Stefan Visscher

A dynamic Bayesian network for diagnosing ventilator-associated pneumonia in ICU patients

Diagnosing ventilator-associated pneumonia in mechanically ventilated patients in intensive care units is seen as a clinical challenge. The difficulty in diagnosing ventilator-associated pneumonia stems from the lack of a simple yet accurate diagnostic test. To assist clinicians in diagnosing and treating patients with pneumonia, a decision-theoretic network had been designed with the help of domain experts. A major limitation of this network is that it does not represent pneumonia as a dynamic process that evolves over time. In this paper, we construct a dynamic Bayesian network that explicitly captures the development of the disease over time. We discuss how probability elicitation from domain experts served to quantify the dynamics involved and how the nature of the patient data helps reduce the computational burden of inference. We evaluate the diagnostic performance of our dynamic model for a number of real patients and report promising results.

Multilevel Bayesian networks for the analysis of hierarchical health care data

OBJECTIVE Large health care datasets normally have a hierarchical structure, in terms of levels, as the data have been obtained from different practices, hospitals, or regions. Multilevel regression is the technique commonly used to deal with such multilevel data. However, for the statistical analysis of interactions between entities from a domain, multilevel regression yields little to no insight. While Bayesian networks have proved to be useful for analysis of interactions, they do not have the capability to deal with hierarchical data. In this paper, we describe a new formalism, which we call multilevel Bayesian networks; its effectiveness for the analysis of hierarchically structured health care data is studied from the perspective of multimorbidity. METHODS Multilevel Bayesian networks are formally defined and applied to analyze clinical data from family practices in The Netherlands with the aim to predict interactions between heart failure and diabetes mellitus. We compare the results obtained with multilevel regression. RESULTS The results obtained by multilevel Bayesian networks closely resembled those obtained by multilevel regression. For both diseases, the area under the curve of the prediction model improved, and the net reclassification improvements were significantly positive. In addition, the models offered considerable more insight, through its internal structure, into the interactions between the diseases. CONCLUSIONS Multilevel Bayesian networks offer a suitable alternative to multilevel regression when analyzing hierarchical health care data. They provide more insight into the interactions between multiple diseases. Moreover, a multilevel Bayesian network model can be used for the prediction of the occurrence of multiple diseases, even when some of the predictors are unknown, which is typically the case in medicine.

Modelling treatment effects in a clinical Bayesian network using Boolean threshold functions

OBJECTIVE Appropriate antimicrobial treatment of infections in critically ill patients should be started as soon as possible, as delay in treatment may reduce a patients prognostic outlook considerably. Ventilator-associated pneumonia (VAP) occurs in patients in intensive care units who are mechanically ventilated and is almost always preceded by colonisation of the respiratory tract by the causative microorganisms. It is very difficult to clinically diagnose VAP and, therefore, some form of computer-based decision support might be helpful for the clinician. MATERIALS AND METHODS As diagnosing and treating VAP involves reasoning with uncertainty, we have used a Bayesian network as the primary tool for building a decision-support system. The effects of usage of antibiotics on the colonisation of the respiratory tract by various pathogens and the subsequent antibiotic choices in case of VAP were modelled using the notion of causal independence. In particular, the conditional probability distribution of the random variable that represents the overall coverage of pathogens by antibiotics was modelled in terms of the conjunctive effect of the seven different pathogens, usually referred to as the noisy-AND model. In this paper, we investigate different coverage models, as well as generalisations of the noisy-AND, called noisy-threshold models, and test them on clinical data of intensive care unit (ICU) patients who are mechanically ventilated. RESULTS Some of the constructed noisy-threshold models offered further improvement of the performance of the Bayesian network in covering present causative pathogens by advising appropriate antimicrobial treatment. CONCLUSIONS By reconsidering the modelling of interactions between the random variables in a Bayesian network using the theory of causal independence, it is possible to refine its performance. This was clearly shown for our Bayesian network concerning VAP, indicating that only specific noisy-threshold models might be appropriate for the modelling of the interaction between pathogens and antimicrobial treatment with respect to susceptibility. The results obtained also provide evidence that the noisy-OR and noisy-AND might not always be the best functions to model interactions among random variables.

Multilevel temporal Bayesian networks can model longitudinal change in multimorbidity

OBJECTIVES Although the course of single diseases can be studied using traditional epidemiologic techniques, these methods cannot capture the complex joint evolutionary course of multiple disorders. In this study, multilevel temporal Bayesian networks were adopted to study the course of multimorbidity in the expectation that this would yield new clinical insight. STUDY DESIGN AND SETTING Clinical data of patients were extracted from 90 general practice registries in the Netherlands. One and half million patient-years were used for analysis. The simultaneous progression of six chronic cardiovascular conditions was investigated, correcting for both patient and practice-related variables. RESULTS Cumulative incidence rates of one or more new morbidities rapidly increase with the number of morbidities present at baseline, ranging up to 47% and 76% for 3- and 5-year follow-ups, respectively. Hypertension and lipid disorders, as health risk factors, increase the cumulative incidence rates of both individual and multiple disorders. Moreover, in their presence, the observed cumulative incidence rates of combinations of cardiovascular disorders, that is, multimorbidity differs significantly from the expected rates. CONCLUSION There are clear synergies between health risks and chronic diseases when multimorbidity within a patient progresses over time. The method used here supports a more comprehensive analysis of such synergies compared with what can be obtained by traditional statistics.

Frequency and Risk Factors for Under- and Over-Treatment in Stroke Prevention for Patients with Non-Valvular Atrial Fibrillation in General Practice

Objective To determine adequacy of antithrombotic treatment in patients with non-valvular atrial fibrillation. To determine risk factors for under- and over-treatment. Design Retrospective, cross-sectional study of electronic health records from 36 general practitioners in 2008. Setting General practice in the Netherlands. Subjects Primary care physicians (n = 36) and patients (n = 981) aged 65 years and over. Main Outcome Measures Rates of adequate, under and over-treatment, risk factors for under and over-treatment. Results Of the 981 included patients with a mean of age 78, 18% received no antithrombotic treatment (under-treatment), 13% received antiplatelet drugs and 69% received oral anticoagulation (OAC). Further, 43% of the included patients were treated adequately, 26% were under-treated, and 31% were over-treated. Patients with a previous ischaemic stroke were at high risk for under-treatment (OR 2.4, CI 1.6–3.5), whereas those with contraindications for OAC were at high risk for over-treatment (OR 37.0, CI 18.1–79.9). Age over 75 (OR 0.2, CI: 0.1–0.3]), diabetes (OR 0.1, CI: 0.1–0.3), heart failure (OR 0.2, CI: 0.1–0.3), hypertension (OR 0.1, CI: 0.1–0.2) and previous ischaemic stroke (OR 0.04, CI: 0.02–0.11) protected against over-treatment. Conclusions In general practice, CHADS2-criteria are being used, but the antithrombotic treatment of patients with atrial fibrillation frequently deviates from guidelines on this topic. Patients with previous stroke are at high risk of not being prescribed OAC. Contraindications for OAC, however, seem to be frequently overlooked.

Improving the therapeutic performance of a medical bayesian network using noisy threshold models

Treatment management in critically ill patients needs to be efficient, as delay in treatment may give rise to deterioration in the patients condition. Ventilator-associated pneumonia (VAP) occurs in patients who are mechanically ventilated in intensive care units. As it is quite difficult to diagnose and treat VAP, some form of computer-based decision support might be helpful. As diagnosing and treating disorders in medicine involves reasoning with uncertainty, we have used a Bayesian network as our primary tool for building a decision-support system for the clinical management of VAP. The effects of antibiotics on colonisation with various pathogens and subsequent antibiotic choices in case of VAP were modelled in the Bayesian network using the notion of causal independence. In particular, the conditional probability distribution of the random variable that represents the overall coverage of pathogens by antibiotics was modelled in terms of the conjunctive effect of the seven different pathogens, usually referred to as the noisy-AND gate. In this paper, we investigate generalisations of the noisy-AND, called noisy threshold models. It is shown that they offer a means for further improvement to the performance of the Bayesian network.

Quality of Co-Prescribing NSAID and Gastroprotective Medications for Elders in The Netherlands and Its Association with the Electronic Medical Record.

Objective To assess guideline adherence of co-prescribing NSAID and gastroprotective medications for elders in general practice over time, and investigate its potential association with the electronic medical record (EMR) system brand used. Methods We included patients 65 years and older who received NSAIDs between 2005 and 2010. Prescription data were extracted from EMR systems of GP practices participating in the Dutch NIVEL Primary Care Database. We calculated the proportion of NSAID prescriptions with co-prescription of gastroprotective medication for each GP practice at intervals of three months. Association between proportion of gastroprotection, brand of electronic medical record (EMR), and type of GP practice were explored. Temporal trends in proportion of gastroprotection between electronic medical records systems were analyzed using a random effects linear regression model. Results We included 91,521 patient visits with NSAID prescriptions from 77 general practices between 2005 and 2010. Overall proportion of NSAID prescriptions to the elderly with co-prescription of gastroprotective medication was 43%. Mean proportion of gastroprotection increased from 27% (CI 25–29%) in the first quarter of 2005 with a rate of 1.2% every 3 months to 55%(CI 52–58%) at the end of 2010. Brand of EMR and type of GP practice were independently associated with co-prescription of gastroprotection. Conclusion Although prescription of gastroprotective medications to elderly patients who receive NSAIDs increased in The Netherlands, they are not co-prescribed in about half of the indicated cases. Brand of EMR system is associated with differences in prescription of gastroprotective medication. Optimal design and utilization of EMRs is a potential area of intervention to improve quality of prescription.

Using Temporal Context-Specific Independence Information in the Exploratory Analysis of Disease Processes

Disease processes in patients are temporal in nature and involve uncertainty. It is necessary to gain insight into these processes when aiming at improving the diagnosis, treatment and prognosis of disease in patients. One way to achieve these aims is by explicitly modelling disease processes; several researchers have advocated the use of dynamic Bayesian networks for this purpose because of the versatility and expressiveness of this time-oriented probabilistic formalism. In the research described in this paper, we investigate the role of context-specific independence information in modelling the evolution of disease. The hypothesis tested was that within similar populations of patients differences in the learnt structure of a dynamic Bayesian network may result, depending on whether or not patients have a particular disease. This is an example of temporal context-specific independence information. We have tested and confirmed this hypothesis using a constraint-based Bayesian network structure learning algorithm which supports incorporating background knowledge into the learning process. Clinical data of mechanically-ventilated ICU patients, some of whom developed ventilator-associated pneumonia, were used for that purpose.

Kwaliteit van ICPC-codering

459 huis art s & we tensch ap 55 ( 10) ok tober 2012 Goed bijgehouden medische dossiers zijn van essentieel belang, zeker nu (onderdelen van) dossiers ook opvraagbaar kunnen zijn door anderen, bijvoorbeeld een zorgverlener op de huisartsenpost. Mede daarom vormt in 2012 en 2013 de kwaliteit van dossiervorming de basis voor de verdeling van de zogenaamde variabiliseringsgelden. Wij gingen na hoe het is gesteld met de kwaliteit van de ICPC-codering van huisartsen tijdens consulten en of er verschillen bestaan tussen huisartsinformatiesystemen (HISsen).

Using a bayesian-network model for the analysis of clinical time-series data

Time is an essential element in the clinical management of patients as disease processes develop in time. A typical example of a disease process where time is considered important is the development of ventilator-associated pneumonia (VAP). A Bayesian network was developed previously to support clinicians in the diagnosis and treatment of VAP. In the research described in this paper, we have investigated whether this Bayesian network can also be used to analyse the temporal data collected in the ICU for patterns indicating development of VAP. In addition, it was studied whether the Bayesian network was able to suggest appropriate antimicrobial treatment. A temporal database with over 17700 patient days was used for this purpose.