We included patients with DM2 of both sexes who met the following inclusion criteria:

• •

  Patients of both sexes between the ages of 29 and 85 years.

• •

  Treated at the University Hospital of Fuenlabrada.

• •

  Who have been diagnosed with DM2, previously or during the study period.

We excluded individuals we could not locate, those who had problems that prevented them from answering the telephone call and those who did not understand Spanish.

After correcting the potential errors in the database, the number of evaluable patients was 2281 (1298 men and 983 women). The sample size was determined by assuming a prevalence of depression associated with DM2 of 25%.17 We assumed a 95% confidence level, a power of 80% and a precision of 2%. In anticipation of study attrition, we increased the sample size by 15%. Based on these assumptions, the sample size was 321 individuals. A randomized sampling was performed, stratified by age and sex. The individuals in the sample were distributed proportionally to the population of each stratum, defined by age and sex. Of the 321 patients in the sample, 46 ultimately did not participate in the study (13 could not be located and 33 did not want to participate); therefore, a total of 275 patients were analyzed.

The patients in the study sample were contacted by telephone to schedule a consultation. After determining whether they met the inclusion criteria and after obtaining their consent, the questionnaires were given to the patients. The information on the demographic variables and life habits was provided during the same interview conducted by a psychiatrist. To obtain the biological measurements, we used the patients’ medical history and the information system established by the University Hospital of Fuenlabrada. During the data analysis, we considered recoding a number of the sociodemographic and medical history variables in other dichotomous types.

The depressive disorder screening was performed using the Beck Depression Inventory (BDI), one of the most widely used instruments for assessing depressive symptoms, which has been validated for Spanish populations.20–23 Given that we already knew the patients who had been diagnosed with depressive disorder by the healthcare system and to assess the discriminative capacity of the BDI scale used to detect the patients with depression in our sample, we constructed a receiver operating characteristic curve. The cutoff of the BDI scale that jointly determines the highest sensitivity and specificity is 16 (Youden's index of 0.809), which coincides with the cutoff of other studies.24 In cases in which the BDI score was ≥16 points, the patients were scheduled in the consultation for a structured clinical assessment. To assess patient drug compliance, we used the Morisky-Green questionnaire.25 The glycated hemoglobin readings were obtained from the patient's medical history, recording the most recent values. Good glycemic control was defined as a glycated hemoglobin level ≤7.

The data were analyzed using the SPSS program v.17.0. The quantitative variables are listed as their mean, standard deviation (SD) and 95% confidence interval (95% CI). We assessed the association between qualitative variables with the chi-squared test. For ordinal variables, we contrasted the hypothesis of an ordinal trend of proportions. We calculated the odds ratio (OR) with its 95% confidence intervals (95% CI). The relationship of the quantitative variables with each of the independent variables (previously transformed into categorical variables) was analyzed using Student's t-test and/or the analysis of variance (ANOVA). We performed a backward stepwise multivariate logistic regression analysis to assess the association of those variables that were significant in the univariate analysis or that were considered clinically relevant. Statistical significance was set at p<.05. Before the start of the study, the potential participants were requested to provide their informed consent. The protocol was submitted for consideration to the Clinical Research Ethics Committee of University Hospital of Fuenlabrada.

A total of 56.4% (95% CI, 50.4–62.1) of the patients were men (Table 1), and the mean age was 64.5 years (standard deviation, 11.6 years). Some 27.6% were smokers at the time of the study, a habit that was more common among the men than among the women (35% and 18%, respectively). A total of 43.3% drank alcohol (men, 73.1%), and 8.7% were on treatment with insulin. Some 88.4% were on treatment with oral diabetic drugs, and 2.9% were on treatment with insulin and oral diabetic drugs.

Overall, the study population had a mean body mass index of 30.3kg/m2 (grade 1 obesity, 29.7kg/m2 [95% CI, 27.5–31.9] in men and 31.1kg/m2 [95% CI, 27.3–34.9] in women).26 A significant increase was observed in the prevalence of obesity as age increased, which was especially notable in the women.

The mean score on the BDI scale for the women (19.03) was significantly higher than that of the men (12.39). The BDI score increased with age, both in the men and women.

The prevalence of depression was calculated at 32.7% (95% CI, 27.4–38.5). This rate was somewhat higher in the women (43.3% [95% CI, 43.8–52.3]) than in the men (24.6% [95% CI, 18.4–31.8]) (Table 2). A significant increase was observed in the prevalence of depression as age increased (Fig. 1). The marital status with the highest prevalence of depression was widowers. In all marital statuses, the prevalence of depression was higher in the women, except for divorces, in which the rate was higher for the men. The risk of experiencing depression was 2.7 times higher among the participants who did not work, which in this case was also higher among the women. The prevalence of depression increased in inverse proportion with the level of education, with a 5-fold lower risk among the participants with university studies.

Figure 1.

Prevalence of depression in patients with type 2 diabetes mellitus, as a function of age and sex.

(0.05MB).

The lifestyle factors associated with a greater prevalence of depression include not performing physical activities, with a 2.5-fold greater risk in this group than in those who performed physical activities. The prevalence of depression was significantly higher in the patient group on treatment with insulin. Obesity was associated with a 2.6-fold greater risk, which was more pronounced in the women. Poor compliance tripled the risk of depression, which was also more pronounced in the women. The prevalence of depression was 2.4-fold greater in the patient group with diabetic complications compared with the group with no complications (40.3% vs. 17%, p<.05). The time from the diabetes diagnosis in the patient group with depression was statistically significant (18.9 years [95% CI, 11.7–26.3]), while for the patient group without depression, this period was 13.0 years (95% CI, 6.5–19.5).

As can be seen in Table 3, 35% (95% CI 26.4–45.8) of the patients who had a BDI score ≥16 had not been diagnosed with depression. The women had been diagnosed more frequently than the men (71.2% vs. 55.3%). Depression was diagnosed less often in the participants who were single, those who had a low social class and those with a high education level. The condition is diagnosed more often as age increases and in smokers, drinkers, individuals who have sedentary lifestyles, individuals with a family history of diabetes, those undergoing treatment with oral diabetes drugs or insulin, people with obesity and those who state that they are in good compliance with the prescription.

Some 91.8% of patients who had been diagnosed with depression were on treatment with psychoactive drugs, and 14.8% were in psychotherapy. The multivariate analysis indicated that the prevalence of depression is related to sex (greater in women), age (increasing prevalence with increasing age), marital status (greater in widowers than in other marital statuses), obesity (greater in patients with diabetes with BMI≥30), and treatment compliance (greater in those who are noncompliant) (Table 4).