A Review of Hot Deck Imputation for Survey Non-response

Social inequalities in patient-reported outcomes among older multimorbid patients--results of the MultiCare cohort study

Introduction

In this article three research questions are addressed: (1) Is there an association between socioeconomic status (SES) and patient-reported outcomes in a cohort of multimorbid patients? (2) Does the association vary according to SES indicator used (income, education, occupational position)? (3) Can the association between SES and patient-reported outcomes (self-rated health, health-related quality of life and functional status) be (partly) explained by burden of disease?

Methods

Analyses are based on the MultiCare Cohort Study, a German multicentre, prospective, observational cohort study of multimorbid patients from general practice. We analysed baseline data and data from the first follow-up after 15 months (N = 2,729). To assess burden of disease we used the patients’ morbidity data from standardized general practitioner (GP) interviews based on a list of 46 groups of chronic conditions including the GP’s severity rating of each chronic condition ranging from marginal to very severe.

Results

In the cross-sectional analyses SES was significantly associated with the patient-reported outcomes at baseline. Associations with income were more consistent and stronger than with education and occupational position. Associations were partly explained (17% to 44%) by burden of disease. In the longitudinal analyses only income (but not education and occupational position) was significantly related to the patient-reported outcomes at follow-up. Associations between income and the outcomes were reduced by 18% to 27% after adjustment for burden of disease.

Conclusions

Results indicate social inequalities in self-rated health, functional status and health related quality of life among older multimorbid patients. As associations with education and occupational position were inconsistent, these inequalities were mainly due to income. Inequalities were partly explained by burden of disease. However, even among patients with a similar disease burden, those with a low income were worse off in terms of the three patient-reported outcomes under study.

Electronic supplementary material

The online version of this article (doi:10.1186/s12939-015-0142-6) contains supplementary material, which is available to authorized users.

Cognitive ability at kindergarten entry and socioeconomic status

OBJECTIVE

To examine how gradients in socioeconomic status (SES) impact US children's reading and math ability at kindergarten entry and determine the contributions of family background, health, home learning, parenting, and early education factors to those gradients.

METHODS

Analysis of 6600 children with cognitive assessments at kindergarten entry from the US Early Childhood Longitudinal Birth Cohort Study. A composite SES measure based on parent's occupation, education, and income was divided into quintiles. Wald F tests assessed bivariate associations between SES and child's cognitive ability and candidate explanatory variables. A decomposition methodology examined mediators of early cognitive gradients.

RESULTS

Average reading percentile rankings increased from 34 to 67 across SES quintiles and math from 33 to 70. Children in lower SES quintiles had younger mothers, less frequent parent reading, less home computer use (27%-84%), and fewer books at home (26-114). Parent's supportive interactions, expectations for their child to earn a college degree (57%-96%), and child's preschool attendance (64%-89%) increased across quintiles. Candidate explanatory factors explained just over half the gradients, with family background factors explaining 8% to 13%, health factors 4% to 6%, home learning environment 18%, parenting style/beliefs 14% to 15%, and early education 6% to 7% of the gaps between the lowest versus highest quintiles in reading and math.

CONCLUSIONS

Steep social gradients in cognitive outcomes at kindergarten are due to many factors. Findings suggest policies targeting levels of socioeconomic inequality and a range of early childhood interventions are needed to address these disparities.

Fibromyalgia patients have reduced hippocampal volume compared with healthy controls

Objective

Fibromyalgia patients frequently report cognitive abnormalities. As the hippocampus plays an important role in learning and memory, we determined whether individuals with fibromyalgia had smaller hippocampal volume compared with healthy control participants.

Methods

T₁-weighted structural magnetic resonance imaging (MRI) scans were acquired from 40 female participants with fibromyalgia and 22 female healthy controls. The volume of the hippocampus was estimated using the software FreeSurfer. An analysis of covariance model controlling for potentially confounding factors of age, whole brain size, MRI signal quality, and Beck Depression Inventory scores were used to determine significant group differences.

Results

Fibromyalgia participants had significantly smaller hippocampi in both left (F[1,56]=4.55, P=0.037, η²_p=0.08) and right hemispheres (F[1,56]=5.89, P=0.019, η²_p=0.10). No significant effect of depression was observed in either left or right hemisphere hippocampal volume (P=0.813 and P=0.811, respectively).

Discussion

Potential mechanisms for reduced hippocampal volume in fibromyalgia include abnormal glutamate excitatory neurotransmission and glucocorticoid dysfunction; these factors can lead to neuronal atrophy, through excitotoxicity, and disrupt neurogenesis in the hippocampus. Hippocampal atrophy may play a role in memory and cognitive complaints among fibromyalgia patients.

A developmental analysis of the factorial validity of the parent-report version of the Adult Responses to Children's Symptoms in children versus adolescents with chronic pain or pain-related chronic illness

The widely used Adult Responses to Children's Symptoms measures parental responses to child symptom complaints among youth aged 7 to 18 years with recurrent/chronic pain. Given developmental differences between children and adolescents and the impact of developmental stage on parenting, the factorial validity of the parent-report version of the Adult Responses to Children's Symptoms with a pain-specific stem was examined separately in 743 parents of 281 children (7-11 years) and 462 adolescents (12-18 years) with chronic pain or pain-related chronic illness. Factor structures of the Adult Responses to Children's Symptoms beyond the original 3-factor model were also examined. Exploratory factor analysis with oblique rotation was conducted on a randomly chosen half of the sample of children and adolescents as well as the 2 groups combined to assess underlying factor structure. Confirmatory factor analysis was conducted on the other randomly chosen half of the sample to cross-validate factor structure revealed by exploratory factor analyses and compare it to other model variants. Poor loading and high cross-loading items were removed. A 4-factor model (Protect, Minimize, Monitor, and Distract) for children and the combined (child and adolescent) sample and a 5-factor model (Protect, Minimize, Monitor, Distract, and Solicitousness) for adolescents was superior to the 3-factor model proposed in previous literature. Future research should examine the validity of derived subscales and developmental differences in their relationships with parent and child functioning.

PERSPECTIVE

This article examined developmental differences in the structure of a widely used measure of caregiver responses to chronic pain or pain-related chronic illness in youth. Results suggest that revised structures that differ across developmental groups can be used with youth with a range of clinical pain-related conditions.

Smoking behavior: a cross-sectional study to assess the dimensionality of the brief Wisconsin inventory of smoking dependence motives and identify different typologies among young daily smokers

INTRODUCTION

The present study aims to investigate the dimensionality of the brief version of the Wisconsin Inventory of Smoking Dependence Motives (B-WISDM) and identify different smoking motivational profiles among young daily smokers (N = 375).

METHODS

We tested 3 measurement models of the B-WISDM using confirmatory factor analysis, whereas cluster analysis was used to identify the smokers' motivational profiles. Furthermore, we compared clusters toward dependence level and the number of cigarettes smoked per day using analysis of variance tests.

RESULTS

The results confirmed that the B-WISDM measures 11 first-order intercorrelated factors. The second-order model, originally proposed for the longer version of the questionnaire, showed adequate fit indices but fitted the data significantly worse than the first-order model. Five motivational clusters were identified and differed in terms of tobacco addiction and the number of cigarettes smoked per day. Although each cluster had specific features, 2 main smoker groups were distinguished: Group A (composed of 3 clusters), which was mainly characterized by high levels of secondary dependence motives, and Group B (composed of 2 clusters), in which the primary and secondary dependence motives reached similar levels. In general, the clusters of Group B were more addicted to cigarettes than Group A clusters.

CONCLUSIONS

Using the B-WISDM to identify different smoking motivational profiles has important practical implications because they might help characterize addiction, which represents the first step to help an individual quit smoking.

The significance of pre-arrest factors in out-of-hospital cardiac arrests witnessed by emergency medical services: a report from the Victorian Ambulance Cardiac Arrest Registry

BACKGROUND

The significance of pre-arrest factors in out-of-hospital cardiac arrests (OHCA) witnessed by emergency medical services (EMS) is not well established. The purpose of this study was to assess the association between prodromal symptoms and pre-arrest clinical observations on the arresting rhythm and survival in EMS witnessed OHCA.

METHODS

Between 1st January 2003 and 31st December 2011, 1056 adult EMS witnessed arrests of a presumed cardiac aetiology were identified from the Victorian Ambulance Cardiac Arrest Registry. Pre-arrest prodromal features and clinical characteristics were extracted from the patient care record. Backward elimination logistic regression was used to identify pre-arrest factors associated with an initial shockable rhythm and survival to hospital discharge.

RESULTS

The median age was 73.0 years, 690 (65.3%) were male, and the rhythm of arrest was shockable in 465 (44.0%) cases. The most commonly reported prodromal symptoms prior to arrest were chest pain (48.8%), dyspnoea (41.8%) and altered consciousness (37.8%). An unrecordable systolic blood pressure was observed in 34.4%, a respiratory rate <13 or >24min(-1) was present in 43.1%, and 45.5% had a Glasgow coma score <15. In the multivariable analysis, the following pre-arrest factors were significantly associated with survival: age, public location, aged care facility, chest pain, arm or shoulder pain, dyspnoea, dizziness, vomiting, ventricular tachycardia, pulse rate, systolic blood pressure, respiratory rate, Glasgow coma score, aspirin and inotrope administration.

CONCLUSION

Pre-arrest factors are strongly associated with the arresting rhythm and survival following EMS witnessed OHCA. Potential opportunities to improve outcomes exist by way of early recognition and management of patients at risk of OHCA.

Determining Sufficient Number of Imputations Using Variance of Imputation Variances: Data from 2012 NAMCS Physician Workflow Mail Survey

How many imputations are sufficient in multiple imputations? The answer given by different researchers varies from as few as 2 - 3 to as many as hundreds. Perhaps no single number of imputations would fit all situations. In this study, η, the minimally sufficient number of imputations, was determined based on the relationship between m, the number of imputations, and ω, the standard error of imputation variances using the 2012 National Ambulatory Medical Care Survey (NAMCS) Physician Workflow mail survey. Five variables of various value ranges, variances, and missing data percentages were tested. For all variables tested, ω decreased as m increased. The m value above which the cost of further increase in m would outweigh the benefit of reducing ω was recognized as the η. This method has a potential to be used by anyone to determine η that fits his or her own data situation.

Missing value imputation in high-dimensional phenomic data: imputable or not, and how?

BACKGROUND

In modern biomedical research of complex diseases, a large number of demographic and clinical variables, herein called phenomic data, are often collected and missing values (MVs) are inevitable in the data collection process. Since many downstream statistical and bioinformatics methods require complete data matrix, imputation is a common and practical solution. In high-throughput experiments such as microarray experiments, continuous intensities are measured and many mature missing value imputation methods have been developed and widely applied. Numerous methods for missing data imputation of microarray data have been developed. Large phenomic data, however, contain continuous, nominal, binary and ordinal data types, which void application of most methods. Though several methods have been developed in the past few years, not a single complete guideline is proposed with respect to phenomic missing data imputation.

RESULTS

In this paper, we investigated existing imputation methods for phenomic data, proposed a self-training selection (STS) scheme to select the best imputation method and provide a practical guideline for general applications. We introduced a novel concept of "imputability measure" (IM) to identify missing values that are fundamentally inadequate to impute. In addition, we also developed four variations of K-nearest-neighbor (KNN) methods and compared with two existing methods, multivariate imputation by chained equations (MICE) and missForest. The four variations are imputation by variables (KNN-V), by subjects (KNN-S), their weighted hybrid (KNN-H) and an adaptively weighted hybrid (KNN-A). We performed simulations and applied different imputation methods and the STS scheme to three lung disease phenomic datasets to evaluate the methods. An R package "phenomeImpute" is made publicly available.

CONCLUSIONS

Simulations and applications to real datasets showed that MICE often did not perform well; KNN-A, KNN-H and random forest were among the top performers although no method universally performed the best. Imputation of missing values with low imputability measures increased imputation errors greatly and could potentially deteriorate downstream analyses. The STS scheme was accurate in selecting the optimal method by evaluating methods in a second layer of missingness simulation. All source files for the simulation and the real data analyses are available on the author's publication website.

Quality of life and functional outcomes 12 months after out-of-hospital cardiac arrest

BACKGROUND

Out-of-hospital cardiac arrest (OHCA) is a significant global health problem. There has been considerable investment in improving the emergency medical response to OHCA, with associated improvements in survival. However, concern remains that survivors have a poor quality of life. This study describes the quality of life of OHCA survivors at 1-year postarrest in Victoria, Australia.

METHODS AND RESULTS

Adult OHCA patients who arrested between 2010 and 2012 were identified from the Victorian Ambulance Cardiac Arrest Registry. Paramedics attended 15 113 OHCA patients of which 46.3% received an attempted resuscitation. Nine hundred and twenty-seven (13.2%) survived to hospital discharge of which 76 (8.2%) died within 12 months. Interviews were conducted with 697 (80.7%) patients or proxies, who were followed-up via telephone interview, including the Glasgow Outcome Scale-Extended, the 12-item short form health survey, and the EuroQol. The majority (55.6%) of respondents had a good recovery via the Glasgow Outcome Scale-Extended≥7 (41.1% if patients who died postdischarge were included and nonrespondents were assumed to have poor recovery). The mean EuroQol index score for respondents was 0.82 (standard deviation, 0.19), which compared favorably with an adjusted population norm of 0.81 (standard deviation, 0.34). The mean 12-item short form Mental Component Summary score for patients was 53.0 (standard deviation, 10.2), whereas the mean Physical Component Summary score was 46.1 (standard deviation, 11.2).

CONCLUSIONS

This is the largest published study assessing the quality of life of OHCA survivors. It provides good evidence that many survivors have an acceptable quality of life 12 months postarrest, particularly in comparison with population norms.

Subjective Well-being of Older African Americans with DSM IV Psychiatric Disorders

This study examined demographic and mental health correlates of subjective well-being (i.e., life satisfaction, happiness) using a national sample of older African Americans with psychiatric disorders. We used a subsample of 185 African Americans, 55 and older with at least one of thirteen lifetime psychiatric disorders from The National Survey of American Life: Coping with Stress in the 21st Century (NSAL). The findings indicated that among this population of older adults who had a lifetime psychiatric disorder, having a lifetime suicidal ideation was associated with life satisfaction but not happiness. Further, having a 12-month anxiety disorder or a lifetime suicidal ideation was not associated with happiness. Having a 12-month mood disorder, however, was negatively associated with an individual's level of happiness, as well as their life satisfaction. Additionally, there were two significant interactions. Among men, employment was positively associated with life satisfaction, and marriage was associated with higher levels of happiness among men but not women. The overall pattern of findings reflects both similarities and departures from prior research confirming that well-being evaluations are associated with multiple factors.

Employment during pregnancy and obstetric intervention without medical reason: labor induction and cesarean delivery

BACKGROUND

Rising rates of labor induction and cesarean delivery, especially when used without a medical reason, have generated concern among clinicians, women, and policymakers. Whether employment status affects pregnant women's childbirth-related care is not known. We estimated the relationship between prenatal employment and obstetric procedures, distinguishing whether women reported that the induction or cesarean was performed for medical reasons.

METHODS

Using data from a nationally representative sample of women who gave birth in U.S. hospitals (n = 1,573), we used propensity score matching to reduce potential bias from nonrandom selection into employment. Outcomes were cesarean delivery and labor induction, with and without a self-reported medical reason. Exposure was prenatal employment status (full-time employment, not employed). We conducted separate analyses for unmatched and matched cohorts using multivariable regression models.

FINDINGS

There were no differences in labor induction based on employment status. In unmatched analyses, employed women had higher odds of cesarean delivery overall (adjusted odds ratio [AOR], 1.45; p = .046) and cesarean delivery without medical reason (AOR, 1.94; p = .024). Adding an interaction term between employment and college education revealed no effects on cesarean delivery without medical reason. There were no differences in cesarean delivery by employment status in the propensity score-matched analysis.

CONCLUSIONS

Full-time prenatal employment is associated with higher odds of cesarean delivery, but this association was not explained by socioeconomic status and no longer existed after accounting for sociodemographic differences by matching women employed full time with similar women not employed during pregnancy.

Family structure transitions and early childhood development in Taiwan

Taiwan has over the past three decades been experiencing demographic changes that may pose important concerns for children’s quality of life. This study examines the relationships and potential pathways between family structure transitions and early childhood development. Our analysis is based on 19,499 children from the 2005 birth cohort who completed three follow-up surveys (at 6, 18, and 36 months of age). Using multiple regressions, we found that children who consistently lived with single parents or who experienced their parents’ divorce/separation did poorer in both cognitive and socioemotional development than those in stable married families. Significance testing for mediation by bootstrapping revealed that the disparities could be explained by the family’s exposure to income poverty, psychosocial functioning and/or parenting quality. It is important to look more deeply into sociocultural factors in future investigations.

Implementation of chronic illness care in German primary care practices--how do multimorbid older patients view routine care? A cross-sectional study using multilevel hierarchical modeling

Background

In primary care, patients with multiple chronic conditions are the rule rather than the exception. The Chronic Care Model (CCM) is an evidence-based framework for improving chronic illness care, but little is known about the extent to which it has been implemented in routine primary care. The aim of this study was to describe how multimorbid older patients assess the routine chronic care they receive in primary care practices in Germany, and to explore the extent to which factors at both the practice and patient level determine their views.

Methods

This cross-sectional study used baseline data from an observational cohort study involving 158 general practitioners (GP) and 3189 multimorbid patients. Standardized questionnaires were employed to collect data, and the Patient Assessment of Chronic Illness Care (PACIC) questionnaire used to assess the quality of care received. Multilevel hierarchical modeling was used to identify any existing association between the dependent variable, PACIC, and independent variables at the patient level (socio-economic factors, weighted count of chronic conditions, instrumental activities of daily living, health-related quality of life, graded chronic pain, no. of contacts with GP, existence of a disease management program (DMP) disease, self-efficacy, and social support) and the practice level (age and sex of GP, years in current practice, size and type of practice).

Results

The overall mean PACIC score was 2.4 (SD 0.8), with the mean subscale scores ranging from 2.0 (SD 1.0, subscale goal setting/tailoring) to 3.5 (SD 0.7, delivery system design). At the patient level, higher PACIC scores were associated with a DMP disease, more frequent GP contacts, higher social support, and higher autonomy of past occupation. At the practice level, solo practices were associated with higher PACIC values than other types of practice.

Conclusions

This study shows that from the perspective of multimorbid patients receiving care in German primary care practices, the implementation of structured care and counseling could be improved, particularly by helping patients set specific goals, coordinating care, and arranging follow-up contacts. Studies evaluating chronic care should take into consideration that a patient’s assessment is associated not only with practice-level factors, but also with individual, patient-level factors.

Trial registration

Current Controlled Trials ISRCTN89818205.

Signs and symptoms of first-onset TMD and sociodemographic predictors of its development: the OPPERA prospective cohort study

Although cross-sectional studies of temporomandibular disorder (TMD) often report elevated prevalence in young women, they do not address the risk of its development. Here we evaluate sociodemographic predictors of TMD incidence in a community-based prospective cohort study of U.S. adults. Symptoms and pain-related disability in TMD cases are also described. People aged 18 to 44 years with no history of TMD were enrolled at 4 study sites when they completed questionnaires about sociodemographic characteristics. During the median 2.8-year follow-up period, 2,737 participants completed quarterly screening questionnaires. Those reporting symptoms were examined clinically and 260 had first-onset TMD. Additional questionnaires asked about severity and impact of their symptoms. Univariate and multivariable Cox regression models quantified associations between sociodemographic characteristics and TMD incidence. First-onset TMD developed in 3.9% of participants per annum, typically producing mild to moderate levels of pain and disability in cases. TMD incidence was positively associated with age, whereas females had only slightly greater incidence than males. Compared to whites, Asians had lower TMD incidence whereas African Americans had greater incidence, although the latter was attenuated somewhat after adjusting for satisfaction with socioeconomic circumstances.

PERSPECTIVE

In this study of 18- to 44-year-olds, TMD developed at a higher rate than reported previously for similar age groups. TMD incidence was positively associated with age but weakly associated with gender, thereby differing from demographic patterns of prevalence found in some cross-sectional studies. Experiences related to aging merit investigation as etiologic influences on development of TMD.

Study protocol, sample characteristics, and loss to follow-up: the OPPERA prospective cohort study

When studying incidence of pain conditions such as temporomandibular disorder (TMD), repeated monitoring is needed in prospective cohort studies. However, monitoring methods usually have limitations and, over a period of years, some loss to follow-up is inevitable. The OPPERA prospective cohort study of first-onset TMD screened for symptoms using quarterly questionnaires and examined symptomatic participants to definitively ascertain TMD incidence. During the median 2.8-year observation period, 16% of the 3,263 enrollees completed no follow-up questionnaires, others provided incomplete follow-up, and examinations were not conducted for one third of symptomatic episodes. Although screening methods and examinations were found to have excellent reliability and validity, they were not perfect. Loss to follow-up varied according to some putative TMD risk factors, although multiple imputation to correct the problem suggested that bias was minimal. A second method of multiple imputation that evaluated bias associated with omitted and dubious examinations revealed a slight underestimate of incidence and some small biases in hazard ratios used to quantify effects of risk factors. Although "bottom line" statistical conclusions were not affected, multiply-imputed estimates should be considered when evaluating the large number of risk factors under investigation in the OPPERA study.

PERSPECTIVE

These findings support the validity of the OPPERA prospective cohort study for the purpose of investigating the etiology of first-onset TMD, providing the foundation for other papers investigating risk factors hypothesized in the OPPERA project.

Additive effects of nutritional supplementation, together with bisphosphonates, on bone mineral density after hip fracture: a 12-month randomized controlled study

Background

After a hip fracture, a catabolic state develops, with increased bone loss during the first year. The aim of this study was to evaluate the effects of postoperative treatment with calcium, vitamin D, and bisphosphonates (alone or together) with nutritional supplementation on total hip and total body bone mineral density (BMD).

Methods

Seventy-nine patients (56 women), with a mean age of 79 years (range, 61–96 years) and with a recent hip fracture, who were ambulatory before fracture and without severe cognitive impairment, were included. Patients were randomized to treatment with bisphosphonates (risedronate 35 mg weekly) for 12 months (B; n=28), treatment with bisphosphonates along with nutritional supplementation (40 g protein, 600 kcal daily) for the first 6 months (BN; n=26), or to controls (C; n=25). All participants received calcium (1,000 mg) and vitamin D3 (800 IU) daily. Total hip and total body BMD were assessed with dual-energy X-ray absorptiometry at baseline, 6, and 12 months. Marker of bone resorption C-terminal telopeptide of collagen I and 25-hydroxy vitamin D were analyzed in serum.

Results

Analysis of complete cases (70/79 at 6 months and 67/79 at 12 months) showed an increase in total hip BMD of 0.7% in the BN group, whereas the B and C groups lost 1.1% and 2.4% of BMD, respectively, between baseline and 6 months (P=0.071, between groups). There was no change in total body BMD between baseline and 12 months in the BN group, whereas the B group and C group both lost BMD, with C losing more than B (P=0.009). Intention-to-treat analysis was in concordance with the complete cases analyses.

Conclusion

Protein-and energy-rich supplementation in addition to calcium, vitamin D, and bisphosphonate therapy had additive effects on total body BMD and total hip BMD among elderly hip fracture patients.

Accuracy of parental reports of children's HPV vaccine status: implications for estimates of disparities, 2009-2010

OBJECTIVES

Since the introduction of the human papillomavirus (HPV) vaccine in 2006, there have been considerable efforts at the national and state levels to monitor uptake and better understand the individual and system-level factors that predict who gets vaccinated. A common method of measuring the vaccination status of adolescents is through parental recall. We examined how the accuracy of parents' reports of their daughters' HPV vaccination status varied by social characteristics.

METHODS

Data were taken from the 2009-2010 National Immunization Survey (NIS)-Teen, which includes a household interview and a provider-completed immunization history. We evaluated concordance between parents' and providers' reports of teens' HPV vaccine initiation (≥1 dose) and completion (≥3 doses). We assessed bivariate associations of sociodemographic characteristics with having a concordant, false-positive (overreporting) or false-negative (underreporting) report, and used multinomial logistic regression to estimate the independent impact of each characteristic.

RESULTS

In bivariate analyses, concordance of parent-reported HPV vaccine initiation was associated with each of the sociodemographic characteristics investigated. In regression models, self-reported nonwhite race, lower household income, and lower education level of the teen's mother were associated with a higher likelihood of having a false-negative parental report than a concordant report.

CONCLUSION

Our results indicate that, while estimates of overall coverage based on parental report may be unbiased, the differences in the accuracy of parental report could result in misleading estimates of disparities in HPV vaccine coverage.

Away-from-home eating: nutritional status and dietary intake among Brazilian adults

Objective

To evaluate the association between eating away from home and BMI and to examine whether dietary intake differs based on the consumption of away-from-home food (AFHF).

Design

Data were obtained from the first Brazilian National Dietary Survey, using food records. The association between the percentage of energy provided by foods consumed away from home and BMI status was tested using logistic regression models. The mean percentages of energy provided by protein, fat, saturated fat and free sugars were calculated based on the consumption of foods away from home among AFHF consumers.

Setting

Urban areas of Brazil.

Subjects

Adults (n13 736) between 25 and 65 years old.

Results

AFHF was not associated with BMI status. Individuals who consumed AFHF had higher intakes of free sugars away from home than at home and had higher intakes of energy-dense foods than AFHF non-consumers.

Conclusion

Although AFHF consumption was not related to overweight or obesity status, individuals who consumed foods away from home had higher intakes of energy-dense foods. Public health policies should be implemented to help people make healthier food choices away from home.

Increasing specificity of correlate research: exploring correlates of children's lunchtime and after-school physical activity

Background

The lunchtime and after-school contexts are critical windows in a school day for children to be physically active. While numerous studies have investigated correlates of children’s habitual physical activity, few have explored correlates of physical activity occurring at lunchtime and after-school from a social-ecological perspective. Exploring correlates that influence physical activity occurring in specific contexts can potentially improve the prediction and understanding of physical activity. Using a context-specific approach, this study investigated correlates of children’s lunchtime and after-school physical activity.

Methods

Cross-sectional data were collected from 423 South Australian children aged 10.0–13.9 years (200 boys; 223 girls) attending 10 different schools. Lunchtime and after-school physical activity was assessed using accelerometers. Correlates were assessed using purposely developed context-specific questionnaires. Correlated Component Regression analysis was conducted to derive correlates of context-specific physical activity and determine the variance explained by prediction equations.

Results

The model of boys’ lunchtime physical activity contained 6 correlates and explained 25% of the variance. For girls, the model explained 17% variance from 9 correlates. Enjoyment of walking during lunchtime was the strongest correlate for both boys and girls. Boys’ and girls’ after-school physical activity models explained 20% variance from 14 correlates and 7% variance from the single item correlate, “I do an organised sport or activity after-school because it gets you fit”, respectively.

Conclusions

Increasing specificity of correlate research has enabled the identification of unique features of, and a more in-depth interpretation of, lunchtime and after-school physical activity behaviour and is a potential strategy for advancing the physical activity correlate research field. The findings of this study could be used to inform and tailor gender-specific public health messages and interventions for promoting lunchtime and after-school physical activity in children.

Analyzing factors associated with women's attitudes and behaviors toward screening mammography using design-based logistic regression

We examined the factors associated with screening mammography adherence behaviors and influencing factors on women's attitudes toward mammography in non-adherent women. Design-based logistic regression models were developed to characterize the influencing factors, including socio-demographic, health related, behavioral characteristics, and knowledge of breast cancer/mammography, on women's compliance with and attitudes toward mammography using the 2003 Health Information National Trends Survey data. Findings indicate significant associations among adherence to mammography and marital status, income, health coverage, being advised by a doctor to have a mammogram, having had Pap smear before, perception of chance of getting breast cancer, and knowledge of mammography (frequency of doing mammogram) in both women younger than 65 and women aged 65 and older. However, number of visits to a healthcare provider per year and lifetime number of smoked cigarettes are only significant for women younger than 65. Factors significantly associated with attitudes toward mammography in non-adherent women are age, being advised by a doctor to have a mammogram, and seeking cancer information. To enhance adherence to mammography programs, physicians need to continue to advise their patients to obtain mammograms. In addition, increasing women's knowledge about the frequency and starting age for screening mammography may improve women's adherence. Financially related factors such as income and insurance are also shown to be significant factors. Hence, healthcare policies aimed at providing breast cancer screening services to underserved women will likely enhance mammography participation.

Association between thyroid function and selected organochlorine pesticides: data from NHANES 2001-2002

Exposure to organochlorine pesticides (OCP) has been shown to be associated with adverse thyroid function. The impact of exposure to selected OCPs on total serum thyroxine (TT4) and thyroid stimulating hormone (TSH) was evaluated by analyzing data from the 2001-2002 National Health and Nutrition Examination Survey. Oxychlordane, p,p'-DDE, trans-nonachlor, and heptachlor epoxide were selected for analysis. Regression models with logs of TSH and TT4 as dependent variables and OCP exposure, race/ethnicity, iodine and smoking status, total lipids and others as independent variables were fitted. TSH levels increased (p < 0.05) with increase in trans-nonachlor exposure for 20-39 year old iodine deficient males. TSH levels were higher when oxychlordane exposure was low than when the exposure was medium or high for 20-39 year old iodine deficient females (p < 0.05). For iodine deficient females, TT4 levels were lower when p,p'-DDE exposure was low than when it was medium (p < 0.05). For non-Hispanic blacks (NHB), TT4 levels decreased with increase in exposure to heptachlor epoxide (p < 0.05). For iodine replete males, TSH levels increased with increase in trans-nonachlor exposure (p < 0.05). For iodine replete females, (i) Mexican Americans (MA) had higher TSH levels when the exposure to oxychlordane was medium than when the exposure was low; (ii) for 60+years old, there was a positive association between TSH and heptachlor epoxide levels; and (iii) TT4 levels had an inverse association with trans-nonachlor and oxychlordane. In general though not always, (i) TSH and TT4 levels were lowest for the 20-39 years old and highest for the 60+years old (p < = 0.05), (ii) TSH and TT4 levels for iodine deficient males and females were lowest for NHB, highest for MA, and in-between for non-Hispanic white, and (iii) non-smokers had higher TSH and TT4 levels than smokers and in general, statistically significantly so.

Media's role in broadcasting acute stress following the Boston Marathon bombings

We compared the impact of media vs. direct exposure on acute stress response to collective trauma. We conducted an Internet-based survey following the Boston Marathon bombings between April 29 and May 13, 2013, with representative samples of residents from Boston (n = 846), New York City (n = 941), and the remainder of the United States (n = 2,888). Acute stress symptom scores were comparable in Boston and New York [regression coefficient (b) = 0.43; SE = 1.42; 95% confidence interval (CI), -2.36, 3.23], but lower nationwide when compared with Boston (b = -2.21; SE = 1.07; 95% CI, -4.31, -0.12). Adjusting for prebombing mental health (collected prospectively), demographics, and prior collective stress exposure, six or more daily hours of bombing-related media exposure in the week after the bombings was associated with higher acute stress than direct exposure to the bombings (continuous acute stress symptom total: media exposure b = 15.61 vs. direct exposure b = 5.69). Controlling for prospectively collected prebombing television-watching habits did not change the findings. In adjusted models, direct exposure to the 9/11 terrorist attacks and the Sandy Hook School shootings were both significantly associated with bombing-related acute stress; Superstorm Sandy exposure wasn't. Prior exposure to similar and/or violent events may render some individuals vulnerable to the negative effects of collective traumas. Repeatedly engaging with trauma-related media content for several hours daily shortly after collective trauma may prolong acute stress experiences and promote substantial stress-related symptomatology. Mass media may become a conduit that spreads negative consequences of community trauma beyond directly affected communities.

Nonlinear multiple imputation for continuous covariate within semiparametric Cox model: application to HIV data in Senegal

Multiple imputation is commonly used to impute missing covariate in Cox semiparametric regression setting. It is to fill each missing data with more plausible values, via a Gibbs sampling procedure, specifying an imputation model for each missing variable. This imputation method is implemented in several softwares that offer imputation models steered by the shape of the variable to be imputed, but all these imputation models make an assumption of linearity on covariates effect. However, this assumption is not often verified in practice as the covariates can have a nonlinear effect. Such a linear assumption can lead to a misleading conclusion because imputation model should be constructed to reflect the true distributional relationship between the missing values and the observed values. To estimate nonlinear effects of continuous time invariant covariates in imputation model, we propose a method based on B-splines function. To assess the performance of this method, we conducted a simulation study, where we compared the multiple imputation method using Bayesian splines imputation model with multiple imputation using Bayesian linear imputation model in survival analysis setting. We evaluated the proposed method on the motivated data set collected in HIV-infected patients enrolled in an observational cohort study in Senegal, which contains several incomplete variables. We found that our method performs well to estimate hazard ratio compared with the linear imputation methods, when data are missing completely at random, or missing at random.

Tensor-based methods for handling missing data in quality-of-life questionnaires

A common problem with self-report quality-of-life questionnaires is missing data. Despite enormous care and effort to prevent it, some level of missing data is common and unavoidable. Missing data can have a detrimental impact on the data analysis. In this paper, a novel approach to imputing missing data in quality-of-life questionnaires is proposed, based on matrix and tensor decompositions. In order to illustrate and assess those methods, two datasets are considered: The first dataset contains the responses of 100 patients to a systemic lupus erythematosus-specific quality-of-life questionnaire; the other contains the responses of 43 patients to a rhino-conjunctivitis quality-of-life questionnaire. The two datasets contain almost no missing data, and for testing purposes, data entries are removed at random to have missing completely at random data. Several proportions of missing values are considered, and for each, the imputation error is assessed through k-fold cross validation. We also evaluate different imputation methods for missing at random and missing not at randomdata. The numerical results demonstrate that the proposed tensor factorization-based methods outperform standard methods in terms of root mean square error with at least 4% improvement, while the bias and variance are similar.

The impact of prenatal employment on breastfeeding intentions and breastfeeding status at 1 week postpartum

BACKGROUND

Postpartum employment is associated with non-initiation and early cessation of breastfeeding, but less is known about the relationship between prenatal employment and breastfeeding intentions and behaviors.

OBJECTIVE

This study aimed to estimate the relationship between prenatal employment status, a strong predictor of postpartum return to work, and breastfeeding intentions and behaviors.

METHODS

Using data from the Listening to Mothers II national survey (N = 1573), we used propensity score matching methods to account for non-random selection into employment patterns and to measure the impact of prenatal employment status on breastfeeding intentions and behaviors. We also examined whether hospital practices consistent with the Baby-Friendly Hospital Initiative (BFHI), assessed based on maternal perception, were differentially associated with breastfeeding by employment status.

RESULTS

Women who were employed (vs unemployed) during pregnancy were older, were more educated, were less likely to have had a previous cesarean delivery, and had fewer children. After matching, these differences were eliminated. Although breastfeeding intention did not differ by employment, full-time employment (vs no employment) during pregnancy was associated with decreased odds of exclusive breastfeeding 1 week postpartum (adjusted odds ratio = 0.48; 95% confidence interval, 0.25-0.92; P = .028). Higher BFHI scores were associated with higher odds of breastfeeding at 1 week but did not differentially impact women by employment status.

CONCLUSION

Women employed full-time during pregnancy were less likely to fulfill their intention to exclusively breastfeed, compared to women who were not employed during pregnancy. Clinicians should be aware that employment circumstances may impact women's breastfeeding decisions; this may help guide discussions during clinical encounters.

The effects of cannabis use expectancies on self-initiated cannabis cessation

AIMS

To prospectively investigate the relation between cannabis use expectancies and cannabis use prior to and during a self-initiated cannabis cessation attempt.

DESIGN

Cohort design that followed participants for 4 weeks following a self-initiated cessation attempt.

SETTING

United States Department of Veterans Affairs medical center.

PARTICIPANTS

One hundred cannabis dependent military veterans.

MEASUREMENTS

The Marijuana Effects Expectancy Questionnaire at baseline; the timeline follow-back procedure at baseline and during the cessation attempt.

FINDINGS

Cannabis use at baseline was associated with positive (P = 0.01), but not negative (P = 0.25), expectancies. Cannabis lapse was associated with positive (P = 0.03) and negative expectancies (P = 0.01), and relapse was associated with positive (P = 0.04), but not negative (P = 0.21), expectancies. The trajectory of average cannabis use during the cessation period was associated with positive (P = 0.03), but not negative (P = 0.96), expectancies. Results were similar in effect and statistical significance when adjusting for demographic factors, motivation to quit cannabis, mental disorder diagnoses, and alcohol and tobacco use, and when analyzing complete data sets obtained through multiple imputation.

CONCLUSIONS

In the USA, cannabis use expectancies, especially those regarding the positive effects of cannabis use, appear to be strongly and consistently linked to cannabis use and quit failure.

Use of long-term antibiotic treatment in COPD patients in the UK: a retrospective cohort study

BACKGROUND

Exacerbations of chronic obstructive pulmonary disease (COPD) are a burden to patients and impose a major cost on health services. Long-term antibiotic therapy may prevent exacerbations, but at present it is not recommended by management guidelines.

AIMS

To identify the type and prevalence of long-term oral antibiotic treatments prescribed to patients with COPD and to assess the patient characteristics associated with long-term antibiotic use.

METHODS

A retrospective cohort using all eligible practices in The Health Improvement Network (THIN) UK primary care database between 2000 and 2009 was studied. We identified patients with COPD and then those who received a course of long-term antibiotics. Long-term courses were defined as >6 months in duration with <50% concomitant oral corticosteroid treatment.

RESULTS

We identified 92,576 patients with COPD, but only 567 patients (0.61%) who received 998 long-term antibiotic courses. Mean follow-up time was 3 years and 10 months. The median long-term antibiotic course length was 280 days (interquartile range 224, 394) and 58 patients (0.06%) were continuously prescribed antibiotics for >2 years. The most commonly used long-term antibiotics were oxytetracycline, doxycycline, and penicillin. Azithromycin, erythromycin, and clarithromycin were less frequently used. There was little evidence of the use of rotating courses of antibiotics. Men, people aged 50-79 years, non-smokers, and patients with poorer lung function were more likely to receive long-term antibiotic treatment.

CONCLUSIONS

Relatively few COPD patients are currently prescribed long-term antibiotics. Further clinical trials are required to determine the efficacy of this therapy. If beneficial, the use of such treatments should be incorporated into clinical guidelines.

From family to friends: does witnessing interparental violence affect young adults' relationships with friends?

PURPOSE

Childhood exposure to violence in one's family of origin has been closely linked to subsequent perpetration and victimization of intimate partner violence. There is, however, little research on the relationship between witnessing violence and subsequent peer violence. This study investigates the effects of witnessing interparental violence among Filipino young adults on their use and experience of psychological aggression with friends.

METHODS

The data source for this study was the Cebu Longitudinal Health and Nutrition Survey. Recent perpetration and victimization of friend psychological aggression among young adults ages 21-22 years was assessed through self-reports from the 2005 survey; witnessing interparental violence during childhood was assessed through self-reports from the 2002 survey. Multinomial logistic regression was used to examine the effects of witnessing interparental violence on subsequent use and experience of friend psychological aggression. Analyses were stratified by gender.

RESULTS

About 13% of females and 4% of males perpetrated psychological aggression toward close friends, and about 4% of females and males were victims. Fourteen percent of females and 3% of males experienced bidirectional psychological aggression. About 44% of females and 47% of males had, during childhood, witnessed their parents physically hurt one another. Witnessing maternal and reciprocal interparental violence during childhood significantly predicted bidirectional friend psychological aggression among males. Among females, witnessing interparental violence did not significantly predict involvement with friend psychological aggression.

CONCLUSIONS

Violence prevention programs should consider using family-centered interventions, and apply a gendered lens to their application. Further research on gender differences in friend aggression is recommended.

The role of pretreatment outcome expectancies and cognitive-behavioral skills in symptom improvement in an acute psychiatric setting

BACKGROUND

Prior research has found that pretreatment expectations of symptom improvement are positively correlated with depressive symptom change. The current investigation extends previous research by examining whether pretreatment outcome expectancies predict symptom change across several diagnostic categories within the context of an acute, naturalistic psychiatric setting.

METHODS

Analyses were conducted to examine whether pretreatment outcome expectancies (credibility/expectancy questionnaire [CEQ]) predicted symptom improvement within major depression (N=420), bipolar disorder (N=120) and psychosis (N=36). Bootstrap mediation analyses were conducted to examine whether acquisition of cognitive behavioral therapy (CBT) skills (cognitive behavior therapy skills questionnaire [CBTSQ]) may mediate expectancy-outcome relations.

RESULTS

Results indicated a differential pattern of associations across diagnoses. Patient CBT skills emerged as a significant mediator of expectancy-outcome relations, but only in the major depression group. Both behavioral and cognitive skills were significantly, and independently, associated with symptom improvement.

LIMITATIONS

Sample sizes were small in the bipolar manic subgroup and psychosis group. CBT skills and symptom measures were assessed at concurrent time points.

CONCLUSIONS

The present findings suggest that patient expectancies and CBT skills may have a differential impact on symptom change as a function of diagnostic category. The implication of these results and directions for future research are discussed.

Emergency department-based brief intervention to reduce risky driving and hazardous/harmful drinking in young adults: a randomized controlled trial

BACKGROUND

Risky driving and hazardous drinking are associated with significant human and economic costs. Brief interventions for more than one risky behavior have the potential to reduce health-compromising behaviors in populations with multiple risk-taking behaviors such as young adults. Emergency department (ED) visits provide a window of opportunity for interventions meant to reduce both risky driving and hazardous drinking.

METHODS

We determined the efficacy of a Screening, Brief Intervention, and Referral to Treatment (SBIRT) protocol addressing risky driving and hazardous drinking. We used a randomized controlled trial design with follow-ups through 12 months. ED patients aged 18 to 44 who screened positive for both behaviors (n = 476) were randomized to brief intervention (BIG), contact control (CCG), or no-contact control (NCG) groups. The BIG (n = 150) received a 20-minute assessment and two 20-minute interventions. The CCG (n = 162) received a 20-minute assessment at baseline and no intervention. The NCG (n = 164) were asked for contact information at baseline and had no assessment or intervention. Outcomes at 3, 6, 9, and 12 months were self-reported driving behaviors and alcohol consumption.

RESULTS

Outcomes were significantly lower in BIG compared with CCG through 6 or 9 months, but not at 12 months: Safety belt use at 3 months (adjusted odds ratio [AOR], 0.22; 95% confidence interval [CI], 0.08 to 0.65); 6 months (AOR, 0.13; 95% CI, 0.04 to 0.42); and 9 months (AOR, 0.18; 95% CI, 0.06 to 0.56); binge drinking at 3 months (adjusted rate ratio [ARR] 0.84; 95% CI, 0.74 to 0.97) and 6 months (ARR, 0.81; 95% CI, 0.67 to 0.97); and ≥5 standard drinks/d at 3 months (AOR, 0.43; 95% CI, 0.20 to 0.91) and 6 months (AOR, 0.41; 95% CI, 0.17 to 0.98). No substantial differences were observed between BIG and NCG at 12 months.

CONCLUSIONS

Our findings indicate that SBIRT reduced risky driving and hazardous drinking in young adults, but its effects did not persist after 9 months. Future research should explore methods for extending the intervention effect.

Relative impact of multimorbid chronic conditions on health-related quality of life--results from the MultiCare Cohort Study

Background

Multimorbidity has a negative impact on health-related quality of life (HRQL). Previous studies included only a limited number of conditions. In this study, we analyse the impact of a large number of conditions on HRQL in multimorbid patients without preselecting particular diseases. We also explore the effects of these conditions on the specific dimensions of HRQL.

Materials and Methods

This analysis is based on a multicenter, prospective cohort study of 3189 multimorbid primary care patients aged 65 to 85. The impact of 45 conditions on HRQL was analysed. The severity of the conditions was rated. The EQ-5D, consisting of 5 dimensions and a visual-analogue-scale (EQ VAS), was employed. Data were analysed using multiple ordinary least squares and multiple logistic regressions. Multimorbidity measured by a weighted count score was significantly associated with lower overall HRQL (EQ VAS), b = −1.02 (SE: 0.06). Parkinson’s disease had the most pronounced negative effect on overall HRQL (EQ VAS), b = −12.29 (SE: 2.18), followed by rheumatism, depression, and obesity. With regard to the individual EQ-5D dimensions, depression (OR = 1.39 to 3.3) and obesity (OR = 1.44 to 1.95) affected all five dimensions of the EQ-5D negatively except for the dimension anxiety/depression. Obesity had a positive effect on this dimension, OR = 0.78 (SE: 0.07). The dimensions “self-care”, OR = 4.52 (SE: 1.37) and “usual activities”, OR = 3.59 (SE: 1.0), were most strongly affected by Parkinson’s disease. As a limitation our sample may only represent patients with at most moderate disease severity.

Conclusions

The overall HRQL of multimorbid patients decreases with an increasing count and severity of conditions. Parkinson’s disease, depression and obesity have the strongest impact on HRQL. Further studies should address the impact of disease combinations which require very large sample sizes as well as advanced statistical methods.

Effects of multiple chronic conditions on health care costs: an analysis based on an advanced tree-based regression model

BACKGROUND

To analyze the impact of multimorbidity (MM) on health care costs taking into account data heterogeneity.

METHODS

Data come from a multicenter prospective cohort study of 1,050 randomly selected primary care patients aged 65 to 85 years suffering from MM in Germany. MM was defined as co-occurrence of ≥3 conditions from a list of 29 chronic diseases. A conditional inference tree (CTREE) algorithm was used to detect the underlying structure and most influential variables on costs of inpatient care, outpatient care, medications as well as formal and informal nursing care.

RESULTS

Irrespective of the number and combination of co-morbidities, a limited number of factors influential on costs were detected. Parkinson's disease (PD) and cardiac insufficiency (CI) were the most influential variables for total costs. Compared to patients not suffering from any of the two conditions, PD increases predicted mean total costs 3.5-fold to approximately € 11,000 per 6 months, and CI two-fold to approximately € 6,100. The high total costs of PD are largely due to costs of nursing care. Costs of inpatient care were significantly influenced by cerebral ischemia/chronic stroke, whereas medication costs were associated with COPD, insomnia, PD and Diabetes. Except for costs of nursing care, socio-demographic variables did not significantly influence costs.

CONCLUSIONS

Irrespective of any combination and number of co-occurring diseases, PD and CI appear to be most influential on total health care costs in elderly patients with MM, and only a limited number of factors significantly influenced cost.

TRIAL REGISTRATION

Current Controlled Trials ISRCTN89818205.

Incidence and treatment patterns in males presenting with lower urinary tract symptoms to the emergency department in the United States

PURPOSE

Due to varying clinical definitions of lower urinary tract symptoms, it has been difficult to determine comparable prevalence and incidence rates of lower urinary tract symptoms and their treatment modalities. We assessed the incidence of emergency department visits in men with lower urinary tract symptoms who presented to emergency departments in the United States and factors associated with an increased likelihood of hospitalization.

MATERIALS AND METHODS

Emergency department visits from 2006 to 2009 associated with a primary diagnosis of lower urinary tract symptoms using established criteria were abstracted from the Nationwide Emergency Department Sample. Age adjusted incidence rates of emergency department visits and charges were calculated. We performed multivariable analysis to examine patient and hospital characteristics of those hospitalized and those with benign prostatic hyperplasia related adverse events.

RESULTS

A weighted estimate of 1,178,423 emergency department visits for lower urinary tract symptoms was recorded with a national incidence of 197.6/100,000 males per year. A total of 112,288 visits (9.5%) resulted in hospitalization. Adverse events were identified in 734,269 patients (62.3%). The most common adverse events were catheterization in 44.6% of cases, infection in 17.4%, hematuria in 9.6%, bladder stones in 1.7%, hydronephrosis in 1.2% and acute renal failure in 0.1%. On multivariable analysis independent predictors of hospital admission included comorbidities, socioeconomic status, hospital characteristics and adverse events such as sepsis, acute renal failure and hydronephrosis. Independent predictors of adverse events included patient age, year of visit, socioeconomic status, hospital characteristics and concomitant neurological disease. In 2009 total emergency department charges for lower urinary tract symptoms were $494,981,922.

CONCLUSIONS

The number of men with lower urinary tract symptoms who visit the emergency department has remained stable, while emergency department charges have increased by 40%. The rate of adverse events increased during the study period. These findings might suggest over reliance on medical and conservative therapy in the contemporary era.

Factor structure of the children's depression inventory in a multisite sample of children and adolescents with chronic pain

This study examined the factor structure of the Children's Depression Inventory (CDI) among children and adolescents with chronic pain using exploratory and confirmatory factor analysis in a large, multisite sample of treatment-seeking youth. Participants included 1,043 children and adolescents (ages 8-18) with a range of chronic pain complaints who presented for initial evaluation at 1 of 3 tertiary care pediatric chronic pain clinics across the United States. They completed the CDI and reported on pain intensity and functional disability. Factor analysis was conducted using a 2-step (exploratory and confirmatory) approach. Results supported a 5-factor model for the CDI with good fit to the data. The distribution and item-total correlations of the somatic items (eg, pain complaints, fatigue) were explored in this sample. Results indicate that the CDI is a useful tool for assessing depressive symptoms in youth with chronic pain, but some caution is warranted in interpreting the clinical significance of scores in light of the overlap of specific symptoms common to both pain and depression.

PERSPECTIVE

The CDI can be considered a valid tool for assessing mood symptoms in children with chronic pain. Caution is encouraged when interpreting the clinical significance of scores due to symptom overlap between chronic pain and depression.

A Bayesian Procedure for File Linking to Analyze End-of-Life Medical Costs

End-of-life medical expenses are a significant proportion of all health care expenditures. These costs were studied using costs of services from Medicare claims and cause of death (CoD) from death certificates. In the absence of a unique identifier linking the two datasets, common variables identified unique matches for only 33% of deaths. The remaining cases formed cells with multiple cases (32% in cells with an equal number of cases from each file and 35% in cells with an unequal number). We sampled from the joint posterior distribution of model parameters and the permutations that link cases from the two files within each cell. The linking models included the regression of location of death on CoD and other parameters, and the regression of cost measures with a monotone missing data pattern on CoD and other demographic characteristics. Permutations were sampled by enumerating the exact distribution for small cells and by the Metropolis algorithm for large cells. Sparse matrix data structures enabled efficient calculations despite the large dataset (≈1.7 million cases). The procedure generates m datasets in which the matches between the two files are imputed. The m datasets can be analyzed independently and results combined using Rubin's multiple imputation rules. Our approach can be applied in other file linking applications.

Reevaluating the relationship between prenatal employment and birth outcomes: a policy-relevant application of propensity score matching

BACKGROUND

Prior research shows an association between prenatal employment characteristics and adverse birth outcomes, but suffers methodological challenges in disentangling women's employment choices from birth outcomes, and little U.S.-based prior research compares outcomes for employed women with those not employed. This study assessed the effect of prenatal employment status on birth outcomes.

METHODS

With data from the Listening to Mothers II survey, conducted among a nationally representative sample of women who delivered a singleton baby in a U.S. hospital in 2005 (n = 1,573), we used propensity score matching to reduce potential selection bias. Primary outcomes were low birth weight (<2,500 g) and preterm birth (gestational age <37 weeks). Exposure was prenatal employment status (full time, part time, not employed). We conducted separate outcomes analyses for each matched cohort using multivariable regression models.

FINDINGS

Comparing full-time employees with women who were not employed, full-time employment was not causally associated with preterm birth (adjusted odds ratio [AOR], 1.37; p = .47) or low birth weight (AOR, 0.73; p = .41). Results were similar comparing full- and part-time workers. Consistent with prior research, Black women, regardless of employment status, had increased odds of low birth weight compared with White women (AOR, 5.07; p = .002).

CONCLUSIONS

Prenatal employment does not independently contribute to preterm births or low birth weight after accounting for characteristics of women with different employment statuses. Efforts to improve birth outcomes should focus on the characteristics of pregnant women (employed or not) that render them vulnerable.

Effects of fluoridated drinking water on dental caries in Australian adults

Systematic reviews produce conflicting conclusions regarding dental caries-preventive effects of water fluoridation in adults. The authors investigated the relationship using data from the nationally representative 2004-2006 Australian National Survey of Adult Oral Health. Effects were compared between the pre-fluoridation cohort born before 1960 (n = 2,270) and the cohort born between 1960 and 1990 (n = 1,509), when widespread implementation of fluoridation increased population coverage from < 1% to 67%. Residential history questionnaires determined the percentage of each person's lifetime exposed to fluoridated water. Examiners recorded decayed, missing, and filled teeth (DMF-Teeth) and decayed and filled tooth surfaces (DF-Surfaces). Socio-demographic and preventive dental behaviors were included in multivariable least-squares regression models adjusted for potential confounding. In fully adjusted models, > 75% of lifetime exposure to fluoridation relative to < 25% of lifetime exposure was associated with 11% and 10% fewer DMF-Teeth in the pre-1960 (p < .0001) and 1960-1990 cohorts (p = .018), respectively. Corresponding reductions in DF-Surfaces were 30% (p < .001) and 21% (p < .001). Findings for intermediate fluoridation exposure suggested a dose-response relationship. Results were consistent in sensitivity analyses accounting for missing data. In this nationally representative sample of Australian adults, caries-preventive effects of water fluoridation were at least as great in adults born before widespread implementation of fluoridation as after widespread implementation of fluoridation.

Preclinical episodes of orofacial pain symptoms and their association with health care behaviors in the OPPERA prospective cohort study

The course of preclinical pain symptoms sheds light on the etiology and prognosis of chronic pain. We aimed to quantify rates of developing initial and recurrent symptoms of painful temporomandibular disorder (TMD) and to evaluate associations with health behaviors. In the OPPERA prospective cohort study, 2,719 individuals aged 18 to 44years with lifetime absence of TMD when enrolled completed 25,103 quarterly (3-monthly) questionnaires during a median 2.3-year follow-up period. Questionnaires documented TMD symptom episodes, headache, other body pain, health care attendance, and analgesic use, and. Kaplan-Meier methods for clustered data estimated symptom-free survival time. Multivariable models assessed demographic variation in TMD symptom rates and evaluated associations with health care and analgesic use. One-third of the study subjects developed TMD symptoms and for a quarter of symptomatic episodes, pain intensity was severe. Initial TMD symptoms developed at an annual rate of 18.8 episodes per 100 persons. The annual rate more than doubled for first-recurrence and doubled again for second or subsequent recurrence such that, 1year after first recurrence, 71% of study subjects experienced a second recurrence. The overall rate increased with age and was greater in African Americans and lower in Asians relative to those of white race/ethnicity. The probability of TMD symptoms was strongly associated with concurrent episodes of headache and body pain and with past episodes of TMD symptoms. Episodes of TMD symptoms, headache, and body pain were associated with increases of ∼10% in probability of analgesic use and health care attendance. Yet, even when TMD, headache, and body pain occurred concurrently, 27% of study subjects neither attended health care nor used analgesics.

A community participatory study of cardiovascular health and exposure to near-highway air pollution: study design and methods

Current literature is insufficient to make causal inferences or establish dose-response relationships for traffic-related ultrafine particles (UFPs) and cardiovascular (CV) health. The Community Assessment of Freeway Exposure and Health (CAFEH) is a cross-sectional study of the relationship between UFP and biomarkers of CV risk. CAFEH uses a community-based participatory research framework that partners university researchers with community groups and residents. Our central hypothesis is that chronic exposure to UFP is associated with changes in biomarkers. The study enrolled more than 700 residents from three near-highway neighborhoods in the Boston metropolitan area in Massachusetts, USA. All participants completed an in-home questionnaire and a subset (440+) completed an additional supplemental questionnaire and provided biomarkers. Air pollution monitoring was conducted by a mobile laboratory equipped with fast-response instruments, at fixed sites, and inside the homes of selected study participants. We seek to develop improved estimates of UFP exposure by combining spatiotemporal models of ambient UFP with data on participant time-activity and housing characteristics. Exposure estimates will then be compared with biomarker levels to ascertain associations. This article describes our study design and methods and presents preliminary findings from east Somerville, one of the three study communities.

Drei Sozialformeln zur Schätzung der (prämorbiden) Intelligenzquotienten nach Wechsler

Im Rahmen neuropsychologischer Untersuchungen dient die Einschätzung der prämorbiden Leistungsfähigkeit der individuellen Defizitdiagnostik ebenso wie der realistischen Behandlungszielsetzung. Die dafür unverzichtbare biographische Anamnese kann durch Schätzmethoden ergänzt werden, deren Vor- und Nachteile skizziert werden. Insbesondere soziodemographische Formeln erlauben die Schätzung prämorbider Intelligenzniveaus unabhängig vom Krankheitsgeschehen, doch sind solche „Sozialformeln” hierzulande nicht verfügbar. Daher wurde eine nach Alter, Geschlecht und Schulbildung bevölkerungsrepräsentativ quotierte Stichprobe von 612 Gesunden zu zahlreichen soziodemographischen Merkmalen befragt. Zusätzlich wurden eine Intelligenztestbatterie (HAWIE-R), ein Wortschatztest (MWT-B) und ein Lesetest (LECTOR) durchgeführt. Mittels Regressionsanalysen wurden Vorhersagemodelle für den Gesamt-, Verbal- und Handlungs-IQ gewonnen und erfolgreich kreuzvalidiert. Trotz unvermeidlicher Regression zur Mitte ergab sich eine ausreichende Vorhersagegüte zumindest für den Gesamt- und Verbal-IQ. Die Schätzformeln sind anwendbar für Bewohner der Alten Bundesrepublik sowie nach 1985 geborene Ostdeutsche. Ein Fallbeispiel und das im Anhang enthaltene Berechnungsblatt erleichtern zusammen mit der Diskussion relevanter Indikations- und Interpretationsfragen die Formelanwendung.

A Typology of Violence against Self and Others and Its Associations with Drinking and Other Drug Use among High School Students in a U.S. General Population Survey

This study examined associations between binge drinking and other substance use and perpetration of violence against self and others. Data were pooled from the 2003, 2005, and was constructed to reflect four categories of behaviors: other-directed violence only, self-directed violence only, combined other- and self-directed violence, and no violence. Results from multinomial logistic regressions show that the frequency of binge drinking and other substance use were significant risk factors for each of the violence categories relative to no-violence. However, the strengths of these associations varied across the violence categories.