Differences between the perspectives of physicians and patients on the potential barriers to optimal diabetes control in China: a multicenter study

This multi-center cross-sectional survey was conducted from December 2015 to March 2016. A multi-stage stratified random sampling method was used to sample representative diabetes physicians and patients in Shaanxi province, China for this survey (Fig. 1).

Step 1 involved sampling physicians and patients in the cities. In addition to the capital city (Xi’an), physicians and patients from an economically developed prefecture-level city (Yulin, annual Gross National Products GDP > 50,000 yuan per capita), a moderately developed prefecture-level city (Baoji, annual GDP 30,000–50,000 yuan per capita), and a developing prefecture-level city (Ankang, annual GDP < 30,000 yuan per capita) were sampled; the GDP was classified according to the 2015 GDP in Shaanxi Province. This step was non-randomized. Step 2 involved stratifying the hospitals in the capital city and the three prefecture-level cities. The hospitals in the capital city were stratified into medical university hospitals, tertiary municipal hospitals, and secondary county hospitals; the hospitals in the prefecture-level cities were stratified into tertiary municipal hospitals and secondary county hospitals. Step 3 involved sampling the hospitals. Given the differences in the city sizes and the distribution of medical resources between the capital city and the prefecture-level cities, we used a simple random sampling (drawing) method to randomly sample three medical university hospitals, three tertiary municipal hospitals, and three secondary county hospitals in the capital city in addition to one tertiary municipal hospital and two secondary county hospitals in each of the prefecture-level cities. Thus, we sampled a total of 18 hospitals, including three medical university hospitals, six tertiary municipal hospitals, and nine secondary county hospitals. Step 4 involved sampling the physicians. According to the distribution of physicians in the different levels of the hospitals, we randomly sampled six residents, three attending physicians, and one chief physician from each medical university hospital; three residents, two attending physicians, and one chief physician from each tertiary municipal hospital; and two residents and one attending or chief physician from each secondary county hospital. Thus, we sampled a total of 93 physicians, including 54 residents, 21 to 31 attending physicians, and nine to 19 chief physicians. Step 5 involved sampling the patients. We used the cluster random sampling method and enrolled all of the patients with type 2 diabetes (based on World Health Organization 1999 criteria) who were hospitalized for diabetes treatment in department of endocrine during a certain time window and who met all of the study criteria, had poor diabetes control (HbA1c ≥ 7%), and were willing to participate in this study, until the planned number of participants was reached. Given the difference in ward capacity between the different levels of hospitals, we planned to sample 40 patients in each medical university hospital and tertiary municipal hospital and 30 patients in each secondary county hospital, for a total of 630 patients, which provided a sufficient sample according to previous literatures [11‐23].

We developed our own questionnaire for data collection. While designing the questionnaire, we referred all the barriers and strategies from previous related literatures, and expanded some potential ones localized for the Chinese, which were from a small-scale pre-survey [11‐23]. Two versions of the survey were used: the physician version and the patient version. (Additional file 1).

The physician questionnaire had two parts. Part 1 was used to collect basic information on the physician, including title and the type and location of the hospital. Part 2 contained two questions. Question 1 asked “According to your own clinical experience, what do you think are the top three potential barriers to optimal diabetes control?” There were three items could be chosen from a list of eight items. The physicians were asked to rank the factors in the order of importance. Question 2 asked “From the perspective of the government/community/hospital/physician, what do you think are the top three areas that require urgent improvement in order to improve diabetes control?” There were three items could be chosen from a list of nine items. Again, the physicians were asked to rank the areas in the order of importance.

The patient questionnaire contained two parts. Part 1 was used to collect patient demographic data, lifestyle, and history of diabetes. Part 2 contained only one question: “What do you (the patient) think are the causes of your uncontrolled blood glucose?” This was a multi-choice question and the patients could choose more than one factors for uncontrolled glucose. The patients were asked to mark the most important item. See the annex for the questionnaire design and baseline information.

EpiData software (version 2.3) was used to enter the data obtained into the database, and the Statistical Package for Social Sciences 20.0 (SPSS Inc., Chicago, IL, USA) was used for the descriptive statistics. The measurement data were expressed as the numerical mean and standard deviation. Count data are expressed as percentages (%).

A total of 93 physicians were sampled and asked to participate, of whom 85 agreed to participate and complete the physician questionnaire (response rate: 91.4%), including 42 (49.4%) residents, 31 (36.5%) attending physicians, and 12 (14.1%) chief physicians; 30 (35.3%) physicians were from medical university hospitals, 35 (41.2%) were from tertiary municipal (non-teaching) hospitals, and 20 (23.5%) were from secondary county hospitals; 45 (52.9%) physicians were from the capital city, 24 (28.2%) were from prefecture-level cities, and 16 (18.8%) were from counties or lower level cities (Table 1).

For question 1 (Figs. 2), “According to your own clinical experience, what do you think are the top three potential barriers to optimal diabetes control?”, 62 (62/85 72.9%) physicians considered patient’s “insufficient understanding of the danger of diabetes” as an important factor, of whom 33 (33/85 38.8%) physicians ranked it as the most important factor, 23 (23/85 27.1%) physicians ranked it as the second most important factor, and 6 (6/85 7.1%) physicians ranked it as the third most important factor.

In addition, 56 (56/85 65.9%) physicians considered patient’s “lack of perseverance to stick to lifestyle intervention (if even understanding the danger of diabetes)” as an important factor, of whom 29 (29/85 34.1%) physicians ranked it as the most important factor, 21 (21/85 24.7%) physicians ranked it as the second most important factor, and 6 (6/85 7.1%) physicians ranked it as the third most important factor.

Moreover, the physicians considered patient’s failure to monitor blood glucose regularly, poor medication compliance, economic reasons, and the uneducated use of herbal medicine or health products as moderately effective potential barriers to optimal diabetes control. Whereas physicians considered factors that failure to attend scheduled follow-ups or inconvenience to purchase drugs as lower effective factors to diabetes control than other factors.

For question 2 (Figs. 3), “From the perspective of the government/community/hospital/physician, what do you think are the top three areas that require urgent improvement in order to improve diabetes control?”, 38 (38/85 44.7%) physicians considered “improving health insurance coverage for diabetes” as an important measure, of whom 20 (20/85 23.5%) physicians ranked it as the most important factor, 12 (12/85 14.1%) physicians ranked it as the second most important factor, and 6 (6/85 7.1%) physicians ranked it as the third most important factor.

In addition, 33 (33/85 38.8%) physicians considered “providing more and easy-to-use diabetes brochures or educational materials for patients” as an important factor, of whom 14 (14/85 16.5%) physicians ranked it as the most important factor, 12 (12/85 14.1%) physicians ranked it as the second most important factor, and 7 (7/85 8.2%) physicians ranked it as the third most important factor.

Moreover, although the physicians considered regulating media campaigns and minimizing false advertising, in addition to engaging or encouraging family members to become involved in the care of diabetic patients as important measures for strengthening public health management of diabetes, few ranked these factors as the most important factor. Furthermore, the physicians did not consider factors such as strengthening the training of diabetes specialists, training more specialist diabetes nurses to provide guidance for patients, and improving multidisciplinary and multi-specialty collaboration for diabetes control as urgent public health measures for improving diabetes control.

A total of 630 patients were sampled and asked to participate, of whom 584 patients agreed to participate and completed the patient questionnaire (response rate: 92.70%). Of the patients who completed the questionnaire, 325 patients were men with a mean age of 56.27 years (± 13.19), and 254 patients were women with a mean age of 58.06 years (± 12.79). Moreover, 242 (44.2%) patients had been diagnosed with diabetes for less than 5 years, 175 (32.0%) patients had been diagnosed with diabetes for 5 to 10 years, and 130 (23.8%) patients had been diagnosed with diabetes for ≥10 years (Table 2).

The data analysis of the 584 valid patient questionnaires showed that for the question, “What do you (the patient) think are the potential barriers to optimal diabetes control?” (Figs. 4), 338 (338/584 57.9%) patients considered “I do not follow a proper diet” as an important barrier, of whom 142 (142/584 24.3%) patients ranked it as the most important reason and 196 (196/584 33.6%) patients ranked it as the second most important barrier. A total of 263 (263/584 45.0%) patients considered “I do not exercise as instructed” as an important reason, of whom 55 (55/584 9.4%) patients ranked it as the most important barrier and 208 (208/584 35.6%) patients ranked it as the second most important barrier.

Moreover, the patients considered failure to monitor blood glucose as instructed as an important potential barrier to optimal diabetes control. However, the patients did not consider high drug cost, inconvenience to purchase drugs, the use of herbal medicine, and incorrect treatment protocol as potential barriers to optimal diabetes control.