Brazilian multicenter study for the evaluation of patients’ satisfaction of blood glucose self-monitoring with BGStar^® blood glucose meter in insulinized patients with diabetes mellitus type 1 and 2

All patients who signed the informed consent form (ICF) were considered as screened. Patient who signed the ICF and were not a screen failure were included in intent-to-treat (ITT) population. All ITT patients who had used at least once the investigational BGM were also considered in the modified intent-to-treat (m-ITT) population. Safety population consisted of all enrolled patients. All efficacy analyses were based on m-ITT population, and all safety analyses on ITT population.

Quantitative data were summarized by descriptive statistics such as mean, standard deviation, minimum and maximum and the categorical data were summarized by absolute (n) and relative (%) frequency of occurrence.

All results were based on total number of patients with available information (i.e. patients who attended the visit and for whom the specific data was not missing). The type I error adopted for the statistical tests was 5 %. All analyses were conducted using SAS version 9.4.

Satisfaction analysis (overall satisfaction with the investigational BGM) were assessed using the Visual Analogue Scale (VAS) and a specific questionnaire of satisfaction with seven items, applied at initial visit (V₁), final visit after 3 months (V₄) and extension visit after 6 months (V₅). For each efficacy endpoint (VAS and the sum of the seven-item satisfaction questionnaire), a model of analysis of variance (ANOVA) was applied with a fixed factor of repeated measures (VISIT), since the same patient is evaluated in two or more visits. The main objective of the ANOVA model was to identify if there was a “VISIT effect”, i.e. to identify if there was any change in the patients satisfaction across V₁, V₄ and V₅.

Global satisfaction was also analyzed in subgroups (age, type of diabetes and insulin delivery system). An ANOVA model was also applied in this case, including the subgroup factor and the interaction with VISIT. The objective was to identify if a possible change observed along the visits in the patients’ satisfaction was similar by each subgroup category. When the factor VISIT was significant in an ANOVA model including the extension period, multiple comparison tests were carried out.

Each of the individual seven-item questionnaire scores were also analyzed in order to see any specific change from V₁ (satisfaction compared with previous BGM) and V₄ (satisfaction after 12 weeks using the investigational BGM), through the non-parametric Wilcoxon rank sum test.

Between 5th April 2013 and 16th June 2014, 293 patients were enrolled in the study. All of them had signed the informed consent form, but one was excluded (screen failure) due to the use of insulin pump therapy. ITT population was similar to the safety population and comprised 292 patients. From this total, 268 patients (91.8 %) constituted the modified-ITT (m-ITT) population, as 8.2 % of the patients did not use at least once the investigational BGM (Fig. 1).

From the ITT population, 270 (92.5 %) patients attended the endpoint visit (V₄), 82.9 % accepted to participate in the extension period and 80.8 % indeed attended V₅. Twenty-nine (9.9 %) patients did not complete the study according to the protocol, wherein seven of them had agreed to proceed to extension visit, but did not perform V₅ (Fig. 1). Reasons for the patients did not complete the study were: lost to follow up (n = 16), consent withdrawal (n = 12) and death (n = 1).

Four patients were considered as protocol deviation. Two of them did not fulfill the criterion “type 2, receiving insulin for at least 2 years with a stable insulin regimen for at least 6 months”, one patient did not fulfill the criterion “use of any BGM for at least 6 months, with ability to perform blood glucose self-monitoring” and the other one did not fulfill the criterion “age ≥18 years old”.

Age of patients ranged from 17.9 to 86.1 years, with mean age of 50.9 (SD = 17.3) and median of 54.7 years. Despite the protocol inclusion criteria “age ≥18 years old”, one patient aging 17.9 years was screened and performed all planned visits; this patient was considered for both ITT and m-ITT analyses. Regarding gender, 124 (42.5 %) patients were male and 168 (57.5 %) female. Body mass index (BMI) ranged between 17.6 and 47.7 kg/m², with mean of 28.9 kg/m² (SD = 5.3). Means of systolic and diastolic pressure were 129.9 (SD = 20.5) and 78.4 (SD = 12.4) mmHg, respectively (Table 1).

Most patients (185; 63.4 %) had type 2 diabetes mellitus. Mean time of diabetes diagnosis was 15.0 years (SD = 9.1). Basal-bolus was the insulin regimen application at baseline for most patients (60.6 %) (Table 1).

Adherence to investigational BGM was evaluated at visits (V₂, V₄ and V₅) by an investigator. Adherence was categorized, according to the protocol definition, as Total (BGM was used at least once a day), Partial (BGM was used at least once a day, between 2 and 6 days of the week) and No Adherence (BGM was used at most once a week). Most of patients (71.3 %) totally adhered to BGM at V₂, but this number decreased at V₄ (62.8 %) and extension visit V₅ (52.8 %), while Partial adherence remained stable (Table 2).

The global satisfaction with the investigational BGM after 12 weeks of use was compared to the global satisfaction with previous BGM, based on VAS score (0–100 mm) and on the sum of a seven-item questionnaire. In the satisfaction questionnaire based on VAS, higher values on the scale are associated to higher satisfaction; moreover, in the analysis of the seven-items questionnaire, a lower score is associated to higher satisfaction. From the 268 patients of m-ITT population, 258 patients attended V₁ and V₄, and answered the global satisfaction questionnaire. Of these, 176 (68.2 %) assigned a higher VAS score after 12 weeks of use of investigational BGM compared with previous BGM, 47 (18.2 %) indicated no change and 35 (13.6 %) assigned a lower score after 12 weeks. A statistically significant improvement of satisfaction was observed when using investigational BGM for 12 weeks: mean VAS score raised from 78.8 mm (SD = 18.0) at V₁ to 90.8 mm (SD = 12.2) at V₄. Mean difference on VAS score was 11.9 (SD = 20.4; p < 0.001; 95 % CI 9.4–14.4 mm) (Table 3).

Mean change of VAS score from V₁ to V₄ was also analyzed for the subgroups: Type of diabetes, Insulin regimen application and Age (Table 4). There was no observed statistically significant effect (p = 0.221) for the interactions VISIT*TYPE, indicating that the increase in mean satisfaction with the use of investigation BGM for 12 weeks was similar for patients with T1DM (14.0 ± 20.8 mm) and T2DM (10.8 ± 20.2 mm). Regarding the subgroup insulin regimen application, interaction VISIT*REGIMEN was not statically significant (p = 0.133), showing that the increase in the satisfaction with the investigational BGM was similar for patients at basal-bolus (13.5 ± 20.0 mm) and at other application regimens (9.5 ± 20.9 mm).

Moreover, a statistically significant effect was observed for the interaction VISIT*AGE (p = 0.023). The increase observed in the satisfaction based on VAS after 12 weeks of use of investigational BGM was inversely proportional to the age. By observing the same three age groups in the visit V₁ (satisfaction with the previous BGM), we observed that younger patients (age <25 years) were more dissatisfied than older patients (age ≥65 years), according to Turkey’s method for multiple comparisons (17.6 ± 19.9 versus 6.2 ± 19.6 mm, respectively; global significance level = 0.05) for the age groups in V₁. In the satisfaction evaluation performed at V₄ (after 12 weeks of use of investigational BGM), there was no statistically significant difference among the three age groups. It is also important to note that in the group of patients with ≥65 years the initial evaluation was already high, thus the difference was not significant. Also, at week 12, all age groups showed very similar satisfaction levels according to VAS scores means, which may indicate that satisfaction with the investigational BGM may increase with its use, and the subsequent raise in the ability to handle it.

Regarding global satisfaction based on a seven-item questionnaire sum, 253 patients from the m-ITT population (n = 268) answered all questionnaires, and were therefore included in the analysis. The mean sum of questionnaire scores was 15.8 (SD = 3.8) at V₁ and 12.0 (SD = 3.4) at V₄, meaning an improvement of satisfaction when using investigational BGM for 12 weeks (p < 0.001). There was no interaction effect of any of the subgroups tested—VISIT*TYPE (p = 0.113), VISIT*REGIMEN (p = 0.439) and VISIT*AGE (p = 0.560)—meaning that changing of satisfaction score from V₁ to V₄ was similar for all age groups, both type 1 and type 2 diabetes, and also basal-bolus versus other insulin application regimens. Global satisfaction with investigational BGM after 12 weeks of use was also compared with previous BGM through each of the individual seven-item questionnaire. All seven-item scores (except item 6: “On average, how often do you stop to take the blood glucose tests recommended by your doctor with the current BGM/investigational BGM?”) decreased significantly (p < 0.001), indicating better satisfaction after 12 weeks usage of investigational BGM.

Global satisfaction levels after 24 weeks using investigational BGM was compared to those obtain with previous BGM and investigational BGM after 12 weeks of use. Levels were based on VAS score (0–100 mm) and on the sum of a seven-item questionnaire. From the 268 patients of m-ITT population, 224 patients attended V₁, V₄ and V₅ and answered the global satisfaction questionnaire based on VAS; and 219 answered the seven-item questionnaire. Mean VAS score raised from 78.6 mm (SD = 18.4) at V₁ to 91.3 mm (SD = 11.3) at V₄ and remained almost the same at V₅, with mean score of 90.3 mm (SD = 13.5) (Table 5). Mean sum of satisfaction questionnaire score was 15.7 (SD = 3.9) at V₁, 11.8 (SD = 3.3) at V₄ and 12.0 (SD = 3.6) at V₅. Multiple comparisons showed higher mean score at V₁ than V₅, indicating an increase in the satisfaction after using the investigational BGM for 24 weeks, but similar mean scores at V₄ and V₅, showing that the satisfaction was maintained from V₄ to V₅ (Table 6).

Ten (3.4 %) patients reported a total of 13 adverse events during the study period, including the 12-week extension. One patient reported four events, and all the others presented only one event each. One (0.3 %) patient had an adverse event considered serious: fall with mild traumatic head injury due to symptoms of hypoglycemia; this event was the only one related to the investigational BGM (product technical complaint, PTC). This case was analyzed by the legal manufacturer; the malfunction of the device was detected and subsequently corrected. Two (0.7 %) patients reported adverse events related to hypoglycemia: 1 (0.3 %) patient at V₅ (considered serious, with blood glucose level ≤36 mg/dl), and 1 (0.3 %) patient at V₂ (considered non-serious, with blood glucose level ≤70 mg/dl). None of the events resulted in permanent discontinuation of BGM use (Table 7).

There were 21 (7.2 %) patients with at least one PTC, resulting in 27 reported PTCs. Nineteen PTCs were reported by the patients, 4 by the investigators and four by the study staff.