SEARCH · Search PubMed

Results for “Predictive”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

SourceResearch connectionFrom yearThrough year

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

At least 829 records · Page 46Linked to original sources

PubMed1996

Culture, age and gender: effects on quality of predicted self and colleague reactions.

Ethnocentrism on the part of health care workers has been documented in the literature and has led to misdiagnosis, mistreatment and undertreatment of culturally diverse individuals worldwide. Aversive Insidious Racism and Ingroup Favoritism theories were used as the guiding framework for this study. Two hundred and sixty-eight female nurses from a large, urban, multi-service hospital in the United States were surveyed to identify those psychosocial variables (age, gender and culture status of the client) which enhanced and/or inhibited their predicted reactions with clients and which have the power to contribute to unethical decision making and less than ethical client care. The findings of this study, which is the first to examine nurses' predicted self and colleague reactions to multiple client variables concurrently, demonstrated that Client Gender as a main effect was not significant in itself when examining self and colleague predictions. Client Age as a main effect was significant for self predictions, p < 0.006, and for colleague predictions, p < 0.000. Client Culture as a main effect was significant for self predictions, p < 0.001 and for colleague predictions, p < 0.001. Many two-way and three-way interaction effects were significant. Subjects consistently predicted more favorable self reactions than colleague reactions, supporting Aversive Insidious Racism theory. Study findings did not consistently support Ingroup Favoritism theory. Subjects did not predict most favorable reactions with Caucasian female clients.

Adult↗

PubMed1995

Muscle lines-of-action affect predicted forces in optimization-based spine muscle modeling.

This study describes the effects of varied torso muscle geometries commonly assumed in optimization-based muscle force prediction models. Specifically, the sensitivity of predicted muscle and spinal forces to assumed muscle lines-of-action (LOA) is systematically examined. The practical significance of varied muscle LOAs is addressed by determining the relative precision needed for individual muscle LOAs and assessing which muscles are more critical to accurate prediction of spinal forces. To perform this analysis a nonlinear optimization model was used to generate muscle force predictions during combined frontal and sagittal plane moment loadings with an assumed erect posture. The LOAs of the erector spinae, rectus abdominus, internal and external oblique, and latissimus dorsi were systematically varied in the frontal and sagittal planes over an anatomically feasible range. The results indicated that moderate changes in the assumed LOA could substantially alter the magnitudes of predicted muscle and spinal forces. The estimated activity level of a muscle, as well as the predicted active/silent state could be affected by the LOA of that muscle and others. The patterns of predicted muscle activity, with respect to load orientation, underwent only minor alterations with changing LOA. The relative activation of several muscles, however, was dependent on LOA, and frequently led to variations in predicted spinal compression (> 100 N change) and shear forces (> 50 N change). This dependence of estimated spinal forces on assumed muscle geometry was most pronounced for the obliques and minimal for the more vertically oriented muscles and when loads were sagittally symmetric. This study suggests that muscle LOAs are critical inputs when interpreting absolute muscle and spinal force values predicted by models of physical exertions.

Algorithms↗

PubMed2004

GPS: a novel group-based phosphorylation predicting and scoring method.

Protein phosphorylation is an important reversible post-translational modification of proteins, and it orchestrates a variety of cellular processes. Experimental identification of phosphorylation site is labor-intensive and often limited by the availability and optimization of enzymatic reaction. In silico prediction may facilitate the identification of potential phosphorylation sites with ease. Here we present a novel computational method named GPS: group-based phosphorylation site predicting and scoring platform. If two polypeptides differ by only two consecutive amino acids, in particular when the two different amino acids are a conserved pair, e.g., isoleucine (I) and valine (V), or serine (S) and threonine (T), we view these two polypeptides bearing similar 3D structures and biochemical properties. Based on this rationale, we formulated GPS that carries greater computational power with superior performance compared to two existing phosphorylation sites prediction systems, ScanSite 2.0 and PredPhospho. With database in public domain, GPS can predict substrate phosphorylation sites from 52 different protein kinase (PK) families while ScanSite 2.0 and PredPhospho offer at most 30 PK families. Using PKA as a model enzyme, we first compared prediction profiles from the GPS method with those from ScanSite 2.0 and PredPhospho. In addition, we chose an essential mitotic kinase Aurora-B as a model enzyme since ScanSite 2.0 and PredPhospho offer no prediction. However, GPS offers satisfactory sensitivity (94.44%) and specificity (97.14%). Finally, the accuracy of phosphorylation on MCAK predicted by GPS was validated by experimentation, in which six out of seven predicted potential phosphorylation sites on MCAK (Q91636) were experimentally verified. Taken together, we have generated a novel method to predict phosphorylation sites, which offers greater precision and computing power over ScanSite 2.0 and PredPhospho.

Algorithms↗

PubMed2006

A computerized method for accurately predicting fetal macrosomia up to 11 weeks before delivery.

OBJECTIVE: To improve the prediction of birth weight and fetal macrosomia by combining sonographically derived fetal biometric data with routinely recorded pregnancy-specific information. STUDY DESIGN: Retrospective data were obtained for 218 normal gravidas who had obstetrical ultrasonography performed within 11 weeks of delivery. Multiple regression was employed to derive a set of equations for predicting birth weight that used different combinations of ultrasonographic and pregnancy-specific variables. RESULTS: A set of 38 unique combination equations was derived to accurately predict birth weight up to 11 weeks before delivery. The equations use different combinations of ultrasonographic and pregnancy-specific variables, so that predictions are still possible in the face of missing data. When ultrasonographic measurements are taken within 3 weeks of delivery, fetal macrosomia is predicted with 75% sensitivity, 93% specificity, and 67% and 95% positive and negative predictive value, respectively. The equations are equally as accurate for primiparous and multiparous women from all racial groups. A jackknifing procedure was used to validate the predictive accuracy of the equations for use with new subjects. CONCLUSION: The combined approach of predicting fetal macrosomia using ultrasonographic fetal measurements and pregnancy-specific characteristics is superior to pre-existing approaches that rely on either method alone. The method can be used up to 11 weeks before delivery, allowing fetal macrosomia to be predicted reliably in low-risk populations sufficiently early for prospective clinical intervention to be undertaken.

Female↗

PubMed2003

A novel MHCp binding prediction model.

Many statistical and molecular mechanics models have been developed and tested for major histocompatibility complex peptide (MHCp) binding predictions during the last decade. The statistical model prediction using pooled peptide sequence data and three-dimensional modeling prediction by molecular mechanics calculations have been assessed for efficiency and human leukocyte antigen diversity coverage. We describe a novel predictive model using information gleaned from 29 human MHCp crystal structures. The validation for the new model is performed using four different sets of data: (1) MHCp crystal structures, (2) peptides with known IC(50) binding values, (3) peptides tested positive by tetramer staining, (4) peptides with known binding information at the MHCBN database. The model produces high prediction efficiencies (average 60 %) with good sensitivity (approximately 50%-73%) and specificity (52%-58%) values. The average positive predictive value of the model is 89%, while the average negative predictive value is only 18%. The efficiency is very high in predicting binders and very low in predicting nonbinders. This model is superior to many existing methods because of its potential application to any given MHC allele whose sequence is clearly defined.

Amino Acid Sequence↗

PubMed2006

Predictive data mining in clinical medicine: current issues and guidelines.

BACKGROUND: The widespread availability of new computational methods and tools for data analysis and predictive modeling requires medical informatics researchers and practitioners to systematically select the most appropriate strategy to cope with clinical prediction problems. In particular, the collection of methods known as 'data mining' offers methodological and technical solutions to deal with the analysis of medical data and construction of prediction models. A large variety of these methods requires general and simple guidelines that may help practitioners in the appropriate selection of data mining tools, construction and validation of predictive models, along with the dissemination of predictive models within clinical environments. PURPOSE: The goal of this review is to discuss the extent and role of the research area of predictive data mining and to propose a framework to cope with the problems of constructing, assessing and exploiting data mining models in clinical medicine. METHODS: We review the recent relevant work published in the area of predictive data mining in clinical medicine, highlighting critical issues and summarizing the approaches in a set of learned lessons. RESULTS: The paper provides a comprehensive review of the state of the art of predictive data mining in clinical medicine and gives guidelines to carry out data mining studies in this field. CONCLUSIONS: Predictive data mining is becoming an essential instrument for researchers and clinical practitioners in medicine. Understanding the main issues underlying these methods and the application of agreed and standardized procedures is mandatory for their deployment and the dissemination of results. Thanks to the integration of molecular and clinical data taking place within genomic medicine, the area has recently not only gained a fresh impulse but also a new set of complex problems it needs to address.

Clinical Medicine↗

PubMed2006

Validation and re-evaluation of a discriminant model predicting anatomic suitability for biventricular repair in neonates with aortic stenosis.

OBJECTIVES: The purpose of this study was to validate and re-evaluate our previously reported scoring systems for predicting optimal management in neonates with aortic stenosis (AS). BACKGROUND: In 1991, we reported a multivariate discriminant equation and an ordinal scoring system for predicting which neonates with AS are suitable for biventricular repair and which are better served by single ventricle management. METHODS: Retrospective analysis was performed to: 1) validate our scoring systems in 89 additional neonates with AS and normal mitral valve area, 2) assess the effects of 5% measurement variation on predictive scores, 3) evaluate our cohort with the Congenital Heart Surgeons' Society scoring system, and 4) repeat the discriminant analysis on the basis of all 126 patients. RESULTS: The original scores each predicted outcome accurately in 68 patients (77%). Minor (5%) measurement variation changed the outcome predicted by the discriminant equation in 8 patients (9%) and by the threshold system in 13 patients (15%). The most accurate model for predicting survival with a biventricular circulation among the full cohort is: 10.98 (body surface area) + 0.56 (aortic annulus z-score) + 5.89 (left ventricular to heart long-axis ratio) - 0.79 (grade 2 or 3 endocardial fibroelastosis) - 6.78. With a cutoff of -0.65, outcome was predicted accurately in 90% of patients. CONCLUSIONS: Both of our original scoring systems are less accurate at predicting outcome than in our original analysis. Revised discriminant analysis yielded a model similar to our original equation that was 90% accurate at predicting survival with a biventricular circulation among neonates with AS and a mitral valve area z-score >-2.

Aortic Valve Stenosis↗

PubMed2004

ProMate: a structure based prediction program to identify the location of protein-protein binding sites.

Is the whole protein surface available for interaction with other proteins, or are specific sites pre-assigned according to their biophysical and structural character? And if so, is it possible to predict the location of the binding site from the surface properties? These questions are answered quantitatively by probing the surfaces of proteins using spheres of radius of 10 A on a database (DB) of 57 unique, non-homologous proteins involved in heteromeric, transient protein-protein interactions for which the structures of both the unbound and bound states were determined. In structural terms, we found the binding site to have a preference for beta-sheets and for relatively long non-structured chains, but not for alpha-helices. Chemically, aromatic side-chains show a clear preference for binding sites. While the hydrophobic and polar content of the interface is similar to the rest of the surface, hydrophobic and polar residues tend to cluster in interfaces. In the crystal, the binding site has more bound water molecules surrounding it, and a lower B-factor already in the unbound protein. The same biophysical properties were found to hold for the unbound and bound DBs. All the significant interface properties were combined into ProMate, an interface prediction program. This was followed by an optimization step to choose the best combination of properties, as many of them are correlated. During optimization and prediction, the tested proteins were not used for data collection, to avoid over-fitting. The prediction algorithm is fully automated, and is used to predict the location of potential binding sites on unbound proteins with known structures. The algorithm is able to successfully predict the location of the interface for about 70% of the proteins. The success rate of the predictor was equal whether applied on the unbound DB or on the disjoint bound DB. A prediction is assumed correct if over half of the predicted continuous interface patch is indeed interface. The ability to predict the location of protein-protein interfaces has far reaching implications both towards our understanding of specificity and kinetics of binding, as well as in assisting in the analysis of the proteome.

Algorithms↗

PubMed2005

Assessing the reliability of a QSAR model's predictions.

Quantitative structure activity relationships (QSAR) are one of the well-developed areas in computational chemistry. In this field, many successful predictive models have been developed for various property, activity or toxicity predictions. However, the predictive power of models for new query compounds is often not well characterized. The breadth of applicability of models is often not characterized. In other words, with a given QSAR model and a specific query compound to be predicted, can the model be used reliably for the desired prediction? In this study, we assessed the reliability of QSAR models' prediction on query compounds. Our approach, employing hierarchical clustering, was developed and tested using a test dataset containing 322 organic compounds with fathead minnow acute aquatic toxicity as the activity of interest. The hypothesis of the approach was that if a query compound is more similar to the compounds used to generate the QSAR model, it should be predicted more accurately. Thus, the core of the approach is to determine the relationship between the similarity of query compounds to the training set compounds of the QSAR model and the prediction accuracy given by that model. This relationship determination was achieved by comparing the results given by the two major components of the approach: objects clustering and activity prediction. With the resultant information from the two steps, a direct relationship was shown.

Models, Chemical↗

PubMed2026

Predicting bloodstream infection by plasma cell-free metagenomic sequencing: a prospective cohort study.

BACKGROUND: Patients receiving myelosuppressive chemotherapy or haematopoietic cell transplantation are at high risk for life-threatening bloodstream infections. A novel pre-emptive treatment paradigm guided by pathogen detection before symptoms appear might reduce this risk, but no validated screening test is available. This study evaluated the sensitivity and specificity of plasma microbial cell-free DNA metagenomic sequencing (mcfDNA-Seq) for predicting bloodstream infections in children and adolescents receiving therapy for high-risk leukaemia. METHODS: In this prospective cohort study, between Aug 9, 2017, and Feb 28, 2022, leftover clinical plasma samples were prospectively collected up to once per day from patients who were younger than 25 years, receiving care for leukaemia at St Jude Children's Research Hospital (Memphis, TN, USA), and at high risk for life-threatening bloodstream infections. mcfDNA-Seq was used to identify pathogen DNA in blood samples obtained during the 7 days before to 1 day after bloodstream infection onset, and in control samples from the same population in the absence of fever or infection. The testing laboratory was masked to sample status. Primary outcomes were predictive sensitivity of mcfDNA-Seq for detecting the expected bloodstream infection pathogen during the 3 days preceding the day of bloodstream infection onset, with a prespecified favourable sensitivity of 50%, and predictive specificity of mcfDNA-Seq in control samples. Exploratory analyses comprised assessing sensitivity and specificity restricted to bacteria or common bloodstream infection pathogens, and after applying a data-derived DNA fragment concentration cutoff; estimating the predictive sensitivity on each of the 7 days before bloodstream infection onset; identifying clinical characteristics that affected predictive sensitivity or specificity; and examining the clinical relevance of additional organisms identified by mcfDNA-Seq during bloodstream infection episodes. Diagnostic sensitivity was also assessed on samples collected on the day of, or day after, diagnosis of bloodstream infection. This study is registered with ClinicalTrials.gov, NCT03226158. FINDINGS: 94 evaluable bloodstream infections occurred in 60 (38%) of 158 enrolled participants; 19 episodes were previously described in the pilot phase of this study. The predictive sensitivity of mcfDNA-Seq was 51·9% (95% CI 40·5-63·1) for all bloodstream infection episodes, 53·8% (42·2-65·2) for bacterial infection only, and 51·9% (40·5-63·1) when applying a DNA fragment concentration cutoff of 140 molecules per μL. Sensitivity was lowest at day -7 and increased daily until the day of diagnosis. Diagnostic sensitivity was 81·3% (95% CI 71·0-89·1) for all bloodstream infection episodes and 83·1% (72·9-90·7) for bacterial infections only. Predictive specificity was 82·7% (95% CI 76·0-88·2), but improved to 88·9% (83·0-93·3) for common bloodstream infection pathogens, and to 93·8% (88·9-97·0) when also applying the DNA fragment concentration cutoff. Predictive sensitivity was higher in participants with acute lymphoblastic leukaemia (adjusted odds ratio [aOR] 11·1 [1·7-74·2] vs those with acute myeloid leukaemia), and it was lower in polymicrobial infections (aOR 0·0 [0·0-0·2] vs monomicrobial Gram-positive infections). Clinical false-positive results were positively associated with gastrointestinal disturbance alone (p=0·037) or combined with recent administration of high-dose cytarabine (p=0·012). Additional organisms identified by mcfDNA-Seq that were not identified by blood culture were less likely than expected organisms to have an increasing DNA concentration during the days preceding bloodstream infection diagnosis. INTERPRETATION: mcfDNA-Seq can detect causative pathogens before the onset of some bloodstream infection episodes in profoundly immunocompromised patients. Predictive specificity might be improved by restricting results to a subgroup of relevant organisms, excluding patients with high risk of false-positive results, or applying a higher concentration cutoff. Clinical trials are needed to evaluate mcfDNA-Seq-guided pre-emptive therapy for preventing life-threatening bloodstream infections in patients with high risk. FUNDING: The National Cancer Institute, American Lebanese Syrian Associated Charities, St Jude Children's Research Hospital, and Karius.

Adolescent↗

PubMed2006

Predicting the impact of population level risk reduction in cardio-vascular disease and stroke on acute hospital admission rates over a 5 year period--a pilot study.

UNLABELLED: The brief for this study was to produce a practical, evidence based, financial planning tool, which could be used to present an economic argument for funding a public health-based prevention programme in coronary heart disease (CHD) related illness on the same basis as treatment interventions. OBJECTIVES: To explore the possibility of using multivariate risk prediction equations, derived from the Framingham and other studies, to estimate how many people in a population are likely to be admitted to hospital in the next 5-10 years with cardio vascular disease (CVD) related events such as heart attacks, strokes, heart failure and kidney disease. To estimate the potential financial impact of reductions in hospital admissions, on an 'invest to save' basis, if primary care trusts (PCTs) were to invest in public health based interventions to reduce cardiovascular risk at a population level. STUDY DESIGN: The populations of five UK PCTs were entered into a spreadsheet based decision tree model, in terms of age and sex (this equated to around 620,000 adults). An estimation was made to determine how many people, in each age group, were likely to be diabetic. Population risk factors such as smoking rates, mean body mass index (BMI), mean total cholesterol and mean systolic blood pressure were entered by age group. The spreadsheet then used a variant of the Framingham equation to calculate how many non-diabetic people in each age group were likely to have a heart attack or stroke in the next 5 years. In addition heart failure and dialysis admission rates were estimated based upon risk factors for incidence. The United Kingdom Prospective Diabetes Study (UKPDS) risk engines 56 and 60 were used to calculate the risk of CHD and stroke, respectively, in people with type 2 diabetes. The spreadsheet deducted the number of people likely to die before reaching hospital and produced a predicted number of hospital admissions for each category over a 5-year period. The final part of the calculation attached a cost to the hospital activity using the UK Health Resource Grouping (HRG) tariffs. The predicted number of events in each of the primary care trusts was then compared with the actual number of events the previous year (2004/2005). METHODOLOGY: The study used a decision tree type model, which was populated with data from the research literature. The model applied the risk equations to population data from five primary care trusts to estimate how many people would suffer from an acute CVD related event over the next 5 years. The predicted number of events was then compared with the actual number of acute admissions for heart attacks, strokes, heart failure, acute hypoglycaemic attacks, renal failure and coronary bypass surgery the previous year. RESULTS: The first outcome of the model was to compare the estimated number of people in each PCT likely to suffer from a heart attack, a stroke, heart failure or chronic kidney failure with the actual number the previous year 2004/2005. The predicted number was remarkably accurate in the case of heart attack and stroke. There was some over-prediction of chronic kidney disease (CKD) which could be accounted for by known under-diagnosis in this illness group and the inability of the model to pick up, at this stage, the fact that many CKD patients die of a CHD related event before they reach the stage of requiring renal replacement. The second outcome of the model was to estimate the financial consequence of risk reduction. Moderate reductions in risk in the order of around 2-4% were estimated to lead to saving in acute admission costs or around pounds sterling 5.4 million over 5 years. More ambitious targets of risk reduction in the order of 5-6% led to estimated savings of around pounds sterling 8.7 million. CONCLUSIONS: This study is not presented as the definitive approach to predicting the economic consequences of investment in public health on the cost of secondary care. It is simply a logical, systematic approach to quantifying these issues in order to present a business case for such investment. The research team do not know if the predicted savings would accrue from such investments; it is theoretical at this stage. The point is, however, that if the predictions are correct then the savings will accrue from over 4000 people, from an adult population of around 185,000 not having a heart attack or a stroke or an acute exacerbation of heart failure.

Adolescent↗

PubMed1995

Accuracy of equations to predict basal metabolic rate in older women.

OBJECTIVE: To assess the accuracy of several published equations for predicting basal metabolic rate (BMR) in older women. DESIGN: BMR was assessed in 116 healthy, older white women, aged 60 to 82 years, on three successive mornings by indirect calorimetry. Body composition was determined by dual energy X-ray absorptiometry or hydrostatic weighing. The measured BMRs were compared with values obtained from eight published prediction equations that used solely, or in various combinations, measures of height, weight, fat-free mass, age, and menopausal status. STATISTICAL ANALYSES PERFORMED: The root mean squared prediction error (RMSPE) was used to determine how accurately predicted BMR matched actual BMR for each subject. In addition, regression analysis was used to evaluate accuracy of predicted BMR vs directly measured BMR. RESULTS: Predicted mean BMR determined using all eight equations was significantly correlated to measured BMR (P = .0001), accounting for 30% to 52% of the variance of measured BMR. When analyzed by RMSPE, however, the equations of Owen et al (1986), Fredrix et al (1990), and Harris-Benedict (1919) predicted actual BMR for each subject within an average of 116 kcal/day, and the equation of Cunningham (1980) resulted in the largest prediction error at 208 kcal/day. APPLICATIONS/CONCLUSIONS: The regression equations of Owen et al (1986), which used body weight, Fredrix et al (1990), which used body weight and age, and Harris-Benedict (1919), which used age, weight, and height as variables, were most accurate in predicting BMR in our sample of healthy older women.

Absorptiometry, Photon↗

PubMed1997

Chronic low back pain: predictions of pain and relationship to anxiety and avoidance.

The present study examined the discrepancy between prediction and experience of pain in 20 chronic low back pain (CLBP) patients undertaking a series of standard physical exercises. Subjects gave ratings of predicted pain and anxiety before each exercise and a rating of the pain experienced during the exercise. Contrary to prediction, the majority of CLBP patients predicted less pain than they subsequently experienced. Moreover, multiple regression analysis showed that, in these patients, the greater the discrepancy between predicted and experienced pain the greater the increase in pain predicted for the subsequent exercise. Anxiety was greater the higher the level of pain predicted, although it was not significantly related to the amount of pain experienced on the previous exercise or to the discrepancy between predicted and experienced pain. However, the discrepancy between predicted and experienced pain was found to be related to subsequent physical performance whereas the other variables were not; the more that pain exceeded expectations on the first exercise the less patients performed on the subsequent exercises. The results of this study are compared with those of previous studies and the implications of the findings to the maintenance of CLBP are discussed.

Adult↗

PubMed1997

Prediction of secondary structures of proteins using the sequence and spectroscopical data.

An algorithm is presented which modifies the parameter of given methods of prediction of secondary protein structures by comparing the predictions with the frequency of secondary structures derived from infrared spectra in a way that the predictions align to the given data. Depending on the prediction method and accuracy of the given secondary structures a 1-6% increase in accuracy can be reached. The algorithm compares the difference between the predicted and real frequency of a given secondary structure in the protein and modifies accordingly the parameter used in the prediction method in order to give a new, more accurate prediction. A correlation between the accuracy of the prediction and increasing correctness between the prediction and infrared data was found using a set of 39 proteins.

Algorithms↗

PubMed2000

Surgeon predictions on growth of minimal invasive therapy: the difficulty of estimating technologic diffusion.

OBJECTIVE: To compare five-year predictions made in 1992 by academic surgeon leaders in UK, US and Canada, with actual experiences in 1997, of increased rates of minimal invasive therapy (MIT) for surgical operations. METHOD: We compared 1992 predictions of percent of operations done by minimal invasive therapy and length of stay in the US with actual 1997 percents found by literature searches. RESULTS: We found sufficient data on 12 operations done by MIT in 1997 of the original 34 operations predicted in 1992 by surgeon experts to be to be amenable to this technique. These 12 operations were among the top 20 most commonly performed procedures in 1992 and 1997. Of these 12 operations, ten had 40-60% lower 1997 percentages than predicted, one had about 10% lower rate, and two had 18% and 100% higher rates of MIT than predicted. Overall mean length of stay (LOS) for all 34 study operations fell from 6.8 days in 1992 to 5.2 days in 1997. Mean LOS in 1997 was 2.5 days by MIT and 6.7 days by open technique (OT). CONCLUSION: Most of the predictions made in 1992 by surgical leaders in Canada, US and UK were incorrect when examined 5 years later. The rate of MIT diffusion and its effect on length of stay were overestimated for most operations, while for two procedures the predictions underestimated extent of diffusion. Also, much of the declines of LOS for surgical care paralleled declines in length of stay for all care, supplemented by the individual contributions of MIT specifically. Relying on expert opinion alone to predict the acceptability, rapidity, scope and extent of technological change is fraught with uncertainty. Unexpected consequences occur when one or a few parts of complex systems are changed. This is a particular problem when predictions are a main basis for informed decision making in the absence of any supporting data from appropriately designed empirical or controlled study.

Attitude of Health Personnel↗

PubMed2004

The second National Toxicology Program comparative exercise on the prediction of rodent carcinogenicity: definitive results.

Chemical carcinogenicity has been the target of a large array of attempts to create alternative predictive models, ranging from short-term biological assays (e.g. mutagenicity tests) to theoretical models. Among the theoretical models, the application of the science of structure-activity relationships (SAR) has earned special prominence. A crucial element is the independent evaluation of the predictive ability. In the past decade, there have been two fundamental comparative exercises on the prediction of chemical carcinogenicity, held under the aegis to the US National Toxicology Program (NTP). In both exercises, the predictions were published before the animal data were known, thus using a most stringent criterion of predictivity. We analyzed the results of the first comparative exercise in a previous paper [Mutat. Res. 387 (1997) 35]; here, we present the complete results of the second exercise, and we analyze and compare the prediction sets. The range of accuracy values was quite large: the systems that performed best in this prediction exercise were in the range 60-65% accuracy. They included various human experts approaches (e.g. Oncologic) and biologically based approaches (e.g. the experimental transformation assay in Syrian hamster embryo (SHE) cells). The main difficulty for the structure-activity relationship-based approaches was the discrimination between real carcinogens, and non-carcinogens containing structural alerts (SA) for genotoxic carcinogenicity. It is shown that the use of quantitative structure-activity relationship models, when possible, can contribute to overcome the above problem. Overall, given the uncertainty linked to the predictions, the predictions for the individual chemicals cannot be taken at face value; however, the general level of knowledge available today (especially for genotoxic carcinogens) allows qualified human experts to operate a very efficient priority setting of large sets of chemicals.

Animals↗

PubMed2002

Artificial neural network predictive model for allergic disease using single nucleotide polymorphisms data.

The purpose of this study was to develop a novel diagnostic prediction method for allergic diseases from the data of single nucleotide polymorphisms (SNPs) using an artificial neural network (ANN). We applied the prediction method to four allergic diseases, such as atopic dermatitis (AD), allergic conjunctivitis (AC), allergic rhinitis (AR) and bronchial asthma (BA), and verified its predictive ability. Almost all the learning data were precisely predicted. Regarding the evaluation data, the learned ANN model could correctly predict a diagnosis with more than 78% accuracy. We also analyzed the SNP data using multiple regression analysis (MRA). Using the MRA model, less than 10% of patients with the above allergic diseases were correctly diagnosed, while this figure was more than 75% for persons without allergic diseases. From these results, it was shown that the ANN model was superior to the MRA model with respect to predictive ability of allergic diseases. Moreover, we used two different methods to convert the genetic polymorphism data into numerical data. Using both methods, diagnostic predictions were quite precise and almost the same predictive abilities were observed. This is the first study showing the application and usefulness of an ANN for the prediction of allergic diseases based on SNP data.

Journal Article↗

PubMed1994

Prediction and site-specific mutagenesis of residues in transmembrane alpha-helices of proton-pumping nicotinamide nucleotide transhydrogenases from Escherichia coli and bovine heart mitochondria.

Nicotinamide nucleotide transhydrogenase from bovine heart consists of a single polypeptide of 109 kD. The complete gene for this transhydrogenase was constructed, and the protein primary structure was determined from the cDNA. As compared to the previously published sequences of partially overlapping clones, three residues differed: Ala591 (previously Phe), Val777 (previously Glu), and Ala782 (previously Arg). The Escherichia coli transhydrogenase consists of an alpha subunit of 52 kD and a beta subunit of 48 kD. Alignment of the protein primary structure of the bovine trashydrogenase with that of the transhydrogenase from E. coli showed an identity of 52%, indicating similarly folded structures. Prediction of transmembrane-spanning alpha-helices, obtained by applying several prediction algorithms to the primary structures of the revised bovine heart and E. coli transhydrogenases, yielded a model containing 10 transmembrane alpha-helices in both transhydrogenases. In E. coli transhydrogenase, four predicted alpha-helices were located in the alpha subunit and six alpha-helices were located in the beta subunit. Various conserved amino acid residues of the E. coli transhydrogenase located in or close to predicted transmembrane alpha-helixes were replaced by site-specific mutagenesis. Conserved negatively charged residues in predicted transmembrane alpha-helices possibly participating in proton translocation were identified as beta Glu82 (Asp655 in the bovine enzyme) and beta Asp213 (asp787 in the bovine enzyme) located close to the predicted alpha-helices 7 and 9 of the beta subunit. beta Glu82 was replaced by Lys or Gln and beta Asp213 by Asn or His. However, the catalytic as well as the proton pumping activity was retained. In contrast, mutagenesis of the conserved beta His91 residue (His664 in the bovine enzyme) to Ser, Thr, and Cys gave an essentially inactive enzyme. Mutation of alpha His450 (corresponding to His481 in the bovine enzyme) to Thr greatly lowered catalytic activity without abolishing proton pumping. Since no other conserved acidic or basic residues were predicted in transmembrane alpha-helices regardless of the prediction algorithm used, proton translocation by transhydrogenase was concluded to involve a basic rather than an acidic residue. The only conserved cysteine residue, beta Cys260 (Cys834 in the bovine enzyme), located in the predicted alpha-helix 10 of the E. coli transhydrogenase, previously suggested to function as a redox-active dithiol, proved not to be essential, suggesting that redox-active dithiols do not play a role in the mechanism of transhydrogenase.

Algorithms↗