Assessment of Health-Related Quality of Life (HRQoL) by EQ-5D in Prediabetes, Diabetes and Normal Glycaemic Population in China


 Aim of the study was to describe and compare differences of health-related quality of life (HRQoL) among populations of normal glycaemic, pre-diabetics and diabetics in China.Methods A quality of life survey based on the EQ-5D-5L scale was conducted through face-to-face interviews and telephone follow-up. A total of 403 respondents with diabetes, 404 respondents with pre-diabetes, and 398 respondents with normal blood glucose were enrolled in the survey. Differences among groups were compared by chi-square test one-way analysis of variance (ANOVA), Kruskal-Wallis test or student's test after adjusting parameters of age and gender by PSM and Covariance analysis.Results In this survey, most of the diabetics were associated with circulatory system or nutritional metabolic diseases. The EQ-5D index of normal glycaemic population, pre-diabetic, and diabetic patients was 0.901, 0.948, and 0.897. And EQ-VAS scores of every groups above were 73.76, 77.45, and 68.34. HRQoL of male was higher than female in three groups of the study. In general, blood sugar control of Chinese diabetics was generally well. Results of covariance analysis and PSM were consistent with above results.Conclusion There was a general trend that patients was associated with a decline of HRQoL from pre-diabetic population, normal glycaemic population to diabetics. However, further and larger longitudinal studies are needed to confirm these findings.

from the out-patients who receiving treatment in several tertiary hospitals.
Prediabetes population were the people who had physical examination in several tertiary hospitals and conformed inclusion criteria. Normal glycemia people were randomly selected from parks and communities in Sichuan province who met the inclusion criteria.

Inclusion and Exclusion Criteria
Eligible diabetic population should fulfill following criteria: the people participated voluntarily; over 18 years of age; and could be diagnosed by the physician or by inquiry that they met the ADA diabetes diagnostic criteria [1] : Hemoglobin A1C(HbA1C)≥6.5% or Fasting Plasma Glucose (FPG)≥126 mg/dL (7.0mmol/L) or 2-h plasma glucose(PG)≥200mg/dL(11.1mmol/L) during an oral glucose tolerance test (OGTT); or a patient with classic symptoms of hyperglycemia or hyperglycemic crisis, a random plasma glucose≥200 mg/dL(11.1mmol/L). We excluded people who had disturbance of consciousness and response (People with hearing loss or tinnitus were also included in the study if they could fill out the questionnaire) or been diagnosed with the gestational diabetes.
Eligible people with pre-diabetes should fulfill following criteria: the people participated voluntarily; over 18 years of age; met the diagnostic criteria for prediabetes published by ADA: FPG between 100mg/dL (5.6mmol/L) to 125mg/dL (6.9mmol/L) or 2-h PG during 75-g OGTT between 140 mg/dL (7.8 mmol/L) to 199 mg/dL (11.0 mmol/L) (IGT) or HbA1C between 5.7% and 6.4%. We excluded people who had disturbance of consciousness and response (People with hearing loss or tinnitus were also included in the study if they could fill out the questionnaire) or been diagnosed with diabetes.
Eligible people with normal glycaemic should fulfill following criteria: the people participated voluntarily; over 18 years of age; normal glycaemia or self-reported without diabetes. We excluded people who had disturbance of consciousness and response (People with hearing loss or tinnitus were also included in the study if they could fill out the questionnaire) or been diagnosed with diabetes and pre-diabetes.

Sample size and sampling
According to sample size formula for simple random sampling(SRS) n= enough sample size shall be no less than 384. Considering 10% ineffective response rate, we should interview 422 people for each types of blood glucose filling out questionnaires. In this study, we collected 403 effective questionnaires from normal glycaemic people, 404 from prediabetic people and 398 from diabetic people, after removing ineffective questionnaires because of missing main information.

Questionnaire Design
Questionnaire was tentatively designed based on literatures as well as inviting suggestions from clinical pharmacists and endocrinologists. The questionnaire consisted of six parts including demographic data, disease situation, economic situation, health insurance, living habits and EQ-5D-5L scale. A pilot survey was conducted among 20 interviewees in April 2015, based on which we modified the questionnaire including 6 parts, 22 questions. EQ-5D-5L instrument was applied in this study after registering on www.euroqol.org. EQ-5D-5L scale consists of descriptive system and EQ Visual Analogue scale (EQ-VAS). The descriptive system including five dimensions: mobility, self-care, usual activities, pain/discomfort, anxiety/depression. Each dimension is divided into 5 levels: no problems, slight problems, moderate problems, severe problems, and extreme problems.

Conducting the Survey
Four trained investigators put the investigation in practice from June 2015 to March 2016. Informed consent was obtained from every respondent before surveyed.
Diabetics were interviewed through face-to-face confabulation method when they were waiting to see the doctor in the hospitals. Normal glycaemic people were surveyed face-to-face in parks or communities. For those people, blood sugar status was asked firstly to exclude respondents who did not know their blood glucose status or who had diabetes. Questionnaires were completed by the investigators according to the patient's dictation when interviewees were not convenient.
People with pre-diabetes were reached by hospital medical examination centers.
Upon approval of the pre-diabetics, surveys were carried on over the phone at the appointed time.

Statistical analysis
The data from questionnaires was recorded by EpiData version 3.0 (EpiData Association, Odense, Denmark) and processed by EXCEL (2013). Reliability of questionnaire was measured by Cronbach's alpha coefficient(α). The statistical package for social science (SPSS 20, Chicago, IL, USA) software was used for statistical analysis. The parametric test was executed as the data followed normal distribution. And non-parametric tests were selected as the data followed abnormal distribution or doesn't meet the requirements of parametric test. EQ-VAS scores and EQ-Index values were expressed as average±SD (x±s). The EQ-Index was calculated by Chinese EQ-5D-5L set. All statistical tests were bilateral. Testing significance of difference and correlation between variables were analyzed by chi-square test, ttest, one-way analysis of variance (ANOVA) or Kruskal-Wallis test. Analysis of covariance was used to adjust covariates, then quality of life of the three groups after correction were compared. Besides age, sex, BMI, and household income of the three groups were matched by PSM. Statistical analysis based on the data above.
Statistical significance was established as P < 0.05.

Socio-demographics
Cronbach's α of questionnaire about diabetics, pre-diabetics and normal glycaemic population were 0.721, 0.748 and 0.753 respectively and the content validity of was 0.673, 0.728 and 0.840 correspondingly, which showed good reliability and validity of questionnaire.
There were statistical differences in age, BMI, gender, alcohol consumption, education level, type of medical insurance, family income, and family history of diabetes for three groups of respondents from different populations. Sociodemographic features of respondents grouped by glucose status were showed in Table 1.

Blood Glucose status
We gathered blood glucose data of pre-diabetics and diabetics, which was displayed in  Table 3.

Quality of life measure
As shown in Table 4, the main health-related problem existing in the respondents with different blood glycaemia status were pain/discomfort, anxiety/depression and self-care which had lowest ranking. The proportion of health-related problems in the five dimensions was consistent among three groups. The sequence of respondents' self-assessments in health-related problem occurrence from high to low was pain/discomfort anxiety/depression mobility usual activities and self-care.
EQ-VAS score and EQ-Index of diabetics, pre-diabetics and normal glycaemic people were shown in Table 5. EQ-VAS scores from high to low were pre-diabetes, normal blood glucose, and diabetes. EQ-Index of the study were calculated by Chinese EQ-5D-5L value sets which was processed by dimension reduction method. Health utility value of diabetics was lower than normal glycaemic people and the pre-diabetes patients have higher health utility values than normal glycaemic people. Among the three blood glucose status groups, the quality of life of women was lower than that of the corresponding male population.
Due to the sample came from different population, we carried out covariance analysis and PSM separately so as to compare the difference in quality of life among three groups. We set age, gender, BMI, alcohol consumption, education situation, medical insurance, household income and family history of diabetes as covariates, then regarding type of different blood-sugar population as independent variable, EQ-5D-5L and EQ-VAS as dependent variable, conducted covariance analysis. When EQ-VAS scores was the independent variable, the pre-analysis showed that there was no significant interaction between the type and the covariates.  [19][20] . The EQ-Index of this study was consistent with Zhou's [21] research which showed EQ-Index value of Chinese diabetic patients ranking from 0.79 to 0.94, which means result of the study was still in the middle range. Compared with the EQ-Index of Spanish [19] and German [3] diabetes whose score was 0.742 and 0.80 respectively, EQ-Index of diabetes in the study was slightly higher. Prevalence of diabetes or prediabetes of male was higher than female in the study, but the differences was not statistically significant. The result of the study was consistent with results of some other studies in which prevalence of diabetes in female is lower than male in Chinese [9,23,24] . In the other words, men should pay more attention to their glycaemia status as well as life-style.
According to WHO statistics, there are hundreds of complications in patients with diabetes, such as diarrhea, periodontitis, coronary heart disease, heart rate irregularity, nephropathy, frequency of urination, sexual dysfunction, arteriosclerosis, numbness of hands and feet, constipation, dry skin, glaucoma, cataract, ulcer, cerebral ischemia, retinopathy, gangrene, etc. Literature shows 30% of chronic renal failure, 40% to 50% of blindness, 50% of cardiovascular and cerebrovascular diseases, and 60% of amputations are caused by diabetes [25] . Our study showed half of diabetics suffered complications, in which 32.25% diabetics had one kind of complication,19.50% diabetics had two kinds of complications, about 11.00% diabetics had more than three kinds of complications. Among all complications, macroangiopathy and microangiopathy were the main causes of disability or death in elderly diabetics. In this study, atherosclerosis accounted for 0.75% of diabetic patients with macrovascular disease, and incidence of retinopathy, neuropathy and anemia was 9.25%, 4.75% and 0.5% respectively.
Therefore, diabetic patients should pay more attention to prevent macroangiopathy and microangiopathy, especially atherosclerosis, retinopathy, neuropathy and anemia.
In the survey, 52% patients had dietary and exercise controlling, but only 14% patients controlled continuously. Most patients misunderstood the way of controlling diet and exercise. For diet controlling, most patients only avoid oily food, sugar and sweets. They were lacked of awareness in calculating calories in foods and practicing separate eating. Therefor patient education of how to get a good dietary and exercise controlling is necessary.
From the survey, we observed that the quality of life in prediabetics was higher than that in normal glycemic people and diabetics. This results were different from a previous German survey showing that the quality of life gradually decreased as glycemic status deteriorated [3] . This may be due to some design defects in this study. The samples of the study were collected from three different populations.
The samples of pre-diabetic group mainly came from people who went to the physical examination centers of tertiary hospitals and had more awareness and capabilities for self-health protection. At the same time, most of them had better jobs and accessed to health insurance easier. In the aspects of family income and education background, they had more advantageous than the diabetics and the normal glycaemic population, who were easier to maintain good physiological and mental state. However, after correcting each parameter of gender, age, BMI, alcohol consumption, family history of diabetes, and type of medical insurance by using covariance analysis, results still showed that the quality of life of pre-diabetic group was better than normal glycaemic population and diabetics. This is probably the result of that asymptomatic pre-diabetic patients have lower rates of anxiety or depression (30.30%) than other populations.
There were other limitations in this study, such as recall bias of the blood sugar indicator, diet, exercise, sleep and illness acquired by questionnaire. Furthermore, the study was a cross-sectional study that does not directly determine the causal relationship between changes in quality of life and the dynamic evolution of diabetes. And we did not discuss EQ-VAS score of diabetes subgroups due to the large differences of sample size in them.

Conclusion
There was a general trend that patients were associated with a decline of HRQoL from pre-diabetic population, normal glycaemic population to diabetics. However, further and larger longitudinal studies are needed to confirm these findings.