Nevin Manimala

J Health Popul Nutr. 2026 Jul 12. doi: 10.1186/s41043-026-01401-y. Online ahead of print.

ABSTRACT

BACKGROUND: Childhood stunting remains a major public health concern, reflecting chronic undernutrition and long-term socioeconomic disadvantage. Among school-aged children, stunting is associated with impaired physical growth, reduced cognitive development, and poorer educational outcomes. Traditional statistical approaches, such as logistic regression, have been widely used to examine factors associated with stunting; however, their ability to capture complex and nonlinear relationships is limited. Machine learning (ML) methods provide a flexible alternative for modeling such relationships.

OBJECTIVE: This study aimed to model and classify stunting among school-aged children in Ethiopia using school- and household-level data, and to compare the performance of machine learning algorithms with multivariable logistic regression.

METHODS: A cross-sectional analysis was conducted using secondary data from Round 5 (2016-2017) of the Young Lives study in Ethiopia. Stunting was defined as a binary outcome based on World Health Organization height-for-age Z-score criteria. Several machine learning algorithms, including Random Forest, Support Vector Machine, and Gradient Boosting Machine, were implemented. Model performance was evaluated using accuracy, sensitivity, specificity, F1-score, and area under the receiver operating characteristic curve (AUC). Variable importance measures were used to identify predictors contributing to model performance. To avoid potential circularity, anthropometric variables closely related to the outcome (e.g., child weight) were excluded from the final models.

RESULTS: The Random Forest model demonstrated modestly improved performance compared with logistic regression and other machine learning methods, and its performance was evaluated using accuracy, sensitivity, specificity, AUC, and F1-score. Key predictors included school type, household wealth index, literacy-related indicators, and region of residence. Notable regional variation in stunting classification was observed, suggesting the influence of broader socioeconomic and environmental conditions.

CONCLUSION: Machine learning models, particularly Random Forest, showed slightly better performance than conventional logistic regression in classifying stunting among school-aged children in Ethiopia. The identified predictors highlight the multifactorial and context-dependent nature of stunting. These findings support the use of ML approaches as complementary analytical tools for understanding patterns of child undernutrition, although their application for prediction should be interpreted within the limitations of cross-sectional data.

PMID:42437956 | DOI:10.1186/s41043-026-01401-y

Cardiovasc Diabetol Endocrinol Rep. 2026 Jul 13;12(1):40. doi: 10.1186/s40842-026-00309-0.

ABSTRACT

BACKGROUND: Although a cross-sectional relation between the atherogenic index of plasma (AIP) and gout has been well documented, the causal roles of its principal components, TG and HDL-C, have not yet been clearly established. Therefore, this study sought to explicate the causal relations between AIP, its individual constituents, and the risk of gout, as well as to investigate the potential molecular mechanisms underlying these associations.

METHODS: We adopted an integrated analytical pipeline. First, mediation and cross-sectional analyses were carried out to quantify the phenotypic associations of TG, HDL-C, and AIP with gout. Second, both multivariable and univariable Mendelian randomization (MR) analyses were applied to assess causal directions. Finally, a network pharmacology approach was applied to construct interaction networks linking lipid-related factors with gout, and subsequent enrichment analyses were done to identify the main biological pathways involved and key genes were identified using the Icelandic database’s pQTLs.

RESULTS: AIP showed a significant positive link with gout prevalence (OR = 1.700, p = 0.010), with this relationship being mediated mainly by HDL-C (45.19%), rather than TG or LDL-C. Univariable MR analyses indicated that TG exerted a substantial causal effect on gout risk (OR = 1.0058, p < 0.001). Although univariable MR showed a nominally protective association for HDLC (OR = 0.623, 95% CI: 0.399-0.974, p = 0.042), this effect was attenuated and became statistically nonsignificant after multivariable adjustment for other lipid traits (p > 0.05). After adjustment for genetic correlations in multivariable MR analyses, TG remained the sole lipid trait with a robust and independent causal effect on gout (OR = 1.0077, p < 0.001), whereas neither HDL-C nor LDL-C reached statistical significance. The TG-specific gout-associated gene set (Group B) was significantly enriched in T-cell receptor and NF-κB signaling pathways, with hub genes including PTPRC, MYD88, and LCK, supporting a direct lipid-immune axis. By contrast, the combined TG/HDL-C gene set (Group F) showed predominant enrichment in PI3K-Akt and cytokine-cytokine receptor pathways, whereas genes uniquely shared between HDL-C and gout (Group C) were mainly enriched in fundamental cellular processes without marked inflammatory pathway involvement.

CONCLUSIONS: AIP exhibits a non-linear positive association with gout, but this association is primarily mediated by HDL‑C, which itself has no independent causal protective effect. TG may play an independent causal role in gout, potentially involving T‑cell receptor/NF‑κB signaling pathways. However, these findings are exploratory and require further validation.

PMID:42437949 | DOI:10.1186/s40842-026-00309-0

Reprod Biol Endocrinol. 2026 Jul 13;24(1):69. doi: 10.1186/s12958-026-01578-1.

ABSTRACT

BACKGROUND: Donated eggs are essential for a range of Medically Assisted Reproduction (MAR) procedures, but relatively little has been written about how donors are recruited. This study examined egg donor recruitment processes at a commercial egg-bank in Florida (USA). It documents what proportion of applicants were ultimately accepted and, if rejected, at what step, whether this was by choice or selection, whether their initial ID-release choice was important, and how this compares to the recruitment of sperm donors.

METHODS: Anonymised records of all egg donor applicants in 2018 and 2019 (n = 2,443) were examined to determine the number passing through (or lost) at each stage of the recruitment process and ultimately how many had successful egg retrieval and cryopreservation. Statistical analysis was carried out to examine differences between the initial ID-release choice made by egg donor applicants (ID-release vs. non-ID release).

RESULTS: Few applicants (2.5%) were accepted and had eggs frozen for donation. This did not differ between applicants who opted at the outset to be ID-release (2.94%) compared to those who didn’t (2.12%) (X₂ = 1.682; Df = 1; Z = 1.297; p = 0.1947). Most were lost during recruitment because they: (i) did not meet the eligibility criteria at the outset (51.17%); (ii) withdrew, failed to respond, did not attend an appointment, or did not return a questionnaire (26.36%); or (iii) reported a disqualifying health issue or failed a screening test (19.69%). There were no significant differences between the initial ID choice of egg donor candidates and the reason for their loss from the process. This differed from what we know about sperm donor recruitment during the same period at the same clinic. Only two women who were accepted to donate failed to do so because of a poor ovarian response. During recruitment, some egg donors decided to change ID-type and it was more common for them to change from non-ID release to ID release (53.57%) than the other way around (9.09%) (X₂ = 14.920; Df = 1; Z = 3.863; p < 0.0001).

CONCLUSION: This study demonstrates how challenging egg donor recruitment processes are, with only a small fraction of those who initially apply ultimately being accepted and having samples certified as safe for use in treatment. The initial ID-release choice of egg donor applicants in the USA has no bearing on whether they were finally accepted as donors or not and is unrelated to their reason for rejection from the program.

PMID:42437948 | DOI:10.1186/s12958-026-01578-1

BMC Psychol. 2026 Jul 13. doi: 10.1186/s40359-026-05176-z. Online ahead of print.

ABSTRACT

BACKGROUND: This study aimed to compare antisocial, narcissistic, and Machiavellian traits, along with attention-deficit/hyperactivity disorder (ADHD) symptoms and psychiatric history, between combat athletes and non-athletes.

METHOD: In this cross-sectional study, a total of 300 participants (150 combat athletes and 150 non-athletic controls) completed online assessments. The Sociodemographic and Clinical Data Form, Adult ADHD Self-Report Scale, Narcissistic Personality Inventory, Antisocial Behaviour Scale, and Machiavellianism Scale were administered. Group comparisons were performed using appropriate statistical tests.

RESULTS: Combat athletes demonstrated significantly higher scores on the Antisocial Behaviour Scale (27.4 ± 6.3 vs. 21.1 ± 5.9; p < 0.001), Narcissistic Personality Inventory (17.6 ± 4.8 vs. 13.2 ± 4.5; p < 0.001), and Machiavellianism Scale (41.3 ± 7.1 vs. 36.0 ± 6.4; p < 0.001). ADHD symptom scores were descriptively higher in combat athletes (31.9 ± 9.2 vs. 25.5 ± 8.6; p < 0.001); however, ADHD symptoms did not remain an independent predictor in the multivariable model. In addition, psychiatric diagnoses, medication use, self-harm behaviors, and suicide attempts were more prevalent among combat athletes.

CONCLUSION: Individuals engaged in combat sports exhibited higher levels of antisocial, narcissistic, and Machiavellian traits, along with descriptively elevated ADHD symptomatology and psychiatric vulnerability. However, ADHD symptoms did not show an independent predictive signal in adjusted models, whereas narcissistic traits emerged as the most robust independent correlate of combat sports participation. These findings highlight the importance of targeted mental health screening and intervention strategies in this population.

PMID:42437937 | DOI:10.1186/s40359-026-05176-z

BMC Res Notes. 2026 Jul 12. doi: 10.1186/s13104-026-07888-2. Online ahead of print.

ABSTRACT

OBJECTIVES: To describe the steps taken in the creation of a Data Pack containing clinical data from the FOCUS4 randomized controlled trial in colorectal cancer, representing a case study of clinical trial data access and sharing.

DATA DESCRIPTION: The FOCUS4 trial recruited 1434 metastatic colorectal cancer patients between January 2017 and March 2020 from 94 sites across the United Kingdom (UK). The trial used an adaptive “umbrella” design, in which tumour samples from recruited patients underwent molecular testing, the results of which determined into which of multiple parallel randomized allocations they would be eligible to enter. Overall, 361 patients were successfully randomised into one of four “sub-trials”: FOCUS4-B (6 patients), FOCUS4-C (67 patients), FOCUS4-D (32 patients) and FOCUS4-N (254 patients). After the trial had concluded, we created a Data Pack containing pseudonymised data organised into registration data and follow up data, marked case report forms (CRFs), data dictionary, study protocols and statistical analysis plan. This will help researchers to easily understand the datasets for study replication, further research analysis or to integrate with other existing datasets or repositories.

PMID:42437936 | DOI:10.1186/s13104-026-07888-2

Malar J. 2026 Jul 12. doi: 10.1186/s12936-026-06031-7. Online ahead of print.

ABSTRACT

BACKGROUND: In October 2024, Sudan introduced the malaria vaccine, beginning in Al-Qadarif and Blue Nile states amidst an ongoing armed conflict. Vaccine acceptance among healthcare providers (HCPs) is critical for a successful national rollout. This study assessed malaria vaccine knowledge, acceptability, concerns, and their independent predictors among HCPs in these conflict-affected regions.

METHODS: A descriptive cross-sectional study was conducted from January to February 2025 et al.-Qadarif and Al-Damazin teaching hospitals. An online questionnaire was administered to 342 HCPs. Multivariable linear regression was employed to identify factors associated with continuous knowledge scores. Subsequently, a multivariable logistic regression model was utilized to determine the independent predictors of high vaccine acceptability, incorporating the knowledge score as a continuous independent variable.

RESULTS: Among the 342 respondents, general vaccine acceptability was promising, yet specific operational knowledge remained limited. Significant overall concern about the vaccine was prevalent (64.0%), primarily regarding adverse effects (57.0%), cold-chain handling conditions in Sudan (56.4%), and vaccine effectiveness (42.7%). Social media (48.5%) and peers (45.0%) were the most common information sources. Multivariable linear regression indicated that older age (p = 0.004)and hospital of work (p < 0.001) significantly predicted higher knowledge. In the multivariable logistic regression, the continuous knowledge score emerged as the only statistically significant independent predictor in this model of high vaccine acceptability (adjusted odds ratio: 1.15, 95% confidence interval: 1.01-1.32, p = 0.033), with no significant influence from demographic variables.

CONCLUSIONS: Despite a generally favorable attitude toward the malaria vaccine, Sudanese HCPs exhibit critical knowledge gaps and valid systemic concerns. Our findings suggest a sequential pathway wherein accumulated clinical experience (older age) enhances technical knowledge, which in turn acts as the primary independent driver of vaccine acceptability. Consequently, health authorities must pivot from generalized promotional campaigns to targeted, technical educational interventions. Leveraging secure digital professional networks is urgently needed to bridge these knowledge gaps, optimize provider confidence, and ensure a successful vaccine rollout in this fragile setting. Trial registration Not applicable.

PMID:42437930 | DOI:10.1186/s12936-026-06031-7

Nutr J. 2026 Jul 13. doi: 10.1186/s12937-026-01364-0. Online ahead of print.

ABSTRACT

BACKGROUND: Short-chain fatty acids produced by microorganisms living in the gastrointestinal tract are known to have numerous beneficial effects. Because certain probiotic strains may influence short-chain fatty acid production, probiotic supplementation has been proposed as a potential strategy to modify faecal short-chain fatty acid concentrations. Our aim was to investigate whether probiotics modify faecal short-chain fatty acid levels in healthy populations via systematic review and meta-analysis.

METHODS: We performed a systematic search of the MEDLINE, Embase, and Cochrane databases on 12/04/2024. We analysed exclusively randomized controlled trials meeting all the following criteria: population: healthy people; intervention: probiotic supplementation; control: no probiotic supplementation; outcome: faecal short-chain fatty acid concentrations. A total of 9217 articles were reviewed, 21 of which met the preset inclusion criteria. Twelve articles, including 541 subjects, were eligible for meta-analysis of changes in faecal short-chain fatty acid concentrations. The standardized mean difference (SMD, Hedges’ g) was chosen as the effect indicator due to the different measurement techniques used. A random-effects model was used to estimate SMD with 95% confidence interval (CI) due to the expected heterogeneity.

RESULTS: No statistically significant differences were detected in the faecal butyrate (SMD = 0.07, 95%CI:[(-)0.16 – 0.30]), acetate (SMD = 0.06, 95%CI:[(-)0.14 – 0.27]), and propionate (SMD = 0.09, 95%CI:[(-)0.07 – 0.25]) levels between individuals taking and not taking probiotics.

CONCLUSIONS: Our results suggest that probiotic intake has a minor effect on faecal short-chain fatty acid concentrations, but this effect is not statistically detectable in healthy individuals. Thus, there is insufficient evidence of statistically significant effect of probiotic supplementation on stool short-chain fatty acid concentrations in healthy individuals.

TRIAL REGISTRATION: Research protocol was registered in PROSPERO (CRD42022286137).

PMID:42437924 | DOI:10.1186/s12937-026-01364-0

BMC Med Educ. 2026 Jul 13. doi: 10.1186/s12909-026-09928-8. Online ahead of print.

ABSTRACT

BACKGROUND: The aim of this study is to compare the performance of three large language models (ChatGPT 5, Microsoft Copilot, and Google Gemini 3), with that of dental students using their responses to multiple-choice questions (MCQs) in restorative dentistry. Accuracy of responses were analyzed across the knowledge and cognitive process dimensions of the revised Bloom’s taxonomy (RBT), as well as across subject areas.

METHODS: The restorative dentistry exam questions used in this study were drawn from Turkish Dentistry Specialization Entrance Exam (DUS) administered between 2020 and 2025. The 90 five-option, single-best-answer MCQs were classified according to the RBT to ensure cognitive diversity. Following the exclusion of one exam question which had been annulled by the examination authority, the data analysis was performed on the remaining 89 exam questions. Accuracy of AI models and dental students was compared using Pearson’s chi-square test and Monte Carlo-corrected Fisher’s exact test. Pairwise comparisons were carried out via Bonferroni-corrected Z-test. The results were presented as frequencies and percentages, and p < 0.050 was considered statistically significant.

RESULTS: Microsoft Copilot and Gemini 3 showed similar performance in answering MCQs, both models achieved higher accuracy than ChatGPT 5 and students (p < 0.001). ChatGPT 5’s overall accuracy was found to be significantly higher than that of the students. Accuracy of responses varied according to Bloom’s taxonomy levels and subject areas. Microsoft Copilot exhibited over 90% accuracy in all categories of Bloom’s knowledge and cognitive process dimensions. At the application level of the cognitive process dimension, all chatbots descriptively achieved 100% accuracy; however, this subgroup difference did not reach statistical significance. Chatbot performance was generally superior to that of students across the subject areas of adhesive dentistry, dentin hypersensitivity, dental caries, tooth whitening, aesthetic restorative procedures, contemporary restorative materials, preventive dentistry, lasers, and saliva.

CONCLUSION: AI-based chatbots demonstrate considerable potential in answering questions about restorative dentistry. At the same time, the differences in performance observed among different models suggest there could be variation in the accuracy values of these systems depending on both the taxonomic level and the subject area of restorative dentistry. The findings support the potential use of chatbots as complementary learning resources in dental education. However, the reliability of such systems should be consistently verified and tested in both an academic and clinical setting under the direct supervision of experts. Moreover, students should be equipped with critical thinking skills to appropriately evaluate and use these systems.

PMID:42437914 | DOI:10.1186/s12909-026-09928-8

BMC Oral Health. 2026 Jul 13. doi: 10.1186/s12903-026-09019-6. Online ahead of print.

ABSTRACT

OBJECTIVES: The purpose of study is to compare early failure rate in patients with self-reported penicillin allergy (SRPA) who take clindamycin after implant surgery with non-allergic patients.

MATERIALS AND METHODS: This retrospective cohort study was conducted on patients who had dental implant surgery. The predictor variable was the type of antibiotics. SRPA patients who received clindamycin were obtained from database. The patients in non-allergic group were randomly selected to be identical to allergic patients in terms of age, gender, diabetes, and surface features of implant. Kruskal-Wallis tests were used analysis early implant failure rate between the two groups implant level. The Pearson chi-square test was used to compare categorical data. Multivariate logistic regression analysis was performed with Hosmer and Lemeshow Test.

RESULTS: The study completed with 406 patients. The mean age was 49.64 ± 13.72 years. 104 were male and 302 were female. The patients in both groups were the same in terms of age, gender, and implant surface characteristics. The failure rate of the implant 10(4.93%) in clindamycin group and 4(1.97%) at patient level. The failure rate of the implant 17(1.56%) in amoxicillin group and 9(0.75%) at implant level. There was no statistically significant difference between the groups in terms of failure rate at patient (p = 0.172) and implant level (p = 0.106). The failure rate in male patients was 4.11 time higher than female patients (p = 0.011).

CONCLUSION: The results showed that clindamycin use in SRPA patients did not significantly increase the implant failure with the limitations of this study.

CLINICAL RELEVANCE: This is first study compared early implant failure rate in patients with who received clindamycin after implant surgery with non-allergic patients with similar age, gender, health status.

PMID:42437912 | DOI:10.1186/s12903-026-09019-6

BMC Musculoskelet Disord. 2026 Jul 13. doi: 10.1186/s12891-026-10199-z. Online ahead of print.

ABSTRACT

BACKGROUND: Medial opening wedge high tibial osteotomy (MOWHTO) is a standard treatment for knee osteoarthritis. However, lateral hinge fractures occur in 19-25% of cases, with Type 2 fractures causing significant instability. While lateral plating is recommended, it requires an additional incision. This study investigated whether medial-only fixation using a 4.5 mm cortical screw inserted into the Tomofix plate’s oval hole could provide sufficient stability for Type 2 hinge fractures.

METHODS: Twelve fresh pig knees underwent MOWHTO with a 6 mm opening. They were divided into two groups: Group I (intact lateral hinge) and Group F (induced Type 2 hinge fracture fixed with a medial 4.5 mm cortical screw in the Tomofix oval hole). Specimens were subjected to 2,000 cycles of axial loading (up to 800 N), representing early postoperative partial weight-bearing conditions. Displacement during cycling, and changes in the anterior gap (AG), posterior gap (PG), and posterior tibial slope (PTS) were measured.

RESULTS: Group F exhibited significantly greater displacement during cyclic loading compared to Group I (p = 0.0029), indicating significantly greater construct displacement. There were no statistically significant differences between the groups regarding changes in AG (p = 0.15), PG (p = 0.53), or PTS (p = 0.22).

CONCLUSIONS: Fixation with a 4.5 mm cortical screw in the Tomofix oval hole did not restore stability comparable to an intact lateral hinge for Type 2 lateral hinge fractures during an ex vivo porcine MOWHTO model.

PMID:42437910 | DOI:10.1186/s12891-026-10199-z