Overall Patient characteristics
We recorded data of 3511 eyes in a total of 1763 patients, of which 15 patients had only one eye. Among them, 326 eyes (9.29%) of 209 patients had poor fundus image quality and were not suitable for grading. A total of 1646 persons (3185 eyes) were graded in this study.The patients suitable for screening of DR (n = 1646) were 58.64 ± 10.88 years old, 58.33% male,the incidence rate of hypertension was 43.16%.
Analysis of efficacy of community general practitioner for screening of diabetes retinopathy in 2020
A total of 1131 eyes of the subjects underwent relevant examinations. Ophthalmologist diagnosed 419 eyes with DR, including 340 R1, 62 R2,17 R3. The prevalence of DR was 37.05% (419/ 1131). Community general practitioner diagnosed 431 eyes with DR, including 338 R1, 55 R2,38 R3. The prevalence of DR was 38.11% (431/ 1131). The Kappa value was 0.597, and the examination results of ophthalmologist and community general practitioner were statistically significant through the consistency test. It is thinked that the consistency of the two doctors is poor. (Table 1)
Table 1
The screening of diabetic retinopathy by different doctors in 2020
Ophthalmologist | Community general practitioner | Total |
R0 | R1 | R2 | R3 |
R0 | 620 | 92 | 0 | 0 | 712 |
R1 | 80 | 237 | 22 | 1 | 340 |
R2 | 0 | 9 | 27 | 26 | 62 |
R3 | 0 | 0 | 6 | 11 | 17 |
Total | 700 | 338 | 55 | 38 | 1131 |
Kappa | 0.597 |
P | 0.000 |
The diagnostic sensitivity of community general practitioner for screening of diabetes retinopathy in diabetic patients was 78.65% (339/431), the specificity was 88.57% (620/700), and the accuracy was 84.79% (959/1131), as shown in Table 2.
Table 2
Sensitivity, specificity and accuracy of community general practitioner in the diagnosis of diabetes retinopathy in 2020
Ophthalmologist | Community general practitioner | Sensitivity% | Specificity% | Accuracy% |
Positive | Negative |
Positive | 339 | 80 | 78.65% (339/431) | 88.57% (620/700) | 84.79% (959/1131) |
Negative | 92 | 620 |
Analysis of efficacy of community general practitioner for screening of diabetes retinopathy in 2021
A total of 1452 eyes of the subjects underwent relevant examinations. Ophthalmologist diagnosed 1051 eyes with DR, including 958 R1, 58 R2,35 R3. The prevalence of DR was 72.38% (1051/ 1452). Community general practitioner diagnosed 1089 eyes with DR, including 963 R1, 88 R2,38 R3. The prevalence of DR was 75.00% (1089 / 1452). The Kappa value was 0.692 which was better than last year, and it was statistically significant through the consistency test. (Table 3).
Table 3
The screening of diabetic retinopathy by different doctors in 2021
Ophthalmologist | Community general practitioner | Total |
R0 | R1 | R2 | R3 |
R0 | 315 | 85 | 1 | 0 | 401 |
R1 | 48 | 864 | 46 | 0 | 958 |
R2 | 0 | 14 | 30 | 14 | 58 |
R3 | 0 | 0 | 11 | 24 | 35 |
Total | 363 | 963 | 88 | 38 | 1452 |
Kappa | 0.692 |
P | 0.000 |
The diagnostic sensitivity of community general practitioner for screening of diabetes retinopathy in diabetic patients was 92.10% (1003/1089), the specificity was 86.78% (315/363), and the accuracy was 90.77% (1318/1452), as shown in Table 4.
Table 4
Sensitivity, specificity and accuracy of community general practitioner in the diagnosis of diabetes retinopathy in 2021
Ophthalmologist | Community general practitioner | Sensitivity% | Specificity% | Accuracy% |
Positive | Negative |
Positive | 1003 | 48 | 92.10% (1003/1089) | 86.78% (315/363) | 90.77% (1318/1452) |
Negative | 86 | 315 |
Analysis of efficacy of community general practitioner for screening of diabetes retinopathy in 2022
A total of 602 eyes of the subjects underwent relevant examinations by June 2022. Ophthalmologist diagnosed 466 eyes with DR, including 426 R1, 24 R2,16 R3. The prevalence of DR was 77.41% (466 / 602). Community general practitioner diagnosed 453 eyes with DR, including 398 R1, 37 R2,18 R3. The prevalence of DR was 75.25% (453 / 602). The Kappa value was 0.754 which is best in these three years, and it was statistically significant through the consistency test. (Table 5).
Table 5
The screening of diabetic retinopathy by different doctors in 2022
Ophthalmologist | Community general practitioner | Total |
R0 | R1 | R2 | R3 |
R0 | 121 | 15 | 0 | 0 | 136 |
R1 | 28 | 379 | 19 | 0 | 426 |
R2 | 0 | 4 | 17 | 3 | 24 |
R3 | 0 | 0 | 1 | 15 | 16 |
Total | 149 | 398 | 37 | 18 | 602 |
Kappa | 0.754 |
P | 0.000 |
The diagnostic sensitivity of community general practitioner for screening of diabetes retinopathy in diabetic patients was 96.69% (438/453), the specificity was 81.21% (121/149), and the accuracy was 92.86% (559/602), as shown in Table 6.
Table 6
Sensitivity, specificity and accuracy of community general practitioner in the diagnosis of diabetes retinopathy in 2022
Ophthalmologist | Community general practitioner | Sensitivity% | Specificity% | Accuracy% |
Positive | Negative |
Positive | 438 | 28 | 96.69% (438/453) | 81.21% (121/149) | 92.86% (559/602) |
Negative | 15 | 121 |
Sensitivity, specificity and accuracy of community general practitioner in the diagnosis of diabetes retinopathy in these three years
The sensitivity and accuracy of community general practitioners in diagnosing diabetes retinopathy had increased year by year in these three years (P < 0.05), as shown in Table 7.
Table 7
Sensitivity, specificity and accuracy of community general practitioner in the diagnosis of diabetes retinopathy in these three years
Time | Cases | Sensitivity% | Specificity% | Accuracy% |
2020 | 1131 | 78.65% (339/431) | 88.57% (620/700) | 84.79% (959/1131) |
2021 | 1452 | 92.10% (1003/1089) | 86.78% (315/363) | 90.77% (1318/1452) |
2022 | 602 | 96.69% (438/453) | 81.21% (121/149) | 92.86% (559/602) |
χ2 | | 91.18 | 6.00 | 34.37 |
P | | 0.000 | 0.05 | 0.000 |