Principled Approaches to Missing Data in Epidemiologic Studies.

In Brief OBJECTIVE To investigate whether the rate of caffeine metabolism influences spontaneous abortion risk. METHODS We studied 101 women with normal karyotype spontaneous abortions and 953 pregnant women at 6–12 gestational weeks. Participants reported on caffeine intake and provided urine for phenotyping cytochrome P4501A2 (CYP1A2) activity and blood for genotyping N-acetylation (NAT2) status. We calculated odds ratios (OR) and 95% confidence intervals (CI) to evaluate the association between each of the two metabolic indices and spontaneous abortion risk and also the potential interaction between caffeine intake and metabolic activity on such risk. In calculating the associations between the metabolic indices and risk of spontaneous abortion, we had 80% power to detect an OR of 2.1, with a Type I error of 0.05. RESULTS Slow acetylators had a nonsignificantly increased risk for spontaneous abortion (OR 1.36, 95% CI 0.84, 2.21) and recurrent spontaneous abortion (OR 2.51, 95% CI 0.81, 7.76). In contrast, low CYP1A2 activity was associated with a significantly decreased risk for spontaneous abortion (OR 0.35, 95% CI 0.20, 0.63). Caffeine was a risk factor for spontaneous abortion among women with high, but not low, CYP1A2 activity (OR 2.42, 95% CI 1.01, 5.80 for 100–299 mg/day; OR 3.17, 95% CI 1.22, 8.22 for 300 mg/day or more, among women with high CYP1A2 activity). CONCLUSION The findings indicate that high CYP1A2 activity may increase the risk of spontaneous abortion, independently or by modifying the effect of caffeine. The results regarding NAT2 are less conclusive but suggest that slow acetylators may be at elevated risk of spontaneous abortion. A rapid rate of caffeine metabolism, as measured by cytochrome P4501A2 phenotype, is associated with an increased risk of first-trimester normal karyotype spontaneous abortion.

Multigenerational smoking during pregnancy and risk of stillbirth: a population-based 3-generation cohort study.

A doubly robust estimation framework to quantify potential bias in linked crash-EMS-trauma data with multi-cohort overlap.

Objective To determine the risk of recurrent stillbirth.Design Systematic review and meta-analysis of cohort and case-control studies.Data sources Embase, Medline, Cochrane Library, PubMed, CINAHL, and Scopus searched systematically with no...

Background:The National Trauma Data Bank (NTDB) is plagued by the problem of missing physiological data. The Glasgow Coma Scale score, Respiratory Rate and Systolic Blood Pressure are an essential part of risk adjustment strategies for trauma system evaluation and clinical research. Missing data on these variables may compromise the feasibility and the validity of trauma group comparisons.Aims:To evaluate the validity of Multiple Imputation (MI) for completing missing physiological data in the National Trauma Data Bank (NTDB), by assessing the impact of MI on 1) frequency distributions, 2) associations with mortality, and 3) risk adjustment.Methods:Analyses were based on 170,956 NTDB observations with complete physiological data (observed data set). Missing physiological data were artificially imposed on this data set and then imputed using MI (MI data set). To assess the impact of MI on risk adjustment, 100 pairs of hospitals were randomly selected with replacement and compared using adjusted Odds Ratios (OR) of mortality. OR generated by the observed data set were then compared to those generated by the MI data set.Results:Frequency distributions and associations with mortality were preserved following MI. The median absolute difference between adjusted OR of mortality generated by the observed data set and by the MI data set was 3.6% (inter-quartile range: 2.4%-6.1%).Conclusions:This study suggests that, provided it is implemented with care, MI of missing physiological data in the NTDB leads to valid frequency distributions, preserves associations with mortality, and does not compromise risk adjustment in inter-hospital comparisons of mortality.

Background Missing data is common in IBD trials. Depending on the volume and nature of missing data, it can reduce statistical power for detecting treatment difference, introduce potential bias and invalidate conclusions. Non-responder imputation (NRI), where patients (patients) with missing data are considered treatment failures, is widely used to handle missing data for dichotomous efficacy endpoints in IBD trials. However, it does not consider the mechanisms leading to missing data and can potentially underestimate the treatment effect. We proposed a hybrid (HI) approach combining NRI and multiple imputation (MI) as an alternative to NRI in the analyses of two phase 3 trials of vedolizumab (VDZ) in patients with moderate-to-severe UC – VISIBLE 11 and VARSITY2. Methods VISIBLE 1 and VARSITY assessed efficacy using dichotomous endpoints based on complete Mayo score. Full methodologies reported previously.1,2 Our proposed HI approach is aimed at imputing missing Mayo scores, instead of imputing the missing dichotomous efficacy endpoint. To assess the impact of dropouts for different missing data mechanisms (categorised as ‘missing not at random [MNAR]’ and ‘missing at random [MAR]’, HI was implemented as a potential sensitivity analysis, where dropouts owing to safety or lack of efficacy were imputed using NRI (assuming MNAR) and other missing data were imputed using MI (assuming MAR). For MI, each component of the Mayo score was imputed via a multivariate stepwise approach using a fully conditional specification ordinal logistic method. Missing baseline scores were imputed using baseline characteristics data. Missing scores from each subsequent visit were imputed using all previous visits in a stepwise fashion. Fifty imputation datasets were computed for each component of Mayo score. The complete Mayo score and relevant efficacy endpoints were derived subsequently. The analysis was performed within each imputed dataset to determine treatment difference, 95% CI and p-value, which were then combined via Rubin’s rules3. Results Tables 1 and 2 show a comparison of efficacy in the two studies using the primary NRI analysis vs. the alternative HI approach for handling missing data. Conclusion HI and NRI approaches can provide consistent efficacy analyses in IBD trials. The HI approach can serve as a useful sensitivity analysis to assess the impact of dropouts under different missing data mechanisms and evaluate the robustness of efficacy conclusions. Reference

BackgroundStillbirth, defined by foetal death at or beyond 28 weeks of gestation, represents a significant challenge in India, contributing to approximately 500,000 foetal deaths each year. The country’s stillbirth rate of 12.2 per 1000 births underscores the imperative to address this preventable occurrence. While maternal metabolic conditions diabetes, and hypertension, are widely recognized as established risk factors for stillbirth worldwide, the extent of their impact on India’s stillbirth burden remains inadequately elucidated due to limited evidence.MethodsThis cross-sectional study utilized NFHS-5 data to examine stillbirths in the most recent pregnancy outcomes of 204,723 women aged 15–49 years, sampled from all states and union territories of India. The primary exposures assessed were diabetes, and hypertension. Descriptive analyses were conducted to determine the prevalence of diabetes, hypertension and stillbirths. Logistic regression was used to quantify the association between diabetes, hypertension and the risk of stillbirth, indicated by adjusted odds ratios (AOR) with 95% confidence intervals (CI). The study also assessed effect modification by maternal age, education, wealth quintile, and social category.ResultsThe prevalence of diabetes and hypertension was 1% and 3% respectively, while the stillbirth rate was 1%. diabetes conferred a significantly higher risk of stillbirth with an increase of 74% (AOR 1.74, CI 1.14–2.67) as compared to women without diabetes. The risk was potential among mothers with hypertension with an increase of 50% (AOR 1.50, CI 1.16–1.95) on contrary to women without hypertension. The combined model (i.e. having diabetes or hypertension) also showed a significant risk of stillbirth with a higher risk of 58% (AOR 1.58, CI 1.25–1.99) indicating a synergistic interaction. Stratified analyses revealed the stillbirth risk among mothers belonging to the scheduled caste category (AOR 1.30, CI 1.10–1.53).ConclusionDiabetes, and hypertension, increase stillbirth risk in India, highlighting the need for better metabolic health management pre- and during pregnancy. Our research highlights the need of integrated care for diabetes and hypertension is crucial. Targeted interventions for high-risk mothers and improved screening are vital to reduce stillbirth rates. More research is needed to understand these risks better. Collaboration across medical fields is essential to save lives and improve pregnancy outcomes.

The effectiveness of interventions to increase the uptake of influenza vaccination in people aged 60 and older is uncertain. To assess access, provider, system and societal interventions to increase the uptake of influenza vaccination in people aged 60 years and older in the community. We searched CENTRAL (2014, Issue 5), MEDLINE (January 1950 to May week 3 2014), EMBASE (1980 to June 2014), AgeLine (1978 to 4 June 2014), ERIC (1965 to June 2014) and CINAHL (1982 to June 2014). Randomised controlled trials (RCTs) of interventions to increase influenza vaccination uptake in people aged 60 and older. Two review authors independently assessed study quality and extracted influenza vaccine uptake data. This update identified 13 new RCTs; the review now includes a total of 57 RCTs with 896,531 participants. The trials included community-dwelling seniors in high-income countries. Heterogeneity limited meta-analysis. The percentage of trials with low risk of bias for each domain was as follows: randomisation (33%); allocation concealment (11%); blinding (44%); missing data (49%) and selective reporting (100%). Increasing community demand (32 trials, 10 strategies)The interventions with a statistically significant result were: three trials (n = 64,200) of letter plus leaflet/postcard compared to letter (odds ratio (OR) 1.11, 95% confidence interval (CI) 1.07 to 1.15); two trials (n = 614) of nurses/pharmacists educating plus vaccinating patients (OR 3.29, 95% CI 1.91 to 5.66); single trials of a phone call from a senior (n = 193) (OR 3.33, 95% CI 1.79 to 6.22), a telephone invitation versus clinic drop-in (n = 243) (OR 2.72, 95% CI 1.55 to 4.76), a free groceries lottery (n = 291) (OR 1.04, 95% CI 0.62 to 1.76) and nurses educating and vaccinating patients (n = 485) (OR 152.95, 95% CI 9.39 to 2490.67).We did not pool the following trials due to considerable heterogeneity: postcard/letter/pamphlets (16 trials, n = 592,165); tailored communications (16 trials, n = 388,164); customised letter/phone-call (four trials, n = 82,465) and client-based appraisals (three trials, n = 4016), although several trials showed the interventions were effective. Enhancing vaccination access (10 trials, six strategies)The interventions with a statistically significant result were: two trials (n = 2112) of home visits compared to clinic invitation (OR 1.30, 95% CI 1.05 to 1.61); two trials (n = 2251) of free vaccine (OR 2.36, 95% CI 1.98 to 2.82) and one trial (n = 321) of patient group visits (OR 24.85, 95% CI 1.45 to 425.32). One trial (n = 350) of a home visit plus vaccine encouragement compared to a home visit plus safety advice was non-significant.We did not pool the following trials due to considerable heterogeneity: nurse home visits (two trials, n = 2069) and free vaccine compared to no intervention (two trials, n = 2250). Provider- or system-based interventions (17 trials, 11 strategies)The interventions with a statistically significant result were: two trials (n = 2815) of paying physicians (OR 2.22, 95% CI 1.77 to 2.77); one trial (n = 316) of reminding physicians about all their patients (OR 2.47, 95% CI 1.53 to 3.99); one trial (n = 8376) of posters plus postcards (OR 2.03, 95% CI 1.86 to 2.22); one trial (n = 1360) of chart review/feedback (OR 3.43, 95% CI 2.37 to 4.97) and one trial (n = 27,580) of educational outreach/feedback (OR 0.77, 95% CI 0.72 to 0.81).Trials of posters plus postcards versus posters (n = 5753), academic detailing (n = 1400) and increasing staff vaccination rates (n = 26,432) were non-significant.We did not pool the following trials due to considerable heterogeneity: reminding physicians (four trials, n = 202,264) and practice facilitators (three trials, n = 2183), although several trials showed the interventions were effective. Interventions at the societal level We identified no RCTs of interventions at the societal level. There are interventions that are effective for increasing community demand for vaccination, enhancing access and improving provider/system response. Heterogeneity limited pooling of trials.

Social network data often contain missing values because of the sensitive nature of the information collected and the dependency among the network actors. As a response, network imputation methods including simple ones constructed from network structural characteristics and more complicated model-based ones have been developed. Although past studies have explored the influence of missing data on social networks and the effectiveness of imputation procedures in many missing data conditions, the current study aims to evaluate a more extensive set of eight network imputation techniques (i.e., null-tie, Reconstruction, Preferential Attachment, Constrained Random Dot Product Graph, Multiple Imputation by Bayesian Exponential Random Graph Models or BERGMs, k-Nearest Neighbors, Random Forest, and Multiple Imputation by Chained Equations) under more practical conditions through comprehensive simulation. A factorial design for missing data conditions is adopted with factors including missing data types, missing data mechanisms, and missing data proportions, which are applied to generated social networks with varying numbers of actors based on 4 different sets of coefficients in ERGMs. Results show that the effectiveness of imputation methods differs by missing data types, missing data mechanisms, the evaluation criteria used, and the complexity of the social networks. More complex methods such as the BERGMs have consistently good performances in recovering missing edges that should have been present. While simpler methods like Reconstruction work better in recovering network statistics when the missing proportion of present edges is low, the BERGMs work better when more present edges are missing. The BERGMs also work well in recovering ERGM coefficients when the networks are complex and the missing data type is actor non-response. In conclusion, researchers analyzing social networks with incomplete data should identify the network structures of interest and the potential missing data types before selecting appropriate imputation methods.

BackgroundMissing data are a common problem in prospective studies with a long follow-up, and the volume, pattern and reasons for missing data may be relevant when estimating the cost of illness. We aimed to evaluate the effects of different methods for dealing with missing longitudinal cost data and for costing caregiver time on total societal costs in Alzheimer’s disease (AD).MethodsGERAS is an 18-month observational study of costs associated with AD. Total societal costs included patient health and social care costs, and caregiver health and informal care costs. Missing data were classified as missing completely at random (MCAR), missing at random (MAR) or missing not at random (MNAR). Simulation datasets were generated from baseline data with 10–40 % missing total cost data for each missing data mechanism. Datasets were also simulated to reflect the missing cost data pattern at 18 months using MAR and MNAR assumptions. Naïve and multiple imputation (MI) methods were applied to each dataset and results compared with complete GERAS 18-month cost data. Opportunity and replacement cost approaches were used for caregiver time, which was costed with and without supervision included and with time for working caregivers only being costed.ResultsTotal costs were available for 99.4 % of 1497 patients at baseline. For MCAR datasets, naïve methods performed as well as MI methods. For MAR, MI methods performed better than naïve methods. All imputation approaches were poor for MNAR data. For all approaches, percentage bias increased with missing data volume. For datasets reflecting 18-month patterns, a combination of imputation methods provided more accurate cost estimates (e.g. bias: −1 % vs −6 % for single MI method), although different approaches to costing caregiver time had a greater impact on estimated costs (29–43 % increase over base case estimate).ConclusionsMethods used to impute missing cost data in AD will impact on accuracy of cost estimates although varying approaches to costing informal caregiver time has the greatest impact on total costs. Tailoring imputation methods to the reason for missing data will further our understanding of the best analytical approach for studies involving cost outcomes.Electronic supplementary materialThe online version of this article (doi:10.1186/s12874-016-0188-1) contains supplementary material, which is available to authorized users.

Large-scale retrospective analysis suggests medications and dietary supplements, such as rosuvastatin, ezetimibe, ubiquinone, risedronate, vitamin D, and magnesium, are associated with a lower rate of severe COVID-19 disease.

Background Sarcopenia, the loss of muscle mass and function, and osteoporosis, a condition of low bone mass and micro-architectural deterioration of bone, frequently coexist and are associated with low functional status in heart failure (HF) patients. Aims We aimed to investigate the impact of coexistence of sarcopenia and osteoporosis on functional status in HF patients. Methods This cross-sectional study was conducted using data from patients who admitted to our institute for the diagnosis and management of HF from 1 November 2015 to 30 April 2021. All patients received the dual-energy X-ray absorptiometry (DEXA) method before discharge. The diagnosis of sarcopenia was made according to the criteria of Asia Working Group for Sarcopenia 2019 recommendation as follows: reduced skeletal muscle mass [appendicular skeletal muscle mass index (ASMI) by DEXA, <7.00 kg/m2 in males and <5.40 kg/m2 in females] plus lower muscle strength (handgrip strength, <28 kg in males and <18 kg in females) and/or poor physical performance (gait speed, <1.0 m/s; chair stand test time. ≥12 s; short physical performance battery ≤9 points). In addition, bone mineral densities (BMDs) at the lumbar spine, femoral neck, and total femur were measured by DEXA, and osteoporosis was diagnosed when BMDs at any of the three sites were less than 70% of Young Adult Mean (YAM). Functional status was assessed by the Barthel Index (BI) within three days before discharge, and patients with a BI score of <85 points was defined as having functional dependence (FD). The missing data were imputed using multiple imputation by chained -equations algorithm. Results Four hundred-thirty eight patients [median age of 74 years (interquartile range, 65–82 years), 37% females] were included in the analyses. Of these, percentage of HF patients with sarcopenia, osteoporosis, and sarcopenia and osteoporosis was 45%, 34%, and 20%, respectively (Figure 1A). The analysis of covariance showed a lower %YAM at any sites in patients with sarcopenia than those without sarcopenia (Figure 1B). When patients were divided into subgroups according to the presence or absence of sarcopenia and osteoporosis, the prevalence of FD was 32%, 34%, and 48% in patients with osteoporosis alone, sarcopenia alone, and sarcopenia and osteoporosis, respectively. Multivariate logistic regression analysis indicated that an increase in adjusted odds ratio (OR) for predicting FD was observed across subgroups in the following order: patients with osteoporosis alone [OR, 1.64; 95% confidence interval (CI), 0.63–4.24; p=0.31], those with sarcopenia alone (OR, 2.44; 95% CI, 1.13–5.25; p=0.02) and those with both conditions (OR, 3.34; 95% CI, 1.52–7.38; p<0.01) (Figure 2). Conclusion There was considerable overlap between sarcopenia and osteoporosis in HF patients, which appeared to be a risk factor for FD. Funding Acknowledgement Type of funding sources: None.

The combination of daptomycin with β-lactam antibiotics is more effective than daptomycin alone for vancomycin-resistant Enterococcus faecium bloodstream infection

To study the current status and risk factors of spontaneous abortion of women with Tibetan ethnicity at reproductive age in rural areas. Pregnant women who lived in Tibet were interviewed on their former reproductive history and being followed on the outcomes of pregnant from 2006 to 2012. Generalized Poisson regression model was used to explore the influencing factors of spontaneous abortion. OR value of the research factors was estimated and its 95% confidence interval counted. There were 1 557 pregnant women under this study, with a total number of 2 687 pregnancies and 2 382 productions. 171 women underwent spontaneous abortion, with a total number of 204 times, 93 women had histories of abortion, with a total number of 101 times. Among all the Tibetan pregnant women, the ratio between spontaneous abortion and pregnant women was 7.6% while the rate of spontaneous abortion was 7.9% . Ratio between the number of women who had experienced spontaneous abortion and the total number of women under survey was 11.0% . Pregnancy appeared the important reason on spontaneous abortion. The risk of having spontaneous abortion would increase among women who had more than 3 pregnancies. Results from Poisson regression analysis revealed that the odds ratio (OR) value before the adjustment was 3.921 (95% CI:2.553-6.021) but after the adjustment, it increased to 4.722 (95% CI:2.834-7.866). The increase of production time could reduce the risk of spontaneous abortion in women of childbearing age. Risks related to spontaneous abortion were associated with the number of pregnancies. Women seemed to have lower risk for spontaneous abortion after 2009, with OR value as 0.419 (95%CI:0.285-0.616) before, compared to aOR value as 0.580 (95%CI:0.380-0.885) after the adjustment Social-demographic characteristics was not found to be particularly associated with the occurrence of spontaneous abortion. Rate of spontaneous abortion among Tibetan women at childbearing age was not particularly high when comparing to those women living in the plain area such as Shanxi. However, in order to further reduce the incidence of spontaneous abortion among Tibetan women, approaches should include the following items:strengthening maternal health care, extending the spacing of pregnancy and reducing the frequency of pregnancy.

Phthalates (PAEs) are widespread persistent organic pollutants and endocrine disruptors. However, the associations between PAE exposure and the risk of miscarriage in humans are unclear, and an insufficient number of studies have evaluated the possible threshold or dose-dependent effects of first trimester PAE exposure on miscarriage risk. Our research measured the levels of mono-methyl phthalate (MMP), mono-ethyl phthalate, mono-isobutyl phthalate, MiBP mono-butyl phthalate (MBP), mono-octyl phthalate, mono-benzyl phthalate, mono(2-ethylhexyl) phthalate, mono(2-ethyl-5-oxohexyl) phthalate, and mono(2-ethyl-5-hydroxyhexyl) phthalate (MEHHP) in maternal urine collected in early gestation between 150 pregnancies ending in miscarriage and 150 pregnancies with live birth. We also estimated the odds ratios (ORs) and 95% confidence intervals (CIs) for miscarriage and each PAE as a continuous variable or quartile. A restrictive cubic splines was used to assess dose-dependent effects after controlling for maternal characteristics (e.g., age, educational level). we identified monotonically increasing dose-dependent effects of MEHHP and MMP on the risk of miscarriage. The largest effect estimates were approximately threefold higher for the highest MBP (OR = 2.57; 95% CI = 1.32-5.01) or MMP quartile (OR = 3.57; 95% CI = 1.82-7.00) and two-fold higher for the highest MEHHP quartile (OR = 2.12; 95% CI = 1.10-4.11). Our research preliminarily obtained possible thresholds of MBP, MEHHP, and MMP which were 18.07, 2.38, and 0.80µg/g Cr for the risk of miscarriage, respectively. First-trimester exposure to MBP, MEHHP, and MMP exceeding certain thresholds increases the risk of miscarriage.