Tag: statistics

BMC Med. 2022 May 11;20(1):168. doi: 10.1186/s12916-022-02366-5.

ABSTRACT

BACKGROUND: The association between vitamin E and cancer risk has been widely investigated by observational studies, but the findings remain inconclusive. Here, we aimed to evaluate the causal effect of circulating vitamin E on the risk of ten common cancers, including bladder, breast, colorectal, esophagus, lung, oral and pharynx, ovarian, pancreatic, prostate, and kidney cancer.

METHODS: A Mendelian randomization (MR) analytic framework was applied to data from a cancer-specific genome-wide association study (GWAS) comprising a total of 297,699 cancer cases and 304,736 controls of European ancestry. Three genetic instrumental variables associated with circulating vitamin E were selected. Summary statistic-based methods of inverse variance weighting (IVW) and likelihood-based approach, as well as the individual genotyping-based method of genetic risk score (GRS) were used. Multivariable IVW analysis was further performed to control for potential confounding effects. Furthermore, the UK Biobank cohort was used as external validation, supporting 355,543 European participants (incident cases ranged from 437 for ovarian cancer to 4882 for prostate cancer) for GRS-based estimation of circulating vitamin E, accompanied by a one-sample MR analysis of dietary vitamin E intake underlying the time-to-event analytic framework.

RESULTS: Specific to cancer GWAS, we found that circulating vitamin E was significantly associated with increased bladder cancer risk (odds ratios [OR]_IVW = 6.23, P_IVW = 3.05×10^-3) but decreased breast cancer risk (OR_IVW = 0.68, P_IVW = 8.19×10^-3); however, the significance of breast cancer was dampened (P_{multivariable IVW} > 0.05) in the subsequent multivariable MR analysis. In the validation stage of the UK Biobank cohort, we did not replicate convincing causal effects of genetically predicted circulating vitamin E concentrations and dietary vitamin E intake on the risk of ten cancers.

CONCLUSIONS: This large-scale population study upon data from cancer-specific GWAS and a longitudinal biobank cohort indicates plausible non-causal associations between circulating vitamin E and ten common cancers in the European populations. Further studies regarding ancestral diversity are warranted to validate such causal associations.

PMID:35538486 | DOI:10.1186/s12916-022-02366-5

BMC Med Educ. 2022 May 10;22(1):353. doi: 10.1186/s12909-022-03434-3.

ABSTRACT

BACKGROUND: Due to the COVID-19 pandemic, the 2021 Harvard Medical School course Clinical Topics in Global Health was offered for the first time as a remote class. We sought to understand student and faculty perceptions of the elective and evaluate the perceived effectiveness of teaching global health using an online education platform.

METHODS: Following the course, students and faculty were invited to complete a combined total of three online surveys, which consisted of closed- and open-response questions assessing the strengths and challenges of online learning. Data analyses included traditional descriptive statistics, Net Promoter Score calculation, and inductive thematic analysis of qualitative data.

RESULTS: Thirty-two students and eighteen guest faculty (including four international faculty) participated in the course. Highly-rated course components included guest lecturers, practical skill sessions, polls, and case studies. The Net Promoter Score for the course was excellent at 92, and students reported a greater likelihood of pursuing a career in global health because of the course. While students and faculty highlighted limitations of the remote learning platform (lack of community and interactivity), they also commented on increased accessibility and faculty diversity. Most faculty and students recommended a hybrid model for future versions of the course and suggested strategies to address current limitations.

CONCLUSIONS: A remote learning platform can effectively deliver global health education, both in the pandemic setting and beyond.

PMID:35538485 | DOI:10.1186/s12909-022-03434-3

BMC Med. 2022 May 11;20(1):174. doi: 10.1186/s12916-022-02369-2.

ABSTRACT

BACKGROUND: Randomized controlled trials (RCTs) and cohort studies are the most common study design types used to assess the treatment effects of medical interventions. To evaluate the agreement of effect estimates between bodies of evidence (BoE) from randomized controlled trials (RCTs) and cohort studies and to identify factors associated with disagreement.

METHODS: Systematic reviews were published in the 13 medical journals with the highest impact factor identified through a MEDLINE search. BoE-pairs from RCTs and cohort studies with the same medical research question were included. We rated the similarity of PI/ECO (Population, Intervention/Exposure, Comparison, Outcome) between BoE from RCTs and cohort studies. The agreement of effect estimates across BoE was analyzed by pooling ratio of ratios (RoR) for binary outcomes and difference of mean differences for continuous outcomes. We performed subgroup analyses to explore factors associated with disagreements.

RESULTS: One hundred twenty-nine BoE pairs from 64 systematic reviews were included. PI/ECO-similarity degree was moderate: two BoE pairs were rated as “more or less identical”; 90 were rated as “similar but not identical” and 37 as only “broadly similar”. For binary outcomes, the pooled RoR was 1.04 (95% CI 0.97-1.11) with considerable statistical heterogeneity. For continuous outcomes, differences were small. In subgroup analyses, degree of PI/ECO-similarity, type of intervention, and type of outcome, the pooled RoR indicated that on average, differences between both BoE were small. Subgroup analysis by degree of PI/ECO-similarity revealed high statistical heterogeneity and wide prediction intervals across PI/ECO-dissimilar BoE pairs.

CONCLUSIONS: On average, the pooled effect estimates between RCTs and cohort studies did not differ. Statistical heterogeneity and wide prediction intervals were mainly driven by PI/ECO-dissimilarities (i.e., clinical heterogeneity) and cohort studies. The potential influence of risk of bias and certainty of the evidence on differences of effect estimates between RCTs and cohort studies needs to be explored in upcoming meta-epidemiological studies.

PMID:35538478 | DOI:10.1186/s12916-022-02369-2

Lipids Health Dis. 2022 May 10;21(1):42. doi: 10.1186/s12944-022-01650-w.

ABSTRACT

BACKGROUND: Monoacetyldiglycerides (MAcDG), are acetylated triglycerides (TG) and an emerging class of bioactive or functional lipid with promising nutritional, medical, and industrial applications. A major challenge exists when analyzing MAcDG from other subclasses of TG in biological matrices, limiting knowledge on their applications and metabolism.

METHODS: Herein a multimodal analytical method for resolution, identification, and quantitation of MAcDG in biological samples was demonstrated based on thin layer chromatography-flame ionization detection complimentary with C30-reversed phase liquid chromatography-high resolution accurate mass tandem mass spectrometry. This method was then applied to determine the MAcDG molecular species composition and quantity in E. solidaginis larvae. The statistical method for analysis of TG subclass composition and molecular species composition of E. solidaginis larvae was one-way analysis of variance (ANOVA).

RESULTS: The findings suggest that the proposed analytical method could simultaneously provide a fast, accurate, sensitive, high throughput analysis of MAcDG from other TG subclasses, including the fatty acids, isomers, and molecular species composition.

CONCLUSION: This method would allow for MAcDG to be included during routine lipidomics analysis of biological samples and will have broad interests and applications in the scientific community in areas such as nutrition, climate change, medicine and biofuel innovations.

PMID:35538477 | DOI:10.1186/s12944-022-01650-w

BMC Public Health. 2022 May 10;22(1):928. doi: 10.1186/s12889-022-13293-8.

ABSTRACT

BACKGROUND: Community Reinforcement and Family Training (CRAFT) is an intervention designed to help the concerned significant others (CSOs) of people with alcohol problems who are reluctant to seek treatment. It aims to improve the well-being of CSOs and teach them how to change their behavior in order to positively influence the “identified patient” (IP) to seek treatment.

METHODS: The aim of the present pragmatic cluster-randomized trial was to compare the effectiveness of three formats for delivering CRAFT in real life settings: group sessions, individual sessions, and written material only (control group). Eighteen public treatment centers for alcohol use disorders were randomly assigned to deliver CRAFT in one of the three formats as part of their daily clinical routine. CSOs were recruited via pamphlets, general practitioners, and advertisements on social media. Trained clinicians delivered CRAFT in individual and group format, and self-administered CRAFT was limited to handing out a self-help book. The primary outcome was treatment engagement of the IP after three months.

RESULTS: A total of 249 CSOs were found to be eligible and randomly assigned to receive CRAFT delivered in group, individual, or self-administered format. The three-month follow-up rate was 60%. At three months follow-up, 29% (n = 32) of the CSOs who received group/individual CRAFT reported that their IP had engaged in treatment. The corresponding rate for the CSOs who received self-administered CRAFT was lower (15%; n = 5) but did not differ significantly from the other group of CSOs (Odds ratio (OR) = 2.27 (95% CI: 0.80, 6.41)).

CONCLUSION: We hypothesized that CSOs receiving CRAFT in a group format would improve the most, but although our findings pointed in this direction, the differences were not statistically significant.

TRIAL REGISTRATION: Clinical trials.gov ID: NCT03281057 . Registration date:13/09/2017.

PMID:35538465 | DOI:10.1186/s12889-022-13293-8

BMC Surg. 2022 May 10;22(1):167. doi: 10.1186/s12893-022-01601-3.

ABSTRACT

AIM: To develop and validate a prediction model to evaluate the perioperative hypoglycemia risk in hospitalized type 2 diabetes mellitus (T2DM) patients undergoing elective surgery.

METHODS: We retrospectively analyzed the electronic medical records of 1410 T2DM patients who had been hospitalized and undergone elective surgery. Regression analysis was used to develop a predictive model for perioperative hypoglycemia risk. The receiver operating characteristic (ROC) curve and the Hosmer-Lemeshow test were used to verify the model.

RESULTS: Our study showed an incidence of 10.7% for level 1 hypoglycemia and 1.8% for level 2 severe hypoglycemia during the perioperative period. A perioperative hypoglycemic risk prediction model was developed that was mainly composed of four predictors: duration of diabetes ≥ 10 year, body mass index (BMI) < 18.5 kg/m², standard deviation of blood glucose (SDBG) ≥ 3.0 mmol/L, and preoperative hypoglycemic regimen of insulin subcutaneous. Based on this model, patients were categorized into three groups: low, medium, and high risk. Internal validation of the prediction model showed high discrimination (ROC statistic = 0.715) and good calibration (no significant differences between predicted and observed risk: Pearson χ² goodness-of-fit P = 0.765).

CONCLUSIONS: The perioperative hypoglycemic risk prediction model categorizes the risk of hypoglycemia using only four predictors and shows good reliability and validity. The model serves as a favorable tool for clinicians to predict hypoglycemic risk and guide future interventions to reduce hypoglycemia risk.

PMID:35538461 | DOI:10.1186/s12893-022-01601-3

BMC Womens Health. 2022 May 10;22(1):151. doi: 10.1186/s12905-022-01736-8.

ABSTRACT

BACKGROUND: Contraceptive knowledge and use has been an emerging topic of interest in adolescents in Asia. This study quantified the contribution of the socioeconomic determinants of inequality in contraceptive use among currently married female adolescents (15-24) in four south Asian countries: India, Bangladesh, Nepal and Pakistan.

DATA AND METHODS: The data of Demographic Health Survey (DHS) for four South Asian countries, i.e. India (NFHS 2015-16), Nepal (DHS 2016), Bangladesh (DHS 2014) and Pakistan (DHS 2012-2013) has been used for examining the contraceptive use and inherent socioeconomic inequality. After employing logistic regression, concentration curves based on decomposition analysis have been made to analyse the socioeconomic inequality.

RESULTS: The results reveal that the use of contraception among female adolescents remains low and factors like education, employment, having one or more children, media exposure were positively associated with it. In terms of socioeconomic inequality, a significant amount of variation has been observed across the countries. In India, poor economic status (95.23%), illiteracy (51.29%) and rural residence (23.06%) contributed maximum in explaining the socioeconomic inequality in contraceptive use among female adolescents. For Bangladesh, the largest contributors to inequalities were rural residence (260%), illiteracy (146.67%) while birth order 3 + (- 173.33%) contributed negatively. Illiteracy (50%), poor economic status (47.83%) and rural residence (16.30%) contributed maximum to the inequalities in contraceptive use in Pakistan while birth order 3 + (- 9.78%) contributed negatively. In Nepal, the important operators of inequalities were unemployment (105.26%), birth order 3 + (52.63%) and poor economic status (47.37%), while rural residence contributed negatively (- 63.16%) to inequalities in contraceptive use.

CONCLUSIONS: Using a cross country perspective, this study presents an socioeconomic inequality analysis in contraceptive use and the important factors involved in the same. Since the factors contributing to inequalities in contraceptive use vary across countries, there is a need to imply country-specific initiatives which will look after the special needs of this age-group.

PMID:35538459 | DOI:10.1186/s12905-022-01736-8

BMC Pediatr. 2022 May 10;22(1):260. doi: 10.1186/s12887-022-03326-x.

ABSTRACT

BACKGROUND: Monitoring body composition and changes in energy expenditure during maturation and growth is significant, as many components can influence body structure in adulthood. In the case of young players, when these changes can influence their strength and power, it seems to be equally important. Our aim was to examine whether resting energy expenditure (REE) and body composition would change after 10 months from baseline in physically active children and adolescents.

METHODS: We obtained data from 80 children and adolescents aged 9 to 17 years at two measurement points: the baseline in September 2018 and after 10 months in July 2019. The study was carried out using a calorimeter (Fitmate MED, Cosmed, Rome, Italy), a device used to assess body composition using by the electrical bioimpedance method by means of a segment analyzer (TANITA MC-980). The Student’s t-test and linear regression analysis were used. Using the stepwise forward regression procedure, the selection of factors in a statistically significant way that describes the level of REE was made.

RESULTS: We noticed that REE was not significantly different between baseline (1596.94 ± 273.01 kcal) and after 10 months (1625.38 ± 253.26 kcal). When analyzing the difference in REE between studies girls, we found body height as a significant predictor. The results of our study show a negative relationship between growth and REE. Differences between sexes and age in REE between baseline and after 10 months were not significant.

CONCLUSIONS: Our study involving physically active children and adolescents, which used repeated objective measures and longitudinal statistical modeling to analyze them, was unable to demonstrate any interaction between body weight change, body composition measurements, and REE.

PMID:35538456 | DOI:10.1186/s12887-022-03326-x

JMIR Med Inform. 2022 May 2. doi: 10.2196/37239. Online ahead of print.

ABSTRACT

BACKGROUND: A primary goal of precision medicine is to identify patient subgroups and infer their underlying disease processes, with the aim of designing targeted interventions. However, while several studies have identified patient subgroups, there is a considerable gap between the identification of patient subgroups, and their modeling and interpretation for clinical applications.

OBJECTIVE: To develop and evaluate a novel analytical framework for modeling and interpreting patient subgroups (MIPS) using a three-step modeling approach. (1) Visual analytical modeling to automatically identify patient subgroups and their co-occurring comorbidities, and determine their statistical significance and clinical interpretability. (2) Classification modeling to classify patients into subgroups and measure its accuracy. (3) Prediction modeling to predict a patient’s risk for an adverse outcome, and compare its accuracy with and without patient subgroup information.

METHODS: The MIPS framework was developed using (1) bipartite networks to identify patient subgroups based on frequently co-occurring high-risk comorbidities; (2) multinomial logistic regression to classify patients into subgroups; and (3) hierarchical logistic regression to predict the risk of an adverse outcome using subgroup membership, compared to standard logistic regression without subgroup membership. The MIPS framework was evaluated on three hospital readmission conditions: chronic obstructive pulmonary disease (COPD), congestive heart failure (CHF), and total hip/knee arthroplasty (THA/TKA). For each condition, we extracted cases defined as patients readmitted within 30 days of hospital discharge, and controls defined as patients not readmitted within 90 days of discharge, matched by age, gender, race, and Medicaid eligibility (n[COPD]=29,016, n[CHF]=51,550, n[THA/TKA]=16,498).

RESULTS: In each condition, the visual analytical model identified patient subgroups that were statistically significant (Q=0.17, 0.17, 0.31; P<.001, <.001, <.05), were significantly replicated (RI=0.92, 0.94, 0.89; P<.001, <.001, <.01), and were clinically meaningful to clinicians. (2) In each condition, the classification model had high accuracy in classifying patients into subgroups (mean accuracy=99.60%, 99.34%, 99.86%). (3) In two conditions (COPD, THA/TKA), the hierarchical prediction model had a small but statistically significant improvement in discriminating between the readmitted and not readmitted patients as measured by net reclassification improvement (NRI=.059, .11), but not as measured by the C-statistic or integrated discrimination improvement (IDI).

CONCLUSIONS: While the visual analytical models identified statistically and clinically significant patient subgroups, the results pinpoint the need to analyze subgroups at different levels of granularity for improving the interpretability of intra- and inter-cluster associations. The high accuracy of the classification models reflects the strong separation of the patient subgroups despite the size and density of the datasets. Finally, the small improvement in predictive accuracy suggests that comorbidities alone were not strong predictors for hospital readmission, and the need for more sophisticated subgroup modeling methods. Such advances could improve the interpretability and predictive accuracy of patient subgroup models for reducing the risk of hospital readmission and beyond.

PMID:35537203 | DOI:10.2196/37239

J Occup Environ Hyg. 2022 May 10:1-15. doi: 10.1080/15459624.2022.2076860. Online ahead of print.

ABSTRACT

Recently, the National Institute for Occupational Safety and Health (NIOSH) released an updated version of the NIOSH Industry and Occupation Computerized Coding System (NIOCCS), which uses supervised machine learning to assign industry and occupational codes based on provided free-text information. However, no efforts have been made to externally verify the quality of assigned industry and job titles when the algorithm is provided with inputs of varying quality. This study sought to evaluate whether the NIOCCS algorithm was sufficiently robust with low quality inputs and how variable quality could impact subsequent job estimated exposures in a large job exposure matrix for noise (NoiseJEM).Using free-text industry and job descriptions from >700,000 noise measurements in the NoiseJEM, three files were created and input into NIOCCS: (1) N1, “raw” industries and job titles; (2) N2, “refined” industries and “raw” job titles; and (3) N3, “refined” industries and job titles. Standardized industry and occupation codes were output by NIOCCS. Descriptive statistics of performance metrics (e.g., misclassification/discordance of occupation codes) were evaluated for each input relative to the original NoiseJEM dataset (N0).Across major Standardized Occupational Classifications (SOC), total discordance rates for N1, N2, and N3 compared to N0 were 53.6%, 42.3%, and 5.0%, respectively. The impact of discordance on major SOC group varied and included both over- and under-estimates of average noise exposure compared to N0. N2 had the most accurate noise exposure estimates (i.e., smallest bias) across major SOC groups compared to N1 and N3. Further refinement of job titles in N3 showed little improvement. Some variation in classification efficacy was seen over time, particularly prior to 1985.Machine learning algorithms can systematically and consistently classify data but are highly dependent on the quality and amount of input data. The greatest benefit for an end-user may come from cleaning industry information before applying this method for job classification. Our results highlight the need for standardized classification methods that remain constant over time.

PMID:35537195 | DOI:10.1080/15459624.2022.2076860