A screening of the general population for the presence of carotid stenosis is not recommended in current guidelines. On the other hand, screening of risk groups is regarded as possibly useful without it being possible to define these risk groups in more detail. In purely mathematical terms, the actual benefit of a screening programme depends not only on the positive and negative predictive value of the procedure or the positive likelihood ratio, but also on the a priori probability of the condition to be detected (e.g. a disease), i.e. here the prevalence of carotid stenosis.
This raises the fundamental question: How high must the a-priori probability (prevalence) of carotid stenosis be in a patient population (i.e. risk group) for a screening programme to be meaningful under the given conditions?
To answer this question, a decision-theoretical model was developed to simulate a screening programme as realistically as possible. Studies on the effectiveness and efficiency of screening programmes for the presence of carotid stenosis have been published several times. However, the present study is the first to overcome several methodological limitations of previous studies. These include the finer graduation of stenosis degrees, the inclusion of a possible progression of ACI stenosis over the course of the observation period, confirmatory examinations by MR angiography, re-screening, interventions in the control group that can be expected in reality (e.g. CEA after TIA) and secondary outcomes such as myocardial infarction and cranial nerve lesions. From a clinical perspective, the model is based on current guideline recommendations (Eckstein et al. 2013; Brett and Levine 2014; Aboyans et al. 2018) as well as expert opinions. Technically, it is a Markov Chain Monte Carlo (MCMC) simulation modeled with TreeAge Pro Health Care®. The necessary model parameters as well as their estimation uncertainty (confidence interval) are derived from an extensive literature research and several meta-analyses. The following three models were compared:
(1) Group 0: Control without screening
(2) Group 1: Single screening with duplex ultrasound and verification by
MR angiography
(3) Group 2: Complex screening with regular re-screening
The analysis showed that screening of a risk group for the presence of carotid stenosis is only effective if the expected a priori probability (prevalence) is at least 5-8 %, i.e. net stroke prevention could be achieved. From a prevalence of about 5 %, the number needed to screen (NNS) falls to about 500 and below. In addition, this study shows that a one-time screening program has no relevant disadvantage in terms of prevented strokes compared to a complex screening program and avoids the significant additional expense of a complex screening program.
Despite the prevention of strokes, the high number of carotid endarterectomies (CEA) required for this leads to an overall loss of quality of life in both simulated screening programs.
In view of the loss of quality of life, the high NNS and the enormous costs, a general screening for the presence of high-grade carotid stenosis does not appear to be appropriate, whereas it might be useful from a purely clinical point of view in collectives with a prevalence of more than 5-10%.
«
A screening of the general population for the presence of carotid stenosis is not recommended in current guidelines. On the other hand, screening of risk groups is regarded as possibly useful without it being possible to define these risk groups in more detail. In purely mathematical terms, the actual benefit of a screening programme depends not only on the positive and negative predictive value of the procedure or the positive likelihood ratio, but also on the a priori probability of the condit...
»