Knowledge-based approach to septic shock patient data using a neural network with trapezoidal activation functions

doi:10.1016/S0933-3657(03)00057-5

Artificial Intelligence in Medicine

Volume 28, Issue 2, June 2003, Pages 207-230

https://doi.org/10.1016/S0933-3657(03)00057-5 Get rights and content

Abstract

In this contribution we present an application of a knowledge-based neural network technique in the domain of medical research. We consider the crucial problem of intensive care patients developing a septic shock during their stay at the intensive care unit. Septic shock is of prime importance in intensive care medicine due to its high mortality rate. Our analysis of the patient data is embedded in a medical data analysis cycle, including preprocessing, classification, rule generation and interpretation. For classification and rule generation we chose an improved architecture based on a growing trapezoidal basis function network for our metric variables. Our results extend those of a black box classification and give a deeper insight in our patient data. We evaluate our results with classification and rule performance measures. For feature selection we introduce a new importance measure.

Introduction

During the last years many scientists have published medical applications of neural networks for classification analysis, e.g. [2], [22], [34], [43]. We have learned that supervised neural networks usually adapt better to data with highly overlapping or nonlinear class borders than statistical regression does [29], [37].

Standard statistical regression and standard neural network techniques like backpropagation do not explain their classification results by rules. Particularly physicians as a main interest group are interested in such rules to get insight in the classification process, e.g. to draw conclusions for therapy. Thus, scientists have developed methods that allow the generation of rules within the classification process or the extraction of rules after the completed classification process. In Section 2.1, we give a short overview of such alternatives. The subsequent layout of the paper is described below.

Since our main goal is the application of a knowledge based method to septic shock patient data, we choose one of the algorithms that are introduced in Section 2.1 for this task, discussing our choice. The main ideas of the chosen algorithm are described in Section 2.2. During the experimental phase we realized that the algorithm could be improved, regarding the overlapping behavior of the neurons and the shrink-mechanism [28], see Section 2.3.

We repeat all the experiments with random partitions of the data into training and test data to get meaningful, statistically reasonable results. We evaluate our classification results with standard performance measures, e.g. classification error on training and test datasets. Each one of the rules is evaluated with a frequency and confidence measure (Section 3.1). Another question concerning rules is the global importance of each variable. We propose an importance measure for this feature selection task in Section 3.2.

Before applying the improved network to our septic shock patient data, we present some results on well-known benchmark datasets in Section 4 to point to our improvements and to clarify the general usefulness of our new measure “importance.”

Septic shock is one of the most common reasons of death in intensive care units (ICUs). Of course, this is reason enough to explore the causes in detail. Our analysis is restricted to abdominal intensive care patients who developed a septic shock during their stay at the ICU. The abdominal septic shock has a high mortality rate in the ICU of up to 50%. Some more details are described in Section 5.1. Our analysis is retrospectively based on a medical database. Thus, preprocessing steps for data quality improvement are necessary (Section 5.2). We present all the results concerning our septic shock patient data in Section 5.3. Some interesting insights are found compared to a mere classification procedure. Finally, we discuss our results in Section 6. We find out that it is not possible to clearly reduce the same dimensions for all the rules if we use septic shock data sampled from the entire time series. But we can generate several performant rules with less dimensions that give insight in the data.

Section snippets

Rule generation for metric data

In principle there are two kinds of medical data: metric (numerical) data (blood pressure, heart frequency, doses of medicaments, etc.) and categorical (symbolic) data (operations, diagnoses, therapies) including binary data (yes/no) or medical codes such as ICD10 or OP301 [42]. Metric data could further be divided into biosignal data (e.g. EEG, MEG), sampled with an adequate sampling rate and measurement data from patient records, recorded by physicians irregularly whenever they considered it

Performance measures

To evaluate the performance of our classification and rule generation results, we reproduce relevant performance measures that are commonly used. A new method to rate the global importance of variables for feature selection (dimension reduction) is introduced. Our aim is to take into account the specific rule structure of the neuro-fuzzy rule set. We do not consider more general or more arbitrary, ad hoc defined rule interestingness measures [15], [46].

An important aspect concerning neural

Application to benchmark datasets

Before applying the algorithm (MGT) to our septic shock patient data, we test it on benchmark datasets. The spiral data serve as a good (two-dimensional) example of nonlinear data to visualize the generated rules by (ALG) and (MGT). “Cancer1” [36] serves as a medical benchmark dataset to evaluate our importance measure.

Application to septic shock patient data

In this section, we review shortly the septic shock problem in intensive care medicine. Then, we describe our preprocessing steps for preparing the data for analysis. In fact, preprocessing of multivariate time series with missing values—the usual case in medical databases—is very time consuming although very important [40]. Finally, we present our results.

Discussion and conclusion

We have presented our data analysis approach for the important medical problem septic shock with an emphasis on rule generation for metric data. The results are a major extension of preliminary work (preprocessing, classification) [11], [33], now providing us with understandable knowledge for classification.

We have reviewed and improved the algorithm [17] with regard to the overlapping behavior and the shrinking mechanism to generate more performant rules. Our results on benchmark data

Acknowledgements

The work was done within the project MEDAN http://www.medan.de (Ref. no. HA 1456/7-2), supported by the German Research Foundation (DFG). The author thanks all the participants of the MEDAN working group especially Dr. Brause and Prof. Hanisch for supporting my work.

References (47)

W. Baxt
Application of artificial neural networks to clinical medicine
Lancet
(1995)
M.R. Berthold et al.
Constructive training of probabilistic neural networks
Neurocomputing
(1998)
D. Nauck et al.
Obtaining interpretable fuzzy classification rules from medical data
Artif Intell Med
(1999)
E. Pesonen et al.
Treatment of missing data values in a neural network based decision support system for acute abdominal pain
Artif Intell Med
(1998)
M. Schumacher et al.
Neural networks and logistic regression: part I
Comput Stat Data Anal
(1996)
Agrawal R, Skrikant R. Fast algorithms for mining association rules. In: Bocca J, Jarke M, Zaniolo C., editors....
Berthold MR. Fuzzy-models and potential outliers. In: Dave RN, Sudkamp TA, editors. Proceedings of the 18th...
Berthold MR, Diamond J. Boosting the performance of RBF networks with dynamic decay adjustment. In: Tesauro G,...
Berthold MR, Huber K-P. From radial to rectangular basis functions: a new approach for rule learning from large...
Brause R, Friedrich F. A neuro-fuzzy approach as medical diagnostic interface. In: Verleysen M, editor. Proceedings of...

Castellano G, Fanelli AM. Modeling fuzzy classification systems with compact rule base. In: Mohammadian M, editor....

Fein AM, et al., editors. Sepsis and multiorgan failure. Baltimore: Lippincott Williams & Wilkins;...

Fritzke B. Incremental neuro-fuzzy systems. In: Bosacchi B, Bezdek JC, Fogel DB, editors. Proceedings of the...

Hamker F, Paetz J, Thöne S, Brause R, Hanisch E. Erkennung kritischer Zustände von Patienten mit der Diagnose...

Hanisch E, Encke A. Intensive care management in abdominal surgical patients with septic complications. In: Faist E,...

R.M. Hardaway

A review of septic shock

Am Surg

(2000)

Haykin S. Neural networks: a comprehensive foundation. 2nd ed. Upper Saddle River: Prentice-Hall;...

Hilderman RJ, Hamilton H. Knowledge discovery and interestingness measures: a survey. Technical Report CS 99-04....

Huber K-P. Datenbasierte Metamodellierung mit automatisch erzeugten Fuzzy-Regeln. VDI-Verlag, Doctoral thesis,...

Huber K-P, Berthold MR. Building precise classifiers with automatic rule extraction. In: Proceedings of the IEEE...

J.-S.R. Jang

ANFIS: adaptive-network-based fuzzy inference system

IEEE Trans Syst—Man and Cybernet

(1993)

J.-S.R. Jang et al.

Functional equivalence between radial basis function networks and fuzzy inference systems

IEEE Trans Neur Networks

(1993)

Jolliffe IT, Principal component analysis. New York: Springer;...

Cited by (38)

Septic shock prediction for ICU patients via coupled HMM walking on sequential contrast patterns
2017, Journal of Biomedical Informatics
Citation Excerpt :
Decision trees were also employed by Gwadry-Sridhar et al. [50] for 20 clinical variables, achieving nearly 100% predictive accuracy. Among the soft computing techniques, numerous wrapper-based feature selection and preprocessing methods, namely Zero-Order-Hold, and missing-value imputation techniques have been employed along with particle swarm optimization, fuzzy models, and neural networks, to improve septic shock classification performance [51,49,55,56,58]. Selecting appropriate clinical features turns out to an important concern for predicting cases of septic shock.
Critical care patient events like sepsis or septic shock in intensive care units (ICUs) are dangerous complications which can cause multiple organ failures and eventual death. Preventive prediction of such events will allow clinicians to stage effective interventions for averting these critical complications.
It is widely understood that physiological conditions of patients on variables such as blood pressure and heart rate are suggestive to gradual changes over a certain period of time, prior to the occurrence of a septic shock. This work investigates the performance of a novel machine learning approach for the early prediction of septic shock. The approach combines highly informative sequential patterns extracted from multiple physiological variables and captures the interactions among these patterns via coupled hidden Markov models (CHMM). In particular, the patterns are extracted from three non-invasive waveform measurements: the mean arterial pressure levels, the heart rates and respiratory rates of septic shock patients from a large clinical ICU dataset called MIMIC-II.
For baseline estimations, SVM and HMM models on the continuous time series data for the given patients, using MAP (mean arterial pressure), HR (heart rate), and RR (respiratory rate) are employed. Single channel patterns based HMM (SCP-HMM) and multi-channel patterns based coupled HMM (MCP-HMM) are compared against baseline models using 5-fold cross validation accuracies over multiple rounds. Particularly, the results of MCP-HMM are statistically significant having a p-value of 0.0014, in comparison to baseline models. Our experiments demonstrate a strong competitive accuracy in the prediction of septic shock, especially when the interactions between the multiple variables are coupled by the learning model.
It can be concluded that the novelty of the approach, stems from the integration of sequence-based physiological pattern markers with the sequential CHMM model to learn dynamic physiological behavior, as well as from the coupling of such patterns to build powerful risk stratification models for septic shock patients.
Missing data in medical databases: Impute, delete or classify?
2013, Artificial Intelligence in Medicine
Citation Excerpt :
This information is new knowledge and we believe will substantively inform study designs of this type. In the final stage following classification, missing segments that are classified as sampling related are imputed by averaging the existing samples on each side of the data gap of the same variable, as in [13]. As previously detailed, our aim is to determine if the true missing data are recoverable or not-recoverable and for this purpose we use fuzzy models.
The multiplicity of information sources for data acquisition in modern intensive care units (ICUs) makes the resulting databases particularly susceptible to missing data. Missing data can significantly affect the performance of predictive risk modeling, an important technique for developing medical guidelines. The two most commonly used strategies for managing missing data are to impute or delete values, and the former can cause bias, while the later can cause both bias and loss of statistical power.
In this paper we present a new approach for managing missing data in ICU databases in order to improve overall modeling performance.
We use a statistical classifier followed by fuzzy modeling to more accurately determine which missing data should be imputed and which should not. We firstly develop a simulation test bed to evaluate performance, and then translate that knowledge using exactly the same database as previously published work by [13].
In this work, test beds resulted in datasets with missing data ranging 10–50%. Using this new approach to missing data we are able to significantly improve modeling performance parameters such as accuracy of classifications by an 11%, sensitivity by 13%, and specificity by 10%, including also area under the receiver–operator curve (AUC) improvement of up to 13%.
In this work, we improve modeling performance in a simulated test bed, and then confirm improved performance replicating previously published work by using the proposed approach for missing data classification. We offer this new method to other researchers who wish to improve predictive risk modeling performance in the ICU through advanced missing data management.
Modified binary PSO for feature selection using SVM applied to mortality prediction of septic patients
2013, Applied Soft Computing Journal
Citation Excerpt :
A comparison was made between the studied wrapper methods and other FS methods already applied to the problem of sepsis outcome prediction. These wrapper approaches are: neuro-fuzzy modeling using only the 12 feature dataset [47], bottom-up (BU) using fuzzy modeling and neural networks [48] and ant feature selection (AFS) using fuzzy modeling and neural networks [48]. The results are depicted in Tables 6 and 7 for the 12 and 28 features datasets respectively.
This paper proposes a modified binary particle swarm optimization (MBPSO) method for feature selection with the simultaneous optimization of SVM kernel parameter setting, applied to mortality prediction in septic patients. An enhanced version of binary particle swarm optimization, designed to cope with premature convergence of the BPSO algorithm is proposed. MBPSO control the swarm variability using the velocity and the similarity between best swarm solutions. This paper uses support vector machines in a wrapper approach, where the kernel parameters are optimized at the same time. The approach is applied to predict the outcome (survived or deceased) of patients with septic shock. Further, MBPSO is tested in several benchmark datasets and is compared with other PSO based algorithms and genetic algorithms (GA). The experimental results showed that the proposed approach can correctly select the discriminating input features and also achieve high classification accuracy, specially when compared to other PSO based algorithms. When compared to GA, MBPSO is similar in terms of accuracy, but the subset solutions have less selected features.
Multi-stage modeling using fuzzy multi-criteria feature selection to improve survival prediction of ICU septic shock patients
2012, Expert Systems with Applications
In many binary medical classification problems, the cost of misclassifying one category is higher than the other, and in these applications it is desirable to employ a classifier with selective sensitivity or specificity. This work explores the utility of a fuzzy multi–criteria function for performance evaluation during knowledge–based medical classification and prediction. The method presented here uses fuzzy optimization to combine the sensitivity, specificity, and accuracy of classification as goals in a single objective function. This approach is used to assign flexible goals, which can be used to maximize the outcome in terms of each one of the goals. The proposed approach significantly increases the sensitivity and the specificity while maintaining or increasing accuracy. The versatility of the method is further exploited in a multi-model approach, using individual structures of multi-objective optimization of sensitivity and specificity separately, and then combining their outcomes through a decision-making module. Among various medical benefits derived from applying this technique, the divergent feature sets selected by high sensitivity and specificity models lend insight into factors more integrally connected to what causes risk of death for patients.
An intelligent model for the classification of children's occupational therapy problems
2012, Expert Systems with Applications
In Taiwan, the classification of real problems of children with appropriate occupational therapy is a difficult job for the therapist. The complexities of 127 attribute values to be evaluated in the assessment, the misleading diagnosis which may be made by the pediatrician and the shortage of manpower cause of high workload for the therapist. The design of an easy to use and effective classification model is therefore an important issue in children’s occupational therapy treatment. This study accordingly applies an artificial neural network (ANN) and classification and regression tree (CART) techniques to skeleton an intelligent classification model in order to provide a comprehensive framework to assist the therapist to raise the accuracy when categorizing children’s problems for occupational therapy. These categories with critical attributes under the guidelines of the American Occupational Therapy Association (AOTA) are discussed, in order to assist the therapist for precise assessment and appropriate treatment. To the best of our knowledge, no research has yet been conducted on the problems’ characteristics in children’s occupational therapy.
Based on the advice and assistance of the therapists and occupational therapy treatment needed, 127 outpatients from a regional hospital in Taiwan between 2007 and 2010 were selected as the data sets for problems in children occupation classification. This study accordingly suggests an intelligent model for the classification which integrates ANN and CART. The major steps in applying the model include: (1) building an ANN higher performance trained model; and (2) adopting CART to the trained model and building in previous steps, to extract the critical attributes of children occupational problems.
The results showed that artificial neural network had a higher accuracy, up to 84%, with evenly distributed datasets. Then high performance of the trained neural network had been extracted for the rules by using the classification tree approach in the classification and regression trees application. Most important of all, this study indicated that some of the rules can correctly identify up to 67% of the problems of the children with 100% confidence, which is much better than the current evaluations being used. Moreover, the tree with a binary variable of age and 8 predicators were found and listed afterward, such as, gross coordination, upper left muscle tone, interpersonal skill, proprioceptive and vestibular, visual, visual stimulus input for influence of emotional and movement, swallowing, and dressing. Actual implementation showed that the intelligent classification model is capable of integrating ANN and CART techniques to clarify children’s occupational therapy problems with considerable accuracy.
The model could be employed as a supporting system in making decisions regarding children problems with occupational therapy classifications and treatment. The rules extracted from CART were helpful to therapists in classifying what category the real problems of the children belonged to. This study expected that more machine learning techniques will certainly play an essential role in future children occupational therapy applications.
Preprocessing of clinical databases to improve classification accuracy of patient diagnosis
2011, IFAC Proceedings Volumes (IFAC-PapersOnline)
In this paper, the prime importance of preprocessing in clinical databases is discussed. Specifically in intensive care units, data is often irregularly recorded, contain a large amount of missing values and sampling times are uneven. This paper proposes a systematic preprocessing procedure that can be generalized to common clinical databases. This procedure is applied to a known septic shock patient database and classification results are compared with previous studies. The goal is to estimate, as accurately as possible, the outcome (survived or deceased) of these septic shock patients. Neural modeling is used for classification. Detailed classification results are presented and show that the preprocessing is crucial to improve classifiers accuracy.

View all citing articles on Scopus

View full text