Iranian risk model as a predictive tool for retinopathy in patients with type 2 diabetes

Prediction of diabetic retinopathy among type 2 diabetic patients in University of Gondar Comprehensive Specialized Hospital, 2006-2021: A prognostic model

BACKGROUND

There has been a paucity of evidence for the development of a prediction model for diabetic retinopathy (DR) in Ethiopia. Predicting the risk of developing DR based on the patient's demographic, clinical, and behavioral data is helpful in resource-limited areas where regular screening for DR is not available and to guide practitioners estimate the future risk of their patients.

METHODS

A retrospective follow-up study was conducted at the University of Gondar (UoG) Comprehensive Specialized Hospital from January 2006 to May 2021 among 856 patients with type 2 diabetes (T2DM). Variables were selected using the Least Absolute Shrinkage and Selection Operator (LASSO) regression. The data were validated by 10-fold cross-validation. Four ML techniques (naïve Bayes, K-nearest neighbor, decision tree, and logistic regression) were employed. The performance of each algorithm was measured, and logistic regression was a well-performing algorithm. After multivariable logistic regression and model reduction, a nomogram was developed to predict the individual risk of DR.

RESULTS

Logistic regression was the best algorithm for predicting DR with an area under the curve of 92%, sensitivity of 87%, specificity of 83%, precision of 84%, F1-score of 85%, and accuracy of 85%. The logistic regression model selected seven predictors: total cholesterol, duration of diabetes, glycemic control, adherence to anti-diabetic medications, other microvascular complications of diabetes, sex, and hypertension. A nomogram was developed and deployed as a web-based application. A decision curve analysis showed that the model was useful in clinical practice and was better than treating all or none of the patients.

CONCLUSIONS

The model has excellent performance and a better net benefit to be utilized in clinical practice to show the future probability of having DR. Identifying those with a higher risk of DR helps in the early identification and intervention of DR.

Machine Learning-Based Diagnosis and Ranking of Risk Factors for Diabetic Retinopathy in Population-Based Studies from South India

This paper discusses the importance of investigating DR using machine learning and a computational method to rank DR risk factors by importance using different machine learning models. The dataset was collected from four large population-based studies conducted in India between 2001 and 2010 on the prevalence of DR and its risk factors. We deployed different machine learning models on the dataset to rank the importance of the variables (risk factors). The study uses a t-test and Shapely additive explanations (SHAP) to rank the risk factors. Then, it uses five machine learning models (K-Nearest Neighbor, Decision Tree, Support Vector Machines, Logistic Regression, and Naive Bayes) to identify the unimportant risk factors based on the area under the curve criterion to predict DR. To determine the overall significance of risk variables, a weighted average of each classifier's importance is used. The ranking of risk variables is provided to machine learning models. To construct a model for DR prediction, the combination of risk factors with the highest AUC is chosen. The results show that the risk factors glycosylated hemoglobin and systolic blood pressure were present in the top three risk factors for DR in all five machine learning models when the t-test was used for ranking. Furthermore, the risk factors, namely, systolic blood pressure and history of hypertension, were present in the top five risk factors for DR in all the machine learning models when SHAP was used for ranking. Finally, when an ensemble of the five machine learning models was employed, independently with both the t-test and SHAP, systolic blood pressure and diabetes mellitus duration were present in the top four risk factors for diabetic retinopathy. Decision Tree and K-Nearest Neighbor resulted in the highest AUCs of 0.79 (t-test) and 0.77 (SHAP). Moreover, K-Nearest Neighbor predicted DR with 82.6% (t-test) and 78.3% (SHAP) accuracy.

Prevalence of diabetic retinopathy in the Eastern Mediterranean Region: a systematic review and meta-analysis

Objectives

Individual studies in the Eastern Mediterranean Region (EMR) have shown the high prevalence of diabetic retinopathy. We conducted a meta-analysis to yield an estimate of the prevalence of diabetic (type 1 and 2) retinopathy in the EMR. Additionally, we explored its potential modulators.

Methods

Two-step screening of relevant articles published from 1 January 2000 to 13 December 2019 was carried out. An estimation of summary proportions, subgroup analysis, meta-regression, and publication bias assessment were performed.

Results

One hundred nine articles were included in the meta-analysis, involving 280,566 patients. The prevalence of diabetic retinopathy was 31% (95% confidence interval [CI] = 28, 33). The highest and lowest diabetic retinopathy prevalence rates were observed in low human development index (HDI) countries (63.6; 95% CI = 52.4, 74.0) and very high HDI countries 22.6 (95% CI = 20.5, 24.7), respectively.

Conclusions

The prevalence of diabetic retinopathy is high in the EMR. Our results provide important information for diverse healthcare surveillance systems in the EMR to implement the modifiable risk factors, diabetes screening to decrease undiagnosed diabetes, early detection of retinopathy, and proper diabetes care to decrease untreated diabetes.

Usefulness of Machine Learning for Identification of Referable Diabetic Retinopathy in a Large-Scale Population-Based Study

Purpose: To development and validation of machine learning-based classifiers based on simple non-ocular metrics for detecting referable diabetic retinopathy (RDR) in a large-scale Chinese population–based survey.

Methods: The 1,418 patients with diabetes mellitus from 8,952 rural residents screened in the population-based Dongguan Eye Study were used for model development and validation. Eight algorithms [extreme gradient boosting (XGBoost), random forest, naïve Bayes, k-nearest neighbor (KNN), AdaBoost, Light GBM, artificial neural network (ANN), and logistic regression] were used for modeling to detect RDR in individuals with diabetes. The area under the receiver operating characteristic curve (AUC) and their 95% confidential interval (95% CI) were estimated using five-fold cross-validation as well as an 80:20 ratio of training and validation.

Results: The 10 most important features in machine learning models were duration of diabetes, HbA1c, systolic blood pressure, triglyceride, body mass index, serum creatine, age, educational level, duration of hypertension, and income level. Based on these top 10 variables, the XGBoost model achieved the best discriminative performance, with an AUC of 0.816 (95%CI: 0.812, 0.820). The AUCs for logistic regression, AdaBoost, naïve Bayes, and Random forest were 0.766 (95%CI: 0.756, 0.776), 0.754 (95%CI: 0.744, 0.764), 0.753 (95%CI: 0.743, 0.763), and 0.705 (95%CI: 0.697, 0.713), respectively.

Conclusions: A machine learning–based classifier that used 10 easily obtained non-ocular variables was able to effectively detect RDR patients. The importance scores of the variables provide insight to prevent the occurrence of RDR. Screening RDR with machine learning provides a useful complementary tool for clinical practice in resource-poor areas with limited ophthalmic infrastructure.

Prognostic models of diabetic microvascular complications: a systematic review and meta-analysis

BACKGROUND

Many prognostic models of diabetic microvascular complications have been developed, but their performances still varies. Therefore, we conducted a systematic review and meta-analysis to summarise the performances of the existing models.

METHODS

Prognostic models of diabetic microvascular complications were retrieved from PubMed and Scopus up to 31 December 2020. Studies were selected, if they developed or internally/externally validated models of any microvascular complication in type 2 diabetes (T2D).

RESULTS

In total, 71 studies were eligible, of which 32, 30 and 18 studies initially developed prognostic model for diabetic retinopathy (DR), chronic kidney disease (CKD) and end stage renal disease (ESRD) with the number of derived equations of 84, 96 and 51, respectively. Most models were derived-phases, some were internal and external validations. Common predictors were age, sex, HbA1c, diabetic duration, SBP and BMI. Traditional statistical models (i.e. Cox and logit regression) were mostly applied, otherwise machine learning. In cohorts, the discriminative performance in derived-logit was pooled with C statistics of 0.82 (0.73‑0.92) for DR and 0.78 (0.74‑0.83) for CKD. Pooled Cox regression yielded 0.75 (0.74‑0.77), 0.78 (0.74‑0.82) and 0.87 (0.84‑0.89) for DR, CKD and ESRD, respectively. External validation performances were sufficiently pooled with 0.81 (0.78‑0.83), 0.75 (0.67‑0.84) and 0.87 (0.85‑0.88) for DR, CKD and ESRD, respectively.

CONCLUSIONS

Several prognostic models were developed, but less were externally validated. A few studies derived the models by using appropriate methods and were satisfactory reported. More external validations and impact analyses are required before applying these models in clinical practice.

SYSTEMATIC REVIEW REGISTRATION

PROSPERO CRD42018105287.

The SMART Framework: Selection of Machine learning Algorithms with ReplicaTions - a Case Study on the Microvascular Complications of Diabetes

Over 34 million people in the US have diabetes, a major cause of blindness, renal failure, and amputations. Machine learning (ML) models can predict high-risk patients to help prevent adverse outcomes. Selecting the 'best' prediction model for a given disease, population, and clinical application is challenging due to the hundreds of health-related ML models in the literature and the increasing availability of ML methodologies. To support this decision process, we developed the Selection of Machine-learning Algorithms with ReplicaTions (SMART) Framework that integrates building and selecting ML models with decision theory. We build ML models and estimate performance for multiple plausible future populations with a replicated nested cross-validation technique. We rank ML models by simulating decision-maker priorities, using a range of accuracy measures (e.g., AUC) and robustness metrics from decision theory (e.g., minimax Regret). We present the SMART Framework through a case study on the microvascular complications of diabetes using data from the ACCORD clinical trial. We compare selections made by risk-averse, -neutral, and -seeking decision-makers, finding agreement in 80% of the risk-averse and risk-neutral selections, with the risk-averse selections showing consistency for a given complication. We also found that the models that best predicted outcomes in the validation set were those with low performance variance on the testing set, indicating a risk-averse approach in model selection is ideal when there is a potential for high population feature variability. The SMART Framework is a powerful, interactive tool that incorporates various ML algorithms and stakeholder preferences, generalizable to new data and technological advancements.

Is central obesity associated with diabetic retinopathy in Chinese individuals? An exploratory study

Objective

To our knowledge, the independent association between central obesity, defined by waist circumference (WC) or waist-to-hip ratio (WHR), and diabetic retinopathy (DR) remains unknown in Chinese individuals.

Method

The study was conducted in two stages. First, the relationship between WC or WHR and DR was estimated in a case-control set (DR vs. non-DR) for the whole population before and after propensity score matching. Subsequently, a systematic review and meta-analysis was performed on evidence from the literature to validate the relationship.

Results

Of 511 eligible patients, DR (N = 156) and non-DR (N = 156) patients with similar propensity scores were included in the propensity score matching analyses. Central obesity (defined by WC) was associated with risk of DR (odds ratio [OR] 1.07, 95% confidence interval [95% CI] (1.03–1.10). The meta-analysis showed that central obesity significantly increased the risk of DR by 12% (OR 1.12, 95% CI 1.02–1.22). Analysis of data from 18 studies showed a significant association between continuous body mass index and risk of proliferative DR (OR 0.95, 95% CI 0.93–0.98; I² = 50%).

Conclusion

Central obesity, particularly as defined by WC, is associated with the risk of DR in the Chinese population.

Prevalence, Incidence and Ecological Determinants of Diabetic Retinopathy in Iran: Systematic Review and Meta-analysis

Purpose

To estimate the pooled prevalence and incidence of diabetic retinopathy (DR) in Iran and to investigate their correlations with the Human Development Index (HDI), healthcare access (i.e., density of specialists and sub-specialists), and methodological issues.

Methods

Electronic databases such as PubMed, Embase, Scopus, Web of Science, Google Scholar, and local databases were searched for cohort and cross-sectional studies published prior to January 2018. Prevalence and incidence rates of DR were extracted from January 2000 to December 2017 and random effects models were used to estimate pooled effect sizes. The Joanna Briggs Institute critical appraisal tool was applied for quality assessment of eligible studies.

Results

A total of 55,445 participants across 33 studies were included. The pooled prevalence (95% CI) of DR in diabetic clinics (22 studies), eye clinics (4 studies), and general population (7 studies) was 31.8% (24.5 to 39.2), 57.8% (50.2 to 65.3), and 29.6% (22.6 to 36.5), respectively. It was 7.4% (3.9 to 10.8) for proliferative DR and 7.1% (4.9 to 9.4) for clinically significant macular edema. The heterogeneity of individual estimates of prevalence was highly significant. HDI (P < 0.001), density of specialists (P = 0.004), subspecialists (P < 0.001), and sampling site (P = 0.041) were associated with heterogeneity after the adjustment for type of DR, duration of diabetes, study year, and proportion of diabetics with controlled HbA1C.

Conclusion

Human development and healthcare access were correlated with the prevalence of DR. Data were scarce on the prevalence of DR in less developed provinces. Participant recruitment in eye clinics might overestimate the prevalence of DR.

How to improve Iranians' vision health: on the national policy of preventing Iranians' blindness

AIM

To review vision health situation of Iranian community, analyze its determinants, and discuss the adopted improvement strategies by the Iran Ministry of Health and Medical Education (MOHME).

METHODS

This was a rapid situation analysis with a qualitative approach in three parts of recognition, orientation and implementation. The data were gathered via review of upstream documents, national and international experiences, and experts and stakeholders' opinions.

RESULTS

Eradicating trachoma, increasing human resources, increasing educational and research centers and promotion of ophthalmic technologies were important achievements in the field of vision health in Iran. Through these achievements, it seemed that the pattern of causes of blindness and low vision was similar to that of the developed countries. However, the review of Iranians' vision health indicators showed that a considerable percent of the blindness and low vision was avoidable through a national program demanding 3 types of interventions in social determinants of health (SDH), community education, and increasing the access to health care services by integrating the necessary services in primary health care system.

CONCLUSION

Managing the issue requires attentions from a national committee for preventing blindness with participation of all stakeholders, implementing a national survey on vision health, preparation of the primary level health centers including employment and education of community health workers (Behvarzes), optometrists and general practitioners, fair distribution of specialized human resources and establishing at least one specialized center in each province for referring patients from the primary levels.

High -density lipoprotein cholesterol as a predictor for diabetes mellitus

Background:

Diabetes is a prevalent chronic disease around the world. To evaluate the risk of diabetes comprehensively, we developed a score model for risk prediction with HDL-C as a protective factor.

Methods:

We extracted physical examination data of 2728 individuals. The data contain 18 demographic and clinical variables. To identify the statistical significant feature variables, the backward stepwise logistic regression was used based on the data of the “exploratory population”. To ascertain the cutoff value of the selected variables, we used the Youden index. Then we assigned each variable level a score according to the estimated regression model coefficients and then calculated the individual’s total score. We gained the cutoff value for the total score through the Youden Index and stratified the total score into four levels. We employed the data of “validation population” to test the performance of the score model based on the area under the ROC curve.

Results:

Age, LDL-C, HDL-C, BMI, family history of diabetes, diastolic blood pressure and TCHO were selected as statistically significant variables. The diabetes risk score range varied from 0 to 17. The risk level categorized by the total score was low, middle, high and extremely high, with a score range of 0-2, 3-7, 8-12 and 13-17, respectively.

Conclusions:

The score model based on physical examination data is an efficient and valuable tool to evaluate and monitor the potential diabetes risk for both healthy and unhealthy people at an individual level.

Body mass index and risk of diabetic retinopathy: A meta-analysis and systematic review

Diabetic retinopathy (DR) is a frequent cause of acquired blindness worldwide. Various studies have reported the effects of body mass index (BMI) on the risk of DR, but the results remain controversial. Therefore, a meta-analysis was performed to evaluate the relationship between BMI and the risk of DR.A systematic search was performed using the Cochrane Library, PubMed, and Embase databases to obtain articles published through December 2016. Articles regarding the association between BMI and the risk of DR were retrieved. The adjusted odds ratios (ORs) and their 95% confidence intervals (CIs) were included and then pooled with a random effects model.A total of 27 articles were included in this meta-analysis. When BMI was analyzed as a categorical variable, neither being overweight (OR = 0.89, 95% CI 0.75-1.07; P = .21; I = 65%) nor obesity (OR = 0.97, 95% CI 0.73-1.30; P = .86) were associated with an increased risk of DR when compared with normal weight. When BMI was analyzed as a continuous variable, a higher BMI was not associated with an increased risk of DR (OR = 0.99, 95% CI 0.97-1.01; P = .25; I2 = 79%). The pooled results did not significantly change after the sensitivity analysis.Based on the current publications, neither being overweight nor obesity is associated with an increased risk of DR. Further studies should confirm these findings.

Prevalence of diabetic retinopathy in Iran: a systematic review and Meta-analysis

AIM

To estimate the overall prevalence of diabetic retinopathy (DR) in Iran by a systematic review and Meta-analysis.

METHODS

We conducted a search of all published literature on diabetic patients for the prevalence of DR using Web of Sciences, PubMed, Scopus, Google Scholar, and national electronic databases SID, Magiran, and Iranmedex from their inception until September 2016 with standard keywords. Pooled estimates of the DR prevalence and the corresponding 95% confidence intervals (CI) were calculated using random effects models.

RESULTS

Thirty-one studies involving 23 729 patients with type I and II diabetes were included. The publication bias assumption for prevalence of DR was rejected by Begg and Egger tests (P=0.825, P=0.057, respectively). The results of Cochran test and I² statistics showed considerable heterogeneity for prevalence of DR (Q=1278.21, d.f.=30, P<0.001 and I²=97.7%). The prevalence of DR, non-proliferative diabetic retinopathy (NPDR) and proliferative diabetic retinopathy (PDR) in Iranian diabetic patients were 41.9% (95% CI: 35.6-48.2), 32.2% (95% CI: 28.7-35.8), and 13.2% (95% CI: 8.3-18.1), respectively.

CONCLUSION

The prevalence of DR in Iran appears a little high. NPDR was more common. This study highlights the necessity for DR screening and management in diabetic patients in Iran.

Glycemic Control and the Risk of Tuberculosis: A Cohort Study

BACKGROUND

Diabetes is a well-known risk factor for tuberculosis (TB) and is increasingly prevalent in low- and middle-income countries, where the burden of TB is high. Glycemic control has the potential to modify the risk of TB. However, there are few studies on the association between glycemic control and TB risk, and the results are inconsistent.

METHODS AND FINDINGS

We assembled a cohort using 123,546 individuals who participated in a community-based health screening service in northern Taiwan from 5 March 2005 to 27 July 2008. Glycemic control was measured using fasting plasma glucose (FPG) at the time of screening. The cohort was followed up to 31 December 2012 for the occurrence of TB by cross-matching the screening database to the national health insurance database. Multiple imputation was used to handle missing information. During a median follow-up of 4.6 y, 327 cases of TB occurred. In the multivariable Cox regression model, diabetic patients with poor glycemic control (FPG > 130 mg/dl) had a significantly higher hazard of TB (adjusted hazard ratio [aHR] 2.21, 95% CI 1.63-2.99, p < 0.001) compared to those without diabetes. The hazard of TB in diabetic patients with good glycemic control (FPG ≤ 130 mg/dl) did not differ significantly from that in nondiabetic individuals (aHR 0.69, 95% CI 0.35-1.36, p = 0.281). In the linear dose-response analysis, the hazard of TB increased with FPG (aHR 1.06 per 10-mg/dl increase in FPG, 95% CI 1.03-1.08, p < 0.001). Assuming the observed association between glycemic control and TB was causal, an estimated 7.5% (95% CI 4.1%-11.5%) of incident TB in the study population could be attributed to poor glycemic control. Limitations of the study include one-time measurement of fasting glucose at baseline and voluntary participation in the health screening service.

CONCLUSIONS

Good glycemic control could potentially modify the risk of TB among diabetic patients and may contribute to the control of TB in settings where diabetes and TB are prevalent.