This study was part of improving the QoL in the CV risk factors management in asymptomatic diabetic subjects outcomes research (CV risk in DM Outcomes Research), which was a prospective, observational study conducted in 24 tertiary hospitals in Korea [15]. The study included patients who were with type 2 DM, aged ≥40, and taken first ever carotid artery ultrasound (CUS) but excluded patients who had previously undergone CUS, who had a history of coronary artery disease, symptomatic congestive heart failure, coronary revascularization, cerebrovascular disease, stroke, transient ischemic attack, or documented peripheral vascular disease (e.g., peripheral artery disease, abdominal aneurysm, or carotid artery stenosis), or who were participating in any interventional studies. Calculation of sample size on basis of statistics is not required since this study was an observational study to describe further on the patients without any hypothesis. Based on the origin of this study, all of the participating investigators made a consensus that 40 patients for 3 months considering the inclusion/exclusion criteria, from each participating hospitals are sufficient to be representative to the target population and thus finally recruited 799 patients.

Prior to the patient enrollment, the study protocol was approved by the Institutional Review Boards of each participating institution of the study (IRB No. 09-104, Hallym University Kangdong Sacred Heart Hospital). Written informed consent was obtained from all individual participants included in the study.

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards.

Physicians' and patients' behavioral changes were observed after they became aware of the patients' CV risks through CUS. Changes in physicians' behaviors were observed based on changes of treatment pattern (antihypertensive, lipid-lowering, and anti-platelet therapy). Changes in physician behaviors were defined as any changes in the dose (including increasing or decreasing) or class of prescribed drugs (including adding on or reducing drug classes of antihypertensive, lipid lowering, and anti-platelet therapy) within the 6-month period.

We defined the changes in patient behaviors based on lifestyle modification (smoking, drinking, exercise, stress, nutrition management, and drug compliance). Smoking, drinking, exercise, and stress were measured using the corresponding items on the Korea National Health and Nutrition Examination Survey (KNHANES) [18], and nutritional management was assessed based on the Customized Visiting Health Management Project Needs Survey by the Ministry of Health and Welfare [19]. Drug compliance was measured using the Korean-translated self-reported questionnaire developed by Morisky et al. [20]. Changes in patients' behaviors were defined as any positive changes in smoking, drinking, exercise, stress, nutritional management, or drug compliance scores within the 6-month period.

The validated K-ADDQoL, a specialized instrument for patients with DM, was used to ensure high reliability [21]. The ADDQoL is widely used to measure QoL of patients with DM, because of its higher sensitivity than other general QoL measures [22].

The ADDQoL begins with two global questions. The first global question, ‘In general, my present quality of life is,’ aims to provide a possible single-item indicator of QoL by assessing the participants' present global QoL. The second global question, ‘If I did not have diabetes, my quality of life would be,’ is a diabetes-specific item that aims to assess the diabetes-dependent global QoL.

The subsequent 19 domain-specific items ask the participants to rate how particular aspects of their lives would be if they did not have DM. For each of these items, participants provide the impact (range, −3 [greatest negative impact] through 0 [no impact] to +1 [positive impact]) and importance (range, 0 [not at all important] to 3 [very important]) scores. Weighted impact (WI) is calculated by multiplying importance and impact ratings, ranging from −9 (negative impact) to +9 (positive impact). The products of these two ratings for each domain are summed and then divided by the number of applicable domains to produce the average weighted impact (AWI) score.

The AWI is the mean WI for each of the 19 domains, ranging from −9 to +9. A lower AWI (closer to −9) reflects that DM has a more negative impact on the patient's life, whereas a higher AWI (closer to +9) reflects that DM has a less negative impact on the patient's life (Supplementary Table 1) [23].

Continuous data were presented as mean±standard deviation, and categorical data were presented as frequency and percentage. K-ADDQoL results were shown with the mean of impact, importance, and WI for each 19-domain and AWI. We compared the changes of WI and AWI in the 6-month period using paired t-tests. Changes in WI in relation to physician and patient behaviors were compared using t-tests.

To identify the factors affecting AWI, we performed a univariate analysis with the participants' demographic and clinical haracteristics. Student's t-test or analysis of variance (ANOVA) was performed for categorical variables, and correlation analysis was performed for continuous variables. Significant variables in the univariate analysis (P>0.1) and clinically meaningful variables were included in a multivariate linear regression. Statistical significance was considered when the two-tailed P value less than 5%. All statistical analyses were performed using the SAS version 9.4 (SAS Institute Inc., Cary, NC, USA).

A total of 799 patients with type 2 DM were enrolled in this study. The mean age was 60.17 years, and 49.56% of the participants were women. The mean duration of DM was 8.12 years and a total of 74.34% of patients had comorbidities. At baseline, the mean fasting glucose level was 145.78 mg/dL and glycosylated hemoglobin (HbA1c) was 7.64%. Mean of BP and laboratory figures were within normal level both at baseline and follow-up and not much differed in subgroups. After perceiving CV risks, 43.2% of physicians changed their behaviors, and 88.7% patients altered their own lifestyles (Table 1).

In global questions of ADDQoL, the score of present QoL (“My present quality of life”) and QoL without DM (“If I did not have diabetes”) between baseline and after 6 months were not significantly different, respectively; score of present QoL was 0.49±0.92 at baseline vs. 0.47±0.86 after 6 months (P=0.613); score of QoL without DM were −1.48±0.98 at baseline and −1.47±0.91 after 6 months (P=0.6723).

The impact ratings at baseline for the 19 domains for K-ADDQoL ranged from −1.76 to −0.69, which were changed to −1.89 and −0.70 after 6 months. DM had the greatest impact on the freedom to eat domain, with scores significantly rising from −1.76±0.93 at baseline to −1.89±0.90 at 6 months (P=0.003). The domain with the least impact was people's reaction (−0.69±0.88 at baseline and −0.70±0.88 at 6 months). The impacts on holidays (P=0.038) and family life (P=0.004) significantly decreased at 6 months from the baseline scores. The impact score of friendship and social life domain decreased at 6 months from the baseline scores in statistically boarder-line (P=0.0550) (Fig. 1A).

The importance ratings at baseline for the 19 domains ranged from 1.54 to 2.27 and 1.55 to 2.19 at 6 months. Family life was perceived as the most important (2.27±0.66, 2.19±0.68), while local or long-distance journey were perceived as relatively less important (1.54±0.81, 1.55±0.75) both at baseline and 6 months. The importance of physical health (P=0.022) and family life (P=0.001) decreased after 6 months from the baseline (Fig. 1B).

WI at baseline ranged from −4.07 to −1.42 and from −4.33 to −1.43 after 6 months. DM had the highest negative impact on freedom to eat (−4.07±2.87 at baseline and −4.33±2.83 at 6 months), and the negative impact significantly increased after 6 months compared to that at baseline (P=0.049). DM had relatively less negative impact on people's reaction (−1.42±2.11 at baseline and −1.43±2.11 at 6 months) compared with other domains. WI on the family life (P=0.001) and friendship and social life (P=0.027) domains changed significantly less negatively over 6 months. AWI score was −2.51±1.36 at baseline and −2.50±1.29 after 6 months. AWI changed insignificantly over 6 months (0.02±1.43) (Fig. 1C).

We compared changes of K-ADDQoL scores in relation to changes of physicians' treatment pattern over 6 months. The impact scores for local or long-distance journey (P=0.001), holidays (P=0.008), physical health (P=0.008) and friendship and social life (P=0.040) were significantly different between patients whose physician's behaviors remained the same and those whose physician's behaviors changed. Compared to the former group, the latter group had less negative impact on these domains after 6 months (Fig. 2A). The importance rating for local or long-distance journey (P=0.042) also significantly differed between the two groups, and the group of patients whose physician behavior changed perceived the domain to be less important after 6 months while perceiving their feelings about the future (P=0.0671) to be more important after 6 months (Fig. 2B). The positive changes of WI score in the most domains were bigger in the group of patients whose physician behavior changed than the other group. In particular, the WI score for the local or long-distance journey (P=0.005), physical health (P=0.045), and holiday (P=0.036) domains significantly differed between the two groups (Fig. 2C). However, changes in several domains such as freedom to eat, freedom to drink, sex life, and feelings about the future showed opposite trends, in which the negative changes were bigger in the group whose physician changed treatments than the other group. Changes of AWI was not significantly differed between groups (Fig. 2C).

Fig. 3 shows the changes of K-ADDQoL scores in relation to patient behaviors for a 6-month period. The changes of the impact rating for sex life (P=0.028) significantly differed between the group of patients who changed their behaviors and those who did not, where the former group had less negative impact on sex life after 6 months (Fig. 3A). With regard to importance ratings, local or long-distance journey (P=0.002) and friendship and social life (P=0.015) domains significantly differed between the two groups, wherein the group whose patients' behaviors had changed perceived these domains to be less important, whereas the other group (no change of patient behaviors) perceived these domains to be more important after 6 months (Fig. 3B). Significant differences were not observed in the changes in WI in any of the domains over 6 months (Fig. 3C).

In univariate analysis, duration of DM (P=0.062), glucose (P= 0.022), HbA1c (P=0.009), insulin therapy (P=0.032), and changes in patient behaviors (P=0.012) showed a statistical significance. Patients who had their treatment altered by physicians showed less negative AWI score compared to those who have no treatment changes although there was no statistical significance. Among the factors indicating a statistical significance in the univariate analysis, we excluded glucose from the multivariate regression model in consideration of its correlation with HbA1c. Table 2 shows the results of multivariate analysis after adjusting for age and gender. The higher HbA1c level (P=0.012) and the longer duration of DM (P=0.044) were associated with the lower AWI. Patients who changed their behaviors showed negative AWI score compared to those who didn't change their behaviors although there was no statistical significance (P=0.085) (Table 2).