We conducted an observational, cross-sectional study between September 2010 and May 2011, which counted on the participation of 661 primary and specialized care centers belonging to Spain's public healthcare system, proportionally distributed among 17 autonomous communities, selected by convenience sampling.

One investigator per center (73% primary care and 27% specialized care) participated in the study. The physician specialists who collaborated were distributed among the following specialties: endocrinology and nutrition, 18%; internal medicine, 8%; and other specialties, 1%. All physicians selected approximately 6 patients (by consecutive sampling) who met the following inclusion criteria: diagnosis of DM2 at least one year prior to their participation in the study, age ≥18 years and undergoing antidiabetic drug treatment. The exclusion criteria were the need to translate any of the study tools, participation in a clinical trial at the time of the study or having participated in a study in the 6 months prior to their inclusion and a clinical situation that prevents study participation, in the investigator's opinion. The data were collected during a single visit not specifically scheduled for the study (according to standard clinical practice).

The protocol and related documents were approved by the Ethics Committee for Clinical Research of the Hospital Clinic of Barcelona. All patients participating in the study received a study information sheet and signed the corresponding informed consent for participation before their inclusion.

The researchers collected the patients’ demographic and clinical characteristics (age, sex, employment status, height, weight, waist circumference, smoking habit, drinking habit, date of diagnosis, family history of DM2, microvascular and macrovascular complications, hospital admissions and emergency department visits associated with hypoglycemic episodes) by consulting their medical history. The patients completed the items concerning hypoglycemic episodes (presence or absence during the 6 months prior to inclusion in the study, number, frequency, severity and absence from work due to hypoglycemia in the last 6 months). To this end, the patient case report forms included the following definition of hypoglycemia: “The presence of low blood sugar levels that occur with discomfort that you probably notice, such as dizziness, lightheadedness, palpitations, cold sweats, tremors, nervousness, headache and voracious hunger. If your blood sugar levels fall significantly, you might experience drowsiness, confusion, behavioral disorders and even loss of consciousness.”

The patients also completed 2 questionnaires: the Audit of Diabetes-Dependent Quality-of-Life (ADDQoL)12 and the Worry Subscale of the Hypoglycemic Fear Survey-II (HFS-II).13 The specific ADDQoL questionnaire consists of 2 general items in which the patient scores their current overall quality of life (from −3=terrible to 3=excellent) and their quality of life if they had no diabetes (from 1=worse to −3=much better), respectively, as well as 19 specific items that explored the impact of diabetes on various aspects of life (from −9=maximum negative impact to 3=maximum positive impact). The mean score achieved on the 19 items indicates the mean impact of diabetes on the patients’ HRQL. The HFS-II subscale contains 18 items that measure the degree of worry about hypoglycemia in the last 6 months (from 0=none to 3=almost always). Their total score can range between 0 and 72 points (higher scores indicate increased fear). Both questionnaires are translations validated by the MAPI Institute14 and are certified by the authors. The ADDQoL questionnaire has been employed in Spain15,16 but has not been used with patients with DM2 with or without hypoglycemia. The HSF-II subscale has not been previously used for the Spanish population; we therefore proceeded to validate the subscale in another phase of this study.17

Four questions were designed for the purposes of this study to explore the knowledge and attitudes of physicians concerning patients’ HRQL and worry about hypoglycemia: (1) “Do you use questionnaires to measure the HRQL of your patients with DM2?”; (2) “To what extent do you believe you know the risk of hypoglycemia associated with treatments for DM2?”; (3) “How often would you say you explore the degree of worry that hypoglycemia causes in your patients?”; and (4) “How often would you say that hypoglycemia requires you to modify the indicated therapeutic regimen?” The physicians answered each of the 4 questions using a visual analog scale of 10cm, with 0 indicating none/never and 10 indicating the maximum/as often as possible.

We proposed to detect a difference between the scores given by the patients who reported hypoglycemic episodes and those who reported none, in terms of the ADDQoL questionnaire and the HFS-II worry subscale, the main outcome variables. After a review of previous studies that used the same instruments, we selected the mean of the scores assigned to the 19 specific items of the ADDQoL18 as a reference for the calculation. Based on the difference between the study groups and assuming a standard deviation (SD) of 0.5, we calculated an effect size of 0.12. Based on this figure and applying a 95% statistical power and an alpha error of 0.05, we determined that a total of 3784 patients were needed. Assuming a 5% loss, the sample size was calculated at 3973 patients.

A descriptive analysis was performed on the sample, which determined the mean, SD, median, minimum, maximum, percentiles (for quantitative variables) and frequency and 95% confidence intervals (CI) (for qualitative variables). We compared the characteristics of the patients who experienced hypoglycemia in the last 6 months with those of the other patients, using the t-test for continuous variables and the chi-squared test for the categorical variables. We calculated the 95% CI for the mean difference of the ADDQoL and HFS-II subscale scores between the 2 populations. Finally, we performed a correlation analysis (Pearson's correlation test) and a test for independence between the mean scores for the 19 specific items of the ADDQoL questionnaire and the overall scores on the HFS-II subscale. The SPSS v.18 statistical program was used for the calculations.

The study included 4054 patients, 242 (6%) of whom were excluded for not meeting the selection criteria (231 patients with a DM2 diagnosis less than 1 year ago, 8 individuals younger than 18 years, 1 patient requiring translation of the tools and 2 participants who failed to meet several of these criteria). Of the 3812 patients included in the analysis, 54% were men, with a mean age of 64 years±11 years and a mean time from diagnosis of 10 years±7 years. Table 1 lists the demographic and clinical characteristics of the population based on the presence (45%) or not (55%) of hypoglycemia during the previous 6 months. The patients who reported hypoglycemic episodes had longer times from disease diagnosis and lower body mass index. The number of participants with a known family history of DM2 and the presence of microvascular and macrovascular complications was higher. For these patients, the mean number of hypoglycemic episodes in the last 6 months was 5±11 years. The morning was the most common time of day for the episodes (41%) and was when they were most severe (43%). For 14% of these patients, the hypoglycemic episodes caused work absences. Eleven percent and 26% of the patients with hypoglycemia required a mean of 1.63±1.34 and 1.86±1.32 hospitalizations and emergency department visits in the last 6 months, respectively.

The mean scores assigned by the patients to the general items of the ADDQoL (Table 2) were between “neither good nor bad” (0 points) and “good” (1 point), both for participants who experienced hypoglycemia and those that did not (0.37±0.98 vs. 0.40±0.98; p=.28). Their assessment of their quality of life had they not had diabetes was located between “a little better” (−1 point) and “much better” (−2 points), indicating a negative impact of diabetes on their overall quality of life, showing no significant differences between the 2 groups (−1.41±0.87 vs. −1.36±0.85; p=.11). The mean scores for each of the 19 specific items were lower for the patients who experienced hypoglycemia (p<.001 in all cases), with the “freedom to eat” and “freedom to drink” items having a more negative impact. The overall impact of diabetes on the quality of life (mean score on the 19 items) was greater among those patients who had experienced hypoglycemia (−2.48±1.61) than on the others (−1.64±1.36) (p<.001).

By comparing the distribution of the scores assigned by the patients in the subscale HFS-II according to whether they reported hypoglycemic episodes or not, we found statistically significant differences (p<.001) between the 2 groups in the 18 items of the subscale. Thus, the patients with hypoglycemia indicated a greater frequency of fear for the issues measured by each item, compared with the patients who did not experience hypoglycemia. Additionally, the mean score on the HFS-II subscale for the patients with hypoglycemia (31.32±15.71) was greater than that achieved by the others (18.85±16.03) (p<.0001; SD, 12.47; 95% CI 11.42–13.52), indicating a greater degree of worry about hypoglycemia in the first group.

Finally, the results showed a correlation between the mean scores on the ADDQoL questionnaire and the HFS-II subscale (correlation value, −0.441), indicating that as the worry grew, the lower the patients’ perceived HRQL (p<.001 in the test for independence).

On a scale from 0 to 10, the scores attributed by the physicians reveal that they do not usually employ questionnaires to measure their patients’ HRQL (mean, 2.83±2.74 points; median, 1.91), even though they state that they understand the risk of hypoglycemia associated with DM2 treatments (mean, 7.80±1.75 points; median, 8). The doctors discreetly explore the worry that hypoglycemia creates in their patients (mean, 5.92±2.52 points; median, 6.26) and occasionally changed the patient's indicated therapeutic regimen due to the presence of hypoglycemia (mean, 5.64±2.78 points; median, 5.74).