1
|
Gopalakrishnan N, Joshi A, Chhablani J, Yadav NK, Reddy NG, Rani PK, Pulipaka RS, Shetty R, Sinha S, Prabhu V, Venkatesh R. Recommendations for initial diabetic retinopathy screening of diabetic patients using large language model-based artificial intelligence in real-life case scenarios. Int J Retina Vitreous 2024; 10:11. [PMID: 38268046 PMCID: PMC10809735 DOI: 10.1186/s40942-024-00533-9] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 12/01/2023] [Accepted: 01/19/2024] [Indexed: 01/26/2024] Open
Abstract
PURPOSE To study the role of artificial intelligence (AI) to identify key risk factors for diabetic retinopathy (DR) screening and develop recommendations based on clinician and large language model (LLM) based AI platform opinions for newly detected diabetes mellitus (DM) cases. METHODS Five clinicians and three AI applications were given 20 AI-generated hypothetical case scenarios to assess DR screening timing. We calculated inter-rater agreements between clinicians, AI-platforms, and the "majority clinician response" (defined as the maximum number of identical responses provided by the clinicians) and "majority AI-platform" (defined as the maximum number of identical responses among the 3 distinct AI). Scoring was used to identify risk factors of different severity. Three, two, and one points were given to risk factors requiring screening immediately, within a year, and within five years, respectively. After calculating a cumulative screening score, categories were assigned. RESULTS Clinicians, AI platforms, and the "majority clinician response" and "majority AI response" had fair inter-rater reliability (k value: 0.21-0.40). Uncontrolled DM and systemic co-morbidities required immediate screening, while family history of DM and a co-existing pregnancy required screening within a year. The absence of these risk factors required screening within 5 years of DM diagnosis. Screening scores in this study were between 0 and 10. Cases with screening scores of 0-2 needed screening within 5 years, 3-5 within 1 year, and 6-12 immediately. CONCLUSION Based on the findings of this study, AI could play a critical role in DR screening of newly diagnosed DM patients by developing a novel DR screening score. Future studies would be required to validate the DR screening score before it could be used as a reference in real-life clinical situations. CLINICAL TRIAL REGISTRATION Not applicable.
Collapse
Affiliation(s)
- Nikhil Gopalakrishnan
- Department of Retina and Vitreous, Narayana Nethralaya Eye Hospital, #121/C, 1st R Block, Chord Road, Rajaji Nagar, Bengaluru, Karnataka, 560010, India
| | - Aishwarya Joshi
- Department of Retina and Vitreous, Narayana Nethralaya Eye Hospital, #121/C, 1st R Block, Chord Road, Rajaji Nagar, Bengaluru, Karnataka, 560010, India
| | - Jay Chhablani
- Medical Retina and Vitreoretinal Surgery, University of Pittsburgh School of Medicine, 203 Lothrop Street, Suite 800, Pittsburg, PA, 15213, USA
| | - Naresh Kumar Yadav
- Department of Retina and Vitreous, Narayana Nethralaya Eye Hospital, #121/C, 1st R Block, Chord Road, Rajaji Nagar, Bengaluru, Karnataka, 560010, India
| | - Nikitha Gurram Reddy
- Anant Bajaj Retina Institute, L V Prasad Eye Institute, Kallam Anji Reddy Campus, Hyderabad, Telangana, 500034, India
| | - Padmaja Kumari Rani
- Anant Bajaj Retina Institute, L V Prasad Eye Institute, Kallam Anji Reddy Campus, Hyderabad, Telangana, 500034, India
| | - Ram Snehith Pulipaka
- Prime Retina Eye Care Center, 3-6-106/1, Street Number 19, Opposite to Vijaya Diagnostic Centre, Himayatnagar, Hyderabad, Telangana, 500029, India
| | - Rohit Shetty
- Department of Cornea and Refractive Services, Narayana Nethralaya Eye Hospital, #121/C, 1st R Block, Chord Road, Rajaji Nagar, Bengaluru, Karnataka, 560010, India
| | - Shivani Sinha
- Department of Vitreo-Retina, Regional Institute of Ophthalmology, Indira Gandhi Institute of Medical Sciences, Sheikhpura, Patna, Bihar, 800014, India
| | - Vishma Prabhu
- Department of Retina and Vitreous, Narayana Nethralaya Eye Hospital, #121/C, 1st R Block, Chord Road, Rajaji Nagar, Bengaluru, Karnataka, 560010, India
| | - Ramesh Venkatesh
- Department of Retina and Vitreous, Narayana Nethralaya Eye Hospital, #121/C, 1st R Block, Chord Road, Rajaji Nagar, Bengaluru, Karnataka, 560010, India.
| |
Collapse
|
2
|
Zang B, Rong S, Li D, Ding X, Zang D, Wang F, Liang Y, Zhai G, Feng K, Zhou Z, Wang Y. Undiagnosed diabetic retinopathy in Northeast China: prevalence and determinants. Front Endocrinol (Lausanne) 2023; 14:1263508. [PMID: 38093961 PMCID: PMC10716530 DOI: 10.3389/fendo.2023.1263508] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Figures] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 07/19/2023] [Accepted: 11/13/2023] [Indexed: 12/18/2023] Open
Abstract
Objective To report the prevalence and contributing factors of undiagnosed diabetic retinopathy (DR) in a population from Northeastern China. Subjects/Methods A total of 800 subjects from the Fushun Diabetic Retinopathy Cohort Study were enrolled. A questionnaire assessing incentives and barriers to diagnosis of DR was administered. Logistic regression was used to identify clinical and sociodemographic factors associated with undiagnosed DR. In a prespecified subgroup analysis, we divided patients into vision-threatening diabetic retinopathy (VTDR) and non-VTDR (NVTDR) subgroups. Results Among 800 participants with DR, 712 (89.0%) were undiagnosed. Among 601 with NVTDR, 566 (94.2%) were undiagnosed. Among 199 with VTDR, 146 (73.4%) were undiagnosed. The risk factors affecting the timely diagnosis of NVTDR and VTDR exhibit significant disparities. In multivariate models, factors associated with undiagnosed VTDR were age over 60 years (OR = 2.966; 95% CI = 1.205-7.299; P = 0.018), duration of diabetes over 10 years (OR = 0.299; 95% CI = 0.118-0753; P = 0.010), visual impairment or blindness (OR = 0.310; 95% CI = 0.117-0.820; P = 0.018), receiving a reminder to schedule an eye examination (OR = 0.380; 95% CI = 0.163-0.883; P = 0.025), and the belief that "people with diabetes are unlikely to develop an eye disease" (OR = 4.691; 95% CI = 1.116-19.724; P = 0.035). However, none of the factors were associated with undiagnosed NVTDR (all P ≥ 0.145). Conclusion Our research has uncovered a disconcerting trend of underdiagnosis in cases of DR within our population. Addressing determinants of undiagnosed DR may facilitate early detection.
Collapse
Affiliation(s)
- Bo Zang
- Department of Ophthalmology, Fushun Eye Hospital, Fushun, Liaoning, China
| | - Shisong Rong
- Mass Eye and Ear, Mass General Brigham, Harvard Medical School, Boston, MA, United States
| | - Dong Li
- Department of Ophthalmology, Fushun Eye Hospital, Fushun, Liaoning, China
| | - Xiaoxia Ding
- Department of Ophthalmology, Fushun Eye Hospital, Fushun, Liaoning, China
| | - Dongxiao Zang
- Department of Ophthalmology, Fushun Eye Hospital, Fushun, Liaoning, China
| | - Fenghua Wang
- Department of Ophthalmology, Beijing Tongren Eye Center, Beijing Tongren Hospital, Capital Medical University, Beijing Ophthalmology and Visual Science Key Laboratory, Beijing, China
| | - Yuanbo Liang
- Eye Hospital, Wenzhou Medical University, Wenzhou, China
| | - Gang Zhai
- Department of Ophthalmology, Fushun Eye Hospital, Fushun, Liaoning, China
| | - Kemi Feng
- Eye Hospital, Wenzhou Medical University, Wenzhou, China
| | - Zixi Zhou
- Department of Ophthalmology, Fushun Eye Hospital, Fushun, Liaoning, China
| | - Yu Wang
- Department of Ophthalmology, Fushun Eye Hospital, Fushun, Liaoning, China
| |
Collapse
|
3
|
Dweib M, El Sharif N. Diabetes-Related Microvascular Complications in Primary Health Care Settings in the West Bank, Palestine. J Clin Med 2023; 12:6719. [PMID: 37959185 PMCID: PMC10649955 DOI: 10.3390/jcm12216719] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 09/28/2023] [Revised: 10/17/2023] [Accepted: 10/20/2023] [Indexed: 11/15/2023] Open
Abstract
BACKGROUND Worldwide, retinopathy, nephropathy, and neuropathy are the major diabetes-related microvascular complications. In Palestine, a low-middle-income country, diabetes is the fourth reason for death. However, a few studies examined diabetes microvascular consequences and its management. Therefore, we carried out a national study that aims to investigate the factors associated with diabetes-related microvascular complications among individuals seeking care in primary healthcare settings of the West Bank of Palestine. METHOD Using a cluster systematic sampling technique, 882 participants with diabetes patients were chosen for a cross-sectional study from primary healthcare facilities operated by the Ministry of Health (PMoH), the United Nations Relief and Works Agency (UNRWA), and the Palestinian Medical Relief Society (PMRS). Data about patients related to diabetes-related complications, medication use, and other diseases were extracted from patients' medical records. In addition, an interview face-to-face questionnaire was used to collect information about patients' sociodemographic variables, medical history, smoking habits, duration of the disease, presence of concurrent conditions previous referrals, and hospital admissions, as well as their level of knowledge regarding diabetes, complications, and treatments. RESULTS Approximately 34.4% of persons with diabetes patients in Palestine encounter at least one microvascular complication associated with diabetes. The most prevalent diabetes-related microvascular complication was retinopathy (17.3%), 23.4% of participants had more than one microvascular complication, and 29% of male patients had erectile dysfunction. A higher probability of having any microvascular complications was associated with older age (over 60 years). Participants with diabetes patients with fundoscopy or ophthalmology reports, according to diabetes follow-up guidelines, were less likely to develop retinopathy. Also, those who performed regular kidney function testing were less likely to have nephropathy, and those who performed a regular foot exam were less likely to develop diabetic foot. CONCLUSIONS Diabetes-related microvascular complications were associated with patient age, low education level, residency location, and adherence to diabetes follow-up guidelines of diabetes management; i.e., having been tested for HbA1c, consulting with specialists, regular kidney function, and foot examination. These factors can be utilized in setting up proper management protocols to prevent or delay microvascular complications in many patients.
Collapse
Affiliation(s)
- Mohammad Dweib
- College of Pharmacy and Medical Sciences, Hebron University, P.O. Box 40, Hebron P720, Palestine;
- School of Public Health, Al-Quds University, Abu Dis, P.O. Box 51000, Jerusalem 20002, Palestine
| | - Nuha El Sharif
- School of Public Health, Al-Quds University, Abu Dis, P.O. Box 51000, Jerusalem 20002, Palestine
| |
Collapse
|
4
|
Appukumran R, Shyamsundar K, Agrawal M, Khurana R, Pannu A, Kumar P. Eight years' experience in mobile teleophthalmology for diabetic retinopathy screening. MEDICAL HYPOTHESIS, DISCOVERY & INNOVATION OPHTHALMOLOGY JOURNAL 2023; 11:162-170. [PMID: 37641607 PMCID: PMC10460246 DOI: 10.51329/mehdiophthal1460] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Subscribe] [Scholar Register] [Received: 11/06/2022] [Accepted: 01/04/2023] [Indexed: 08/31/2023]
Abstract
Background Screening for diabetic retinopathy in the community without compromising the routine work of ophthalmologists at hospitals is the essence of teleophthalmology. This study was aimed at investigating the efficacy of teleophthalmology practice for screening diabetic retinopathy from 2012 to 2020. It was also aimed at comparing the 2-year prevalence of camps organized by a district hospital in South India, as well as the footfall, reporting, follow-up, patient response, and diagnostic efficacy at these camps. Methods All patients with diabetes and unexplained vision deterioration attending the mobile camp units underwent non-dilated fundus photography. Patients underwent teleconsultation with the ophthalmologist at the district hospital, and those requiring intervention were called to the district hospital. Trends were studied for the number of patients reporting to the hospital. Patient satisfaction was recorded based on a questionnaire. Results A total of 682 camps were held over 8 years, and 30 230 patients were examined. Teleconsultation was done for 12 157 (40.21%) patients. Patients requiring further investigations, intervention for diabetic retinopathy, or further management of other ocular pathologies were urgently referred to the district hospital (n= 3293 [10.89%] of 30 230 examined patients). The severity and presence of clinically significant macular edema increased significantly with an increased duration of diabetes mellitus (P < 0.001). The percentage of teleconsultations showed an increasing trend over the years (P = 0.001). Similarly, considering trends of patients reporting to the hospital, the attrition rate decreased over the years (P < 0.05). A total of 10 974 of 12 157 (90.27%) patients who underwent teleophthalmic consultation were satisfied with the service. Conclusions Teleconsultations over the years showed an increasing trend, and the attrition rate decreased over the years. Teleophthalmology is achieving success in providing high-quality service, easy access to care, and in increasing patient satisfaction. Future studies on the role of teleophthalmology for other leading preventable causes of blindness seem possible and necessary.
Collapse
Affiliation(s)
| | | | - Mohini Agrawal
- Department of Ophthalmology, Military Hospital, Jalandhar, Punjab, India
| | - Rolli Khurana
- Department of Ophthalmology, Military Hospital, Ahmedabad, India
| | - Anju Pannu
- Department of Community Medicine, Armed Forces Medical College, Pune, India
| | - Praveen Kumar
- Department of Community Medicine, Armed Forces Medical College, Pune, India
| |
Collapse
|