Impact of Artificial Intelligence Assessment of Diabetic Retinopathy on Referral Service Uptake in a Low-Resource Setting

Mathenge, Wanjiku; Whitestone, Noelle; Nkurikiye, John; Patnaik, Jennifer L; Piyasena, Prabhath; Uwaliraye, Parfait; Lanouette, Gabriella; Kahook, Malik Y.; Cherwek, David H.; Congdon, Nathan; Jaccard, Nicolas

doi:10.1016/j.xops.2022.100168

Cited by 30 publications

(40 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We came across a rather large number of LMICbased publications of research protocols related to planned or on-going AI evaluations, as well as studies since the time we our literature search that would have met our inclusion criteria. For instance, recent ophthalmology studies from Thailand and Rwanda have demonstrated the potential of AIassisted diabetic retinopathy screening in LMICs while also flagging issues similar to those of our included studies, such as the challenge of integrating AI systems into existing workflows and infrastructure 31,32 . The private sector is also highly active in developing AI tools for healthcare, as our grey literature search revealed (see Table 5).…”

Section: Discussionmentioning

confidence: 80%

Artificial intelligence for strengthening healthcare systems in low- and middle-income countries: a systematic scoping review

et al. 2022

View full text Add to dashboard Cite

In low- and middle-income countries (LMICs), AI has been promoted as a potential means of strengthening healthcare systems by a growing number of publications. We aimed to evaluate the scope and nature of AI technologies in the specific context of LMICs. In this systematic scoping review, we used a broad variety of AI and healthcare search terms. Our literature search included records published between 1st January 2009 and 30th September 2021 from the Scopus, EMBASE, MEDLINE, Global Health and APA PsycInfo databases, and grey literature from a Google Scholar search. We included studies that reported a quantitative and/or qualitative evaluation of a real-world application of AI in an LMIC health context. A total of 10 references evaluating the application of AI in an LMIC were included. Applications varied widely, including: clinical decision support systems, treatment planning and triage assistants and health chatbots. Only half of the papers reported which algorithms and datasets were used in order to train the AI. A number of challenges of using AI tools were reported, including issues with reliability, mixed impacts on workflows, poor user friendliness and lack of adeptness with local contexts. Many barriers exists that prevent the successful development and adoption of well-performing, context-specific AI tools, such as limited data availability, trust and evidence of cost-effectiveness in LMICs. Additional evaluations of the use of AI in healthcare in LMICs are needed in order to identify their effectiveness and reliability in real-world settings and to generate understanding for best practices for future implementations.

show abstract

Section: Discussionmentioning

confidence: 80%

Artificial intelligence for strengthening healthcare systems in low- and middle-income countries: a systematic scoping review

et al. 2022

View full text Add to dashboard Cite

show abstract

“…35,36 Innovative strategies such as personalized interventions and immediate feedback on referral status based on artificial intelligence-supported screening have shown promise in recent randomized controlled trials for improving adherence to recommended follow-up primary eye care exams. 37,38 Similar strategies could be extended to a teleretinal screening setting since TRI is often captured in primary care settings as in our study.…”

Section: Discussionmentioning

confidence: 99%

Factors Associated with Ophthalmology Referral and Adherence in a Teleretinal Screening Program: Insights from a Federally Qualified Health Center

Song¹,

Johnson²,

Mirzania³

et al. 2022

OPTH

View full text Add to dashboard Cite

Background: Early detection of ophthalmic conditions such as diabetic retinopathy (DR) and glaucoma is crucial to preventing vision loss. Previous studies have evaluated teleretinal screening programs for DR in well-insured populations. The purpose of this retrospective study was to evaluate a teleretinal screening program in a population of uninsured and underinsured patients seen in a Federally Qualified Health Center (FQHC). Methods: We conducted a retrospective chart review of patients (age ≥18) who underwent teleretinal imaging (TRI) at a FQHC between January 2015 and September 2019. TRI gradings and patient demographic and clinical information were abstracted. Factors associated with referral for a dilated eye exam by an ophthalmologist, adherence to recommended follow-up dilated eye exam, and ophthalmology visit attendance were examined. Results: 3130 TRIs were graded in 2216 eyes (1107 patients). 65.2% (N = 722) self-identified as Hispanic and 56.3% (N = 623) required interpreter services. Follow-up dilated fundus examination (DFE) was recommended for 388 TRIs, 49% (N = 190) of which were completed within 1 year. Adherence to the recommended ophthalmology exam was not associated with any baseline clinical or demographic characteristics (p > 0.05). Older age, male sex, hypertension, proteinuria, and higher A1c were significantly associated with greater odds of ophthalmology referral based on TRI (all p < 0.05), after adjusting for covariates. Less severe diabetic retinopathy, no insurance coverage, and Hispanic ethnicity were associated with lower odds of attending an ophthalmology visit, regardless of follow-up recommendations based on TRI (all p < 0.05). Conclusion:In an FQHC serving predominantly uninsured and underinsured patients, only 49% of recommended DFE were completed within one year. Less severe diabetic retinopathy, lack of insurance coverage, and Hispanic ethnicity were associated with a lower likelihood of having a DFE regardless of recommendation. These results suggest that greater system-level efforts are needed to increase adherence to follow-up eye exams after TRI to ensure sight-saving care for underserved populations.

show abstract

“…The advantages of our predictive diagnostic system —such as increased objectivity and efficiency in determining early risk factors for DR by the Random Forest system compared with health-care professionals, higher referral adherence from real-time point-of-care screening recommendations 42 , more efficient resource allocation towards prevention and treatments due to the Random Forest system offloading tasks from human graders, and reduction in the prevalence of RD in the mid to long term— are implied but unproven. Also, although translation of our system into clinics seems feasible, real-world testing and acceptation is in its infancy.…”

Section: Discussionmentioning

confidence: 99%

“…The clinical community is experiencing an explosion of machine learning-guided (tele)diagnostics, particularly in the context of detecting treatment-requiring DR, and the positive impact on care delivery is becoming evident 41,42 . Risk-scoring algorithms for undiagnosed diabetes are also available 43 , but there is no such approach for the main risk factors of type 2 diabetes, i.e., overweight, obesity, and MetS.…”

Section: Discussionmentioning

confidence: 99%

Spontaneous electroretinogram signals to screen people with early risk factors for diabetic retinopathy

Imm

Muñoz-Benitez

Medina

et al. 2022

Preprint

View full text Add to dashboard Cite

Given the ever-increasing prevalence of type 2 diabetes and obesity, the pressure on global healthcare is expected to be colossal, especially in terms of blindness. Electroretinogram (ERG) has long been perceived as a first-use technique for diagnosing eye diseases, but existing methods are insufficient to screen early risk factors of diabetic retinopathy (DR). Here, we introduce non-evoked ERG as a simple, fast modality to record spontaneous activity, from which we developed a single random forest-based model that predicts disease cases in rodent models of obesity and in people with overweight, obesity, and metabolic syndrome. Classification performance was validated using a dataset from an independent eye center. Our algorithm can be coupled with different ERG sensors, including ones working with portative, non-mydriatic devices. Principal component and discriminant analysis suggest slow spontaneous ERG frequencies as main discriminators for our predictive model. Our study will facilitate the implementation of interventions for the prevention of overweight and obesity by providing a robust, quantitative, and non-invasive identification and follow-up approach, which should ultimately reduce DR incidence.

show abstract

Impact of Artificial Intelligence Assessment of Diabetic Retinopathy on Referral Service Uptake in a Low-Resource Setting

Cited by 30 publications

References 21 publications

Artificial intelligence for strengthening healthcare systems in low- and middle-income countries: a systematic scoping review

Artificial intelligence for strengthening healthcare systems in low- and middle-income countries: a systematic scoping review

Factors Associated with Ophthalmology Referral and Adherence in a Teleretinal Screening Program: Insights from a Federally Qualified Health Center

Spontaneous electroretinogram signals to screen people with early risk factors for diabetic retinopathy

Contact Info

Product

Resources

About