Youth and Parent Satisfaction With Clinical Use of the GlucoWatch G2 Biographer in the Management of Pediatric Type 1 Diabetes

The study was conducted by a network of five clinical centers, a coordinating center, and a central laboratory. An external data and safety monitoring board and the institutional review board at each center approved the study protocol, informed consent form, and assent form. A parent or guardian gave written consent and children gave written assent before any study procedures were performed. The protocol described in a separate publication (14) is summarized below.

To be eligible, a subject had to be between 7 and 18 years old, have type 1 diabetes with use of insulin for at least 1 year, have an HbA_1c (A1C) level between 7.0 and 11.0% inclusive, and be on a stable insulin regimen (either insulin pump or two or more daily injections) for the prior 2 months with no plans to switch modality of insulin administration during the next 6 months. Exclusion criteria included prior home use of a GW2B, use of corticosteroids within the last 6 months, or presence of another chronic illness. The sample size for the trial was based on the desire to detect a difference in A1C of 0.5% between the GW2B or usual care groups after 6 months in their respective treatments.

Between July 2003 and January 2004, 200 subjects entered the trial (40 at each of the five clinical centers), with 99 randomized to the GW2B group and 101 to the usual care group. Since families were selected for enrollment by the participating physicians, virtually all who were approached agreed to enroll. The average age of youths in the GW2B group was 12.3 ± 2.7 years, 45% were female and 85% were Caucasian, 6% were African American, 2% were Asian, 2% Hispanic, 4% multiple races, and 1% unknown race. Educational attainment for fathers and mothers was 24 and 16% postgraduate degrees, 35 and 35% bachelors degrees, 16 and 17% associates, 23 and 31% high school diploma, and 1 and 1% less than high school, respectively. Mean diabetes duration was 5.3 ± 3.4 years. Insulin modality was multiple daily injections for 54% and insulin pump for 46%. Mean baseline A1C was 8.0%, with 41% ≥8.0%.

Subjects in the GW2B group were provided with a GW2B and an unlimited number of sensors. The GW2B resembles a large wristwatch that adheres to the skin via a sensor that draws interstitial fluid through the skin and displays the estimated glucose level every 10 min for up to 13 h. Audible alarms signal high, low, and rapidly falling glucose levels. The data are stored in memory for computer analysis. Subjects were encouraged to use sensors as often as desired, using a minimum of four sensors during the 1st week and then at least two sensors weekly thereafter (with use of at least one sensor weekly during sleep). Subjects were given a personal computer for home use to download SMBG and GW2B data and view glucose results weekly. All subjects were asked to complete a weekly computerized hypoglycemia questionnaire. The GW2B group was also asked to report problems with using the device.

Phone contacts were made after 1, 2, and 4 weeks and then monthly through 6 months to review diabetes management. Follow-up visits occurred at 3 and 6 months. At each visit, measures included A1C, frequency of GW2B use, problems encountered with its use, and episodes of hypoglycemia. A standardized assessment of skin reactions attributable to GW2B use was performed.

An initial pool of 34 items was compiled by one of the authors (T.W.) after consultation with other psychologists who were familiar with pediatric diabetes and its management regarding possible affective, behavioral, or cognitive effects of CGM use. The intent in designing the scale was to create a measure of satisfaction and perceived therapeutic impact of current and future CGM devices. The scale is relevant to CGM devices that provide frequent and immediate glucose information and alarms to signal unwanted glucose fluctuations. Based upon the researchers’ collective knowledge of the technology in development, it was the group consensus that these features would characterize most future CGM devices. The initial scale items were drafted and then reviewed for content, format, and wording by the entire DirecNet research group, which includes 11 physicians, 10 registered nurses, a psychologist, and 4 other research professionals. Following several iterations of this review and editing process, a CGM-SAT scale consisting of 37 items was finalized, with a Flesch-Kincaid grade level of 5.3, as calculated by Microsoft Office 2000 (Microsoft, Redmond, WA).

At 6 months, the CGM-SAT was completed. The CGM-SAT consists of 37 statements about CGM use, and respondents rate their agreement or disagreement with each statement on a five-point Likert scale. Among the 37 items, 19 are negatively worded (e.g., “the CGM is uncomfortable or painful”) and 18 are positively worded (e.g., “the CGM makes adjusting insulin easier”). For the latter 18 items, responses were “reverse scored” such that higher scores indicate more favorable satisfaction or greater benefit. Positively and negatively worded items were distributed nonsystematically throughout the sequence of questions. CGM-SAT scores could range from 37 to 185. The questionnaire was completed independently using a tablet personal computer by parents and youths aged ≥11 years on the coordinating center’s secure website. Based on prior experience with similar measures, the researchers felt that many children aged <11 years would have difficulty interpreting certain CGM-SAT items, and so it was administered only to older children. The computerized administration of the CGM-SAT required entry of a response to every item before submission. Among GW2B participants completing the trial, 97 of 98 parents and 66 of 67 youths aged ≥11 years completed the CGM-SAT at the 6-month visit.

At baseline and 6 months, parents and youths aged ≥11 years also completed the Diabetes Worry Scale (15) and the age-appropriate form of the PedsQL diabetes module (16), a diabetes-specific quality-of-life scale. Finally, parents and older youths were interviewed by telephone at baseline and 6 months with the Diabetes Self-Management Profile (17), a structured interview about diabetes self-management. Previous studies (15–17) have demonstrated that each of these measures has adequate psychometric properties.

CGM-SAT scores were summarized with means ± SD. The Kolmogorov-Smirnov test was used to verify normality of distributions. We calculated internal consistency (Cronbach’s α), average interitem correlation, and split-half reliability (Pearson correlation between the even-numbered and odd-numbered items) separately for parents and youths. Associations between CGM-SAT scores and other continuous variables were calculated using Pearson correlation. Since multiple comparisons were performed, only P values ≤0.01 were considered statistically significant. With regard to associations between CGM-SAT scores and the other psychological measures that were obtained, we anticipated that higher CGM-SAT scores (greater satisfaction) would be associated with more favorable quality of life (PedsQL), better overall treatment adherence (Diabetes Self-Management Profile), and lower diabetes-related anxiety (Diabetes Worry Scale). Finally, change in the PedsQL, Diabetes Self-Management Profile, and Diabetes Worry Scale scores during the 6-month study were examined as correlates of CGM-SAT scores.

As described elsewhere (14), 98 (99%) of 99 subjects in the GW2B group completed the 6-month trial. Mean A1C was 8.0% at baseline and 8.1% at 6 months (P = NS). During the 1st month, the number of sensors used averaged 2.1 ± 0.8 weekly, while the number of sensors with ≥8 h of use averaged 1.2 ± 0.7 weekly. Only 16% of the subjects averaged at least 2.0 uses of ≥8 h per week, and none averaged >4.0 per week. By the 3rd month, 7 (7%) of 99 subjects stopped using the GW2B, and by the 6th month, 26 (27%) of 98 subjects had done so. Skin irritation from the GW2B was reported at least once during the 6 months by all 99 subjects. At the 6-month visit, 54 (55%) subjects were considered to have acute GW2B-related changes (mild in 36%, moderate in 19%, and severe in 0). Nonacute changes (e.g., scabbing, dry skin, hypopigmentation, hyperpigmentation, or scarring) occurred in 50 (51%) children. Reasons given for declining use included skin irritation (76%), frequent skips (periods of varying duration in which no glucose estimate is displayed) (56%), excessive alarms (47%), and inaccurate readings (33%).

Means ± SD total scores on the CGM-SAT (maximum possible score 185) were 98.5 ± 23.1 for parents and 98.3 ± 24.1 for youths. Total scores ranged from 44 to 154 for parents and 54 to 154 for youths. Total scores were normally distributed for parents and youths according to the Kolmogorov-Smirnov test.

Table 1 presents means and SD scores (after reverse scoring) for each CGM-SAT item for youths and parents, as well as the correlation coefficient for each item with the total score. Parents’ mean responses were <3.0 (neutral) for 30 of 37 items (81%), and for youths, 27 of 37 mean responses (73%) met this criterion. The items that obtained the three lowest mean scores (lower satisfaction) and the three highest mean scores (higher satisfaction) were identified to further describe the CGM-SAT findings. The three lowest mean scores (1.9 in each case) for parents, indicating low satisfaction with the GW2B, were on item 33 (interferes a lot with sports, playing outside, etc.), item 16 (is uncomfortable or painful), and item 32 (shows more “glitches” and “bugs” than it should). The three lowest mean scores for youths were for item 16 (is uncomfortable or painful; mean = 1.8), item 22 (allows more freedom in daily life; mean = 2.0), item 18 (is more trouble than it is worth; mean = 2.2), item 19 (has helped my family to get along better about diabetes; mean = 2.2), item 33 (interferes a lot with sports, playing outside, etc.; mean = 2.2), and item 36 (causes too many interruptions during the day; mean = 2.2). The items with the three highest mean scores for parents, indicating the aspects of CGM use that were less dissatisfying, were item 1 (causes me to be more worried about controlling blood sugars; mean = 3.6), item 29 (causes our family to talk about blood sugars too much; mean = 3.5), item 14 (sometimes gives too much information to work with; mean = 3.5), and item 5 (makes me think about diabetes too much; mean = 3.5). The items with the three highest mean scores for youths, indicating the aspects of CGM use that they found less dissatisfying, were item 29 (causes our family to talk about blood sugars too much; mean = 3.5), item 25 (has caused more family arguments; mean = 3.4), and item 1 (causes me to be more worried about controlling blood sugars; mean = 3.4).

Cronbach’s α was 0.95 for parents and 0.94 for youths, and average interitem correlations were 0.33 for parents and 0.30 for youths. Split-half reliability was r = 0.91 for parents and 0.93 for youths. Table 1 also presents correlation coefficients for parents and adolescents for each item with the CGM-SAT total score. The average item-total correlation was 0.59 for parents and 0.57 for adolescents, and the range across items was 0.14–0.82 for parents and −0.03 to 0.77 for adolescents. Items with extremely high item-total correlation coefficients could be considered for deletion in subsequent studies using this instrument.

For the 66 cases where both parent and youth (aged ≥11 years) completed the CGM-SAT, the correlation between their scores was 0.68 (P < 0.001).

For parent CGM-SAT responses, a suggestive trend that did not meet the P < 0.01 level for significance was observed with frequency of GW2B use (94 ± 24 vs. 96 ± 22 vs. 110 ± 22 among patients averaging <1.0, 1.0 to <2.0, and ≥2.0 weekly GW2B uses, respectively; ρ = 0.23, P = 0.02). Neither CGM-SAT total or subscale scores correlated significantly with patient age or change in A1C from baseline to 6 months. CGM-SAT total scores did not correlate significantly with either 6 months or change in total scores from baseline on either the PedsQL or Diabetes Worry Scale among parents or adolescents.

A 6-month randomized trial comparing the glycemic effects of current diabetes management with and without supplemental use of an FDA-approved CGM device (GW2B) provided the context for initial validation of a measure of satisfaction with and perceived therapeutic impact of CGM devices such as the GW2B and similar devices as they become available.

The results of the study supported the psychometric properties of the CGM-SAT. The reliability of the instrument was supported by high α coefficients from parents (α = 0.95) and youths aged ≥11 years (α = 0.94). Modest average interitem correlations and estimates of strong split-half reliability were also obtained with parents’ and youths’ responses. Also, parent and adolescent scores were strongly correlated (r = 0.68, P < 0.001).

The high internal consistency coefficients suggest that the scale may contain redundant items. Table 1 presents the item-total correlations for each item for parents and adolescents separately. Those items with very high item-total correlation coefficients may be candidates for deletion in subsequent iterations of the CGM-SAT. However, item deletion is probably inappropriate until the performance of the CGM-SAT can be evaluated with other CGM devices.

The convergent validity of the instrument was supported weakly by marginal associations between parent scores and frequency of GW2B use and adolescent scores with overall diabetes treatment adherence. CGM-SAT scores were not associated with scores on a measure of diabetes-related anxiety (Diabetes Worry Scale) or diabetes-specific quality of life (PedsQL diabetes module) following 6 months of GW2B use or with changes in scores on those measures over 6 months. In the trial, use of the GW2B did not improve metabolic control. The failure to observe a significant effect of the GW2B on diabetes control may be because subjects used the sensor too infrequently to affect glycemic control. At the start of the study, participants were given the device and offered an unlimited free supply of sensors, and they were encouraged during telephone contacts to use the GW2B frequently. However, GW2B use was considerably less than we had hoped. The CGM-SAT provides important insights into why this occurred.

The data presented in this report yielded a detailed description of parents’ and older youths’ satisfaction with use of the device and provided substantial descriptive data that can serve as a standard for comparison in future evaluations of other CGM devices. Mean total scores on the CGM-SAT were 98.5 for parents and 98.3 for youths, yielding mean item scores of ∼2.7 for both. For most items (81% for parents and 73% for youths), the mean satisfaction rating was <3.0, indicating less than favorable reactions to GW2B use. These data suggest that the sample, in aggregate, was mildly dissatisfied with use of the GW2B. The absence of glycemic benefit from using the GW2B and the occurrence of skin irritation, frequent skips, excessive alarms, and inaccurate readings reported by study participants all probably contributed to the declining frequency of GW2B use to a mean of 1.5 times weekly during the last month. Glycemic benefit from use of a CGM device may require that it be used more continuously than occurred here.

CGM-SAT items reflecting the least satisfaction among parents and youths concerned functional aspects of the instrument (skin problems, skipped readings, false alarms, inaccuracy). In contrast, those items that yielded mean ratings conveying less dissatisfaction related to psychological or behavioral effects of CGM use (e.g., anxiety, intrusiveness, family conflict, and information overload). Participants in the study were therefore more dissatisfied with the technical functioning of the device than they were about the psychological ramifications of using it. Psychological effects of CGM use may become more apparent as new devices are introduced and the existing devices are refined.

Limitations of the report include the fact that it was based on data obtained from a sample of patients and parents who may not have been representative of all such families in terms of parental education, socioeconomic status, and motivation to optimize glycemic control. Further studies of the CGM-SAT with a more broadly representative sample of patients and parents will be needed to verify that the present findings apply more generally.

The results point to aspects of the GW2B in particular, and perhaps CGM devices in general, that should be targeted for improvement in order for the technology to be clinically acceptable and effective. These include reducing skin irritation consequent to use of the device, decreasing the frequency of “skips,” reducing “false alarms” that alert the patient to erroneous glucose fluctuations, and improving the overall accuracy of the device. Since the GW2B is among the first such devices to have received FDA approval for clinical use, it is likely that future iterations of this technology will be designed in an effort to overcome these initial obstacles.

Tim Wysocki, PhD, ABPP; Roy W. Beck, MD, PhD; William V. Tamborlane, MD; Rosanna Fiallo-Scharer, MD; Michael J. Tansey, MD; Stuart A. Weinzimer, MD; Craig Kollman, PhD; Katrina J. Ruedy, MSPH; and Dongyuan Xing, MPH.

1) Barbara Davis Center for Childhood Diabetes, University of Colorado, Denver, CO: H. Peter Chase, MD (PI); Rosanna Fiallo-Scharer, MD (I); Jennifer H. Fisher, ND, RN (C); Barbara Tallant, RN, MA (C). 2) Department of Pediatrics, University of Iowa Carver College of Medicine, Iowa City, IA: Eva Tsalikian, MD (PI); Michael J. Tansey, MD (I); Linda F. Larson, RN (C); Julie Coffey, MSN (C). 3) Nemours Children’s Clinic, Jacksonville, FL: Tim Wysocki, PhD, ABPP (PI); Nelly Mauras, MD (I); Larry A. Fox, MD (I); Keisha Bird, MSN (C); Kelly L. Lofton, RN (C). 4) Division of Pediatric Endocrinology and Diabetes, Stanford University, Stanford, CA: Bruce A. Buckingham, MD (PI); Darrell M. Wilson, MD (I); Jennifer M. Block, RN, CDE (C); Paula Clinton, RD, CDE (C). 5) Department of Pediatrics, Yale University School of Medicine, New Haven, CT: Stuart A. Weinzimer, MD (PI); William V. Tamborlane, MD (I); Elizabeth A. Doyle, MSN (C); Kristin Sikes, MSN (C).

Jaeb Center for Health Research, Tampa, FL: Roy W. Beck, MD, PhD; Katrina J. Ruedy, MSPH; Craig Kollman, PhD; Dongyuan Xing, MPH; Andrea Kalajian, MS; Cynthia R. Stockdale.

Gilman D. Grave, MD; Barbara Linder MD, PhD; Karen K. Winer, MD.

Michael W. Steffes, MD, PhD; Jean M. Bucksa, CLS; Maren L. Nowicki, CLS; Carol A. Van Hale, CLS.

Dorothy M. Becker, MBBCh; Christopher Cox, PhD; Christopher M. Ryan, PhD; Neil H. White, MD, CDE; Perrin C. White, MD.

Support provided by National Institutes of Health/National Institute of Child Health and Human Development Grants HD041919-01, HD041915-01, HD041890, HD041918-01, HD041908-01, and HD041906-01 and by Nemours Research Programs. GlucoWatch G2 Biographers were purchased from Cygnus at a discounted price.