Clinical guidelines for using two dichotomous tests

FFTrees: A toolbox to create, visualize, and evaluate fast-and-frugal decision trees

Fast-and-frugal trees (FFTs) are simple algorithms that facilitate efficient and accurate decisions based on limited information. But despite their successful use in many applied domains, there is no widely available toolbox that allows anyone to easily create, visualize, and evaluate FFTs. We fill this gap by introducing the R package FFTrees. In this paper, we explain how FFTs work, introduce a new class of algorithms called fan for constructing FFTs, and provide a tutorial for using the FFTrees package. We then conduct a simulation across ten real-world datasets to test how well FFTs created by FFTrees can predict data. Simulation results show that FFTs created by FFTrees can predict data as well as popular classification algorithms such as regression and random forests, while remaining simple enough for anyone to understand and use.

Temporomandibular joint internal derangement: detection with 12.5 MHz ultrasonography

The purpose of this study was to determine whether 12.5 MHz ultrasonography (US) could be used to assess the presence or absence of temporomandibular joint (TMJ) internal derangement (ID). In 48 consecutive patients with TMJ disorders 192 TMJ positions were investigated by US to analyse the functional disc-condyle relationship (DCR). In order to compare the respective findings with those of a diagnostic method offering a high accuracy, coronal and sagittal magnetic resonance imaging was carried out immediately afterwards. With US showing a sensitivity of 0.58 and 0.75, and a specificity of 0.92 and 0.84 for disc displacement with and without reduction, the data revealed US to be marginal in detecting the presence, but sensitive in detecting the absence of the respective types of a TMJ ID. In addition, with a positive predictive value of 0.83 and 0.71, and a negative predictive value of 0.81 and 0.87 for disc displacement with and without reduction, the results indicate that US may be insufficient in establishing a correct diagnosis for the presence or absence of the respective types of TMJ ID. Regarding the diagnosis of absence or presence of TMJ ID, the results demonstrate high-resolution (HR)-US to be sensitive in detecting the absence, and reliable in predicting the presence of TMJ ID. In view of the fact that the 12.5 MHz US technique proved to be a reliable diagnostic aid for the detection of normal, and the prediction of abnormal DCR, the results of this study should be of further interest and encourage research in its potential uses and diagnostic capabilities.

Temporomandibular joint osteoarthrosis diagnosed with high resolution ultrasonography versus magnetic resonance imaging: how reliable is high resolution ultrasonography?

The purpose of this study was to determine whether 12.5 MHz ultrasonography (US) could be used to assess the presence or absence of temporomandibular joint (TMJ) condylar osteoarthrosis. In 40 consecutive patients with TMJ disorders, 80 TMJs were investigated by US to analyse the condylar morphology. In order to compare the respective findings with those of a diagnostic method offering a high accuracy, coronal and sagittal magnetic resonance imaging was carried out immediately afterwards. With US showing a sensitivity of 87%, and a specificity of 20%, the data revealed US to be sensitive in detecting the presence, but insufficient in detecting the absence of osteoarthrosis. In addition, with a positive predictive value of 88%, and a negative predictive value of 18%, the results indicate that US may be valuable in diagnosing the presence, but insufficient in diagnosing the absence of osteoarthrosis. In view of the fact that the 12.5 MHz US technique proved to be a valuable diagnostic aid for the detection and prediction of abnormal condylar morphology, the results of this study should be of further interest and encourage research in its potential uses and diagnostic capabilities.

Validation of the clinical diagnostic criteria for temporomandibular disorders for the diagnostic subgroup of degenerative joint disease

Research is needed to assess the validity of the Clinical Diagnostic Criteria for Temporomandibular Disorders (CDC/TMD). The purpose of this study was to test the reliability of the clinical diagnosis of temporomandibular joint (TMJ) degenerative joint disease (DJD) as compared with the magnetic resonance imaging (MRI) 'gold standard'. The TMJ DJD group comprised 48 joints in 24 consecutive patients who were assigned a clinical bilateral diagnosis of TMJ DJD. The TMJ non-DJD group consisted of 82 joints in 41 consecutive patients without a TMJ-related diagnosis of TMD. Bilateral sagittal and coronal MR images were obtained subsequently to establish the corresponding diagnosis of degenerative joint changes. An MRI diagnosis of osteoarthrosis (OA) was defined by the presence of flattening, subchondral sclerosis, surface irregularities, and erosion of the condyle or presence of condylar deformities associated with flattening, subchondral sclerosis, surface irregularities, erosion and osteophyte. For the CDC/TMD interpretations, the positive predictive of DJD for OA was 67%, and for the presence of degenerative joint changes 88%. The overall diagnostic agreement for DJD was 44.6% with a corresponding K-value of 0.01. Most of the disagreement was due to false-negative interpretations of asymptomatic joints. The results suggest CDC/TMD to be predictive for degenerative joint changes but insufficient for determination of OA. Patients assigned a clinical TMJ-related diagnosis of DJD may need to be supplemented by evidence from MRI to determine the presence or absence of OA.

Are high-resolution ultrasonographic signs of disc displacement valid?

PURPOSE

The study goal was to define the positive predictive value (PPV) for the diagnosis of disc displacement of the temporomandibular joint (TMJ) with the use of high-resolution ultrasonography (HR-US).

MATERIALS AND METHODS

The HR-US reports for 29 patients aged 19 to 62 years who had a prospective clinical diagnosis of TMJ internal derangement were correlated with magnetic resonance results.

RESULTS

For the prospective HR-US interpretation, the PPV for disc displacement was 97% at the closed-mouth position and 88% at the maximum-mouth opening position.

CONCLUSION

The PPV for the HR-US diagnosis of disc displacement is high.

Comparing methods for diagnosing temporomandibular joint disk displacement without reduction

BACKGROUND

Research is needed to assess the validity of the Clinical Diagnostic Criteria for Temporomandibular Disorders, or CDC/TMD. The authors conducted a prospective, double-blind study to determine whether applying the specific CDC/TMD diagnosis of TMJ internal derangement, or ID, type III would demonstrate good agreement with diagnoses obtained by MRI.

METHODS

The study comprised 138 TMJs in 69 subjects who had a clinical diagnosis of unilateral TMJ ID type III (disk displacement without reduction). The authors obtained bilateral sagittal and coronal MRIs to establish the corresponding diagnosis of disk-condyle relationship.

RESULTS

For the CDC/TMD interpretations, the positive predictive value of ID type III for disk displacement without reduction was 86 percent, and for the presence of an ID it was 91 percent. The overall diagnostic agreement for ID type III was 78.3 percent with a corresponding K value of 0.57. Most of the disagreement was due to false-positive interpretations of an absence of ID.

CONCLUSIONS

The results suggest that using CDC/TMD for ID type III is predictive for the presence of an ID but is not sufficiently reliable for determining disk displacement without reduction. Parameters other than the functional disk-condyle relationship may need to be addressed to account for the biological plausibility of this entity.

CLINICAL IMPLICATIONS

A clinical TMJ-related diagnosis of ID type III may need to be supplemented by evidence from an MRI to determine the functional disk-condyle relationship. Investigation of longitudinal evidence, including risk factors, history and response to treatment, appears to be warranted.

Bayesian analysis and inference from QSAR predictive model results

QSAR models have been under development for decades but acceptance and utilization of model results have been slow, in part, because there is no widely accepted metric for assessing their reliability. We reapply a method commonly used in quantitative epidemiology and medical decision-making for evaluating the results of screening tests to assess reliability of a QSAR model. It quantifies the accuracy (expressed as sensitivity and specificity) of QSAR models as conditional probabilities of correct and incorrect classification of chemical characteristic, given a true characteristic. Using Bayes formula, these conditional probabilities are combined with prior information to generate a posterior distribution to determine the probability a specific chemical has a particular characteristic, given a model prediction. As an example, we apply this approach to evaluate the predictive reliability of a CATABOL model and base on it a "ready" and "not ready" biodegradability classification. Finally, we show how predictive capability of the model can be improved by sequential use of two models, the first one with high sensitivity and the second with high specificity.

Error patterns and observer variations in the high-resolution ultrasonography imaging evaluation of the disk position of the temporomandibular joint

OBJECTIVES

The purpose of this study was (1) to describe errors in the interpretation of high-resolution ultrasonography (HR-US) images of the temporomandibular joint (TMJ) and analyze the discrepancies between HR-US imaging and diagnostic magnetic resonance imaging findings, and (2) to measure interobserver and intraobserver variations with respect to reporting HR-US images of the TMJ.

STUDY DESIGN

The sonographic reports of 208 patients aged 13 to 78 years who had a prospective clinical diagnosis of TMJ internal derangement were correlated with magnetic resonance imaging results. Prospective and retrospective readings were used to classify types of disk displacement. Assessment of observer variation was based on a set of 200 HR-US images in 50 randomly selected patients. The images were interpreted independently by 2 trained radiologists at 2 sessions, 1 to 2 weeks apart.

RESULTS

Prospective interpretation of TMJ HR-US images at the closed-mouth position had a sensitivity of 80%, specificity of 87%, and accuracy of 82%. For images at the maximum mouth-opening position, the sensitivity was 68%, specificity 93%, and accuracy 82%. The number of diagnostic errors at retrospective analysis was 60.3% (n = 73) of the number at prospective analysis. The intraobserver agreement for reporting the disk position at the closed-mouth position was 93% and at the maximum mouth-opening position was 88%. The agreement between the 2 observers was 89% on the position of the disk at the closed-mouth position and 84% at the maximum mouth-opening position.

CONCLUSIONS

The high diagnostic accuracy, together with the low observer variation reported in this study, suggest that HR-US of the TMJ is a valuable imaging technique for assessment of disk displacement. Many errors that occur at prospective analysis didn't occur at retrospective analysis.

Clinical predictability of temporomandibular joint disc displacement

Single items from a typical clinical examination have proved disappointing in their predictive value for temporomandibular joint (TMJ) disc displacement. Only one criterion (the 12 o'clock) is used to diagnose normal disc position. According to this criterion, the posterior band of the disc should be located at the top of the condyle, at the 12 o'clock position. The purpose of this study was to determine which signs and symptoms provide a valid prediction of the condition of the joint based on 4 magnetic resonance imaging (MRI) criteria used to define normal disc position. Sagittal MRI and clinical findings of 137 temporomandibular disorder patients and 23 normal asymptomatic volunteers were used. Three calibrated and blinded observers interpreted the images. Disc position with the mouth closed was evaluated based on 4 MRI criteria: 12, 11, 10 o'clock, and the intermediate zone. Disc position with the mouth open was determined based on one criterion. It was considered normal if the intermediate zone of the disc was located between the condyle and the articular eminence. Joints were classified as normal or as having disc displacement with or without reduction. The sensitivity and specificity of multiple clinical parameters for predicting the condition of the joint established by each of these 4 gold-standard MRI criteria were then determined. Regarding disc displacement with reduction, significant differences were observed in the sensitivity and specificity of all of the clinical parameters used to predict the imaging diagnosis established by each of the criteria. Concerning disc displacement without reduction, no significant differences were observed. The intermediate zone criterion was the criterion that most accurately reflected the condition of the joint. The clinical predictability of the disorder diagnosed according to this criterion suggests that clinical findings alone are too often nonspecific as predictors of the imaging stage of disc displacement. However, we found that combining the most sensitive clinical items to predict the disorder and using an overall criterion for positivity to interpret the results led to an impressive increase in the specificity of the combination, enabling false-positive diagnoses to be excluded.

Genetic screening for reproductive planning: methodological and conceptual issues in policy analysis

OBJECTIVES

This paper explores several critical assumptions and methodological issues arising in cost-effectiveness analyses of genetic screening strategies in the reproductive setting.

METHODS

Seven issues that arose in the development of a decision analysis of alternative strategies for cystic fibrosis carrier screening are discussed. Each of these issues required a choice in technique.

RESULTS

The presentations of these analyses frequently mask underlying assumptions and methodological choices. Often there is no best choice. In the case of genetic screening in the reproductive setting, these underlying issues often touch on deeply felt human values.

CONCLUSIONS

Space limitations for published papers often preclude explaining such choices in detail; yet these decisions determine the way the results should be interpreted. Those who develop these analyses need to make sure that the implications of important assumptions are understood by the clinicians who will use them. At the same time, clinicians need to enhance their understanding of what these models truly mean and how they address underlying clinical, ethical, and economic issues.

Evaluation of medical technologies: a generalized ROC analysis

The author describes a new methodology to solve medical decision problems involving a choice between alternatives under conditions of risk and uncertainty when imperfect empirical information from diagnostic technologies is available. The main new concept is the generalized ROC (GROC) curve, which extends considerably the scope of analysis of the traditional ROC curve as well as the threshold approach to medical decision making. Both techniques become special cases of the new general approach. The author shows how to apply the technique and derive comprehensive clinical guidelines to solve the medical decision problem involving one patient in the most general situation. The larger problem of evaluating the performances of diagnostic technologies for a population of patients with varying prior probabilities of disease is then addressed. A new measure of performance is proposed that goes beyond the well-accepted area under the ROC curve index. This measure is used to compare different technologies for a population of patients, and a new methodology is proposed to carry out such comparisons.

Cost-effective diagnostic test sequencing

Proper diagnostic test selection and logical sequencing can minimize cost without compromising care. This study analyzes the logic for sequencing tests either in series (one after another) or in parallel (simultaneously). A model is created using 2 diagnostic tests for 2 diseases. Tests are assumed to have perfect performance characteristics. Factors involved in cost-effective test sequencing include cost of each test, test performance characteristics, per diem hospital charge, and the clinician's prior probability of the suspected diagnosis based on history, physical examination, and previous laboratory data. At a given point of clinical suspicion, cost-effective sequencing strategy should shift. In general, a cheap test can effectively be used in parallel even at low diagnostic probabilities; an expensive test should, in general, be used in series only when there is a high prior probability of diagnosis. High per diem hospitalization costs--such as for a patient in the intensive care unit--favor in parallel testing. Clinical acumen and suspicion of diagnosis (prior probability) primarily direct cost-effective diagnostic test sequencing. Decision analysis of diagnostic test sequencing represents a technique for making cost-effective decisions for sequencing and, thereby, minimizing costs while achieving optimal patient care.

Threshold analysis using diagnostic tests with multiple results

Clinical problems represented by decision trees can be analyzed in terms of the probability threshold model, which provides management recommendations based on the prior probability of disease, the test threshold, and the test-treatment threshold. As originally proposed, the threshold model assumes that diagnostic tests provide information about a single event that is relevant to the decision. For some problems, however, a diagnostic test may provide information about more than one such event (e.g., a computed tomography [CT] scan gives information about both mediastinal and hilar metastases in lung cancer). The authors extend the probability threshold model to cases in which a single test provides information about two events that are relevant to the decision. They derive four thresholds that determine the best strategy for any combination of test results. The approach is illustrated for the decision to use a CT scan to stage lung cancer. The analysis reveals that: 1) the range of prior probabilities for which testing is optimal increases; 2) for some prior probabilities only test results about one event are important; 3) for some prior probabilities test results about both events are important; and 4) failure to account fully for information provided by a test can lead to erroneous test and treatment recommendations.

Optimal selection of a battery of tests: a multiobjective optimization methodology

The general problem of selecting a battery of tests for diagnostic purposes is discussed and multiobjective optimization methodology is applied to solve it, with battery selection being based on performance indices such as sensitivity, specificity, and the cost of testing. For a battery of tests, the extended majority rule is developed and used to interpret the compound test results. The major advantage of the model developed in this paper is that it can generate a set of noninferior batteries without requiring the calculation of all possible combinations of tests. An example in which the method is applied to a real problem--the selection of short-term tests to detect the carcinogenicity of chemicals--is discussed.

The importance of considering single testing when two tests are available

When considering two dichotomous tests in combination for reaching a treatment decision, the choice between single and multiple testing depends, in part, on the pretest probability of disease. The authors show that two tests are never preferred to a single test for all disease probabilities, regardless of whether the two tests are performed in parallel or in series.

Understanding laboratory test results. Conditions for appropriate use of laboratory tests

The appropriate use of laboratory tests requires that valid, reliable, and reproducible data be obtained and that the clinician know both how to interpret the information provided by diagnostic tests and how to apply it to individual clinical situations. Appropriate interpretation and clinical use of diagnostic tests requires that clinicians understand the principles of laboratory testing, the information provided and not provided by laboratory tests, and how to evaluate the clinical efficacy of laboratory tests. As laboratory testing moves from regional and hospital laboratories into office laboratories, clinicians need to become more knowledgeable regarding the technical and quality control factors that affect diagnostic test accuracy. This will require the establishment of stronger, more effective links between clinicians and laboratory pathologists.

Evaluating rapid tests for streptococcal pharyngitis: the apparent accuracy of a diagnostic test when there are errors in the standard of comparison

The evaluation of new rapid antigen detection tests for streptococcal pharyngitis is complicated by the presence of errors in the standard of comparison. Even "gold standard" throat cultures are not perfectly accurate. This article is an analysis of how errors in the comparison cultures affect the apparent accuracy of new tests for streptococcal pharyngitis. One of the questions addressed is whether throat cultures performed in a physician's office are accurate enough to use as a standard in a clinical trial of a new rapid test. While the accuracy of office cultures is accepted by many to be adequate for diagnostic purposes, this analysis shows that when evaluating a new test stricter conditions must be placed on the accuracy of the cultures used for comparison.

Multiple test analyzer (MTA): a microcomputer program for determining preferred strategies with two diagnostic tests

There are ten distinct management strategies in clinical situations that involve two diagnostic tests with dichotomous outcomes. The authors describe a microcomputer program, based on a previously described model, that can be used to identify test and test-treatment thresholds and to compute preferred strategies. The program provides tables and graphs of the results, which can be viewed or printed, and there is an optimization routine that facilitates comprehensive analysis. It can be used by decision-analytic researchers and policy analysts, medical educators who teach decision analysis, and clinicians who use decision analysis in their practices.