SAS macros for point and interval estimation of area under the receiver operating characteristic curve for non-proportional and proportional hazards Weibull models

Aims and objectives  For prediction of risk of cardiovascular end points using survival models the proportional hazards assumption is often not met. Thus, non-proportional hazards models are more appropriate for developing risk prediction equations in such situations. However, computer program for evaluating the prediction performance of such models has been rarely addressed. We therefore developed SAS macro programs for evaluating the discriminative ability of a non-proportional hazards Weibull model developed by Anderson (1991) and that of a proportional hazards Weibull model using the area under receiver operating characteristic (ROC) curve.

Method  Two SAS macro programs for non-proportional hazards Weibull model using Proc NLIN and Proc NLP respectively and model validation using area under ROC curve (with its confidence limits) were written with SAS IML language. A similar SAS macro for proportional hazards Weibull model was also written.

Results  The computer program was applied to data on coronary heart disease incidence for a Framingham population cohort. The five risk factors considered were current smoking, age, blood pressure, cholesterol and obesity. The predictive ability of the non-proportional hazard Weibull model was slightly higher than that of its proportional hazard counterpart. An advantage of SAS Proc NLP in terms of the example provided here is that it provides significance level for the parameter estimates whereas Proc NLIN does not.

Conclusion  The program is very useful for evaluating the predictive performance of non-proportional and proportional hazards Weibull models.

The impact of sleep restriction while performing simulated physical firefighting work on cortisol and heart rate responses

PURPOSE: Physical work and sleep restriction are two stressors faced by firefighters, yet the combined impact these demands have on firefighters' acute stress responses is poorly understood. The purpose of the present study was to assess the effect firefighting work and sleep restriction have on firefighters' acute cortisol and heart rate (HR) responses during a simulated 3-day and 2-night fire-ground deployment. METHODS: Firefighters completed multiple days of simulated physical work separated by either an 8-h (control condition; n = 18) or 4-h sleep opportunity (sleep restriction condition; n = 17). Salivary cortisol was sampled every 2 h, and HR was measured continuously each day. RESULTS: On day 2 and day 3 of the deployment, the sleep restriction condition exhibited a significantly higher daily area under the curve cortisol level and an elevated cortisol profile in the afternoon and evening when compared with the control condition. Firefighters' HR decreased across the simulation, but there were no significant differences found between conditions. CONCLUSION: Findings highlight the protective role an 8-h sleep opportunity between shifts of firefighting work has on preserving normal cortisol levels when compared to a 4-h sleep opportunity which resulted in elevated afternoon and evening cortisol. Given the adverse health outcomes associated with chronically high cortisol, especially later in the day, future research should examine how prolonged exposure to firefighting work (including restricted sleep) affects firefighters' cortisol levels long term. Furthermore, monitoring cortisol levels post-deployment will determine the minimum recovery time firefighters need to safely return to the fire-ground.

Model to predict inpatient mortality from information gathered at presentation to an emergency department: The Triage Information Mortality Model (TIMM)

OBJECTIVES: To derive and validate a mortality prediction model from information available at ED triage. METHODS: Multivariable logistic regression of variables from administrative datasets to predict inpatient mortality of patients admitted through an ED. Accuracy of the model was assessed using the receiver operating characteristic area under the curve (ROC-AUC) and calibration using the Hosmer-Lemeshow goodness of fit test. The model was derived, internally validated and externally validated. Derivation and internal validation were in a tertiary referral hospital and external validation was in an urban community hospital. RESULTS: The ROC-AUC for the derivation set was 0.859 (95% CI 0.856-0.865), for the internal validation set was 0.848 (95% CI 0.840-0.856) and for the external validation set was 0.837 (95% CI 0.823-0.851). Calibration assessed by the Hosmer-Lemeshow goodness of fit test was good. CONCLUSIONS: The model successfully predicts inpatient mortality from information available at the point of triage in the ED.

Hormonal and metabolic responses to repeated cycling sprints under different hypoxic conditions

OBJECTIVE: Sprint exercise and hypoxic stimulus during exercise are potent factors affecting hormonal and metabolic responses. However, the effects of different hypoxic levels on hormonal and metabolic responses during sprint exercise are not known. Here, we examined the effect of different hypoxic conditions on hormonal and metabolic responses during sprint exercise. DESIGN: Seven male subjects participated in three experimental trials: 1) sprint exercise under normoxia (NSE); 2) sprint exercise under moderate normobaric hypoxia (16.4% oxygen) (HSE 16.4); and 3) sprint exercise under severe normobaric hypoxia (13.6% oxygen) (HSE 13.6). The sprint exercise consisted of four 30s all-out cycling bouts with 4-min rest between bouts. Glucose, free fatty acids (FFA), blood lactate, growth hormone (GH), epinephrine (E), norepinephrine (NE), and insulin concentrations in the HSE trials were measured before exposure to hypoxia (pre 1), 15 min after exposure to hypoxia (pre 2), and at 0, 15, 30, 60, 120, and 180 min after the exercise performed in hypoxia. The blood samples in the NSE trial were obtained in normoxia at the same time points as the HSE trials. RESULTS: Circulating levels of glucose, FFA, lactate, GH, E, NE, and insulin significantly increased after all three exercise trials (P < 0.05). The area under the curve (AUC) for GH was significantly higher in the HSE 13.6 trial than in the NSE and HSE 16.4 trials (P < 0.05). A maximal increase in FFA concentration was observed at 180 min after exercise and was not different between trials. CONCLUSION: These findings suggest that severe hypoxia may be an important factor for the enhancement of GH response to all-out sprint exercise.

Identifying the location of fire refuges in wet forest ecosystems

The increasing frequency of large, high-severity fires threatens the survival of old-growth specialist fauna in fire-prone forests. Within topographically diverse montane forests, areas that experience less severe or fewer fires compared with those prevailing in the landscape may present unique resource opportunities enabling old-growth specialist fauna to survive. Statistical landscape models that identify the extent and distribution of potential fire refuges may assist land managers to incorporate these areas into relevant biodiversity conservation strategies. We used a case study in an Australian wet montane forest to establish how predictive fire simulation models can be interpreted as management tools to identify potential fire refuges. We examined the relationship between the probability of fire refuge occurrence as predicted by an existing fire refuge model and fire severity experienced during a large wildfire. We also examined the extent to which local fire severity was influenced by fire severity in the surrounding landscape. We used a combination of statistical approaches, including generalized linear modeling, variogram analysis, and receiver operating characteristics and area under the curve analysis (ROC AUC). We found that the amount of unburned habitat and the factors influencing the retention and location of fire refuges varied with fire conditions. Under extreme fire conditions, the distribution of fire refuges was limited to only extremely sheltered, fire-resistant regions of the landscape. During extreme fire conditions, fire severity patterns were largely determined by stochastic factors that could not be predicted by the model. When fire conditions were moderate, physical landscape properties appeared to mediate fire severity distribution. Our study demonstrates that land managers can employ predictive landscape fire models to identify the broader climatic and spatial domain within which fire refuges are likely to be present. It is essential that within these envelopes, forest is protected from logging, roads, and other developments so that the ecological processes related to the establishment and subsequent use of fire refuges are maintained.

Spatial prediction of rufous bristlebird habitat in a coastal heathland: a GIS-based approach

1. To develop a conservation management plan for a species, knowledge of its distribution and spatial arrangement of preferred habitat is essential. This is a difficult task, especially when the species of concern is in low   abundance. In south-western Victoria, Australia, populations of the rare rufous bristlebird Dasyornis broadbenti are threatened by fragmentation of suitable habitat. In order to improve the conservation status of this species, critical habitat requirements must be identified and a system of corridors must be established to link known populations. A predictive spatial model of rufous bristlebird habitat was developed in order to identify critical areas requiring preservation, such as corridors for dispersal.
2. Habitat models generated using generalized linear modelling techniques can assist in delineating the specific habitat requirements of a species. Coupled with geographic information system (GIS) technology, these models can be extrapolated to produce maps displaying the spatial configuration of suitable habitat.
3. Models were generated using logistic regression, with bristlebird presence or absence as the dependent variable and landscape variables, extracted from both GIS data layers and multispectral digital imagery, as the predictors. A multimodel inference approach based on Akaike’s information criterion was used and the resulting model was applied in a GIS to extrapolate predicted likelihood of occurrence across the entire area of concern. The predictive performance of the selected model was evaluated using the receiver operating characteristic (ROC) technique. A hierarchical partitioning protocol was used to identify the predictor variables most likely to influence variation in the dependent variable. Probability of species presence was used as an index of habitat suitability.
4. Negative associations between rufous bristlebird presence and  increasing elevation, 'distance to cree', 'distance to coast' and sun index were evident, suggesting a preference for areas relatively low in altitude, in close proximity to the coastal fringe and drainage lines, and receiving less direct sunlight. A positive association with increasing habitat complexity also suggested that this species prefers areas containing high vertical density of vegetation.
5. The predictive performance of the selected model was shown to be high (area under the curve 0·97), indicating a good fit of the model to the data. Hierarchical partitioning analysis showed that all the variables considered had significant  independent contributions towards explaining the variation in the dependent variable. The proportion of the total study area that was predicted as suitable habitat for the rufous bristlebird (using probability of occurrence at a ≥0·5 level ) was 16%.
6. Synthesis and applications. The spatial model clearly delineated areas predicted as highly suitable rufous bristlebird habitat, with evidence of potential corridors linking coastal and inland populations via gullies. Conservation of this species will depend on management actions that protect the critical habitats identified in the model. A multi-scale  approach to the modelling process is recommended whereby a spatially explicit model is first generated using landscape variables extracted from a GIS, and a second model at site level is developed using fine-scale habitat variables measured on the ground. Where there are constraints on the time and cost involved in measuring finer scale variables, the first step alone can be used for conservation planning.

Landscape characteristics associated with species richness and occurrence of small native mammals inhabiting a coastal heathland: a spatial modelling approach

In the coastal region of south-western Victoria, Australia, populations of native small mammal species are restricted to patches of suitable habitat in a highly fragmented landscape. The size and spatial arrangement of these patches is likely to influence both the occupancy and richness of species at a location. Geographic Information System (GIS)-based habitat models of the species richness of native small mammals, and individual species  occurrences, were developed to produce maps displaying the spatial  configuration of suitable habitat. Models were generated using either generalised linear Poisson regression (for species richness) or logistic regression (for species occurrences) with species richness or  presence/absence as the dependent variable and landscape variables, extracted from both GIS data layers and multi-spectral digital imagery, as the predictor variables. A multi-model inference approach based on the Akaike Information Criterion was used and the resulting model was applied in a GIS framework to extrapolate predicted richness/likelihood of occurrence across the entire area of the study. A negative association between species  richness and elevation, habitat complexity and sun index indicated that richness within the study area decreases with increasing altitude, vertical vegetation structure and exposure to solar radiation. Landform  characteristics were important (to varying degrees) in determining habitat occupancy for all of the species examined, while the influence of habitat complexity was important for only one of the species. Performance of all but one of the models generated using presence/absence data was high, as indicated by the area under the curve of a receiver-operating characteristic plot. The effective conservation of the small mammal species in the area of concern is likely to depend on management actions that promote the protection of the critical habitats identified in the models.

Palatability and glucose, insulin and satiety responses of chickpea flour and extruded chickpea flour bread eaten as part of a breakfast

Objective: To determine the effect of adding chickpea flour or extruded chickpea flour to white bread on palatability and postprandial glycaemia, insulinaemia and satiety.

Design: A randomised, single-blind, cross-over study of four 50 g available carbohydrate breakfasts.

Setting: School of Exercise and Nutrition Sciences, Deakin University.

Subjects: In all, 12 healthy subjects were recruited through posted notices. Totally, 11 (nine male, two female) completed the study (meanplusminuss.e.m.; age 32±2 y; body mass index, 24.7±0.8 kg/m²).

Intervention: After overnight fasting, subjects consumed a control (white) bread (WB) breakfast twice, a chickpea bread (CHB) breakfast once and an extruded chickpea bread (EXB) breakfast once. Palatability and postprandial blood glucose, insulin and satiety responses were determined. Following this, food intakes from an ad libitum buffet and for the remainder of the day were assessed.

Results: A trend towards a lower incremental area under the curve (IAUC) of glucose for the CHB breakfast compared to the WB breakfast was observed (P=0.087). The IAUC of insulin and insulinaemic index (II) of the CHB breakfast were higher (P<0.05) than for the WB breakfast. No differences in glycaemic index (GI), satiety response, food intake or palatability were observed.

Conclusions: CHB and EXB demonstrated acceptable palatability. CHB demonstrated some hypoglycaemic effect compared to WB, but neither CHB nor EXB demonstrated effects on satiety or food intake. The hyperinsulinaemic effect of CHB observed in this study requires further investigation.

Assessing risk for sexual offenders in New Zealand : development and validation of a computer-scored risk measure

A growing research base supports the predictive validity of actuarial methods of risk assessment with sexual offenders. These methods use clearly defined variables with demonstrated empirical association with re-offending. The advantages of actuarial measures for screening large numbers of offenders quickly and economically are further enhanced when the variables used can be extracted from existing electronic databases. This study reports the results of applying a computerized set of historical variables with a sample of 1,133 male sexual offenders released from prison by the New Zealand Department of Corrections. Area under the curve figures of 0.70-0.78 were obtained over periods of 5 to 15 years, reflecting a significant level of association with sexual recidivism. Detected rates of re-offending across risk levels were comparable to those previously reported for the Static-99. Rates of sexual re-offending by child molesters for all sexual offences and offences against child victims are reported separately.

VVP808 is a novel agent that improves glucose tolerance in DIO mice

As the prevalence of diabetes mellitus continues to increase, there is an urgent need to discover new, effective treatment strategies to combat this disorder. In this study, we tested a novel agent, VVP808, which we previously demonstrated has insulin-sensitising properties (as measured by an increase in insulin-stimulated glucose uptake in 3T3-L1 adipocytes). A dose-ranging study was performed (10-100mg/kg/d) in C57BL/6J mice that had been fed a high-fat diet (45% of energy) for 12 weeks. VVP808 was administered by single daily oral gavage for a period of 16 days. Body weight, food intake and water intake were measured daily, whilst fasting blood glucose and plasma insulin levels were measured at the beginning and end of the study, with an intra-peritoneal glucose tolerance test (ipGTT) performed on day -1 and day 13. Administration of VVP808 to diet-induced obese (DIO) mice caused a strong dose-dependent improvement in glucose tolerance. There was a 34-42% reduction in the blood glucose area under the curve (AUC) at doses of 20mg/kg, 50mg/kg and 100mg/kg VVP808 (p=0.02-0.005). Administration of VVP808 resulted in a small but significant reduction in body weight in the 50mg/kg and 100mg/kg treated animals relative to vehicle (p=0.01 and 0.001 respectively). This decrease in body weight was associated with a reduction in food intake for the 100mg/kg treated animals only. Epididymal fat pad weight was significantly reduced in animals treated with 100mg/kg VVP808 (p=0.01). Furthermore, treatment with VVP808 for 16 days resulted in a highly significant dose-dependent reduction in fasting blood glucose levels relative to vehicle treated animals (p= 0.01-0.001). In conclusion, our data showed that VVP808 acts in a dose-dependent manner to reduce fasting blood glucose levels and improve glucose tolerance. These data suggest that VVP808 is an interesting new agent with potential for development as a novel therapeutic for type 2 diabetes.

Assessment of community reintegration planning for sex offenders

Poor planning for reintegrating child molesters from prison to the community is a likely risk factor for sexual recidivism. The quality of reintegration planning was retrospectively measured for groups of recidivist (n = 30) and nonrecidivist (n = 30) child molesters who were individually matched on static risk level and time since release. Recidivists had significantly poorer reintegration planning scores than nonrecidivists, consistent with a previous study by the authors. Data from both studies were combined (total N = 141), and survival analyses showed that poor reintegration planning predicted an increased rate of recidivism. Accommodation, employment, and social support planning combined to predict recidivism, with predictive validity comparable to static risk models (area under the curve = .71). Summing these items yielded a scale of reintegration planning quality that differentiated well between recidivists and nonrecidivists and may have practical utility for risk assessment as an adjunct to static models.

Physiological responses to psychological stress: importance of adiposity in men aged 50-70 years

We tested the hypothesis that overweight/obese men aged 50–70 years will have a greater salivary cortisol, salivary alpha amylase and heart rate (HR) responses to psychological stress compared with age matched lean men. Lean (BMIZ20–25 kg/m2; nZ19) and overweight/obese (BMIZ27–35 kg/m2; nZ17) men (50–70 years) were subjected to a well-characterised psychological stress (Trier Social Stress Test, TSST) at 1500 h. Concentrations of cortisol and alpha amylase were measured in saliva samples collected every 7–15 min from 1400 to 1700 h. HR was recorded using electrocardiogram. Body weight, BMI, percentage body fat, resting systolic and diastolic blood pressure and mean arterial pressure were significantly higher (P!0.05) in overweight/obese men compared with lean men. Both groups responded to the TSST with a substantial elevation in salivary cortisol (372%), salivary alpha amylase (123%) and HR (22%). These responses did not differ significantly between the groups (time!treatment interaction for salivary cortisol, salivary alpha amylase and HR; PZ0.187, PZ0.288, PZ0.550, respectively). There were no significant differences between the groups for pretreatment values, peak height, difference between pretreatment values and peak height (reactivity) or area under the curve for salivary cortisol, salivary alpha amylase or HR (PO0.05 for all). The results showed that, for men with a moderate level of overweight/obesity who were otherwise healthy, the response of salivary cortisol, salivary alpha amylase and HR to acute psychological stress was not impaired.

Endurance training in obese humans improves glucose tolerance and mitochondrial fatty acid oxidation and alters muscle lipid content

Muscle fatty acid (FA) metabolism is impaired in obesity and insulin resistance, reflected by reduced rates of FA oxidation and accumulation of lipids. It has been suggested that interventions that increase FA oxidation may enhance insulin action by reducing these lipid pools. Here, we examined the effect of endurance training on rates of mitochondrial FA oxidation, the activity of carnitine palmitoyltransferase I (CPT I), and the lipid content in muscle of obese individuals and related these to measures of glucose tolerance. Nine obese subjects completed 8 wk of moderate-intensity endurance training, and muscle biopsies were obtained before and after training. Training significantly improved glucose tolerance, with a reduction in the area under the curve for glucose (P< 0.05) and insulin (P = 0.01) during an oral glucose tolerance test. CPT I activity increased 250% (P = 0.001) with training and became less sensitive to inhibition by malonyl-CoA. This was associated with an increase in mitochondrial FA oxidation (+120%, P < 0.001). Training had no effect on muscle triacylglycerol content; however, there was a trend for training to reduce both the total diacylglcyerol (DAG) content (−15%, P = 0.06) and the saturated DAG-FA species (−27%, P = 0.06). Training reduced both total ceramide content (−42%, P = 0.01) and the saturated ceramide species (−32%, P < 0.05). These findings suggest that the improved capacity for mitochondrial FA uptake and oxidation leads not only to a reduction in muscle lipid content but also a to change in the saturation status of lipids, which may, at least in part, provide a mechanism for the enhanced insulin action observed with endurance training in obese individuals.

Detection of respiratory arousals using photoplethysmography (PPG) signal in sleep apnea patients

Respiratory events during sleep induce cortical arousals and manifest changes in autonomic markers in sleep disorder breathing (SDB). Finger photoplethysmography (PPG) has been shown to be a reliable method of determining sympathetic activation. We hypothesize that changes in PPG signals are sufficient to predict the occurrence of respiratory-event-related cortical arousal. In this study, we develop a respiratory arousal detection model in SDB subjects by using PPG features. PPG signals from 10 SDB subjects (9 male, 1 female) with age range 43-75 years were used in this study. Time domain features of PPG signals, such as 1) PWA--pulse wave amplitude, 2) PPI--peak-to-peak interval, and 3) Area--area under peak, were used to detect arousal events. In this study, PWA and Area have shown better performance (higher accuracy and lower false rate) compared to PPI features. After investigating possible groupings of these features, combination of PWA and Area (PWA + Area) was shown to provide better accuracy with a lower false detection rate in arousal detection. PPG-based arousal indexes agreed well across a wide range of decision thresholds, resulting in a receiver operating characteristic with an area under the curve of 0.91. For the decision threshold (PC(thresh) = 25%) chosen for the final analyses, a sensitivity of 68.1% and a specificity of 95.2% were obtained. The results showed an accuracy of 84.68%, 85.15%, 86.93%, and 50.79% with a false rate of 21.80%, 55.41%, 64.78%, and 50.79% at PC(thresh) = 25% or PPI, PWA, Area , and PWA + Area features, respectively. This indicates that combining PWA and Area features reduced the false positive rate without much affecting the sensitivity of the arousal detection system. In conclusion, the PPG-based respiratory arousal detection model is a simple and promising alternative to the conventional electroencephalogram (EEG)-based respiratory arousal detection system.

Instruments that prospectively predict bipolar disorder - a systematic review

BACKGROUND: Identification of earlier stages of Bipolar Disorder (BD), even prior to the first manic episode, may help develop interventions to prevent or delay the onset of BD. However, reliable and valid instruments are necessary to ascertain such earlier stages of BD. The aim of the current review was to identify instruments that had predictive validity and utility for BD for use in early intervention (EI) settings for the prevention of BD. METHODS: We undertook a systematic examination of studies that examined participants without BD I or II at baseline and prospectively explored the predictive abilities of instruments for BD onset over a period of 6 months or more. The instruments and the studies were rated with respect to their relative validity and utility predicting onset of BD for prevention or early intervention. Odds ratios and area under the curve (AUC) values were derived when not reported. RESULTS: Six studies were included, identifying five instruments that examined sub-threshold symptoms, family history, temperament and behavioral regulation. Though none of the identified instruments had been examined in high-quality replicated studies for predicting BD, two instruments, namely the Child Behavioral Checklist - Pediatric BD phenotype (CBCL-PBD) and the General Behavioral Inventory - Revised (GBI-R), had greater levels of validity and utility. LIMITATION: Non-inclusion of studies and instruments that incidentally identified BD on follow-up limited the breadth of the review. CONCLUSION: Instruments that test domains such as subthreshold symptoms, behavioral regulation, family history, and temperament hold promise in predicting BD onset.