SEARCH · PubMed Health

Results for “algorithms”

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 289 records · Page 16Linked to original sources

PubMed1996

Thyroid function testing based on assay of thyroid-stimulating hormone: assessing an algorithm's reliability.

OBJECTIVE: To assess the ability of an algorithm for thyroid-function testing (based on assay of thyroid-stimulating hormone [TSH]) to discern euthyroidism in patients with and without conditions affecting thyroid function. DESIGN: The Australian Health Insurance Commission (HIC) specifies clinical categories for which Medicare rebate is given for assay of both TSH and free thyroxine (FT4), but otherwise rebates for thyroid function testing are given for TSH assay only. A prospective study was made of paired TSH and FT4 results of 1000 consecutive assays categorised by indication for testing. An FT4 value within the reference range was accepted as indicating euthyroidism; the reliability of an initial TSH measurement as the sole indicator of thyroid disease was assessed against this criterion standard. SETTING: A large suburban teaching hospital. OUTCOME MEASURE: Success or failure of the algorithm, with failure defined as an abnormal FT4 level missed because the TSH level was normal. RESULTS: The algorithm failure rate both overall and in the patients not in the HIC clinical categories was 2.7%, and there was no significant difference in algorithm failure rate in the patients in the various HIC clinical categories. The categories and failure rates were: patients being monitored for thyroid disease, 3.4%; patients with the "sick euthyroid" syndrome, nil; patients with psychosis or dementia, 1.1%; patients taking drugs affecting thyroid function, 2.1%; and patients with pituitary dysfunction, one of six cases. The range of FT4 values in patients in whom the algorithm failed was 6.4-29.5 pmol/L in those without thyroid disease and 3.4-27.4 pmol/L in those with thyroid disease. In patients being monitored for thyroid disease, the proportion of abnormal values of TSH alone was significant (P<0.001). CONCLUSION: We have shown that the HIC's imposition of a TSH-based algorithm by financial fiat is also scientifically acceptable. Use of this algorithm in hospitals (including psychiatric hospitals) will result in substantial savings.

Algorithms↗

PubMed1999

Mental health care from the public perspective: the Texas Medication Algorithm Project.

Medication treatment algorithms have been suggested as a strategy to provide uniform care at predictable costs. The Texas Medication Algorithm Project is a 3-phase study designed to provide solid data on the usefulness of medication algorithms. In phase 1, medication algorithms for the treatment of schizophrenia, major depressive disorder, and bipolar disorder were developed. Phase 2 was a feasibility study of these algorithms, and phase 3, now underway, compares the costs and outcome in 3 groups, one using a combination of an algorithm and patient/family education, a second using treatment as usual in a clinic that uses an algorithm for a different disorder, and a third using treatment as usual in a nonalgorithm clinic.

Algorithms↗

PubMed1999

Neural networks for visual field analysis: how do they compare with other algorithms?

PURPOSE: To compare the performance of a neural network in identifying visual field defects with the performance of other available algorithms. METHODS: A feed-forward neural network with a single hidden layer was trained to recognize visual field defects previously collected in a longitudinal follow-up glaucoma study, and then tested on fields taken from the same study but not used in the training. The receiver operating characteristics of the network then were compared with the previously determined performance of other algorithms on the same data set. RESULTS: At a specificity greater than 90%, the neural network was more sensitive than any of the available algorithms (although only the global indices were available for comparison, as the cluster and cross-meridional algorithms did not achieve such high specificity at their current settings). At a lower specificity (80-85%), the neural network was unable to attain the high sensitivity of the cluster or cross-meridional algorithms; in fact, the cluster algorithm from the Low-Tension Glaucoma study was significantly more sensitive. CONCLUSION: The receiver operating characteristics of a feed-forward neural network designed to detect visual field defects were explored. At a very high specificity (90-95%) a neural network performed better than the global indices. However, at a lower specificity (78%-88%), the neural network performed worse than cluster and cross-meridional algorithms.

Algorithms↗

PubMed1999

Comparison of clinical staging algorithms and 111indium-capromab pendetide immunoscintigraphy in the prediction of lymph node involvement in high risk prostate carcinoma patients.

BACKGROUND: The pretherapy prediction of occult lymph node involvement and the avoidance of otherwise futile and potentially morbid definitive local therapy is paramount in men with newly diagnosed prostate carcinoma. To identify patients with prostate carcinoma who likely have lymph node involvement and would benefit from staging lymphadenectomy prior to definitive local therapy, the authors compared the ability of several predictive staging algorithms and a radiolabeled monoclonal antibody scan to predict lymphatic metastases prior to treatment. METHODS: Between August 1991 and June 1994, 198 men with clinical T2 or T3 classified (TNM) prostate carcinoma (bone scan negative) who were at high risk of lymph node involvement underwent a 111In-capromab pendetide scan prior to staging lymphadenectomy. Several predictive models based on preoperative prostate specific antigen level, biopsy Gleason score, and clinical stage were selected to predict those men having a > or =20% probability of lymph node involvement. The ability to predict pathologic stage using several clinical algorithms and the monoclonal antibody scan was compared with pathologic examination of the lymph nodes. RESULTS: Overall, 39% of the pelvic lymph node specimens were positive for metastatic disease by pathologic analysis. Published algorithms predicting lymph node metastases had a positive predictive value (PPV) ranging from 40.5% to 46.6% and an area under the receiver operating characteristic curve (AUC) ranging from 0.52 to 0.61. The monoclonal antibody scan had a PPV of 66.7% and an AUC of 0.71. The differences between the PPV and the AUC for the individual clinical algorithms when compared with immunoscintigraphy were statistically significant. Combining the radiolabeled monoclonal antibody scan with clinical predictive models, a PPV of up to 72.1% could be obtained. CONCLUSIONS: These data suggest that the PPVs for the clinical predictive algorithms are similar and that the PPV of the radiolabeled monoclonal antibody scan alone or in combination with the algorithms has additional value in predicting lymph node involvement in prostate carcinoma patients at high risk of regional disease spread. These algorithms and the 111In-capromab pendetide scan may be used for the appropriate selection of candidates for definitive local therapy in men with clinically localized prostate carcinoma and significant risk of lymph node involvement.

Aged↗

PubMed1999

Leap-frog is a robust algorithm for training neural networks.

Optimization of perceptron neural network classifiers requires an optimization algorithm that is robust. In general, the best network is selected after a number of optimization trials. An effective optimization algorithm generates good weight-vector solutions in a few optimization trial runs owing to its inherent ability to escape local minima, where a less effective algorithm requires a larger number of trial runs. Repetitive training and testing is a tedious process, so that an effective algorithm is desirable to reduce training time and increase the quality of the set of available weight-vector solutions. We present leap-frog as a robust optimization algorithm for training neural networks. In this paper the dynamic principles of leap-frog are described together with experiments to show the ability of leap-frog to generate reliable weight-vector solutions. Performance histograms are used to compare leap-frog with a variable-metric method, a conjugate-gradient method with modified restarts, and a constrained-momentum-based algorithm. Results indicate that leap-frog performs better in terms of classification error than the remaining three algorithms on two distinctly different test problems.

Algorithms↗

PubMed1999

[Comparison of EyeSys videokeratoscope algorithms in the evaluation of idiopathic and postoperative astigmatism].

PURPOSE: To compare the accuracy and reproducibility of the Eye Sys videokeratoscope algorithms for analyzing idiopathic and surgery-induced astigmatism analysis. METHODS: Refractive astigmatism, videokeratoscopy (axial, tangential and refractive power), autorefractometry, autokeratometry, and keratometry were recorded in 20 patients with idiopathic astigmatism, 40 patients who had undergone cataract surgery and 40 patients who had undergone penetrating keratoplasty. For each eye, 2 successive videokeratoscopy were recorded. RESULTS: Both cylinder and axis provided by the tangential algorithm are significantly less reproducible than the cylinder and axis provided by the axial and refractive algorithms (P < 0.001). Cylinders provided by the axial and refractive algorithms showed a stronger correlation with subjective cylinder (rs > 0.89; p < 0.001) than the cylinder provided by the tangential algorithm (rs = 0.66; p < 0.001). Both keratometric axis and autokeratometric axis showed the strongest correlation with subjective axis (rs > 0.92; p < 0.001). The accuracy and reproducibility were higher for the topographic "bow tie" patterns than for the other topographic patterns. CONCLUSION: The axial and refractive algorithms of the Eye Sys videokeratoscope are more accurate and reproducible than the tangential algorithm for analyzing idiopathic or surgery-induced astigmatism.

Algorithms↗

PubMed2000

Biological sequence compression algorithms.

Today, more and more DNA sequences are becoming available. The information about DNA sequences are stored in molecular biology databases. The size and importance of these databases will be bigger and bigger in the future, therefore this information must be stored or communicated efficiently. Furthermore, sequence compression can be used to define similarities between biological sequences. The standard compression algorithms such as gzip or compress cannot compress DNA sequences, but only expand them in size. On the other hand, CTW (Context Tree Weighting Method) can compress DNA sequences less than two bits per symbol. These algorithms do not use special structures of biological sequences. Two characteristic structures of DNA sequences are known. One is called palindromes or reverse complements and the other structure is approximate repeats. Several specific algorithms for DNA sequences that use these structures can compress them less than two bits per symbol. In this paper, we improve the CTW so that characteristic structures of DNA sequences are available. Before encoding the next symbol, the algorithm searches an approximate repeat and palindrome using hash and dynamic programming. If there is a palindrome or an approximate repeat with enough length then our algorithm represents it with length and distance. By using this preprocessing, a new program achieves a little higher compression ratio than that of existing DNA-oriented compression algorithms. We also describe new compression algorithm for protein sequences.

Algorithms↗

PubMed1991

Evaluation of a comprehensive algorithm for blunt and penetrating thoracic and abdominal trauma.

The objective was to develop a single branched-chain decision tree for both blunt and penetrating thoracic and abdominal trauma and to test its feasibility to track clinical decisions. The algorithm consisted of 14 specific patient management loops and 31 decision nodes. During a 4-month period, the management decisions and clinical course of 434 trauma patients were prospectively observed. Thirty-four patients had no signs of life on arrival to the emergency department (ED) and were excluded from the statistical evaluation; the remaining 400 patients constituted the study group. The mean Injury Severity Score (ISS), Penetrating Abdominal Trauma Index (PATI), and Trauma Score (TS) scores in the series were 21 +/- 10, 34 +/- 12, and 13 +/- 3. The overall patient mortality of the study group was 17 per cent; it was 61 per cent in those patients with major deviations from the algorithm and 6 per cent in patients who complied with the algorithm. The ISS, PATI, and TS scores were 29 +/- 9, 32 +/- 12, and 13 +/- 2 in patients with deviations and 20 +/- 10, 37 +/- 12, and 14 +/- 2 in patients who complied with the algorithm. Of the 37 patients who died with major deviations from the algorithm, the deviation was directly contributory to death in 21 cases (57%) and probably contributory in another 14 cases (38%). There were 108 patients with ISS scores between 20 and 50. In this group, mortality was 55 per cent when a major deviation occurred and 5 per cent without major deviations from the algorithm. The authors conclude that the survival of trauma patients may be improved by following the specific management criteria outlined by the algorithm.

Abdominal Injuries↗

PubMed1989

A selective mapping algorithm for computer analysis of voided urine cell images.

One of the fundamental targets of the automated image analysis of cytologic preparations is the reduction of computer classification errors due to cells or other objects that do not lend themselves to image segmentation or that have morphologic features that may mislead the cell classification schemes. In prior work from this laboratory, the achievement of this goal was attempted by hierarchical analysis of sequential microscopic objects at high resolution. This paper reports on the successful development and implementation of an automated "selective mapping algorithm" that selects cells at low power for further analysis and eliminates a large proportion of unwanted "objects." The algorithm classifies the objects and extracts appropriate features from a 256 X 240 digital image obtained via a 10 X planachromatic objective. The five-node binary tree classifier used in this triage is described. The algorithm was trained and tested initially on 501 visually classified microscopic "objects," resulting in a correct acceptance rate of 61.3% and correct rejection rate of 81.3%. The selective mapping algorithm was subsequently integrated into the video-based image analysis system constructed at the Montefiore Medical Center for the diagnostic evaluation of sediments of voided urine. The algorithm was then tested on ten cytocentrifuge preparations for a preliminary evaluation of its performance. Up to 100 "objects" per case were selected by the algorithm for further classification by the computer at high power. Of the 810 "objects" selected by the selective mapping algorithm, 344 (42.5%) were classified by the computer at high resolution as cells of diagnostic value ("WELL" cells) and 466 were rejected.(ABSTRACT TRUNCATED AT 250 WORDS)

Algorithms↗

PubMed1996

Algorithms for verbal autopsies: a validation study in Kenyan children.

The verbal autopsy (VA) questionnaire is a widely used method for collecting information on cause-specific mortality where the medical certification of deaths in childhood is incomplete. This paper discusses review by physicians and expert algorithms as approaches to ascribing cause of deaths from the VA questionnaire and proposes an alternative, data-derived approach. In this validation study, the relatives of 295 children who had died in hospital were interviewed using a VA questionnaire. The children were assigned causes of death using data-derived algorithms obtained under logistic regression and using expert algorithms. For most causes of death, the data-derived algorithms and expert algorithms yielded similar levels of diagnostic accuracy. However, a data-derived algorithm for malaria gave a sensitivity of 71% (95% Cl: 58-84%), which was significantly higher than the sensitivity of 47% obtained under an expert algorithm. The need for exploring this and other ways in which the VA technique can be improved are discussed. The implications of less-than-perfect sensitivity and specificity are explored using numerical examples. Misclassification bias should be taken into consideration when planning and evaluating epidemiological studies.

Algorithms↗

PubMed1996

[Comparative study of 3 algorithms to localize the accessory pathway in Wolff-Parkinson-White syndrome].

BACKGROUND AND OBJECTIVES: Some electrocardiographic algorithms have been developed to predict the location of the accessory pathway in the WPW syndrome. Few studies address the interobserver variability of such algorithms and the possible observer-dependent changes of accuracy. This study analyzes three algorithms to localize accessory pathways recently published, comparing the inter-observer variability, their predictive value and the most frequent problems observed during their application. METHODS: Ninety-six electrocardiograms from patients who underwent successful ablation of a single accessory pathway were reviewed. The location of each pathway was predicted by two independent observers according to three different reported electrocardiographic algorithms. The interobserver agreement, percentage of correct predictions and critical steps of each algorithm were analyzed. RESULTS: The interobserver agreement varied between 64 and 79% and the accuracy between 38 and 67%. The best results were obtained in the left lateral accessory pathways (69 to 89% correctly located). All the algorithms presented critical steps at which more than 20% of pathways were incorrectly classified. CONCLUSIONS: The analyzed algorithms present a high interobserver variability. The accuracy obtained is clearly lower than that reported by the corresponding authors. These facts should be considered when being used them in clinical settings.

Adult↗

PubMed1996

Management of four arterial blood gas problems in adult mechanical ventilation: decision-making algorithms and rationale for their use.

Use of these algorithms does not eliminate the need to think. One must always evaluate each patient to determine if the algorithms are applicable. The algorithms provide a learning framework for any practitioner who is responsible for managing patients receiving mechanical ventilation. The effectiveness of any reference tool depends, to an extent, on the context in which it is applied. The use of these algorithms without an adequate understanding of the principles of gas exchange, acid-base balance, and the function of mechanical ventilators will probably not benefit the patient or the practitioner. The portable nature of these algorithms allows them to be used in the clinical setting. The ultimate goal, of course, is to replace the algorithms with the ability to make and justify mechanical ventilation decisions. Experience with these algorithms will also assist users in applying this approach with unfamiliar problems to find viable solutions.

Adult↗

PubMed1997

An evaluation of the application of the genetic algorithm to the problem of ordering genetic loci on human chromosomes using radiation hybrid data.

We consider the problem of ordering detectable genetic loci along a chromosome by minimizing the number of obligatory breaks that can be inferred from radiation hybrid data. The problem bears some resemblance to the travelling-salesman problem, for which genetic algorithms have been used with considerable success. We find that the results from other studies on closely related problems are not directly transferable, and although we did find a genetic algorithm that performed well in this application it would appear that this algorithm is highly sensitive to any changes in the problem. Moreover, a very simple stochastic algorithm performed almost as well as our much more complicated and computer-intensive genetic algorithm and it did so in a fraction of the time. While we do not dispute that genetic algorithms can work on large complicated problems, the various modifications and fine-tuning necessary for good performance tend to be highly problem specific and they are often only arrived at after an exhaustive exploration of possibilities. Thus, we would dispute any claim that genetic algorithms are robust in their form and range of applicability.

Algorithms↗

PubMed1998

Screening for sexually transmitted diseases in rural women in Papua New Guinea: are WHO therapeutic algorithms appropriate for case detection?

The presence of a large reservoir of untreated sexually transmitted diseases (STDs) in developing countries has prompted a number of suggestions for improving case detection, including the use of clinical algorithms and risk assessments to identify women likely to be infected when they present to clinics for other reasons. We used data from a community-based study of STDs to develop and evaluate algorithms for detection of cervical infection with Chlamydia trachomatis or Neisseria gonorrhoeae, and for detection of vaginal infection with Trichomonas vaginalis or bacterial vaginosis. The algorithms were derived using data from 192 randomly selected women, then evaluated on 200 self-selected women. We evaluated the WHO algorithm for vaginal discharge in both groups. The prevalences of cervical and vaginal infection in the randomly selected group were 27% and 50%, respectively, and 23% and 52%, respectively, in the self-selected group. The derived algorithms had high sensitivities in both groups, but poor specificities in the self-selected women, and the positive predictive values were unacceptably low. The WHO algorithms had extremely low sensitivity for detecting either vaginal or cervical infection because relatively few women reported vaginal discharge. Simple algorithms and risk assessments are not valid for case detection in this population.

Adolescent↗

PubMed1998

Deterministic annealing EM algorithm.

This paper presents a deterministic annealing EM (DAEM) algorithm for maximum likelihood estimation problems to overcome a local maxima problem associated with the conventional EM algorithm. In our approach, a new posterior parameterized by `temperature' is derived by using the principle of maximum entropy and is used for controlling the annealing process. In the DAEM algorithm, the EM process is reformulated as the problem of minimizing the thermodynamic free energy by using a statistical mechanics analogy. Since this minimization is deterministically performed at each temperature, the total search is executed far more efficiently than in the simulated annealing. Moreover, the derived DAEM algorithm, unlike the conventional EM algorithm, can obtain better estimates free of the initial parameter values. We also apply the DAEM algorithm to the training of probabilistic neural networks using mixture models to estimate the probability density and demonstrate the performance of the DAEM algorithm.

Journal Article↗

PubMed1998

A unified algorithm for principal and minor components extraction.

Principal component and minor component extractions provide powerful techniques in many information-processing fields. However, by conventional algorithms minor component extraction is much more difficult than principal component extraction. A unified algorithm which can be used to extract both principal and minor component eigenvectors is proposed. This 'unified' algorithm can extract true principle components (eigenvectors) and if altered simply by the sign, it can also serve as a true minor components extractor. This is of practical significance in neural network implementation. It is shown how the present algorithms are related to Oja's principal subspace algorithm, Xu's algorithm and the Brockett flow. It is also shown that the algorithms are based on the natural gradient ascend/descent methods (a potential flow in a Riemannian space).

Journal Article↗

PubMed1983

Clinical aspects of resuscitation with and without an algorithm: relative importance of various decisions.

Clinical description was made of a series of hypotensive patients resuscitated with and without an algorithm. Of 603 hypotensive patients, there were 114 (19%) deaths and 169 (28%) patients with complications; the average low MAP was 53 +/- 25 mm Hg. Of 169 patients with complications, 48 (28%) had shock-related (SR) complications; 25 (52%) of these patients died. There were 265 (44%) patients who had severe associated diseases and these patients comprised the group most vulnerable to complications and death; in this group, there were 41 patients with SR-complications who had significantly higher mortality, longer resuscitation times, lower MAP, more deviations from the algorithm, and more delays in resuscitation than did those with nonshock-related complications. Multiple deviations from the algorithm were associated with longer resuscitation times, and higher incidence of SR complications. Most of the delays in resuscitation of these patients and most of the SR complications could have been prevented. Seventeen percent of hypotensive patients entering the emergency department (ED) normally carried low arterial pressures, which averaged 75 +/- 3 (SD) mm Hg; this was more common, but not confined to young females. Of the 603 patients who actually were hypotensive, 6% were admitted in arrest (phase I), 18% in severe shock (phase II, MAP less than 60 mm Hg), 52% in moderate shock (phase III, MAP less than 80 mm Hg), and 24% were normotensive but subsequently became hypotensive (phase IV). The importance of various decision nodes of the algorithm were evaluated. The present algorithm, designed for these hypotensive emergency patients, provides a framework for fluid management that expedites resuscitation and reduces complications related to shock. We conclude that: (a) delays in resuscitation can be clearly related to an increased incidence of SR complications; (b) when the algorithm was satisfactorily followed, there was faster resuscitation and less SR complications; and (c) when the algorithm was satisfactorily followed in patients with severe associated illnesses, there was also shorter ICU stay, shorter hospitalization and decreased mortality.

Adolescent↗

PubMed2003

Directed geometrical worm algorithm applied to the quantum rotor model.

We discuss the implementation of a directed geometrical worm algorithm for the study of quantum link-current models. In this algorithm the Monte Carlo updates are made through the biased reptation of a worm through the lattice. A directed algorithm is an algorithm where, during the construction of the worm, the probability for erasing the immediately preceding part of the worm, when adding a new part, is minimal. We introduce a simple numerical procedure for minimizing this probability. The procedure only depends on appropriately defined local probabilities and should be generally applicable. Furthermore, we show how correlation functions C(r,tau) can be straightforwardly obtained from the probability of a worm to reach a site (r,tau) away from its starting point independent of whether or not a directed version of the algorithm is used. Detailed analytical proofs of the validity of the Monte Carlo algorithms are presented for both the directed and undirected geometrical worm algorithms. Results for autocorrelation times and Green's functions are presented for the quantum rotor model.

Journal Article↗