An assessment of the statistical methods used to analyse toxicology studies

The Statisticians in the Pharmaceutical Industry Toxicology Special Interest Group has collated and compared statistical analysis methods for a number of toxicology study types including general toxicology, genetic toxicology, safety pharmacology and carcinogenicity. In this paper, we present the study design, experimental units and analysis methods.     

Review of the statistical analysis of the dog telemetry study

In 2008, the PSI Toxicology Special Interest Group met to discuss the design and analysis of dog telemetry studies. The dog telemetry study is one component of the integrated cardiovascular assessment required by regulatory bodies. Although there are guidelines for these studies, little is said about the statistical analysis. With parameters of interest measured continually over time, in studies typically involving four dogs, the analysis is not straightforward. This has led to many different types of analysis being proposed in the literature, with many different methods applied within the pharmaceutical industry itself. This paper summarizes the PSI Toxicology group's discussions and recommendations around these issues.     

Performance and risk management

Abstract

Performance and risk management are seen by some as two ends of the same spectrum. Performance measurement and management is about steering an enterprise towards a profitable and viable future, whilst risk management is about avoiding the pitfalls that can overwhelm and ultimately put an enterprise out of business. But should the functions and processes of performance measurement and management be integrated with those of risk management? What are the consequences of this integration? How should this be done in practice? In this editorial we briefly chart the debate between those who propose it is important to keep the functions separate and those who advocate integration before presenting the empirical research that informs this conversation.     

Losses Disguised as Wins Affect Game Selection on Multiline Slots

Multiline slots are exciting games that contain features which make them alluring. One such feature is a loss disguised as a win (LDW); wherein, players win less than they wager (e.g., bet 2 dollars, win back 50 cents), but this net loss is disguised by flashing graphics and winning sounds. Research to date concludes that LDWs are both rewarding and reinforcing. Here, we investigated whether LDWs affect players’ game selection. Thirty-two undergraduate students with experience playing slot machines played 100 spins on four games—two had positive payback percentages (115%) and two had negative payback percentages (85%) after 100 spins. For each payback percentage condition, there was a game with no LDWs and a game with a moderate number of LDWs. For the 100 spins, players could choose to play whichever game they wished. They then rated their preference for each game following the 100-spins and chose a game to continue playing. The majority of players preferred playing the positive payback percentage game with LDWs and chose to continue playing this game over the three other games. We conclude that in addition to LDWs being reinforcing and rewarding, LDWs do in fact influence game selection. We conclude that responsible gambling initiatives should educate players about LDWs.     

A Risk Assessment Framework for the Socioeconomic Impacts of Electricity Transmission Infrastructure Failure Due to Space Weather: An Application to the United Kingdom

Space weather phenomena have been studied in detail in the peer‐reviewed scientific literature. However, there has arguably been scant analysis of the potential socioeconomic impacts of space weather, despite a growing gray literature from different national studies, of varying degrees of methodological rigor. In this analysis, we therefore provide a general framework for assessing the potential socioeconomic impacts of critical infrastructure failure resulting from geomagnetic disturbances, applying it to the British high‐voltage electricity transmission network. Socioeconomic analysis of this threat has hitherto failed to address the general geophysical risk, asset vulnerability, and the network structure of critical infrastructure systems. We overcome this by using a three‐part method that includes (i) estimating the probability of intense magnetospheric substorms, (ii) exploring the vulnerability of electricity transmission assets to geomagnetically induced currents, and (iii) testing the socioeconomic impacts under different levels of space weather forecasting. This has required a multidisciplinary approach, providing a step toward the standardization of space weather risk assessment. We find that for a Carrington‐sized 1‐in‐100‐year event with no space weather forecasting capability, the gross domestic product loss to the United Kingdom could be as high as £15.9 billion, with this figure dropping to £2.9 billion based on current forecasting capability. However, with existing satellites nearing the end of their life, current forecasting capability will decrease in coming years. Therefore, if no further investment takes place, critical infrastructure will become more vulnerable to space weather. Additional investment could provide enhanced forecasting, reducing the economic loss for a Carrington‐sized 1‐in‐100‐year event to £0.9 billion.     

UK Paramedics Confidence in Identifying Child Sexual Abuse: A Mixed-Methods Investigation

United Kingdom (UK) paramedics are in a prime position to identify children and young people who are victims or at risk of sexual abuse. Paramedics have access, by phone, or in person, to unprepared homes and communities which other health professionals such as social workers may not. Little research exists however, investigating UK paramedic confidence in identifying child sexual abuse. This mixed-method explanatory sequential investigation used the self-reported confidence levels of 276 UK paramedics to inform the design of seven semi-structured focus groups with 25 UK paramedics from a large ambulance service with operating models similar to all UK services. Multiple factors contribute to a lack of confidence in identifying child sexual abuse, child sexual exploitation, and female genital mutilation, including a perceived lack of exposure to sexual abuse, the perceived hidden nature of sexual abuse, and the lack of physical symptoms and examination. An overarching lack of knowledge is the most significant contributor to a lack of confidence which in turn perpetuates misinformation surrounding prevalence, location, and the signs and symptoms of sexual abuse. These findings suggest a lack of sufficient training and a need for further research evaluating the content of current training and its method of delivery.     

Bayesian CV@R/super-quantile regression

In this paper we provide a Bayesian interpretation of the conditional value at risk, CV@R, or super-quantile regression recently developed by Rockafellar et al. [Super-quantile regression with applications to buffered reliability, uncertainty quantification, and conditional value-at-risk, Eur. J. Oper. Res. 234 (2014), pp. 140–154]. Computations are based on particle filtering using a special posterior distribution consistent with the super-quantile concept. An empirical application to data used by RRM as well to another data set on energy prices confirms their results and shows the applicability of the new techniques.     

Does welfare reform affect fertility? Evidence from the UK

This paper provides evidence on the effect of welfare reform on fertility, focusing on UK reforms in 1999 that increased per-child spending by 50% in real terms. We use a difference-in-differences approach, exploiting the fact that the reforms were targeted at low-income households. The reforms were likely to differentially affect the fertility of women in couples and single women because of the opportunity cost effects of the welfare-to-work element. We find no increase in births among single women, but evidence to support an increase in births (by around 15%) among coupled women.     

An evaluation of methods for testing hypotheses relating to two endpoints in a single clinical trial

The issues and dangers involved in testing multiple hypotheses are well recognised within the pharmaceutical industry. In reporting clinical trials, strenuous efforts are taken to avoid the inflation of type I error, with procedures such as the Bonferroni adjustment and its many elaborations and refinements being widely employed. Typically, such methods are conservative. They tend to be accurate if the multiple test statistics involved are mutually independent and achieve less than the type I error rate specified if these statistics are positively correlated. An alternative approach is to estimate the correlations between the test statistics and to perform a test that is conditional on those estimates being the true correlations. In this paper, we begin by assuming that test statistics are normally distributed and that their correlations are known. Under these circumstances, we explore several approaches to multiple testing, adapt them so that type I error is preserved exactly and then compare their powers over a range of true parameter values. For simplicity, the explorations are confined to the bivariate case. Having described the relative strengths and weaknesses of the approaches under study, we use simulation to assess the accuracy of the approximate theory developed when the correlations are estimated from the study data rather than being known in advance and when data are binary so that test statistics are only approximately normally distributed.