Glycemic status and health-related quality of life (HRQOL) in populations at risk of diabetes in two Latin American cities

A cross-sectional study was designed from the study Evaluation of a community health program to prevent type 2 diabetes and other cardiometabolic risk factors in adults (–PREDICOL-ClinicalTrials.gov identifier NCT03049839). The methodology of this study has been previously published [24]. In summary, 1135 participants were recruited in the PREDICOL clinical trial, developed in neighborhoods with low socioeconomic strata in two Colombian cities between 2018 and 2019. The cities included in this study were Bogotá (large city) and Barranquilla, the first and fourth cities, respectively, with the largest number of inhabitants in the country [34, 35].

HRQOL was assessed using the generic EQ-5D-3L questionnaire, which comprises five dimensions: Mobility (MO), Self-care (SC), Usual activities (UA), Pain/discomfort (P/D), and Anxiety/depression (A/D). Each dimension consists of three severity levels corresponding to no problem (1), moderate problems (2) and extreme or severe problems (3). The combination of the digits of all dimensions generates a 5-digit number describing the participant's health status, there being 243 possible health states [30, 36]. For example, state (11111) represents that the participant reports no problems in any dimension and (33333) represents extreme or severe problems in all dimensions, while state (11121) indicates moderate pain/discomfort problems and no problems for the other dimensions. Moreover, the participant should mark the point on the vertical line of a 20 cm Visual Analog Scale (VAS) rating of his or her overall health status denoted as worst imaginable health status (0) to best imaginable health status (100) [30].

Because of Colombia does not currently have an EQ-5D-3L value set [37, 38], the EQ-5D score in this study was calculated using the Latin American value set [39] as suggested by recent guidelines from agencies in charge of health economic evaluations [40, 41]. This set has 243 utility values which were obtained by the time trade off valuation technique "TTO" where the states (11111) and (33333) represent utility values of 1 and − 0.101, respectively. Negative HRQOL values mean that some health states are considered worse than death.

The risk of T2D was identified and stratified by means of the Finnish Diabetes Risk Score (FINDRISC) questionnaire. The FINDRISC is a questionnaire with eight questions, in which each response is assigned a score, 8 items are considered with categorized responses on Body Mass Index (BMI), age, waist circumference, self-reported physical activity, daily consumption of fruits or vegetables, history of glycemia alteration, and family history of T2D [3, 4]. This instrument has been validated and adapted for various populations worldwide as in different Latin American countries, including Colombia, where the weighting and stratification of the total risk is obtained from the simple sum of the individual weights of each response. The FINDRISC score ranges from 0 (minimum score) to 26 (maximum score) points, and, for Colombia, a FINDRISC score ≥ 12 defines a risk of developing T2D in the next 10 years [5‐7].

An oral glucose tolerance test (OGTT) was performed on study participants identified as being at risk for T2D (score ≥ 12 on the FINDRISC questionnaire). According to the recommendations and diagnostic classification criteria issued by the World Health Organization, an initial fasting glycemia was performed, followed by providing the patient with a solution containing 75 g of glucose; where two hours after ingestion of the glucose serum, a second glycemia sample is taken (Glycemia 2 h) [42]. Participants who had a fasting plasma glucose (FPG) level ≥ 126 mg/dl or plasma glucose at 2 h (2hPG) ≥ 200 mg/dl were classified as UDT2. Those with 2hPG ≥ 140 mg/dl, but < 200 mg/dl and FPG < 126 mg/dl were classified as impaired glucose tolerance (IGT). Impaired fasting glucose (IFG) was defined as FPG ≥ 110 but < 126 mg/dl and 2hPG < 140 mg/dl. Participants with FPG ≥ 110 but < 126 mg/dl and 2hPG ≥ 140 mg/dl but < 200 mg/dl were defined as combined IGT and IFG. Finally, participants with values below 110 mg/dl in basal glycemia and 2hPG < 140 mg/dl respectively were classified as normotolerant (NGT). Prediabetes was defined as participants with some abnormal glucose tolerance (IFG, IGT, or IFG and IGT) [42].

To determine the level of physical activity (PA), the Physical Activity Questionnaire-Short Version (IPAQ-SV) was used, which consists of 7 questions that address the frequency, intensity, and duration of PA performed in the last 7 days [43]. Additionally, a survey where data on socioeconomic factors such as marital status, occupation, educational level, sex, health regimen, and cardiovascular history were collected was applied.

This study was approved by the ethics committee of the Universidad del Norte by means of Evaluation Act 141 of April 28, 2016, under the norms of good clinical practice and the guidelines of the Declaration of Helsinki. Each participant was provided with basic information about the study, confidentiality, and data protection; they signed the informed consent before participation and could withdraw from the study whenever they considered it necessary.

Qualitative characteristics of the participants were expressed as absolute numbers and percentages, while quantitative characteristics were expressed as means with standard deviations. The nonparametric distribution of the variables was determined by Kolmogorov–Smirnov or Shapiro–Wilk tests. To evaluate significant differences in quantitative characteristics according to glycemic status, Mann–Whitney and Kruskal–Wallis U tests and the Chi-square test (X²) were used for categorical variables. The percentage of problems reported by the participants in the 5 dimensions of quality of life are shown in double bar graphs according to glycemic classification. We examined significant differences in mean EQ-5D and VAS scores among individuals with NGT, prediabetes, and UT2D. Binary logistic regression models by the glycemic group were used to establish statistically significant factors associated with presenting problems in the EQ-5D-3L dimensions. The dependent variable for the logistic regression models was each quality-of-life questionnaire dimension because the percentages of severe or extreme problems (level 3) in the dimensions were very small or null. Thus, we dichotomized the levels of each dimension into "no problems" for level 1 and "some problems" for levels 2 and 3. We then included all variables that had shown statistical significance (< 0.05). However, to summarize the results, only variables that had a statistically significant effect with at least one dimension of quality of life were reported. Finally, to find factors that affected participants' HRQOL, a Tobit regression model was constructed for each glycemic group [44], using the EQ-5D score as the dependent variable. We considered the Tobit regression model appropriate for two reasons. Firstly, the EQ-5D scores are biased and censored at − 0.101 and 1. And secondly a high percentage of the scores (30.6% n = 347) are censored at the upper limit of 1. P-values below standard significance (P-value < 0.05) were considered statistically significant. Data processing and analysis were performed in R statistical software version 4.0.0.

Data from 1135 participants (76.4% female) were analyzed. Table 1 shows the baseline characteristics of the participants overall and by glycemic classification. The majority were classified as NGT (n = 851, 75.0%), while 16% (n = 182) had prediabetes (IFG, IGT, or IFG and IGT), and 9.0% (n = 102) had undetected type two diabetes (UT2D). Mean age was significantly higher in the prediabetes or UT2D groups (P-value < 0.001). Most participants with high glucose levels (Prediabetes or UT2D) were female and from the city of Barranquilla. Participants with UT2D had a higher BMI (31.1 kg/m²) compared to NGT (29.6 kg/m²). Likewise, mean systolic and diastolic blood pressure was significantly higher in individuals with higher glucose levels. In addition, participants with prediabetes or UT2D had a larger waist circumference than NGTs. The percentage of participants with NGT who received treatment for hypertension was 34.9%, being higher than the percentage presented in participants with prediabetes (51.6%) and UT2D (45.1%). Comparisons between glycemic status and education level, marital status, daily fruit intake, physical activity, and total cholesterol levels were not significant (P > 0.05).

The percentages of moderate or severe problems reported by participants in each glycemic group in the five dimensions of the EQ-5D-3L description system are shown in Fig. 1. Most participants reported having no problems in several dimensions of the quality-of-life questionnaire. The dimensions with the least reported problems were self-care (NGT = 1.6%; prediabetes = 3.3%; and UT2D = 2.9%) and usual daily activities (NGT = 14.2%; prediabetes = 16.5%; and UT2D = 19.6%). However, at a high percentage the participants in this study reported some problem (moderate or severe) in the dimensions of mobility (NGT = 21.3%; prediabetes = 20.3%; and UT2D = 27.4%), pain/discomfort (NGT = 60.5%; prediabetes = 59.9%, and UT2D = 64.7%), and anxiety/depression (NGT = 32.2%; prediabetes = 25.3%; and UT2D = 29.4%). The percentage of problems in each of the dimensions did not differ significantly when compared by glycemic status.

Table 2 shows the mean scores of the EQ-5D-3L and EQ-VAS according to demographic and clinical characteristics of the participants by glycemic status. Mean utility scores were statistically similar in the NGT (0.80; 95% CI 0.79–0.81), prediabetes (0.81; 95% CI 0.79–0.83), and those with UT2D (0.79; 95% CI 0.76–0.82) groups. Likewise, mean EQ-VAS scores did not differ significantly by glycemic status. However, mean EQ-5D and EQ-VAS scores were significantly lower in NGT (0.79 vs. 0.83 for EQ-5D and 74.3 vs. 77.4 for VAS score) and prediabetic (0.78 vs. 0.86 for EQ-5D and 72.1 vs. 78.3 for VAS score) women compared with men in the same glycemic group. Participants older than 64 years with NGT had significantly lower EQ-5D and EQ-VAS scores compared to younger NGTs. In terms of city, mean HRQOL scores were lower in Bogotá residents in each of the glycemic groups. Divorced or widowed NGTs or UT2D reported statistically lower EQ-5D and VAS scores compared to married people in the same glycemic group. The EQ-5D and EQ-VAS scores of those NGT participants receiving treatment for hypertension were statistically lower compared to those who reported no hypertension problems. However, mean EQ scores across the different glycemic groups were similar in terms of participant BMI classification.

Table 3 shows the results of the logistic regression models by glycemic group for each of the five dimensions of the EQ-5D-3L. The factors that were significantly associated with a higher report of problems in some dimension of quality of life in participants with NGT were the largest city, oldest age, female sex, hypertension, education, and low physical activity. In prediabetes, they were largest city, oldest age, female sex, and low physical activity; while, in those with UT2D, only largest city and female sex were significantly associated. In the prediabetes group, larger city (OR 2.3; 95% CI 1.2–4.5) and female sex (OR 2.4; 95% CI 1.3–4.6) were significantly associated with higher report of problems in the P/D dimension. In the UT2D group, larger city (OR 5.2; 95% CI 1.9–13.8) and older age (OR 5.3; 95% CI 1.4–19.9) were significantly associated with higher report of problems in the MO dimension; likewise, larger city was associated with the UA and A/D dimensions. In the SC dimension, no significant associations were found in the prediabetes and UT2D groups. Likewise, no significant associations were found in participants with UT2D in P/D.

The results of the Tobit regression models for each of the glycemic states, like the model for the total population, are shown in Table 4. The variables for the EQ-5D score in the total population showed statistical significance in women), in participants who were aged 45–54 years, 55–64 years, older than 64 years, and for those living in the smallest city. The results of the Tobit regression model for the NGT group showed significant differences in quality of life in women, in participants aged 45–54 years, those aged 55–64 years, those living in the smallest city, and participants with primary education. In prediabetes, the significant results were in women, in those living in the smallest city, and in participants receiving treatment for hypertension), while, in UT2D, only city and marital status yielded significant results.