SEARCH · PubMed Health

Results for “Decision Trees”

Explore indexed PubMed citations for clinical trials, systematic reviews and public health research. Read source abstracts and follow each citation to its original PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 73 records · Page 4Linked to original sources

PubMedDate not supplied

Screening blood donations for hepatitis C in Central Africa: analysis of a risk- and cost-based decision tree.

Four screening strategies (no testing, HC Abbott, HC Pasteur, and a combined test) for the detection of hepatitis C virus (HCV) antibody in donated blood were considered in a formal decision tree. Decision criteria included residual risk of infection and overall monetary cost. Tree parameters were determined using data from the Central African Republic. The prevalences observed among blood donors for HIV infection, hepatitis B, syphilis, and hepatitis C varied between 6% and 15%. The current residual risk of transfusion-transmitted infections is very high (8.4%). Screening for HCV would bring that risk down to about 3% with either the HC Pasteur, the HC Abbott, or the combined test. Even though baseline analysis gives preference to the HC Abbott test (the combined test coming out last), Monte Carlo sensitivity and uncertainty analyses showed that Abbott's and Pasteur's tests are interchangeable, on the basis or either risk or cost considerations.

Blood Donors↗

PubMed1998

Evaluation of a statistically derived decision tree for the cytodiagnosis of fine needle aspirates of the breast (FNAB).

A decision tree for the diagnosis of FNAB was derived from defined human observations using a rule induction method, C4.5 (a derivative of the ID3 algorithm). This algorithm is an implementation of the top-down induction method where the tree is determined iteratively by adding those nodes and branches which maximize the information gain at each step. The tree was derived from a training set of 200 FNAB with known outcome using 10 defined features (from one observer) and patient age. The tree contained a total of seven nodes (six observable features and patient age) with eight endpoints (four benign, four malignant). The tree was applied to a test set of 400 further FNAB with observations from the training observer and produced a sensitivity of 95%, specificity of 93% and a positive predictive value (PPV) of a malignant result of 89%. Four trainee pathologists were given a training session on the observable features and then used the tree to determine outcome in a further 50 FNAB. The observers were blind to clinical details apart from age and the endpoints were coded with letters and not labelled benign or malignant. The results from these observers produced ranges of sensitivity 80-96%, specificity 64-92%, PPV 73-92% and kappa statistics (with known outcome) 0.6-0.8. Reported difficulties in using the tree included estimation of nuclear size. These results were worse than the performance of the observers on a further 50 cases without using the decision tree (sensitivity 80-100%, specificity 72-100%, PPV 78-100%, kappa 0.72-0.92). The original 50 case test set was rerandomized and the four trainee observers made all 10 defined observations on each specimen without using the decision tree; these observations were then used to derive decisions from the tree. The performance from this method was similar to that using selected features from the tree, suggesting that observation of all features together does not improve the reliability of each specific observation. The poor performance of this tree suggests that this methodology may be unsuitable for producing decision support aids for diagnostic or training purposes in this domain.

Biopsy, Needle↗

PubMed2006

Suicide attempts among substance use disorder patients: an initial step toward a decision tree for suicide management.

BACKGROUND: Little empirical data are available to develop profiles of patients who attempt suicide or to formulate a decision tree for suicide management. This study identifies profiles of patients who have a high risk of suicide attempt and takes a first step toward developing a decision tree to classify high-risk patients. METHODS: Based on a cross-sectional, nationwide cohort of substance use disorder patients (N=34,251) in 150 Veterans Affairs (VA) facilities, a total of 5,671 patients who reported suicidal ideation in the 30 days before intake assessment were included in receiver operating characteristic (ROC) analyses to identify the 30-day risk of an actual suicide attempt. Clinical diagnostic and Addiction Severity Index interview data were used. RESULTS: Results provide an initial decision tree to classify high-risk patients with sensitivity ranging from 0.33 to 0.89, and specificity, from 0.42 to 0.87. The factors included in the decision tree encompass history of prior suicide attempts, current drinking to intoxication, current cocaine use, first occasion of suicidal ideation, and difficulty controlling violent behavior. CONCLUSIONS: To our knowledge, this is the first attempt to use empirical data to provide information to eventually establish a decision tree for clinical management of patients with suicidal ideation. The findings show that profiles of patients who are at high risk of suicide attempts can be effectively identified using ROC, with relatively good sensitivity and specificity.

Adult↗

PubMed1996

Effects of pruning of a decision-tree for the ear, nose, and throat realm in primary health care based on case-notes.

In the primary health care center of Mjölby a sample of case notes in the ear-nose- and-throat realm (N = 425) was computer processed using an inductive rule-based decision-tree generating program. As a result of incomplete information in the case-files, the decision trees were "noisy," e.g., had branches and leaves without meaning. This led to a need for "pruning." Various methods were tried. The effects of different methods of decision-tree generating and pruning are discussed. The choice of root argument and branching of the decision-trees suggested by the software was the most clinically applicable. The "statistic" approach to pruning gave the most compact and still most clinically relevant decision-tree. The pruned and edited decision trees are compared with a previously published preliminary essential data set for the ear-nose- and-throat realm in primary health care and then discussed as a possible decision support system for various primary health care groups in a practice setting.

Algorithms↗

PubMed2002

Decision trees: an overview and their use in medicine.

In medical decision making (classification, diagnosing, etc.) there are many situations where decision must be made effectively and reliably. Conceptual simple decision making models with the possibility of automatic learning are the most appropriate for performing such tasks. Decision trees are a reliable and effective decision making technique that provide high classification accuracy with a simple representation of gathered knowledge and they have been used in different areas of medical decision making. In the paper we present the basic characteristics of decision trees and the successful alternatives to the traditional induction approach with the emphasis on existing and possible future applications in medicine.

Clinical Medicine↗

PubMed2000

[A new decision tree method for statistical analysis of quantitative data obtained in toxicity studies on rodents].

Regarding the statistical analysis of the quantitative data obtained in control and dosage groups in toxicity studies, we tried to simplify the decision tree method. In a new decision tree presented in this article, one-way analysis of variance and Kruskal-Wallis nonparametric analysis of variance are excluded from the traditional decision tree: (1) Bartlett's test is used as a test for the equality of k variances: (2) Then, if the k sampled populations have equal variances (p > 0.05 by the Bartlett's test), Dunnett's multiple comparison test is performed: otherwise, Steel's test is used. This new method, which increases the power in some conditions, may serve as an alternative to the traditional decision tree method.

Analysis of Variance↗

PubMed2000

Risk of extreme events in multiobjective decision trees. Part 1. Severe events

Earlier work with decision trees identified nonseparability as an obstacle to minimizing the conditional expected value, a measure of the risk of extreme events, by the well-known method of averaging out and folding back. This first of two companion papers addresses the conditional expected value that is defined as the expected outcome assuming the exceedance of a threshold beta, where beta is preselected by the decision maker. An approach is proposed to overcome the need to evaluate all policies in order to identify the optimal policy. The approach is based on the insight that the conditional expected value is separable into two constituent elements of risk and can thus be optimized along with other objectives, including the unconditional expected value of the outcome, by using a multiobjective decision tree. An example of sequential decision making for improving highway capacity is given.

Journal Article↗

PubMed2000

Risk of extreme events in multiobjective decision trees. Part 2. Rare events

Earlier work with decision trees identified nonseparability as an obstacle to minimizing the conditional expected value, a measure of the risk of extreme events, by the well-known method of averaging out and folding back. This second of two companion papers addresses the conditional expected value that is defined as the expected outcome assuming that a random variable is observed only in the upper 100 (1 - alpha) percent of potential outcomes, where alpha is a cumulative probability preselected by the decision maker. An approach is proposed to overcome the need to evaluate all policies in order to identify the optimal policy. The approach is based in part on approximating the conditional expected value by using statistics of extremes. An existing convenient approximation of the conditional expected value is shown to be separable into two constituent elements of risk and can thus be optimized, along with other objectives including the unconditional expected value of the outcome, in a multiobjective decision tree. An example of sequential decision making for remediation or environmental contamination is provided. The importance of the results for risk analyis beyond the minimization of conditional expected values is pointed out.

Journal Article↗

PubMed2003

Cost-effective treatment for ureteropelvic junction obstruction: a decision tree analysis.

PURPOSE: We determined the optimal treatment for primary ureteropelvic junction obstruction based on cost using a decision tree model. MATERIALS AND METHODS: A comprehensive literature search for articles addressing surgical correction of ureteropelvic junction obstruction was performed and data were abstracted on operative time, hospital stay, complications and success rate. The overall cost and individual cost centers at our institution for antegrade endopyelotomy, retrograde ureteroscopic endopyelotomy, Acucise (Applied Medical Resources, Laguna Hills, California) endopyelotomy, laparoscopic pyeloplasty and open pyeloplasty were compared. A decision tree model estimated the cost of treatment and followup for each modality using commercially available software. Sensitivity analyses were performed to evaluate the effect of individual treatment variables on overall cost. RESULTS: Based on cost center review retrograde ureteroscopic endopyelotomy was the least costly procedure ($2,891). In the decision tree model the rank order of overall treatment costs was: retrograde ureteroscopic endopyelotomy ($3,842), Acucise endopyelotomy ($4,427), antegrade endopyelotomy ($5,297), laparoscopic pyeloplasty ($7,026) and open pyeloplasty ($7,119). Despite various hospital stay, operative time, equipment cost and success rate data 1-way sensitivity analysis revealed that antegrade endopyelotomy, laparoscopic pyeloplasty and open pyeloplasty were never cost effective compared with retrograde ureteroscopic endopyelotomy or Acucise endopyelotomy, while 2-way sensitivity analysis favored retrograde ureteroscopic endopyelotomy. CONCLUSIONS: Primary cost variables for ureteropelvic junction obstruction treatments include operative time, hospital stay, equipment cost and success rate. Decision tree analysis showed that retrograde ureteroscopic or Acucise endopyelotomy is the most cost-effective treatment modality at our institution. However, cost is only 1 of a number of factors that are considered when deciding on an optimal course of treatment.

Cost-Benefit Analysis↗

PubMed2005

An evaluation of the decision tree format of the American College of Rheumatology 1987 classification criteria for rheumatoid arthritis: performance over five years in a primary care-based prospective study.

OBJECTIVE: The American College of Rheumatology (ACR) 1987 criteria for rheumatoid arthritis (RA) can be applied in 2 formats, a standard "x/y" list and a decision tree. This study evaluated the performance of the decision tree compared with the list approach in the ascertainment of RA in subjects with new-onset inflammatory polyarthritis (IP) over the first 5 years of observation. Moreover, the use of clinical surrogates to substitute for missing rheumatoid factor (RF) and radiologic erosion data was assessed for validity and for its influence on the resulting RA prevalence estimates. METHODS: In this population-based prospective study, 848 subjects with new-onset IP were interviewed and examined at baseline, with followup at 1, 2, 3, and 5 years. RF and erosion status were determined at prespecified time points. The list criteria were applied cumulatively, while the decision tree was applied cross-sectionally using either data surrogates or the actual reported data. RA prevalence in the 848 subjects and agreement in classification between the 2 methods was assessed at each time point. The influence of using clinical surrogates on RA prevalence estimates at 5 years and the agreement between surrogate and real results were also analyzed. RESULTS: At baseline, RA prevalence was higher using the decision tree compared with the list approach (63% versus 47%; P = 0.0001); by 5 years of followup, RA estimates were approximately equal (69% versus 72%) and agreement between the approaches was good (kappa = 0.67). The use of surrogates had little influence on RA prevalence at 5 years, although substitution of metacarpophalangeal joint swelling for erosion produced a higher RA prevalence estimate (78% versus 70%). Although there was only weak agreement between surrogate and real data, the use of the surrogate data provided good to very good agreement between the approaches in categorizing subjects as RA positive (kappa = 0.61-0.72). CONCLUSION: Over 5 years, the 2 formats of the ACR criteria for RA performed similarly, with no important differences between them. The use of surrogates for missing radiologic and serologic data did not have any major influence on disease classification. Although the RA criteria were not originally derived from subjects with early disease in a population setting, this study shows that the use of the decision tree approach with the option of substituting clinical surrogates for missing laboratory data is an appropriate alternative to the conventional list approach.

Adult↗

PubMed2005

Automated decision tree classification of corneal shape.

PURPOSE: The volume and complexity of data produced during videokeratography examinations present a challenge of interpretation. As a consequence, results are often analyzed qualitatively by subjective pattern recognition or reduced to comparisons of summary indices. We describe the application of decision tree induction, an automated machine learning classification method, to discriminate between normal and keratoconic corneal shapes in an objective and quantitative way. We then compared this method with other known classification methods. METHODS: The corneal surface was modeled with a seventh-order Zernike polynomial for 132 normal eyes of 92 subjects and 112 eyes of 71 subjects diagnosed with keratoconus. A decision tree classifier was induced using the C4.5 algorithm, and its classification performance was compared with the modified Rabinowitz-McDonnell index, Schwiegerling's Z3 index (Z3), Keratoconus Prediction Index (KPI), KISA%, and Cone Location and Magnitude Index using recommended classification thresholds for each method. We also evaluated the area under the receiver operator characteristic (ROC) curve for each classification method. RESULTS: Our decision tree classifier performed equal to or better than the other classifiers tested: accuracy was 92% and the area under the ROC curve was 0.97. Our decision tree classifier reduced the information needed to distinguish between normal and keratoconus eyes using four of 36 Zernike polynomial coefficients. The four surface features selected as classification attributes by the decision tree method were inferior elevation, greater sagittal depth, oblique toricity, and trefoil. CONCLUSION: Automated decision tree classification of corneal shape through Zernike polynomials is an accurate quantitative method of classification that is interpretable and can be generated from any instrument platform capable of raw elevation data output. This method of pattern classification is extendable to other classification problems.

Cornea↗

PubMed2003

Assessment of the potential allergenicity of ice structuring protein type III HPLC 12 using the FAO/WHO 2001 decision tree for novel foods.

The introduction of novel proteins into foods carries a risk of eliciting allergic reactions in individuals sensitive to the introduced protein. Therefore, decision trees for evaluation of the risk have been developed, the latest being proposed by WHO/FAO early in 2001. Proteins developed using modern biotechnology and derived from fish are being considered for use in food and other applications, and since allergy to fish is well established, a potential risk from such proteins to susceptible human beings exists. The overall aim of the study was to investigate the potential allergenicity of an Ice Structuring Protein (ISP) originating from an arctic fish (the ocean pout, Macrozoarces americanus) using the newly developed decision tree proposed by FAO/WHO. The methods used were those proposed by FAO/WHO including amino acid sequence analysis for sequence similarity to known allergens, methods for assessing degradability under standardised conditions, assays for detection of specific IgE against the protein (Maxisorb RAST) and histamine release from human basophils. In the present paper we describe the serum screening phase of the study and discuss the overall application of the decision tree to the assessment of the potential allergenicity of ISP Type III. In an accompanying paper [Food Chem. Toxicol. 40 (2002) 965], we detail the specific methodology used for the sequence analysis and assessment of resistance to pepsin-catalysed proteolysis of this protein. The ISP showed no sequence similarity to known allergens nor was it stable to proteolytic degradation using standardised methods. Using sera from 20 patients with a well-documented clinical history of fish allergy, positive in skin prick tests to ocean pout, eel pout and eel were used, positive IgE-binding in vitro to extracts of the same fish was confirmed. The sera also elicited histamine release in vitro in the presence of the same extracts. The ISP was negative in all cases in the same experiments. Using the proposed decision tree, we demonstrated the safety of the ISP to patients already sensitised to fish, as well as to individuals potentially susceptible to producing IgE responses to proteins. Furthermore, the practicability of the new decision tree was confirmed.

Adolescent↗

PubMed2006

[Correlations between diagnostic information and therapeutic efficacy in rheumatoid arthritis analyzed with decision tree model].

OBJECTIVE: To explore the correlations between diagnostic information and therapeutic efficacy in rheumatoid arthritis (RA) with decision tree model analysis. METHODS: Three hundred and ninety seven patients came from 9 clinical centers were randomly divided into the Western medicine (WM) group (n=194) treated with non-steroidal anti-inflammatory drugs and slow-acting antirheumatic drug and the Chinese medicine (CM) group (n=203) with basic therapy and syndrome-differentiation dependant TCM treatment. TCM and WM diagnostic information were collected. The ACR 20 was used for efficacy evaluation and the information of patients before treatment was analyzed by SAS 8.2 statistical package. Through single-factor exploratory analysis, odds ratio of efficacy and variable was calculated taken P < 0.2 as the including criteria for data mining analysis with decision tree model. All data were classified into the training set (75%) and verifying set (25%) with efficacy as the variable for layering to make further verification of the data-mining analysis. RESULTS: Twenty variables were included in the CM group and 26 in the WM group in the data-mining model. In the former, 9 variables were positively correlated to the efficacy, including degree of arthralgia, tenderness and morning stiffness, number of swollen joint, and joint with tenderness, levels of IgM, rheumatoid factor (RF), C-reactive protein (CRP), and total assessment from doctor; and disease duration and degree of nocturnal polyuria were negatively correlated to that. While in the latter, 8 were positively correlated to the efficacy, including erythrocyte sedimentation rate (ESR), sour and weak waist and knees, white fur in tongue, joint ache and stiffness, swollen joint, and total assessment from doctor and patient, and red tongue with yellow fur and leucocyte count negatively correlated to it. Data mining with decision tree analysis revealed that different combinations of morning stiffness, slight red tongue, joint tenderness and nocturnal polyuria in the CM group, and those of white fur in tongue, CRP level, leucocyte count and morning stiffness in the WM group showed different efficacy, which were also verified in the randomly chosen verifying set. CONCLUSION: To analyze the correlations between diagnostic information and therapeutic efficacy with decision tree analysis is conformed to the theory of TCM in applying treatment according to syndrome differentiation individually, thus it would contribute to elevate the accuracy of therapy.

Adolescent↗

PubMed1995

Decision tree model describing alternate health care choices made by oncology patients.

What common sense decision strategies do patients with cancer use when they are making health care choices that include alternate therapies? Existing research indicates that oncology patients are making alternate choices while associated with biomedicine. Often patients' decision strategies are exploited by the alternate system to promote and market alternate products. Although some of these practices are benign, others are dangerous or may interfere or delay successful treatment in biomedicine. Therefore, it seems important for biomedical professionals to understand patients' common sense decision patterns. A decision tree model, outlining patients' decision strategies, has recently been developed through intensive interviews with 300 patients who were diagnosed with cancer of the respiratory and digestive systems. The two-phase methodology included, first, a context sensitive approach to develop the model, followed by a predictive approach testing the model developed in the first phase on a separate yet similar random sample of patients. The discussion in this article focuses on the research, the patterns of the decision tree model, and the implications and adaptability of this model to nursing practice.

Adult↗

PubMed1996

Decision tree sensitivity analysis for cost-effectiveness of FDG-PET in the staging and management of non-small-cell lung carcinoma.

UNLABELLED: Preliminary studies have shown that PET is more accurate than CT for the staging of non-small-cell lung carcinoma (NSCLC). However, the potential effect of PET on the management of these patients and its cost-effectiveness has not been rigorously studied. Thus, we have used decision tree sensitivity analysis to assess the cost-effectiveness of a PET based strategy for staging of NSCLC. METHODS: Two decision strategies for selection of potential surgical candidates were compared; thoracic CT alone or thoracic CT and thoracic PET. The first decision tree was conservatively constructed by requiring mediastinoscopy (biopsy) to confirm imaging results so that no patient with surgically curable disease would miss the opportunity for surgery in either strategy. A second less conservative tree in which only nonconcordant results are biopsied was also tested. The various paths of each strategy are dependent on numerous parameters which were determined from a review of the medical literature. Life expectancy was calculated using the declining exponential approximation of life expectancy and reduced based on procedural mortality. Costs were based on mean costs at our institution. For all possible outcomes of each strategy, the expected cost and projected life expectancy were determined. The effect of changing one or more parameters on the expected cost and life expectancy were studied using a sensitivity analysis. RESULTS: The CT + PET strategy in the conservative decision tree showed a saving of $1154 per patient without a loss of life expectancy (increase of 2.96 days) as compared to the alternate strategy of CT alone. Both these effects were the result of improved staging of lung carcinoma prior to the decision for surgery. The CT + PET strategy in the less conservative decision tree showed a savings of $2267 per patient but misses 1.7% of potentially operable patients. CONCLUSION: These results show through rigorous decision tree analysis, the potential cost-effectiveness of using FDG PET in the management of NSCLC. These results form a basis for detailed study of the results obtained from multicenter trials on the accuracy of PET in NSCLC management. Furthermore, the techniques utilized for decision tree analysis have broad range of applicability to the entire field of nuclear medicine.

Carcinoma, Non-Small-Cell Lung↗

PubMed2000

Predicting cesarean delivery with decision tree models.

OBJECTIVE: The purpose of this study was to determine whether decision tree-based methods can be used to predict cesarean delivery. STUDY DESIGN: This was a historical cohort study of women delivered of live-born singleton neonates in 1995 through 1997 (22,157). The frequency of cesarean delivery was 17%; 78 variables were used for analysis. Decision tree rule-based methods and logistic regression models were each applied to the same 50% of the sample to develop the predictive training models and these models were tested on the remaining 50%. RESULTS: Decision tree receiver operating characteristic curve areas were as follows: nulliparous, 0.82; parous, 0.93. Logistic receiver operating characteristic curve areas were as follows: nulliparous, 0.86; parous, 0.93. Decision tree methods and logistic regression methods used similar predictive variables; however, logistic methods required more variables and yielded less intelligible models. Among the 6 decision tree building methods tested, the strict minimum message length criterion yielded decision trees that were small yet accurate. Risk factor variables were identified in 676 nulliparous cesarean deliveries (69%) and 419 parous cesarean deliveries (47.6%). CONCLUSION: Decision tree models can be used to predict cesarean delivery. Models built with strict minimum message length decision trees have the following attributes: Their performance is comparable to that of logistic regression; they are small enough to be intelligible to physicians; they reveal causal dependencies among variables not detected by logistic regression; they can handle missing values more easily than can logistic methods; they predict cesarean deliveries that lack a categorized risk factor variable.

Adolescent↗

PubMed2000

Treatment of epileptic seizures as medical emergencies: a prospective analysis of a decision tree for nonmedically trained staff.

The aim of this study was to evaluate the benefit of treating every epileptic seizure as an emergency and to propose a decision tree for use by nonmedical staff. Published literature on emergency medical treatment for seizures was reviewed. A detailed decision tree was developed for use by nonmedical staff. Its effectiveness was prospectively analysed over 16 months. Each seizure was logged, and its duration and severity were recorded. Type, dose, and time of antiepileptic treatment were noted. Results were compared with a 6-month retrospective review. Initiation of antiepileptic drug treatment 5 or 7 minutes after seizure onset was indicated in the tree. Rectal diazepam or lorazepam, or fast-dissolving lorazepam was used. Use of the decision tree resulted in decreased duration of seizures, which lasted up to 18 hours in the retrospective review compared with only 37 minutes in the prospective study. Using this tree, nonmedical staff can appropriately treat adults experiencing seizures. Duration, severity, and frequency of seizures decreased when every seizure was treated as a medical emergency. Implementation of a decision tree is recommended for establishments with patients prone to seizures. It can also be taught to caregivers and parents of patients with well-known epilepsy.

Anticonvulsants↗

PubMed2006

Rule generation for protein secondary structure prediction with support vector machines and decision tree.

Support vector machines (SVMs) have shown strong generalization ability in a number of application areas, including protein structure prediction. However, the poor comprehensibility hinders the success of the SVM for protein structure prediction. The explanation of how a decision made is important for accepting the machine learning technology, especially for applications such as bioinformatics. The reasonable interpretation is not only useful to guide the "wet experiments," but also the extracted rules are helpful to integrate computational intelligence with symbolic AI systems for advanced deduction. On the other hand, a decision tree has good comprehensibility. In this paper, a novel approach to rule generation for protein secondary structure prediction by integrating merits of both the SVM and decision tree is presented. This approach combines the SVM with decision tree into a new algorithm called SVM_ DT, which proceeds in three steps. This algorithm first trains an SVM. Then, a new training set is generated through careful selection from the output of the SVM. Finally, the obtained training set is used to train a decision tree learning system and to extract the corresponding rule sets. The results of the experiments of protein secondary structure prediction on RS126 data set show that the comprehensibility of SVM_DT is much better than that of the SVM. Moreover, the generalization ability of SVM_DT is better than that of C4.5 decision trees and is similar to that of the SVM. Hence, SVM_DT can be used not only for prediction, but also for guiding biological experiments.

Algorithms↗