http://repub.eur.nl/res/aut/10735/ List of Publications en http://repub.eur.nl/static-eur/img/logo.png http://repub.eur.nl http://repub.eur.nl/res/pub/39382/ 2013-03-22T00:00:00Z Purpose: We determined whether FGFR3 mutation analysis of voided urine samples would be cost-effective to partly replace cystoscopy in the surveillance of patients treated for nonmuscle invasive urothelial carcinoma. Materials and Methods: In this decision analytical study we analyzed data on 70 Dutch patients with FGFR3 positive primary tumors and a median followup of 8.8 years. Surveillance strategies were compared in a Markov model. Modified surveillance consisted of FGFR3 mutation analysis of voided urine samples every 3 months, and cystoscopy at 3, 12 and 24 months. Standard surveillance was defined as cystoscopy every 3 months and minimal surveillance was defined as cystoscopy at 3, 12 and 24 months. Analysis was stratified for 3 risk profiles, including surveillance after 1) the primary tumor, 2) the first to third recurrence and 3) the fourth recurrence or more. Sensitivity analysis was performed to evaluate the impact of variations in cost, sensitivity and specificity. Results: The probability of no recurrence after 2 years of surveillance after a primary tumor was higher for modified surveillance than for standard and minimal surveillance, eg after primary tumors (95.7% vs 95.0% and 93.9%, respectively). The total cost of surveillance after the primary tumor was lower for minimal and modified surveillance (€2,254 and €2,558, respectively) than for standard surveillance (€5,861). Results were robust to changing inputs over plausible ranges and consistent for each of the 3 risk profiles. Conclusions: Surveillance in which cystoscopy is partly replaced by FGFR3 mutation analysis of urine seems a safe, effective and cost-effective surveillance strategy. Further validation in larger cohorts is required. http://repub.eur.nl/res/pub/37395/ 2012-10-15T00:00:00Z During the recent decades, interest in prediction models has substantially increased, but approaches to synthesize evidence from previously developed models have failed to keep pace. This causes researchers to ignore potentially useful past evidence when developing a novel prediction model with individual participant data (IPD) from their population of interest. We aimed to evaluate approaches to aggregate previously published prediction models with new data. We consider the situation that models are reported in the literature with predictors similar to those available in an IPD dataset. We adopt a two-stage method and explore three approaches to calculate a synthesis model, hereby relying on the principles of multivariate meta-analysis. The former approach employs a naive pooling strategy, whereas the latter accounts for within-study and between-study covariance. These approaches are applied to a collection of 15 datasets of patients with traumatic brain injury, and to five previously published models for predicting deep venous thrombosis. Here, we illustrated how the generally unrealistic assumption of consistency in the availability of evidence across included studies can be relaxed. Results from the case studies demonstrate that aggregation yields prediction models with an improved discrimination and calibration in a vast majority of scenarios, and result in equivalent performance (compared with the standard approach) in a small minority of situations. The proposed aggregation approaches are particularly useful when few participant data are at hand. Assessing the degree of heterogeneity between IPD and literature findings remains crucial to determine the optimal approach in aggregating previous evidence into new prediction models. http://repub.eur.nl/res/pub/37396/ 2012-10-15T00:00:00Z Diagnostic problems in medicine are sometimes polytomous, meaning that the outcome has more than two distinct categories. For example, ovarian tumors can be benign, borderline, primary invasive, or metastatic. Extending the main measure of binary discrimination, the c-statistic or area under the ROC curve, to nominal polytomous settings is not straightforward. This paper reviews existing measures and presents the polytomous discrimination index (PDI) as an alternative. The PDI assesses all sets of k cases consisting of one case from each outcome category. For each category i (i=1,...,k), it is assessed whether the risk of category i is highest for the case from category i. A score of 1/k is given per category for which this holds, yielding a set score between 0 and 1 to indicate the level of discrimination. The PDI is the average set score and is interpreted as the probability to correctly identify a case from a randomly selected category within a set of k cases. This probability can be split up by outcome category, yielding k category-specific values that result in the PDI when averaged. We demonstrate the measures on two diagnostic problems (residual mass histology after chemotherapy for testicular cancer; diagnosis of ovarian tumors). We compare the behavior of the measures on theoretical data, showing that PDI is more strongly influenced by simultaneous discrimination between all categories than by partial discrimination between pairs of categories. In conclusion, the PDI is attractive because it better matches the requirements of a measure to summarize polytomous discrimination. http://repub.eur.nl/res/pub/30928/ 2011-10-01T00:00:00Z Background: Several clinical risk scores (CRSs) for the outcome of patients with colorectal liver metastases have been validated, but not in patients undergoing neoadjuvant chemotherapy. Therefore, this study evaluates the predictive value of these CRSs in this specific group. Methods: Between January 2000 and December 2008, all patients undergoing a metastasectomy were analyzed and divided into two groups: 193 patients did not receive neoadjuvant chemotherapy (group A), and 159 patients received neoadjuvant chemotherapy (group B). In group B, the CRSs were calculated before and after administration of neoadjuvant chemotherapy. Results were evaluated by using the CRSs proposed by Nordlinger et al., Fong et al., Nagashima et al., and Konopke et al. Results: In groups A and B, the overall median survival was 43 and 47 months, respectively (P = 0.648). In group A, all CRSs used were of statistically significant predictive value. Before administration of neoadjuvant chemotherapy, only the Nordlinger score was of predictive value. After administration of neoadjuvant chemotherapy, all CRSs were of predictive value again, except for the Konopke score. Conclusions: Traditional CRSs are not a reliable prognostic tool when used in patients before treatment with neoadjuvant chemotherapy. However, CRSs assessed after the administration of neoadjuvant chemotherapy are useful to predict prognosis. http://repub.eur.nl/res/pub/23419/ 2011-02-01T00:00:00Z http://repub.eur.nl/res/pub/27535/ 2010-10-15T00:00:00Z Various performance measures related to calibration and discrimination are available for the assessment of risk models. When the validity of a risk model is assessed in a new population, estimates of the model's performance can be influenced in several ways. The regression coefficients can be incorrect, which indeed results in an invalid model. However, the distribution of patient characteristics (case mix) may also influence the performance of the model. Here the authors consider a number of typical situations that can be encountered in external validation studies. Theoretical relations between differences in development and validation samples and performance measures are studied by simulation. Benchmark values for the performance measures are proposed to disentangle a case-mix effect from incorrect regression coefficients, when interpreting the model's estimated performance in validation samples. The authors demonstrate the use of the benchmark values using data on traumatic brain injury obtained from the International Tirilazad Trial and the North American Tirilazad Trial (1991-1994). http://repub.eur.nl/res/pub/27563/ 2010-07-01T00:00:00Z Aim: The cost-effectiveness of passive immunisation against respiratory syncytial virus (RSV) in the Netherlands was studied by assessing incremental costs to prevent one hospitalisation in high-risk children using a novel individualised monthly approach. Methods: Cost-effectiveness analysis was performed by combining estimates of individual hospitalisation costs and monthly hospitalisation risks, with immunisation costs, parental costs and efficacy of passive immunisation for a reference case with the highest hospitalisation risks and costs of hospitalisation during the RSV season (male, gestational age ≤28 weeks, birth weight ≤2500 g, having bronchopulmonary dysplasia (BPD), aged 0 months at the beginning of the season (October)). Various sensitivity analyses and a cost-neutrality analysis were performed. Results: Cost-effectiveness of passive immunisation varied widely by child characteristics and seasonal month. For the reference case it was most cost effective in December at €13 190 per hospitalisation averted. Cost-effectiveness was most sensitive to changes in hospitalisation risk. For the reference case, cost neutrality was reached in December, if acquisition costs of passive immunisation decreased from €930 to €375, monthly hospitalisation risk increased from 7.6% to 17%, or hospitalisation costs increased from €10 250 to €23 250 per hospitalisation. Even if passive immunisation prevented all hospitalisations, costs per hospitalisation averted in December would still exceed €2645. Conclusions: Although cost-effectiveness of passive immunisation varied strongly by child characteristics and seasonal month, incremental costs per hospitalisation averted were always high. A restrictive immunisation policy only immunising children with BPD in high-risk months is therefore recommended. The costs of passive immunisation would have to be considerably reduced to achieve cost-effectiveness. http://repub.eur.nl/res/pub/29729/ 2008-10-01T00:00:00Z http://repub.eur.nl/res/pub/29815/ 2008-02-01T00:00:00Z Objective: Physicians commonly consider the presence of all differential diagnoses simultaneously. Polytomous logistic regression modeling allows for simultaneous estimation of the probability of multiple diagnoses. We discuss and (empirically) illustrate the value of this method for diagnostic research. Study Design and Setting: We used data from a study on the diagnosis of residual retroperitoneal mass histology in patients presenting with nonseminomatous testicular germ cell tumor. The differential diagnoses include benign tissue, mature teratoma, and viable cancer. Probabilities of each diagnosis were estimated with a polytomous logistic regression model and compared with the probabilities estimated from two consecutive dichotomous logistic regression models. Results: We provide interpretations of the odds ratios derived from the polytomous regression model and present a simple score chart to facilitate calculation of predicted probabilities from the polytomous model. For both modeling methods, we show the calibration plots and receiver operating characteristics curve (ROC) areas comparing each diagnostic outcome category with the other two. The ROC areas for benign tissue, mature teratoma, and viable cancer were similar for both modeling methods, 0.83 (95% confidence interval [CI] = 0.80-0.85) vs. 0.83 (95% CI = 0.80-0.85), 0.78 (95% CI = 0.75-0.81) vs. 0.78 (95% CI = 0.75-0.81), and 0.66 (95% CI = 0.61-0.71) vs. 0.64 (95% CI = 0.59-0.69), for polytomous and dichotomous regression models, respectively. Conclusion: Polytomous logistic regression is a useful technique to simultaneously model predicted probabilities of multiple diagnostic outcome categories. The performance of a polytomous prediction model can be assessed similarly to a dichotomous logistic regression model, and predictions by a polytomous model can be made with a user-friendly method. Because the simultaneous consideration of the presence of multiple (differential) conditions serves clinical practice better than consideration of the presence of only one target condition, polytomous logistic regression could be applied more often in diagnostic research. http://repub.eur.nl/res/pub/10708/ 2006-07-14T00:00:00Z Objectives: Surgical resection of postchemotherapy retroperitoneal lymph nodes is often performed in patients with advanced nonseminomatous testicular germ cell cancer. We previously developed a model to predict the probability that the lymph nodes contain only necrotic or fibrotic (benign) tissue versus mature teratoma and viable cancer (tumour) to identify patients who actually need resection. The present study used an updated model with new patient data and studied the validity of the updated model across various settings. Methods: We combined data of 544 patients from the original model with data of 550 new patients and performed a new logistic regression analysis, which included the same six predictors: histology of the primary tumour, prechemotherapy serum levels of a-fetoprotein, human chorionic gonadotropin, lactate dehydrogenase, residual mass size measured on computed tomography, and change in mass size. The validity of the updated model was studied in individual centres. Calibration of the predicted probabilities was assessed graphically and with the Hosmer-Lemeshow test. Discrimination was studied with the concordance (c)-statistic. Results: The updated model had slightly different, although more precise, regression coefficients. Statistically nonsignificant Hosmer-Lemeshow tests confirmed good calibration in most centres. The c-statistic for all centres except one exceeded 0.80. The updated model was valid over the complete range of predicted probabilities across a broad spectrum of centres. Conclusions: This finding gives confidence in the applicability of the model to select patients for resection, particularly patients with small residual masses and low predicted probabilities of benign tissue (i.e., substantial predicted risks of residual tumour). http://repub.eur.nl/res/pub/1169/ 2003-09-03T00:00:00Z linical prediction models combine patient characteristics to predict the probability of having a certain disease (diagnosis) or the probability that a particular disease state will occur (prognosis). The predicted probability of the diagnostic or prognostic outcome may assist the clinician in decision making for patient care. Before a prediction model can reliably be applied in clinical practice, the performance of the model in new patients needs to be studied (‘external validity’). This thesis described several theoretical and practical aspects of the external validation of clinical prediction models. The objectives were (i) to describe aspects of model validity and relevant performance measures; (ii) to estimate the power of these performance measures; (iii) to externally validate a prediction model for residual mass histology in testicular cancer; and (iv) to update this model with all available information. Three aspects of model performance are discussed: calibration, discrimination, and clinical usefulness. The external validity of a model does not only depend on the new patients for whom the model is applied, but also on the development process of the model. Therefore, this thesis contained also some illustrations of model development aspects. The development process of the prediction model for residual mass histology in nonseminomatous testicular germ cell cancer was described; the model predicts the probability that a residual retroperitoneal mass contains only benign tissue after cis-platin based chemotherapy for metastatic tumour.