**2.3.1 Adolescents and parent**

**Clinical, Socio-Demographic Questionnaire** (Pereira et al., 2010) that reports gender and age in adolescents and their family members as well as metabolic control (glycated hemoglobin) and duration of disease, in the adolescent.

**Brief Illness Perception Questionnaire** – Brief-IPQ – Broadbent et al. (2006), (Portuguese version of Figueiras & Alves, 2007). The Brief-IPQ is a 9 items questionnaire, measuring cognitive and emotional representations of illness, that includes nine dimensions of illness perceptions: consequences, timeline, personal control, treatment control, identity, concern, coherence, emotional representation and causal representations. Both adolescents and parents answered the questionnaire. *Higher results indicate a more threatening perception of illness.* Due to the fact that each subscale includes only one item, it is not possible to calculate an alpha. As a result, like in the original version, pearson correlations between dimensions were calculated. In adolescents, significant correlations were present between consequences and emotional representation (r=.635), personal control and coherence (r=.511) and personal control and treatment control (r=.371). In the family sample, significant correlations were obtained between consequences and emotional representation (r=.558), personal control and coherence (r=.522) and between concern and coherence (r=.324).

#### **2.3.2 Adolescents**

*Self Care Inventory – SCI* - La Greca, A. (1992), (Portuguese version of Almeida & Pereira, 2010). It´s a 14 items questionnaire assessing adherence to diabetes treatment's

Predictors of Adherence, Metabolic Control and Quality of Life in Adolescents with Type 1 Diabetes 127

First, descriptive statistics were performed to find the rate of adherence to self-care, metabolic control and quality of life. Hierarchical regression analyses were later performed to identify the best predictors of adherence to self-care, metabolic control and quality of life. Due to the size of the sample, regression analysis were first performed taking in consideration all variables ,except illness perceptions, and later including only them in the regression equation. The first regression was performed using the method *enter* since the selection of variables was based on previous research. The second regression, due to its

For both regressions, the variables considered in the first step were socio-demographic and clinical variables i.e. gender of the adolescent, duration of disease and values of glycated hemoglobin. In the first regression analysis, the second step included adolescents' psychosocial variables i.e. family support, quality of life, adherence and school support. The third step included family variables i.e. family functioning and coping. In the second regression analysis, the second step included adolescents' illness perceptions and the third

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

**Control** 

M SD M SD M SD M SD M SD M SD

**Male** 7.05 4.10 4.00 0.59 9.00 1.72 75.91 16.96 106.63 13.15 27.93 6.34

**Female** 6.17 3.19 4.13 0.40 9.13 1.44 83.55 18.19 107.81 11.73 28.21 5.92

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

**Quality of Life** 

**Family Support** 

**School Support** 

exploratory nature, was performed using the stepwise method.

step included family member's illness perceptions.

difference was significant (t(83)=-2.004; p=.048) (table 1).

**Diabetes Adherence Metabolic** 

**3. Data analysis** 

**4. Results** 

**4.1 Sample caracteristics** 

Variables **Duration of** 

Psychosocial variables

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

recommendations regarding self care that includes four subscales: blood glucose regulation, insulin and food regulation, exercise and emergency precautions. *Higher results indicate more adherence.* Only the full scale was considered in the present study. Internal consistency in the original version was .80 and in this sample was .73.

**Diabetes Family Behaviour Scale** – DFBS – McKelvey et al., (1993), (Portuguese version of Almeida & Pereira (in press). DFBS is a 47 items questionnaire that assesses family support given to the adolescent in diabetes self care. It is composed of two subscales: Guidance-Control *(*15 items) and Warmth-Caring (15 items). The remaining 17 items do not belong to any of the subscales. *High results indicate less social support.* Internal consistency, in the original version, was .86, .81 and .79 for the full scale, guidance-control and warmth-caring, respectively. The Portuguese version showed an alpha of .91 (total scale), .76 (guidance-control) and .81 (warmth-caring.). In this study only the full scale was considered (alpha of .75).

**Diabetes Quality of Life** – DQoL - Ingersoll & Marrero (1991), (Portuguese version of Almeida & Pereira (2008). DQol is a 52 items questionnaire that assesses quality of life in patients with diabetes that includes three subscales: impact of diabetes (23 items); worries towards diabetes (11 items) and satisfaction (towards treatment: 7 items; towards life in general: 10 items) and one item that assesses health and quality of life. Higher results indicate lower quality of life. In the original version, the alpha for the total subscale was .92, followed by .86 (satisfaction), .85 (impact of diabetes) and .82 (worries towards diabetes). In this sample alphas were .89 (total scale), .71 (impact on diabetes), .82 (worries towards diabetes) and .87 (satisfaction). All the subscales were considered in the hypothesis testing.

**School Support** (Pereira & Almeida, 2009). School Support is a 6 items questionnaire that measures school support (e.g. healthy snacks available in cafeteria) and peer support regarding daily diabetes' management (e.g. feeling supported by fiends regarding diabetes). *Higher results indicate more school support. T*he alpha in this sample was .81.

#### **2.3.3 Parent**

**Family Assessment Device** – FAD – Epstein et al., (1983), (Portuguese version provided by Ryan et al., 2005). It´s a 60 items questionnaire distributed by seven subscales: Problems Solving, Communication, Roles, Affective Responsiveness; Affective Involvement; Behavior control and General Functioning. *Higher results indicate low family functioning.* In the original version, Epstein, Baldwin and Bishop (1983) found the following results: Problem solving: .74; Communication: .75; Roles: .72; Affective responsiveness: .83; Affective involvement: .78; Behavior Control: .72 and General Functioning: .92. Only the full scale was used in the present study and the alpha, in the present sample, was .93.

**Coping Health Inventory for Parents** – CHIP – McCubbin et al., (1983), (Portuguese version of Pereira & Almeida, 2001). CHIP is a 45 items questionnaire that measures parents' response to management of family life when they have a child who is seriously and/or chronically ill. It includes three subscales: 1) Maintaining family integration, cooperation and an optimist definition of the situation; 2) Maintaining social support, selfesteem and psychological stability; and 3) Understanding the medical situation through communication with other parents and consultation with medical staff. *Higher results indicate better coping*. In the original version, the alpha for the first and second subscale was .79 and .71 for the third. In this sample, alphas were: .65 for the first subscale, .79 for the second and .71 for the last subscale.
