The influence of disease risk on the optimal time interval between screens for the early detection of cancer: a mathematical approach.

The intervals between screens for the early detection of diseases such as breast and colon cancer suggested by screening guidelines are typically based on the average population risk of disease. With the emergence of ever more biomarkers for cancer risk prediction and the development of personalized medicine, there is a need for risk-specific screening intervals. The interval between successive screens should be shorter with increasing cancer risk. A risk-dependent optimal interval is ideally derived from a cost-effectiveness analysis using a validated simulation model. However, this is time-consuming and costly. We propose a simplified mathematical approach for the exploratory analysis of the implications of risk level on optimal screening interval. We develop a mathematical model of the optimal screening interval for breast cancer screening. We verified the results by programming the simplified model in the MISCAN-Breast microsimulation model and comparing the results. We validated the results by comparing them with the results of a full, published MISCAN-Breast cost-effectiveness model for a number of different risk levels. The results of both the verification and validation were satisfactory. We conclude that the mathematical approach can indicate the impact of disease risk on the optimal screening interval.

Accuracy of self-reported family history is strongly influenced by the accuracy of self-reported personal health status of relatives.

OBJECTIVE: We investigated the accuracy of self-reported family history for diabetes, hypertension, and overweight against two reference standards: family history based on physician-assessed health status of relatives and on self-reported personal health status of relatives. STUDY DESIGN AND SETTING: Subjects were participants from the Erasmus Rucphen Family study, an extended family study among descendants of 20 couples who lived between 1850 and 1900 in a southwest region of the Netherlands and their relatives (n=1,713). Sensitivity and specificity of self-reported family history were calculated. RESULTS: Sensitivity of self-reported family history was 89.2% for diabetes, 92.2% for hypertension, and 78.4% for overweight when family history based on relatives' self-reported personal health status was used as reference and 70.8% for diabetes, 67.4% for hypertension, and 77.3% for overweight when physician-assessed health status of relatives was used. Sensitivity and specificity of self-reported personal health status were 76.8% and 98.8% for diabetes, 38.9% and 98.0% for hypertension, and 80.9% and 75.7% for overweight, respectively. CONCLUSION: The accuracy of self-reported family history of diabetes and hypertension is strongly influenced by the accuracy of self-reported personal health status of relatives. Raising awareness of personal health status is crucial to ensure the utility of family history for the assessment of risk and disease prevention.

Predicting outcome after traumatic brain injury: development and international validation of prognostic scores based on admission characteristics.

BACKGROUND: Traumatic brain injury (TBI) is a leading cause of death and disability. A reliable prediction of outcome on admission is of great clinical relevance. We aimed to develop prognostic models with readily available traditional and novel predictors. METHODS AND FINDINGS: Prospectively collected individual patient data were analyzed from 11 studies. We considered predictors available at admission in logistic regression models to predict mortality and unfavorable outcome according to the Glasgow Outcome Scale at 6 mo after injury. Prognostic models were developed in 8,509 patients with severe or moderate TBI, with cross-validation by omission of each of the 11 studies in turn. External validation was on 6,681 patients from the recent Medical Research Council Corticosteroid Randomisation after Significant Head Injury (MRC CRASH) trial. We found that the strongest predictors of outcome were age, motor score, pupillary reactivity, and CT characteristics, including the presence of traumatic subarachnoid hemorrhage. A prognostic model that combined age, motor score, and pupillary reactivity had an area under the receiver operating characteristic curve (AUC) between 0.66 and 0.84 at cross-validation. This performance could be improved (AUC increased by approximately 0.05) by considering CT characteristics, secondary insults (hypotension and hypoxia), and laboratory parameters (glucose and hemoglobin). External validation confirmed that the discriminative ability of the model was adequate (AUC 0.80). Outcomes were systematically worse than predicted, but less so in 1,588 patients who were from high-income countries in the CRASH trial. CONCLUSIONS: Prognostic models using baseline characteristics provide adequate discrimination between patients with good and poor 6 mo outcomes after TBI, especially if CT and laboratory findings are considered in addition to traditional predictors. The model predictions may support clinical practice and research, including the design and analysis of randomized controlled trials.

Effects of Glasgow Outcome Scale misclassification on traumatic brain injury clinical trials.

The Glasgow Outcome Scale (GOS) is the primary endpoint for efficacy analysis of clinical trials in traumatic brain injury (TBI). Accurate and consistent assessment of outcome after TBI is essential to the evaluation of treatment results, particularly in the context of multicenter studies and trials. The inconsistent measurement or interobserver variation on GOS outcome, or for that matter, on any outcome scales, may adversely affect the sensitivity to detect treatment effects in clinical trial. The objective of this study is to examine effects of nondifferential misclassification of the widely used five-category GOS outcome scale and in particular to assess the impact of this misclassification on detecting a treatment effect and statistical power. We followed two approaches. First, outcome differences were analyzed before and after correction for misclassification using a dataset of 860 patients with severe brain injury randomly sampled from two TBI trials with known differences in outcome. Second, the effects of misclassification on outcome distribution and statistical power were analyzed in simulation studies on a hypothetical 800-patient dataset. Three potential patterns of nondifferential misclassification (random, upward and downward) on the dichotomous GOS outcome were analyzed, and the power of finding treatments differences was investigated in detail. All three patterns of misclassification reduce the power of detecting the true treatment effect and therefore lead to a reduced estimation of the true efficacy. The magnitude of such influence not only depends on the size of the misclassification, but also on the magnitude of the treatment effect. In conclusion, nondifferential misclassification directly reduces the power of finding the true treatment effect. An awareness of this procedural error and methods to reduce misclassification should be incorporated in TBI clinical trials.

A systematic review finds methodological improvements necessary for prognostic models in determining traumatic brain injury outcomes.

OBJECTIVES: To describe the modeling techniques used for early prediction of outcome in traumatic brain injury (TBI) and to identify aspects for potential improvements. STUDY DESIGN AND SETTING: We reviewed key methodological aspects of studies published between 1970 and 2005 that proposed a prognostic model for the Glasgow Outcome Scale of TBI based on admission data. RESULTS: We included 31 papers. Twenty-four were single-center studies, and 22 reported on fewer than 500 patients. The median of the number of initially considered predictors was eight, and on average five of these were selected for the prognostic model, generally including age, Glasgow Coma Score (or only motor score), and pupillary reactivity. The most common statistical technique was logistic regression with stepwise selection of predictors. Model performance was often quantified by accuracy rate rather than by more appropriate measures such as the area under the receiver-operating characteristic curve. Model validity was addressed in 15 studies, but mostly used a simple split-sample approach, and external validation was performed in only four studies. CONCLUSION: Although most models agree on the three most important predictors, many were developed on small sample sizes within single centers and hence lack generalizability. Modeling strategies have to be improved, and include external validation.

IMPACT database of traumatic brain injury: design and description.

The objective of this report is to describe the design and content of the International Mission for Prognosis And Clinical Trial (IMPACT) database of traumatic brain injury which contains the complete dataset from most clinical trials and organized epidemiologic studies conducted over the past 20 years. This effort, funded by the U.S. National Institutes of Health, has led to the accumulation thus far of data from 9205 patients with severe and moderate brain injuries from eight randomized placebo controlled trials and three observational studies. Data relevant to the design and analysis of pragmatic Phase III clinical trials, including pre-hospital, admission, and post-resuscitation assessments, information on the acute management, and short- and long-term outcome were merged into a top priority data set (TPDS). The major emphasis during the first phase of study is on information from time of injury to post-resuscitation and outcome at 6 months thereby providing a unique resource for prognostic analysis and for studies aimed at optimizing the design and analysis of Phase III trials in traumatic brain injury.

Statistical approaches to the univariate prognostic analysis of the IMPACT database on traumatic brain injury.

The univariate prognostic analysis of the IMPACT database on traumatic brain injury (TBI) poses the formidable challenge of how best to summarize a highly complex set of data in a way which is accessible without being overly simplistic. In this paper, we describe and illustrate the battery of statistical methods that have been used. Boxplots, histograms, tabulations, and splines were used for initial data checking and in identifying appropriate transformations for more formal statistical modeling. Imputation techniques were used to minimize the problems associated with the analysis of incomplete data due to missing values. The associations between covariates and outcome (Glasgow Outcome Scale [GOS] assessed at 6 months) were expressed as odds ratios with supporting confidence intervals when the GOS was collapsed to a dichotomous scale. This was extended to use common odds ratios from proportional odds models to express associations over the full range of the GOS. Forest plots were used to illustrate the consistency of results from study to study within the IMPACT database. The overall prognostic strength of the prognostic factors was expressed as the proportion of variance explained (Nagelkerke's R(2) statistic). Many of our approaches are based on simple graphical displays of the data, but, where appropriate, we have also used methods that although established in the statistical literature are relatively novel in their application to TBI.

Prognostic value of demographic characteristics in traumatic brain injury: results from the IMPACT study.

Outcome following traumatic brain injury (TBI) is not only dependent on the nature and severity of injury and subsequent treatment, but also on constituent characteristics of injured individuals. We aimed to describe and quantify the relationship between demographic characteristics and six month outcome assessed by the Glasgow Outcome Scale (GOS) after TBI. Individual patient data on age (n = 8719), gender (n = 8720), race (n = 5320), and education (n = 2201) were extracted from eight therapeutic Phase III randomized clinical trials and three surveys in moderate or severe TBI, contained in the IMPACT database. The strength of prognostic effects was analyzed with binary and proportional odds regression analysis and expressed as an odds ratio. Age was analyzed as a continuous variable with spline functions, and the odds ratio calculated over the difference between the 75 th and 25 th percentiles. Associations with other predictors were explored. Increasing age was strongly related to poorer outcome (OR 2.14; 95% CI 2.00-2.28) in a continuous fashion that could be approximated by a linear function. No gender differences in outcome were found (OR: 1.01; CI 0.92-1.11), and exploratory analysis failed to show any gender/age interaction. The studies included predominantly Caucasians (83%); outcome in black patients was poorer relative to this group (OR 1.30; CI 1.09-1.56). This relationship was sustained on adjusted analyses, and requires further study into mediating factors. Higher levels of education were weakly related to a better outcome (OR: 0.70; CI 0.52-0.94). On multivariable analysis adjusting for age, motor score, and pupils, the prognostic effect of race and education were sustained. We conclude that outcome following TBI is dependent on age, race, to a lesser extent on education, but not on gender.

Prognostic value of the Glasgow Coma Scale and pupil reactivity in traumatic brain injury assessed pre-hospital and on enrollment: an IMPACT analysis.

We studied the prognostic strength of the individual components of the Glasgow Coma Scale (GCS) and pupil reactivity to Glasgow Outcome Score (GOS) at 6 months post-injury. A total of 8721 moderate or severe traumatic brain injury (TBI) patient data from the IMPACT database on traumatic brain injury comprised the study cohort. The associations between motor score and pupil reactivity and 6-month GOS were analyzed by binary logistic regression and proportional odds methodology. The strength of prognostic effects were expressed as the unadjusted odds ratios presented for all individual studies as well as in meta-analysis. We found a consistent strong association between motor score and 6-month GOS across all studies (OR 1.74-7.48). The Eye and Verbal components were also strongly associated with GOS. In the pooled population, one or both un-reactive pupils and lower motor scores were significantly associated with unfavorable outcome (range 2.71-7.31). We also found a significant change in motor score from pre-hospital direct to study hospital enrollment ( p < 0.0001) and from the first in-hospital to study enrollment scores (p < 0.0001). Pupil reactivity was more robust between these time points. It is recommended that the study hospital enrollment GCS and pupil reactivity be used for prognostic analysis.

Prognostic value of cause of injury in traumatic brain injury: results from the IMPACT study.

We aimed to describe and quantify the relationship between cause of injury and final outcome following traumatic brain injury (TBI). Individual patient data (N = 8708) from eight therapeutic Phase III randomized clinical trials in moderate or severe TBI, and three TBI surveys were used to investigate the relationship between cause of injury and outcome, as assessed by the Glasgow Outcome Scale (GOS) at 6 months. Proportional odds methodology was applied to quantify the strength of the association and expressed as an odds ratio in a meta-analysis. Heterogeneity across studies was assessed and associations with other predictive factors explored. In a univariate analysis, a strong association between the cause of injury and long-term outcome in moderate to severe TBI patients was observed, with consistent results across the studies. Road traffic accidents (OR 0.66, 95% CI 0.60-0.73), assaults (OR 0.66, 95% CI 0.52-0.84), and injuries sustained during sporting or recreational activities (OR 0.45, 95% CI 0.28-0.71) were all associated with better outcomes than the reference category of falls. Falls were found to be associated with an older age and with a higher incidence of mass lesions. Following adjustment for age in the analysis, the relationship between cause of injury and outcome was lost.

Prognostic value of secondary insults in traumatic brain injury: results from the IMPACT study.

We determined the relationship between secondary insults (hypoxia, hypotension, and hypothermia) occurring prior to or on admission to hospital and 6-month outcome after traumatic brain injury (TBI). A meta-analysis of individual patient data, from seven Phase III randomized clinical trials (RCT) in moderate or severe TBI and three TBI population-based series, was performed to model outcome as measured by the Glasgow Outcome Scale (GOS). Proportional odds modeling was used to relate the probability of a poor outcome to hypoxia (N = 5661), hypotension ( N = 6629), and hypothermia ( N = 4195) separately. We additionally analyzed the combined effects of hypoxia and hypotension and performed exploratory analysis of associations with computerized tomography (CT) classification and month of injury. Having a pre-enrollment insult of hypoxia, hypotension or hypothermia is strongly associated with a poorer outcome (odds ratios of 2.1 95% CI [1.7-2.6], 2.7 95% CI [2.1-3.4], and 2.2 95% CI [1.6-3.2], respectively). Patients with both hypoxia and hypotension had poorer outcomes than those with either insult alone. Radiological signs of raised intracranial pressure (CT class III or IV) were more frequent in patients who had sustained hypoxia or hypotension. A significant association was observed between month of injury and hypothermia. The occurrence of secondary insults prior to or on admission to hospital in TBI patients is strongly related to poorer outcome and should therefore be a priority for emergency department personnel.

Prognostic value of admission blood pressure in traumatic brain injury: results from the IMPACT study.

Hypotension following traumatic brain injury (TBI) is recognized as an important secondary insult that is associated with adverse outcome. We aimed to describe the relationship between actual levels of admission blood pressure and Glasgow Outcome Scale (GOS) at 6 months. Individual patient data from the IMPACT database were available on systolic (N = 6801) and mean arterial (N = 6647) blood pressure. Regression models with restricted cubic spline functions were used to explore the shape of the relationships between blood pressure and outcome in unadjusted and adjusted analyses. Proportional odds methodology was applied to quantify the strength of the associations across the full range of the GOS. Analyses were performed to search for threshold values. A smooth U-shaped relationship was observed between systolic (SBP) and mean arterial (MABP) blood pressures and outcome, without any evidence of an abrupt threshold effect. Best outcomes were observed for values of SBP of the order of 135 mm Hg and for values of MABP of the order of 90 mm Hg. Both lower (OR 1.53; 95% CI: 1.31-1.80) and higher levels (OR 1.42; CI: 1.20-1.68) of SBP and lower (OR 1.30; CI 1.12-1.51) and higher levels of MABP (OR 1.45; CI 1.19-1.76) were associated with poorer outcome. These findings were consistent across studies. The relationship between high blood pressure level and poorer outcome largely disappeared on adjusted analysis. Current guidelines for the management of blood pressure in TBI focus on the avoidance of hypotension as defined by SBP < 90 mm Hg. Our finding of a smooth relationship with improving outcome as SBP increases up to 135 mm Hg, while not supporting a strong causal inference, does suggest that current guidelines need to be reconsidered.

Prognostic value of computerized tomography scan characteristics in traumatic brain injury: results from the IMPACT study.

Computerized tomography (CT) scanning provides an objective assessment of the structural damage to the brain following traumatic brain injury (TBI). We aimed to describe and quantify the relationship between CT characteristics and 6-month outcome, assessed by the Glasgow Outcome Scale (GOS). Individual patient data from the IMPACT database were available on CT classification (N = 5209), status of basal cisterns ( N = 3861), shift ( N = 4698), traumatic subarachnoid hemorrhage (tSAH) ( N = 7407), and intracranial lesions ( N = 7613). We used binary logistic and proportional odds regression for prognostic analyses. The CT classification was strongly related to outcome, with worst outcome for patients with diffuse injuries in CT class III (swelling; OR 2.50; CI 2.09-3.0) or CT class IV (shift; OR 3.03; CI 2.12-4.35). The prognosis in patients with mass lesions was better for patients with an epidural hematoma (OR 0.64; CI 0.56-0.72) and poorer for an acute subdural hematoma (OR 2.14; CI 1.87-2.45). Partial obliteration of the basal cisterns (OR 2.45; CI 1.88-3.20), tSAH (OR 2.64; CI 2.42-2.89), or midline shift (1-5 mm-OR 1.36; CI 1.09-1.68); >5 mm-OR 2.20; CI 1.64-2.96) were strongly related to poorer outcome. Discrepancies were found between the scoring of basal cisterns/shift and the CT classification, indicating observer variation. These were less marked in studies that had used a central review process. Multivariable analysis indicated that individual CT characteristics added substantially to the prognostic value of the CT classification alone. We conclude that both the CT classification and individual CT characteristics are important predictors of outcome in TBI. For clinical trials, a central review process is advocated to minimize observer variability in CT assessment.

Prognostic value of admission laboratory parameters in traumatic brain injury: results from the IMPACT study.

Abnormalities in laboratory parameters are frequent following traumatic brain injury (TBI), but few studies have investigated their predictive value. We aimed to describe and quantify the relation between laboratory parameters that are routinely determined on admission and final outcome following TBI. Individual patient data were available in the IMPACT database from six Phase III randomized controlled trials and one observational study in TBI. We studied glucose (N = 4834), sodium ( N = 5270), pH ( N = 3398), hemoglobin (Hb, N = 3875), platelet count ( N = 1629), and prothrombin time (PT; N = 840) for their associations with outcome at 6 months (Glasgow Outcome Scale [GOS]). We used logistic regression models with linear, quadratic, and restricted cubic spline functions. The strength of the associations was expressed as an unadjusted odds ratio, calculated over the shift in outcome between the 25th and 75th percentiles. Proportional odds methodology was further applied to quantify the strength of the associations across the full range of the GOS. All parameters were consistently associated with outcome in a continuous relationship: glucose and prothrombin time showed a positive linear relation to outcome (i.e., increasing values associated with poorer outcome) and Hb, platelets, and pH an inverse linear relation (i.e., low values associated with poorer outcome). Sodium demonstrated a U-shaped relation to outcome, with low levels being more strongly related to poorer outcome. Effects were strongest for increasing levels of glucose (odds ratio 1.7; 95% CI 1.54-1.83) and decreasing levels of Hb (odds ratio 0.7; CI 0.60-0.78). Higher glucose values were associated with increasing age, but on adjusted analysis, the strength of the association with outcome remained. Whether treatment of abnormal values may improve outcome needs further rigorous study.

Multivariable prognostic analysis in traumatic brain injury: results from the IMPACT study.

We studied the prognostic value of a wide range of conventional and novel prognostic factors on admission after traumatic brain injury (TBI) using both univariate and multivariable analysis. The outcome measure was Glasgow Outcome Scale at 6 months after injury. Individual patient data were available on a cohort of 8686 patients drawn from eight randomized controlled trials and three observational studies. The most powerful independent prognostic variables were age, Glasgow Coma Scale (GCS) motor score, pupil response, and computerized tomography (CT) characteristics, including the Marshall CT classification and traumatic subarachnoid hemorrhage. Prothrombin time was also identified as a powerful independent prognostic factor, but it was only available for a limited number of patients coming from three of the relevant studies. Other important prognostic factors included hypotension, hypoxia, the eye and verbal components of the GCS, glucose, platelets, and hemoglobin. These results on prognostic factors will underpin future work on the IMPACT project, which is focused on the development of novel approaches to the design and analysis of clinical trials in TBI. In addition, the results provide pointers to future research, including further analysis of the prognostic value of prothrombin time, and the evaluation of the clinical impact of intervening aggressively to correct abnormalities in hemoglobin, glucose, and coagulation.

Adjustment for strong predictors of outcome in traumatic brain injury trials: 25% reduction in sample size requirements in the IMPACT study.

The aim of this study was to quantify the potential reduction in sample size that can be achieved by adjustment for predictors of outcome in traumatic brain injury (TBI) trials. We used individual patient data from seven therapeutic phase III randomized clinical trials (RCTs; n = 6166) in moderate or severe TBI, and three TBI surveys (n = 2238). The primary outcome was the dichotomized Glasgow Outcome Scale at 6 months (favorable/unfavorable). Baseline predictors of outcome considered were age, motor score, pupillary reactivity, computed tomography (CT) classification, traumatic subarachnoid hemorrhage, hypoxia, hypotension, glycemia, and hemoglobin. We calculated the potential sample size reduction obtained by adjustment of a hypothetical treatment effect for one to seven predictors with logistic regression models. The distribution of predictors was more heterogeneous in surveys than in trials. Adjustment of the treatment effect for the strongest predictors (age, motor score, and pupillary reactivity) yielded a reduction in sample size of 16-23% in RCTs and 28-35% in surveys. Adjustment for seven predictors yielded a reduction of about 25% in most studies: 20-28% in RCTs and 32-39% in surveys. A major reduction in sample size can be obtained with covariate adjustment in TBI trials. Covariate adjustment for strong predictors should be incorporated in the analysis of future TBI trials.