SEARCH · PubMed Health

Results for “measurement error”

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 163 records · Page 9Linked to original sources

PubMed2006

Comparing methods to measure error in gynecologic cytology and surgical pathology.

CONTEXT: Both gynecologic cytology and surgical pathology use similar methods to measure diagnostic error, but differences exist between how these methods have been applied in the 2 fields. OBJECTIVE: To compare the application of methods of error detection in gynecologic cytology and surgical pathology. DATA SOURCES: Review of the literature. CONCLUSIONS: There are several different approaches to measuring error, all of which have limitations. Measuring error using reproducibility as the gold standard is a common method to determine error. While error rates in gynecologic cytology are well characterized and methods for objectively assessing error in the legal setting have been developed, meaningful methods to measure error rates in clinical practice are not commonly used and little is known about the error rates in this setting. In contrast, in surgical pathology the error rates are not as well characterized and methods for assessing error in the legal setting are not as well defined, but methods to measure error in actual clinical practice have been characterized and preliminary data from these methods are now available concerning the error rates in this setting.

Diagnostic Errors↗

PubMed1994

Computer program for the proportional hazards measurement error model.

The Cox-regression analysis based on the partial likelihood assumes that the covariates, or independent variables, are exactly measured without error. If the covariates are subject to measurement error and the error-prone observed values are used in the analysis by simply ignoring the measurement error, the results are generally biased and misleading; the bias does not diminish as the sample size is increased. The objective of the paper is to briefly describe a method searching for asymptotically unbiased estimates of the parameters correcting for the measurement error in the Cox-regression model and to present a FORTRAN program to perform the correction method; asymptotic standard errors of the corrected estimates are also obtained. The measurement error distribution, that is the conditional distribution of the observed values given the true value, must be specified. An advantage of the method described is that it does not require any assumption on the distribution of the true values; in other words, <--> values are treated as unknown fixed constants. It can accommodate tied failure times unless ties are very frequent, and any censorship or loss to follow-up are allowed as long as they are 'independent of survival'.

Likelihood Functions↗

PubMed1984

Some effects of random dose measurement errors on analyses of atomic bomb survivor data.

The effects of random dose measurement errors on analyses of atomic bomb survivor data are described and quantified for several analytical procedures. It is found that the ways in which measurement error is most likely to mislead are through downward bias in the estimated regression coefficients and through distortion of the shape of the dose-response curve. The magnitude of the bias and the power for testing the hypothesis of no effect are evaluated for several dose treatments including the use of grouped and ungrouped data, analyses with and without substituting 600 rad for estimated doses exceeding this value, and analyses which exclude doses exceeding 200 rad. The calculations are based on a model in which the error distributions are assumed to be log normal with standard deviations that are 0, 30, and 50%, respectively, of the true dose values. Results are limited to a dose-response function which is linear on total dose. It is found that the commonly applied practice of substituting 600 rad for doses exceeding this value definitely reduces bias in the presence of error. Restricting analyses to doses less than 200 rad reduces bias even more but at the price of considerable loss of power. Both the bias and the power for analyses based on grouped data are very close to the respective bias and power with ungrouped data.

Dose-Response Relationship, Radiation↗

PubMed1989

Covariance analysis in generalized linear measurement error models.

We summarize some of the recent work on the errors-in-variables problem in generalized linear models. The focus is on covariance analysis, and in particular testing for and estimation of treatment effects. There is a considerable difference between the randomized and non-randomized models when testing for an effect. In randomized studies, simple techniques exist for testing for a treatment effect. In some instances, such as linear and multiplicative regression, simple methods exist for estimating the treatment effect. In other examples such as logistic regression, estimating a treatment effect requires careful attention to measurement error. In non-randomized studies, there is no recourse to understanding and modelling measurement error. In particular ignoring measurement error can lead to the wrong conclusions, for example the true but unobserved data may indicate a positive effect for treatment, while the observed data indicate the opposite. Some of the possible methods are outlined and compared.

Analysis of Variance↗

PubMed2006

Recursive versus sequential multiple error measures reduction: a curve simplification approach to ECG data compression.

Previous time domain compression methods have been tackled by sequential one point at a time sub-optimal selection strategies running in approximately O(N) or all points at a time optimal strategies running in approximately O(N3) temporal complexities. Yet basically, the selected dominant points (DPs) are locally only significant in these methods, which may lead to inaccurate reconstruction or even loss of clinical data. Alternatively, the recursive one point at a time selection strategy, through different variants of the Douglas-Peucker line simplification algorithm, computes globally significant DPs for an approximately O(N.log2(N)) temporal complexity. We illustrate that the recursive strategy performs numerically almost two times better than the sequential one. The piecewise linear approximation of the input ECG is formally expressed as a curve simplification problem, through reduction of the Hausdorff error measure. We also illustrate that reduction of two error measures performs better than reduction of one error measure. An additional compression option is proposed in the case of the recursive strategy through simplification of the sorted distances associated to the selected set of points. The outcome is a compression algorithm that yields compression ratios ranging from 8:1 to 22:1 for a perceptually good reconstruction quality and near linear execution time. The tests have been conducted on the MIT-BIH public ECG database. Results show that the proposed recursive algorithm is an excellent compromise on compression ratio - computational time - reconstruction quality.

Algorithms↗

PubMed1991

Adjustment of provisional mortality series: the dynamic linear model with structured measurement errors.

The authors "consider the problem of adjusting provisional time series using a bivariate structural model with correlated measurement errors. Maximum likelihood estimators and a minimum mean squared error adjustment procedure are derived for a provisional and final series containing common trend and seasonal components. The model also includes measurement errors common to both series and errors that are specific to the provisional series. [The authors] illustrate the technique by using provisional data to forecast ischemic heart disease mortality."

Cause of Death↗

PubMed2000

Measuring the effectiveness of diagnostic markers in the presence of measurement error through the use of ROC curves.

The area under the receiver operating characteristic curve is frequently used to assess the effectiveness of diagnostic markers in distinguishing between diseased and healthy individuals. These markers are generally subject to measurement error. In this paper confidence intervals for the area under the curve are developed which take measurement error into account. These intervals depend on the availability of replicated observations for the subject. Both equal and unequal numbers of replicates per subject are considered. The practice of averaging over replicates and then ignoring measurement error is examined and found wanting.

Adolescent↗

PubMed1987

Estimation of the mid-upper arm circumference measurement error.

The mid-upper arm circumference measurement error was estimated in order to assess the usefulness of this measurement for detecting small changes in a patient's arm circumference during hospitalization. In a study having a 3 X 3 X 2 factorial design, 10 trained observers recorded 1,560 measurements from 52 healthy adult subjects selected to represent three factors: sex, age, and weight for height. Each observer obtained three independent measurements from every subject. By analysis of variance, the variation between different observers' measurements of the same subject was estimated with 95% confidence to be less than or equal to 2.8% for men and less than or equal to 3.8% for women. This variation was unaffected by age but decreased with decreasing weight for height in women (p less than .05). Variation between each observer's repeated measurements of the same subject was estimated with 95% confidence to be less than or equal to 1.1% and was unaffected by sex, age, or weight for height. It was concluded that the smallest detectable change in arm circumference is 7.8% for men and 9.8% for women when successive measurements are recorded by different observers. However, arm circumference changes as small as 2.2% can be detected in both men and women when successive measurements are obtained by the same observer.

Adult↗

PubMed1994

Corrected likelihood for proportional hazards measurement error model and its application.

Consider the case where the exact values of covariates in the proportional hazards model may not be observed but instead, only surrogates for them involving measurement errors are available. The maximum likelihood estimate based on the partial likelihood with the true covariate replaced by the observed surrogate is even asymptotically biased and may cause seriously misleading results in covariance analysis based on the partial likelihood. These facts are illustrated by Monte Carlo simulation. A correction to partial likelihood proposed by the first author is studied to gain insight into its merits and limitations in practical applications. The results indicate that when the "effective magnitude of the measurement error" as defined in this article is small, which is indeed the case for most applications, the method will be useful. Some other correction methods for the measurement error in censored survival models are also reviewed and discussed.

Adult↗

PubMed2003

Analysis of associations with change in a multivariate outcome variable when baseline is subject to measurement error.

A simple general algorithm is described for correcting for bias caused by measurement error in independent variables in multivariate linear regression. This algorithm, using standard software, is then applied to several approaches to the analysis of change from baseline as a function of baseline value of the outcome measure plus other covariates, any of which might have measurement error. The algorithm may also be used when the independent variables differ by component of the multivariate independent variable. Simulations indicate that under various conditions bias is much reduced, as is mean squared error, and coverage of 95 per cent confidence intervals is good.

Algorithms↗

PubMed1995

Correction of risk estimates for measurement error in epidemiology.

Epidemiologists are under considerable pressure to acknowledge the presence of measurement error in the determination of risk factors. Repeatability and validation studies are often prescribed in conjunction with epidemiological studies. We describe some practical uses for repeatability and validation study data, in terms of correcting risk estimates for measurement error. Commonly available methods are described, with their advantages and shortcomings. A user-friendly computer program to carry out the analyses described accompanies the paper.

Case-Control Studies↗

PubMed1994

Implications of measurement error in exposure for the sample sizes of case-control studies.

In this paper, recent results describing the effects of measurement error on estimation of the association between an exposure and a disease are applied to sample size calculation in case-control studies. Models of the relation between true exposure and a surrogate exposure measure assessed with error are used to derive equations for sample size determination. The results show that the sample size of a study based on an exposure variable which is measured with error must be larger by a factor of 1/rho 2 than if exposure were measured without error, where rho is the correlation between the true exposure and the surrogate exposure measure. Review of the magnitude of measurement error in dietary assessments suggests that failure to account for measurement error in sample size determination for case-control studies of diet and disease could lead to marked underestimation of the required sample size.

Bias↗

PubMed2003

Estimation in capture-recapture models when covariates are subject to measurement errors.

We consider estimation problems in capture-recapture models when the covariates or the auxiliary variables are measured with errors. The naive approach, which ignores measurement errors, is found to be unacceptable in the estimation of both regression parameters and population size: it yields estimators with biases increasing with the magnitude of errors, and flawed confidence intervals. To account for measurement errors, we derive a regression parameter estimator using a regression calibration method. We develop modified estimators of the population size accordingly. A simulation study shows that the resulting estimators are more satisfactory than those from either the naive approach or the simulation extrapolation (SIMEX) method. Data from a bird species Prinia flaviventris in Hong Kong are analyzed with and without the assumption of measurement errors, to demonstrate the effects of errors on estimations.

Analysis of Variance↗

PubMed2006

Repeatability (test-retest variability) of refractive error measurement in clinical settings.

PURPOSE: To estimate the repeatability of refractive error measurement (REM) in a clinical environment in cataractous, pseudophakic and healthy eyes. METHODS: The refractive error of patients referred for cataract surgery or consultation measured by ophthalmic professionals was re-examined and the measurement results were compared. A total of 99 eyes from 99 persons (41 cataractous, 36 pseudophakic and 22 healthy eyes) with visual acuity (VA) of 0.3-1.3 (logMAR 0.52 to - 0.11) were included. The differences between measurements 1 and 2 were calculated as 3-dimensional vector values and spherical equivalents (SEs) and expressed as the coefficient of repeatability (CR). The mean time interval between the first and second examinations was 45 days. RESULTS: The CRs for all eyes for vertical (V), torsional (T) and horizontal (H) vectors were 0.74 D, 0.34 D and 0.93 D, respectively. The CR of SE for all eyes was 0.74 D. Eyes with lower VA (0.3-0.45) had larger variability in vector and SE values but the differences between VA groups were not statistically significant. The difference in the mean defocus equivalent (DE) between measurements 1 and 2 was, however, significantly greater in the group with lower VA. In all VA groups the mean difference vector was very close to the zero vector, which means that there was no systematic difference. CONCLUSIONS: Repeatability of refractive error measurements in clinical settings has a certain degree of variability. In this series, the variability in eyes with better VA was not great and was in accordance with earlier findings in healthy eyes. Eyes with lower VA had greater variability due to greater tolerance to defocus. Thus, conclusions concerning changes in the refractive state and the need to make changes in the refractive correction of eyes with poorer vision should be made with caution.

Adult↗

PubMed2000

Neuroretinal rim measurement error using PC-based stereo software.

The neuroretinal rims of a set of glaucoma patients were measured using digitized stereo photographs, to determine the reproducibility of computerized stereo measurements of the neuroretinal rim. Each rim was measured five times at 18 locations, with measurement error (ME) defined as the mean of standard deviations of each set of measurements. The following ME were determined: (i) inter-sessional variability (n = 27 right and 24 left eyes, at t1 and t2); (ii) inter-assessor variability (n = 9, 2 assessors); and (iii) variability after colour adjustment algorithms were applied (n = 15). The results were as follows: (i) inter-sessional variability was 3.41+/-1.08 for t1 and 3.22+/-0.84 for t2; (ii) there was a significant difference between the two assessors, although the ME was still low; and (iii) there was no significant differences between the ME of unadjusted and adjusted images. With a measurement error of up to 11% of rim width, these results show that lowcost rim measurements can be made using PC-based software.

Algorithms↗

PubMedDate not supplied

Uncertainty analysis of field placement error measurements using digital portal and simulation image correlations.

An uncertainty analysis of a field placement error measurement method using correlations between digital portal and simulation images has been performed. This analysis used the Monte Carlo method to simulate random uncertainties arising in image registration and portal field edge detection. The uncertainty analysis used data from a field placement error study to predict the dependences of the measurement method upon image matchpoint number and spatial distribution, image registration, and uncertainties arising in the detection and description of the field edge in a portal image. It was shown that the uncertainty in the field registration had the most significant effect upon the measurement uncertainty, whereas the number of matchpoints had an extremely limited effect (beyond the minimum required for the mathematical procedure used in registration). In contrast, it was shown that an accurate field placement error measurement required that the number of pixels describing a field edge used for comparisons should exceed 20, thus automatically setting a minimum field length necessary for field placement comparisons. It was also shown that the spatial uniformity of the matchpoints in an image pair had a limited effect upon the field placement error measurement, provided that the points are distributed over about 15% of the image area.

Humans↗

PubMed1995

Estimating degradation by a Wiener diffusion process subject to measurement error.

Most materials and components degrade physically before they fail. Engineering degradation tests are designed to measure these degradation processes. Measurements in the tests reflect the inherent randomness of degradation itself as well as measurement errors created by imperfect instruments, procedures and environments. This paper describes a statistical model for measured degradation data that takes both sources of variation into account. The degradation process in the model is taken to be a Wiener diffusion process. The measurement errors are assumed to be independent normal random outcomes that are independent of the degradation process. The paper describes inference procedures for the model and discusses some practical issues that must be considered in dealing with the statistical problem. A case study is presented.

Engineering↗

PubMed1996

Quasilikelihood estimation in measurement error models with correlated replicates.

We consider quasilikelihood models when some of the predictors are measured with error. In many cases, the true but fallible predictor is impossible to measure, and the best one can do is to obtain replicates of the fallible predictor. We consider the case that the replicates are not independent. If one assumes that replicates are independent and they are not, one typically underestimates the extent of the measurement error, leading to an inconsistent errors in variables correction. We devise techniques for estimating the measurement error covariance matrix. In addition, we discuss how one might perform a quasilikelihood analysis by computing the mean and variance functions of the observed data, both using approximations and also exactly through a Monte Carlo method. The methods are illustrated on a data set involving systolic blood pressure and urinary sodium chloride, where the measurement errors appear to be approximately normally distributed but highly correlated, and the distribution of the true predictor is reasonably modeled as a mixture of normals.

Analysis of Variance↗