Damon Berridge

Background: Travelling away from home can be associated with fewer limits on behavior, particularly for students who participate in exchange programs. Aims: To examine the effects of eight moderators on change in alcohol use and related negative outcomes, drug use and unprotected sexual behavior in European study abroad students before, during, and after their time abroad. Methods: A three wave (before departure, while abroad, and after their return) longitudinal design collecting data on the frequency and volume of alcohol consumed, heavy episodic drinking, alcohol-related outcomes, drug use, and unprotected casual sex. Results: The baseline survey was completed by 1145 students participating in one or two semester exchange programs (67.5% spent up to a semester abroad), of which 906 participated in two or more waves, representing 42 and 33 countries of origin and destination, respectively. Mean age was 22.2 years (SD = 2.28) and 72.7% were female. Students increased the amount of alcohol consumed by 35% (B = 0.32; 95% CI 0.287–0.349) and experienced more alcohol-related consequences (B = 0.15; 95% CI 0.089–0.219) during the study abroad experience, though levels fell below pre-departure levels when they returned home. Factors related to greater alcohol use while abroad include pre-departure expectations about alcohol use during the study abroad experience, psychological adjustment to the host country, academic involvement, and host country living costs. No statistically meaningful change in drug use and unprotected sexual behavior was observed. Conclusions: Studying abroad exposes European students to additional time-limited alcohol-related health risks.

The paper explores competitive balance in the top tier of English league football during the era of the Premier League. This era is compared with both the preceding twenty-five year period and also with the era before World War I. Using several statistical measures, the analysis shows that the current high levels of competitive imbalance are not new phenomena. An explanation is proffered in terms of unequal resources between clubs which are inherent to the ‘longue durée’ of professional league football in England.

Background: Each year, approximately 800,000 people die by suicide worldwide, accounting for 1–2 in every 100 deaths. It is always a tragic event with a huge impact on family, friends, the community and health professionals. Unfortunately, suicide prevention and the development of risk assessment tools have been hindered by the complexity of the underlying mechanisms and the dynamic nature of a person’s motivation and intent. Many of those who die by suicide had contact with health services in the preceding year but identifying those most at risk remains a challenge. Objective: To explore the feasibility of using artificial neural networks with routinely collected electronic health records to support the identification of those at high risk of suicide when in contact with health services. Methods: Using the Secure Anonymised Information Linkage Databank UK, we extracted the data of those who died by suicide between 2001 and 2015 and paired controls. Looking at primary (general practice) and secondary (hospital admissions) electronic health records, we built a binary feature vector coding the presence of risk factors at different times prior to death. Risk factors included: general practice contact and hospital admission; diagnosis of mental health issues; injury and poisoning; substance misuse; maltreatment; sleep disorders; and the prescription of opiates and psychotropics. Basic artificial neural networks were trained to differentiate between the suicide cases and paired controls. We interpreted the output score as the estimated suicide risk. System performance was assessed with 10x10-fold repeated cross-validation, and its behavior was studied by representing the distribution of estimated risk across the cases and controls, and the distribution of factors across estimated risks. Results: We extracted a total of 2604 suicide cases and 20 paired controls per case. Our best system attained a mean error rate of 26.78% (SD 1.46; 64.57% of sensitivity and 81.86% of specificity). While the distribution of controls was concentrated around estimated risks < 0.5, cases were almost uniformly distributed between 0 and 1. Prescription of psychotropics, depression and anxiety, and self-harm increased the estimated risk by ~0.4. At least 95% of those presenting these factors were identified as suicide cases. Conclusions: Despite the simplicity of the implemented system, the proposed methodology obtained an accuracy like other published methods based on specialized questionnaire generated data. Most of the errors came from the heterogeneity of patterns shown by suicide cases, some of which were identical to those of the paired controls. Prescription of psychotropics, depression and anxiety, and self-harm were strongly linked with higher estimated risk scores, followed by hospital admission and long-term drug and alcohol misuse. Other risk factors like sleep disorders and maltreatment had more complex effects.

Background: Poor-quality housing adversely affects residents’ health but there is a paucity of high-quality evidence to support this.Objective: This research investigated the health impact of bringing housing to a national quality standard.Design: A natural experiment of improvements to housing quality analysed using repeated measures of health-care utilisation and economic outcomes at an individual person level.Setting: Carmarthenshire, UK.Participants: A total of 32,009 residents registered for a minimum of 60 days at 8558 social homes that received housing improvements between January 2005 and March 2015.Interventions: Multiple internal and external housing improvements, including wall and loft insulation, windows and doors, heating system upgrades, new kitchens and bathrooms, garden path safety improvements and electrical system upgrades (adding power sockets, and extractor fans in kitchens and bathrooms).Main outcome measures: Emergency hospital admissions for cardiorespiratory conditions and injuries. Primary health-care utilisation for respiratory and common mental health disorders, emergency department injury attendances and health-care utilisation costs.Data sources: Carmarthenshire County Council home address and intervention records were anonymously linked within the Secure Anonymous Information Linkage databank to demographic information from the Welsh Demographic Service data set; hospital admission data from the Patient Episode Dataset for Wales; primary care contacts and prescribed medications from general practice data; emergency department attendances from the Emergency Department Data Set; and deaths from the Office for National Statistics mortality register.Methods: The study used a longitudinal panel design to examine changes in standard of eight housing cointervention from intervention records, and linked to individuals registered at intervention homes. Health outcomes were obtained retrospectively for each individual in a dynamic cohort and were captured for up to 123 consecutive months. An additional local authority region could not be utilised as a comparator owing to different reporting pressures resulting in the recording of a different intervention. The exposure group for each cointervention was compared with an internal reference group of people living in homes that did not receive the cointervention during their tenancy. A multilevel modelling approach was used to account for repeated observations for individuals living in intervention homes. Counts of health outcomes were analysed using negative binomial regression models to determine the effect of each cointervention that reached housing quality standards during an individual’s period of tenancy, compared with those living in properties that did not. We adjusted for potential confounding factors and for background trends in the regional general population. A cost–consequences analysis was conducted as part of the health economic evaluation.Results: Residents aged ≥ 60 years living in homes in which electrical systems were upgraded were associated with 39% fewer admissions than those living in homes in which they were not [incidence rate ratio (IRR) 0.61, 95% confidence interval (CI) 0.53 to 0.72;p < 0.01]. Reduced admissions were also associated with windows and doors (IRR 0.71, 95% CI 0.63 to 0.81;p < 0.01), wall insulation (IRR 0.75, 95% CI 0.67 to 0.84;p < 0.01) and gardens and estates (IRR 0.73, 95% CI 0.64 to 0.83;p < 0.01) for those living in homes in which these cointervention were upgraded. There were no associations of change in emergency admissions with upgrading heating (IRR 0.91, 95% CI 0.82 to 1.01;p = 0.072), loft insulation (IRR 0.98, 95% CI 0.86 to 1.11;p = 0.695), kitchens (IRR 0.98, 95% CI 0.83 to 1.17;p = 0.843) or bathrooms (IRR 0.93, 95% CI 0.81 to 1.06;p = 0.287).Limitations: There was no randomisation, there were incomplete data on the scale of the intervention for individual households and there were no estimates for the impact of the whole programme.Conclusions: This complex interdisciplinary study found that hospital admissions could be avoided through improving housing quality standards.Future work: At their initiation, future non-health projects should have a built-in evaluation to allow intervention exposures to be randomly allocated to residents, with the simultaneous analysis of multiple health outcomes in one statistical model.Funding: The National Institute for Health Research Public Health Research programme.

BACKGROUND Each year, approximately 800,000 people die by suicide worldwide, accounting for 1–2 in every 100 deaths. It is always a tragic event with a huge impact on family, friends, the community and health professionals. Unfortunately, suicide prevention and the development of risk assessment tools have been hindered by the complexity of the underlying mechanisms and the dynamic nature of a person’s motivation and intent. Many of those who die by suicide had contact with health services in the preceding year but identifying those most at risk remains a challenge. OBJECTIVE To explore the feasibility of using artificial neural networks with routinely collected electronic health records to support the identification of those at high risk of suicide when in contact with health services. METHODS Using the Secure Anonymised Information Linkage Databank UK, we extracted the data of those who died by suicide between 2001 and 2015 and paired controls. Looking at primary (general practice) and secondary (hospital admissions) electronic health records, we built a binary feature vector coding the presence of risk factors at different times prior to death. Risk factors included: general practice contact and hospital admission; diagnosis of mental health issues; injury and poisoning; substance misuse; maltreatment; sleep disorders; and the prescription of opiates and psychotropics. Basic artificial neural networks were trained to differentiate between the suicide cases and paired controls. We interpreted the output score as the estimated suicide risk. System performance was assessed with 10x10-fold repeated cross-validation, and its behavior was studied by representing the distribution of estimated risk across the cases and controls, and the distribution of factors across estimated risks. RESULTS We extracted a total of 2604 suicide cases and 20 paired controls per case. Our best system attained a mean error rate of 26.78% (SD 1.46; 64.57% of sensitivity and 81.86% of specificity). While the distribution of controls was concentrated around estimated risks < 0.5, cases were almost uniformly distributed between 0 and 1. Prescription of psychotropics, depression and anxiety, and self-harm increased the estimated risk by ~0.4. At least 95% of those presenting these factors were identified as suicide cases. CONCLUSIONS Despite the simplicity of the implemented system, the proposed methodology obtained an accuracy like other published methods based on specialized questionnaire generated data. Most of the errors came from the heterogeneity of patterns shown by suicide cases, some of which were identical to those of the paired controls. Prescription of psychotropics, depression and anxiety, and self-harm were strongly linked with higher estimated risk scores, followed by hospital admission and long-term drug and alcohol misuse. Other risk factors like sleep disorders and maltreatment had more complex effects.

This paper explores competitive balance in top-tier English league football from its inception in 1888. It examines the extent to which finishing in the top four positions in successive seasons is the preserve of a small number of clubs. Using a range of statistical measures, the analysis shows that the current high levels of competitive imbalance are not new phenomena. The overall pattern approximates a ‘U curve’: current patterns parallel those in the 1890s. In the early years of English league football, differences in resources between clubs soon became apparent. Clubs from the larger conurbations generated consistently larger revenues than their counterparts in the smaller industrial towns. The introduction of the maximum wage in 1901 and the transfer system helped to stem these increasing inequalities between clubs. This coincided with a massive wave of new stadia construction which enabled all the clubs to compete on an increasingly level playing field. These conjunctural changes to English football before 1915 produced the era of relatively competitive football during the inter-war years. This continued until the abolition of the maximum wage in 1961. Since that time, competitive balance has reversed and become increasingly restricted. English top-tier football has re-entered an era of extreme competitive imbalance.

In this paper, the problem of identifying differentially expressed genes under different conditions using gene expression microarray data, in the presence of outliers, is discussed. For this purpose, the robust modeling of gene expression data using some powerful distributions known as normal/independent distributions is considered. These distributions include the Student’s t and normal distributions which have been used previously, but also include extensions such as the slash, the contaminated normal and the Laplace distributions. The purpose of this paper is to identify differentially expressed genes by considering these distributional assumptions instead of the normal distribution. A Bayesian approach using the Markov Chain Monte Carlo method is adopted for parameter estimation. Two publicly available gene expression data sets are analyzed using the proposed approach. The use of the robust models for detecting differentially expressed genes is investigated. This investigation shows that the choice of model for differentiating gene expression data is very important. This is due to the small number of replicates for each gene and the existence of outlying data. Comparison of the performance of these models is made using different statistical criteria and the ROC curve. The method is illustrated using some simulation studies. We demonstrate the flexibility of these robust models in identifying differentially expressed genes.

In this paper, a joint model for analyzing multivariate mixed ordinal and continuous responses, where continuous outcomes may be skew, is presented. For modeling the discrete ordinal responses, a continuous latent variable approach is considered and for describing continuous responses, a skew-normal mixed effects model is used. A Bayesian approach using Markov Chain Monte Carlo (MCMC) is adopted for parameter estimation. Some simulation studies are performed for illustration of the proposed approach. The results of the simulation studies show that the use of the separate models or the normal distributional assumption for shared random effects and within-subject errors of continuous and ordinal variables, instead of the joint modeling under a skew-normal distribution, leads to biased parameter estimates. The approach is used for analyzing a part of the British Household Panel Survey (BHPS) data set. Annual income and life satisfaction are considered as the continuous and the ordinal longitudinal responses, respectively. The annual income variable is severely skewed, therefore, the use of the normality assumption for the continuous response does not yield acceptable results. The results of data analysis show that gender, marital status, educational levels and the amount of money spent on leisure have a significant effect on annual income, while marital status has the highest impact on life satisfaction.