Journal of Pediatric Psychology, Vol. 27, No. 8, 2002, pp.

665676

The Diabetes Social Support Questionnaire-Family Version: Evaluating Adolescents Diabetes-Specic Support From Family Members
Downloaded from http://jpepsy.oxfordjournals.org/ by guest on March 10, 2012

Annette M. La Greca, PhD, and Karen J. Bearman, MS

University of Miami
Objective: To develop and evaluate the Diabetes Social Support Questionnaire-Family Version (DSSQ-Family) for adolescents with type 1 diabetes. Methods: Normative and individualized approaches to scoring were examined. Also examined were associations between diabetes-specic family support and adolescents age, disease duration, gender, emotional support from family and friends, and treatment adherence. The most supportive family behaviors were identied as well. Adolescents (n = 74) rated 58 DSSQ-Family behaviors on their supportiveness and frequency and completed measures of emotional support from family and friends and treatment adherence. After eliminating nonsupportive items, the Total DSSQ-Family and ve areas of diabetes care (insulin, blood testing, meals, exercise, emotions) were scored for frequency (normative approach) and frequency support (individualized approach). The upper quartile of the DSSQ-Family items was identied as most supportive. Results: Scores from the DSSQ-Family had high internal consistency. Higher frequency and individualized ratings were related to younger adolescent age and to more family emotional support and cohesion, but not to friend support or family conict (in general). The individualized ratings were signicant predictors of adolescents adherence, even when controlling for age and general levels of family support. The most supportive family behaviors reected emotional support for diabetes. Conclusions: The DSSQ-Family is a useful clinical and research tool for measuring adolescents perceptions of diabetes-specic family support. Future interventions should stress family support for management tasks, taking into account the adolescents perceptions of supportive behaviors. Additional research is needed with culturally diverse adolescents and with other chronic pediatric conditions. Key words: type 1 diabetes; adolescents; family; social support; adherence; friends.

Type 1 or insulin-dependent diabetes is a complex and challenging disease to manage, especially for
All correspondence should be sent to Annette M. La Greca, Department of Psychology, P. O. Box 249229, University of Miami, Coral Gables, Florida 33124. E-mail: alagreca@umiami.edu.

adolescents, who have been found to be less adherent and in poorer metabolic control than preadolescent youths (e.g., Anderson, Auslander, Jung, Miller, & Santiago, 1990; Anderson, Miller, Auslander, & Santiago, 1981; La Greca, Follansbee, & Skyler, 1990).

2002 Society of Pediatric Psychology

666

La Greca and Bearman

Given that intensive self-management may be necessary to prevent or forestall serious health complications (e.g., retinopathy, renal disease) associated with diabetes (Diabetes Control and Complications Trial, 1993), efforts to understand and promote optimal self-management behaviors among adolescents are needed. In this regard, family support for diabetes may be critical to adolescents disease management. Diabetes has been referred to as a family disease (La Greca, 1998), as family members are involved in daily management tasks, such as meal planning, well into youngsters teenage years (Follansbee, 1989; La Greca et al., 1990). Moreover, adolescents who have more supportive, cohesive families have been found to have better metabolic control (e.g., Anderson et al., 1981), better treatment adherence (La Greca et al., 1995), and better psychosocial adaptation (Hanson, De Guire, Schinkel, Henggeler, & Burghen, 1992; see also Burroughs, Harris, Pontius, & Santiago, 1997). Despite the importance of family support for adolescents diabetes management, little is known about the specific ways in which family members provide support for adolescents diabetes care. Such information would be extremely useful for designing supportive family interventions for adolescents with diabetes. In fact, a number of prominent pediatric researchers have noted the dearth of effective family interventions for youths with chronic disease (Drotar, 1997; Kaslow et al., 1997; Kazak, 1997). Efforts to enhance family support for adolescents diabetes care may be facilitated by the development of tools that assess the specific ways that family members provide support for diabetes. Along these lines, La Greca et al. (1995) developed a structured interview to assess family members support for diabetes care, the Diabetes Social Support Interview (DSSI-Family). Those authors found that older adolescents reported less diabetes-specific support from their families than younger adolescents and that family support for diabetes care was a significant predictor of adolescents treatment adherence. The results highlighted the potential value of keeping families involved in adolescents diabetes care, even as adolescents mature. Despite these positive findings using the DSSIFamily, a standard questionnaire to assess family support for diabetes management may offer several advantages relative to a structured interview. First, by providing a list of supportive behaviors in questionnaire format, recall problems may be minimized. In contrast, a structured interview (e.g., the DSSI-

Family) requires adolescents to generate the family behaviors they find to be supportive and, thus, depends on adolescents ability to recall supportive behaviors. Second, scoring is substantially simplified with a standard questionnaire, in comparison to the complex and time-consuming scoring for a structured interview. Simplified scoring could be advantageous in pediatric research when time is at a premium (La Greca & Lemanek, 1996). Finally, the use of a standard set of supportive family behaviors would also allow researchers to compare adolescents responses across pediatric samples and across informants (e.g., parent and adolescent) more readily than would be the case with a structured interview and idiosyncratic adolescent responses. Thus, the primary objective of this study was to develop and evaluate the Diabetes Social Support Questionnaire-Family Version (DDSQ-Family), a paper-and-pencil measure of perceived family support for adolescents diabetes care. Items were generated primarily from adolescent reports of supportive family behaviors and reflect five key areas of diabetes management: insulin administration, blood glucose testing, meals, exercise, and emotional support (Chase, 1992). Four specific goals were identified to aid in the development and evaluation of this measure; they involved examining (1) scoring options, (2) construct validity, (3) predictive validity (i.e., how well scores predict treatment adherence), and (4) clinical utility (i.e., identifying the most supportive behaviors as a guide to intervention). Two scoring methods were examined for the DSSQ-Family: one based on a normative approach to evaluating social support and the other based on an individualized approach. Both methods have been used in prior research (e.g., La Greca et al., 1995; Procidano & Heller, 1983), although it was hypothesized that the individualized approach might be a better predictor of adolescents treatment adherence. Specifically, a normative approach assumes that all items reflect supportive behaviors; thus, only the frequencies of the behaviors are examined. An individualized approach allows for individual interpretations of each items supportiveness; thus, the frequency of each item is adjusted for its perceived supportiveness (i.e., frequency support). Although the individualized approach is more timeconsuming (i.e., two sets of ratings are obtained), there may be some benefit to considering the individuals perception of what is supportive. Interviews have shown that adolescents vary substantially in their perceptions of supportive behaviors (La Greca

Downloaded from http://jpepsy.oxfordjournals.org/ by guest on March 10, 2012

Diabetes Social Support Questionnaire

667

et al., 1995). Therefore, these two approaches to scoring the DSSQ-Family were evaluated with respect to their internal consistencies, intercorrelations, and associations with other measures. To examine construct validity, associations between perceived family support for diabetes care and demographic variables were evaluated. We hypothesized that younger adolescents would report more support from their families than older adolescents, as prior research has found family members to be more involved in the diabetes management of younger adolescents (Anderson et al., 1990; La Greca et al., 1990). In addition, associations between family support, disease duration, and gender were examined, although no predictions were made. Previous work has found few, if any, associations between family support for diabetes care and adolescents gender or disease duration (La Greca et al., 1995). The construct validity of the DSSQ-Family also was examined using other measures of support. We hypothesized that adolescents perceptions of family support for diabetes care would be substantially and significantly related to their reports of general (i.e., non-disease-specific) family support and cohesion, but not to perceptions of emotional support from friends or to family conflict. Such patterns would provide support for the convergent and discriminant validity of the DSSQ-Family. To examine predictive validity, we evaluated the DSSQ-Family as a predictor of adolescents treatment adherence. Based on prior research showing that more supportive or cohesive families have adolescents with better adherence or metabolic control (e.g., Anderson et al., 1981; Hansen, Henggeler, & Burgen, 1987; Hauser et al., 1990), we hypothesized that higher levels of perceived family support for diabetes care would predict better adherence. Moreover, as a stringent test of predictive validity, we examined whether the DSSQ-Family would incrementally predict adherence after first controlling for general levels of perceived family support and cohesion. If the DSSQ-Family added significantly to the prediction of adherence, it would support the incremental and predictive validity of the DSSQ-Family and highlight the importance of family support for adolescents diabetes care. Finally, to evaluate the clinical utility of the DSSQ-Family, we examined the specific family behaviors that adolescents perceived to be most supportive for their diabetes care. Identifying the most supportive behaviors could provide health care professionals and family members with concrete ideas

regarding the kinds of family behaviors adolescents typically find useful and helpful. Thus, we examined the content of the most supportive items and also evaluated whether the frequency or individualized ratings for these most supportive items predicted adolescents treatment adherence. In summary, the primary objective of this study was to develop a questionnaire measure of perceived family support for diabetes care (DSSQ-Family). In evaluating the measure, the specific study goals were (1) to compare a normative and individualized approach to scoring; (2) to examine the concurrent validity of the DSSQ-Family, expecting that younger adolescents would report more perceived family support for diabetes care than older adolescents, but not expecting any differences as a function of disease duration or gender; the DSSQ-Family also was expected to be related to general measures of family support and cohesion, but not to measures of friend support or family conflict; (3) to examine the predictive validity of the DSSQ-Family, expecting that more perceived family support would be associated with better adherence, even when controlling for general levels of family support and cohesion; and (4) to identify the types of diabetes-specific family behaviors that adolescents perceived as most supportive for their diabetes care. The ultimate purpose of this line of research is to use the information for developing empirically supported family interventions for adolescents with diabetes.

Downloaded from http://jpepsy.oxfordjournals.org/ by guest on March 10, 2012

Method
Participants Seventy-four adolescents (45 boys, 29 girls) with type 1 diabetes were interviewed during routine appointments for diabetes care at the pediatric endocrinology clinic of a large childrens hospital in the Midwest (also described in Bearman & La Greca, 2002). Adolescents ranged from 11 to 18 years (M = 14.2 years, SD = 2.3 years), with a mean diabetes duration of 5.2 years (SD = 3.5). The sample was 83.8% Caucasian, 10.8% African American, and 4.4% other or mixed ethnicities. Most adolescents resided in twoparent homes (78%); some came from single-parent families (12%) or had other living arrangements (9%), such as a parent and an adult relative. The average household size (including the adolescent) was 4.3 persons (range = 2 to 9; median = 4), with an average of 2.05 adults (range = 1 to 5; median = 2) and

668

La Greca and Bearman

2.26 children (range = 1 to 7; median = 2). The sample was primarily lower to upper middle class, with 1.8 adults (median = 2) employed outside the home, and with an average family income of $49,400 (range = $7,900 to $200,000; median = $40,000). Procedure All consecutive adolescents who received treatment for their diabetes over a 4-month period (for routine check-ups or problems with diabetes) at a childrens hospital in the Midwest were invited to participate. Adolescents 11 to 18 years of age were recruited if they had had diabetes for at least 6 months and were accompanied by a parent or guardian who could provide informed consent. Ninety percent of the eligible families agreed to participate, and those who declined most commonly cited time constraints as the reason. Parents and adolescents gave written consent prior to participation, and adolescents were compensated $20.00 for their time. Measures Diabetes Social Support Questionnaire-Family Version. The DSSQ-Family (copy available from first author) was developed to assess adolescents perceptions of family behaviors that are supportive for their diabetes care (see Table I for a list of the items.) The 58 items in the initial version were developed from several sources, including focus interviews with adolescents, interviews with health care providers, and prior research (e.g., La Greca et al., 1995). Items reflected five key areas of diabetes care: insulin (10 items); blood testing/reactions (14 items); meals (20 items); exercise (9 items); and emotional support (5 items). Adolescents rated the frequency of each behavior (How often does a family member . . . ?), with 0 = never, 1 = less than 2 times a month, 2 = twice a month, 3 = once a week, 4 = several times a week, or 5 = at least once a day. The frequency ratings (1 to 5) were identical to those used in the diabetes social support interview (DSSI-Family; La Greca et al., 1995); a rating of 0 (never) was added, in the event that the behavior never occurred. Adolescents also provided ratings of supportiveness (How does this make you feel?), with 1 = not supportive, 0 = neutral, 1 = a little supportive, 2 = supportive, 3 = very supportive. The positive ratings (1 to 3) were identical to those used in the DSSI-Family; nonsupportive (1) and neutral (0) ratings were added, as the questionnaire assessed a wide range of family behaviors;

not all of them may be viewed as supportive to adolescents. (The DSSI-Family specifically asks the adolescent about supportive behaviors, so that ratings of neutral or nonsupportive are not necessary.) Based on prior research (e.g., La Greca et al., 1995; Procidano & Heller, 1983), two scoring methods were examined. One was based on the frequency of the family behaviors (a normative approach), and one was based on the frequency adjusted for the ratings of supportiveness (i.e., frequency support) (an individualized approach). These scores were calculated for all the DSSQ-Family items (Total) and for the five areas of diabetes care (see the Results section for scoring details.) Perceived Social Support (PSS; Procidano & Heller, 1983). The PSS assesses adolescents perceived emotional support from family (PSS-Family) and friends (PSS-Friends). Each subscale contains 20 items answered in a yes/no/dont know format. Scores can range from 0 to 20, with higher scores reflecting greater emotional support. The PSS has been used with adolescents and adults who have diabetes (La Greca et al., 1995; Lyons, Perrotta, & Hancher-Kvam, 1988). Previous research supports the reliability and validity of the instrument. Internal consistencies have ranged from .84 to .92 across other samples (Lyons et al., 1988). Emotional support from the family on the PSS-Family also has been found to correlate significantly with diabetes-specific support from family members (La Greca et al., 1995). In this sample, the internal consistencies were .75 for the PSS-Family and .77 for the PSS-Friends. Family Environment Scale (FES; Moos & Moos, 1986). The FES is a widely used instrument that assesses family environment. It consists of 10 subscales, each with 9 true-false items; subscale scores can range from 0 to 9. The FES subscales have adequate internal consistency (range from .61 to .78), and good test-retest reliabilities over a 2-month period (.52 to .89) (Moos & Moos, 1986). Due to time constraints, only the Cohesion and Conflict subscales were used in this study, as they were of most interest. In this sample, their internal consistencies were .81 (Cohesion) and .74 (Conflict). Adherence to Diabetes Care. Adherence was assessed using a structured interview developed by Hanson et al. (1987, 1992) that included multiple aspects of diabetes care (insulin, glucose testing, meals, and treating hypoglycemia). A total score (range = 0 to 41) was calculated, with higher scores reflecting better adherence. Test-retest reliabilities over 3 and 6 months have been reported to be .70 and .73, respec-

Downloaded from http://jpepsy.oxfordjournals.org/ by guest on March 10, 2012

Diabetes Social Support Questionnaire

669

Table I. Item

DSSQ-Family: Adolescent Ratings of Supportiveness, Frequency, and Individualized Ratings (Means and SD) Supportiveness Frequency Individualized Rating

Insulin administration 1. Give you your shots. 2. Remind you to take your shots. 3. Praise you for giving yourself shots correctly or on time. 4. Help out when you give yourself shots. 5. Wake you up so you can take your morning shot on time.a 6. Change their own schedule to get an early start too, when you give yourself a morning shot. 7. Check after youve taken your shot to make you have done it. 8. Let you know they appreciate how difcult it is to take insulin shots. Total insulin administration Blood glucose testing 9. Ask you about the results of your blood tests. 10. Watch you test your blood sugars to see what the values are. 11. Test your blood sugar for you. 12. Remind you to test your blood sugar. 13. Make sure you have materials needed for blood testing.a 14. Let you know that they appreciate how hard it is to test blood sugars every day. 15. Set up materials you need for testing your blood sugar. 16. Praise you for testing your blood sugar on your own. 17. Help out when you test your blood sugar. 18. Keep track of testing results for you. 19. Watch for signs that your blood sugar is low.a 20. Help out when you might be having a reaction.a Total blood glucose testing Meals 21. Encourage you to eat the right foods.a 22. Let you know they understand how important it is for you to eat right.a 23. Ask if certain foods are okay for you to eat, before serving them. 24. Do the grocery shopping for your meals. 25. Schedule meals at the times you need to eat.a 26. Remind you about sticking to your meal plan. 27. Suggest foods you can eat on your meal plan. 28. Join you in eating the same foods as you. 29. Get on your case after you ate something you shouldnt. 30. Avoid tempting you with food or drinks that you shouldnt have. 31. Watch what you eat to make sure that you eat the right foods. 32. Cook meals for you that t your meal plan.a 33. Choose restaurants that serve food you can eat. 34. Eat at the same time you do.a 35. Praise you for following your diet. 36. Tell you when youve eaten too much or too little. 37. Show theyre pleased when youve eaten right. 38. Keep track of your mean plan for you. 39. Buy special foods that you can eat.a 40. Tell you not to eat something you shouldnt. Total meals Exercise 41. Suggest ways you can get exercise. 42. Remind you to exercise. 43. Invite you to join in exercising with them. 44. Congratulate or praise you for exercising regularly. 45. Encourage you to join an organized sports activity. 1.47 (1.2) 1.19 (1.3) 1.39 (1.3) 1.35 (1.3) 1.40 (1.3) 2.52 (1.9) 2.12 (1.8) 2.03 (1.7) 1.66 (1.8) 1.77 (1.7) 5.18 (5.4) 3.65 (5.2) 4.27 (5.0) 3.72 (5.1) 3.74 (5.1) 1.82 (1.3) 1.82 (1.2) 1.38 (1.3) 1.74 (1.3) 1.95 (1.2) 1.61 (1.4) 1.60 (1.2) 1.68 (1.3) 1.32 (1.4) 1.67 (1.3) 1.64 (1.3) 2.07 (1.1) 1.64 (1.2) 1.77 (1.2) 1.45 (1.3) 1.31 (1.3) 1.38 (1.2) 1.55 (1.3) 2.04 (1.1) 1.47 (1.4) 1.68 (1.0) 3.70 (1.5) 3.19 (1.5) 2.03 (1.9) 3.04 (1.4) 3.89 (1.6) 3.12 (1.6) 2.76 (1.7) 3.63 (1.7) 2.73 (1.7) 2.76 (1.9) 3.03 (1.9) 4.04 (1.4) 2.41 (1.6) 4.03 (1.5) 2.59 (1.9) 2.55 (1.6) 2.28 (1.8) 2.70 (2.0) 3.04 (1.5) 3.00 (1.7) 3.15 (1.1) 7.58 (6.0) 6.72 (5.6) 4.28 (5.6) 6.07 (5.2) 8.41 (5.9) 5.99 (6.1) 5.81 (5.6) 7.33 (6.1) 3.76 (5.4) 5.59 (5.8) 5.95 (6.2) 9.00 (5.6) 5.05 (5.0) 7.82 (6.0) 4.28 (5.3) 3.68 (5.2) 4.51 (5.1) 5.46 (6.1) 6.84 (5.0) 5.18 (5.8) 5.96 (4.3) 1.34 (1.3) 1.15 (1.4) 1.38 (1.4) 1.23 (1.5) 1.97 (1.1) 1.30 (1.4) 1.44 (1.4) 1.43 (1.3) 1.35 (1.4) 1.43 (1.3) 1.90 (1.1) 2.31 (1.0) 1.52 (0.9) 4.12 (1.3) 2.60 (1.8) 1.44 (1.9) 3.11 (1.8) 3.11 (1.5) 1.88 (1.8) 1.64 (1.9) 1.84 (1.7) 1.78 (1.9) 2.26 (2.1) 3.39 (1.7) 2.75 (1.7) 2.50 (1.1) 5.85 (6.3) 4.20 (5.8) 3.43 (5.2) 3.51 (5.9) 6.96 (5.5) 4.23 (5.4) 4.08 (5.2) 3.97 (4.9) 4.28 (5.5) 5.08 (5.8) 7.84 (5.5) 6.99 (5.1) 5.05 (4.0) 1.59 (1.3) 1.31 (1.4) 1.34 (1.4) 1.30 (1.3) 1.80 (1.4) 1.46 (1.4) 1.14 (1.4) 1.30 (1.4) 1.40 (1.0) 2.34 (2.1) 2.95 (1.7) 1.51 (1.7) 1.86 (2.0) 3.43 (1.9) 2.44 (1.9) 2.35 (1.9) 2.01 (1.9) 2.39 (1.1) 4.95 (6.0) 4.04 (5.7) 3.38 (4.7) 3.86 (5.2) 7.80 (6.3) 5.51 (5.9) 3.99 (5.9) 4.74 (5.7) 4.78 (3.8)

Downloaded from http://jpepsy.oxfordjournals.org/ by guest on March 10, 2012

670

La Greca and Bearman

Table I. Item

Continued Supportiveness 1.55 (1.4) 1.28 (1.3) 1.37 (0.9) 1.86 (1.2) 1.47 (1.3) 1.62 (1.2) 1.78 (1.2) 1.78 (1.2) 1.80 (1.1) 1.55 (0.9) Frequency 1.46 (1.4) 1.55 (1.7) 1.90 (1.2) 3.50 (1.6) 1.85 (1.5) 2.46 (1.6) 3.15 (1.6) 2.40 (1.5) 3.00 (1.4) 2.59 (1.0) Individualized Rating 3.19 (4.0) 3.68 (5.0) 3.90 (3.9) 7.84 (6.0) 3.59 (4.3) 5.16 (5.0) 6.65 (5.5) 5.08 (4.8) 5.68 (4.3) 5.08 (3.6)

46. Buy sports equipment for you. 47. Exercise with you. Total exercise Emotional support 48. Are available to listen to concerns or worries about your diabetes care.a 49. Give you things to read on diabetes care. 50. Tell you how well youve been doing with your diabetes care. 51. Encourage you to do a good job of taking care of your diabetes.a 52. Understand when you sometimes make mistakes in taking care of your diabetes.a Total emotional support Overall total
a

Most Supportive Item; Individualized Rating = Frequency Supportiveness.

Downloaded from http://jpepsy.oxfordjournals.org/ by guest on March 10, 2012

tively (Hanson et al., 1987, 1992). Hanson et al. have provided data on the validity of this measure for adolescents with diabetes; in particular, higher levels of adherence have been significantly related to metabolic control (Hb A1 assays), with correlations in the range of .30.

plied by the corresponding supportiveness score (i.e., the frequency was adjusted for the items perceived as supportiveness). Individualized scores (averages) were calculated for the Total DSSQ-Family and the five areas of diabetes care and could range from 5 (not supportive but very frequent) to 15 (very supportive and very frequent). Psychometric Considerations An initial study goal was to evaluate the two methods for scoring the DSSQ-Family. Internal consistencies2 (Cronbachs ) were calculated. For the frequency ratings, internal consistencies were .95 for Total, .75 for insulin, .85 for blood testing, .93 for meals, .85 for exercise, and .83 for emotions. Internal consistencies for the individualized ratings were slightly higher: .98 for Total, .82 for insulin, .91 for blood testing, .96 for meals, .89 for exercise, and .89 for emotions. Intercorrelations among the DSSQ-Family scores were examined next. Within the five areas of diabetes care, intercorrelations among the frequency ratings were moderate (range = .51 to.76; median = .57; all ps < .001). The intercorrelations among the individualized ratings were a bit higher (range = .62 to .85; median = .72; all ps < .001). The frequency ratings were highly correlated with the corresponding individualized rating (which was a combination of fre2 Although it was not possible to evaluate test-retest reliability in this study, data obtained over a 2-week period for a sample of 25 adolescents, using the Friends version of the DSSQ , yielded the following results. For the frequency ratings, retest reliabilities were .89 for the Total, .87 for insulin, .92 for blood testing, .93 for meals, .85 for exercise, and .90 for emotions. For the individualized ratings, retest reliabilities were .94 for the Total, .85 for insulin, .84 for blood testing, .82 for meals, .78 for exercise, and .88 for emotions. All were significant at p < .001.

Results
Scoring of the DSSQ-Family The primary study objective was to develop and evaluate the DSSQ-Family. Prior to the main analyses, the 58 items were examined to determine their appropriateness for inclusion on the final questionnaire. Six items1 were eliminated because the majority of the adolescents (50% or more) rated them as nonsupportive (1) or neutral (0). That is, most adolescents did not view these items as supportive. The remaining 52 items that were retained on the DSSQ-Family appear in Table I. To examine a normative scoring approach, we calculated average frequency scores for the Total DSSQFamily (all 52 items) and for the five areas of diabetes care (insulin, blood testing, meals, exercise, emotions). Scores could range from 0 to 5 (see Table I). To examine an individualized approach, for each adolescent, the frequency score for each item was multi1 The six items eliminated from the questionnaire focused mainly on nagging. These items were Nag you about your shots, Get on your case about taking insulin, after you were late or forgot, Nag you until you do your testing, Remind you to test urine for ketones, Get on your case when you havent exercised, and Bug you about exercising. The mean rating of supportiveness for each of these six items fell below 1.00 (a little supportive).

Diabetes Social Support Questionnaire

671

Table II.

DSSQ-Family Frequency and Individualized Ratings: Correlations With Age, Disease Duration, and Other Support Measures Age Duration PSS-Fa PSS-Fr Coh Con

Frequency ratings DSSQ Family (all) Insulin administration Blood glucose testing Meals Exercise Emotional support Individualized ratings DSSQ Family (all) Insulin administration Blood glucose testing Meals Exercise Emotional support .38*** .34** .37*** .40*** .36*** .21* .07 .06 .09 .02 .13 .02 .41*** .29** .35*** .41*** .28** .49*** .03 .01 .01 .02 .03 .09 .26* .26* .24* .23* .12 .29** .12 .09 .13 .06 .01 .52*** .48*** .52*** .46*** .46*** .26* .18 .13 .18 .10 .20 .13 .47*** .25* .37*** .42*** .36*** .49*** .08 .08 .04 .05 .04 .10 .34** .23* .33** .26* .18 .37*** .08 .06 .07 .00 .07 .20*

Downloaded from http://jpepsy.oxfordjournals.org/ by guest on March 10, 2012

.24*

PSS-Fa = Perceived Support Scale for Family; PSS-Fr = Perceived Support Scale for Friends; Coh = Cohesion subscale of Family Environment Scale; Con = Conflict subscale of the Family Environment Scale. *p < .05. **p < .01. ***p < .001.

quency and supportiveness); these correlations ranged from .75 to .91 (median = .84; all ps < .001). Similarly, the Total DSSQ-Family scores for the frequency and the individualized ratings were significantly interrelated (r = .88, p < .001). Family Support on DSSQ-Family: Correspondence With Age, Disease Duration, and Gender A second study goal pertained to concurrent validity. The associations between perceived family support for diabetes care and demographic variables (age, disease duration, gender) were examined, but only agerelated differences in perceived family support for diabetes care were expected. Pearson correlations were computed for age and disease duration with the frequency and the individualized ratings on the DSSQ-Family (see Table II). As expected, for the Total score, younger adolescents reported receiving more frequent support from family members for their diabetes care. In addition, for all five areas of diabetes management, the frequency of family support was significantly related to age, with younger adolescents perceiving more support than older adolescents. Identical findings were obtained for the individualized ratings. As expected, diabetes duration was unrelated to perceived family support. Gender differences were evaluated using one-way analyses of variance (ANOVAs) for each of the DSSQ-Family scores listed in Table II (left side of the table), using an

alpha level of .05 to identify a significant difference. As expected, none of the measures of family support differed significantly for adolescent boys and girls. (The means for the total sample were reported in Table I.) DSSQ-Family: Associations With Other Measures of Support As another way of evaluating concurrent validity, the associations between the DSSQ-Family and other measures of support from family and friends were examined. We hypothesized that the DSSQ-Family would be related to the general measures of family support (PSS-Family, FES-Cohesion), but not to friends support (PSS-Friends) or to family conflict (FES-Conflict). In support of the concurrent validity of the DSSQ-Family, adolescents who reported more frequent family support for diabetes care viewed their families as more emotionally supportive (PSS-Family; r = .47, p < .001), and more cohesive (FES-Cohesion; r = .34, p < .01), than adolescents who reported less frequent family support (see Table II, right side.) This pattern was identical for the individualized ratings. Similar results were obtained for the five areas of diabetes management. In general, adolescents who reported more frequent family support for the specific areas of diabetes care also viewed their families as more emotionally supportive and more cohesive. The one exception to this pattern was that family

672

La Greca and Bearman

Table III.

Perceived Family Support for Diabetes Care as a Predictor of Adolescents Treatment Adherence Total R R2 Change Step F (change) Partial r a

DV: Treatment adherence Step 1: Adolescent age Step 2: PSS-Family FES-Cohesion Step 3: (equation #1) Total DSSQ: Frequency Step 3: (equation #2) Total DSSQ: Individualized
a

.24 .40

.06 .10

4.45** 7.84***

.24** .11ns .32***

.44 .47

.04 .06

2.99 5.29**

.20 .27**

Partial correlation controlling for all the preceding variables in the regression analyses. **p < .05. ***p < .01.

Downloaded from http://jpepsy.oxfordjournals.org/ by guest on March 10, 2012

support for exercise was not significantly related to family cohesion (r = .18), although the correlation was in the same direction as the others. An identical pattern was observed for the individualized ratings. Thus, regardless of whether the frequency or individualized ratings were used, more emotionally supportive and cohesive families were perceived as providing more diabetes-specific family support. In support of the discriminant validity of the DSSQ-Family, none of the frequency or individualized ratings was related to support from friends (PSSFriends) or to family conflict (FES-conflict). The only exception was that perceptions of emotional support for diabetes care were negatively related to family conflict. Specifically, adolescents who reported less emotional support for diabetes care (using either the frequency or individualized ratings) perceived their families as more conflictual (rs = .20, .24, respectively) than did adolescents who reported more emotional support for their diabetes care. DSSQ-Family: Predicting Adolescents Treatment Adherence A third study goal was to examine the predictive validity of the DSSQ-Family, hypothesizing that greater perceived family support for adolescents diabetes care would predict better adherence, even when controlling for general levels of families emotional support. Two hierarchical regression analyses were conducted (see Table III), with adherence as the dependent variable, and using either the Total frequency ratings or the Total individualized ratings as predictors. On the first step, adolescent age was entered, because younger adolescents were more adherent than older adolescents (r = .24, p < .05). On the second step, family support and cohesion (PSSFamily and FES-Cohesion) were entered to control

for general levels of family support and cohesion and to determine if more supportive, cohesive families had more adherent adolescents. In the third step, the diabetes-specific support scores were entered. Table III shows that younger adolescents had better adherence (Step 1), as did adolescents who perceived their families as more cohesive (Step 2). Together, these two sets of variables predicted 16% of the variance in adolescents adherence. Interestingly, family cohesion (partial r = .32), but not general family support (partial r = .11), was related to adherence. On the third step (first analysis), the Total frequency score from the DSSQ-Family was not a significant predictor of adolescents treatment adherence; however, the Total individualized score (second analysis) predicted 6% additional variance (p < .025). Adolescents with greater perceived family support for diabetes care reported better adherence. As a follow-up to this analysis, the regression was repeated with the individualized ratings for the five areas of diabetes management entered on the third step; this set predicted 17% additional unique variance in adolescents adherence, Fchange = 3.29, p < .01, for a total R2 of .33. Partial correlations (controlling for age, family support, and cohesion) indicated that greater family support for insulin administration (.24, p < .05), blood glucose testing (.32, p < .01), and meals (.28, p < .05) primarily contributed to this effect, suggesting that family support for daily management tasks may be more closely related to adolescents adherence than family support for exercise or emotions. Analysis of the Most Supportive Family Behaviors A final study goal was to examine the clinical utility of the DSSQ-Family by identifying the specific family

Diabetes Social Support Questionnaire

673

behaviors that adolescents perceived as most supportive for their diabetes care. The most supportive behaviors were defined as those in the upper quartile (25%) of the DSSQ-Family items, based on supportiveness ratings. These 13 items are marked in Table I with a superscript a. The mean perceived supportiveness of the 13 items was 1.91 (SD = .88) and was significantly higher than for the remaining 39 items on the DSSQ-Family (M = 1.35, SD =.93; t = 7.52, p < .0001). Included in the most supportive behaviors were 60% of the items that dealt with emotional support (e.g., are available to listen to your concerns about diabetes); this area of diabetes care was overrepresented in the upper quartile. Also among the most supportive items were 30% of the items that dealt with meals (cook meals that fit your meal plan, eat at the same time you do), and 25% of the items that dealt with blood glucose testing (help out when you might be having a reaction, make sure you have materials needed for testing). Underrepresented among the most supportive items were those dealing with insulin (12.5%) (wake you up so you can take your shot on time) or exercise (0%). Across the 13 most supportive items, an average frequency score and an average individualized score were calculated. Their internal consistencies were .83 and .92, respectively. Regression analyses (identical to those described) were conducted for the 13 most supportive items as predictors of adolescents adherence. As with the Total DSSQ-Family, in Step 3, the frequency of the most supportive items did not predict adherence (partial r = .12, controlling for age, family support, and cohesion), but the individualized rating did significantly predicted 6% unique variance (partial r = .25, p < .035).3

Discussion
Little research has examined the specific family behaviors associated with youngsters disease management (Drotar, 1997), even though families play an important role in disease management and adaptation for youths with chronic pediatric conditions. Thus, information on family behaviors that relate to
3 The patterns of correlations for the frequency and individualized ratings for the 13 most supportive items were identical to those for the Total DSSQ-Family items. Specifically, the frequency of the most supportive items correlated with other variables as follows: age (r = .44, p < .001), diabetes duration (r = .09, ns), PSS-Family (r = .48, p < .001), PSSFriend (r = .02, ns), cohesion (r = .38, p < .01), and conflict (r = .12, ns). For the individualized ratings the correlations were age (r = .33, p < .001), diabetes duration (r = .04, ns), PSS-Family (r = .44, p < .001), PSSFriend (r = .01, ns), cohesion (r = .29, p < .05), and conflict (r = .14, ns).

better treatment adherence for adolescents with diabetes has the potential to inform the next generation of family interventions for youths with diabetes. In this regard, this study presents a new measure and provides useful information on the family behaviors that adolescents perceived to be supportive for their diabetes care. This information may be useful for enhancing adolescents treatment adherence. The primary study objective was to develop and examine the utility of a new measure, the DSSQFamily, to assess adolescents perceptions of family support for diabetes care. The results provided promising support for this measure. In particular, internal consistencies for the various DSSQ-Family scores were high, and the patterns of relationships with other measures were consistent with predictions. The results also provided support for the incremental and predictive validity of the individualized ratings from the DSSQ-Family, which predicted adolescents adherence above and beyond general levels of family emotional support and cohesion. One important clinical implication of these findings is that the DSSQ-Family appears to be a useful measure of perceived family support for adolescents diabetes care. Drotar (1997) has recommended that future research on family functioning among youths with chronic pediatric conditions make greater use of illnessspecific measures involving family variables (p. 161). In this regard, the DSSQ-Family may be useful to include in future studies of adaptation and disease management for youths with diabetes. In the process of evaluating the DSSQ-Family, two different scoring methods were examined: one based on a normative approach that utilizes frequency ratings for supportive behaviors, and one based on an individualized approach that adjusts the frequency ratings for the individual adolescents perceptions of supportiveness. Although the findings were very similar for the two methods, the results appeared to favor the individualized ratings. Their advantage was most clearly observed in the analyses of adolescents treatment adherence where the individualized scores, but not the normative/frequency scores, predicted adolescents adherence, even when only the 13 most supportive items from the DSSQFamily were used. These findings suggest that the individualized approach may be more useful than the normative approach in clinical settings. In particular, efforts to increase family support for adolescents diabetes care may be better served by including adolescents own perspectives on what they view as supportive, rather than relying on what adolescents typically view as supportive.

Downloaded from http://jpepsy.oxfordjournals.org/ by guest on March 10, 2012

674

La Greca and Bearman

Another key finding from this study, supporting the concurrent validity of the DSSQ-Family, was that older adolescents perceived their family members to provide less diabetes-specific support than did younger adolescents. Others (e.g., Anderson et al., 1990; Follansbee, 1989) have expressed concern that, as adolescents get older, family members become less involved in and supportive of their diabetes care. One of the potential benefits of a measure such as the DSSQ-Family is that it may be used to identify family behaviors that adolescents do find to be supportive, so that family members can provide appropriate kinds of support and maintain involvement in diabetes care as adolescents mature. In contrast to the findings for age, disease duration was not related to perceived family support for diabetes care. Thus, the relationship between age and perceived family support cannot be explained by the fact that younger adolescents typically have had diabetes for a shorter period of time and, therefore, need more assistance with their diabetes care. In further support of the concurrent validity of the DSSQ-Family, the results revealed that adolescents who perceived their families as providing more diabetes-specific support also viewed their families as more cohesive and emotionally supportive. In contrast, adolescents perceptions of family support for diabetes care were not related to adolescents support from friends or to reports of family conflict, in general. Although further replication of these findings would be desirable, these data do provide good preliminary support for the convergent and discriminant validity of the DSSQ-Family. Furthermore, from a clinical perspective, it was interesting to note that the correlations between diabetes-specific family support and general family support and cohesion were moderate (.23 to .49). This suggests that even cohesive, emotionally supportive families do not necessarily provide high levels of diabetes-specific support. This was especially true for family support for adolescents exercise, which was unrelated to family cohesion. A clinical implication of these findings is that even supportive, cohesive families may need help in identifying specific ways to support adolescents diabetes care. With respect to the predictive and incremental validity of the DSSQ-Family, a key finding was that the individualized ratings from the DSSQ-Family predicted adherence above and beyond the more generic measures of family support and cohesion. In fact, the general measure of family support was not significantly related to adherence, when adolescent

age was controlled. Pediatric investigators have emphasized that disease-specific measures may be helpful in understanding youngsters disease management and disease adaptation (e.g., Drotar, 1997; La Greca & Lemanek, 1996), and these findings are consistent with this point. Furthermore, the specific areas of diabetes-related family support that were most associated with adolescents adherence involved daily management tasks (meals, glucose testing, and insulin administration), rather than exercise or emotions. Although family members support of management tasks may be important for adherence, their provision of emotional support may also be critical. In this study, families support for the emotional aspects of diabetes care was unrelated to adherence, but this does not rule out the possibility that it may be associated with other indices of disease adaptation, such as adolescents quality of life (Grey, Boland, Yu, Sullivan-Bolyai, & Tamborlane, 1998) or self-esteem (Varni, Babani, Wallander, Roe, & Frasier, 1989). Moreover, family members emotional support may help to prevent or minimize feelings of depression, which appear to be high among individuals with diabetes (Anderson, Freedland, Clouse, & Lustman, 2001). Most of the emotional support items from the DSSQ-Family were among those viewed as most supportive by the adolescents. In the future, it may be fruitful to examine linkages between families emotional support for diabetes and other indices of disease adaptation and adjustment. Although our findings were promising, continued study of the DSSQ-Family is desirable. In particular, several study limitations suggest directions for further investigation. First, in this study it was not possible to obtain retest reliability for the DSSQFamily, and this will be important in future work. Second, and also important for future research, would be a replication of the study results with a larger sample, especially one that has sufficient power to factor analyze the DSSQ-Family. Third, this study relied on adolescents reports of family support and treatment adherence. Adolescents are considered to be the best informants for evaluating their social support and their daily treatment adherence (e.g., La Greca & Lemanek, 1996); in addition, the pattern of findings cannot be explained solely by shared informant variance (e.g., family support for diabetes was related to family support, but not to friend support). Nevertheless, in future studies it will be desirable to evaluate support and adherence from multiple perspectives (e.g., adolescent and parent),

Downloaded from http://jpepsy.oxfordjournals.org/ by guest on March 10, 2012

Diabetes Social Support Questionnaire

675

and to incorporate other objective measures of disease status, such as hemoglobin assays, in addition to measures of adherence. Fourth, this study focused on a primarily middle-class sample, with a relatively small number of minority youths, and thus can best be generalized to similar groups of adolescents. In the future, work on the DSSQ-Family should be extended to multiethnic, low-income adolescents with diabetes. One advantage of a checklist measure, such as the DSSQ-Family, is that it may facilitate comparisons across different ethnic groups because of its standard set of items. Finally, this study focused on a particular chronic disease, although it is likely that the finding might be applicable to other chronic conditions for which medications, exercise, or meal management play a role. Future studies may wish to adapt the DSSQ-Family for other chronic pediatric conditions. In conclusion, this study offers the DSSQ-Family as a useful instrument for understanding family members support of adolescents diabetes care. An important clinical implication is that continued family involvement in the day-to-day management of diabetes may be critical for youngsters disease management. At least when the goal of intervention is to promote treatment adherence, it may be most

productive to focus on ways that families can support daily management tasks, taking into account what the individual adolescent perceives to be supportive in these areas. This may be especially critical for older adolescents, as families are substantially less likely to be supportive of older teens management tasks, as the data in this study have shown. The key issue is not whether families should be involved and supportive, but how best to do this, especially as adolescents mature.

Acknowledgments
Preparation of this paper was supported, in part, by grants from the National Heart, Lung, and Blood Institute (HL 36588) and the National Institute of Child Health and Development (T32 HD07510). We thank the following individuals for their input on the initial development of the DSSQ-Family: Edwin Fisher, Jr., Wendy Auslander, Peggy Greco, and Dante Spetter. Received July 16, 2001; revisions received November 15, 2001, and February 25, 2002; accepted March 1, 2002
Downloaded from http://jpepsy.oxfordjournals.org/ by guest on March 10, 2012

References
Anderson, B. J., Auslander, W. F., Jung, K. C., Miller, J. P., & Santiago, J. V. (1990). Assessing family sharing of diabetes responsibilities. Journal of Pediatric Psychology, 15, 477492. Anderson, B. J., Miller, J. P., Auslander, W. F., & Santiago, J. V. (1981). Family characteristics of diabetic adolescents: Relationship to metabolic control. Diabetes Care, 4, 586594. Anderson, R. J., Freedland, K. E., Clouse, R. E., & Lustman, P. J. (2001). The prevalence of comorbid depression in adults with diabetes: A meta-analysis. Diabetes Care, 24, 106978. Bearman, K. J., & La Greca, A. M. (2002). Assessing friend support of adolescents diabetes care: The Diabetes Social Support Questionnaire-Friends Version. Journal of Pediatric Psychology, 27, 417428. Burroughs, T. E., Harris, M. A., Pontious, S. L., & Santiago, J. V. (1997). Research on social support in adolescents with IDDM: A critical review. Diabetes Educator, 23, 438448. Chase, H. P. (1992). Understanding insulin dependent diabetes. Denver: University of Colorado Health Sciences Center. Diabetes Control and Complications Trial. (1993). New England Journal of Medicine, 329, 683689. Drotar, D. (1997). Relating parent and family functioning to the psychological adjustment of children with chronic health conditions: What have we learned? Journal of Pediatric Psychology, 22, 149165. Follansbee, D. S. (1989). Assuming responsibility for diabetes management: What age? What price? Diabetes Educator, 15, 347352. Grey, M., Boland, E. A., Yu, C., Sullivan-Bolyai, S., & Tamborlane, W. V. (1998). Personal and family factors associated with quality of life in adolescents with diabetes. Diabetes Care, 21, 909914. Hanson, C. L., De Guire, M. J., Schinkel, A. M., Henggeler, S. W., & Burgen, G. A. (1992). Comparing social learning and family systems correlates of adaptation in youths with IDDM. Journal of Pediatric Psychology, 17, 555572. Hanson, C. L., Henggeler, S. W., & Burghen, G. A. (1987). Social competence and parental support as mediators of the link between stress and metabolic control in adolescents with insulin-dependent diabetes mellitus. Journal of Consulting and Clinical Psychology, 55, 529 533. Hauser, S. T., Jacobson, A. M., Lavori, P., Wolfsdorf, J. I., Herskowitz, R. D. Milley, J. E., Bliss, R., Wertlieb, D., &

AQ1

AQ2

676

La Greca and Bearman

Stein, J. (1990). Adherence among children and families with insulin-dependent diabetes over a four-year longitudinal follow-up. Journal of Pediatric Psychology, 15, 527542. Kaslow, N.J., Collins, M. H., Loundy, M. R., Brown, F., Hollins, L. D., & Eckman, J. (1997). Empirically validated family interventions for pediatric psychology: Sickle cell disease as an exemplar. Journal of Pediatric Psychology, 22, 213227. Kazak, A. E. (1997). A contextual family/systems approach to pediatric psychology: Introduction to the special issue. Journal of Pediatric Psychology, 22, 141148. La Greca, A. M. (1998). Its All in the Family: Responsibility for diabetes care. Journal of Pediatric Endocrinology & Metabolism, 11, 379385. La Greca, A. M., Auslander, W. F., Greco, P., Spetter, D., Fisher, E. B., & Santiago, J. V. (1995). I get by with a little help from my family and friends: Adolescents support for diabetes care. Journal of Pediatric Psychology, 20, 449 476. La Greca, A. M., Follansbee, D. J., & Skyler, J. S. (1990). De-

velopmental and behavioral aspects of diabetes in children and adolescents. Childrens Health Care, 19, 132 139. La Greca, A. M., & Lemanek, K. L. (1996). Assessment as a process in pediatric psychology. Journal of Pediatric Psychology, 21, 137151. Lyons, J. S., Perrotta, P., & Hancher-Kvam, S. (1988). Perceived social support from family and friends: Measurement across disparate samples. Journal of Personality Assessment, 52, 4247. Moos, R., & Moos, B. (1986). Family environment scale (2nd ed.). Palo Alto: Consulting Psychologists Press. Procidano, M., & Heller, K. (1983). Measures of perceived social support from friends and from family: Three validation studies. American Journal of Community Psychology, 11, 124. Varni, J. W., Babani, L., Wallander, J. L., Roe, T. F., & Frasier, S. D. (1989). Social support and self-esteem effects on psychological adjustment in children and adolescents with insulin-dependent diabetes mellitus. Child and Family Behavior Therapy, 11, 116.

Downloaded from http://jpepsy.oxfordjournals.org/ by guest on March 10, 2012