Exact and approximate Bayesian inference for low count time series models with intractable likelihoods

In this paper we present a new simulation methodology in order to obtain exact or approximate Bayesian inference for models for low-valued count time series data that have computationally demanding likelihood functions. The algorithm fits within the framework of particle Markov chain Monte Carlo (PMCMC) methods. The particle filter requires only model simulations and, in this regard, our approach has connections with approximate Bayesian computation (ABC). However, an advantage of using the PMCMC approach in this setting is that simulated data can be matched with data observed one-at-a-time, rather than attempting to match on the full dataset simultaneously or on a low-dimensional non-sufficient summary statistic, which is common practice in ABC. For low-valued count time series data we find that it is often computationally feasible to match simulated data with observed data exactly. Our particle filter maintains $N$ particles by repeating the simulation until $N+1$ exact matches are obtained. Our algorithm creates an unbiased estimate of the likelihood, resulting in exact posterior inferences when included in an MCMC algorithm. In cases where exact matching is computationally prohibitive, a tolerance is introduced as per ABC. A novel aspect of our approach is that we introduce auxiliary variables into our particle filter so that partially observed and/or non-Markovian models can be accommodated. We demonstrate that Bayesian model choice problems can be easily handled in this framework.

Comparison of moped, scooter and motorcycle crash risk and crash severity

The increased popularity of mopeds and motor scooters in Australia and elsewhere in the last decade has contributed substantially to the greater use of powered two-wheelers (PTWs) as a whole. As the exposure of mopeds and scooters has increased, so too has the number of reported crashes involving those PTW types, but there is currently little research comparing the safety of mopeds and, particularly, larger scooters with motorcycles. This study compared the crash risk and crash severity of motorcycles, mopeds and larger scooters in Queensland, Australia. Comprehensive data cleansing was undertaken to separate motorcycles, mopeds and larger scooters in police-reported crash data covering the five years to 30 June 2008. The crash rates of motorcycles (including larger scooters) and mopeds in terms of registered vehicles were similar over this period, although the moped crash rate showed a stronger downward trend. However, the crash rates in terms of distance travelled were nearly four times higher for mopeds than for motorcycles (including larger scooters). More comprehensive distance travelled data is needed to confirm these findings. The overall severity of moped and scooter crashes was significantly lower than motorcycle crashes but an ordered probit regression model showed that crash severity outcomes related to differences in crash characteristics and circumstances, rather than differences between PTW types per se. Greater motorcycle crash severity was associated with higher (>80 km/h) speed zones, horizontal curves, weekend, single vehicle and nighttime crashes. Moped crashes were more severe at night and in speed zones of 90 km/h or more. Larger scooter crashes were more severe in 70 km/h zones (than 60 km/h zones) but not in higher speed zones, and less severe on weekends than on weekdays. The findings can be used to inform potential crash and injury countermeasures tailored to users of different PTW types.

Random breath testing in Queensland and Western Australia : Examination of how the random breath testing rate influences alcohol related traffic crash rates

In this paper we explore the relationship between monthly random breath testing (RBT) rates (per 1000 licensed drivers) and alcohol-related traffic crash (ARTC) rates over time, across two Australian states: Queensland and Western Australia. We analyse the RBT, ARTC and licensed driver rates across 12 years; however, due to administrative restrictions, we model ARTC rates against RBT rates for the period July 2004 to June 2009. The Queensland data reveals that the monthly ARTC rate is almost flat over the five year period. Based on the results of the analysis, an average of 5.5 ARTCs per 100,000 licensed drivers are observed across the study period. For the same period, the monthly rate of RBTs per 1000 licensed drivers is observed to be decreasing across the study with the results of the analysis revealing no significant variations in the data. The comparison between Western Australia and Queensland shows that Queensland's ARTC monthly percent change (MPC) is 0.014 compared to the MPC of 0.47 for Western Australia. While Queensland maintains a relatively flat ARTC rate, the ARTC rate in Western Australia is increasing. Our analysis reveals an inverse relationship between ARTC RBT rates, that for every 10% increase in the percentage of RBTs to licensed driver there is a 0.15 decrease in the rate of ARTCs per 100,000 licenced drivers. Moreover, in Western Australia, if the 2011 ratio of 1:2 (RBTs to annual number of licensed drivers) were to double to a ratio of 1:1, we estimate the number of monthly ARTCs would reduce by approximately 15. Based on these findings we believe that as the number of RBTs conducted increases the number of drivers willing to risk being detected for drinking driving decreases, because the perceived risk of being detected is considered greater. This is turn results in the number of ARTCs diminishing. The results of this study provide an important evidence base for policy decisions for RBT operations.

Prognostic factors for malignant mesothelioma in 142 patients : validation of CALGB and EORTC prognostic scoring systems

Background The incidence of malignant mesothelioma is increasing. There is the perception that survival is worse in the UK than in other countries. However, it is important to compare survival in different series based on accurate prognostic data. The European Organisation for Research and Treatment of Cancer (EORTC) and the Cancer and Leukaemia Group B (CALGB) have recently published prognostic scoring systems. We have assessed the prognostic variables, validated the EORTC and CALGB prognostic groups, and evaluated survival in a series of 142 patients. Methods Case notes of 142 consecutive patients presenting in Leicester since 1988 were reviewed. Univariate analysis of prognostic variables was performed using a Cox proportional hazards regression model. Statistically significant variables were analysed further in a forward, stepwise multivariate model. EORTC and CALGB prognostic groups were derived, Kaplan-Meier survival curves plotted, and survival rates were calculated from life tables. Results Significant poor prognostic factors in univariate analysis included male sex, older age, weight loss, chest pain, poor performance status, low haemoglobin, leukocytosis, thrombocytosis, and non-epithelial cell type (p<0.05). The prognostic significance of cell type, haemoglobin, white cell count, performance status, and sex were retained in the multivariate model. Overall median survival was 5.9 (range 0-34.3) months. One and two year survival rates were 21.3% (95% CI 13.9 to 28.7) and 3.5% (0 to 8.5), respectively. Median, one, and two year survival data within prognostic groups in Leicester were equivalent to the EORTC and CALGB series. Survival curves were successfully stratified by the prognostic groups. Conclusions This study validates the EORTC and CALGB prognostic scoring systems which should be used both in the assessment of survival data of series in different countries and in the stratification of patients into randomised clinical studies.

EGFR expression : associations with outcome and clinicopathological variables in malignant pleural mesothelioma

Malignant mesothelioma (MM) is a fatal tumour of increasing incidence which is related to asbestos exposure. This work evaluated expression in MM of Epidermal Growth Factor Receptor (EGFR) by immunohistochemistry in 168 tumour sections and its correlations with clinicopathological and biological factors. The microvessel density (MVD) was derived from CD34 immunostained sections. Hematoxylin and eosin stained sections were examined for intratumoural necrosis. COX-2 protein expression was evaluated with semi-quantitative Western blotting of homogenised tumour supernatants (n = 45). EGFR expression was correlated with survival by Kaplan-Meier and log rank analysis. Univariate and multivariate Cox proportional hazards models were used to compare the effects of EGFR with clinicopathological and biological prognostic factors and prognostic scoring systems. EGFR expression was identified in 74 cases (44%) and correlated with epithelioid cell type (p < 0.0001), good performance status (p < 0.0001), the absence of chest pain (p < 0.0001) and the presence of TN (p = 0.004), but not MVD or COX-2. EGFR expression was a good prognostic factor in univariate analysis (p = 0.01). Independent indicators of poor prognosis in multivariate analysis were non-epithelioid cell type (p = 0.0001), weight loss, performance status and WBC > 8.3 × 10 9 L -1. EGFR status was not an independent prognostic factor. EGFR expression in MM correlates with epithelioid histology and TN. EGFR may be a target for selective therapies in MM. © 2006 Elsevier Ireland Ltd. All rights reserved.

Galectin-1 : a link between tumor hypoxia and tumor immune privilege

Purpose: To identify a 15-KDa novel hypoxia-induced secreted protein in head and neck squamous cell carcinomas (HNSCC) and to determine its role in malignant progression. Methods: We used surface-enhanced laser desorption ionization time-of-flight mass spectrometry (SELDI-TOF-MS) and tandem MS to identify a novel hypoxia-induced secreted protein in FaDu cells. We used immunoblots, real-time polymerase chain reaction (PCR), and enzyme-linked immunoabsorbent assay to confirm the hypoxic induction of this secreted protein as galectin-1 in cell lines and xenografts. We stained tumor tissues from 101 HNSCC patients for galectin-1, CA IX (carbonic anhydrase IX, a hypoxia marker) and CDS (a T-cell marker). Expression of these markers was correlated to each other and to treatment outcomes. Results: SELDI-TOF studies yielded a hypoxia-induced peak at 15 kDa that proved to be galectin-1 by MS analysis. Immunoblots and PCR studies confirmed increased galectin-1 expression by hypoxia in several cancer cell lines. Plasma levels of galectin-1 were higher in tumor-bearing severe combined immunodeficiency (SCID) mice breathing 10% O 2 compared with mice breathing room air. In HNSCC patients, there was a significant correlation between galectin-1 and CA IX staining (P = .01) and a strong inverse correlation between galectin-1 and CDS staining (P = .01). Expression of galectin-1 and CDS were significant predictors for overall survival on multivariate analysis. Conclusion: Galectin-1 is a novel hypoxia-regulated protein and a prognostic marker in HNSCC. This study presents a new mechanism on how hypoxia can affect the malignant progression and therapeutic response of solid tumors by regulating the secretion of proteins that modulate immune privilege. © 2005 by American Society of Clinical Oncology.

SRL 172 (killed Mycobacterium vaccae) in addition to standard chemotherapy improves quality of life without affecting survival, in patients with advanced non-small-cell lung cancer : phase III results

Background: This open-label, randomised phase III study was designed to further investigate the clinical activity and safety of SRL172 (killed Mycobacterium vaccae suspension) with chemotherapy in the treatment of non-small-cell lung cancer (NSCLC). Patients and methods: Patients were randomised to receive platinum-based chemotherapy, consisting of up to six cycles of MVP (mitomycin, vinblastine and cisplatin or carboplatin) with (210 patients) or without (209 patients) monthly SRL172. Results: There was no statistical difference between the two groups in overall survival (primary efficacy end point) over the course of the study (median overall survival of 223 days versus 225 days; P = 0.65). However, a higher proportion of patients were alive at the end of the 15-week treatment phase in the chemotherapy plus SRL172 group (90%), than in the chemotherapy alone group (83%) (P = 0.061). At the end of the treatment phase, the response rate was 37% in the combined group and 33% in the chemotherapy alone group. Patients in the chemotherapy alone group had greater deterioration in their Global Health Status score (-14.3) than patients in the chemotherapy plus SRL172 group (-6.6) (P = 0.02). Conclusion: In this non-placebo controlled trial, SRL172 when added to standard cancer chemotherapy significantly improved patient quality of life without affecting overall survival times. © 2004 European Society for Medical Oncology.

Cetuximab plus chemotherapy in patients with advanced non-small-cell lung cancer (FLEX) : an open-label randomised phase III trial

Background: Use of cetuximab, a monoclonal antibody targeting the epidermal growth factor receptor (EGFR), has the potential to increase survival in patients with advanced non-small-cell lung cancer. We therefore compared chemotherapy plus cetuximab with chemotherapy alone in patients with advanced EGFR-positive non-small-cell lung cancer. Methods: In a multinational, multicentre, open-label, phase III trial, chemotherapy-naive patients (≥18 years) with advanced EGFR-expressing histologically or cytologically proven stage wet IIIB or stage IV non-small-cell lung cancer were randomly assigned in a 1:1 ratio to chemotherapy plus cetuximab or just chemotherapy. Chemotherapy was cisplatin 80 mg/m 2 intravenous infusion on day 1, and vinorelbine 25 mg/m 2 intravenous infusion on days 1 and 8 of every 3-week cycle) for up to six cycles. Cetuximab-at a starting dose of 400 mg/m 2 intravenous infusion over 2 h on day 1, and from day 8 onwards at 250 mg/m 2 over 1 h per week-was continued after the end of chemotherapy until disease progression or unacceptable toxicity had occurred. The primary endpoint was overall survival. Analysis was by intention to treat. This study is registered with ClinicalTrials.gov, number NCT00148798. Findings: Between October, 2004, and January, 2006, 1125 patients were randomly assigned to chemotherapy plus cetuximab (n=557) or chemotherapy alone (n=568). Patients given chemotherapy plus cetuximab survived longer than those in the chemotherapy-alone group (median 11·3 months vs 10·1 months; hazard ratio for death 0·871 [95% CI 0·762-0·996]; p=0·044). The main cetuximab-related adverse event was acne-like rash (57 [10%] of 548, grade 3). Interpretation: Addition of cetuximab to platinum-based chemotherapy represents a new treatment option for patients with advanced non-small-cell lung cancer. Funding: Merck KGaA. © 2009 Elsevier Ltd. All rights reserved.

Randomized phase II study of cetuximab plus cisplatin/vinorelbine compared with cisplatin/vinorelbine alone as first-line therapy in EGFR-expressing advanced non-small-cell lung cancer

Background: The Lung Cancer Cetuximab Study is an open-label, randomized phase II pilot study of cisplatin and vinorelbine combined with the epidermal growth factor receptor (EGFR)-targeted monoclonal antibody cetuximab versus cisplatin and vinorelbine alone, in patients with advanced EGFR-expressing, non-small-cell lung cancer (NSCLC). End points of the study are activity, safety and pharmacokinetics. Patients and methods: Following randomization, for a maximum of eight cycles, patients received three-weekly cycles of cisplatin (80 mg/m2, day 1) and vinorelbine (25 mg/m2 on days 1 and 8) alone or following cetuximab treatment (initial dose 400 mg/m, followed by 250 mg/m2 weekly thereafter). Results: Eighty-six patients were randomly allocated to the study (43 per arm). Confirmed response rates were 28% in the cisplatin/vinorelbine arm (A) and 35% in the cetuximab plus cisplatin/vinorelbine arm (B). Median progression-free survival (PFS) was 4.6 months in arm A and 5.0 months in arm B, with PFS rates at 12 months of 0% and 15%, respectively. Median survival was 7.3 months in arm A and 8.3 months in arm B. The 24-month survival rates were 0% and 16%, respectively. The cetuximab combination was well tolerated. Conclusion: In the first-line treatment of advanced NSCLC, the combination of cetuximab plus cisplatin/vinorelbine demonstrated an acceptable safety profile and the potential to improve activity over cisplatin/vinorelbine alone. © 2007 European Society for Medical Oncology.

Maternal exposure to heatwave and preterm birth in Brisbane, Australia

Objective To quantify the short-term effects of maternal exposure to heatwave on preterm birth. Design An ecological study. Setting: A population-based study in Brisbane, Australia. Population All pregnant women who had a spontaneous singleton live birth in Brisbane between November and March in 2000–2010 were studied. Methods Daily data on pregnancy outcomes, meteorological factors, and ambient air pollutants were obtained. The Cox proportional hazards regression model with time-dependent variables was used to examine the short-term impact of heatwave on preterm birth. A series of cut-off temperatures and durations were used to define heatwave. Multivariable analyses were also performed to adjust for socio-economic factors, demographic factors, meteorological factors, and ambient air pollutants. Main outcome measure Spontaneous preterm births. Results The adjusted hazard ratios (HRs) ranged from 1.13 (95% CI 1.03–1.24) to 2.00 (95% CI 1.37–2.91) by using different heatwave definitions, after controlling for demographic, socio-economic, and meteorological factors, and air pollutants. Conclusions Heatwave was significantly associated with preterm birth: the associations were robust to the definitions of heatwave. The threshold temperatures, instead of duration, could be more likely to influence the evaluation of birth-related heatwaves. The findings of this study may have significant public health implications as climate change progresses.

Exploring retinal markers of diabetic neuropathy

Purpose To investigate the application of retinal nerve fibre layer (RNFL) thickness as a marker for severity of diabetic peripheral neuropathy (DPN) in people with Type 2 diabetes. Methods This was a cross-sectional study whereby 61 participants (mean age 61 [41-75 years], mean duration of diabetes 14 [1-40 years], 70% male) with Type 2 diabetes and DPN underwent optical coherence tomography (OCT) scans. Global and 4 quadrant (TSNI) RNFL thicknesses were measured at 3.45mm around the optic nerve head of one eye. Neuropathy disability score (NDS) was used to assess the severity of DPN on a 0 to 10 scale. Participants were divided into three age-matched groups representing mild (NDS=3-5), moderate (NDS=6-8) and severe (NDS=9-10) neuropathy. Two regression models were fitted for statistical analysis: 1) NDS scores as co-variate for global and quadrant RNFL thicknesses, 2) NDS groups as a factor for global RNFL thickness only. Results Mean (SD) RNFL thickness (µm) was 103(9) for mild neuropathy (n=34), 101(10) for moderate neuropathy (n=16) and 95(13) in the group with severe neuropathy (n=11). Global RNFL thickness and NDS scores were statistically significantly related (b=-1.20, p=0.048). When neuropathy was assessed across groups, a trend of thinner mean RNFL thickness was observed with increasing severity of neuropathy; however, this result was not statistically significant (F=2.86, p=0.065). TSNI quadrant analysis showed that mean RNFL thickness reduction in the inferior quadrant was 2.55 µm per 1 unit increase in NDS score (p=0.005). However, the regression coefficients were not statistically significant for RNFL thickness in the superior (b=-1.0, p=0.271), temporal (b=-0.90, p=0.238) and nasal (b=-0.99, p=0.205) quadrants. Conclusions RNFL thickness was reduced with increasing severity of DPN and the effect was most evident in the inferior quadrant. Measuring RNFL thickness using OCT may prove to be a useful, non-invasive technique for identifying severity of DPN and may also provide additional insight into common mechanisms for peripheral neuropathy and RNFL damage.

Simulation-based fully Bayesian experimental design for mixed effects models

In this paper, we present fully Bayesian experimental designs for nonlinear mixed effects models, in which we develop simulation-based optimal design methods to search over both continuous and discrete design spaces. Although Bayesian inference has commonly been performed on nonlinear mixed effects models, there is a lack of research into performing Bayesian optimal design for nonlinear mixed effects models that require searches to be performed over several design variables. This is likely due to the fact that it is much more computationally intensive to perform optimal experimental design for nonlinear mixed effects models than it is to perform inference in the Bayesian framework. In this paper, the design problem is to determine the optimal number of subjects and samples per subject, as well as the (near) optimal urine sampling times for a population pharmacokinetic study in horses, so that the population pharmacokinetic parameters can be precisely estimated, subject to cost constraints. The optimal sampling strategies, in terms of the number of subjects and the number of samples per subject, were found to be substantially different between the examples considered in this work, which highlights the fact that the designs are rather problem-dependent and require optimisation using the methods presented in this paper.

Integrating the impact of climate into event based failure models for water pipes

This paper presents an event-based failure model to predict the number of failures that occur in water distribution assets. Often, such models have been based on analysis of historical failure data combined with pipe characteristics and environmental conditions. In this paper weather data have been added to the model to take into account the commonly observed seasonal variation of the failure rate. The theoretical basis of existing logistic regression models is briefly described in this paper, along with the refinements made to the model for inclusion of seasonal variation of weather. The performance of these refinements is tested using data from two Australian water authorities.

Exploring Retinal Markers of Diabetic Neuropathy

Purpose : To investigate the application of retinal nerve fibre layer (RNFL) thickness as a marker for severity of diabetic peripheral neuropathy (DPN) in people with Type 2 diabetes. Methods : This was a cross-sectional study whereby 61 participants (mean age 61 [41-75 years], mean duration of diabetes 14 [1-40 years], 70% male) with Type 2 diabetes and DPN underwent optical coherence tomography (OCT) scans. Global and 4 quadrant (TSNI) RNFL thicknesses were measured at 3.45mm around the optic nerve head of one eye. Neuropathy disability score (NDS) was used to assess the severity of DPN on a 0 to 10 scale. Participants were divided into three age-matched groups representing mild (NDS=3-5), moderate (NDS=6-8) and severe (NDS=9-10) neuropathy. Two regression models were fitted for statistical analysis: 1) NDS scores as co-variate for global and quadrant RNFL thicknesses, 2) NDS groups as a factor for global RNFL thickness only. Results : Mean (SD) RNFL thickness (µm) was 103(9) for mild neuropathy (n=34), 101(10) for moderate neuropathy (n=16) and 95(13) in the group with severe neuropathy (n=11). Global RNFL thickness and NDS scores were statistically significantly related (b=-1.20, p=0.048). When neuropathy was assessed across groups, a trend of thinner mean RNFL thickness was observed with increasing severity of neuropathy; however, this result was not statistically significant (F=2.86, p=0.065). TSNI quadrant analysis showed that mean RNFL thickness reduction in the inferior quadrant was 2.55 µm per 1 unit increase in NDS score (p=0.005). However, the regression coefficients were not statistically significant for RNFL thickness in the superior (b=-1.0, p=0.271), temporal (b=-0.90, p=0.238) and nasal (b=-0.99, p=0.205) quadrants. Conclusions : RNFL thickness was reduced with increasing severity of DPN and the effect was most evident in the inferior quadrant. Measuring RNFL thickness using OCT may prove to be a useful, non-invasive technique for identifying severity of DPN and may also provide additional insight into common mechanisms for peripheral neuropathy and RNFL damage.

Bridging experiments, models and simulations : an integrative approach to validation in computational cardiac electrophysiology

Computational models in physiology often integrate functional and structural information from a large range of spatio-temporal scales from the ionic to the whole organ level. Their sophistication raises both expectations and scepticism concerning how computational methods can improve our understanding of living organisms and also how they can reduce, replace and refine animal experiments. A fundamental requirement to fulfil these expectations and achieve the full potential of computational physiology is a clear understanding of what models represent and how they can be validated. The present study aims at informing strategies for validation by elucidating the complex interrelations between experiments, models and simulations in cardiac electrophysiology. We describe the processes, data and knowledge involved in the construction of whole ventricular multiscale models of cardiac electrophysiology. Our analysis reveals that models, simulations, and experiments are intertwined, in an assemblage that is a system itself, namely the model-simulation-experiment (MSE) system. Validation must therefore take into account the complex interplay between models, simulations and experiments. Key points for developing strategies for validation are: 1) understanding sources of bio-variability is crucial to the comparison between simulation and experimental results; 2) robustness of techniques and tools is a pre-requisite to conducting physiological investigations using the MSE system; 3) definition and adoption of standards facilitates interoperability of experiments, models and simulations; 4) physiological validation must be understood as an iterative process that defines the specific aspects of electrophysiology the MSE system targets, and is driven by advancements in experimental and computational methods and the combination of both.