Knowledge, attitude and practice regarding screening and managing diabetic microvascular complications among general practitioners of community health centres: a cross-sectional study in Shanghai, China

Abstract

Objective

To evaluate the knowledge, attitude and practice (KAP) regarding screening and managing diabetic microvascular complications, encompassing diabetic retinopathy (DR), diabetic kidney disease (DKD) and diabetic neuropathy (DN), among general practitioners (GPs).

Design

Cross-sectional study.

Setting

The online questionnaire survey was conducted between April and July 2023.

Participants

GPs from community health centres (CHCs) in all 16 districts of Shanghai were recruited.

Primary and secondary outcome measures

The data of sociodemographic characteristics, KAP scales, training experience and screening instruments for community screening and managing diabetic microvascular complications were collected. Multiple stepwise linear regression was used to explore the influencing factors of KAP. Restricted cubic spline curves with four knots (5%, 35%, 65%, 95%) were used to determine the association between KAP score and duration of general practice.

Results

A total of 1243 questionnaires were included in the analysis. The total KAP score was 66.6±8.8/100, and the knowledge, attitude and practice scores were 64.7±8.7, 83.5±10.5 and 51.6+17.8, respectively. Male (β=−2.419, p=0.012), shorter practice duration (β=−1.033, p=0.031), practice in rural area (β=3.230, p=0.001), not attending training in diabetic microvascular complications (β=−6.346, p<0.001), not managing diabetic patients (β=−4.503, p<0.001), less number of diabetes patients under management (β=−0.007, p=0.035), less number of screening instruments based on self-report of GP (β=−1.681, p<0.001), lower knowledge score (β=−0.190, p<0.001) and lower attitude score (β=−0.414, p<0.001) were associated with lower practice score of GPs. The KAP total score increased with the working years of general practice; however, this effect was no longer observed in knowledge score after 15 years, while the attitude and practice scores showed a continuously increasing trend.

Conclusions

GPs showed insufficient knowledge and poor clinical practice on screening and managing diabetic microvascular complications. There is an urgent need to improve their capacity to provide better care for those with diabetic microvascular complications through targeted training.

Strengths and limitations of this study.

• A large number of participants were included for cross-sectional analysis; they covered all 16 districts in Shanghai, China, accounting for 12% of the city’s total number of general practitioners (GPs).

• Restricted cubic spline curves with four knots (5%, 35%, 65%, 95%) were used to determine the association between knowledge, attitude and practice (KAP) score and duration of general practice.

• A potential methodological limitation of our study was the selection bias that may arise from the use of convenience sampling.

• There may have been some information bias on the instruments of community health centres (CHCs) due to self-reported data.

• This study was cross-sectional, meaning only associations and not causation could be assessed.

Introduction

Due to epidemiological changes, such as shifts in nutrition patterns, urbanisation and sedentary lifestyles, the global incidence of diabetes mellitus (DM) has been steadily increasing.¹ As the prevalence of DM increases, so does the occurrence of diabetic microvascular complications, encompassing diabetic retinopathy (DR), diabetic kidney disease (DKD) and diabetic neuropathy (DN), which are common complications resulting from chronic hyperglycaemia damaging capillaries.^{2 3} This condition significantly impacts the quality of life of diabetic patients, leading to disability and imposing a substantial financial expenditure on individuals and the healthcare system, thus, creating a significant social burden.^{4 5}

Early screening and managing diabetic microvascular complications are crucial to prevent the progression of diabetes and improve patient outcomes, such as preventing vision loss, end-stage renal disease and amputation.^{6 7} Lin et al showed that screening for DR is cost-saving and 90% effective in preventing vision loss, emphasising the importance of early detection and treatment.⁸ Previous research highlighted the importance of regular screening, early diagnosis and treatment of DKD, the leading cause of renal disease in dialysis patients.^{9 10} In their review, Sorber et al reported that early detection and regular surveillance can potentially improve the outcomes of diabetic foot (DF) disease.¹¹ However, diagnosing diabetic microvascular complications poses a significant challenge as patients often do not exhibit symptoms until the disease has reached an advanced stage, resulting in underdiagnosis and suboptimal screening practices.¹² For example, about 30% of adults with severe CKD did not know they had CKD.¹³ In a similar study, Shaki et al noted that about 59.4% of people with diabetes never received a foot examination since their initial diagnosis.¹⁴ In their review, Egunsola et al stated that a substantial proportion of patients with diabetes remain unscreened.¹⁵ For example, 40% of people with diabetes in the USA were not screened annually for DR, and 38% of patients in Canada were never screened for it.

It has been suggested that screening for diabetic microvascular complications should be carried out by community general practitioners (GPs).^{11 16} Tan et al found that the initial diagnosis of DF made by a primary care physician rather than a specialist reduces the risk of amputation.¹⁷ GPs are often the initial point of contact for patients seeking affordable healthcare services and are considered the backbone of community health centres (CHCs). Since 2009, grassroots-level diabetes prevention and management has been promoted and implemented nationwide as part of China’s Basic National Public Health Service Project.¹⁸ In 2015, diabetes was selected as one of the first pilot diseases for hierarchical diagnosis and treatment, with the initiative promoting initial diagnosis, basic treatment, and ongoing prevention and management of diabetes at the grassroots level through the family doctor contracting system.¹⁹ As this project has been implemented and expanded, an increasing number of diabetic patients are opting to seek care at CHCs. GPs play a crucial role as ‘gatekeepers of health,’ responsible for the prevention, screening, diagnosis, treatment and long-term monitoring of microvascular complications associated with diabetes.

The knowledge, attitude and practice (KAP) model is a crucial theory used to explore the underlying mechanisms of healthcare practice behaviour.²⁰ According to this model, the practices of medical professionals are influenced by their knowledge and attitude. It has been demonstrated that attitudes have an impact on observable behaviour, and GPs’ attitudes constitute the basis for achieving better healthcare.²¹ Therefore, GPs must possess sufficient knowledge, attitude and practice to effectively fulfil their role in community management. Given the scarcity of studies examining the KAP of GPs on screening and managing diabetic microvascular complications, it is imperative to assess their present KAP in order to develop comprehensive strategies for the future management of diabetic microvascular complications in the communities. This study aimed to assess the KAP regarding screening and managing diabetic microvascular complications among GPs in CHCs.

Research design and methods

Study participants and data collection

The cross-sectional study was conducted between April and July 2023, using an online questionnaire (Online link: https://www.wjx.cn/vm/rDZ9YYo.aspx#) to collect data. Before the formal investigation, a pilot test was conducted to estimate the time required to complete the questionnaire. In the initial section of the questionnaire, we collected respondents’ full names and names of their respective CHCs. Furthermore, the online survey enabled us to collect the Internet Protocol (IP) addresses of the respondents. The respondents’ names and IP addresses served as unique identifiers. In cases where multiple responses originated from the same IP address and bore the same name, only the first submission was considered valid. To prevent respondents from seeking answers or submitting random responses during the survey, we implemented a time limit for each section of the questionnaire. Once a respondent completed a section, they could not return to previous sections to make changes. The convenience sampling method was employed to recruit study participants. Online questionnaires were distributed in the following two ways: (1) at the annual meeting of the Cross-Straits Medicine Exchange Association-Committee of General Practice (CSMEA), a grand annual academic meeting of general practice that was held on April 2023, when we sent the Quick Response (QR) code of the questionnaire to GPs from Shanghai; (2) we requested GPs attending the meeting mentioned above in Shanghai to forward the QR code of the questionnaire to their colleagues. The inclusion criteria were the following: (1) registered GPs who worked in CHCs of Shanghai, China; (2) consented to participate in the survey. All available eligible samples were included. The survey was conducted via the online platform WJX (Changsha Ranxing Information Technology Limited Company).

This study was approved by the Ethics Committee at the Xiamen Branch of Zhongshan Hospital, Fudan University (B2022-013). A consent statement explaining the purpose and procedure of the study was attached to the online questionnaire for the participants to read before deciding whether to proceed with the survey. The online questionnaire could be withdrawn at any time and without giving any reason.

Survey instruments

The online questionnaire used in this study was divided into four sections, that is, sociodemographic characteristics, scale of KAP, training experience, and instruments for screening and managing diabetic microvascular complications. KAP was adapted from a previous study validated for the Chinese population (online supplemental file).²² The KAP scale’s reliability and validity were tested. Confirmatory factor analysis indicated that the scale had a χ² degree of freedom ratio (χ²/df) of 2.729, a Comparative Fit Index (CFI) of 0.784, and a root mean square error of approximation (RMSEA) of 0.072, suggesting an acceptable model fit. In terms of discriminant validity, the differences in total and subscale scores between high-score and low-score groups were statistically significant (t values ranging from 10.82 to 44.49, all p value <0.001). In terms of reliability, the Cronbach’s α coefficient of the total scale was 0.891; the intraclass correlation coefficient for test-retest reliability was 0.76.²³

The knowledge section contained 24 items across 6 dimensions related to the knowledge of general knowledge of diabetic microvascular complications, DR, DKD, DN, DF and medication use; this section required 10 min to complete. Each item of the scale was scored by the GPs as true, false or do not know, and 1 point was allocated for a correct response and 0 points for an incorrect or ‘do not know’ response. The total score was converted into a hundred percentage point from 0 to 100. The higher scores indicated a better knowledge of participants.

The attitude section contained 12 items, each scored by the GPs with 1–5 points indicating strongly disagree, disagree, neutral, agree and strongly agree. The total score was converted into a hundred percentage point from 0 to 100. The higher scores indicated a better attitude of participants.

The practice section contained 14 items across 3 dimensions related to screening suggestions, screening action and patient education. Four of the 14 items were open questions and were not scored. The remaining 10 items of the scale were scored by the GPs with 1–5 points indicating none, minority of diabetic patients, half of diabetic patients, majority of diabetic patients and all diabetic patients. The total score was converted into a hundred percentage point from 0 to 100. The higher scores indicated better practices of participants.

The GPs’ training experience in diabetic microvascular complications contained three items. The GPs’ self-reported information about screening instruments for diabetic microvascular complications in their CHCs contained four items.

Statistical analysis

Data were coded and recorded on a spreadsheet. SPSS software version 25 was used for statistical analysis (SPSS Inc., IL, USA). Descriptive statistics were reported as proportions (%) for categorical variables and mean±SD for continuous variables. The respondents’ KAP and sociodemographic characteristics were compared using a T-test or analysis of variance. The association between KAP and demographic characteristics, training experience and screening instruments was assessed using multiple stepwise linear regression analysis. Restricted cubic spline curves with four knots (5%, 35%, 65%, 95%) were used to determine the association between KAP score and duration of general practice.²⁴ Restricted cubic spline curves were constructed and mapped using R software (version 4.3.0). A two-sided p<0.05 was considered statistically significant.

Patient and public involvement

Patients or public were not involved in the design, conduct, reporting or dissemination plans of our research.

Results

Among a total of 1477 participants who completed the questionnaire, 94 were not GPs, 74 did not work in CHCs of Shanghai and 66 did not complete the knowledge part within 10 min, so they were excluded. The remaining 1243 GPs were included in the final analysis based on the inclusion criteria; they covered all 16 districts in Shanghai, accounting for 12% of the city’s total number of GPs.²⁵

Socio-demographic characteristics

Most GPs were women (70.8%), with a mean age of 39.9±7.7 years. The overall mean duration of practice was 16.2±8.9 years, while the duration of practice as a GP was 10.4±6.1. Only 2.4% of the respondents did not have a bachelor’s degree. Most respondents (84.8%) held at least a middle professional title. Regarding income, 74% reported earning between 100 000 and 200 000 Chinese Yuan annually. Half of the respondents received GP Structured Vocational Training. Furthermore, 72.6% were contracted family doctors, with an average of 1603.8±868.6 community residents per doctor. Among the contracted family doctors, 66.7% managed patients with diabetes, with an average of 197.1±159.2 diabetic patients per doctor. The characteristics of the GPs involved in the study are presented in table 1.

Table 1. Sociodemographic characteristics with the knowledge, attitude and practice scores of general practitioner (GPs).

Variables (n=1243)	N (%)	Knowledge score		Attitude score		Practice score
		Mean±SD	P value	Mean±SD	P value	Mean±SD	P value
Gender, n (%)			0.109		0.633		0.062
Male	363 (29.2)	64.8±8.3		81.3±9.5		50.1±17.0
Female	880 (70.8)	65.6±8.6		81.1±9.5		52.2±18.1
Age, n (%)			0.018		0.859		0.002
<30	110 (8.8)	64.2±10.3		80.8±10.0		48.5±17.4
30–39	499 (40.1)	65.8±8.4		81.0±9.6		49.7±17.4
40–49	487 (39.2)	65.7±8.2		81.2±9.3		53.9±18.2
50–59	140 (11.3)	63.9±8.0		81.4±9.6		52.2±17.2
≥60	7 (0.6)	59.4±6.3		84.6±11.4		53.7±19.9
District			0.007		0.237		<0.001
Urban	403 (32.4)	66.3±8.3		81.6±9.2		54.7±18.3
Suban	840 (67.6)	65.0±8.6		80.9±9.6		50.1±17.4
Duration of practice (years), n (%)			0.475		0.015		0.002
<10	328 (26.4)	65.3±9.2		81.4±9.4		49.7±17.3
10–19	453 (36.4)	65.7±8.6		80.4±9.8		50.5±18.2
20–29	359 (28.9)	65.4±8.2		81.1±9.1		54.0±17.4
≥30	103 (8.3)	64.2±6.7		83.7±9.6		54.2±18.2
Educational level, n (%)			<0.001		0.649		0.772
Below bachelor degree	30 (2.4)	58.9±10.9		81.7±12.4		49.7±20.7
Bachelor degree	1002 (80.6)	65.3±8.2		81.2±9.3		51.5±17.4
Master’s degree	211 (17.0)	66.7±9.3		80.6±10.0		52.1±19.2
Professional title, n (%)			<0.001		0.006		<0.001
Resident	189 (15.2)	63.5±10.5		80.7±10.2		48.9±18.0
Attending physician	711 (57.2)	65.3±7.9		80.9±9.5		50.7±17.6
Associate senior physician	305 (24.5)	66.4±8.3		81.4±8.9		53.8±17.0
Senior physician	38 (3.1)	67.3±8.4		86.4±9.9		63.7±20.3
Annual income (thousand yuan)*, n (%)			0.024		0.001		0.542
<100	116 (9.5)	63.6±8.7		81.2±9.6		52.7±19.2
100–149	430 (35.4)	64.8±8.2		80.6±9.6		51.4±18.1
150–199	469 (38.6)	65.7±8.4		81.1±9.2		50.8±17.0
200–299	182 (15.0)	66.7±9.1		82.3±9.7		52.9±17.3
≥300	19 (1.6)	66.5±6.8		88.4±7.6		56.8±22.7
Duration of general practice (years), n (%)			0.372		0.018		<0.001
<5	211 (17.0)	65.0±9.0		81.1±10.3		48.1±18.9
5–9	349 (28.1)	65.2±8.5		80.7±9.3		50.3±17.4
10–14	360 (29.0)	65.4±8.7		80.4±9.6		51.5±16.9
15–20	204 (16.4)	66.4±7.9		81.9±8.2		55.5±18.1
≥20	119 (9.6)	64.7±7.7		83.5±9.8		55.0±17.6
Received GP Structured Vocational Training, n (%)			0.123		0.036		0.854
Yes	622 (50.0)	65.8±9.1		81.7±9.5		51.5±17.9
No	621 (50.0)	65.0±8.0		80.6±9.5		51.7±17.7
Contracted family doctor, n (%)			<0.001		0.214		<0.001
Yes	903 (72.6)	66.0±8.2		81.4±9.3		53.1±16.5
No	340 (27.4)	63.6±9.0		80.6±10.1		47.5±20.3
Managing diabetic patients, n (%)			<0.001		0.008		<0.001
Yes	829 (66.7)	66.4±8.2		81.7±9.1		54.5±16.2
No	414 (33.3)	63.3±8.7		80.1±10.2		45.8±19.4
Training experience on diabetic microvascular complications
Yes	845 (68.0%)	65.6±8.0	0.212	82.77±9.2	<0.001	55.1±17.2	<0.001
No	398 (32%)	64.9±9.4		77.7±9.2		44.1±16.7
Number of screening instrument, n (%)			<0.001		<0.001		<0.001
0	13 (1.0%)	60.0±7.3		71.0±11.3		33.2±15.6
1	6 (0.5%)	64.7±3.9		74.4±12.1		31.7±6.0
2	60 (4.8%)	63.7±7.7		76.6±8.1		40.7±12.2
3	164 (13.2%)	62.5±8.1		78.0±10.3		44.3±15.5
4	164 (13.2%)	66.2±9.5		79.8±9.3		48.5±16.1
5	252 (20.3%)	65.1±8.4		81.5±9.1		52.1±16.9
6	232 (18.7%)	65.4±7.9		82.2±8.5		54.7±17.0
7	193 (15.5%)	67.9±8.3		82.9±8.6		55.7±18.6
8	159 (12.8%)	66.1±8.6		84.2±9.9		58.2±19.3

The bold data indicate values where p<0.05, which represent statistically significant results.

^*Number of respondents included in analysis.

KAP score and factors associated with knowledge, attitude, and practice

The total score of KAP was 66.6±8.8/100, and the knowledge, attitude and practice scores were 64.7±8.7, 83.5±10.5 and 51.6+17.8, respectively (table 2). The multiple stepwise linear regression models confirmed that younger age (β=−1.138, p=0.002), higher educational level (β=1.568, p=0.010), higher professional title (β=1.519, p<0.001), managing diabetic patients (β=2.695, p<0.001), more screening instruments based on self-report of GP (β=0.440, p=0.001) and higher practice score (β=0.048, p=0.001) were associated with better knowledge score of GPs. Lower educational level (β=−1.459, p=0.029), higher annual income (β=0.739, p=0.003), received GP Structured Vocational Training (β=2.049, p<0.001), attended training in diabetic microvascular complications (β=3.386, p<0.001), more screening instruments based on self-report of GP (β=0.649, p<0.001) and higher practice score (β=0.157, p<0.001) were associated with better attitude score of GPs. Female (β=2.419, p=0.012), longer practice duration (β=1.033, p=0.031), practice in urban area (β=−3.230, p=0.001), attended training in diabetic microvascular complications (β=6.346, p<0.001), managing diabetic patients (β=4.503, p<0.001), more diabetes patients under management (β=0.007, p=0.035), more screening instruments based on self-report of GP (β=1.681, p<0.001), higher knowledge score (β=0.190, p<0.001) and higher attitude score (β=0.414, p<0.001) were associated with better practice score of GPs. An overview of factors associated with KAP of GPs of diabetic microvascular complications is presented in table 3.

Table 2. Overview of the knowledge, attitude and practice scores of general practitioner (GPs) of diabetic microvascular complications.

Variables		Score (mean±SD)
Knowledge	General knowledge	14.6 (26.4)
	Diabetic retinopathy	52.8 (17.7)
	Diabetic kidney disease	77.9 (14.1)
	Diabetic neuropathy	65.6 (13.3)
	Diabetic foot	98.2 (10.1)
	Medication	69.1 (19.1)
	Total score	64.7 (8.7)
Attitude	Awareness of management	97.1 (7.5)
	Self-ability evaluation	75.6 (15.4)
	Attitude on participating training and research	85.5 (14.7)
	Total score	83.5 (10.5)
Practice	Screening suggestion	53.3 (19.2)
	Screening action	46.5 (18.8)
	Patient education	58.3 (21.8)
	Total score	51.6 (17.8)
Total		66.6 (8.8)

Table 3. Factors associated with knowledge, attitude and practice of general practitioner (GPs) of diabetic microvascular complications.

	Regression coefficients	95% CI		T value	P value
Knowledge
Age	−1.138	−1.860	−0.417	−3.098	0.002
Educational level	1.568	0.373	2.764	2.573	0.010
Professional title	1.519	0.696	2.343	3.619	<0.001
Managing diabetes patients or not	2.695	1.681	3.708	5.218	<0.001
Number of screening instrument	0.440	0.172	0.709	3.216	0.001
Practice score	0.048	0.020	0.076	3.368	0.001
Attitude
Educational level	−1.459	−2.771	−0.147	−2.181	0.029
Annual income	0.739	0.254	1.224	2.988	0.003
Received GP Structured Vocational Training	2.049	0.956	3.142	3.676	<0.001
Attending training on diabetic microvascular complications or not	3.386	2.180	4.592	5.510	<0.001
Number of screening instrument	0.649	0.338	0.961	4.09	<0.001
Practice score	0.157	0.124	0.189	9.481	<0.001
Practice
Gender	2.419	0.529	4.308	2.511	0.012
Duration of practice	1.033	0.092	1.973	2.154	0.031
Location of practice	−3.230	−5.081	−1.378	−3.422	0.001
Attending training on diabetic microvascular complications or not	6.346	4.412	8.280	6.437	<0.001
Managing diabetes patients or not	4.503	2.224	6.782	3.877	<0.001
Number of diabetic patients under management	0.007	0.000	0.014	2.106	0.035
Number of screening instrument	1.681	1.184	2.178	6.636	<0.001
Knowledge score	0.190	0.088	0.291	3.649	<0.001
Attitude score	0.414	0.328	0.500	9.453	<0.001

Association between KAP score and general practice working years

We used restricted cubic splines and a model adjusted for age, gender and education level to explore the association between KAP score and general practice working years. Our analysis revealed that the KAP total score increased with the working years of general practice (figure 1A). However, the knowledge score did not increase after 15 years, while the attitude and practice scores showed a continuously increasing trend (figure 1B).

Figure 1. (A) Association between knowledge, attitude and practice (KAP) score and duration of general practice, adjusted age, gender and education level. (B) Association between knowledge, attitude, practice score and duration of general practice, adjusted age, gender and education level.

The training experience on diabetic microvascular complications and screening instruments equipped by CHCs

Most GPs learnt about diabetic microvascular complications through lectures, courses, guides and literature. However, it was concerning that 32.0% of GPs had not received training on diabetic microvascular complications in the past year. The majority of GPs expressed their preference for acquiring knowledge about diabetic microvascular complications through online workshops or courses. Only about half of GPs reported available tonometers, ophthalmoscopes and fundus cameras for DR screening in their CHCs. Nearly all GPs reported that their CHCs were equipped with blood creatinine and urine tests to screen for DKD, while most CHCs also had access to urine albumin creatinine ratio (UACR) tests. Only about half of GPs reported that their CHCs were equipped with a 128 Hz tuning fork and a 10 g nylon wire, which are essential for screening DN. In most CHCs, GPs reported that DM can be followed up with continuous glucose monitoring (CGM) and glycosylated haemoglobin. An overview of the training experience of GPs on diabetic microvascular complications and the screening instruments for diabetic microvascular complications in their CHCs are shown in table 4.

Table 4. The training experience in diabetic microvascular complications and screening instrument based on self-report of general practitioner (GPs) in their community health centres (CHCs).

	N (%)
Where did you get the knowledge about diabetic microvascular complications?
Lectures and courses	1116 (89.8)
Peer communication	829 (66.7)
Guide and literature	998 (80.3)
Professional book	790 (63.6)
How many times had you attended training in diabetic microvascular complications in the past year?
1	388 (31.2)
2	272 (21.9)
3	74 (6.0)
4 or more times	111 (8.9)
None	398 (32.0)
In what form would you like to acquire knowledge about diabetic microvascular complications?
Self-study	540 (43.3)
Peer communication	761 (61.2)
Online courses	1072 (86.2)
Offline courses	668 (53.7)
Others
To screening for diabetic retinopathy, which of the following instrument are available in your CHC?
Tonometer	575 (46.3)
Ophthalmoscope	810 (65.2)
Fundus camera	538 (43.3)
Unknown	254 (20.4)
To screening for diabetic kidney disease, which of the following instrument are available in your CHC?
Urine albumin creatinine ratio	904 (72.7)
Serum creatinine	1199 (96.5)
Routine urine test	1198 (96.4)
Unknown	19 (1.5)
To screening for diabetic neuropathy, which of the following instrument are available in your CHC?
128 Hz tuning fork	654 (52.6)
10 g nylon wire	671 (54.0)
Unknown	425 (34.2)
To follow-up diabetes mellitus, which of the following instrument are available in your CHC?
Continuous glucose monitoring devices	1040 (83.7)
Glycosylated haemoglobin	1224 (98.5)
Oral glucose tolerance test	825 (66.4)

Discussion

While several studies about the KAP of GPs toward DR, DKD or DF have been previously reported,1116 26,28 studies on the KAP of GPs toward diabetic microvascular complications are lacking. The present research showed that GPs in Shanghai had insufficient knowledge of screening and managing diabetic microvascular complications. Despite demonstrating a strong awareness of managing diabetic microvascular complications and a positive self-assessment of their capabilities in this area, as well as expressing a willingness to participate in relevant training and research, GPs infrequently provided screening recommendations or patient education, or conduct screenings in practice. Furthermore, the knowledge score of GPs did not increase after 15 years of general practice, while the attitude score and practice score showed a continuously increasing trend.

Our results revealed insufficient knowledge regarding diabetic microvascular complications among GPs. The three questions with the lowest accuracy were: ‘diabetic microvascular complications mainly involve the heart, kidney, retina, and peripheral nerves’ (7.4%), ‘screening for diabetic peripheral neuropathy is feasible with 15 g nylon wire test’ (11.1%) and ‘microaneurysms belong to proliferative diabetic retinopathy’ (13.4%). The knowledge score of DR was lower than that of DKD, DN and DF; however, it was impossible to make the comparisons with the existing research as these studies only reported single complication. A study on KAP of DR conducted by Lestari et al in Jakarta, Indonesia, showed that 89.1% of GPs demonstrated good knowledge, defined by the correct response rate of >75%.¹⁶ On the contrary, Al Rasheed et al reported that the mean overall knowledge score of DR in Riyadh, Saudi Arabia, was 57±14/100 in 216 primary care physicians,²⁹ which is similar to our results of 52.8±17.7. This could be because GPs believed DR screening and managing did not constitute primary healthcare and tended to refer patients to ophthalmologists.¹⁶ In the present study, the DN score was the second lowest in the knowledge score of diabetic microvascular complications, while the DF score was the highest. There are few KAP studies of provider’s DN knowledge, especially in GPs. In their study of DF, Alsheikh et al found that the knowledge of 152 primary care providers on diagnosing and managing DF infection was suboptimal.²⁷ Only 19.7% of the primary care physicians evaluated themselves as above average when asked to self-evaluate their knowledge of therapies and prevention of DF. In the present study, the average knowledge score of DKD was 77.9±14.1/100, which is more than in the previous study, reporting 62.7% accurate rates on knowledge of DKD among 138 GPs in Shanghai.²⁶ Generally speaking, in the present study, the GPs’ knowledge of diabetic microvascular complications was insufficient. Training on diabetic microvascular complications may help improve the knowledge of GPs about diabetic microvascular complications. DM management and its complications have been a priority in Jakarta since 2020, and many webinars and training opportunities have been held to raise awareness of GPs.¹⁶ However, in our study, nearly one-third of GPs had not received training on diabetic microvascular complications in the past year. It is noteworthy that healthcare delivery is a knowledge-driven process, and a lack of adequate knowledge may lead to inefficient healthcare delivery.³⁰ In the future, we must conduct more training on diabetic microvascular complications for GPs, especially in diabetic retinopathy and DN. Based on the present study’s findings, most GPs preferred acquiring knowledge through online workshops or courses. For GPs, the virtual training method may be more convenient as it allows them to choose the time and place of study flexibly.

According to the results of this study, GPs recognised the importance of managing diabetic microvascular complications, showed a positive attitude and preferred to participate in training and research on diabetic microvascular complications. Similar to our study, GPs in Indonesia showed a positive attitude toward DR screening.¹⁶ However, other studies have come to the contrary conclusions. Negative attitude was found among primary care physicians in Saudi Arabia.³¹ Furthermore, although GPs in the present study had a generally positive attitude, they still lacked confidence in managing diabetic microvascular complications, which was concluded based on the self-ability evaluation item: “I fully understand my clinical ability and know how to choose the right way to care for microvascular complications in diabetic patients”. Similar to Lestari et al, who reported that only 65.2% of GPs in Indonesia agreed or strongly agreed with the question: “Ophthalmology training in a medical school adequately equips the GP to manage patients with eye complaints”.¹⁶

In the present study, GPs showed poor practice in managing diabetic microvascular complications, especially screening for patients with diabetes and giving screening suggestions to such patients. The average scores of the two dimensions were 46.5 and 53.3/100, respectively, which is similar to some previous studies. A study by Al-Rashidi et al revealed that only 37% of GPs in Qassim Province, Saudi Arabia, performed dilated fundus exams.³² A qualitative study by Watson et al revealed that GPs faced role confusion in DR screening as some GPs considered themselves a referral source.³³ In addition, we also investigated the availability of diabetes-related instruments, including screening instruments for diabetic microvascular complications in CHCs. Glycosylated haemoglobin could be done in nearly all CHCs, and >80% of CHCs were equipped with CGM devices. However, instruments for screening diabetic microvascular complications were insufficient, as only 65.2% had ophthalmoscope, 43.3% had fundus camera, 72.2% had UACR, 52.6% had 128 Hz tuning fork and 54.0% had 10 g nylon wire. Some GPs did not know what kind of screening instruments their CHCs had, which indirectly suggested that they had never done screening for microvascular complications in diabetic patients.

In the present study, we further explored the influencing factors of knowledge, attitude and practice of GPs. Interestingly, we found that the number of screening instruments was associated with the KAP of GPs. More screening instruments mean more opportunities for GPs to conduct screening, which requires more knowledge and gives GPs more confidence to manage their patients. In order to urge GPs to screen diabetic microvascular complications more effectively, the CHCs should be equipped with sufficient screening instruments and carry out targeted training.

Another interesting finding was that attending training on diabetic microvascular complications was associated with the attitude and practice of GPs but not with knowledge. Received GP Structured Vocational Training was associated with the attitude of GPs but not knowledge or practice. We also found that the knowledge score did not increase after 15 working years of general practice, while the attitude score and practice score continuously increased. These results differed from a study by Al Rasheed et al, which reported that family medicine subspecialty training and years of practice were predictors of knowledge on DR.³⁰ The evaluation of knowledge can be affected by many aspects, while practice evaluation is relatively more objective. Although continuing medical education (CME) on diabetic microvascular complications did not improve GPs’ knowledge, it did improve their practical ability, which is the ultimate scope of CME.

Strengths and limitations

To the best of our knowledge, this is the first study that estimated the KAP of GPs toward the screening and managing diabetic microvascular complications. Previous studies on diabetes and its microvascular complications have predominantly centred on patient populations. Gaining insight into the knowledge, attitudes and practices of general practitioners could offer valuable perspectives for enhancing the management of such conditions. The findings of this study could provide a solid basis for improving the management of diabetic microvascular complications in community settings. However, the present study has several limitations. First, a potential methodological limitation of our study was the selection bias that may arise from the use of convenience sampling. To mitigate this issue, we implemented several measures to enhance the representativeness of the sample, resulting in the inclusion of approximately 12% of the city’s total GP population was included in the current study. Second, there might have been some information bias in the instruments of CHCs due to reliance on self-reported data. Third, it is crucial to note that this study was cross-sectional, meaning only associations and not causation could be assessed. Lastly, the KAP scale used in our research was based on the context of Chinese GPs. Despite its scientifically rigorous design, making it suitable for evaluating the KAP of most primary care GPs, it is recommended that other regions consider the unique aspects of their healthcare systems, resource allocation and diabetes prevalence when adapting the scale items for local use. Future longitudinal studies are necessary to better understand the cause-effect relationships between KAP and influencing factors.

Conclusions

GPs in Shanghai had insufficient knowledge of community screening and managing diabetic microvascular complications. Despite demonstrating a strong awareness of managing diabetic microvascular complications and a positive self-assessment of their capabilities in this area, as well as expressing a willingness to participate in relevant training and research, GPs infrequently provided screening recommendations or patient education, or conduct screenings in practice. Equipping the CHCs with sufficient screening instruments and carrying out targeted training for GPs may help improve the practice capability on diabetic microvascular complications.

Data availability statement

All data relevant to the study are included in the article or uploaded as supplementary information.