Bayesian imputation of non-chosen attribute values in revealed preference surveys

Obtaining attribute values of non-chosen alternatives in a revealed preference context is challenging because non-chosen alternative attributes are unobserved by choosers, chooser perceptions of attribute values may not reflect reality, existing methods for imputing these values suffer from shortcomings, and obtaining non-chosen attribute values is resource intensive. This paper presents a unique Bayesian (multiple) Imputation Multinomial Logit model that imputes unobserved travel times and distances of non-chosen travel modes based on random draws from the conditional posterior distribution of missing values. The calibrated Bayesian (multiple) Imputation Multinomial Logit model imputes non-chosen time and distance values that convincingly replicate observed choice behavior. Although network skims were used for calibration, more realistic data such as supplemental geographically referenced surveys or stated preference data may be preferred. The model is ideally suited for imputing variation in intrazonal non-chosen mode attributes and for assessing the marginal impacts of travel policies, programs, or prices within traffic analysis zones.

Lumbar epidural steroids are not cost effective in sciatica

Background Lumbar Epidural Steroids Injections (ESI’s) have previously been shown to provide some degree of pain relief in sciatica. Number Needed To Treat (NNT) to achieve 50% pain relief has been estimated at 7 from the results of randomised controlled trials. Pain relief is temporary. They remain one of the most commonly provided procedures in the UK. It is unknown whether this pain relief represents good value for money. Methods 228 patients were randomised into a multi-centre Double Blind Randomised Controlled Trial. Subjects received up to 3 ESI’s or intra-spinous saline depending on response and fall off with the first injection. All other treatments were permitted. All received a review of analgesia, education and physical therapy. Quality of life was assessed using the SF36 at 6 points and compared using independent sample t-tests. Follow up was up to 1 yr. Missing data was imputed using last observation carried forward (LOCF). QALY’s (Quality of Life Years) were derived from preference based heath values (summary health utility score). SF-6D health state classification was derived from SF-36 raw score data. Standard gambles (SG) were calculated using Model 10. SG scores were calculated on trial results. LOCF was not used for this. Instead average SG were derived for a subset of patients with observations for all visits up to week 12. Incremental QALY’s were derived as the difference in the area between the SG curve for the active group and placebo group. Results SF36 domains showed a significant improvement in pain at week 3 but this was not sustained (mean 54 Active vs 61 Placebo P<0.05). Other domains did not show any significant gains compared with placebo. For derivation of SG the number in the sample in each period differed. In week 12, average SG scores for active and placebo converged. In other words, the health gain for the active group as measured by SG was achieved by the placebo group by week 12. The incremental QALY gained for a patient under the trial protocol compared with the standard care package was 0.0059350. This is equivalent to an additional 2.2 days of full health. The cost per QALY gained to the provider from a patient management strategy administering one epidural as suggested by results was £25 745.68. This result was derived assuming that the gain in QALY data calculated for patients under the trial protocol would approximate that under a patient management strategy based on the trial results (one ESI). This is above the threshold suggested by some as a cost effective treatment. Conclusions The transient benefit in pain relief afforded by ESI’s does not appear to be cost-effective. Further work is needed to develop more cost-effective conservative treatments for sciatica.

Towards a phylogeny of the Tenebrionoidea (Coleoptera)

The phylogenetic relationships of the beetle superfamily Tenebrionoidea are investigated using the most comprehensive genetic data set compiled to date. With ∼34,000 described species in approximately 1250 genera and 28 families, Tenebrionoidea represent one of the most diverse and species-rich superfamilies of beetles. The interfamilial relationships of the Tenebrionoidea are poorly known; previous morphological and molecular phylogenies recovered few well-supported and often conflicting relationships between families. Here we present a molecular phylogeny of Tenebrionoidea based on genes commonly used to resolve family and superfamily-level phylogenies of beetles (18S, 28S, 16S, 12S, tRNA Val and COI). The alignment spanned over 6.5 KB of DNA sequence and over 300 tenebrionoid genera from 24 of the 28 families were sampled. Maximum Likelihood and Bayesian analysis could not resolve deeper level divergences within the superfamily and very few relationships between families were supported. Increasing gene coverage in the alignment by removing taxa with missing data did not improve clade support but when rogue taxa were removed increased resolution was recovered. Investigation of signal strength suggested conflicting phylogenetic signal was present in the standard genes used for beetle phylogenetics, even when rogue taxa were removed. Our study of Tenebrionoidea highlights that even with relatively comprehensive taxon sampling within a lineage, this standard set of genes is unable to resolve relationships within this superfamily.

Development and validation of the ankle fracture outcome of rehabilitation measure (A-FORM)

Study Design Delphi panel and cohort study. Objective To develop and refine a condition-specific, patient-reported outcome measure, the Ankle Fracture Outcome of Rehabilitation Measure (A-FORM), and to examine its psychometric properties, including factor structure, reliability, and validity, by assessing item fit with the Rasch model. Background To our knowledge, there is no patient-reported outcome measure specific to ankle fracture with a robust content foundation. Methods A 2-stage research design was implemented. First, a Delphi panel that included patients and health professionals developed the items and refined the item wording. Second, a cohort study (n = 45) with 2 assessment points was conducted to permit preliminary maximum-likelihood exploratory factor analysis and Rasch analysis. Results The Delphi panel reached consensus on 53 potential items that were carried forward to the cohort phase. From the 2 time points, 81 questionnaires were completed and analyzed; 38 potential items were eliminated on account of greater than 10% missing data, factor loadings, and uniqueness. The 15 unidimensional items retained in the scale demonstrated appropriate person and item reliability after (and before) removal of 1 item (anxious about footwear) that had a higher-than-ideal outfit statistic (1.75). The “anxious about footwear” item was retained in the instrument, but only the 14 items with acceptable infit and outfit statistics (range, 0.5–1.5) were included in the summary score. Conclusion This investigation developed and refined the A-FORM (Version 1.0). The A-FORM items demonstrated favorable psychometric properties and are suitable for conversion to a single summary score. Further studies utilizing the A-FORM instrument are warranted. J Orthop Sports Phys Ther 2014;44(7):488–499. Epub 22 May 2014. doi:10.2519/jospt.2014.4980

Association between Australian-Indian mothers' controlling feeding practices and children's appetite traits

This cross-sectional study examined the association between controlling feeding practices and children's appetite traits. The secondary aim studied the relationship between controlling feeding practices and two proxy indicators of diet quality. Participants were 203 Australian-Indian mothers with children aged 1-5 years. Controlling feeding practices (pressure to eat, restriction, monitoring) and children's appetite traits (. food approach traits: food responsiveness, enjoyment of food, desire to drink, emotional overeating; food avoidance traits: satiety responsiveness, slowness in eating, fussiness and emotional undereating) were measured using self-reported, previously validated scales/questionnaires. Children's daily frequency of consumption of core and non-core foods was estimated using a 49-item list of foods eaten (yes/no) in the previous 24 hours as an indicator of diet quality. Higher pressure to eat was associated with higher scores for satiety responsiveness, slowness in eating, fussiness and lower score for enjoyment of food. Higher restriction was related to higher scores for food responsiveness and emotional overeating. Higher monitoring was inversely associated with fussiness, slowness in eating, food responsiveness and emotional overeating and positively associated with enjoyment of food. Pressure to eat and monitoring were related to lower number of core and non-core foods consumed in the previous 24 hours, respectively. All associations remained significant after adjusting for maternal and child covariates (n = 152 due to missing data). In conclusion, pressure to eat was associated with higher food avoidance traits and lower consumption of core foods. Restrictive feeding practices were associated with higher food approach traits. In contrast, monitoring practices were related to lower food avoidance and food approach traits and lower non-core food consumption.

Global environmental emissions estimate: Application of multiple imputation

A new database called the World Resource Table is constructed in this study. Missing values are known to produce complications when constructing global databases. This study provides a solution for applying multiple imputation techniques and estimates the global environmental Kuznets curve (EKC) for CO2, SO2, PM10, and BOD. Policy implications for each type of emission are derived based on the results of the EKC using WRI. Finally, we predicted the future emissions trend and regional share of CO2 emissions. We found that East Asia and South Asia will be increasing their emissions share while other major CO2 emitters will still produce large shares of the total global emissions.

Viewing systematic reviews and meta-analysis in social research through different lenses

Systematic reviews and meta-analyses are used to combine results across studies to determine an overall effect. Meta-analysis is especially useful for combining evidence to inform social policy, but meta-analyses of applied social science research may encounter practical issues arising from the nature of the research domain. The current paper identifies potential resolutions to four issues that may be encountered in systematic reviews and meta-analyses in social research. The four issues are: scoping and targeting research questions appropriate for meta-analysis; selecting eligibility criteria where primary studies vary in research design and choice of outcome measures; dealing with inconsistent reporting in primary studies; and identifying sources of heterogeneity with multiple confounded moderators. The paper presents an overview of each issue with a review of potential resolutions, identified from similar issues encountered in meta-analysis in medical and biological sciences. The discussion aims to share and improve methodology in systematic reviews and meta-analysis by promoting cross-disciplinary communication, that is, to encourage 'viewing through different lenses'.

Living well with diabetes: 24-month outcomes from a randomized trial of telephone-delivered weight loss and physical activity intervention to improve glycemic control

OBJECTIVE: To evaluate the effectiveness of a telephone-delivered behavioral weight loss and physical activity intervention targeting Australian primary care patients with type 2 diabetes. RESEARCH DESIGN AND METHODS: Pragmatic randomized controlled trial of telephone counseling (n = 151) versus usual care (n = 151). Reported here are 18-month (end-of-intervention) and 24-month (maintenance) primary outcomes of weight, moderate-to-vigorous-intensity physical activity (MVPA; via accelerometer), and HbA1c level. Secondary outcomes include dietary energy intake and diet quality, waist circumference, lipid levels, and blood pressure. Data were analyzed via adjusted linear mixed models with multiple imputation of missing data. RESULTS: Relative to usual-care participants, telephone counseling participants achieved modest, but significant, improvements in weight loss (relative rate [RR] -1.42% of baseline body weight [95% CI -2.54 to -0.30% of baseline body weight]), MVPA (RR 1.42 [95% CI 1.06-1.90]), diet quality (2.72 [95% CI 0.55-4.89]), and waist circumference (-1.84 cm [95% CI -3.16 to -0.51 cm]), but not in HbA1c level (RR 0.99 [95% CI 0.96-1.02]), or other cardio-metabolic markers. None of the outcomes showed a significant change/deterioration over the maintenance period. However, only the intervention effect for MVPA remained statistically significant at 24 months. CONCLUSIONS: The modest improvements in weight loss and behavior change, but the lack of changes in cardio-metabolic markers, may limit the utility, scalability, and sustainability of such an approach.

Separable spatiotemporal priors for convex reconstruction of time-varying 3D point clouds

Reconstructing 3D motion data is highly under-constrained due to several common sources of data loss during measurement, such as projection, occlusion, or miscorrespondence. We present a statistical model of 3D motion data, based on the Kronecker structure of the spatiotemporal covariance of natural motion, as a prior on 3D motion. This prior is expressed as a matrix normal distribution, composed of separable and compact row and column covariances. We relate the marginals of the distribution to the shape, trajectory, and shape-trajectory models of prior art. When the marginal shape distribution is not available from training data, we show how placing a hierarchical prior over shapes results in a convex MAP solution in terms of the trace-norm. The matrix normal distribution, fit to a single sequence, outperforms state-of-the-art methods at reconstructing 3D motion data in the presence of significant data loss, while providing covariance estimates of the imputed points.

The development and piloting of the ISPCAN child abuse screening tool-parent version (ICAST-P)

Objective Child maltreatment is a problem that has longer recognition in the northern hemisphere and in high-income countries. Recent work has highlighted the nearly universal nature of the problem in other countries but demonstrated the lack of comparability of studies because of the variations in definitions and measures used. The International Society for the Prevention of Child Abuse and Neglect has developed instrumentation that may be used with cross-cultural and cross-national benchmarking by local investigators. Design and sampling The instrument design began with a team of expert in Brisbane in 2004. A large bank of questions were subjected to two rounds of Delphi review to develop the fielded version of the instrument. Convenience samples included approximately 120 parent respondents with children under the age of 18 in each of six countries (697 total). Results This paper presents an instrument that measures parental behaviors directed at children and reports data from pilot work in 6 countries and 7 languages. Patterns of response revealed few missing values and distributions of responses that generally were similar in the six countries. Subscales performed well in terms of internal consistency with Cronbach's alpha in very good range (0.77–0.88) with the exception of the neglect and sex abuse subscales. Results varied by child age and gender in expected directions but with large variations among the samples. About 15% of children were shaken, 24% hit on the buttocks with an object, and 37% were spanked. Reports of choking and smothering were made by 2% of parents. Conclusion These pilot data demonstrate that the instrument is well tolerated and captures variations in, and potentially harmful forms of child discipline. Practice implications The ISPCAN Child Abuse Screening Tool – Parent Version (ICAST-P) has been developed as a survey instrument to be administered to parents for the assessment of child maltreatment in a multi-national and multi-cultural context. It was developed with broad input from international experts and subjected to Dephi review, translation, and pilot testing in six countries. The results of the Delphi study and pilot testing are presented. This study demonstrates that a single instrument can be used in a broad range of cultures and languages with low rates of missing data and moderate to high internal consistency.

ISPCAN child abuse screening tool children's version (ICAST-C): Instrument development and multi-national pilot testing

Objective To develop a child victimization survey among a diverse group of child protection experts and examine the performance of the instrument through a set of international pilot studies. Methods The initial draft of the instrument was developed after input from scientists and practitioners representing 40 countries. Volunteers from the larger group of scientists participating in the Delphi review of the ICAST P and R reviewed the ICAST C by email in 2 rounds resulting in a final instrument. The ICAST C was then translated and back translated into six languages and field tested in four countries using a convenience sample of 571 children 12–17 years of age selected from schools and classrooms to which the investigators had easy access. Results The final ICAST C Home has 38 items and the ICAST C Institution has 44 items. These items serve as screeners and positive endorsements are followed by queries for frequency and perpetrator. Half of respondents were boys (49%). Endorsement for various forms of victimization ranged from 0 to 51%. Many children report violence exposure (51%), physical victimization (55%), psychological victimization (66%), sexual victimization (18%), and neglect in their homes (37%) in the last year. High rates of physical victimization (57%), psychological victimization (59%), and sexual victimization (22%) were also reported in schools in the last year. Internal consistency was moderate to high (alpha between .685 and .855) and missing data low (less than 1.5% for all but one item). Conclusions In pilot testing, the ICAST C identifies high rates of child victimization in all domains. Rates of missing data are low, and internal consistency is moderate to high. Pilot testing demonstrated the feasibility of using child self-report as one strategy to assess child victimization. Practice implications The ICAST C is a multi-national, multi-lingual, consensus-based survey instrument. It is available in six languages for international research to estimate child victimization. Assessing the prevalence of child victimization is critical in understanding the scope of the problem, setting national and local priorities, and garnering support for program and policy development aimed at child protection.

Discussion of “Generalized estimating equations: Notes on the choice of the working correlation matrix”

Objective To discuss generalized estimating equations as an extension of generalized linear models by commenting on the paper of Ziegler and Vens "Generalized Estimating Equations. Notes on the Choice of the Working Correlation Matrix". Methods Inviting an international group of experts to comment on this paper. Results Several perspectives have been taken by the discussants. Econometricians have established parallels to the generalized method of moments (GMM). Statisticians discussed model assumptions and the aspect of missing data Applied statisticians; commented on practical aspects in data analysis. Conclusions In general, careful modeling correlation is encouraged when considering estimation efficiency and other implications, and a comparison of choosing instruments in GMM and generalized estimating equations, (GEE) would be worthwhile. Some theoretical drawbacks of GEE need to be further addressed and require careful analysis of data This particularly applies to the situation when data are missing at random.

Personality, health and lifestyle in a questionnaire family study: A comparison between highly cooperative and less cooperative families

The effect of nonresponse on health and lifestyle measures has received extensive study, showing at most relatively modest effects. Nonresponse bias with respect to personality has been less thoroughly investigated. The present study uses data from responding individuals as a proxy for the missing data of their nonresponding family members to examine the presence of nonresponse bias for personality traits and disorders as well as health and lifestyle traits. We looked at the Big Five personality traits, borderline personality disorder (BPD) features, attention-deficit/hyperactivity disorder, Anger, and several measures of health (Body Mass Index, migraine) and lifestyle (smoking, alcohol use). In general, outcomes tend to be slightly more favorable for individuals from highly cooperative families compared to individuals from less cooperative families. The only significant difference was found for BPD features (p = .001). However, the absolute difference in mean scores is very small, less than 1 point for a scale ranging from 0 to 72. In conclusion, survey data on personality, health and lifestyle are relatively unbiased with respect to nonresponse.

Nineteenth century narratives reveal historic catch rates for Australian snapper (Pagrus auratus)

Snapper (Pagrus auratus) is widely distributed throughout subtropical and temperate southern oceans and forms a significant recreational and commercial fishery in Queensland, Australia. Using data from government reports, media sources, popular publications and a government fisheries survey carried out in 1910, we compiled information on individual snapper fishing trips that took place prior to the commencement of fisherywide organized data collection, from 1871 to 1939. In addition to extracting all available quantitative data, we translated qualitative information into bounded estimates and used multiple imputation to handle missing values, forming 287 records for which catch rate (snapper fisher−1 h−1) could be derived. Uncertainty was handled through a parametric maximum likelihood framework (a transformed trivariate Gaussian), which facilitated statistical comparisons between data sources. No statistically significant differences in catch rates were found among media sources and the government fisheries survey. Catch rates remained stable throughout the time series, averaging 3.75 snapper fisher−1 h−1 (95% confidence interval, 3.42–4.09) as the fishery expanded into new grounds. In comparison, a contemporary (1993–2002) south-east Queensland charter fishery produced an average catch rate of 0.4 snapper fisher−1 h−1 (95% confidence interval, 0.31–0.58). These data illustrate the productivity of a fishery during its earliest years of development and represent the earliest catch rate data globally for this species. By adopting a formalized approach to address issues common to many historical records – missing data, a lack of quantitative information and reporting bias – our analysis demonstrates the potential for historical narratives to contribute to contemporary fisheries management.

Construction of a global map of human gene expression : the process, tools and analysis

This thesis studies human gene expression space using high throughput gene expression data from DNA microarrays. In molecular biology, high throughput techniques allow numerical measurements of expression of tens of thousands of genes simultaneously. In a single study, this data is traditionally obtained from a limited number of sample types with a small number of replicates. For organism-wide analysis, this data has been largely unavailable and the global structure of human transcriptome has remained unknown. This thesis introduces a human transcriptome map of different biological entities and analysis of its general structure. The map is constructed from gene expression data from the two largest public microarray data repositories, GEO and ArrayExpress. The creation of this map contributed to the development of ArrayExpress by identifying and retrofitting the previously unusable and missing data and by improving the access to its data. It also contributed to creation of several new tools for microarray data manipulation and establishment of data exchange between GEO and ArrayExpress. The data integration for the global map required creation of a new large ontology of human cell types, disease states, organism parts and cell lines. The ontology was used in a new text mining and decision tree based method for automatic conversion of human readable free text microarray data annotations into categorised format. The data comparability and minimisation of the systematic measurement errors that are characteristic to each lab- oratory in this large cross-laboratories integrated dataset, was ensured by computation of a range of microarray data quality metrics and exclusion of incomparable data. The structure of a global map of human gene expression was then explored by principal component analysis and hierarchical clustering using heuristics and help from another purpose built sample ontology. A preface and motivation to the construction and analysis of a global map of human gene expression is given by analysis of two microarray datasets of human malignant melanoma. The analysis of these sets incorporate indirect comparison of statistical methods for finding differentially expressed genes and point to the need to study gene expression on a global level.