Category: Statistics

BioData Min. 2024 Jun 22;17(1):18. doi: 10.1186/s13040-024-00370-4.

ABSTRACT

Deep learning shows great promise for medical image analysis but often lacks explainability, hindering its adoption in healthcare. Attribution techniques that explain model reasoning can potentially increase trust in deep learning among clinical stakeholders. In the literature, much of the research on attribution in medical imaging focuses on visual inspection rather than statistical quantitative analysis.In this paper, we proposed an image-based saliency framework to enhance the explainability of deep learning models in medical image analysis. We use adaptive path-based gradient integration, gradient-free techniques, and class activation mapping along with its derivatives to attribute predictions from brain tumor MRI and COVID-19 chest X-ray datasets made by recent deep convolutional neural network models.The proposed framework integrates qualitative and statistical quantitative assessments, employing Accuracy Information Curves (AICs) and Softmax Information Curves (SICs) to measure the effectiveness of saliency methods in retaining critical image information and their correlation with model predictions. Visual inspections indicate that methods such as ScoreCAM, XRAI, GradCAM, and GradCAM++ consistently produce focused and clinically interpretable attribution maps. These methods highlighted possible biomarkers, exposed model biases, and offered insights into the links between input features and predictions, demonstrating their ability to elucidate model reasoning on these datasets. Empirical evaluations reveal that ScoreCAM and XRAI are particularly effective in retaining relevant image regions, as reflected in their higher AUC values. However, SICs highlight variability, with instances of random saliency masks outperforming established methods, emphasizing the need for combining visual and empirical metrics for a comprehensive evaluation.The results underscore the importance of selecting appropriate saliency methods for specific medical imaging tasks and suggest that combining qualitative and quantitative approaches can enhance the transparency, trustworthiness, and clinical adoption of deep learning models in healthcare. This study advances model explainability to increase trust in deep learning among healthcare stakeholders by revealing the rationale behind predictions. Future research should refine empirical metrics for stability and reliability, include more diverse imaging modalities, and focus on improving model explainability to support clinical decision-making.

PMID:38909228 | DOI:10.1186/s13040-024-00370-4

BMC Pediatr. 2024 Jun 22;24(1):403. doi: 10.1186/s12887-024-04879-9.

ABSTRACT

BACKGROUND: Diarrhea is considered to be one of the major public health concerns in developing countries. It has a detrimental impact, reflecting one of the highest child mortality rates globally, especially in Sub-Saharan Africa, where 2 out of every 10 children in Uganda under the age of five die. The objective of this study was to investigate the factors associated with time to treatment seeking by caretakers of children under-five with Diarrhea in Uganda.

METHOD: DOVE dataset of 745 caretakers in a prospective and retrospective incidence-based study using multi-stage sampling design was used in the assessment. The analysis was done using a time-to-event approach using life tables, Kaplan Meier survival analysis and multilevel proportional hazards model.

RESULTS: Kaplan-Meier survival analysis indicated the median time to seeking treatment among 745 caretakers of children under-Five after onset of diarrhea was 2 days. The multi-level proportional hazards model of a Weibull distribution showed that the estimated frailty variance was 0.13, indicating heterogeneity of treatment seeking time by caretakers of under-five children with diarrhea across regions in Uganda. Significant factors found to influence time to treatment-seeking by caretakers of children under-five with diarrhea were, male children (HR = 0.82; 95% CI = 0.71-0.95, p = 0.010), belonging to richest wealth quintile (HR = 1.37; 95% CI = 1.05-1.78, p = 0.022), and residing more than 5 km away from a health facility (HR = 0.68; 95% CI = 0.56-0.84, p = 0.000).

CONCLUSIONS: There are delays in seeking diarrhea treatment in Uganda because two days are enough to claim a life after dehydration.The policymakers should pay attention to formulate effective intervention to sensitize caregivers on the importance of early treatment-seeking behavior to avoid severe malnutrition caused by diarrhea. Community awareness program should also be encouraged particularly in areas of more than 5 km from the health facility to make people aware of the necessity to take prompt action to seek care in the early stage.

PMID:38909217 | DOI:10.1186/s12887-024-04879-9

BMC Med Res Methodol. 2024 Jun 22;24(1):136. doi: 10.1186/s12874-024-02257-8.

ABSTRACT

BACKGROUND: Generating synthetic patient data is crucial for medical research, but common approaches build up on black-box models which do not allow for expert verification or intervention. We propose a highly available method which enables synthetic data generation from real patient records in a privacy preserving and compliant fashion, is interpretable and allows for expert intervention.

METHODS: Our approach ties together two established tools in medical informatics, namely OMOP as a data standard for electronic health records and Synthea as a data synthetization method. For this study, data pipelines were built which extract data from OMOP, convert them into time series format, learn temporal rules by 2 statistical algorithms (Markov chain, TARM) and 3 algorithms of causal discovery (DYNOTEARS, J-PCMCI+, LiNGAM) and map the outputs into Synthea graphs. The graphs are evaluated quantitatively by their individual and relative complexity and qualitatively by medical experts.

RESULTS: The algorithms were found to learn qualitatively and quantitatively different graph representations. Whereas the Markov chain results in extremely large graphs, TARM, DYNOTEARS, and J-PCMCI+ were found to reduce the data dimension during learning. The MultiGroupDirect LiNGAM algorithm was found to not be applicable to the problem statement at hand.

CONCLUSION: Only TARM and DYNOTEARS are practical algorithms for real-world data in this use case. As causal discovery is a method to debias purely statistical relationships, the gradient-based causal discovery algorithm DYNOTEARS was found to be most suitable.

PMID:38909216 | DOI:10.1186/s12874-024-02257-8

BMC Womens Health. 2024 Jun 22;24(1):366. doi: 10.1186/s12905-024-03218-5.

ABSTRACT

BACKGROUND: Insulin resistance (IR) induces hyperinsulinemia, which activates downstream signaling pathways such as the phosphatidylinositol-3-kinase/protein kinase B (PI3K/AKT) pathway, ultimately leading to abnormal proliferation and apoptosis of endometrial cells. This is thought to be a key pathogenic mechanism underlying the development of endometrial polyps (EP). This study aims to investigate the relationship between IR and the development of EP, the expression levels of downstream signaling molecules, including PI3K and AKT, and related laboratory parameters were examined.

METHODS: A total of 100 patients who visited the gynecology outpatient clinic of Zhongda Hospital affiliated with Southeast University from May 2021 to March 2023 and were diagnosed with abnormal endometrial echoes by vaginal ultrasound and underwent hysteroscopic diagnostic curettage were enrolled in this study. General data and relevant hematological indicators were compared, and intraoperative specimens were obtained for pathological examination. Possible factors influencing the development of endometrial polyps were analyzed using Pearson correlation analysis and logistic regression analysis.

RESULTS: In terms of body mass index, waist circumference, fasting insulin, insulin resistance index, serum total testosterone, and free testosterone index, women of childbearing age in the endometrial polyp group had higher values than those in the non-polyp group, while sex hormone-binding globulin in the endometrial polyp group was lower than that in the non-polyp group, and the differences were statistically significant (P < 0.05). The expression scores and mRNA expression levels of PI3K and AKT proteins were higher in the EP group than in the non-EP group (p < 0.05). Pearson correlation analysis showed a positive correlation between HOMA-IR and the expression scores of PI3K and AKT proteins (p < 0.01).

CONCLUSIONS: Insulin resistance and abnormal activation of the phosphatidylinositol 3-kinase/protein kinase B signaling pathway may be potential pathogenic mechanisms for the development of endometrial polyps.

PMID:38909214 | DOI:10.1186/s12905-024-03218-5

BMC Public Health. 2024 Jun 22;24(1):1664. doi: 10.1186/s12889-024-19074-9.

ABSTRACT

BACKGROUND: Obesity is a global health problem, and its connection with social and environmental factors is well-established. Social factors, such as urban segregation, may impact obesity through various mechanisms, including food and physical activity environments, as well as social norms and networks. This multilevel study aims to examine the effect of socio-economic residential segregation of Latin American cities on the obesity of individuals within those cities.

METHODS: We analyzed data from national surveys for a total of 59,340 individuals of 18-70 years of age, conducted in 156 cities across Brazil, Chile, Colombia, and Mexico between 2007 and 2013. We adjusted two-level linear mixed models for body mass index (BMI) stratified by sex and country, controlling for age, educational level and poverty. Separate models were built for dissimilarity and isolation segregation indices.

RESULTS: The relationships between segregation indices and BMI were mostly not statistically significant, and in some cases, they were opposite to what was expected. The only significant relationships were observed in Colombian men, using the dissimilarity index (-7.5 [95% CI: -14.4, -0.5]) and in Colombian women, using the isolation index (-7.9 [95% CI: -14.1, -1.7]).

CONCLUSIONS: While individual-level factors cannot fully explain differences among people in the same city, segregation indices may help. However, we found that in some cases, the relationship between BMI and segregation indices is opposite to what is expected based on prior literature. This should be considered in examining the phenomenon. Further research on obesogenic environments in segregated neighborhoods could provide valuable evidence.

PMID:38909210 | DOI:10.1186/s12889-024-19074-9

BMC Public Health. 2024 Jun 22;24(1):1665. doi: 10.1186/s12889-024-19139-9.

ABSTRACT

BACKGROUND: The related literature has primarily addressed cigarette smoking control. It seems that researchers have failed to explore the determinants of hookah smoking (HS) control. In an attempt to fill this gap, the present study explores experts’ views about aspects of HS control in Bandar Abbas, a city in the south of Iran.

METHODS: The present qualitative study, conducted in 2022 and 2023, used a content analysis. To this aim, 30 experts in tobacco prevention and control were invited to participate in the research. Twenty seven accepted the invitation. In-depth, semi-structured, and face-to-face interviews were held with the experts. A purposive sampling was used and the data collection continued until data saturation. The interviews lasted between 18 and 65 min. MAXQDA 10.0 was used for data management and analysis.

RESULTS: The expert interviewees had a mean age of 44.77 ± 6.57 years and a mean work experience of 18.6 ± 6.8 years. A total number of six main categories were extracted from the data, including usin influential figures to control HS, controlling HS by alternative activities, changing beliefs and attitudes toward HS, taking administrative and regulatory measures, and facilitating HS cessation.

CONCLUSION: This qualitative study explored the multifaceted ways people adopt to quit HS. Using influential figures to control hookah smoking, promoting alternative activities as a means of control, changing beliefs and attitudes, enforcing administrative regulations, and facilitating quit attempts all play an important role in tackling the prevalence of hookah smoking. These findings emphasize the importance of a comprehensive and multifaceted approach to integrate various interventions to effectively address hookah smoking behavior.

PMID:38909209 | DOI:10.1186/s12889-024-19139-9

BMC Pediatr. 2024 Jun 22;24(1):405. doi: 10.1186/s12887-024-04848-2.

ABSTRACT

BACKGROUND: Nine in every thousand children born in the United Kingdom have congenital heart disease, and 250,000 adults are living with the condition. This study aims to investigate the associations between congenital heart disease and educational outcomes among school-aged children in Scotland.

METHODS: Routine health and education databases were linked to produce a cohort of all singleton children born in Scotland and attending a local authority run primary, secondary, or special school in Scotland at some point between 2009 and 2013. Children with congenital heart disease within this cohort were compared with children unaffected by congenital conditions. Outcomes investigated were special educational need (SEN), absenteeism, exclusion, academic attainment, and unemployment. All analyses were adjusted for sociodemographic and maternity confounders. Absenteeism was investigated as a mediating factor in the associations with attainment and unemployment.

RESULTS: Of the 715,850 children, 6,295 (0.9%) had congenital heart disease and 4,412 (6.1%) had isolated congenital heart disease. Congenital heart disease and isolated congenital heart disease were both significantly associated with subsequent special educational need (OR 3.45, 95% CI 3.26-3.65, p < 0.001 and OR 1.98, 95% CI 1.84-2.13, p < 0.001 respectively), absenteeism (IRR 1.13, 95% CI 1.10-1.16, p < 0.001 and IRR 1.10, 95% CI 1.06-1.13, p < 0.001 respectively), and low academic attainment (OR 1.69, 95% CI 1.39-2.07, p < 0.001 and OR 1.35, 95% CI 1.07-1.69, p = 0.011 respectively). Neither congenital heart disease nor isolated congenital heart disease were associated with school exclusion. Only congenital heart disease (OR 1.21, 95% CI 1.03-1.42, p = 0.022) but not isolated congenital heart disease was associated with unemployment. When days absent were included in the analyses investigating attainment and unemployment, the conclusions were not altered.

CONCLUSION: Children with congenital heart disease have greater special educational need, lower school attendance, attain lower examination grades and have greater unemployment compared to peers. In addition to healthcare support, affected children need educational support to avoid additional impact on their long-term wellbeing.

PMID:38909207 | DOI:10.1186/s12887-024-04848-2

BMC Urol. 2024 Jun 22;24(1):131. doi: 10.1186/s12894-024-01493-w.

ABSTRACT

OBJECTIVE: The incidence of recurrent hernia after radical resection of prostate cancer is high, so this article discusses the incidence and risk factors of inguinal hernia after radical resection of prostate cancer.

METHODS: This case control study was conducted in The First People’s Hospital of Huzhou clinical data of 251 cases underwent radical resection of prostate cancer in this hospital from March 2019 to May 2021 were retrospectively analyzed. According to the occurrence of inguinal hernia, the subjects were divided into study group and control group, and the clinical data of each group were statistically analyzed, Multivariate Logistic analysis was performed to find independent influencing factors for predicting the occurrence of inguinal hernia. The Kaplan-Meier survival curve was drawn according to the occurrence and time of inguinal hernia.

RESULTS: The overall incidence of inguinal hernia after prostate cancer surgery was 14.7% (37/251), and the mean time was 8.58 ± 4.12 months. The average time of inguinal hernia in patients who received lymph node dissection was 7.61 ± 4.05 (month), and that in patients who did not receive lymph node dissection was 9.16 ± 4.15 (month), and there was no significant difference between them (P > 0.05). There were no statistically significant differences in the incidence of inguinal hernia with age, BMI, hypertension, diabetes, PSA, previous abdominal operations and operative approach (P > 0.05), but there were statistically significant differences with surgical method and pelvic lymph node dissection (P < 0.05). The incidence of pelvic lymph node dissection in the inguinal hernia group was 24.3% (14/57), which was significantly higher than that in the control group 11.8% (23/194). Logistic regression analysis showed that pelvic lymph node dissection was a risk factor for inguinal hernia after prostate cancer surgery (OR = 0.413, 95%Cl: 0.196-0.869, P = 0.02). Kaplan-Meier survival curve showed that the rate of inguinal hernia in the group receiving pelvic lymph node dissection was significantly higher than that in the control group (P < 0.05).

CONCLUSION: Pelvic lymph node dissection is a risk factor for inguinal hernia after radical resection of prostate cancer.

PMID:38909202 | DOI:10.1186/s12894-024-01493-w

BMC Womens Health. 2024 Jun 22;24(1):363. doi: 10.1186/s12905-024-03204-x.

ABSTRACT

INTRODUCTION: Intimate partner violence (IPV) can be described as a violation of human rights that results from gender inequality. It has arisen as a contemporary issue in societies from both developing and industrialized countries and an impediment to long-term development. This study evaluates the prevalence of IPV and its variants among the empowerment status of women and identify the associated sociodemographic parameters, linked to IPV.

METHODS: This study is based on data from the National Family Health Survey (NFHS) of India, 2019-21 a nationwide survey that provides scientific data on health and family welfare. Prevalence of IPV were estimated among variouss social and demographic strata. Pearson chi-square test was used to estimate the strength of association between each possible covariate and IPV. Significantly associated covariates (from univariate logistic regression) were further analyzed through separate bivariate logistic models for each of the components of IPV, viz-a-viz sexual, emotional, physical and severe violence of the partners.

RESULTS: The prevalence of IPV among empowered women was found to be 26.21%. Among those who had experienced IPV, two-thirds (60%) were faced the physical violence. When compared to highly empowered women, less empowered women were 74% more likely to face emotional abuse. Alcohol consumption by a partner was established to be attributing immensely for any kind of violence, including sexual violence [AOR: 3.28 (2.83-3.81)].

CONCLUSIONS: Our research found that less empowered women experience all forms of IPV compared to more empowered women. More efforts should to taken by government and other stakeholders to promote women empowerment by improving education, autonomy and decision-making ability.

PMID:38909198 | DOI:10.1186/s12905-024-03204-x

BMC Public Health. 2024 Jun 22;24(1):1667. doi: 10.1186/s12889-024-19112-6.

ABSTRACT

BACKGROUND: HALE is now a regular strategic planning indicator for all levels of the Chinese government. However, HALE measurements necessitate comprehensive data collection and intricate technology. Therefore, effectively converting numerous diseases into the years lived with disability (YLD) rate is a significant challenge for HALE measurements. Our study aimed to construct a simple YLD rate measurement model with high applicability based on the current situation of actual data resources within China to address challenges in measuring HALE target values during planning.

METHODS: First, based on the Chinese YLD rate in the Global Burden of Disease (GBD) 2019, Pearson correlation analysis, the global optimum method, etc., was utilized to screen the best predictor variables from the current Chinese data resources. Missing data for predictor variables were filled in via spline interpolation. Then, multiple linear regression models were fitted to construct the YLD rate measurement model. The Sullivan method was used to measure HALE. The Monte Carlo method was employed to generate 95% uncertainty intervals. Finally, model performances were assessed using the mean absolute error (MAE) and mean absolute percentage error (MAPE).

RESULTS: A three-input-parameter model was constructed to measure the age-specific YLD rates by sex in China, directly using the incidence of infectious diseases, the incidence of chronic diseases among persons aged 15 and older, and the addition of an under-five mortality rate covariate. The total MAE and MAPE for the combined YLD rate were 0.0007 and 0.5949%, respectively. The MAE and MAPE of the combined HALE in the 0-year-old group were 0.0341 and 0.0526%, respectively. There were slightly fewer males (0.0197, 0.0311%) than females (0.0501, 0.0755%).

CONCLUSION: We constructed a high-accuracy model to measure the YLD rate in China by using three monitoring indicators from the Chinese national routine as predictor variables. The model provides a realistic and feasible solution for measuring HALE at the national and especially regional levels, considering limited data.

PMID:38909195 | DOI:10.1186/s12889-024-19112-6