Computing probable maximum loss in catastrophe reinsurance portfolios on multi-core and many-core architectures

In the reinsurance market, the risks natural catastrophes pose to portfolios of properties must be quantified, so that they can be priced, and insurance offered. The analysis of such risks at a portfolio level requires a simulation of up to 800 000 trials with an average of 1000 catastrophic events per trial. This is sufficient to capture risk for a global multi-peril reinsurance portfolio covering a range of perils including earthquake, hurricane, tornado, hail, severe thunderstorm, wind storm, storm surge and riverine flooding, and wildfire. Such simulations are both computation and data intensive, making the application of high-performance computing techniques desirable.

In this paper, we explore the design and implementation of portfolio risk analysis on both multi-core and many-core computing platforms. Given a portfolio of property catastrophe insurance treaties, key risk measures, such as probable maximum loss, are computed by taking both primary and secondary uncertainties into account. Primary uncertainty is associated with whether or not an event occurs in a simulated year, while secondary uncertainty captures the uncertainty in the level of loss due to the use of simplified physical models and limitations in the available data. A combination of fast lookup structures, multi-threading and careful hand tuning of numerical operations is required to achieve good performance. Experimental results are reported for multi-core processors and systems using NVIDIA graphics processing unit and Intel Phi many-core accelerators.

Risk adjustment for hospital use using social security data: cross sectional small area analysis

Objectives: To identify demographic and socioeconomic determinants of need for acute hospital treatment at small area level. To establish whether there is a relation between poverty and use of inpatient services. To devise a risk adjustment formula for distributing public funds for hospital services using, as far as possible, variables that can be updated between censuses. Design: Cross sectional analysis. Spatial interactive modelling was used to quantify the proximity of the population to health service facilities. Two stage weighted least squares regression was used to model use against supply of hospital and community services and a wide range of potential needs drivers including health, socioeconomic census variables, uptake of income support and family credit, and religious denomination. Setting: Northern Ireland. Main outcome measure: Intensity of use of inpatient services. Results: After endogeneity of supply and use was taken into account, a statistical model was produced that predicted use based on five variables: income support, family credit, elderly people living alone, all ages standardised mortality ratio, and low birth weight. The main effect of the formula produced is to move resources from urban to rural areas. Conclusions: This work has produced a population risk adjustment formula for acute hospital treatment in which four of the five variables can be updated annually rather than relying on census derived data. Inclusion of the social security data makes a substantial difference to the model and to the results produced by the formula.

Association analysis of apolipoprotein E genotype and risk of depressive symptoms in Alzheimer's disease

OBJECTIVES: Behavioural and psychological symptoms of dementia (BPSD) are potent predictors of carer distress and admission to institutional care. In Alzheimer's disease (AD), depressive symptoms are one of the most common complaints affecting around 50% of all patients. There is speculation these symptoms result from known genetic risk factors for AD, therefore we investigated the role of apolipoprotein E epsilon4 in the aetiology of depression in AD. METHODS: In this well-characterised cohort (n = 404) from the relatively genetically homogeneous Northern Ireland population, we tested the hypothesis that genetic variants of apolipoprotein E influence the risk for depressive symptoms in AD patients using the Neuropsychiatric Inventory (NPI-D) to determine the presence of depressive symptoms during the dementing illness. RESULTS: A total of 55% of patients exhibited a history of depression/dysphoria during the course of the illness as gathered by the NPI-D questionnaire. Forty-six percent were suffering from depression/dysphoria when the analysis was restricted to the month prior to interview. No statistically significant association between genotypes or alleles of apolipoprotein E and depression/dysphoria in AD was observed, nor was any association noted between the presence of severe symptoms and genotypes/alleles of apolipoprotein E. CONCLUSIONS: These results suggest apolipoprotein E genotype creates no additional risk for depressive symptoms in AD.

Dietary glycaemic index, glycaemic load and endometrial and ovarian cancer risk: a systematic review and meta-analysis.

Long-term consumption of a high glycaemic index (GI) or glycaemic load (GL) diet may lead to chronic hyperinsulinaemia, which is a potential risk factor for cancer. To date, many studies have examined the association between GI, GL and cancer risk, although results have been inconsistent, therefore our objective was to conduct a systematic review of the literature. Medline and Embase were systematically searched using terms for GI, GL and cancer to identify studies published before December 2007. Random effects meta-analyses were performed for endometrial cancer, combining maximally adjusted results that compared risk for those in the highest versus the lowest category of intake. Separate analysis examined risk by body mass index categories. Five studies examining GI and/or GL intake and endometrial cancer risk were identified. Pooled effect estimates for endometrial cancer showed an increased risk for high GL consumers (RR 1.20; 95% CI: 1.06-1.37), further elevated in obese women (RR 1.54; 95% CI: 1.18-2.03). No significant associations were observed for GI. Only two studies examined ovarian cancer and therefore no meta-analysis was performed, but results indicate positive associations for GL also. A high GL, but not a high GI, diet is positively associated with the risk of endometrial cancer, particularly among obese women. © 2008 Cancer Research UK All rights reserved.

Dietary glycaemic index, glycaemic load and breast cancer risk: a systematic review and meta-analysis

This systematic review aimed to examine if an association exists between dietary glycaemic index (GI) and glycaemic load (GL) intake and breast cancer risk. A systematic search was conducted in Medline and Embase and identified 14 relevant studies up to May 2008. Adjusted relative risk estimates comparing breast cancer risk for the highest versus the lowest category of GI/GL intake were extracted from relevant studies and combined in meta-analyses using a random-effects model. Combined estimates from six cohort studies show non-significant increased breast cancer risks for premenopausal women (relative risk (RR) 1.14, 95% CI 0.95-1.38) and postmenopausal women (RR 1.11, 95% CI 0.99-1.25) consuming the highest versus the lowest category of GI intake. Evidence of heterogeneity hindered analyses of GL and premenopausal risk, although most studies did not observe any significant association. Pooled cohort study results indicated no association between postmenopausal risk and GL intake (RR 1.03, 95% CI 0.94-1.12). Our findings do not provide strong support of an association between dietary GI and GL and breast cancer risk. © 2008 Cancer Research UK.


--------------------------------------------------------------------------------

Reaxys Database Information|

Glycemic index, glycemic load, and risk of digestive tract neoplasms: a systematic review and meta-analysis

Background: Habitual consumption of diets with a high glycemic index (GI) and a high glycemic load (GL) may influence cancer risk via hyperinsulinemia and the insulin-like growth factor axis.
Objective: The objective was to conduct a systematic review to assess the association between GI, GL, and risk of digestive tract cancers.
Design: Medline and Embase were searched for relevant publications from inception to July 2008. When possible, adjusted results from a comparison of cancer risk of the highest compared with the lowest category of GI and GL intake were combined by using random-effects meta-analyses.
Results: Cohort and case-control studies that examined the risk between GI or GL intake and colorectal cancer (n = 12) and adenomas (n = 2), pancreatic cancer (n = 6), gastric cancer (n = 2), and squamous-cell esophageal carcinoma (n = 1) were retrieved. Most case-control studies observed positive associations between GI and GL intake and these cancers. However, pooled cohort study results showed no associations between colorectal cancer risk and GI intake [relative risk (RR): 1.04; 95% CI: 0.92, 1.12; n = 7 studies] or GL intake (RR: 1.06; 95% CI: 0.95, 1.17; n = 8 studies). Furthermore, no significant associations were observed in meta-analyses of cohort study results of colorectal cancer subsites and GI and GL intake. Similarly, no significant associations emerged between pancreatic cancer risk and GI intake (RR: 0.99; 95% CI: 0.83, 1.19; n = 5 studies) or GL intake (RR: 1.01; 95% CI: 0.86, 1.19; n = 6 studies) in combined cohort studies.
Conclusions: The findings from our meta-analyses indicate that GI and GL intakes are not associated with risk of colorectal or pancreatic cancers. There were insufficient data available regarding other digestive tract cancers to make any conclusions about GI or GL intake and risk.

Is hyperfiltration associated with the future risk of developing diabetic nephropathy? A meta-analysis.

Aims/hypothesis Glomerular hyperfiltration is a well established phenomenon occurring early in some patients with type 1 diabetes. However, there is no consistent answer regarding whether hyperfiltration predicts later development of nephropathy. We performed a systematic review and meta-analysis of observational studies that compared the risk of developing diabetic nephropathy in patients with and without glomerular hyperfiltration and also explored the impact of baseline GFR.

Methods A systematic review and meta-analysis was carried out. Cohort studies in type 1 diabetic participants were included if they contained data on the development of incipient or overt nephropathy with baseline measurement
of GFR and presence or absence of hyperfiltration.

Results We included ten cohort studies following 780 patients. After a study median follow-up of 11.2 years, 130 patients had developed nephropathy. Using a random effects model, the pooled odds of progression to a minimum
of microalbuminuria in patients with hyperfiltration was 2.71 (95% CI 1.20–6.11) times that of patients with normofiltration. There was moderate heterogeneity (heterogeneity test p=0.05, measure of degree of inconsistency=48%) and some evidence of funnel plot asymmetry, possibly due to publication bias. The pooled weighted mean difference in baseline GFR was 13.8 ml min-1 1.73 m-2 (95% CI 5.0–22.7) greater in the group progressing to nephropathy than in those not progressing (heterogeneity test p<0.01).

Conclusions/interpretation In published studies, individuals with glomerular hyperfiltration were at increased risk of progression to diabetic nephropathy using study level data. Further larger studies are required to explore this relationship and the role of potential confounding variables.

Homocysteine, Methylenetetrahydrofolate Reductase C677T Polymorphism, and Risk of Retnal Vein Occlusion: A Meta-analysis

Objective: To assess the role of plasma total homocysteine (tHcy) concentrations and homozygosity for the thermolabile variant of the methylenetetrahydrofolate reductase (MTHFR) C677T gene as risk factors for retinal vascular occlusive disease.

Design: Retinal vein occlusion (RVO) is an important cause of vision loss. Early meta-analyses showed that tHcy was associated with an increased risk of RVO, but a significant number of new studies have been published. Participants and/or Controls: RVO patients and controls.

Methods: Data sources included MEDLINE, Web of Science, and PubMed searches and searching reference lists of relevant articles and reviews. Reviewers searched the databases, selected the studies, and then extracted data. Results were pooled quantitatively using meta-analytic methods.

Main Outcome Measures: tHcy concentrations and MTHFR genotype.

Results: There were 25 case-control studies for tHcy (1533 cases and 1708 controls) and 18 case-control studies for MTHFR (1082 cases and 4706 controls). The mean tHcy was on average 2.8 mol/L (95% confidence
interval [CI], 1.8 –3.7) greater in the RVO cases compared with controls, but there was evidence of between-study heterogeneity (P0.001, I2 93%). There was funnel plot asymmetry suggesting publication bias. There was no evidence of association between homozygosity for the MTHFR C677T genotype and RVO (odds ratio [OR] 1.20; 95% CI, 0.84–1.71), but again marked heterogeneity (P 0.004, I2 53%) was observed.

Conclusions: There was some evidence that elevated tHcy was associated with RVO, but not homozygosity for the MTHFR C677T genotype. Both analyses should be interpreted cautiously because of marked heterogeneity between the study estimates and possible effect of publication bias on the tHcy findings.

Financial Disclosure(s): The author(s) have no proprietary or commercial interest in any materials discussed in this article.

Birthweight and the risk of childhood-onset type 1 diabetes: a meta-analysis of observational studies using individual patient data

Aims/hypothesis: We investigated whether children who are heavier at birth have an increased risk of type 1 diabetes. Methods: Relevant studies published before February 2009 were identified from literature searches using MEDLINE, Web of Science and EMBASE. Authors of all studies containing relevant data were contacted and asked to provide individual patient data or conduct pre-specified analyses. Risk estimates of type 1 diabetes by category of birthweight were calculated for each study, before and after adjustment for potential confounders. Meta-analysis techniques were then used to derive combined ORs and investigate heterogeneity between studies. Results: Data were available for 29 predominantly European studies (five cohort, 24 case-control studies), including 12,807 cases of type 1 diabetes. Overall, studies consistently demonstrated that children with birthweight from 3.5 to 4 kg had an increased risk of diabetes of 6% (OR 1.06 [95% CI 1.01-1.11]; p=0.02) and children with birthweight over 4 kg had an increased risk of 10% (OR 1.10 [95% CI 1.04-1.19]; p=0.003), compared with children weighing 3.0 to 3.5 kg at birth. This corresponded to a linear increase in diabetes risk of 3% per 500 g increase in birthweight (OR 1.03 [95% CI 1.00-1.06]; p=0.03). Adjustments for potential confounders such as gestational age, maternal age, birth order, Caesarean section, breastfeeding and maternal diabetes had little effect on these findings. Conclusions/interpretation: Children who are heavier at birth have a significant and consistent, but relatively small increase in risk of type 1 diabetes. © 2010 Springer-Verlag.


--------------------------------------------------------------------------------

Reaxys Database Information|

--------------------------------------------------------------------------------

Dietary patterns and breast cancer risk: a systematic review and meta-analysis

Background: Dietary patterns, which represent whole-diet and possible food and nutrient interactions, have been linked to the risk of various cancers. However, the associations of these dietary patterns with breast cancer remain unclear. Objective: We critically appraised the literature and conducted meta-analyses to pool the results of studies to clarify the relation between dietary patterns and breast cancer risk.
Design: MEDLINE and EMBASE were searched for relevant articles that identified common dietary patterns published up to November 2009. Multivariable-adjusted odds ratios (ORs) comparing highest and lowest categories of dietary pattern scores and multi-variable-adjusted ORs for a 20th-percentile increase in dietary pattern scores were combined by using random-effects meta-analyses. Results: Case-control and cohort studies were retrieved that identified prudent/healthy (n = 18), Western/unhealthy (n = 17), and drinker (n = 4) dietary patterns. There was evidence of a decrease in the risk of breast cancer in the highest compared with the lowest categories of prudent/healthy dietary patterns (OR = 0.89; 95% CI: 0.82, 0.99; P = 0.02) in all studies and in pooled cohort studies alone. An increase in the risk of breast cancer was shown for the highest compared with the lowest categories of a drinker dietary pattern (OR = 1.21; 95% CI: 1.04, 1.41; P = 0.01). There was no evidence of a difference in the risk of breast cancer between the highest and the lowest categories of Western/unhealthy dietary patterns (OR = 1.09; 95% CI: 0.98, 1.22; P = 0.12). Conclusion: The results of this systematic review and meta-analysis indicate that some dietary patterns may be associated with breast cancer risk.

Can physical activity modulate pancreatic cancer risk?:a systematic review and meta-analysis

Numerous epidemiological studies have examined the association between physical activity and pancreatic cancer; however, findings from individual cohorts have largely not corroborated a protective effect. Among other plausible mechanisms, physical activity may reduce abdominal fat depots inducing metabolic improvements in glucose tolerance and insulin sensitivity, thereby potentially attenuating pancreatic cancer risk. We performed a systematic review to examine associations between physical activity and pancreatic cancer. Six electronic databases were searched from their inception through July 2009, including MEDLINE and EMBASE, seeking observational studies examining any physical activity measure with pancreatic cancer incidence/mortality as an outcome. A random effects model was used to pool individual effect estimates evaluating highest vs. lowest categories of activity. Twenty-eight studies were included. Pooled estimates indicated a reduction in pancreatic cancer risk with higher levels of total (five prospective studies, RR: 0.72, 95% CI: 0.52-0.99) and occupational activity (four prospective studies, RR: 0.75, 95% CI: 0.59-0.96). Nonsignificant inverse associations were seen between risks and recreational and transport physical activity. When examining exercise intensity, moderate activity appeared more protective (RR: 0.79, 95% CI: 0.52-1.20) than vigorous activity (RR: 0.97, 95% CI: 0.85-1.11), but results were not statistically significant and the former activity variable incorporated marked heterogeneity. Despite indications of an inverse relationship with higher levels of work and total activity, there was little evidence of such associations with recreational and other activity exposures.