PubMed Health⌕ Search

SEARCH · PubMed Health

Results for “Models, Statistical”

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 1,333 records · Page 74Linked to original sources
PubMed2005

Evaluating ELISPOT summary measures with criteria for obtaining reliable estimates.

The ELISPOT assay is a commonly used technique for quantifying the occurrence of T lymphocyte cells secreting a cytokine after stimulation with an antigen or peptide. The assay endpoint, the number of spot-forming cells (SFC) at a specific concentration of effector cells, is typically estimated using either a simple arithmetic mean or the predicted value from a linear regression model. We compare statistical modeling approaches for summarizing these assays using data from the Pediatric AIDS Clinical Trial Group (PACTG) study 299. A simulation study was conducted to compare methods under controlled conditions. Assuming the optimal effector cell concentration is known, we demonstrate that the simple mean is appropriate if assays are conducted at the same concentration for all samples. Normalizing simple means to a summary concentration using results from a range of concentrations is not valid. A random effects or mixed model is superior to the simple mean when a large within-assay (subject) variance relative to between-subject variance exists. When assays are conducted over a range of effector cell concentrations for each individual, the theoretical linearity assumption of the regression model is often violated and can result in biased estimates. In this case, nonlinear models provide more accurate estimation. Collecting data over a range of concentrations allows reassessment of the optimal cell concentration after the data are generated.

Computer Simulation↗
PubMed2006

Effective Hamiltonian models and unimolecular decomposition.

Partitioning Hilbert space into two subspaces by using orthogonal projection operators yields compact forms for effective Hamiltonians for each of the subspaces. When one (the Q space) contains molecular bound states and the other (the P space) contains dissociative continua, a simple form for the non-Hermitian Q-space effective Hamiltonian, H(eff), can be obtained, subject to reasonable approximations. Namely, H(eff) = H0 - ivariant Planck's/2pi Gamma/2, where H0 is Hermitian, and the width operator variant Planck's/2pi Gamma accounts for couplings of the Q-space levels to the P-space continua. The P/Q partitioning procedure has been applied in many areas of atomic, molecular, and nuclear physics with widespread success. Inputting into this formalism ideas from random matrix theory in order to model independent open channels yields the random matrix H(eff) model. Despite numerous efforts, this model has failed to model satisfactorily the statistical transition-state theory of unimolecular decomposition (hereafter referred to as TST) in the regime of overlapping resonances, where nearly all such reactions occur. All statistical models of unimolecular decomposition are premised on rapid intramolecular vibrational redistribution (IVR) for a given set of good quantum numbers. The phase space thus accessed results in a threshold reaction rate of 1/h rho, and for K independent open channels, the rate is K/h rho. This reaction rate corresponds to a resonance width of K/2pi rho, and when K increases, the resonances (which are rho(-1) apart) overlap. In this regime, the random matrix H(eff) model fails because it does not introduce independent open channels. To illustrate the source of the problem, an analysis is carried out of a simple model that is obviously and manifestly inconsistent with TST. This model is solved exactly, and it is then put in the form of the random matrix H(eff) model, illustrating the one-to-one correspondence. This reveals the deficiencies of the latter. In manipulating this model into the form H0 - ivariant Planck's/2pi Gamma/2, it becomes clear that the independent open channels in the random matrix H(eff) model are inconsistent with TST. Rather, this model is one of gateway states (i.e., bound states that are coupled to their respective continua as well as to a manifold of zero-order bound states, none of which are coupled directly to the continua). Despite the fact that the effective Hamiltonian method is, by itself, beyond reproach, the random matrix H(eff) model is flawed as a model of unimolecular decomposition in several respects, most notably, bifurcations of the distributions of resonance widths in the regime of overlapping resonances.

Journal Article↗
PubMed2002

An analysis of factors that influence personal exposure to nitrogen oxides in residents of Richmond, Virginia.

Nitrogen oxides (NO(x)) are ubiquitous pollutants in outdoor and indoor air. However, epidemiologic studies that evaluate health effects associated with NO(x) commonly rely upon outdoor concentrations of NO(x), nitrogen dioxide (NO(2)), or residence characteristics as surrogates for personal exposure. In this study, personal exposures (48 h) and corresponding indoor and outdoor concentrations of nitric oxide (NO), NO(2), and NO(x) were measured (July-September) in 39 adults and 9 children from 23 households in Richmond, Virginia, using Ogawa passive NO(x) monitors. Demographic, time-activity patterns, and household data were collected by questionnaire and used to develop exposure prediction models. Adults had higher NO(2), NO, and NO(x) exposures (means: 16, 63, and 79 ppb, respectively) than children (13, 49, and 62 ppb). Measurements taken in bedrooms (18, 57, and 75 ppb) and living rooms (19, 65, and 84 ppb) surpassed measurements taken outdoors (15, 21, and 36 ppb). In indoor locations, NO(x) concentrations were influenced largely by NO, and consequently, personal exposure prediction models for NO(x) were reflective of models for NO. Statistical models that best predicted personal exposures included indoor measurements; outdoor measurements contributed relatively little to personal exposure. Close to 70% of the variation in personal NO(2) and NO(x) exposure was explained by two variable models (bedroom NO(2) and time spent in other indoor locations; bedroom NO(x) and time spent in kitchen). Given appropriate resources, measurement error in epidemiologic studies can be reduced significantly with the use of personal exposure measurements or prediction models developed from indoor measurements and survey data.

Activities of Daily Living↗
PubMed2004

Statistical inference and model selection for the 1861 Hagelloch measles epidemic.

A stochastic epidemic model is proposed which incorporates heterogeneity in the spread of a disease through a population. In particular, three factors are considered: the spatial location of an individual's home and the household and school class to which the individual belongs. The model is applied to an extremely informative measles data set and the model is compared with nested models, which incorporate some, but not all, of the aforementioned factors. A reversible jump Markov chain Monte Carlo algorithm is then introduced which assists in selecting the most appropriate model to fit the data.

Adolescent↗
PubMed2006

Intensive care unit length of stay: Benchmarking based on Acute Physiology and Chronic Health Evaluation (APACHE) IV.

OBJECTIVE: To revise and update the Acute Physiology and Chronic Health Evaluation (APACHE) model for predicting intensive care unit (ICU) length of stay. DESIGN: Observational cohort study. SETTING: One hundred and four ICUs in 45 U.S. hospitals. PATIENTS: Patients included 131,618 consecutive ICU admissions during 2002 and 2003, of which 116,209 met inclusion criteria. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: The APACHE IV model for predicting ICU length of stay was developed using ICU day 1 patient data and a multivariate linear regression procedure to estimate the precise ICU stay for randomly selected patients who comprised 60% of the database. New variables were added to the previous APACHE III model, and advanced statistical modeling techniques were used. Accuracy was assessed by comparing mean observed and mean predicted ICU stay for the excluded 40% of patients. Aggregate mean observed ICU stay was 3.86 days and mean predicted 3.78 days (p < .001), a difference of 1.9 hrs. For 108 (93%) of 116 diagnoses, there was no significant difference between mean observed and mean predicted ICU stay. The model accounted for 21.5% of the variation in ICU stay across individual patients and 62% across ICUs. Correspondence between mean observed and mean predicted length of stay was reduced for patients with a short (< or =1.7 days) or long (> or =9.4 days) ICU stay and a low (<20%) or high (>80%) risk of death on ICU day 1. CONCLUSIONS: The APACHE IV model provides clinically useful ICU length of stay predictions for critically ill patient groups, but its accuracy and utility are limited for individual patients. APACHE IV benchmarks for ICU stay are useful for assessing the efficiency of unit throughput and support examination of structural, managerial, and patient factors that affect ICU stay.

APACHE↗
PubMed1993

Statistical moments in pharmacokinetics: models and assumptions.

The modelling basis of statistical moments in pharmacokinetics is considered and the associated assumptions and restrictions highlighted. Both deterministic and statistical models are described and they are seen to give identical results on the basis of equivalent assumptions. It is shown that it is not necessary to assume that all kinetic processes are first order and that some of the pharmacokinetic parameters may be dependent on the dose of administered drug and on the route of administration.

Models, Statistical↗
PubMed2006

Mean blood glucose and biological variation have greater influence on HbA1c levels than glucose instability: an analysis of data from the Diabetes Control and Complications Trial.

OBJECTIVE: Mean blood glucose (MBG) over 2-3 months is a strong predictor of HbA(1c) (A1C) levels. Glucose instability, the variability of blood glucose levels comprising the MBG, and biological variation in A1C (BV) have also been suggested as predictors of A1C independent of MBG. To assess the relative importance of MBG, BV, and glucose instability on A1C, we analyzed patient data from the Diabetes Control and Complications Trial (DCCT). RESEARCH DESIGN AND METHODS: A glucose profile set and sample for A1C were collected quarterly over the course of the DCCT from each participant (n = 1,441). The glucose profile set consisted of seven samples, one each drawn before and 90 min after breakfast, lunch, and dinner and one before bedtime. MBG and glucose instability (SD of blood glucose [SDBG]) were calculated as the arithmetic mean and SD of glucose profile set samples for each visit, respectively. A statistical model was developed to predict A1C from MBG, SDBG, and BV, adjusted for diabetes duration, sex, treatment group, stratum, and race. RESULTS: Data from 32,977 visits were available. The overall model was highly statistically significant (log likelihood = -41,818.75, likelihood ratio chi2[7] = 7,218.71, P > chi2 = 0.0000). MBG and BV had large influences on A1C based on their standardized coefficients. SDBG had only 1/14 of the impact of MBG and 1/10 of the impact of BV. CONCLUSIONS: MBG and BV have a large influence on A1C, whereas SDBG is relatively unimportant. Consideration of BV as well as MBG in the interpretation of A1C may enhance our ability to monitor diabetes management and predict complications.

Blood Glucose↗
PubMed2001

Identification of patients at high cardiovascular risk: a critical appraisal of applicability of statistical risk prediction models.

Assessment of cardiovascular risk is widely proposed as a basis for taking management decisions about patients presenting with hypertension or hypercholesterolemia. Our aim was to critically assess the use of risk equations derived from epidemiological studies for the purpose of identifying high-risk patients. Risk equations were retrieved from the MEDLINE database and then applied to a data set of 118 patients. This data set was an evaluation study of the clinical value of the World Health Organization 1993 hypertension guidelines for the decision to treat mild hypertensive patients. We calculated agreement: 1) between equations and 2) between equations and the decision to treat taken by the physician. Most models were not applicable to our population, mainly because the original population had a narrow age range or comprised only males. Between-model agreement was better for the lower and upper risk quintiles than for the three other risk quintiles (0.58, 0.33, 0.34, 0.45, 0.70, from the lower to the upper risk quintile). When using an arbitrary threshold for defining high-risk patients (i.e. > 2% per year), we observed a huge variation of the proportion of patients classified at high risk (from 0 to 17%). There was a poor agreement between risk models and the decision to treat taken by the physician. These results suggest that risk-based guidelines should be validated before their diffusion.

Antihypertensive Agents↗
PubMed2004

Providing more up-to-date estimates of patient survival: a comparison of standard survival analysis with period analysis using life-table methods and proportional hazards models.

OBJECTIVE: Standard survival methods can yield out-of-date estimates of long-term survival. Period analysis, based on life-table methodology, provides more up-to-date survival estimates by exploring survival during a restricted recent period of interest. It excludes the short-term survival of patients recruited at the start of the study. We use statistical models to further develop the method of period analysis, providing more up-to-date estimates of survival and the ability to explore differences in survival by covariates and adjust for case mix. METHODS: We use cancer registry data for colorectal cancer in Leicestershire, UK, to illustrate the use of Cox proportional hazards (CPH) models to estimate period and standard survival. We compare these estimates with those obtained using life-table methodology. RESULTS: Period estimates were slightly higher than the standard estimates as they reflect recent improvements in short-term survival. The results for period analysis using the life-table approach and using CPH models were similar. However, CPH models allowed further investigation of other risk factors and the ability to control for potential confounding variables. CONCLUSION: Using period survival estimates, more up-to-date information is available to clinicians and others with an interest in monitoring survival. Period CHP models offer all the advantages of statistical modeling, and are straightforward to fit in standard statistical packages.

Colorectal Neoplasms↗
PubMed1989

[Study of a comparative statistical analytical model for nucleotide sequences in the genome of viruses of the Papovaviridae family].

In the present study a model for comparative analysis of nucleotide sequences has been developed, in order to evaluate statistical features of nucleotide distribution in DNA strands without any genetic relationship. Every DNA strand has been considered as a finite Markov chain; a matrix, whose elements represent the number of couplings between a nucleotide and the following one in 5'-3' direction, has been used for every DNA strand, and the statistical relationship has been detected by using Kendall's test. The genomes of Polyomavirus (strain A2) and DPV have been analysed by the proposed model; a substantial likeness between the behaviour of nucleotide distribution on all four DNA strands analysed has been shown; the strongest likeness concerned the complementary strands of Polyomavirus as well as the homologous sense strands of both viruses.

Base Sequence↗
PubMed2006

A strategy to discover genes that carry multi-allelic or mono-allelic risk for common diseases: a cohort allelic sums test (CAST).

A method is described to discover if a gene carries one or more allelic mutations that confer risk for any specified common disease. The method does not depend upon genetic linkage of risk-conferring mutations to high frequency genetic markers such as single nucleotide polymorphisms. Instead, the sums of allelic mutation frequencies in case and control cohorts are determined and a statistical test is applied to discover if the difference in these sums is greater than would be expected by chance. A statistical model is presented that defines the ability of such tests to detect significant gene-disease relationships as a function of case and control cohort sizes and key confounding variables: zygosity and genicity, environmental risk factors, errors in diagnosis, limits to mutant detection, linkage of neutral and risk-conferring mutations, ethnic diversity in the general population and the expectation that among all exonic mutants in the human genome greater than 90% will be neutral with regard to any effect on disease risk. Means to test the null hypothesis for, and determine the statistical power of, each test are provided. For this "cohort allelic sums test" or "CAST", the statistical model and test are provided as an Excel program, CASTAT(c) at . Based on genetics, technology and statistics, a strategy of enumerating the mutant alleles carried in the exons and splice sites of the estimated approximately 25,000 human genes in case cohort samples of 10,000 persons for each of 100 common diseases is proposed and evaluated: A wide range of possible conditions of multi-allelic or mono-allelic and monogenic, multigenic or polygenic (including epistatic) risk are found to be detectable using the statistical criteria of 1 or 10 "false positive" gene associations approximately 25,000 gene-disease pair-wise trials and a statistical power of >0.8. Using estimates of the distribution of both neutral and gene-inactivating nondeleterious mutations in humans and the sensitivity of the test to multigenic or multicausal risk, it is estimated that about 80% of nullizygous, heterozygous and functionally dominant gene-common disease associations may be discovered. Limitations include relative insensitivity of CAST to about 60% of possible associations given homozygous (wild type) risk and, more rarely, other stochastic limits when the frequency of mutations in the case cohort approaches that of the control cohort and biases such as absence of genetic risk masked by risk derived from a shared cultural environment.

Alleles↗
PubMed2005

Statistical methods for testing plaque removal efficacy in clinical trials.

OBJECTIVES: To evaluate the ability of different statistical approaches in finding a statistically significant difference in plaque removal efficiency between brushes in clinical trials. MATERIALS AND METHODS: The approaches, which are evaluated, concern the scores after brushing only, the difference in scores before and after brushing and the relative difference scores (i.e. score before minus score after brushing divided by the score before brushing). In each case the scores before brushing may be included as a covariate. Except for the relative difference scores, the power of the test statistics of the approaches has been compared by assuming a simple statistical model. These theoretical results have been compared with the numerical results of two particular clinical trials--one with a between-subject design and one with a within-subject design. RESULTS: The numerical results of these clinical trials show that the calculated p-values support the conclusions drawn from the statistical model, i.e. the power of the F-test is highest when evaluating the data after brushing with the data before brushing included as a covariate. Using the differences in scores before and after brushing--again with the data before brushing as a covariate--does not add additional power to the test. Omitting the data before brushing as a covariate only gives satisfactory results when the variance over the subjects or the error variance is zero, which in general is not the case. CONCLUSIONS: This investigation reveals that in general the approach of analysing the scores after brushing with the scores before brushing as a covariate yields the highest chance of finding a statistically significant difference between two brushes.

Adult↗