When all variables were included in the model, adherence was predicted by gender of adolescent (p<.05), glycated hemoglobin (p<.05) and family support (p<.001), explaining 30% of the total variance. None of the family variables predicted adherence. Taking in consideration what a high score means, in each instrument, results showed that low perception of family support, gender (being male) and high glycated hemoglobin (bad

Metabolic Control was predicted by family support (total) (p<.05), adherence (total) (p<.05), quality of life (total) (p<.05) and parental coping (understanding the medical situation) (p<.05), explaining 15.9% of total variance. As a result, higher adherence of adolescent to self-care and parental understanding of the medical situation predicted lower levels of glycated hemoglobin (better metabolic control). On the other hand, low quality of life and low perception of family support predicted high values of glycated hemoglobin (poor

Quality of life was predicted by gender (p<.05), glycated hemoglobin (p<.05) and school support (total) (p<.01) explaining 26.5% of the total variance. Higher values of glycated hemoglobin (poor metabolic control) predicted lower quality of life. On the other hand, higher adherence and a higher school support predicted better quality of life. Like in adherence, none of the family variables predicted quality of life, in adolescents. Table 2

**4.3 Predictors of adherence, metabolic control and quality of life in adolescents on** 

Overall, adherence was predicted by personal control of adolescent's illness representations (p<.001) and family's representation of timeline (p<.05) explaining 20.3% of the total variance. Thus, lower adolescents' perception of personal control predicted lower adherence to self care and higher family perception of diabetes duration (timeline) predicted higher

Metabolic control, in adolescents, was predicted by emotional representation of adolescents' illness perceptions (p<.001) and by family's perceptions of illness coherence (p<.05), explaining 16.6% of the total variance. Therefore, higher adolescents' perception of emotional representation (diabetes seen as a threatening disease) predicted higher values of glycated hemoglobin (poor metabolic control) and lower family's comprehension of diabetes

Quality of life was predicted by glycated hemoglobin (p<.05), adolescent's perception of consequences (p<.05) and emotional representation (p<.05) explaining 31.6% of the total variance. Higher perception of the consequences of diabetes by adolescents and higher perception of emotional representation (diabetes seen as a threatening disease) predicted lower quality of life. None of the family variables predicted adolescent's quality of life.

fathers was 46 years (SD=4.55) and for mothers was 44 years (SD=6.19).

metabolic control) predicted lower adherence to diabetes self-care.

**glycated hemoglobin and illness representations** 

predicted higher values of glycated hemoglobin.

adherence to self care, in adolescents.

Table 3 shows the results.

**parental coping** 

metabolic control).

shows the results.


Table 3. Predictors of Adherence, Metabolic Control and Quality of Life in Adolescents on Glycated Hemoglobin and Illness Representations (N=85 adolescents; N= 85 fam. members) Predictors of Adherence, Metabolic Control and Quality of Life in Adolescents with Type 1 Diabetes 131

In this study, adolescent's gender (i.e. being male) predicted lower adherence to diabetes self-care and higher quality of life. An association between gender and low adherence to diabetes, in adolescents girls, particularly regarding exercise, has been found in the literature (Patino et al., 2005). Girls with diabetes show lower quality of life than boys because they seemed to worry more regarding their illness (Grey et al., 1998; Rocha, 2010; Hoey et al., 2001). In fact, low quality of life, in girls, has been associated to more difficulties and worries regarding diabetes and less satisfaction with metabolic control. Girls enter puberty earlier than boys and a weak metabolic control may be associated to physiological changes, normal to adolescence, such as increased levels of hormones responsible for insulin

In terms of predictors of **a**dherence, taking in consideration the final model, higher values of glycated hemoglobin (poor metabolic control) predicted lower adherence to diabetes selfcare and lower quality of life. These results are in accordance with the literature. Adolescents have more difficulties with metabolic control suggesting that hormonal changes, associated with puberty and the decline on adherence to self-care, were responsible for these results (Helgeson et al., 2009). In another study, glycated hemoglobin explained a small variance of quality of life in adolescents with diabetes suggesting that higher levels of glycated hemoglobin (poor metabolic control) had negative effects on the adolescent's perception of quality of life (Malik & Koot, 2009). In a study that addressed metabolic control and quality of life, good metabolic control (measured by glycated hemoglobin) was a

Higher family support predicted higher adherence and better metabolic control (lower levels of glycated hemoglobin). These results are in accordance with the literature. Family support has been found to be a predictor of good metabolic control (Lewin et al., 2006). In fact, low family support was associated to low adherence to diabetes self-care and, indirectly, to a poor metabolic control. La Greca and Bearman, in 2002, suggested that family support predicts adolescents' adherence to diabetes self-care because family support is an important factor on the daily management of diabetes' self-care tasks in adolescents. Higher family support was found to be a predictor of higher adherence to self-care and good metabolic control suggesting the direct impact of parental support on diabetes' management tasks influencing , as a result, adherence and metabolic control, in the adolescent (Duke et al., 2008; Ellis et al., 2007). In a Portuguese sample of adolescents, family support was found

In the present study, a lower perception of personal control predicted lower adherence to diabetes self-care in adolescents. Beliefs in the effectiveness of treatment (control over the illness) were found to predict adherence to dietary self-care (Delamater, 2009). When the benefits, compared to costs of following the diabetes regimen were considered lower, diabetes was perceived as a less threatening disease and adherence to self care in diabetes ,

Higher family perception of diabetes' duration, as an illness, predicted higher adherence of adolescents to diabetes self-care. In an attempt to understand if there were differences between illness representations in adults with type 2 diabetes and their partners, a relationship was found between partner's perceptions of the duration of diabetes (timeline) and treatment suggesting that partners' perceptions could influence positively patients' adherence to diabetes self-care (Searle et al., 2007). Based on these result, the same may be true for the dyads parent-adolescent. In fact, parent's perception as a long last condition in

to predict adherence in adolescents with type 1 diabetes (Pereira et al., 2008).

**5. Discussion** 

resistance (Carroll & Shade, 2005).

predictor of better quality of life (Hoey et al.,2005).

as a result, was poor (Patino et al., 2005).
