A Review of Hot Deck Imputation for Survey Non-response

Peripheral magnetic stimulation for chronic peripheral neuropathic pain: A systematic review and meta-analysis

OBJECTIVES

To provide a systematic review of the literature on the effects of peripheral magnetic stimulation (PMS) in the treatment of chronic peripheral neuropathic pain.

METHODS

A systematic search of MEDLINE, EMBASE, CENTRAL, CINHAL, Web of Science, and ProQuest was conducted from inception to July 2023 to identify studies of any design published in English language that enrolled adult patients (≥18 years) that received PMS for treatment of a chronic peripheral neuropathic pain disorder (pain > 3 months).

RESULTS

Twenty-three studies were identified which included 15 randomized controlled trials (RCTs), five case series, two case reports, and one non-randomized trial. PMS regimens varied across studies and ranged from 5 to 240 min per session over 1 day to 1 year of treatment. Results across included studies were mixed, with some studies suggesting benefits while others showing no significant differences. Of nine placebo-controlled RCTs, four reported statistically significant findings in favor of PMS use. In the meta-analysis, PMS significantly reduced pain scores compared to control within 0-1 month of use (mean difference -1.64 on a 0-10 numeric rating scale, 95% confidence interval -2.73 to -0.56, p = 0.003, I2 = 94%, 7 studies [264 participants], very low quality of evidence), but not at the 1-3 months and >3 months of PMS use (very low and low quality of evidence, respectively). Minimal to no adverse effects were reported with PMS use.

DISCUSSION

There is limited and low-quality evidence to make definitive recommendations on PMS usage, however, the available data is encouraging, especially for short-term applications of this novel modality. Large high-quality randomized controlled trials are required to establish definitive efficacy and safety effects of PMS.

Treatments and Outcomes Among Patients with Sydenham Chorea: A Meta-Analysis

IMPORTANCE

Sydenham chorea is the most common acquired chorea of childhood worldwide; however, treatment is limited by a lack of high-quality evidence.

OBJECTIVES

To evaluate historical changes in the clinical characteristics of Sydenham chorea and identify clinical and treatment factors at disease onset associated with chorea duration, relapsing disease course, and functional outcome.

DATA SOURCES

The systematic search for this meta-analysis was conducted in PubMed, Embase, CINAHL, Cochrane Library, and LILACS databases and registers of clinical trials from inception to November 1, 2022 (search terms: [Sydenham OR Sydenham's OR rheumatic OR minor] AND chorea).

STUDY SELECTION

Published articles that included patients with a final diagnosis of Sydenham chorea (in selected languages).

DATA EXTRACTION AND SYNTHESIS

This study followed the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) reporting guideline. Individual patient data on clinical characteristics, treatments, chorea duration, relapse, and final outcome were extracted. Data from patients in the modern era (1945 through 2022) were entered into multivariable models and stratified by corticosteroid duration for survival analysis of chorea duration.

MAIN OUTCOMES AND MEASURES

The planned study outcomes were chorea duration at onset, monophasic course (absence of relapse after ≥24 months), and functional outcome (poor: modified Rankin Scale score 2-6 or persisting chorea, psychiatric, or behavioral symptoms at final follow-up after ≥6 months; good: modified Rankin Scale score 0-1 and no chorea, psychiatric, or behavioral symptoms at final follow-up).

RESULTS

In total, 1479 patients were included (from 307 articles), 1325 since 1945 (median [IQR] age at onset, 10 [8-13] years; 875 of 1272 female [68.8%]). Immunotherapy was associated with shorter chorea duration (hazard ratio for chorea resolution, 1.51 [95% CI, 1.05-2.19]; P = .03). The median chorea duration in patients receiving 1 or more months of corticosteroids was 1.2 months (95% CI, 1.2-2.0) vs 2.8 months (95% CI, 2.0-3.0) for patients receiving none (P = .004). Treatment factors associated with monophasic disease course were antibiotics (odds ratio [OR] for relapse, 0.28 [95% CI, 0.09-0.85]; P = .02), corticosteroids (OR, 0.32 [95% CI, 0.15-0.67]; P = .003), and sodium valproate (OR, 0.33 [95% CI, 0.15-0.71]; P = .004). Patients receiving at least 1 month of corticosteroids had significantly lower odds of relapsing course (OR, 0.10 [95% CI, 0.04-0.25]; P < .001). No treatment factor was associated with good functional outcome.

CONCLUSIONS AND RELEVANCE

In this meta-analysis of treatments and outcomes in patients with Sydenham chorea, immunotherapy, in particular corticosteroid treatment, was associated with faster resolution of chorea. Antibiotics, corticosteroids and sodium valproate were associated with a monophasic disease course. This synthesis of retrospective data should support the development of evidence-based treatment guidelines for patients with Sydenham chorea.

Relationship Between Central Sensitization-Related Symptoms and Pain-Related Disability After Cervical Spine Surgery: A Structural Equation Model

BACKGROUND

It is unknown if central sensitization (CS)-related symptoms have an intermediate role that might explain how disability develops from pain after cervical spinal surgery.

AIMS

The study aim was to investigate the role of CS-related symptoms in the relationship between pain and disability reported after cervical spinal surgery.

DESIGN

Cross-sectional study.

SETTINGS

Tertiary care spinal surgery center.

PARTICIPANTS/SUBJECTS

The participants included individuals with a cervical degenerative condition who had undergone surgery.

METHODS

The following patient-reported outcome measures were evaluated: (1) Numerical Rating Scale; (2) Neck Disability Index; and (3) Short Form of the Central Sensitization Inventory. A hypothesized model containing the CS-related symptoms and the relationships between pain and disability was constructed and tested by structural equation modeling.

RESULTS

Questionnaires were mailed to 280 individuals, and responses were obtained from 145 participants. Of these respondents, 99 (68.3%) were males and 46 (31.7%) were females, with a mean age of 64.4 ± 12.3 years. The latent variable for pain, represented by the neck (coefficient: 0.856, p < .001) and upper limb pain (0.568, p < .001), influenced CS-related symptoms (coefficient: 0.504, p < .001). Pain directly affected disability (coefficient: 0.497, p < .001) and indirectly through CS-related symptoms. Bootstrap analysis confirmed this indirect effect (point estimate: 2.85, 95% confidence interval: 1.04 to 6.30, p = .04).

CONCLUSIONS

The results revealed that neck and upper limb pain affected disabilities both directly and through CS-related symptoms. Future research should focus on the efficacy of biopsychosocial approaches for patients after cervical spine surgery with a high risk of disability due to CS-related symptoms.

Predicting absolute risk for a person with missing risk factors

We compared methods to project absolute risk, the probability of experiencing the outcome of interest in a given projection interval accommodating competing risks, for a person from the target population with missing predictors. Without missing data, a perfectly calibrated model gives unbiased absolute risk estimates in a new target population, even if the predictor distribution differs from the training data. However, if predictors are missing in target population members, a reference dataset with complete data is needed to impute them and to estimate absolute risk, conditional only on the observed predictors. If the predictor distributions of the reference data and the target population differ, this approach yields biased estimates. We compared the bias and mean squared error of absolute risk predictions for seven methods that assume predictors are missing at random (MAR). Some methods imputed individual missing predictors, others imputed linear predictor combinations (risk scores). Simulations were based on real breast cancer predictor distributions and outcome data. We also analyzed a real breast cancer dataset. The largest bias for all methods resulted from different predictor distributions of the reference and target populations. No method was unbiased in this situation. Surprisingly, violating the MAR assumption did not induce severe biases. Most multiple imputation methods performed similarly and were less biased (but more variable) than a method that used a single expected risk score. Our work shows the importance of selecting predictor reference datasets similar to the target population to reduce bias of absolute risk predictions with missing risk factors.

The effects of living arrangements and leisure activities on depressive symptoms of Chinese older adults: Evidence from panel data analysis

BACKGROUND

Livable environment and ways, as the necessary conditions for the elderly to enjoy their old age, have a significant impact on their mental health and happiness. It's crucial to understand how living arrangements affect depressive symptoms in China. Studies on how various leisure activities modify this association are yet limited.

METHODS

This study relies on panel data derived from the Chinese Longitudinal Healthy Longevity Survey (CLHLS), collected during waves spanning 2008/2009, 2011/2012, 2014, and 2018. The primary objective is to examine the relationship between living arrangements, leisure activities, and depressive symptoms of elderly individuals. Linear mixed-effects models were used to analyze the data.

RESULTS

A total of 26,342 observations aged 65 and over were included in this paper. Older adults living alone (β = 0.66, 95 % CI: 0.55, 0.76) or living in institutions (β = 0.69, 95 % CI: 0.40, 0.98) had more depressive symptoms than those living with family. Leisure activities were negatively associated with depressive symptoms (β = -0.16, 95 % CI: -0.18, -0.15). Moreover, there was significant interactions between living arrangements and leisure activities. No matter which kind of living arrangements, participating in physical, productive or social activity was associated with a lower risk of depressive symptoms.

LIMITATIONS

Study design might introduce bias, and it cannot establish causality between the tested variables.

CONCLUSIONS

Older adults living alone or in institutions have more possibility to develop depressive symptoms than those living with family, and such a relationship among Chinese older adults can be moderated by participating in leisure activities.

Decomposition of medical imaging spending growth between 2010 and 2021 in the US employer-insured population

Medical imaging, identified as a potential driver of unsustainable US health care spending growth, was subject to policies to reduce prices and use in low-value settings. Meanwhile, the Affordable Care Act increased access to preventive services-many involving imaging-for employer-sponsored insurance (ESI) beneficiaries. We used a large insurance claims database to examine imaging spending trends in the ESI population between 2010 and 2021-a period of considerable policy and benefits changes. Nominal spending on imaging increased 35.9% between 2010 and 2021, but as a share of total health care spending fell from 10.5% to 8.9%. The 22.5% growth of nominal imaging prices was below inflation, 24.3%, as measured by the Consumer Price Index. Other key contributors to imaging spending growth were increased use (7.4 percentage points [pp]), shifts toward advanced modalities (4.0 pp), and demographic changes (3.5 pp). Shifts in care settings and provider network participation resulted in 2.5-pp and 0.3-pp imaging spending decreases, respectively. In sum, imaging spending decreased as a share of all health care spending and relative to inflation, as intended by concurrent cost-containment policies.

Machine Learning Multicenter Risk Model to Predict Right Ventricular Failure After Mechanical Circulatory Support: The STOP-RVF Score

Importance

The existing models predicting right ventricular failure (RVF) after durable left ventricular assist device (LVAD) support might be limited, partly due to lack of external validation, marginal predictive power, and absence of intraoperative characteristics.

Objective

To derive and validate a risk model to predict RVF after LVAD implantation.

Design, Setting, and Participants

This was a hybrid prospective-retrospective multicenter cohort study conducted from April 2008 to July 2019 of patients with advanced heart failure (HF) requiring continuous-flow LVAD. The derivation cohort included patients enrolled at 5 institutions. The external validation cohort included patients enrolled at a sixth institution within the same period. Study data were analyzed October 2022 to August 2023.

Exposures

Study participants underwent chronic continuous-flow LVAD support.

Main Outcome and Measures

The primary outcome was RVF incidence, defined as the need for RV assist device or intravenous inotropes for greater than 14 days. Bootstrap imputation and adaptive least absolute shrinkage and selection operator variable selection techniques were used to derive a predictive model. An RVF risk calculator (STOP-RVF) was then developed and subsequently externally validated, which can provide personalized quantification of the risk for LVAD candidates. Its predictive accuracy was compared with previously published RVF scores.

Results

The derivation cohort included 798 patients (mean [SE] age, 56.1 [13.2] years; 668 male [83.7%]). The external validation cohort included 327 patients. RVF developed in 193 of 798 patients (24.2%) in the derivation cohort and 107 of 327 patients (32.7%) in the validation cohort. Preimplant variables associated with postoperative RVF included nonischemic cardiomyopathy, intra-aortic balloon pump, microaxial percutaneous left ventricular assist device/venoarterial extracorporeal membrane oxygenation, LVAD configuration, Interagency Registry for Mechanically Assisted Circulatory Support profiles 1 to 2, right atrial/pulmonary capillary wedge pressure ratio, use of angiotensin-converting enzyme inhibitors, platelet count, and serum sodium, albumin, and creatinine levels. Inclusion of intraoperative characteristics did not improve model performance. The calculator achieved a C statistic of 0.75 (95% CI, 0.71-0.79) in the derivation cohort and 0.73 (95% CI, 0.67-0.80) in the validation cohort. Cumulative survival was higher in patients composing the low-risk group (estimated <20% RVF risk) compared with those in the higher-risk groups. The STOP-RVF risk calculator exhibited a significantly better performance than commonly used risk scores proposed by Kormos et al (C statistic, 0.58; 95% CI, 0.53-0.63) and Drakos et al (C statistic, 0.62; 95% CI, 0.57-0.67).

Conclusions and Relevance

Implementing routine clinical data, this multicenter cohort study derived and validated the STOP-RVF calculator as a personalized risk assessment tool for the prediction of RVF and RVF-associated all-cause mortality.

Health-Related Quality of Life is Associated With Pain, Kinesiophobia, and Physical Activity in Individuals Who Underwent Cervical Spine Surgery

OBJECTIVE

To determine the association between health-related quality of life (HRQOL) and neck pain, kinesiophobia, and modalities of physical activity in individuals with postoperative degenerative cervical myelopathy and radiculopathy (DCM/R) because postoperative pain after cervical spine surgery is likely to persist, causing kinesiophobia and avoidance of physical activity.

METHODS

A questionnaire was distributed to 280 individuals with DCM/R. The questionnaire comprised the following four items: HRQOL (EuroQol 5-dimensions 5-level), neck pain (numerical rating scale [NRS]), kinesiophobia (11-item Tampa Scale for Kinesiophobia [TSK-11]), and physical activity (paid work, light exercise, walking, strength training, and gardening). Hierarchical multiple regression analysis was performed using the NRS, TSK-11, and physical activity as independent variables.

RESULTS

In total, 126 individuals provided analyzable responses (45.0%). After including the NRS score as an independent variable to the multiple regression equation for participants' background, the independent rate of the regression equation significantly improved by only 4.1% (R2=0.153). The addition of the TSK-11 score significantly improved this effect by 11.1% (R2=0.264). Finally, the addition of physical activity also significantly improved the explanatory rate by 9.9% (R2=0.363).

CONCLUSION

Neck pain, kinesiophobia, and physical activity (specifically paid work and walking) were independently associated with HRQOL in individuals with postoperative DCM/R.

Femicide, intimate partner femicide, and non-intimate partner femicide in South Africa: An analysis of 3 national surveys, 1999-2017

BACKGROUND

In most countries, reliable national statistics on femicide, intimate partner femicide (IPF), and non-intimate partner femicide (NIPF) are not available. Surveys are required to collect robust data on this most extreme consequence of intimate partner violence (IPV). We analysed 3 national surveys to compare femicide, IPF, and NIPF from 1999 to 2017 using age-standardised rates (ASRs) and incidence rate ratios (IRRs).

METHODS AND FINDINGS

We conducted 3 national mortuary-based retrospective surveys using weighted cluster designs from proportionate random samples of medicolegal laboratories. We included females 14 years and older who were identified as having been murdered in South Africa in 1999 (n = 3,793), 2009 (n = 2,363), and 2017 (n = 2,407). Further information on the murdered cases were collected from crime dockets during interviews with police investigating officers. Our findings show that South Africa had an IPF rate of 4.9/100,000 female population in 2017. All forms of femicide among women 14 years and older declined from 1999 to 2017. For IPF, the ASR was 9.5/100,000 in 1999. Between 1999 and 2009, the decline for NIPF was greater than for IPF (IRR for NIPF 0.47 (95% confidence interval (CI) 0.42 to 0.53) compared to IRR for IPF 0.69 (95% CI 0.63 to 0.77). Rates declined from 2009 to 2017 and did not differ by femicide type. The decline in IPF was initially larger for women aged 14 to 29, and after 2009, it was more pronounced for those aged 30 to 44 years. Study limitations include missing data from the police and having to use imputation to account for missing perpetrator data.

CONCLUSIONS

In this study, we observed a reduction in femicide overall and different patterns of change in IPF compared to NIPF. The explanation for the reductions may be due to social and policy interventions aimed at reducing IPV overall, coupled with increased social and economic stability. Our study shows that gender-based violence is preventable even in high-prevalence settings, and evidence-based prevention efforts must be intensified globally. We also show the value of dedicated surveys in the absence of functional information systems.

Genomic prediction and allele mining of agronomic and morphological traits in pea (Pisum sativum) germplasm collections

Well-performing genomic prediction (GP) models for polygenic traits and molecular marker sets for oligogenic traits could be useful for identifying promising genetic resources in germplasm collections, setting core collections, and establishing molecular variety distinction. This study aimed at (i) defining GP models and key marker sets for predicting 15 agronomic or morphological traits in germplasm collections, (ii) verifying the GP model usefulness also for selection in breeding programs, (iii) investigating the consistency between molecular and phenotypic diversity patterns, and (iv) identifying genomic regions associated with to the target traits. The study was based on phenotyping data and over 41,000 genotyping-by-sequencing-generated SNP markers of 220 landraces or old cultivars belonging to a world germplasm collection and 11 modern cultivars. Non-metric multi-dimensional scaling (NMDS) and an analysis of population genetic structure indicated a high level of genetic differentiation of material from Western Asia, a major West-East diversity gradient, and quite limited genetic diversity of the improved germplasm. Mantel's test revealed a low correlation (r = 0.12) between phenotypic and molecular diversity, which increased (r = 0.45) when considering only the molecular diversity relative to significant SNPs from genome-wide association analyses. These analyses identified, inter alia, several areas of chromosome 6 involved in a largely pleiotropic control of vegetative or reproductive organ pigmentation. We found various significant SNPs for grain and straw yield under severe drought and onset of flowering, and one SNP on chromosome 5 for grain protein content. GP models displayed moderately high predictive ability (0.43 to 0.61) for protein content, grain and straw yield, and onset of flowering, and high predictive ability (0.76) for individual seed weight, based on intra-population, intra-environment cross-validations. The inter-population, inter-environment assessment of the models trained on the germplasm collection for breeding material of three recombinant inbred line (RIL) populations, which was challenged by much narrower diversity of the material, over eight-fold less available markers and quite different test environments, led to an overall loss of predictive ability of about 40% for seed weight, 50% for protein content and straw yield, and 60% for onset of flowering, and no prediction for grain yield. Within-RIL population predictive ability differed among populations.

Changes in life satisfaction among middle-aged adults living alone over a 12-year span

This secondary analysis used data collected for the Korean Longitudinal Study of Aging from 2006 to 2018 to examine changes in life satisfaction among middle-aged adults living alone in South Korea. Individuals who were over 45 years of age, lived alone at the time of the first data collection wave, and responded at least twice to the survey over the 12-year study period were included in the final linear mixed model (N = 124). Life satisfaction increased for those who had increased assets, were widowed, and had more frequent contact with acquaintances (i.e., once a month and once a week compared with once a year). Life satisfaction decreased as the number of chronic illnesses increased for underweight individuals compared with normal weight or overweight individuals and for depressed versus non-depressed individuals. This study's findings indicate that increased social support is beneficial for middle-aged marginalized individuals, including those who are economically disadvantaged, have few social interactions, are underweight, and have chronic illnesses.

Prevalence of COVID-19 Mitigation Behaviors in US Adults (August-December 2020): Nationwide Household Probability Survey

BACKGROUND

COVID-19 mitigation behaviors, such as wearing masks, maintaining social distancing, and practicing hand hygiene, have been and will remain vital to slowing the pandemic.

OBJECTIVE

This study aims to describe the period prevalence of consistent mask-wearing, social distancing, and hand hygiene practices during the peak of COVID-19 incidence (August-December 2020) and just before COVID-19 vaccine availability, overall and in demographic subgroups.

METHODS

We used baseline survey data from a nationwide household probability sample to generate weighted estimates of mitigation behaviors: wearing masks, maintaining social distancing, and practicing hand hygiene. Weighted logistic regression explored differences in mitigation behaviors by demographics. Latent class analysis (LCA) identified patterns in mitigation behaviors.

RESULTS

Among 4654 participants, most (n=2727, 58.6%) were female, were non-Hispanic White (n=3063, 65.8%), were aged 55 years or older (n=2099, 45.1%), lived in the South (n=2275, 48.9%), lived in metropolitan areas (n=4186, 89.9%), had at least a bachelor's degree (n=2547, 54.7%), had an income of US $50,000-$99,000 (n=1445, 31%), and were privately insured (n=2734, 58.7%). The period prevalence of consistent mask wearing was 71.1% (sample-weighted 95% CI 68.8-73.3); consistent social distancing, 42.9% (95% CI 40.5-45.3); frequent handwashing, 55.0% (95% CI 52.3-57.7); and frequent hand sanitizing, 21.5% (95% CI 19.4-23.8). Mitigation behaviors were more prevalent among women, older persons, Black or Hispanic persons, those who were not college graduates, and service-oriented workers. LCA identified an optimal-mitigation class that consistently practiced all behaviors (n=2656, 67% of US adults), a low-mitigation class that inconsistently practiced all behaviors (n=771, 20.6%), and a class that had optimal masking and social distancing but a high frequency of hand hygiene (n=463, 12.4%).

CONCLUSIONS

Despite a high prevalence of COVID-19 mitigation behaviors, there were likely millions who did not consistently practice these behaviors during the time of the highest COVID-19 incidence. In future infectious disease outbreak responses, public health authorities should also consider addressing disparities in mitigation practices through more targeted prevention messaging.

Imputation-based Q-learning for optimizing dynamic treatment regimes with right-censored survival outcome

Q-learning has been one of the most commonly used methods for optimizing dynamic treatment regimes (DTRs) in multistage decision-making. Right-censored survival outcome poses a significant challenge to Q-Learning due to its reliance on parametric models for counterfactual estimation which are subject to misspecification and sensitive to missing covariates. In this paper, we propose an imputation-based Q-learning (IQ-learning) where flexible nonparametric or semiparametric models are employed to estimate optimal treatment rules for each stage and then weighted hot-deck multiple imputation (MI) and direct-draw MI are used to predict optimal potential survival times. Missing data are handled using inverse probability weighting and MI, and the nonrandom treatment assignment among the observed is accounted for using a propensity-score approach. We investigate the performance of IQ-learning via extensive simulations and show that it is more robust to model misspecification than existing Q-Learning methods, imputes only plausible potential survival times contrary to parametric models and provides more flexibility in terms of baseline hazard shape. Using IQ-learning, we developed an optimal DTR for leukemia treatment based on a randomized trial with observational follow-up that motivated this study.

The Effectiveness of the StaySafe Intervention Using a Paradigm for Predicting Missing Outcome Data

Sample attrition is a confounding issue in the analysis of data collected in follow-up studies. The present study uses a regression procedure that includes a propensity score as a predictor in estimating imputed data. The utility of the procedure was addressed by comparing results from this augmented data with those from the original data. Data were from a randomized controlled study testing the utility of a tablet-based intervention designed to improve decision-making with respect to health risk behaviors. Outcomes included self-reported testing for HIV, STD, and hepatitis. Two samples were used (163 in community facilities and 348 in residential facilities). Seventy-eight in the community sample and 238 in the residential sample completed follow-up surveys. Propensity scores based on a stepwise logistic regression were used to make the calibration sample and the missing data sample as close as possible. Multilevel analysis was performed for each outcome and multiple imputation compared estimated mean differences for the augmented and original analyses. The model imputing missing data was effective for the three outcomes and increased power. Least square mean differences between augmented and original data appeared to be essentially the same for most of the outcomes. This protocol has been registered with https://www.clinicaltrials.gov/(NCT02777086).

Development and internal validation of a risk prediction model for HIV disease severity among people living with HIV and mental illness or substance use disorder

PURPOSE

Mental illness (MI) and substance use disorders (SUD) are highly prevalent among people living with HIV (PLWH), and have been linked to poor HIV clinical outcomes. Innovative tools for early risk identification can facilitate timely interventions for PLWH and MI/SUD to improve their health outcomes, however, this is currently lacking in Texas, a state with the 4th largest population of PLWH in the United States. To address this gap, we developed a predictive model to estimate the risk of suboptimal HIV clinical outcomes among PLWH and MI/SUD in Texas.

METHODS

The Texas Medical Monitoring Project data obtained from June 2015-May 2020 were used to develop and internally validate the predictive model. Univariate descriptive and bivariate inferential statistics were performed to describe the characteristics of the study population and unadjusted associations with HIV clinical outcomes. Multivariable logistic regression was used to develop the prediction model. Internal validation was performed using the bootstrap method.

RESULTS

A total of 518 respondents aged 18 years and above, representing 27,255 adults living with HIV and mental illness or substance use disorders in Texas were included. Most participants were male (77.0%), less than 50 years of age (60.0%), and had mild diagnosed mental illness and substance use disorder (54.8%). The risk predictive model contained eight predictors, which together yielded an area under the receiver operating characteristic (ROC) curve of 0.727. Non-retention in care appeared to be the strongest risk predictor for having suboptimal HIV clinical outcome (adjusted odds ratio (aOR) = 3.27; 95% confidence interval (CI) = 1.45, 7.42).

CONCLUSIONS

The predictive model had good discrimination between persons at risk of poor HIV clinical outcomes and those not at risk.

A Hybrid Missing Data Imputation Method for Batch Process Monitoring Dataset

Batch process monitoring datasets usually contain missing data, which decreases the performance of data-driven modeling for fault identification and optimal control. Many methods have been proposed to impute missing data; however, they do not fulfill the need for data quality, especially in sensor datasets with different types of missing data. We propose a hybrid missing data imputation method for batch process monitoring datasets with multi-type missing data. In this method, the missing data is first classified into five categories based on the continuous missing duration and the number of variables missing simultaneously. Then, different categories of missing data are step-by-step imputed considering their unique characteristics. A combination of three single-dimensional interpolation models is employed to impute transient isolated missing values. An iterative imputation based on a multivariate regression model is designed for imputing long-term missing variables, and a combination model based on single-dimensional interpolation and multivariate regression is proposed for imputing short-term missing variables. The Long Short-Term Memory (LSTM) model is utilized to impute both short-term and long-term missing samples. Finally, a series of experiments for different categories of missing data were conducted based on a real-world batch process monitoring dataset. The results demonstrate that the proposed method achieves higher imputation accuracy than other comparative methods.

A cross-sectional examination of post-myocardial infarction physical activity levels among US rural and urban residents: Findings from the 2017-2019 Behavioral Risk Factor Surveillance System

BACKGROUND

This study sought to examine the relationship between rural residence and physical activity levels among US myocardial infarction (MI) survivors.

METHODS

We conducted a cross-sectional study using nationally representative Behavioral Risk Factor Surveillance System surveys from 2017 and 2019. We determined the survey-weighted percentage of rural and urban MI survivors meeting US physical activity guidelines. Logistic regression models were used to examine the relationship between rural/urban residence and meeting physical activity guidelines, accounting for sociodemographic factors.

RESULTS

Our study included 22,732 MI survivors (37.3% rural residents). The percentage of rural MI survivors meeting physical activity guidelines (37.4%, 95% CI: 35.1%-39.7%) was significantly less than their urban counterparts (45.6%, 95% CI: 44.0%-47.2%). Rural residence was associated with a 28.8% (95% CI: 20.0%-36.7%) lower odds of meeting physical activity guidelines, with this changing to a 19.3% (95% CI: 9.3%-28.3%) lower odds after adjustment for sociodemographic factors.

CONCLUSIONS

A significant rural/urban disparity in physical activity levels exists among US MI survivors. Our findings support the need for further efforts to improve physical activity levels among rural MI survivors as part of successful secondary prevention in US high-MI burden rural areas.

Multiple imputation approaches for epoch-level accelerometer data in trials

Clinical trials that investigate physical activity interventions often use accelerometers to measure step count at a very granular level, for example in 5-second epochs. Participants typically wear the accelerometer for a week-long period at baseline, and for one or more week-long follow-up periods after the intervention. The data is aggregated to provide daily or weekly step counts for the primary analysis. Missing data are common as participants may not wear the device as per protocol. Approaches to handling missing data in the literature have defined missingness on the day level using a threshold on daily weartime, which leads to loss of information on the time of day when data are missing. We propose an approach to identifying and classifying missingness at the finer epoch-level and present two approaches to handling missingness using multiple imputation. Firstly, we present a parametric approach which accounts for the number of missing epochs per day. Secondly, we describe a non-parametric approach where missing periods during the day are replaced by donor data from the same person where possible, or data from a different person who is matched on demographic and physical activity-related variables. Our simulation studies show that the non-parametric approach leads to estimates of the effect of treatment that are least biased while maintaining small standard errors. We illustrate the application of these different multiple imputation strategies to the analysis of the 2017 PACE-UP trial. The proposed framework is likely to be applicable to other digital health outcomes and to other wearable devices.

Development and validation of a real-world model to predict 1-year Level 3 (severe) hypoglycaemia risk in adults with diabetes (the iNPHORM study, United States)

AIMS

We aimed to develop and internally validate a real-world prognostic model for Level 3 hypoglycaemia risk compatible with outpatient care in the United States.

MATERIALS AND METHODS

iNPHORM is a 12-month, US-based panel survey. Adults (18-90 years old) with type 1 diabetes mellitus or insulin- and/or secretagogue-treated type 2 diabetes mellitus were recruited from a nationwide, probability-based internet panel. Among participants completing ≥ 1 follow-up questionnaire(s), we modelled 1-year Level 3 hypoglycaemia risk using Andersen and Gill's Cox survival and penalized regression with multiple imputation. Candidate variables were selected for their clinical relevance and ease of capture at point-of-care.

RESULTS

In total, 986 participants [type 1 diabetes mellitus: 17%; men: 49.6%; mean age: 51 (SD: 14.3) years] were analysed. Across follow-up, 035.1 (95% CI: 32.2-38.1)% reported ≥1 Level 3 event(s), and the rate was 5.0 (95% CI: 4.1-6.0) events per person-year. Our final model showed strong discriminative validity and parsimony (optimism corrected c-statistic: 0.77). Numerous variables were selected: age; sex; body mass index; marital status; level of education; insurance coverage; race; ethnicity; food insecurity; diabetes type; glycated haemoglobin value; glycated haemoglobin variability; number, type and dose of various medications; number of SH events requiring hospital care (past year and over follow-up); type and number of comorbidities and complications; number of diabetes-related health care visits (past year); use of continuous/flash glucose monitoring; and general health status.

CONCLUSIONS

iNPHORM is the first US-based primary prognostic study on Level 3 hypoglycaemia. Future model implementation could potentiate risk-tailored strategies that reduce real-world event occurrence and overall diabetes burden.

Out-of-Pocket Medical Expenditures in the Redesigned Current Population Survey: Evaluating Improvements to Data Processing

Household surveys are an important source of information on medical spending and burden. We examine how recently implemented post-processing improvements to the Current Population Survey Annual Social and Economic Supplement (CPS ASEC) affected estimates of medical expenditures and medical burden. The revised data extraction and imputation procedures mark the second stage of the CPS ASEC redesign and the beginning of a new time series for studying household medical expenditures. Using data for the calendar year 2017, we find that median family medical expenditures are not statistically different from legacy methods; however, updated processing does significantly reduce the percentage of families estimated to have a high medical burden (medical expenses are at least 10% of family income). The updated processing system also changes the characteristics of families with high medical spending and is primarily driven by changes in imputation of health insurance and medical spending.

Catastrophic health expenditure and its association with socioeconomic status in China: evidence from the 2011-2018 China Health and Retirement Longitudinal Study

BACKGROUND

An increase in healthcare utilization in response to universal health coverage may leave massive economic burden on individuals and households. Identifying catastrophic health expenditure helps us understand such burden. This study aims to examine the incidence of catastrophic health expenditure at various thresholds, explore its trend over years, and investigate whether it varies across socioeconomic status (SES).

METHODS

Data used in this study were from four waves of the China Health and Retirement Longitudinal Study (CHARLS): 2011, 2013, 2015, and 2018. SES was measured by annual per-capita household expenditure, which was then divided into quintiles (Quintile 1 (Q1): the poorest - Quintile 5 (Q5): the wealthiest). Catastrophic health expenditure was measured at both a fixed threshold (40%) and a set of variable thresholds, where the thresholds for other quintiles were estimated by multiplying 40% by the ratio of average food expenditure in certain quintile to that in the index quintile. Multilevel mixed-effects logistic regression models were used to analyze the determinants of catastrophic health expenditure at various thresholds.

RESULTS

A total of 6,953 households were included in our study. The incidence of catastrophic health expenditure varied across the thresholds set. At a fixed threshold, 10.90%, 9.46%, 13.23%, or 24.75% of households incurred catastrophic health expenditure in 2011, 2013, 2015, and 2018, respectively, which were generally lower than those at variable thresholds. Catastrophic health expenditure often decreased from 2011 to 2013, and an increasing trend occurred afterwards. Compared to households in Q5, those in lower quintiles were more likely to suffer catastrophic health expenditure, irrespective of the thresholds set. Similarly, having chronic diseases and healthcare utilization increased the odds of catastrophic health expenditure.

CONCLUSIONS

The financial protection against catastrophic health expenditure shocks remains a challenge in China, especially for the low-SES and those with chronic diseases. Concerted efforts are needed to further expand health insurance coverage across breadth, depth, and height, optimize health financing mechanism, redesign cost-sharing arrangements and provider payment methods, and develop more efficient expenditure control strategies.

A national examination of suicidal ideation, planning, and attempts among United States adults: Differences by military veteran status, 2008-2019

There is a widening disparity in suicide deaths between United States (U.S.) military veterans and nonveterans. However, it is unclear if there are similar differences in suicidal ideation, planning, and attempts that often precipitate these deaths. A better understanding of trends in suicidal thoughts and behaviors could illuminate opportunities for prevention. We examined pooled cross-sectional data (N = 479,801 adults) from the 2008 to 2019 National Survey on Drug Use and Health. We examined differences in past-year suicidal ideation, suicide planning, and suicide attempts between U.S. veterans (n = 26,508) and nonveterans (n = 453,293). We conducted post hoc analyses to examine for differences in these relationships by race/ethnicity and sex. Lastly, we examined trends in these outcomes over time and tested for differences in trends by veteran status. Overall, veterans had significantly greater odds of past-year suicidal ideation (aOR = 1.33, 95% CI 1.20 to 1.47) and suicide planning (aOR = 1.52, 95% CI 1.30 to 1.78) compared to nonveterans. However, the association between veteran status and past-year suicide attempt was not statistically significant (aOR = 1.29, 95% CI 1.00 to 1.68). These relationships did not differ by race/ethnicity or sex (ps > 0.05). Among all adults, there were significant linear increases in past-year suicidal ideation, planning, and attempts (ps < 0.001). However, these trends did not differ between veterans and nonveterans (ps > 0.05). Veterans may be more likely to experience suicidal thoughts and behaviors than nonveteran adults. Upward trends in suicidal thoughts and behaviors among both veterans and nonveterans from 2008 to 2019 highlight opportunities for intervention.

A comparison of multiple imputation strategies to deal with missing nonnormal data in structural equation modeling

Missing data and nonnormality are two common factors that can affect analysis results from structural equation modeling (SEM). The current study aims to address a challenging situation in which the two factors coexist (i.e., missing nonnormal data). Using Monte Carlo simulation, we evaluated the performance of four multiple imputation (MI) strategies with respect to parameter and standard error estimation. These strategies include MI with normality-based model (MI-NORM), predictive mean matching (MI-PMM), classification and regression trees (MI-CART), and random forest (MI-RF). We also compared these MI strategies with robust full information maximum likelihood (RFIML), a popular (non-imputation) method to deal with missing nonnormal data in SEM. The results suggest that MI-NORM had similar performance to RFIML. MI-PMM outperformed the other methods when data were not missing on the heavy tail of a skewed distribution. Although MI-CART and MI-RF do not require any distribution assumption, they did not perform well compared with the others. Based on the results, practical guidance is provided.

Disparities in Food Security and Glycemic Control Among People with Type 2 Diabetes During the COVID-19 Pandemic

Background

Little is known about the differing impacts of food insecurity on HbA1c by race in type 2 diabetes (T2D). Predictions around increased food insecurity from COVID-19 exacerbating racial disparities led us to estimate its prevalence and associations with HbA1c by race during the COVID-19 pandemic.

Methods

Data came from medical records and surveys among a clinic-based sample of T2D patients. Linear regression models estimated associations between food insecurity and HbA1c and between change in food insecurity and change in HbA1c. Likelihood ratio tests and examination of stratum-specific estimates assessed effect modification by race.

Results

Our sample was 59% White, 59% female, and mean age was 60.8 ± 12.6. During the pandemic, food insecurity prevalence and HbA1c were significantly (p < .05) higher among non-Whites (39%, 8.4% ± 2.1) compared to Whites (15%, 7.8% ±1.6). HbA1c among those who were very food insecure was 1.00% (95% CI: 0.222, 1.762, p = .01) higher than those who were food secure. Those with increased food insecurity had a 0.58% (95% CI: 0.024, 1.128, p = .04) higher HbA1c increase than among those experiencing no change. No effect modification was detected.

Limitations

Convenience sampling in an endocrinology clinic, recall bias, and inadequate power may underlie null effect modification results.

Conclusion

Although effect modification was not detected, racial disparities in HbA1c and food insecurity warrant further investigation. These disparities, combined with the significant impact of food insecurity on HbA1c, suggest that prioritization of resources to high-risk populations should be considered early during public emergencies to minimize short- and long-term health consequences.

Comparative effectiveness of safety planning intervention with instrumental support calls (ISC) versus safety planning intervention with two-way text message caring contacts (CC) in adolescents and adults screening positive for suicide risk in emergency departments and primary care clinics: Protocol for a pragmatic randomized controlled trial

BACKGROUND

Suicide is a leading cause of death in adolescents and adults in the US. Follow-up support delivered when patients return home after an emergency department (ED) or primary care encounter can significantly reduce suicidal ideation and attempts. Two follow-up models to augment usual care including the Safety Planning Intervention have high efficacy: Instrumental Support Calls (ISC) and Caring Contacts (CC) two-way text messages, but they have never been compared to assess which works best. This protocol for the Suicide Prevention Among Recipients of Care (SPARC) Trial aims to determine which model is most effective for adolescents and adults with suicide risk.

METHODS

The SPARC Trial is a pragmatic randomized controlled trial comparing the effectiveness of ISC versus CC. The sample includes 720 adolescents (12-17 years) and 790 adults (18+ years) who screen positive for suicide risk during an ED or primary care encounter. All participants receive usual care and are randomized 1:1 to ISC or CC. The state suicide hotline delivers both follow-up interventions. The trial is single-masked, with participants unaware of the alternative treatment, and is stratified by adolescents/adults. The primary outcome is suicidal ideation and behavior, measured using the Columbia Suicide Severity Rating Scale (C-SSRS) screener at 6 months. Secondary outcomes include C-SSRS at 12 months, and loneliness, return to crisis care for suicidality, and utilization of outpatient mental health services at 6 and 12 months.

DISCUSSION

Directly comparing ISC and CC will determine which follow-up intervention is most effective for suicide prevention in adolescents and adults.

Personalized Longitudinal Assessment of Multiple Sclerosis Using Smartphones

Personalized longitudinal disease assessment is central to quickly diagnosing, appropriately managing, and optimally adapting the therapeutic strategy of multiple sclerosis (MS). It is also important for identifying idiosyncratic subject-specific disease profiles. Here, we design a novel longitudinal model to map individual disease trajectories in an automated way using smartphone sensor data that may contain missing values. First, we collect digital measurements related to gait and balance, and upper extremity functions using sensor-based assessments administered on a smartphone. Next, we treat missing data via imputation. We then discover potential markers of MS by employing a generalized estimation equation. Subsequently, parameters learned from multiple training datasets are ensembled to form a simple, unified longitudinal predictive model to forecast MS over time in previously unseen people with MS. To mitigate potential underestimation for individuals with severe disease scores, the final model incorporates additional subject-specific fine-tuning using data from the first day. The results show that the proposed model is promising to achieve personalized longitudinal MS assessment; they also suggest that features related to gait and balance as well as upper extremity function, remotely collected from sensor-based assessments, may be useful digital markers for predicting MS over time.

Comparing Individualized Survival Predictions From Random Survival Forests and Multistate Models in the Presence of Missing Data: A Case Study of Patients With Oropharyngeal Cancer

Background

In recent years, interest in prognostic calculators for predicting patient health outcomes has grown with the popularity of personalized medicine. These calculators, which can inform treatment decisions, employ many different methods, each of which has advantages and disadvantages.

Methods

We present a comparison of a multistate model (MSM) and a random survival forest (RSF) through a case study of prognostic predictions for patients with oropharyngeal squamous cell carcinoma. The MSM is highly structured and takes into account some aspects of the clinical context and knowledge about oropharyngeal cancer, while the RSF can be thought of as a black-box non-parametric approach. Key in this comparison are the high rate of missing values within these data and the different approaches used by the MSM and RSF to handle missingness.

Results

We compare the accuracy (discrimination and calibration) of survival probabilities predicted by both approaches and use simulation studies to better understand how predictive accuracy is influenced by the approach to (1) handling missing data and (2) modeling structural/disease progression information present in the data. We conclude that both approaches have similar predictive accuracy, with a slight advantage going to the MSM.

Conclusions

Although the MSM shows slightly better predictive ability than the RSF, consideration of other differences are key when selecting the best approach for addressing a specific research question. These key differences include the methods' ability to incorporate domain knowledge, and their ability to handle missing data as well as their interpretability, and ease of implementation. Ultimately, selecting the statistical method that has the most potential to aid in clinical decisions requires thoughtful consideration of the specific goals.

Association between maternal labour epidural analgesia and autistic traits in offspring

STUDY OBJECTIVE

Studies investigating associations between maternal epidural analgesia (MEA) and autism spectrum disorder (ASD) in the offspring are conflicting and lack prospective neurobehavioral follow-up assessments for autistic traits. We aim to prospectively investigate associations between MEA and autistic traits in the offspring.

DESIGN

Prospective neurobehavioral observational cohort study.

SETTING

Singaporean tertiary healthcare institutions.

PATIENTS

Participants recruited were singleton non-IVF children, >36 weeks gestation, delivered via normal vaginal delivery by mothers >18 years of age, delivered in Singapore from June 2009-September 2010 and followed up over 7 years.

INTERVENTIONS

Exposure to maternal epidural analgesia during delivery.

MEASUREMENTS

The primary outcome is an abnormal Social Responsiveness Scale (SRS) T score at 7 years (≥60 points). Secondary outcomes include the diagnosis of ASD and abnormal scores for autistic traits assessed via a neurobehavioral battery comprising: CBCL (child behavioural checklist), Q-CHAT (Quantitative Checklist for Autism in Toddlers), and Bayley-III. Multivariable analyses adjusting for maternal and offspring characteristics were performed.

MAIN RESULTS

704 out of 769 mother-child dyads recruited fulfilled the criteria for analysis. 365/704 mothers received MEA. The incidence of an abnormal SRS score at 7 years in offspring exposed to MEA was 19.9%, and 26.1% in non-exposed offspring (p = 0.154). Multivariable analysis did not demonstrate a significant association between MEA and abnormal SRS scores at 7 years (O.R.0.726, 95% C·I. 0.394-1.34, p = 0.305). After adjustment for maternal and fetal demographics, exposure to MEA was not significantly associated with an abnormal screen in all other tests for autistic traits. The clinical incidence of ASD was 1.76% in children without exposure to MEA, and 2.32% in children with MEA exposure (p = 0.506).

CONCLUSIONS

MEA is not significantly associated with the development of ASD and autistic traits in offspring, assessed over 7 years. Results should be taken into perspective given our wide confidence intervals and small cohort size.

Missing data approaches in longitudinal studies of aging: A case example using the National Health and Aging Trends Study

PURPOSE

Missing data is a key methodological consideration in longitudinal studies of aging. We described missing data challenges and potential methodological solutions using a case example describing five-year frailty state transitions in a cohort of older adults.

METHODS

We used longitudinal data from the National Health and Aging Trends Study, a nationally-representative cohort of Medicare beneficiaries. We assessed the five components of the Fried frailty phenotype and classified frailty based on their number of components (robust: 0, prefrail: 1-2, frail: 3-5). One-, two-, and five-year frailty state transitions were defined as movements between frailty states or death. Missing frailty components were imputed using hot deck imputation. Inverse probability weights were used to account for potentially informative loss-to-follow-up. We conducted scenario analyses to test a range of assumptions related to missing data.

RESULTS

Missing data were common for frailty components measured using physical assessments (walking speed, grip strength). At five years, 36% of individuals were lost-to-follow-up, differentially with respect to baseline frailty status. Assumptions for missing data mechanisms impacted inference regarding individuals improving or worsening in frailty.

CONCLUSIONS

Missing data and loss-to-follow-up are common in longitudinal studies of aging. Robust epidemiologic methods can improve the rigor and interpretability of aging-related research.

Matrix completion under complex survey sampling

Multivariate nonresponse is often encountered in complex survey sampling, and simply ignoring it leads to erroneous inference. In this paper, we propose a new matrix completion method for complex survey sampling. Different from existing works either conducting row-wise or column-wise imputation, the data matrix is treated as a whole which allows for exploiting both row and column patterns simultaneously. A column-space-decomposition model is adopted incorporating a low-rank structured matrix for the finite population with easy-to-obtain demographic information as covariates. Besides, we propose a computationally efficient projection strategy to identify the model parameters under complex survey sampling. Then, an augmented inverse probability weighting estimator is used to estimate the parameter of interest, and the corresponding asymptotic upper bound of the estimation error is derived. Simulation studies show that the proposed estimator has a smaller mean squared error than other competitors, and the corresponding variance estimator performs well. The proposed method is applied to assess the health status of the U.S. population.

Medication use and contextual factors associated with meeting guideline-based glycemic levels in diabetes among a nationally representative sample

Introduction

Based on the long-lasting diabetes management challenges in the United States, the objective was to examine glycemic levels among a nationally representative sample of people with diabetes stratified by prescribed antihyperglycemic treatment regimens and contextual factors.

Methods

This serial cross-sectional study used United States population-based data from the 2015 to March 2020 National Health and Nutrition Examination Surveys (NHANES). The study included non-pregnant adults (≥20 years old) with non-missing A1C and self-reported diabetes diagnosis from NHANES. Using A1C lab values, we dichotomized the outcome of glycemic levels into <7% versus ≥7% (meeting vs. not meeting guideline-based glycemic levels, respectively). We stratified the outcome by antihyperglycemic medication use and contextual factors (e.g., race/ethnicity, gender, chronic conditions, diet, healthcare utilization, insurance, etc.) and performed multivariable logistic regression analyses.

Results

The 2042 adults with diabetes had a mean age of 60.63 (SE = 0.50), 55.26% (95% CI = 51.39-59.09) were male, and 51.82% (95% CI = 47.11-56.51) met guideline-based glycemic levels. Contextual factors associated with meeting guideline-based glycemic levels included reporting an "excellent" versus "poor" diet (aOR = 4.21, 95% CI = 1.92-9.25) and having no family history of diabetes (aOR = 1.43, 95% CI = 1.03-1.98). Contextual factors associated with lower odds of meeting guideline-based glycemic levels included taking insulin (aOR = 0.16, 95% CI = 0.10-0.26), taking metformin (aOR = 0.66, 95% CI = 0.46-0.96), less frequent healthcare utilization [e.g., none vs. ≥4 times/year (aOR = 0.51, 95% CI = 0.27-0.96)], being uninsured (aOR = 0.51, 95% CI = 0.33-0.79), etc.

Discussion

Meeting guideline-based glycemic levels was associated with medication use (taking vs. not taking respective antihyperglycemic medication classes) and contextual factors. The timely, population-based estimates can inform national efforts to optimize diabetes management.

Impact of Smoking, Body Weight, Diabetes, Hypertension and Kidney Dysfunction on Survival in Pancreatic Cancer Patients-A Single Center Analysis of 2323 Patients within the Last Decade

In addition to being risk factors for pancreatic cancer, parameters such as smoking, diabetes, or obesity might also act as potential prognostic factors for the survival of patients initially diagnosed with pancreatic cancer. By implementing one of the largest retrospective study cohorts of 2323 pancreatic adenocarcinoma (PDAC) patients treated at a single high-volume center, potential prognostic factors for survival were evaluated on the basis of 863 cases. Since parameters such as smoking, obesity, diabetes, and hypertension can cause severe chronic kidney dysfunction, the glomerular filtration rate was also considered. In the univariate analyses, albumin (p < 0.001), active smoking (p = 0.024), BMI (p = 0.018), and GFR (p = 0.002) were identified as metabolic prognostic markers for overall survival. In multivariate analyses, albumin (p < 0.001) and chronic kidney disease stage 2 (GFR < 90 mL/min/1.37 m²; p = 0.042) were identified as independent metabolic prognostic markers for survival. Smoking presented a nearly statistically significant independent prognostic factor for survival with a p-value of 0.052. In summary, low BMI, status of active smoking, and reduced kidney function at the time of diagnosis were associated with lower overall survival. No prognostic association could be observed for presence of diabetes or hypertension.

Perceived Caregiving Trajectories and their Relationship with Caregivers' Burdens and Gains

The study explores the meanings that family caregivers of people with dementia ascribe to the past, present, and future of their role as a caregiver, and how their integration into caregiving trajectories is related to caregivers' burdens and gains. The sample was made up of 197 family caregivers (M_age = 62.1, SD = 12.3, 70.1% females). They completed three incomplete sentences regarding their past, present, and future caring role, the Zarit Burden Interview and the Gains Associated with Caregiving scale. Sentence completions were content analyzed, and the associations between the resulting trajectories and burdens and gains were studied by means of a one-way ANOVA. Caregivers differed in the meanings ascribed to past, present, and future of their role. Stable-negative (M = 43.6, SD = 13.3), regressive (M = 43.3, SD = 12.7), and present-enhancing (M = 37.4, SD = 13.7) trajectories showed higher levels of burdens than progressive (M = 31.3, SD = 12.3) and/or stable-positive trajectories (M = 26.1, SD = 13.7). Progressive trajectories (M = 38.9, SD = 15.7) were related to more gains than regressive trajectories (M = 28.6, SD = 12.7). Family caregivers' evaluations of their past, present, and future are not only important separately, but their combination into caregiving trajectories is also relevant. Such trajectories might be relevant when designing interventions to help caregivers reduce their burden levels and increase the benefits ascribed to their experience. The most adaptive trajectory identified was the progressive one, whereas the regressive trajectory was the most dysfunctional.

Prior Incarceration Is Associated with Poor Mental Health at Midlife: Findings from a National Longitudinal Cohort Study

BACKGROUND

People with mental illnesses and people living in poverty have higher rates of incarceration than others, but relatively little is known about the long-term impact that incarceration has on an individual's mental health later in life.

OBJECTIVE

To evaluate prior incarceration's association with mental health at midlife.

DESIGN

Retrospective cohort study PARTICIPANTS: Participants from the National Longitudinal Survey of Youth 1979 (NLSY79)-a nationally representative age cohort of individuals 15 to 22 years of age in 1979-who remained in follow-up through age 50.

MAIN MEASURES

Midlife mental health outcomes were measured as part of a health module administered once participants reached 50 years of age (2008-2019): any mental health history, any depression history, past-year depression, severity of depression symptoms in the past 7 days (Center for Epidemiologic Studies Depression [CES-D] scale), and mental health-related quality of life in the past 4 weeks (SF-12 Mental Component Score [MCS]). The main exposure was any incarceration prior to age 50.

KEY RESULTS

Among 7889 participants included in our sample, 577 (5.4%) experienced at least one incarceration prior to age 50. Prior incarceration was associated with a greater likelihood of having any mental health history (predicted probability 27.0% vs. 16.6%; adjusted odds ratio [aOR] 1.9 [95%CI: 1.4, 2.5]), any history of depression (22.0% vs. 13.3%; aOR 1.8 [95%CI: 1.3, 2.5]), past-year depression (16.9% vs. 8.6%; aOR 2.2 [95%CI: 1.5, 3.0]), and high CES-D score (21.1% vs. 15.4%; aOR 1.5 [95%CI: 1.1, 2.0]) and with a lower (worse) SF-12 MCS (-2.1 points [95%CI: -3.3, -0.9]; standardized mean difference -0.24 [95%CI: -0.37, -0.10]) at age 50, when adjusting for early-life demographic, socioeconomic, and behavioral factors.

CONCLUSIONS

Prior incarceration was associated with worse mental health at age 50 across five measured outcomes. Incarceration is a key social-structural driver of poor mental health.

Food insecurity, SNAP participation and glycemic control in low-income adults with predominantly type 2 diabetes: a cross-sectional analysis using NHANES 2007-2018 data

INTRODUCTION

Diabetes, characterized by elevated blood glucose levels, affects 13% of US adults, 95% of whom have type 2 diabetes (T2D). Social determinants of health (SDoH), such as food insecurity, are integral to glycemic control. The Supplemental Nutrition Assistance Program (SNAP) aims to reduce food insecurity, but it is not clear how this affects glycemic control in T2D. This study investigated the associations between food insecurity and other SDoH and glycemic control and the role of SNAP participation in a national socioeconomically disadvantaged sample.

RESEARCH DESIGN AND METHODS

Adults with likely T2D and income <185% of the federal poverty level (FPL) were identified using cross-sectional National Health and Nutrition Examination Survey (NHANES) data (2007-2018). Multivariable logistic regression assessed the association between food insecurity, SNAP participation and glycemic control (defined by HbA1c 7.0%-8.5% depending on age and comorbidities). Covariates included demographic factors, clinical comorbidities, diabetes management strategies, and healthcare access and utilization.

RESULTS

The study population included 2084 individuals (90% >40 years of age, 55% female, 18% non-Hispanic black, 25% Hispanic, 41% SNAP participants, 36% low or very low food security). Food insecurity was not associated with glycemic control in the adjusted model (adjusted OR (aOR) 1.181 (0.877-1.589)), and SNAP participation did not modify the effect of food insecurity on glycemic control. Insulin use, lack of health insurance, and Hispanic or another race and ethnicity were among the strongest associations with poor glycemic control in the adjusted model.

CONCLUSIONS

For low-income individuals with T2D in the USA, health insurance may be among the most critical predictors of glycemic control. Additionally, SDoH associated with race and ethnicity plays an important role. SNAP participation may not affect glycemic control because of inadequate benefit amounts or lack of incentives for healthy purchases. These findings have implications for community engaged interventions and healthcare and food policy.

Using machine-learning algorithms to improve imputation in the medical expenditure panel survey

OBJECTIVE

To assess the feasibility of applying machine learning (ML) methods to imputation in the Medical Expenditure Panel Survey (MEPS).

DATA SOURCES

All data come from the 2016-2017 MEPS.

STUDY DESIGN

Currently, expenditures for medical encounters in the MEPS are imputed with a predictive mean matching (PMM) algorithm in which a linear regression model is used to predict expenditures for events with (donors) and without (recipients) data. Recipient events and donor events are then matched based on the smallest distance between predicted expenditures, and the donor event's expenditures are used as the recipient event's imputation. We replace linear regression algorithm in the PMM framework with ML methods to predict expenditures. We examine five alternatives to linear regression: Gradient Boosting, Random Forests, Extreme Random Forests, Deep Neural Networks, and a Stacked Ensemble approach. Additionally, we introduce an alternative matching scheme, which matches on a vector of predicted expenditures by sources of payment instead of a single total expenditure prediction to generate potentially superior matches.

DATA COLLECTION

Study data is derived from a large federal survey.

PRINCIPAL FINDINGS

ML algorithms perform better at both prediction and matching imputation than Ordinary Least Squares (OLS), the most common prediction algorithm used in PMM. On average, the Stacked Ensemble approach that combines all the ML algorithms performs best, improving expenditure prediction R2 by 108% (0.156 points) and final imputation R2 by 227% (0.397 points). Matching on a prediction vector also improves alignment of sources of payments between donor and recipient events.

CONCLUSIONS

ML algorithms and an alternative matching scheme improve the overall quality of expenditure PMM imputation in the MEPS. These methods may have additional value in other national surveys that currently rely on PMM or similar methods for imputation.

Psychometric Properties of the Prenatal Opioid Use Perceived Stigma Scale and Its Use in Prenatal Care

OBJECTIVE

To examine the psychometric properties of the Prenatal Opioid Use Perceived Stigma (POPS) scale and to assess the relationship of POPS scores to adequate prenatal care.

DESIGN

Prospective cohort study.

SETTING

Medical centers in Alabama, Ohio, and Pennsylvania (N = 4).

PARTICIPANTS

Women (N = 127) who took opioids during pregnancy and whose infants participated in the Outcomes of Babies With Opioid Exposure Study.

METHODS

Participants reported their perceptions of stigma during pregnancy by responding to the eight items on the POPS scale. We evaluated the instrument's internal consistency reliability (Cronbach's alpha), structural validity (factor analysis), and convergent validity (relationship with measures of similar constructs). In addition, to assess construct validity, we used logistic regression to examine the relationship of POPS scores to the receipt of adequate prenatal care.

RESULTS

The internal consistency of the POPS scale was high (Cronbach's α = .88), and all item-total correlations were greater than 0.50. The factor analysis confirmed that the items cluster into one factor. Participants who reported greater perceived stigma toward substance users and everyday discrimination in medical settings had higher POPS scores, which supported the convergent validity of the scale. POPS scores were significantly associated with not receiving adequate prenatal care, adjusted OR = 1.47, 95% confidence interval [1.19, 1.83], p < .001.

CONCLUSION

The psychometric testing of the POPS scale provided initial support for the reliability and validity of the instrument. It may be a useful tool with which to assess perceived stigma among women who take opioids, a potential barrier to seeking health care during pregnancy.

Evaluation of a Social Media Campaign Designed to Increase Awareness of Thirdhand Smoke among California Adults

Despite a growing body of research outlining the harms of thirdhand smoke (THS), the public remains generally unaware of risks and exposure routes. This project built on past tobacco prevention campaigns and the tenants of McGuire's input-output model to implement and evaluate a seven-month Facebook-disseminated campaign seeking to improve THS awareness among California adults (n = 1087). Multilinear regression showed that THS-related knowledge (χ2[6] = 19.31, p < .01), attitude (χ2[6] = 13.88, p < .05), and efficacy (χ2[6] = 13.81, p < .05) significantly increased by the campaign's end, with messages highlighting children's health (r = .110, p < .05), pets (r = .145, p < .01), and dust reservoirs (r = .144, p < .01) as the most persuasive. Path analysis modeling found campaign recall to be associated with changes in knowledge (β = .161, p < .01), which predicated attitude change (β = .614, p < .001) and, in turn, behavior change (β = .149, p < .05). Findings suggest social media campaigns should continue to educate diverse populations about new tobacco risks and that tobacco control advocates should consider integrating educational THS messages.

Multiple Imputation with Massive Data: An Application to the Panel Study of Income Dynamics

Multiple imputation (MI) is a popular and well-established method for handling missing data in multivariate data sets, but its practicality for use in massive and complex data sets has been questioned. One such data set is the Panel Study of Income Dynamics (PSID), a longstanding and extensive survey of household income and wealth in the United States. Missing data for this survey are currently handled using traditional hot deck methods because of the simple implementation; however, the univariate hot deck results in large random wealth fluctuations. MI is effective but faced with operational challenges. We use a sequential regression/chained-equation approach, using the software IVEware, to multiply impute cross-sectional wealth data in the 2013 PSID, and compare analyses of the resulting imputed data with those from the current hot deck approach. Practical difficulties, such as non-normally distributed variables, skip patterns, categorical variables with many levels, and multicollinearity, are described together with our approaches to overcoming them. We evaluate the imputation quality and validity with internal diagnostics and external benchmarking data. MI produces improvements over the existing hot deck approach by helping preserve correlation structures, such as the associations between PSID wealth components and the relationships between the household net worth and sociodemographic factors, and facilitates completed data analyses with general purposes. MI incorporates highly predictive covariates into imputation models and increases efficiency. We recommend the practical implementation of MI and expect greater gains when the fraction of missing information is large.

Neighbourhood socioeconomic disadvantage and body size in Australia's capital cities: The contribution of obesogenic environments

Residents of socioeconomically disadvantaged neighbourhoods have higher rates of overweight and obesity and chronic disease than their counterparts from advantaged neighbourhoods. This study assessed whether associations between neighbourhood disadvantage and measured body mass index (BMI) and waist circumference, are accounted for by obesogenic environments (i.e., residential distance to the Central Business District [CBD], supermarket availability, access to walkable destinations). The study used 2017-18 National Health Survey data for working-aged adults (aged ≥18 years, n = 9,367) residing in 3,454 neighbourhoods across Australia's state and territory capital cities. In five of eight cities (i.e., Sydney, Melbourne, Brisbane, Adelaide, and Perth) residents of disadvantaged neighbourhoods had significantly higher BMI and a larger waist circumference than residents of more advantaged areas. There was no association between neighbourhood disadvantage and body size in Hobart, Darwin, and Canberra. Associations between neighbourhood disadvantage and body size were partially explained by neighbourhood differences in distance to the CBD but not supermarket availability or walkable amenities. The results of this study point to the role of urban design and city planning as mechanisms for addressing social and economic inequities in Australia's capital cities, and as solutions to this country's overweight and obesity epidemic and associated rising rates of chronic disease.

Racial and Ethnic Differences in COVID-19 Vaccination Coverage Among Children and Adolescents Aged 5-17 Years and Parental Intent to Vaccinate Their Children - National Immunization Survey-Child COVID Module, United States, December 2020-September 2022

Some racial and ethnic groups are at increased risk for COVID-19 and associated hospitalization and death because of systemic and structural inequities contributing to higher prevalences of high-risk conditions and increased exposure (1). Vaccination is the most effective prevention intervention against COVID-19-related morbidity and mortality*; ensuring more equitable vaccine access is a public health priority. Differences in adult COVID-19 vaccination coverage by race and ethnicity have been previously reported (2,3), but similar information for children and adolescents is limited (4,5). CDC analyzed data from the National Immunization Survey-Child COVID Module (NIS-CCM) to describe racial and ethnic differences in vaccination status, parental intent to vaccinate their child, and behavioral and social drivers of vaccination among children and adolescents aged 5-17 years. By August 31, 2022, approximately one third (33.2%) of children aged 5-11 years, more than one half (59.0%) of children and adolescents aged 12-15 years, and more than two thirds (68.6%) of adolescents aged 16-17 years had received ≥1 COVID-19 vaccine dose. Vaccination coverage was highest among non-Hispanic Asian (Asian) children and adolescents, ranging from 63.4% among those aged 5-11 years to 91.8% among those aged 16-17 years. Coverage was next highest among Hispanic or Latino (Hispanic) children and adolescents (34.5%-77.3%). Coverage was similar for non-Hispanic Black or African American (Black), non-Hispanic White (White), and non-Hispanic other race^† or multiple race (other/multiple race) children and adolescents aged 12-15 and 16-17 years. Among children aged 5-11 years, coverage among Black children was lower than that among Hispanic, Asian, and other/multiple race children. Enhanced public health efforts are needed to increase COVID-19 vaccination coverage for all children and adolescents. To address disparities in child and adolescent COVID-19 vaccination coverage, vaccination providers and trusted messengers should provide culturally relevant information and vaccine recommendations and build a higher level of trust among those groups with lower coverage.

Using machine learning to modify and enhance the daily living questionnaire

The Daily Living Questionnaire (DLQ) constitutes one of a number of functional cognitive measures, commonly employed in a range of medical and rehabilitation settings. One of the drawbacks of the DLQ is its length which poses an obstacle to conducting efficient and widespread screening of the public and which incurs inaccuracies due to the length and fatigue of the subjects.

Objective

This study aims to use Machine Learning (ML) to modify and abridge the DLQ without compromising its fidelity and accuracy.

Method

Participants were interviewed in two separate research studies conducted in the United States of America and Israel, and one unified file was created for ML analysis. An ML-based Computerized Adaptive Testing (ML-CAT) algorithm was applied to the DLQ database to create an adaptive testing instrument-with a shortened test form adapted to individual test scores.

Results

The ML-CAT approach was shown to reduce the number of tests required on average by 25% per individual when predicting each of the seven DLQ output scores independently and reduce by over 50% when predicting all seven scores concurrently using a single model. These results maintained an accuracy of 95% (5% error) across subject scores. The study pinpoints which DLQ items are more informative in predicting DLQ scores.

Conclusions

Applying the ML-CAT model can thus serve to modify, refine and even abridge the current DLQ, thereby enabling wider community screening while also enhancing clinical and research utility.

U.S. Blue Water Navy Veterans of the Vietnam War: Comparisons from the Vietnam Era Health Retrospective Observational Study (VE-HEROeS)

Background

US Vietnam War Blue Water Navy veterans (BWN) conducted military operations on Vietnam's offshore waters and likely experienced various war-related exposures. The overall health of the BWN has never been systematically studied.

Purpose

Describe and compare BWN's health with other servicemembers and non-veterans of the Vietnam era.

Materials and methods

Survey of 45 067 randomly selected US Vietnam War theatre and non-theatre veterans and 6885 non-veterans.

Results

For 22 646 male respondents, self-reported health was contrasted by veteran status defined as BWN (n=985), theatre veterans (n=6717), non-theatre veterans (n=10 698) and non-veterans (n=4246). Exposure was service in the Vietnam War theatre. Collected were demographics, military service characteristics, lifestyle factors and health conditions. Adjusted odds ratios (aOR) were calculated using multivariable logistic regression. Controlling for cigarette smoking and other covariates, respiratory cancer risk was highest in BWN vs other veterans (theatre: aOR 1.65; 95% CI 1.09, 2.50; non-theatre: aOR 1.77; 1.13, 2.77) and to non-veterans (aOR 1.78; 1.15, 2.74). Other findings showed BWN's health risks between theatre and non-theatre veterans.

Conclusion

There was a higher risk for respiratory cancers in BWN. Other risks were less than theatre veterans but greater than non-theatre or non-veterans, indicating a potential role of military exposures in BWN's health.

Review and evaluation of imputation methods for multivariate longitudinal data with mixed-type incomplete variables

Estimating relationships between multiple incomplete patient measurements requires methods to cope with missing values. Multiple imputation is one approach to address missing data by filling in plausible values for those that are missing. Multiple imputation procedures can be classified into two broad types: joint modeling (JM) and fully conditional specification (FCS). JM fits a multivariate distribution for the entire set of variables, but it may be complex to define and implement. FCS imputes missing data variable-by-variable from a set of conditional distributions. In many studies, FCS is easier to define and implement than JM, but it may be based on incompatible conditional models. Imputation methods based on multilevel modeling show improved operating characteristics when imputing longitudinal data, but they can be computationally intensive, especially when imputing multiple variables simultaneously. We review current MI methods for incomplete longitudinal data and their implementation on widely accessible software. Using simulated data from the National Health and Aging Trends Study, we compare their performance for monotone and intermittent missing data patterns. Our simulations demonstrate that in a longitudinal study with a limited number of repeated observations and time-varying variables, FCS-Standard is a computationally efficient imputation method that is accurate and precise for univariate single-level and multilevel regression models. When the analyses comprise multivariate multilevel models, FCS-LMM-latent is a statistically valid procedure with overall more accurate estimates, but it requires more intensive computations. Imputation methods based on generalized linear multilevel models can lead to biased subject-level variance estimates when the statistical analyses involve hierarchical models.

Development and validation of the Chinese Geriatric Depression Risk calculator (CGD-risk): A screening tool to identify elderly Chinese with depression

BACKGROUND

The prevalence of depression among China's elderly is high, but stigma surrounding mental illness and a shortage of psychiatrists limit widespread screening and diagnosis of geriatric depression. We sought to develop a screening tool using easy-to-obtain and minimally sensitive predictors to identify elderly Chinese with depressive symptoms (depression hereafter) for referral to mental health services and determine the most important factors for effective screening.

METHODS

Using nationally representative survey data, we developed and externally validated the Chinese Geriatric Depression Risk calculator (CGD-Risk). CGD-Risk, a gradient boosting machine learning model, was evaluated based on discrimination (Concordance (C) statistic), calibration, and through a decision curve analysis. We conducted a sensitivity analysis on a cohort of middle-aged Chinese, a sub-group analysis using three data sets, and created predictor importance and partial dependence plots to enhance interpretability.

RESULTS

A total of 5681 elderly Chinese were included in the development data and 12,373 in the external validation data. CGD-Risk showed good discrimination during internal validation (C: 0.81, 95 % CI 0.79 to 0.84) and external validation (C: 0.77, 95 % CI: 0.76, 0.78). Compared to an alternative screening strategy CGD-Risk would correctly identify 17.8 more elderly with depression per 100 people screened.

LIMITATIONS

We were only able to externally validate a partial version of CGD-Risk due to differences between the internal and external validation data.

CONCLUSIONS

CGD-Risk is a clinically viable, minimally sensitive screening tool that could identify elderly Chinese at high risk of depression while circumventing issues of response bias from stigma surrounding emotional openness.

Cultural patterns of evasive answer bias in surveys

Complete and accurate survey data are key input for research, policy, and decision making in many disciplines. However, survey respondents do not always fully cooperate, such that they skip some items or overuse the “don’t know” answer option. Evasive answer bias reflects different information than overall survey response rates, leading to item missing data and causing substantial inaccuracies in survey results. Using data from the World Values Survey, this article identifies the magnitude of the problem, then relies on individual data and country-level cultural values to derive patterns of and reasons for this evasive answer bias. While skipping answers happens less often in collectivistic and low power distance cultures, the choice of the “don’t know” option is not significantly influenced by any cultural dimension. Across countries, the effect of cultural values is stronger for female than for male respondents. Accordingly, cross-cultural researchers are advised to use advanced imputation rather than deletion methods for handling missing data.

Classification of breast cancer recurrence based on imputed data: a simulation study

Several studies have been conducted to classify various real life events but few are in medical fields; particularly about breast recurrence under statistical techniques. To our knowledge, there is no reported comparison of statistical classification accuracy and classifiers' discriminative ability on breast cancer recurrence in presence of imputed missing data. Therefore, this article aims to fill this analysis gap by comparing the performance of binary classifiers (logistic regression, linear and quadratic discriminant analysis) using several datasets resulted from imputation process using various simulation conditions. Our study aids the knowledge about how classifiers' accuracy and discriminative ability in classifying a binary outcome variable are affected by the presence of imputed numerical missing data. We simulated incomplete datasets with 15, 30, 45 and 60% of missingness under Missing At Random (MAR) and Missing Completely At Random (MCAR) mechanisms. Mean imputation, hot deck, k-nearest neighbour, multiple imputations via chained equation, expected-maximisation, and predictive mean matching were used to impute incomplete datasets. For each classifier, correct classification accuracy and area under the Receiver Operating Characteristic (ROC) curves under MAR and MCAR mechanisms were compared. The linear discriminant classifier attained the highest classification accuracy (73.9%) based on mean-imputed data at 45% of missing data under MCAR mechanism. As a classifier, the logistic regression based on predictive mean matching imputed-data yields the greatest areas under ROC curves (0.6418) at 30% missingness while k-nearest neighbour tops the value (0.6428) at 60% of missing data under MCAR mechanism.

Population-based correlates of COVID-19 infection: An analysis from the DFW COVID-19 prevalence study

BACKGROUND

COVID-19 has resulted in over 1 million deaths in the U.S. as of June 2022, with continued surges after vaccine availability. Information on related attitudes and behaviors are needed to inform public health strategies. We aimed to estimate the prevalence of COVID-19, risk factors of infection, and related attitudes and behaviors in a racially, ethnically, and socioeconomically diverse urban population.

METHODS

The DFW COVID-19 Prevalence Study Protocol 1 was conducted from July 2020 to March 2021 on a randomly selected sample of adults aged 18-89 years, living in Dallas or Tarrant Counties, Texas. Participants were asked to complete a 15-minute questionnaire and COVID-19 PCR and antibody testing. COVID-19 prevalence estimates were calculated with survey-weighted data.

RESULTS

Of 2969 adults who completed the questionnaire (7.4% weighted response), 1772 (53.9% weighted) completed COVID-19 testing. Overall, 11.5% of adults had evidence of COVID-19 infection, with a higher prevalence among Hispanic and non-Hispanic Black persons, essential workers, those in low-income neighborhoods, and those with lower education attainment compared to their counterparts. We observed differences in attitudes and behaviors by race and ethnicity, with non-Hispanic White persons being less likely to believe in the importance of mask wearing, and racial and ethnic minorities more likely to attend social gatherings.

CONCLUSION

Over 10% of an urban population was infected with COVID-19 early during the pandemic. Differences in attitudes and behaviors likely contribute to sociodemographic disparities in COVID-19 prevalence.

Diabetes management behaviors associated with depression in the U.S

BACKGROUND

There is a lack of nationally representative evidence from the U.S. investigating the relationships between depression and diabetes management behaviors. Our study aimed to assess the associations between diabetes management behaviors and depression status, and to compare U.S. population-level percentages of diabetes management behaviors among patients with and without depression.

METHODS

A cross-sectional study was conducted using population-based survey data to assess patient-reported variables retrospectively. We used the Behavioral Risk Factor Surveillance System (BRFSS) data and included states in the U.S. that continuously adopted the diabetes optional modules in 2013, 2015, 2017, and 2019. We included U.S. adults (≥ 18 years old) with self-reported diabetes in our analysis. Main outcomes were diabetes management behaviors (i.e., self-check for blood glucose and feet sores/irritation, regular diabetes clinical visit, HbA1c check, professional feet check, and dilated eye examination) and lifestyle behaviors (i.e., exercise, smoking, and alcohol consumption).

RESULTS

Among the 74,011 respondents with diabetes, patients with depression had a higher likelihood of performing routine HbA1c checks (adjusted odds ratio (AOR) = 1.12; 95% CI 1.01-1.23) but had a lower likelihood to perform regular self-check for blood glucose (AOR = 0.91; 95% CI 0.84-0.99), receive professional feet checks (AOR = 0.87; 95% CI 0.79-0.95), and receive a dilated eye examination (AOR = 0.89; 95% CI 0.82-0.98). For lifestyle behaviors, patients with depression were more likely to smoke (No smoking (AOR) = 0.65; 95% CI = 0.59-0.72) and less likely to engage in sufficient exercise time (AOR = 0.69; 95% CI 0.63-0.75). There were no significant associations between depression and other behaviors, including self-check for feet sores/irritation (AOR = 0.99; 95% CI 0.92-1.08), regular diabetes clinical visit (AOR = 1.03, 95% CI 0.94-1.13), and alcohol consumption (AOR = 1.01, 95% CI 0.92-1.10).

CONCLUSIONS

The association between depression status and diabetes management behaviors varied. People with depression were positively associated with HbA1c checks. However, less uptake of other behaviors may indicate the needs for improvement in diabetes management.

No associations between C-reactive protein and spinal pain trajectories in children and adolescents (CHAMPS study-DK)

Preliminary evidence points to a link between C-reactive protein (CRP) and spinal pain in adults. However, there is a paucity of research in younger populations. Therefore, we aimed to determine associations between CRP and spinal pain in childhood and adolescence. We identified trajectories of spinal pain from childhood to adolescence and investigated the associations between CRP and trajectory subgroups. Six- to 11-year-old children from 13 primary schools, were followed from October 2008 and until 2014. High-sensitivity CRP collected at baseline (2008) was measured using serum samples. The outcome was the number of weeks with non-traumatic spinal pain between November 2008 and June 2014. We constructed a trajectory model to identify different spinal pain trajectory subgroups. The associations between CRP and spinal pain trajectory subgroups were modelled using mixed-effects multinominal logistic regression. Data from 1556 participants (52% female), with a mean age of 8.4 years at baseline, identified five spinal pain trajectory subgroups: "no pain" (55.3%), "rare" (23.7%), "rare, increasing" (13.6%), "moderate, increasing" (6.1%), and "early onset, decreasing" (1.3%). There were no differences in baseline high-sensitivity CRP levels between spinal pain trajectory subgroups. Thus, the heterogeneous courses of spinal pain experienced were not defined by differences in CRP at baseline.