Development of a New Fear of Hypoglycemia Scale: Preliminary Results Jodi L. Kamps, 1 PHD, Michael C. Roberts, 2 PHD, ABPP, and R. Enrique Varela, 3 PHD 1 Children s Hospital of New Orleans, 2 University of Kansas and 3 Tulane University Objective To provide preliminary psychometric data on the Children s Hypoglycemia Index (CHI), a new scale was designed to measure fear of hypoglycemia (FH) in children. Methods One hundred and nine children with type 1 diabetes attending a diabetes summer camp completed the CHI, and 57 of these children completed the CHI 2 weeks later. All participants also completed the Children s Hypoglycemia Fear Survey (CHFS), the Revised Children s Manifest Anxiety Scale (RCMAS), and a Hypoglycemia History Form. Parents of the children also completed a Hypoglycemia History Form. Results The CHI demonstrated good factor structure, internal consistency, test-retest reliability, and criterion-related validity with established measures. Frequency of severe hypoglycemic episodes since diabetes diagnosis also predicted scores on the Situation- and General-Fear subscales of the CHI. Conclusions The CHI demonstrated reliability and validity as a measurement of children s FH. The use of this measure offers important clinical and research implications. Key words assessment; fear of hypoglycemia; type 1 diabetes. The negative consequences of hypoglycemia in individuals with type 1 diabetes have been reliably documented in the literature (e.g., McCrimmon, Gold, Deary, Kelnar, & Frier, 1995) with physical symptoms being most prominent (e.g., shaking, sweating, drowsiness, nausea, and headache). Other symptoms of hypoglycemia may include poor motor coordination, confusion, and negative mood states, with argumentativeness and irritability being primary features in children (McCrimmon et al., 1995). Although usually transient, if left untreated, hypoglycemia can result in more severe consequences including seizures, coma, or death (McCrimmon et al., 1995). In some individuals with diabetes, the aversive symptoms of hypoglycemia may result in phobic-like fear of hypoglycemia (FH) (Cox, Irvine, Gonder-Frederick, Nowacek, & Butterfield, 1987; Irvine, Cox, & Gonder- Frederick, 1994), and this fear may have significant clinical implications for metabolic control, and consequently, for the treatment of diabetes. For instance, studies have found that patients who report FH engage in overcompensatory behaviors (e.g. taking too little insulin and eating excessively) to avoid hypoglycemia, resulting in poor metabolic control (Irvine et al., 1994). The association between FH and poor glycemic control is of particular concern in the context of current recommendations for tighter metabolic control by the Diabetes Control and Complications Trial (Diabetes Control and Complications Trial Research Group, 1994). Tighter diabetic control is achieved through intensive insulin therapy; however, an unwanted consequence of this therapy is a threefold increase in severe episodes of hypoglycemia (Diabetes Control and Complications Trial Research Group, 1994), thus increasing the risk of FH. Fear of hypoglycemia may be an even greater concern with children, considering that higher rates of severe hypoglycemia are found in children and adolescents than in adults, with most clinical researchers reporting a range of 0.1 0.2 episodes per child per year (Diabetes Control and Complications Trial Research Group, 1994; Soltesz, 1993). Developmentally, children may also lack the All correspondence should be addressed to Jodi Kamps, Children s Hospital, Department of Psychology, 200 Henry Clay Avenue, New Orleans, Louisiana 70118. E-mail: jkamps21@yahoo.com Journal of Pediatric Psychology 30(3) pp. 287 291, 2005 doi:10.1093/jpepsy/jsi038 Journal of Pediatric Psychology vol. 30 no. 3 Society of Pediatric Psychology 2005; all rights reserved.
288 Kamps, Roberts, and Varela cognitive and emotional sophistication to develop adaptive coping strategies in the face of threatening experiences related to hypoglycemia. Examining factors related to FH such as the risk and severity of hypoglycemia has become an important consideration in the study of FH (Irvine et al., 1994). For instance, high fear in the presence of high risk of hypoglycemia might be adaptive by motivating appropriate behaviors to avoid hypoglycemic episodes. However, low fear, or denial, when risk of hypoglycemia is high, or excessive fear when risk is low may lead to maladaptive responses (Irvine et al., 1994). Severe hypoglycemia [characterized by blood glucose below 50 mg/dl, or seizure/loss of consciousness (SLC)] may result in more fear than mild hypoglycemia (characterized by blood glucose between 50 and 70 mg/dl). Although several variables may impact the relationship between FH and metabolic control, FH is considered to play a significant role in diabetes management (Irvine et al., 1994), particularly for children with diabetes (Green, Wysocki, & Reineck, 1990). However, a review of the relevant literature indicates that only one scale, the Children s Hypoglycemia Fear Survey (CHFS) (Green et al., 1990), has been widely employed in the assessment of FH in children. Although the CHFS is a promising scale with good psychometric properties, some aspects of the scale might limit the progress of understanding the role of FH in the functioning of children with diabetes. For instance, the CHFS assesses only two areas related to FH, specific worries associated with hypoglycemia (Worry subscale) and behaviors aimed at maintaining adequate blood sugar control (Behavior subscale). In order to provide a greater understanding of FH, this phenomenon would benefit from assessing more than two areas related to its conceptualization. The construct validity of the behavior subscale also has been questioned because it measures behaviors that all children with diabetes typically engage in, independent of FH. In addition, as acknowledged by Green et al. (1990), the CHFS was modified from the original Hypoglycemia Fear Survey (HFS) (Cox et al., 1987), a measure designed to examine FH in adults. Although items relevant to adults such as, Having a reaction while driving, were removed when modifying the HFS for children (T. Wysocki, personal communication, March 2000), the remaining items were taken directly from the original measure. Given that the CHFS is a downward extension of an adult measure, the items may not fully capture the characteristics of FH in children. Given the clinical relevance of FH in the treatment of diabetes coupled with the limitations of the CHFS, there is a need for the development of alternative assessment tools for FH in children with diabetes. Reliable, valid, and convenient ways of measuring FH in children will facilitate investigation of this phenomenon in relation to metabolic control, regimen adherence, and adjustment to diabetes. Improved understanding of these relationships will result in interventions designed to promote improved metabolic control and adaptive patterns of coping with hypoglycemia in children with diabetes. The purpose of the current study was the development of a new measure of FH that explores the characteristics of this construct in children and provides an alternative to the existing CHFS. Preliminary data focusing on the children s hypoglycemia index (CHI) s factor structure, internal consistency, test-retest reliability, and convergent validity are presented here. Method Participants were 109 children with type 1 diabetes and their parents. The children were 42 boys (39%) and 67 girls (61%) ranging in age from 8 to 16 years (M = 11.9 years, SD = 2.3) who attended week-long sessions of an American Diabetes Association summer camp in a midwestern state. All families of eligible children registered for the camp were mailed informed consent and assent forms and the measures described here 1 week prior to the start of camp. Sixty-five percent of those contacted (n = 109) returned the signed consent forms and the completed measures during camp registration. Fifty-two percent of these participants (n = 57) also completed the CHI and the CHFS at 1-week post camp. The majority of children (n = 95) were Caucasian and most parents had at least some college education. Children completed the CHI, CHFS, Revised Children s Manifest Anxiety Scale (RCMAS), and a Hypoglycemia History Form, which asked about lifetime incidence of severe hypoglycemia and number of severe hypoglycemic episodes within the past year. Parents also completed the Hypoglycemia History Form. Procedure To develop the CHI, the authors first compiled a list of items considered relevant in three areas of FH: general fears related to hypoglycemia and its consequences, fear related to specific situations in which hypoglycemia might occur, and specific behaviors designed to address hypoglycemia or avoid it out of fear for its consequences. Items were chosen to assess specific situations related to hypoglycemia because children s fear might differ if specific situations are used as referents rather
Fear of Hypoglycemia 289 Table I. Means and Standard Deviations for children s hypoglycemia index (CHI) and Sample Items Measure M (SD) Range Total 51.74 (13.2) 25 87 Situation 13.28 (4.6) 7 27 Having a low blood sugar when I am at school makes me feel. Having a low blood sugar when I am asleep at night makes me feel. General Fears 19.94 (6.1) 9 37 Not realizing when my blood sugar gets low makes me feel. Passing out in public when my blood sugar is low makes me feel. Behavior 18.61 (4.7) 9 32 I take less insulin than I should because I don t want my blood sugar to get low. I check my blood sugar often because I am afraid it might be low. Note. CHI = children s hypoglycemia index; Pre-camp n = 109. than a global notion of hypoglycemia. In specifying that fear was the motivating factor behind enacting particular behaviors to address or avoid hypoglycemia, the CHI differs from the CHFS. Most items were derived from either a review of the theoretical and empirical literature in the area (Irvine et al., 1994) or based on the authors clinical experience with children with diabetes. All items were meant to have relevance to child or adolescent practices and situations particular to these age groups. Three advanced doctoral students in pediatric psychology sorted all items with relevance to one of the three theoretical areas of content. Only items for which there was 100% agreement about the conceptual area to which it belonged were kept (Table I). Nine items were considered relevant for assessment of general fears related to hypoglycemia and its consequences (General Fears), seven items for assessing fears in specific situations (Situation), and nine for assessing how often children engage in behaviors to avoid and/or prevent hypoglycemia out of fear (Behavior). Responses are rated on a 5-point Likert scale (Not Afraid to Extremely Afraid or Never to All the Time) with higher scores representing greater levels of fear. Results Means and standard deviations for the pre-camp CHI subscales and total scores are presented in Table I. Children s scores on the CHI and CHFS were not significantly different between the two administrations. Further, children who did not complete post-camp measures did not differ from those who completed precamp measures. Validity Construct validity of the CHI was examined by first conducting a maximum-likelihood confirmatory factor analysis on pre-camp data to test the hypothesis that the CHI measures three independent factors indicating three separate constructs (i.e., general fears, situation, and behavior). Because of the modest sample size, we used the Bentler-Bonett non-normed fit index (NNFI) (Bentler, 1990). Results indicated that this three-factor model was not rejected, X 2 (24) = 35.37, p >.05, and that this was an excellent fit, NNFI =.98. To further examine the construct validity of the CHI, we conducted a hierarchical regression with parent and child report of frequency of severe hypoglycemic (experienced SLC) and mild hypoglycemic (no SLC) episodes since diagnosis of diabetes and during the past year entered as predictors of the CHI subscales. We hypothesized that increased experience with severe hypoglycemia would result in higher FH than experiencing mild hypoglycemic episodes, given the particularly aversive nature of severe hypoglycemia. For all regressions, parent and child report of frequency of severe hypoglycemic episodes since diagnosis was a significant predictor of FH as measured by the General and Situations subscale (Table II). The CHI demonstrated good convergent validity evidenced by the significant correlations among its different subscales. The Situation subscale was positively correlated with the General Fears (r =.79, p <.01) and Behavior (r =.50, p <.01) subscales. In turn, the Behavior and General Fears subscales were positively correlated with each other (r =.47, p <.01). To assess criterion-related validity, the relationships between the CHI and established measures of anxiety and FH were examined. To assess anxiety, we used the RCMAS, a widely-used 37-item self-report measure that assesses the level and nature of anxiety in 6 19-year-old youth and has excellent psychometric properties (Reynolds & Richmond, 1997). The RCMAS consists of five scores including a Total score, three anxiety subscales measuring Physiological Anxiety, Worry/Oversensitivity, and Social Concerns/Concentration, and a Lie subscale. To assess FH, we used the CHFS, a 23-item self-report instrument that measures worries related to hypoglycemia and behaviors associated with diabetes care (Green et al., 1990). The CHI total score was positively correlated with the RCMAS Total (r =.25, p <.05), the CHFS Total (r =.69, p <.01), the CHFS Worry (r =.63, p <.01), and the CHFS Behavior subscales (r =.38, p <.01). Further,
290 Kamps, Roberts, and Varela Table II. Summary of Hierarchical Regression Analyses for Variables Predicting the CHI Subscales Note. CHI = children s hypoglycemia index; hypo = hypoglycemia; Pre-camp n = 109. *p <.05. p <.01. the CHI General Fears subscale was positively correlated with the RCMAS Total (r =.27, p <.01), RCMAS Worry subscale (r =.36, p <.01), and CHFS Worry subscale (r =.59, p <.01). Reliability Internal consistency was calculated using Cronbach s alpha coefficient. Cronbach s alpha coefficients for the CHI total score, the Situation, General Fears, and the Behavior subscales were.89,.80,.82, and.68, respectively, for the first administration and.93,.87,.89,.75, respectively, for the post-camp administration. Thus, the CHI demonstrated good internal consistency at both times of administration. Two-week test-retest reliability using Pearson correlation coefficients was.76 (p <.01) for the CHI Total score,.71 (p <.01) for the Situation subscale,.68 (p <.01) for the General Fears scale, and.68 (p <.01) for the Behavior subscale. Discussion Parent Participant Child Scale R 2 β F R 2 β F Situation Step 1.00.00.00.40 Mild hypo.00.07 Step 2.10 3.60*.11 4.04 Severe hypo Lifetime.31*.40* Severe hypo Past year.01.08 General Fears Step 1.03 2.92.00.06 Mild hypo.17.03 Step 2.09* 4.16.09* 3.07* Severe hypo Lifetime.34*.45 Severe hypo Past year.05.25 Behavior Step 1.01 0.77.00.02 Mild hypo.09.02 Step 2.01 0.43.01.16 Severe hypo Lifetime.08.11 Severe hypo Past year.008.09 The purpose of the present study was to develop a new measure of FH, the CHI. The results indicated that the CHI is a valid and reliable instrument in the measurement of children s FH and improves upon existing methods of assessment. The validity of the CHI was supported in a variety of ways. Confirmatory factor analyses demonstrated an excellent fit between the actual factor structure and proposed factor structure. Additionally, the CHI was significantly correlated with established measures of anxiety and FH in children. Of note is that children and parents experiences with severe hypoglycemic episodes since onset of diabetes were found to be consistent predictors of scores on the Situation and General Fear scales of the CHI. This pattern of results is not surprising given that diabetic seizures represent a more severe adverse event than an episode of mild hypoglycemia. The severity or intensity of a negative event might make this event more salient than less severe or intense events. However, the Behavior scale of the CHI was found to be unrelated to the frequency of severe hypoglycemic episodes. Perhaps children s cognitions and emotions regarding diabetes are more sensitive to change by severe hypoglycemic episodes compared to their diabetes management behavior, which may be subject to greater influence from caregivers. Results from the Behavior scale of the CHI appear similar to findings from previous studies utilizing the CHFS, which have demonstrated less clinical utility for this scale (Green et al., 1990). The reliability of the CHI was also supported, as the measure demonstrated adequate internal consistency and test-retest reliability. Not only does the CHI demonstrate adequate psychometric properties but also it further develops the assessment of FH. For instance, the CHI expands upon the conceptualization of FH by exploring three areas related to this phenomenon. Additionally, the Behavior scale was modified such that it represents behavior that was motivated by FH rather than behavior that individuals with diabetes perform independent of FH. The CHI also utilized items that were empirically derived from both the literature and clinical practice with children and adolescents with diabetes. A discussion of the study s limitations is warranted. Due to the procedure used in participant recruitment, independent completion of the measure by child participants could not be guaranteed. The sample also included participants covering a wide range of ages, and younger children may have had more difficulty reading the measure than their older counterparts. Children in the study also represented a relatively high level of SES, and results might not be relevant for all children with diabetes. Additionally, measures of adherence were not collected in the present study, although, the relationship of this variable with FH remains an important area of study. Despite these limitations, use of the CHI has several important clinical implications. Utilizing assessment
Fear of Hypoglycemia 291 measures with strong psychometric properties, such as the CHI, will help to accurately identify individuals who experience FH. Screening for FH will be particularly relevant for families who have children with a history of seizure or loss of consciousness. Additionally, FH is an important area to consider as families pursue intensive insulin therapy. Addressing FH may be another factor to consider in the treatment of children exhibiting problems with adherence to the diabetes regimen. Treatments that include cognitive restructuring may be promising venues for addressing attitudes and feelings related to FH. Ultimately, accurate assessment of FH is a critical first step toward developing relevant and effective interventions when such fear is maladaptive. Received September 12, 2003; revisions received January 4, 2004, and May 4, 2004; accepted June 8, 2004 References Bentler, P. M. (1990). Comparative fit indexes in structural models. Psychological Bulletin, 107, 238 246. Cox, D. J., Irvine, A., Gonder-Frederick, L., Nowacek, G., & Butterfield, J. (1987). Fear of hypoglycemia: Quantification, validation, and utilization. Diabetes Care, 10, 617 621. Diabetes Control and Complications Trial Research Group. (1994). 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. Journal of Pediatrics, 125, 177 188. Green, L. B., Wysocki, T., & Reineck, B. M. (1990). Fear of hypoglycemia in children and adolescents with diabetes. Journal of Pediatric Psychology, 15, 633 641. Irvine, A., Cox, D., & Gonder-Frederick, L. (1994). The fear of hypoglycaemia scale. In C. Bradley (Ed.), Handbook of Psychology and Diabetes (pp. 133 155). Great Britain: Harwood Academic Publishers. McCrimmon, R. J., Gold, A. E., Deary, I. J., Kelnar, C. J., & Frier, B. M. (1995). Symptoms of hypoglycemia in children with IDDM. Diabetes Care, 18, 858 861. Reynolds, C. R., & Richmond, B. O. (1997). Revised Children s Manifest Anxiety Scale (RCMAS) Manual, 6th ed. Los Angeles: Western Psychological Services. Soltesz, G. (1993). Hypoglycaemia in the diabetic child. Baillere s Clinical Endocrinology and Metabolism, 7, 741 755.