Kevser Sak
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Kevser Sak has a total of 33 co-authors. Here are the top co-authors by number of shared publications:
- Ugurcan Sayili
Department of Public Health, Cerrahpasa Faculty of Medicine, Istanbul University-Cerrahpasa, 34098 Istanbul, Turkey
3 shared publications - Busra Kara
Department of Public Health, Cerrahpasa Faculty of Medicine, Istanbul University-Cerrahpasa, Istanbul, Türkiye
2 shared publications - Günay Can
Department of Public Health, Cerrahpasa Medical Faculty, Istanbul University-Cerrahpasa, Istanbul, Turkey
2 shared publications - Abdulkerim Uygur
Cerkezkoy District Health Diroctorate, 59500 Tekirdag, Türkiye
1 shared publication Abdülkerim UygurAffiliationsİstanbul Üniversitesi Cerrahpaşa Tıp Fakültesi
1 shared publication- Aysem Kaya
Department of Biochemistry , Institute of Cardiology, Istanbul University-Cerrahpaşa , Istanbul , Türkiye
1 shared publication
Latest Publications
- 8 June 2026
- Published by MediHealth Academy in Journal of Cardiology & Cardiovascular Surgery
- Vol. 4 (2) , 28-35
Comprehensive comparison of Turkish hypertension patient education texts generated by large language models: a cross-sectional content analysis
Aims: This study aimed to compare Turkish patient education texts on hypertension generated by large language models in terms of quality, understandability, actionability, readability, guideline adherence, and patient safety. Methods: In a cross-sectional descriptive content analysis design, standard questions covering eight core domains of hypertension patient education were posed to ChatGPT, Gemini, and Claude using the same Turkish prompt. For each model, only the first response from a separate new chat session was analyzed, yielding a total of 24 texts. The quality of responses containing treatment or self-management information was evaluated using DISCERN, while the understandability and actionability of all texts were assessed with PEMAT-P, and readability was evaluated using the Ateşman formula. In addition, adherence to the 2025 AHA/ACC hypertension guideline was scored, and patient safety was classified dichotomously. Friedman, Wilcoxon, and Cochran’s Q tests were used for statistical analysis. Results: No significant difference was found among the models in total DISCERN scores (p=0.066). There was an overall difference in the DISCERN overall quality score (p=0.022), but no significant pairwise difference was identified in post hoc comparisons. PEMAT-P understandability scores differed significantly among the models (p=0.004); Gemini and Claude scored higher than ChatGPT, and the ChatGPT–Gemini difference remained significant in post hoc analysis (p=0.012). PEMAT-P actionability scores were similar across models (p=0.472). For Ateşman readability scores, Claude scored higher than both ChatGPT and Gemini (p=0.030; p=0.017 for both pairwise comparisons). No significant difference was found in guideline adherence (p=0.717) or patient safety (p=0.135). Conclusion: Large language models may be useful for generating Turkish hypertension patient education materials; however, their performance varies according to the evaluation domain. While differences were observed among models in understandability and readability, no clear superiority was demonstrated in guideline adherence or patient safety. Therefore, LLM-generated outputs should be used under physician supervision, particularly with respect to clinical content and safety.
- 8 June 2026
- Published by MediHealth Academy in Journal of Cardiology & Cardiovascular Surgery
- Vol. 4 (2) , 28-35
Comprehensive comparison of Turkish hypertension patient education texts generated by large language models: a cross-sectional content analysis
Aims: This study aimed to compare Turkish patient education texts on hypertension generated by large language models in terms of quality, understandability, actionability, readability, guideline adherence, and patient safety. Methods: In a cross-sectional descriptive content analysis design, standard questions covering eight core domains of hypertension patient education were posed to ChatGPT, Gemini, and Claude using the same Turkish prompt. For each model, only the first response from a separate new chat session was analyzed, yielding a total of 24 texts. The quality of responses containing treatment or self-management information was evaluated using DISCERN, while the understandability and actionability of all texts were assessed with PEMAT-P, and readability was evaluated using the Ateşman formula. In addition, adherence to the 2025 AHA/ACC hypertension guideline was scored, and patient safety was classified dichotomously. Friedman, Wilcoxon, and Cochran’s Q tests were used for statistical analysis. Results: No significant difference was found among the models in total DISCERN scores (p=0.066). There was an overall difference in the DISCERN overall quality score (p=0.022), but no significant pairwise difference was identified in post hoc comparisons. PEMAT-P understandability scores differed significantly among the models (p=0.004); Gemini and Claude scored higher than ChatGPT, and the ChatGPT–Gemini difference remained significant in post hoc analysis (p=0.012). PEMAT-P actionability scores were similar across models (p=0.472). For Ateşman readability scores, Claude scored higher than both ChatGPT and Gemini (p=0.030; p=0.017 for both pairwise comparisons). No significant difference was found in guideline adherence (p=0.717) or patient safety (p=0.135). Conclusion: Large language models may be useful for generating Turkish hypertension patient education materials; however, their performance varies according to the evaluation domain. While differences were observed among models in understandability and readability, no clear superiority was demonstrated in guideline adherence or patient safety. Therefore, LLM-generated outputs should be used under physician supervision, particularly with respect to clinical content and safety.
- Open Access
- 1 January 2026
- Published by Walter de Gruyter GmbH in Open Medicine
- Vol. 21 (1)
Impact of lipid profiles and risk factors on all-cause mortality in males and females: insights from a Turkish cohort study
Objectives: This study examined associations between lipid profiles, anthropometric measures, and all-cause mortality in a Turkish adult cohort, stratified by sex. Methods: This prospective study analyzed 2,378 participants from the TEKHARF cohort over a median follow-up of 256 months (up to 25 years). All-cause mortality was evaluated using sex-specific Cox proportional hazards regression models and Restricted Cubic Spline analysis to assess non-linear associations with baseline lipid parameters, anthropometric measures, and clinical risk factors. Results: During follow-up, 811 deaths (34.1 %) occurred, with significantly higher mortality in males vs. females (37.5 vs. 30.6 %, p<0.001). Low income, physical inactivity, hypertension, and diabetes were strongly associated with mortality in both sexes. In multivariate models, age, hypertension, and diabetes emerged as the most robust independent predictors of mortality. Although total cholesterol and LDL-C quartiles showed associations with mortality in unadjusted analyses, these became non-significant after adjustment. In males, HDL-C in the third quartile (Q3) was associated with a lower risk of mortality compared to the lowest quartile (HR: 0.759, p=0.045), whereas the highest triglyceride quartile independently predicted mortality only in males (HR: 1.464, p=0.008). Conclusions: Traditional cardiometabolic risk factors, particularly age, hypertension, and diabetes, were the dominant mortality determinants. Notably, in males, triglycerides demonstrated a sex-specific independent effect on mortality risk, while intermediate HDL-C levels were associated with reduced mortality, emphasizing the importance of comprehensive risk assessment integrating lipid parameters with established clinical factors.
- 1 January 2026
- Published by Walter de Gruyter GmbH in Open Medicine
- Vol. 21 (1)
Open AccessImpact of lipid profiles and risk factors on all-cause mortality in males and females: insights from a Turkish cohort study
Objectives: This study examined associations between lipid profiles, anthropometric measures, and all-cause mortality in a Turkish adult cohort, stratified by sex. Methods: This prospective study analyzed 2,378 participants from the TEKHARF cohort over a median follow-up of 256 months (up to 25 years). All-cause mortality was evaluated using sex-specific Cox proportional hazards regression models and Restricted Cubic Spline analysis to assess non-linear associations with baseline lipid parameters, anthropometric measures, and clinical risk factors. Results: During follow-up, 811 deaths (34.1 %) occurred, with significantly higher mortality in males vs. females (37.5 vs. 30.6 %, p<0.001). Low income, physical inactivity, hypertension, and diabetes were strongly associated with mortality in both sexes. In multivariate models, age, hypertension, and diabetes emerged as the most robust independent predictors of mortality. Although total cholesterol and LDL-C quartiles showed associations with mortality in unadjusted analyses, these became non-significant after adjustment. In males, HDL-C in the third quartile (Q3) was associated with a lower risk of mortality compared to the lowest quartile (HR: 0.759, p=0.045), whereas the highest triglyceride quartile independently predicted mortality only in males (HR: 1.464, p=0.008). Conclusions: Traditional cardiometabolic risk factors, particularly age, hypertension, and diabetes, were the dominant mortality determinants. Notably, in males, triglycerides demonstrated a sex-specific independent effect on mortality risk, while intermediate HDL-C levels were associated with reduced mortality, emphasizing the importance of comprehensive risk assessment integrating lipid parameters with established clinical factors.
- Open Access
- 21 October 2025
- Published by Springer Nature in BMC Public Health
- Vol. 25 (1) , 1-10
Do scales measuring internet addiction give compatible results? Comparison of two internet addiction scales among medical students
Internet addiction; excessive and uncontrolled use of the internet is a global problem that affects different segments of the population at different levels. There are many measurement tools in the literature that assess individuals’ internet addiction. The aim of this study was to analyze the agreement and compare the characteristics and results of the two most commonly used scales developed by Young and Chen to measure internet addiction. This cross-sectional study was conducted between March 2023 and June 2023 among 5th- and 6th-year medical school students. The questionnaire consisted of a section that included questions about the sociodemographic and life characteristics of the students and the Young Internet Addiction Test (IAT) and the Chen Internet Addiction Scale (CIAS-R). The prevalence of Internet addiction in 244 medical students was 15% according to the Young IAT with 50 as the cutoff, 56% according to the Young IAT with 30 as the cutoff and 11% according to the CIAS-R. There was moderate agreement (kappa coefficient: 0.50) between the CIAS-R and Young IAT (50); there was slight agreement (kappa coefficient: 0.17) between the CIAS-R and Young IAT (30). In addition, there was no significant difference between the Young IAT (50) and CIAS-R results. There was a significant, strong correlation (r = 0.81) between the Young IAT (50) and the CIAS-R. Although there was moderate agreement between the Young IAT (50) and CIAS, the agreement between the Young IAT (30) and CIAS-R was poor. This study showed that the Young IAT and CIAS-R, which are most commonly used for determining internet addiction, did not function in as high agreement as expected. We think that the selection of the population to be applied in new scale development studies will contribute to the importance of selecting people with clinical diagnoses and generalizable to the society.
- 21 October 2025
- Published by Springer Nature in BMC Public Health
- Vol. 25 (1) , 1-10
Open AccessDo scales measuring internet addiction give compatible results? Comparison of two internet addiction scales among medical students
Internet addiction; excessive and uncontrolled use of the internet is a global problem that affects different segments of the population at different levels. There are many measurement tools in the literature that assess individuals’ internet addiction. The aim of this study was to analyze the agreement and compare the characteristics and results of the two most commonly used scales developed by Young and Chen to measure internet addiction. This cross-sectional study was conducted between March 2023 and June 2023 among 5th- and 6th-year medical school students. The questionnaire consisted of a section that included questions about the sociodemographic and life characteristics of the students and the Young Internet Addiction Test (IAT) and the Chen Internet Addiction Scale (CIAS-R). The prevalence of Internet addiction in 244 medical students was 15% according to the Young IAT with 50 as the cutoff, 56% according to the Young IAT with 30 as the cutoff and 11% according to the CIAS-R. There was moderate agreement (kappa coefficient: 0.50) between the CIAS-R and Young IAT (50); there was slight agreement (kappa coefficient: 0.17) between the CIAS-R and Young IAT (30). In addition, there was no significant difference between the Young IAT (50) and CIAS-R results. There was a significant, strong correlation (r = 0.81) between the Young IAT (50) and the CIAS-R. Although there was moderate agreement between the Young IAT (50) and CIAS, the agreement between the Young IAT (30) and CIAS-R was poor. This study showed that the Young IAT and CIAS-R, which are most commonly used for determining internet addiction, did not function in as high agreement as expected. We think that the selection of the population to be applied in new scale development studies will contribute to the importance of selecting people with clinical diagnoses and generalizable to the society.
- Open Access
- 21 August 2024
- Published by Springer Nature in Discover Public Health
- Vol. 21 (1) , 1-13
Development, validity and reliability of the healthy lifestyle behavior scale
- 21 August 2024
- Published by Springer Nature in Discover Public Health
- Vol. 21 (1) , 1-13
Open AccessDevelopment, validity and reliability of the healthy lifestyle behavior scale
- 18 December 2023
- Published by Springer Nature in Langenbeck's Archives of Surgery
- Vol. 409 (1) , 1-7
Evaluation of the ACS NSQIP surgical risk calculator in patients undergoing common bile duct exploration
Purpose The American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) surgical risk calculator is a risk stratification tool to help predict risks of postoperative complications, which is important for informed decision-making. The purpose of this study was to evaluate the accuracy of the calculator in predicting postoperative complications in patients undergoing common bile duct (CBD) exploration. Methods A retrospective chart review was completed for 305 patients that underwent open and laparoscopic CBD exploration at a single institution from 2010 to 2018. Patient demographics and preoperative risk factors were entered into the calculator, and the predicted complication risks were compared with observed complication rates. Brier score, C-statistic, and Hosmer–Lemeshow regression analysis were used to assess discrimination and calibration. Results The observed rate exceeded the predicted rate for any complication (35.1% vs. 21%), return to operating room (5.9% vs. 3.6%), death (3.3% vs. 1%), and sepsis (3% vs. 2.4%). The model performed best in predicting serious complication (Brier 0.087, C-statistic 0.818, Hosmer–Lemeshow 0.695), surgical site infection (Brier 0.068, C-statistic 0.670, Hosmer–Lemeshow 0.292), discharge to rehabilitation facility (Brier 0.041, C-statistic 0.907, Hosmer–Lemeshow 0.638), and death (Brier 0.028, C-statistic 0.898, Hosmer–Lemeshow 0.004). In multivariable analysis, there was no statistically significant predicted complication type that affected the type of surgery. Conclusion The calculator was accurate in predicting serious complication, surgical site infection, discharge to rehabilitation facility, and death. However, the model displayed poor predictive ability in all other complications that were analyzed.
- 18 December 2023
- Published by Springer Nature in Langenbeck's Archives of Surgery
- Vol. 409 (1) , 1-7
Evaluation of the ACS NSQIP surgical risk calculator in patients undergoing common bile duct exploration
Purpose The American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) surgical risk calculator is a risk stratification tool to help predict risks of postoperative complications, which is important for informed decision-making. The purpose of this study was to evaluate the accuracy of the calculator in predicting postoperative complications in patients undergoing common bile duct (CBD) exploration. Methods A retrospective chart review was completed for 305 patients that underwent open and laparoscopic CBD exploration at a single institution from 2010 to 2018. Patient demographics and preoperative risk factors were entered into the calculator, and the predicted complication risks were compared with observed complication rates. Brier score, C-statistic, and Hosmer–Lemeshow regression analysis were used to assess discrimination and calibration. Results The observed rate exceeded the predicted rate for any complication (35.1% vs. 21%), return to operating room (5.9% vs. 3.6%), death (3.3% vs. 1%), and sepsis (3% vs. 2.4%). The model performed best in predicting serious complication (Brier 0.087, C-statistic 0.818, Hosmer–Lemeshow 0.695), surgical site infection (Brier 0.068, C-statistic 0.670, Hosmer–Lemeshow 0.292), discharge to rehabilitation facility (Brier 0.041, C-statistic 0.907, Hosmer–Lemeshow 0.638), and death (Brier 0.028, C-statistic 0.898, Hosmer–Lemeshow 0.004). In multivariable analysis, there was no statistically significant predicted complication type that affected the type of surgery. Conclusion The calculator was accurate in predicting serious complication, surgical site infection, discharge to rehabilitation facility, and death. However, the model displayed poor predictive ability in all other complications that were analyzed.
- Open Access
- 30 September 2023
- Published by Doc Design and Informatics Co. Ltd. in Klimik Dergisi/Klimik Journal
- Vol. 36 (3) , 208-215
Morbidity and Mortality in Immunocompromised COVID-19 Patients: Which Factors Are Responsible?
Objective: The COVID-19 pandemic has caused significant morbidity and mortality and has challenged health systems worldwide. Immunosuppression is considered a risk factor for development of severe COVID-19. In this study, we aimed to determine the factors affecting morbidity and mortality in immunocompromised patients diagnosed with COVID-19. Methods: In this retrospective, cross-sectional study, laboratory findings and outcomes of 340 COVID-19 patients with secondary immunosuppression in a university hospital in Turkey from March 2020 to April 2022 were analyzed according to the 28-day clinical outcomes. Mortality-related factors were investigated in 3 patient groups with different forms of immunodeficiency: 1) malignancy (hematologic/solid), 2) organ transplantation, and 3) rheumatological diseases. Results: 172 (51%) of the patients were female, and the mean age was 61±15 years. A total of 122 (35.8%) patients died during hospitalization. The mortality rate (73.4%) in patients admitted to the intensive care unit (ICU) was significantly higher than those not admitted (pppp=0.029) and nasal oxygen support (p<0.001) led to a decrease in the risk of mortality. Conclusion: Immunocompromised patients are at risk for serious COVID-19 complications. LDH, D-dimer, CRP, and lymphocyte values can be used as markers to evaluate survival, and standard treatment is essential in immunosuppressive COVID-19 patients.
- 30 September 2023
- Published by Doc Design and Informatics Co. Ltd. in Klimik Dergisi/Klimik Journal
- Vol. 36 (3) , 208-215
Open AccessMorbidity and Mortality in Immunocompromised COVID-19 Patients: Which Factors Are Responsible?
Objective: The COVID-19 pandemic has caused significant morbidity and mortality and has challenged health systems worldwide. Immunosuppression is considered a risk factor for development of severe COVID-19. In this study, we aimed to determine the factors affecting morbidity and mortality in immunocompromised patients diagnosed with COVID-19. Methods: In this retrospective, cross-sectional study, laboratory findings and outcomes of 340 COVID-19 patients with secondary immunosuppression in a university hospital in Turkey from March 2020 to April 2022 were analyzed according to the 28-day clinical outcomes. Mortality-related factors were investigated in 3 patient groups with different forms of immunodeficiency: 1) malignancy (hematologic/solid), 2) organ transplantation, and 3) rheumatological diseases. Results: 172 (51%) of the patients were female, and the mean age was 61±15 years. A total of 122 (35.8%) patients died during hospitalization. The mortality rate (73.4%) in patients admitted to the intensive care unit (ICU) was significantly higher than those not admitted (pppp=0.029) and nasal oxygen support (p<0.001) led to a decrease in the risk of mortality. Conclusion: Immunocompromised patients are at risk for serious COVID-19 complications. LDH, D-dimer, CRP, and lymphocyte values can be used as markers to evaluate survival, and standard treatment is essential in immunosuppressive COVID-19 patients.