Prediction of Diabetes Using Data Mining and Machine Learning Algorithms: A Cross-Sectional Study.

OBJECTIVES: This study aimed to develop a model to predict fasting blood glucose status using machine learning and data mining, since the early diagnosis and treatment of diabetes can improve outcomes and quality of life. METHODS: This crosssectional study analyzed data from 3376 adults over 30 years old at 16 comprehensive health service centers in Tehran, Iran who participated in a diabetes screening program. The dataset was balanced using random sampling and the synthetic minority over-sampling technique (SMOTE). The dataset was split into training set (80%) and test set (20%). Shapley values were calculated to select the most important features. Noise analysis was performed by adding Gaussian noise to the numerical features to evaluate the robustness of feature importance. Five different machine learning algorithms, including CatBoost, random forest, XGBoost, logistic regression, and an artificial neural network, were used to model the dataset. Accuracy, sensitivity, specificity, accuracy, the F1-score, and the area under the curve were used to evaluate the model. RESULTS: Age, waist-to-hip ratio, body mass index, and systolic blood pressure were the most important factors for predicting fasting blood glucose status. Though the models achieved similar predictive ability, the CatBoost model performed slightly better overall with 0.737 area under the curve (AUC). CONCLUSIONS: A gradient boosted decision tree model accurately identified the most important risk factors related to diabetes. Age, waist-to-hip ratio, body mass index, and systolic blood pressure were the most important risk factors for diabetes, respectively. This model can support planning for diabetes management and prevention.

Motivating non-physician health workers to reduce the behavioral risk factors of non-communicable diseases in the community: a field trial study.

BACKGROUND: Non-communicable diseases behavioral risk factors can be improved if effective interventions are designed considering the health system's capabilities and local resources. This study evaluated the effectiveness of interventions that aimed at increasing non-physician community health workers' motivation in reducing non-communicable diseases behavioral risk factors in the community. METHODS: A randomized field trial study was conducted in 32 community health centers in 4 Iranian districts after a baseline population survey on the status of NCDs of 30-70-year-old individuals (n = 1225). The interventions were performed to improve insufficient physical activity, insufficient fruit consumption, insufficient vegetable consumption, high salt intake, and tobacco use. Four intervention packages were implemented in 24 community health centers; the other 8 centers were used as control groups. The non-physician community health workers performed the interventions. The packages additively included goal-setting, evidence-based education, operational planning, and incentive payments. A second survey was conducted 1 year after the start of the interventions to identify the effects on an independent random sample of 30-70-year-old individuals (n = 1221). Difference-in-difference method was used to quantify the interventions' effects. RESULTS: The average age of participants in both surveys was about 49 years. Also, about half of the participants were female, and about 43% were illiterate or had a primary school education. The interventions had statistically significant effects only on decreasing the prevalence of insufficient physical activity. The package with all the intervention components decreased the odds of insufficient physical activity to 0.24 (95% CI, 0.08, 0.72). The package with operational planning but no performance-based financing did not change the odds of insufficient physical activity. CONCLUSIONS: This study highlighted the importance of components, design, and implementation details of interventions intended to reduce NCDs behavioral risk factors. Some risk factors, such as insufficient physical activity, seem more easily modifiable with limited low-cost interventions in a one-year horizon. However, risk factors related to healthy food consumption and tobacco use need more extensive interventions. TRIAL REGISTRATION: This trial was registered on the Iranian Registry of Clinical Trials (IRCT20081205001488N2) on 3 June 2018 ( https://en.irct.ir/trial/774 ).

The impact of intraoperative radiotherapy on breast cancer: focus on the levels of angiogenic factors.

OBJECTIVE: Angiogenesis is one of the hallmarks of cancers that is involved in tumor progression. Angiogenic factors induce the formation of new blood vessels and tumor extension, and finally reduce the survival of patients. Intraoperative radiotherapy (IORT), in which radiation is delivered to the tumor bed can kill cells and change tumor microenvironment. Here, we compared the impact of IORT on the levels of angiogenic factors in the blood and surgical wound fluids (SWF) of the breast cancer patients. PATIENTS AND METHODS: Three hundred sixty patients, who had undergone breast-conserving surgery between 2013 and 2018, were enrolled in IORT and non-IORT groups non-randomly. Blood and drained wound fluid (WF) samples were collected from the patients before and after surgery, followed by quantification of the amounts of TGF-ß, EGF, FGF, VEGF, and DLL4 in the patients using ELISA. RESULTS: Our results were indicative of significant differences between the pre-surgery and post-surgery serum levels of EGF, DLL4, and VEGF. Furthermore, ROC analyses showed that TGF-ß and DLL4 can differentiate of the early-stage from late-stage of the disease. Interestingly, the rate of the death and recurrence was reduced in IORT group. CONCLUSIONS: In summary, IORT is a safe and effective treatment that can affect angiogenic factors and improve the overall- and recurrence-free survival of breast cancer patients.

Adverse effects following COVID-19 vaccination in Iran.

BACKGROUND: Vaccination is a key intervention to prevent COVID-19. Many vaccines are administered globally, yet there is not much evidence regarding their safety and adverse effects. Iran also faces this challenge, especially as data regarding the Sputnik V vaccine is sparse. Therefore, the aim of this study is to determine the adverse effects of the most commonly used vaccines in Iran. METHODS: Using a retrospective cohort study design, 6600 subjects aged 18 years or older who had received two doses of any of the three COVID-19 vaccines (Sinopharm, AstraZeneca, and Sputnik V) were selected using a random sampling method between March and August 2021. Subjects were asked about any adverse effects of the vaccines by trained interviewers via telephone interview. Vaccine-related adverse effects in individuals during the first 72 h and subsequently following both doses of the vaccines were determined. The demographic variables, type of administered vaccine, adverse effects, and history of the previous infection with COVID-19 were collected. Descriptive statistics (mean, standard deviation) and analytical statistics (Chi-squared and Wilcoxon tests) were performed at a 95% significance level using STATA software version 15 (STATA Corp, College Station, TX, USA). RESULTS: From 6600 participants, 4775 responded (response rate = 72.3%). Of the participants, 1460 (30.6%) received the AstraZeneca vaccine, 1564 (32.8%) received the Sinopharm vaccine and 1751 (36.7%) received the Sputnik V vaccine. 2653 participants (55.56%) reported adverse effects after the first dose and 1704 (35.7%) after the second dose. Sputnik V caused the most adverse effects with 1449 (82.7%) vaccine recipients reporting symptoms after the first or second dose, compared with 1030 (70.5%) for AstraZeneca and only 585 (37.4%) for the Sinopharm vaccine. The most common adverse effects after the first dose were fatigue (28.37%), chill/fever (26.86%), and skeletal pain (22.38%). These three adverse effects were the same for the second dose, although their prevalence was lower. CONCLUSIONS: In this study, we demonstrate that the Sputnik V vaccine has the highest rate of adverse effects, followed by the AstraZeneca and Sinopharm vaccines. COVID-19 vaccines used in Iran are safe and there were no reports of serious adverse effects.

Evaluating the impact of results-based motivating system on noncommunicable diseases risk factors in Iran: Study protocol for a field trial.

Background: Risk factors of noncommunicable diseases (NCD) are increasingly contributing to morbidity and mortality in Iran. Health care providers' competencies and motivation are essential factors for the success and efficiency of primary health care. This field trial aims to evaluate the impact of a results-based motivating system on population level of the NCD risk factors field trial (IRPONT) in Iran. Methods: Population groups of 24 rural or urban catchment areas from 3 provinces were randomized to 1 of the 4 types of study groups. The groups were defined based on a set of 4 intervention packages. Extra 8 rural or urban catchment areas in a separate city were considered as independent nonintervention (control) group. Population levels of major NCD risk factors in all 32 population groups were measured at the beginning of the trial, at the end of the first year, and will be measured in the second year through standardized population surveys. As the outcome measure, the difference in population levels of the risk factors will be compared among the study groups. Study group IV will be compared with combined control groups (study groups I, II, and III). Also, we will conduct subgroup analysis to determine the effects of interventions 2, 3, and 4. Ethics: This trial has received ethical approval from National Institute for Medical Research Development in Iran (IR.NIMAD.REC.1396.084) in 2017. Trial Registration Number: This trial has been registered on the Iranian Registry of Clinical Trials (identifier: IRCT20081205001488N2). Registered on 3 June 2018 and updated on 12 April 2020.

Geo-epidemiological reporting and spatial clustering of the 10 most prevalent cancers in Iran.

Cancer is a problem of both global and local concern. We determined the geo-epidemiological and spatial distribution of the 10 most common cancers in Iran. We used the data of the Iranian Cancer Registry for the year 2014 analysing the prevalence of 112,131 registered cancer cases with the aim of detecting potential geographical underlying causes. The geographic distribution of cancers is reported as standardized incidence rates at the provincial level considering risk with respect to sex and age. A geographical information systems (GIS) approach based on Anselin Local Moran's index method was used to map clusters and spatial autocorrelation patterns. The mean age of the patients was 55.6 (±17.8) and 61.7 (±18.2) for females and males, respectively, in the database which showed 46.1% (n=51,665) of all cases to be female. Analysis of the spatial distribution of cancers showed significant differences among the different provinces. Stomach and breast cancers were the most prevalent cancers in men and females, respectively. The highest incidence rates of stomach cancer were found in Ardabil and Zanjan provinces, with 48.38 and 48.08 per 100,000 population, respectively, while Tehran and Yazd provinces had the highest incidences of breast cancer, 51.0 and 47.5 per 100,000 population, respectively. Strong clustering patterns for stomach and breast cancers were identified in the north-western provinces and in Semnan Province, respectively. These patterns indicate a diversity of geo-epidemiological contributing factors to cancer incidence in Iran.

Population-Based Approaches to Prevent Domestic Violence against Women Using a Systematic Review.

Objective: In this systematic review, we aimed to evaluate the existing strategies and interventions in domestic violence prevention to assess their effectiveness. Method : To select studies, Pubmed, ISI, CINAHL, PsycINFO, Cochrane, Scopus, Embase, Ovid, Science Direct, ProQuest, and Elsevier databases were searched. Two authors reviewed all papers using established inclusion/ exclusion criteria. Finally, 18 articles were selected and met the inclusion criteria for assessment. Following the Cochrane quality assessment tool and AHRQ Standards, the studies were classified for quality rating based on design and performance quality. Two authors separately reviewed the studies and categorized them as good, fair, and poor quality. Results: Most of the selected papers had fair- or poor-quality rating in terms of methodology quality. Different intervention methods had been used in these studies. Four studies focused on empowering women; 3, 4, and 2 studies were internet-based interventions, financial interventions, and relatively social interventions, respectively. Four interventions were also implemented in specific groups. All authors stated that interventions were effective. Conclusion: Intervention methods should be fully in line with the characteristics of the participants. Environmental and cultural conditions and the role of the cause of violence are important elements in choosing the type of intervention. Interventions are not superior to each other because of their different applications.

How Much Hypertension is Attributed to Overweight, Obesity, and Hyperglycemia Using Adjusted Population Attributable Risk in Adults?

BACKGROUND: High blood pressure (HBP) is a proven risk factor for cardiovascular diseases. So, determining the extent of the contribution of the factors associated with HBP seems to be necessary. Accordingly, this study aimed to investigate how much the prevalence of HBP attributed to obesity and high blood glucose (HBG). METHODS: Data were collected from 7612 participants extracted from a screening program in 2018, in Iran, which was conducted on the subjects with the age of 30 years old and older to investigate the prevalence of HBP and their associated risk factors. To collect data, we used a standard checklist in terms of the WHO STEPS manual, and a stratified multistage sampling method was also applied. The adjusted population attributable risk of overweight, obesity, and HBG for HBP was calculated by the logistic regression model using the aflogit module. RESULTS: Among the studied people, 7.4% of male and 10.8% of female subjects were hypertensive. The adjusted analysis showed that, in men, 27% and 41% and, among women, 19% and 37% of HBP prevalence rates were attributable to obesity (BMI ≥ 30) and fast blood sugar (FBS) (≥126), respectively. In people with both obesity and HBG, 59% of the prevalence rate of HBP in men and 46% of the prevalence in women were due to the abovementioned risk factors altogether. The results show that, if obesity and HBG were eliminated, the prevalence of HBP could be theoretically decreased from 7.4% to 5.4% and 4.3% in male subjects and from 10.8% to 8.7% and 6.8% in female subjects, respectively. CONCLUSIONS: Our findings indicate that how much the prevalence of HBP attributes to obesity and HBG in middle-age and older population. It seems that the prevention programs should be administered in the general population, and excess body weight prevention programs should also be implemented in childhood.

Stillbirth in Iran and associated factors (2014-2016): A population-based study.

Background: Under Every Newborn Action Plan (ENAP), stillbirth rate in every country should be reduced by 12 or fewer per 1000 total births by 2030. The aims of this study were to determine stillbirth rate at national and subnational levels and to investigate its associated risk factors in Iran. Methods: Using all data from Iranian Maternal and Neonatal Network (IMaN), we calculated stillbirth rate of Iran from 2014-2016. This network registers information of almost all births across the country. The logistic regression was used to estimate the adjusted odds ratio (aOR) with 95% confidence intervals (CIs) for stillbirth. Results: In 2014, still birth rate was 7.40 per 1000 births. In 2015 and 2016, stillbirth rates were 7.22 per 1000 births and 7.63 per 1000 births, respectively. The most important related factors of stillbirth were preterm birth (aOR= 62.53, 95% CI; 60.77-64.34), sexual ambiguity (aOR= 14.51, 95% CI; 12.76-16.50), and post term birth (aOR= 3.31, 95% CI; 2.66-4.13). Conclusion: Under Every Newborn Action Plan (ENAP), stillbirth rate in every country should be reduced by 12 or fewer per 1000 total births by 2030. Iran has achieved stillbirth target of ENAP at national level. It is important for the health care system to establish and improve specific and focused policies, interventions, and programs for achieving this target even in the most deprived areas.

Early Cessation of Breastfeeding and Determinants: Time to Event Analysis.

BACKGROUND: The onset of breastfeeding has a high success rate in most countries, but the time for termination of breastfeeding varies between countries. OBJECTIVE: This survey was aimed to determine the effective factors on the early termination of breastfeeding. METHODS: This study was conducted in 2018, in Iran. About 410 mothers were enrolled in the study. All considered factors were evaluated as factors influencing the continuity of breastfeeding. Survival analysis was used to analyze data. RESULTS: The mean age of the mothers was equal to 29.48 ± 5.8 years. The frequency of termination of breastfeeding before the first 2 years was equal to 34%. The mean of breastfeeding duration was equal to 21.49 ± 5.3 months. The percentage of infants who had been breastfed for 24 months was equal to 65.8%. An infant's birth weight (2500-4000 gr) (hazard ratio: 0.54), neonatal birth order (hazard ratio: 0.69), neonatal pathologic jaundice (hazard ratio: 1.52), starting time of using complementary food (hazard ratio: 2.45), using pacifier (hazard ratio: 2.82), and the status of using artificial milk (hazard ratio: 3.29) were significantly associated with cessation of breastfeeding before 24 months of age. The probability of termination of breastfeeding at 6, 12, 18, and 24 months of age was reported by 6%, 8%, 15%, and 34%, respectively. CONCLUSIONS: There were notifiable variations in breastfeeding rates both in national and international levels. Nevertheless, in this study, the mean of breastfeeding duration was longer compared to a number of countries and previous national studies.

Early estimation of the epidemiological parameters of novel coronavirus disease (COVID-2019) outbreak in Iran: 19 Feb-15 March, 2020.

AIM: To estimate the epidemiological parameters related to the Covid-19 outbreak in Iran. BACKGROUND: Estimating the epidemiological parameters of new public health threat (COVID-19) is essential to support and inform public health decision-making in different communities including Iran. METHODS: We established a mathematical model to estimate the epidemiological parameters from 19 Feb to 15 March based on daily COVID-19 confirmed cases in Iran. Then, we estimated the effect of early traffic restriction on our estimation. RESULTS: We estimated the R0 at 2.11 (95% CI, 1.87-2.50) and the infected number at 92,260 (95% CI: 59,263 -152,212) by 15 March. Our estimate for the ascertainment rate was about 1.2% (95% CI: 1.1-1.4). The latent period estimation was 4.24 (95% CI: 2.84-6.65). We observed a decline in our estimate after considering the traffic restriction. CONCLUSION: Our results suggest that health authorities in Iran must take impactful strategies to control the COVID-19 outbreak to reach R0<1. Therefore, the establishment of complementary, multilateral, and cost-effective measures for the treatment of symptomatic and early diagnosis and isolation of asymptomatic cases/contacts are strongly recommended because of low ascertainment rate and large number of infected cases. We additionally recommend that traffic restriction be combined with other controlling measures.

Breast Cancer Survival Analysis Based on Immunohistochemistry Subtypes (ER/PR/HER2): a Retrospective Cohort Study.

BACKGROUND: We conducted this study to estimate the prevalence of biomarkers, including estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER2) among patients with breast cancer and to explore their effects on disease mortality. METHODS: We conducted this registry-based retrospective cohort study in Tehran, in 2014, using the data on 1622 patients with breast cancer, diagnosed pathologically and registered with the Comprehensive Cancer Control Center from 1998 to 2013. The outcome of interest was the survival probability of patients with breast cancer based on receptor status along with other prognostic factors such as age, histopathology, stage/grade of tumor, metastatic status, and surgical procedures using the life table, Kaplan-Meier curves, and multivariate Cox proportional hazard model. We generated different subtypes based on expression of ER, PR, and HER2, positive (+) and/or negative (-). RESULTS: ER+/PR+/HER2- subtype (51.5%) was the most common form of breast cancer cells. Compared to the ER+/PR+/HER- subtype, the hazard ratio (95% confidence interval) of cancer mortality was 2.14 (1.13, 4.03) for ER-/PR-/HER2- subtype, 1.92 (1.03, 3.59) for ER-/PR-/HER2+ subtype and 5.19 (1.51, 17.86) for ER-/PR+/HER2+ subtype. CONCLUSION: In this study, breast cancer cases with ER-/HER2+ tumors had shorter survival than those with ER+/PR+/HER2- tumors. Triple negative tumors were the only other subtype with a statistically significant poorer prognosis. The results of this study in a middle-income country further indicate the importance of receptor status, in particular HER2 status, in the prognosis of breast cancer.