SEARCH · PubMed Health

Results for “Model performance”

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 721 records · Page 40Linked to original sources

PubMed1999

The validity of medical history, classic symptoms, and chest radiographs in predicting pulmonary tuberculosis: derivation of a pulmonary tuberculosis prediction model.

STUDY OBJECTIVE: To improve the respiratory isolation policy for patients with suspected pulmonary tuberculosis (TB). DESIGN: Prospective, descriptive, French multicenter study. SETTING: Emergence of nosocomial outbreaks of TB. PATIENTS: All consecutive patients admitted with suspicion of pulmonary TB. MEASUREMENTS AND RESULTS: Medical history, social factors, symptoms, and chest radiograph (CXR) pattern (symptoms and CXR both scored as typical of pulmonary TB, compatible, negative, or atypical) were obtained on admission. Serial morning sputa were collected. Of the 211 patients, 47 (22.3%) had culture-proven pulmonary TB, including 31 (14.7%) with a positive smear. Mean age was 46.2 years; 52 patients were HIV positive (24.6%). The sensitivity of the respiratory isolation policy was 71.4%, specificity was 51.7%, negative predictive value (NPV) was 88.2%, and positive predictive value (PPV) was 26.3%. On univariate analysis, predictive factors of culture-proven pulmonary TB were CXR (p < 0.00001), symptoms (p = 0.0004), age (mean, 40.8 years for TB patients vs 47.5 years for non-TB patients; p = 0.04), absence of HIV infection (89.4% vs 71.3%; p = 0.01), immigrant status (72% vs 55%; p = 0.03), and bacillus Calmette-Guérin status (p = 0.025). On multivariate analysis, CXR pattern (p < 0.00001), HIV infection (p = 0.002), and symptoms (p = 0.009) remained independently predictive. Based on these data, a model was proposed using a receiver operating characteristics curve. In the derivation cohort, the sensitivity and NPV of the model in detecting smear-positive pulmonary TB would have been 100%. The specificity and PPV would have been 48.4% and 25%, respectively. The model performed less well when evaluated on two retrospective groups, but its sensitivity remained above that of the current respiratory isolation policy (91.1% and 82.4% for the retrospective groups vs 71.1% for the current policy). CONCLUSIONS: Improved interpretation of clinical and radiologic data available on patient admission could improve adequacy of respiratory isolation. A prediction model is proposed.

BCG Vaccine↗

PubMed2006

E-DIMEM: an economic model to estimate the costs of e-disease management.

In order to estimate the cost of generic clinical guidelines based on telemedicine services for chronic patients, an economic model (E-DIMEM) has been defined. This model is designed to help health suppliers and managers to plan for the future. It is essentially a workflow composed by diagnosis and therapeutic activities. The cost of each workflow activity is related to healthcare providers (activity agents), drugs, instruments, telemedicine services, and type of specialized healthcare centers used (hospital, clinics, etc.). This model performs multidimensional analysis. A web service based on the E-DIMEM has been implemented.

Clinical Protocols↗

PubMed2005

Comparison of the predictive validity of diagnosis-based risk adjusters for clinical outcomes.

OBJECTIVES: Many possible methods of risk adjustment exist, but there is a dearth of comparative data on their performance. We compared the predictive validity of 2 widely used methods (Diagnostic Cost Groups [DCGs] and Adjusted Clinical Groups [ACGs]) for 2 clinical outcomes using a large national sample of patients. METHODS: We studied all patients who used Veterans Health Administration (VA) medical services in fiscal year (FY) 2001 (n = 3,069,168) and assigned both a DCG and an ACG to each. We used logistic regression analyses to compare predictive ability for death or long-term care (LTC) hospitalization for age/gender models, DCG models, and ACG models. We also assessed the effect of adding age to the DCG and ACG models. RESULTS: Patients in the highest DCG categories, indicating higher severity of illness, were more likely to die or to require LTC hospitalization. Surprisingly, the age/gender model predicted death slightly more accurately than the ACG model (c-statistic of 0.710 versus 0.700, respectively). The addition of age to the ACG model improved the c-statistic to 0.768. The highest c-statistic for prediction of death was obtained with a DCG/age model (0.830). The lowest c-statistics were obtained for age/gender models for LTC hospitalization (c-statistic 0.593). The c-statistic for use of ACGs to predict LTC hospitalization was 0.783, and improved to 0.792 with the addition of age. The c-statistics for use of DCGs and DCG/age to predict LTC hospitalization were 0.885 and 0.890, respectively, indicating the best prediction. CONCLUSIONS: We found that risk adjusters based upon diagnoses predicted an increased likelihood of death or LTC hospitalization, exhibiting good predictive validity. In this comparative analysis using VA data, DCG models were generally superior to ACG models in predicting clinical outcomes, although ACG model performance was enhanced by the addition of age.

Adolescent↗

PubMed2004

ErbB-1 expression in experimental model of inflammatory bowel disease in rats.

BACKGROUND: Ulcerative colitis (UC) and Crohn's disease (CD) belong to inflammatory bowel disease (IBD). The etiology of IBD is still unknown. Therapy remains empiric or is used for the relief of specific symptoms. The erbB-1 oncogene coding epidermal growth factor receptor (EGFR) is typed as a prognostic marker in several benign and malignant tissues. The aim of our study was to examine the erbB-1 expression in experimental surgically performed model of IBD in rats and to find out if there is any correlation between severity of the intestinal inflammation and altered level of erbB-1 expression. After inducing an experimental colitis samples were taken from different parts of the intestine in all studied groups of rats for histopathology. ErbB-1 mRNA expression was estimated by RT-PCR. PCR products were separated on a 1.5% TBE-agarose gel and visualized with ethidium bromide. The integrated optical density (IOD) of electrophoretically separated amplification products was measured using a video densitometer and Gel-Pro 3.0 software. Nonparametrical statistical test has been used throughout in analyzing the results. Relative erbB-1 expression was determined by comparing to cyclophiline expression. RESULTS: Microscopic changes were similar to those observed in IBD. ErbB-1 expression was significantly higher in inflamed tissues of the bowel ( P = 0.04 for the transverse colon and P = 0.027 for the cecum). Significantly higher erbB-1 expression in inflamed tissues of the bowel suggests that EGFR overexpression may play a role in the pathogenesis of IBD. Overexpression of erbB-1 correlates with the severity of inflammation in bowel tissues.

Animals↗

PubMed2006

New probabilistic network models and algorithms for oncogenesis.

Chromosomal aberrations in solid tumors appear in complex patterns. It is important to understand how these patterns develop, the dynamics of the process, the temporal or even causal order between aberrations, and the involved pathways. Here we present network models for chromosomal aberrations and algorithms for training models based on observed data. Our models are generative probabilistic models that can be used to study dynamical aspects of chromosomal evolution in cancer cells. They are well suited for a graphical representation that conveys the pathways found in a dataset. By allowing only pairwise dependencies and partition aberrations into modules, in which all aberrations are restricted to have the same dependencies, we reduce the number of parameters so that datasets sizes relevant to cancer applications can be handled. We apply our framework to a dataset of colorectal cancer tumor karyotypes. The obtained model explains the data significantly better than a model where independence between the aberrations is assumed. In fact, the obtained model performs very well with respect to several measures of goodness of fit and is, with respect to repetition of the training, more or less unique.

Algorithms↗

PubMed2006

Combined use of the EPA-QUAL2E simulation model and factor analysis to assess the source apportionment of point and non point loads of nutrients to surface waters.

Diffuse pollution is generally indirectly estimated by area and specific emission factors as function of land use. However in many cases these estimates were proven to be remarkably inaccurate. Aim of this study was to combine a water quality simulation model, (USEPA-QUAL2E) and Factor Analysis to increase the understanding of the water pollutants source apportionment. The study concerned two different watersheds, an upland area characterised by a very scarce agricultural use, and another area covering both the upland and the lowland physiographic regions. Particularly the lowland region is included in one of the most productive agricultural areas in Italy. By comparing instream measurements with QUAL2E simulations during dry and wet weather conditions, a good fit (errors +/-20%) was found for the dry weather scenario, whereas very poor was the model performance on the wet weather scenario. This was in the same way expected since the rainfall-driven pollutants scenario deviates significantly from QUAL2E general assumptions of constant emissions in steady state streamflow conditions. However the poor fit was also due to the scarcer reliability of the adopted non point emission estimates. Despite of approximations the model wet weather simulations enabled to estimate the non point contribution to the instream load at the rainfall event scale resolution. Such diffuse sources contribution was found around 80% in the area of extensive agricultural land use, and around 40% in the upland region. Factor Analysis applied to the instream measurements data shed light on the exchange from the groundwater to the surface water system that occurred in the upland region. The hypothesis of a groundwater contribution to the instream total load of nitrates was also supported by QUAL2E simulations that, when considering only the point loads, systematically underestimate the dry weather nitrate concentrations. The same pattern was not observed for the lowland region.

Environmental Monitoring↗

PubMed1997

Epidermal aminopeptidase activity and metabolism as observed in an organized HaCaT cell sheet model.

Metabolism studies on organized HaCaT keratinocyte cell sheets are reported. Cells were grown on porous membranes to form organized cell sheets of several cell layers, which were considered as a model of viable epidermis. Metabolism was studied by reflection kinetics, with the top side of the cell sheets in contact with a donor solution and the basal side closed by an impermeable backing layer. Metabolite formation was followed by HPLC of substrate and metabolite in the donor. For comparison, studies with homogenized HaCaT cells were also performed. Model substrates were amino acid amides of 4-methoxy-2-naphthylamine (MNA) (i.e., Ala-MNA, Arg-MNA, Glu-MNA, and Leu-MNA). Also Leu-enkephalin was studied as a model peptide. Except for Glu-MNA, all substrates were metabolized in both the organized cell sheet and in the homogenates. In homogenate studies, saturation of the metabolic reaction was reached at <100 nmol mL(-1) substrate, whereas metabolism in organized cell sheets was below saturation (up to 500 nmol mL(-1)) except for Leu-enkephalin that showed saturation at >100 nmol mL(-1). In homogenates, substrate inhibition was found with Leu-MNA (> approximately 20 nmol mL(-1)) but not with Ala-MNA and Arg-MNA, both of which showed saturation. Differences of homogenates versus organized cell sheets are due to the intact organization and enzyme compartmentation of the cell sheets as opposed to the loss of organization and compartmentation in homogenates. Also, diffusion of substrate into cell sheets may be rate limiting.

2-Naphthylamine↗

PubMed2006

A combined approach for the evaluation of a volatile organic compound emissions inventory.

Emissions inventories significantly affect photochemical air quality model performance and the development of effective control strategies. However, there have been very few studies to evaluate their accuracy. Here, to evaluate a volatile organic compound (VOC) emissions inventory, we implemented a combined approach: comparing the ratios of carbon bond (CB)-IV VOC groups to nitrogen oxides (NOx) or carbon monoxide (CO) using an emission preprocessing model, comparing the ratios of VOC source contributions from a source apportionment technique to NOx or CO, and comparing ratios of CB-IV VOC groups to NOx or CO and the absolute concentrations of CB-IV VOC groups using an air quality model, with the corresponding ratios and concentrations observed at three sites (Maryland, Washington, DC, and New Jersey). The comparisons of the ethene/NOx ratio, the xylene group (XYL)/NOx ratio, and ethene and XYL concentrations between estimates and measurements showed some differences, depending on the comparison approach, at the Maryland and Washington, DC sites. On the other hand, consistent results at the New Jersey site were observed, implying a possible overestimation of vehicle exhaust. However, in the case of the toluene group (TOL), which is emitted mainly from surface coating and printing sources in the solvent utilization category, the ratios of TOL/ NOx or CO, as well as the absolute concentrations revealed an overestimate of these solvent sources by a factor of 1.5 to 3 at all three sites. In addition, the overestimate of these solvent sources agreed with the comparisons of surface coating and printing source contributions relative to NOx from a source apportionment technique to the corresponding value of estimates at the Maryland site. Other studies have also suggested an overestimate of solvent sources, implying a possibility of inaccurate emission factors in estimating VOC emissions from surface coating and printing sources. We tested the impact of these overestimates with a chemical transport model and found little change in ozone but substantial changes in calculated secondary organic aerosol concentrations.

Air Pollutants↗

PubMed2002

Combined clinical and imaging information as an early stroke outcome measure.

BACKGROUND AND PURPOSE: Imaging information has been proposed as a potential surrogate outcome in stroke clinical trials. The purpose of this study was to determine whether an early outcome measure combining clinical and imaging information is better than either alone in predicting 3-month outcome in acute ischemic stroke patients. METHODS: Clinical information (National Institutes of Health Stroke Scale) and imaging information (CT infarct volume), measured at 1 week from 201 patients from the Randomized Trial of Tirilazad Mesylate in Acute Stroke (RANTTAS) study, were used in a multivariable logistic regression analysis to predict excellent and devastating 3-month outcome. The combined models were compared with the infarct volume models and the clinical models. Discrimination, calibration, and change in global model chi-square were assessed. RESULTS: The combined models and models using clinical information alone had areas under the receiver operating characteristic curves that did not differ significantly (probability value = 0.092 to 0.4), ranging from 0.83 to 0.95. The imaging alone models performed less well (P<0.005) and had areas under the receiver operating characteristic curves that ranged from 0.70 to 0.80. CONCLUSIONS: The National Institutes of Health Stroke Scale at 1 week is highly predictive of 3-month outcome in ischemic stroke patients. The addition of 1-week infarct volume does not improve the accuracy of the predictive model.

Activities of Daily Living↗

PubMedDate not supplied

Diagnostic value of signs and symptoms associated with peripheral arterial occlusive disease seen in general practice: a multivariable approach.

OBJECTIVES: To assess the diagnostic values of single and combined data from the history, physical examination, and medical record with regard to peripheral arterial occlusive disease (PAOD) in patients with leg complaints; to construct a multivariable model for the clinical diagnosis of PAOD by primary care physicians. SETTING: 18 general practice centers in The Netherlands. DESIGN: Cross-sectional comparison of signs, symptoms, and data from the medical record with the independently assessed ankle-brachial systolic pressure index (ABPI; cutoff point < 0.90); analysis: bivariate, multiple logistic regression (MLR). POPULATION: 2,455 individuals with leg complaints, aged 40.7-78.4 years; ABPI < 0.90 present in 9.2% of legs (11.7% of individuals). OUTCOME MEASURES: Clinical variables: sensitivity, specificity, positive and negative predictive values (PV+, PV-), diagnostic odds ratio (OR); models: likelihood ratio test, area under the receiver operating characteristic curve (AUC). RESULTS: Bivariate analysis: highest sensitivity: age more than 60 years (77.3%); highest specificity: wounds or sores on toes and foot (99.7%); highest PV+: typical intermittent claudication (IC) (45.0%) (abnormal foot pulses 41.3%); highest PV-: strong pulses of both foot arteries (97.7%). MRL: the best-performing model (AUC 0.89) consisted of ten clinical variables: gender (OR 1.5), age more than 60 (OR 2.2); IC (OR 3.5); palpation of the skin temperature of the feet (OR 2.5), palpation of both foot pulses [OR 16.4 (abnormal) and 7.0 (doubtful)], auscultation of the femoral artery (OR 3.5); previous diagnosis of IHD (OR 1.7) or diabetes (OR 1.6), history of smoking (OR 2.1), and elevated blood pressure (OR 1.5). The range of predicted probabilities was 0.4-98%. The Hosmer-Lemeshow goodness-of-fit test indicated good overall fit (p = 52). CONCLUSIONS: Palpation of both foot pulses is the key procedure for the clinical diagnosis of PAOD. Traditional clinical evaluation enables the general practitioner to exclude the diagnosis of PAOD in many individuals with a high degree of certainly, to establish the diagnosis in a small group of patients, and to define a limited group of patients where supplementary noninvasive testing is appropriate. The MLR model can be used as a diagnostic checklist and as a reference for the physician's clinical hypothesis.

Adult↗

PubMed2002

Using a biology-based model (DEBtox) to analyze bioassays in ecotoxicology: opportunities and recommendations.

The conventional analysis of bioassays does not account for biological significance. However, mathematical models do exist that are realistic from a biological point of view and describe toxicokinetics and effects on test organisms of chemical compounds. Here we studied a biology-based model (DEBtox) that provides an estimate of a no-effect concentration, and we demonstrated the ability of such a model to adapt to different situations. We showed that the basic model can be extended to deal with problems usually faced during bioassays like time-varying concentrations or unsuitable choices of initial concentrations. To reach this goal, we report experimental data from Daphnia magna exposed to zinc. These data also showed the potential benefit of the model in understanding the influence of food on toxicity. We finally make some recommendations about the choice of initial concentrations, and we propose a test with a depuration period to check the relevance and the predictive capacity of the DEBtox model. In our experiments, the model performed well and proved its usefulness as a tool in risk assessment.

Animals↗

PubMed2002

The oral glucose minimal model: estimation of insulin sensitivity from a meal test.

Recently, a new approach has been proposed to estimate insulin sensitivity (S(I)) from an oral glucose tolerance test or a meal using an "integral equation". Here, we improve on the "integral equation" by resorting to a "differential equation" approach. The classic glucose kinetics minimal model was used with the addition of a parametric model for the rate of appearance into plasma of oral glucose (Ra). Three behavioral models of Ra were proposed: piecewise-linear (P), spline (S) and dynamic (D). All three models performed satisfactorily allowing a precise estimation of S(I) and a plausible reconstruction of Ra. Mean S(I) estimates were virtually identical: S(I)P = 6.81 +/- 0.87 (SE); S(I)S = 6.53 +/- 0.80; and S(I)D = 6.62 +/- 0.79. S(I) strongly correlated with the integral-equation index (I) S(I)I: r = 0.99, p < 0.01 for models D and S, and r 0.97, p < 0.01 for P. Also, SI compared well with insulin sensitivity estimated from intravenous glucose tolerance test in the same subjects (r = 0.75, p < 0.01; r = 0.71, p < 0.01; r = 0.73, p < 0.01, respectively, for P, S, and D models versus s(I)IVGTT). Finally, the novel approach allows estimation of SI from a shorter test (120 min): model P yielded S(I)R = 7.16 +/- 1.0 (R for reduced) which correlated very well with S(I)P and S(I)I (respectively, r = 0.94, p < 0.01; r = 0.95, p < 0.01) and still satisfactorily with S(I)IVGTT (r = 0.77, p < 0.01).

Blood Glucose↗

PubMed2003

Predicting effective continuous positive airway pressure in Taiwanese patients with obstructive sleep apnea syndrome.

BACKGROUND AND PURPOSE: Race/ethnicity plays an important role in determining body size and the severity of obstructive sleep apnea syndrome (OSAS), perhaps affecting the lowest effective continuous positive airway pressure (CPAP) [P(eff)] required to abolish sleep apneas. This study aimed to determine P(eff), using an optimal regression model to facilitate the titration of CPAP in Taiwanese people and to compare the findings of a regression model to the previously reported models developed in Caucasian populations. METHODS: 121 patients with moderate to severe OSAS were studied. All were titrated for nasal CPAP at a university hospital. The patients underwent separate polysomnographic studies to diagnose OSAS and to determine the P(eff) for home use. Seven polysomnographic variables were included in the analysis: apnea/hypopnea index (AHI), desaturation index (DI), desaturation duration, mean oxyhemoglobin saturation (S(P)O(2)), mean desaturation, nadir S(P)O(2), and snore index. Multiple linear regression was used to model the effects of anthropometric and polysomnographic variables on P(eff). A data-splitting cross-validation study and model fit criteria were used to select the predictors and assess the model performance. RESULTS: Most of the subjects were obese men who suffered severe OSAS with significant oxyhemoglobin desaturation and daytime sleepiness, and who required 8 cm H(2)O of CPAP. Age, body mass index (BMI), neck circumference (NC), AHI, DI, mean S(P)O(2), and daytime somnolence score were significantly associated with P(eff). The final prediction model (n = 121) was estimated as P(eff) = 0.52 + 0.174 x BMI + 0.042 x AHI, while the previously reported equation uses BMI, NC, and AHI. The prediction equation was more accurate in predicting P(eff) in this Taiwanese population as compared to the previously reported equations designed from a Caucasian sample. CONCLUSIONS: Obesity and severity of sleep apnea are the 2 most important predictors of CPAP setting effectively abolishing the apneas. After inclusion of BMI and AHI in the model, other variables (such as NC) do not significantly improve the prediction of P(eff), suggesting that ethic differences may play a role in predicting P(eff).

Adult↗

PubMed2026

Genetic Ancestry and Colorectal Cancer in the All of Us Dataset.

IMPORTANCE: Genetic ancestry may complement biological, behavioral, and clinical factors in understanding colorectal cancer (CRC) disparities; yet, ancestry-informed analyses in CRC remain limited. OBJECTIVE: To characterize associations of genetic ancestry with CRC burden, age at diagnosis, and age-specific risk, and to develop a multiethnic CRC risk-prediction model. DESIGN, SETTING, AND PARTICIPANTS: This retrospective cohort study used All of Us data from July 1986 to October 2023, with follow-up through last visit or death (median [IQR], 133.1 [57.1-186.5] months); analyses were conducted from February to June 2026. All of Us is a US research cohort with linked electronic health record (EHR) and short-read whole-genome sequencing (srWGS) data. All of Us Research Program participants with srWGS and linked EHR data were included, except those with hereditary polyposis or Lynch syndrome. EXPOSURES: Genetically inferred ancestry categories and principal components. MAIN OUTCOMES AND MEASURES: Any CRC was the primary outcome. Associations were evaluated using Fisher exact tests, cumulative incidence functions with Gray tests, cause-specific and Fine-Gray subdistribution hazard models, and pooled multivariable logistic regression. Prediction models used penalized least absolute shrinkage and selection operator and extreme gradient boosting (XGBoost). RESULTS: Among 316 624 participants (median [IQR] age, 56.3 [40.2-68.2] years; 172 327 [54.4%] of European ancestry; 191 705 female [61.2%]; 121 585 male [38.8%]), 2914 (0.9%) developed CRC. European ancestry was associated with higher odds of CRC vs all other ancestries combined (odds ratio, 1.50; 95% CI, 1.39-1.62). The median age at CRC diagnosis was older in European (63.4 [53.9-71.2] years) than in American admixed-Latino, African, East Asian, and Other ancestry groups. In cause-specific hazard models on the attained-age scale, American admixed-Latino (hazard ratio, 1.30; 95% CI, 1.14-1.47) and East Asian (hazard ratio, 1.43; 95% CI, 1.06-1.94) ancestry had higher age-specific CRC hazard than European ancestry, with consistent findings on the subdistribution scale accounting for competing death. The multiethnic XGBoost model performed best (receiver operating characteristic area under the curve, 0.898; 95% CI, 0.882-0.912; precision-recall area under the curve, 0.338; 95% CI, 0.296-0.379) and was well calibrated. CONCLUSIONS AND RELEVANCE: In this cohort study, genetic ancestry was associated with meaningful differences in CRC burden and age-specific risk. These findings suggest that a multiethnic XGBoost model may complement CRC screening as a risk-enrichment tool.

Aged↗

PubMed2000

Building a Web-based drug ordering system for hospitals: from requirements engineering to prototyping.

Web-based drug ordering allows a growing number of hospitals without pharmacy to communicate seamlessly with their external pharmacy. Business process analysis and object oriented modelling performed together with the users at a pilot hospital resulted in a comprehensive picture of the user and business requirements for electronic drug ordering. The user requirements were further validated with the help of a software prototype. In order to capture the needs of a large number of users CAP10, a new method making use of pre-built models, is proposed. Solutions for coping with the technical requirements (interfacing the business software at the pharmacy) and with the legal requirements (signing the orders) are presented.

Computer Simulation↗

PubMed2005

Bio-medical entity extraction using support vector machines.

OBJECTIVE: Support vector machines (SVMs) have achieved state-of-the-art performance in several classification tasks. In this article we apply them to the identification and semantic annotation of scientific and technical terminology in the domain of molecular biology. This illustrates the extensibility of the traditional named entity task to special domains with large-scale terminologies such as those in medicine and related disciplines. METHODS AND MATERIALS: The foundation for the model is a sample of text annotated by a domain expert according to an ontology of concepts, properties and relations. The model then learns to annotate unseen terms in new texts and contexts. The results can be used for a variety of intelligent language processing applications. We illustrate SVMs capabilities using a sample of 100 journal abstracts texts taken from the {human, blood cell, transcription factor} domain of MEDLINE. RESULTS: Approximately 3400 terms are annotated and the model performs at about 74% F-score on cross-validation tests. A detailed analysis based on empirical evidence shows the contribution of various feature sets to performance. CONCLUSION: Our experiments indicate a relationship between feature window size and the amount of training data and that a combination of surface words, orthographic features and head noun features achieve the best performance among the feature sets tested.

Algorithms↗

PubMed2003

The Society of Thoracic Surgeons: 30-day operative mortality and morbidity risk models.

BACKGROUND: Although 30 day risk-adjusted operative mortality (ROM) has been used for quality assessment, it is not sufficient to describe the outcomes after coronary artery bypass grafting (CABG) surgery. Risk-adjusted major morbidity may differentially impact quality of care (as complications occur more frequently than death) and enhance a surgical team's ability to assess their quality. This study identified the preoperative risk factors associated with several complications and a composite outcome (the presence of any major morbidity or 30-day operative mortality or both). METHODS: For CABG procedures, the 1997 to 1999 Society of Thoracic Surgeons (STS) National Adult Cardiac Surgery Database was used to develop ROM and risk-adjusted morbidity (ROMB) models. Risk factors were selected using standard STS univariate screening and multivariate logistic regression approaches. Risk model performance was assessed. Across STS participating sites, the association of observed-to-expected (O/E) ratios for ROM and ROMB was evaluated. RESULTS: The 30-day operative death and major complication rates for STS CABG procedures were 3.05% and 13.40%, respectively (503,478 CABG procedures), including stroke (1.63%), renal failure (3.53%), reoperation (5.17%), prolonged ventilation (5.96%), and sternal infection (0.63%). Risk models were developed (c-indexes for stroke [0.72], renal failure [0.76], reoperation [0.64], prolonged ventilation [0.75], sternal infection [0.66], and the composite endpoint [0.71]). Only a slight correlation was found, however, between ROMB and ROM indicators. CONCLUSIONS: Used in combination, ROMB and ROM may provide the surgical team with additional information to evaluate the quality of their care as well as valuable insights to allow them to focus on areas for improvement.

Aged↗

PubMed2003

Adapting the Charlson Comorbidity Index for use in patients with ESRD.

BACKGROUND: Accurate prediction of survival for patients with end-stage renal disease (ESRD) and multiple comorbid conditions is difficult. In nondialysis patients, the Charlson Comorbidity Index has been used to adjust for comorbidity. The purpose of this study is to assess the validity of the Charlson index in incident dialysis patients and modify the index for use specifically in this patient population. METHODS: Subjects included all incident hemodialysis and peritoneal dialysis patients starting dialysis therapy between July 1, 1999, and November 30, 2000. These 237 patients formed a cohort from which new integer weights for Charlson comorbidities were derived using Cox proportional hazards modeling. Performance of the original Charlson index and the new ESRD comorbidity index were compared using Kaplan-Meier survival curves, change in likelihood ratio, and the c statistic. RESULTS: After multivariate analysis and conversion of hazard ratios to index weights, only 6 of the original 18 Charlson variables were assigned the same weight and 6 variables were assigned a weight higher than in the original Charlson index. Using Kaplan-Meier survival curves, we found that both the original Charlson index and the new ESRD comorbidity index were associated with and able to describe a wide range of survival. However, the new study-specific index had better validated performance, indicated by a greater change in the likelihood ratio test and higher c statistic. CONCLUSION: This study indicates that the original Charlson index is a valid tool to assess comorbidity and predict survival in patients with ESRD. However, our modified ESRD comorbidity index had slightly better performance characteristics in this population.

Aged↗