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 145 records · Page 8Linked to original sources

PubMed1991

DNA measurement errors with a scanning microdensitometer in cytologic and histologic samples of breast cancers.

The influence of various DNA measurement errors using a commercially available scanning microdensitometer was evaluated on Feulgen-stained cytologic and histologic samples prepared from paraffin blocks containing invasive ductal breast cancers. The overall average total measurement error was 5.5% for the cytologic specimens and 10.9% for the 4 micron histologic sections. Components of the error included microscopic adjustment variation and focussing errors (3.5% and 1.1%, respectively, for both cytologic and histologic samples) and background intensity estimation errors (3.0% for the cytologic samples and 10.0% for the histologic samples). Measurements of the integrated optical density had a minimal error of 0.5% and an average error of 1.0%. Limitations due to the histologic architecture and/or heterogeneous cell population gave rise to large differences in the selection of nuclei when differently sized scanning masks were used. To improve the reproducibility, masks used should be based on the individual cell size, and background intensity values should be carefully estimated in the vicinity of the selected cells. Overall, the cytologic tumor samples were preferable to the histologic samples for static DNA measurements. It was easier to select cells suitable for measurement in the cytologic samples, and the cytologic measurements were less time consuming and produced a smaller measurement error.

Breast Neoplasms↗

PubMed1997

Measurement error and its impact on the estimated relationship between dust lead and children's blood lead: Members of the Rochester Lead-in-Dust Study Group.

OBJECTIVE: Lead-contaminated house dust is a major contributor to lead intake among urban children, but the reliabilities of various dust lead measurement methods and their impact on the estimated correlations between dust lead and children's blood lead levels are unknown. METHODS: Repeated field measurements of lead-contaminated dust from children's homes were taken from 16 housing units using five dust lead measurement methods. Estimates of measurement error were used to obtain reliability ratios for the dust lead measurements, which were then used to correct estimated correlations between lead-contaminated dust and children's blood lead. RESULTS: Reliability varied over methods and surface types (from 0.0 to 0.8), but wipe loading and the BRM vacuum loading methods generally had greater reliability. Technician effects, inadvertent field exposure to lead, contamination of collection equipment, and laboratory instrument error were found to contribute little to total measurement error. Corrected correlations between blood lead and wipe loading measurements were 7 to 104% higher than uncorrected correlations. The multiple R2 and partial R2 for a wipe composite measurement in a multivariate regression model increased from 0.43 to 0.64 and from 0.053 to 0.26, respectively, after correction for measurement error bias. CONCLUSIONS: Variation in lead deposition within small areas and variations in collection inherent to the devices are major contributors to measurement error. Measurement error causes dramatic underestimation of correlation between lead-contaminated house dust and children's blood lead.

Air Pollution, Indoor↗

PubMed1998

The problem of measurement error in multisite clinical trials.

The implementation of a multisite, randomized, clinical psychopharmacologic trial involves a substantial investment of time and effort on the part of all participants. Because of their complexity, such clinical trials present unique methodological and design challenges. Indeed, it is not uncommon for such studies to conclude with uninterpretable results, due in part to such methodological pitfalls. It has been suggested that clarification of such methodologic dilemmas is one of the most important challenges facing the future of industry-sponsored psychopharmacologic drug development. Among the many factors that may contribute to problematic clinical trial results, error in measuring the phenomena being studied is of particular concern. In this article, we describe the outcome of an intensive series of interrater reliability training sessions for the 17-item Hamilton Depression Rating Scale conducted at the start of a Phase II multisite clinical drug trial. The data underscore the magnitude of error present in such a test setting and provide preliminary evidence for the potential effect of this problem on the detection of clinical change.

Adult↗

PubMed2007

Estimating the response rate in the presence of measurement error.

In clinical research, it is often of interest to estimate the response rate (i.e. the proportion of subjects who achieve a clinically meaningful threshold) for a particular variable. The standard estimator of the response rate is generally biased in the presence of measurement error. The estimation accounting for the measurement error utilizing fully nonparametric (NP) methods is complicated and may not be efficient. Therefore, we propose a model-based approach assuming a parametric model for the true value and only the first few moments for the measurement error. The estimator for the true response rate and the variance for the estimator are derived. An innovative method using bootstrap simulation is proposed to check the model assumption. Simulations show that the proposed estimator outperforms a fully NP estimator if the model assumption for X holds. This method is applied to address a commonly occurring question in osteoporosis regarding response to treatment in terms of longitudinal changes in bone mineral density (BMD). Bootstrap simulations showed that the model utilized is appropriate. The proposed method can also be applied in other fields of clinical research.

Absorptiometry, Photon↗

PubMed1990

The effects of measurement error on previously reported mathematical relationships between indicator organism density and swimming-associated illness: a quantitative estimate of the resulting bias.

Several recent epidemiological studies seeking associations between swimming in recreational waters contaminated with domestic sewage and increased illness among such swimmers have reported mathematical relationships relating indicator organism densities to illness among swimmers. Common to the design of all of these studies is the failure to adequately control for large amounts of measurement error contained in estimates of exposure, i.e. estimated indicator organism densities. The limited precision of current methods of indicator organism enumeration, coupled with temporal and spacial variation in indicator organism densities at the locations studied, are responsible for a substantial portion of this measurement error. The failure to control adequately for these sources of measurement error has resulted in a significant amount of bias being present in the mathematical relationships reported by these previously published epidemiological studies. In order to explore the magnitude and direction of this bias, computer simulations were conducted using data in which estimation of indicator organism density was obtained by the two most widely used techniques of enumeration: the Multiple Tube Fermentation Technique and the Membrane Filtration Technique. The results of the computer simulations show that the bias caused by this measurement error is non-differential causing the mathematical relationships between indicator organism density and swimming-associated illness reported in previous epidemiological studies to underestimate true risk by a minimum of approximately 14%, and that this underestimate could range as high as approximately 30 to 57%. This study also demonstrates that substantial reduction of this bias can be easily accomplished by incorporating a formal water quality sampling strategy, based on statistical principles of experimental design and analysis, into the design of future epidemiological studies seeking mathematical relationships between indicator organism density and swimming-associated illness.

Analysis of Variance↗

PubMed2004

Musculoskeletal responses of 2-year-old Thoroughbred horses to early training. 2. Measurement error and effect of training stage on the relationship between objective and subjective criteria of training workload.

AIM: To identify the magnitude and sources of error involved in simply quantifying the workload of a group of Thoroughbred racehorses in training. METHODS: Data were obtained from seven 2-year-old fillies trained for 13 weeks on sand and grass tracks at various cantering velocities, and at the gallop in the last 4 weeks. Four fillies completed training and comprised a 'high exercise intensity' subgroup; three fillies did not complete all gallops due to lameness or injury and comprised a 'medium exercise intensity' subgroup. The trainer recorded the distance, time, track used, and subjective gait (SG; viz canter, 1/4-pace, 1/2-pace and gallop) of each horse each day. Data were analysed to describe the training programme and sources of variation in workload. A cumulative workload index (CWI) was devised to simply express distance and velocity aspects of the training programme. RESULTS: The magnitude of errors in measurement of timing and distance were small but significant (<0.8%, p=0.01). SG was an inaccurate descriptor of velocity. Mean velocity of 1/2-pace decreased from 8.92 (SE 0.049) in Weeks 6-8 to 6.33 (SE 0.061) m/sec after the start of galloping sessions in Week 9 (p=0.01). Track surface had a significant effect on the velocity of 1/2-pace (p=0.001). The fastest velocity was recorded on the grass (9.06, SE 0.292 m/sec), followed by 'plough' (8.23, SE 0.152 m/sec), and sand (7.90, SE 0.132 m/sec) tracks. The CWI of the horses ranged between 1,085,000 and 1,634,000 (mean=1,428,000; SE 78,000). The average exercise intensity per week was 110,000 (SE 1,048). There were no significant differences between high and medium exercise intensity subgroups for the canter and 1/4-pace. The high exercise subgroup had a greater CWI for the 1/2-pace (839,000 vs 559,000; p=0.004), gallop (71,000 vs 31,000; p=0.001), and total index (1,580,000 vs 1,242,000; p=0.001). CONCLUSIONS: The workload history of a horse can be accurately described using an index based on manual measurement of the time and distance worked. CLINICAL RELEVANCE: Accurate and precise data on the training programme of performance horses are likely to enable greater differentiation of aspects of workload which are associated with both performance and injury.

Journal Article↗

PubMed1993

Measurement error in dietary assessment: an investigation using covariance structure models. Part II.

In Part I we presented a covariance structure model for analysing measurement error in the assessment of nitrogen intake. In this paper we include data on urine nitrogen excretion which allows a critical assessment of the model proposed. Inclusion of urine nitrogen data produces more pessimistic estimates of the quality of dietary intake measurements and shows that previous assumptions about independence of measurement error were wrong. This underscores the need for well founded assumptions in the analysis of measurement error.

Bayes Theorem↗

PubMed1999

The problem of measurement error in modelling the effect of compliance in a randomized trial.

This paper explores the implications of measurement error in the analysis of compliance-response relationships in data from randomized trials. Given that compliance measures are rarely, if ever, error-free indicators of exposure it is argued that both the designs for the collection of compliance data and the statistical models for their resulting analysis should be changed to take the possibility of measurement error into account. An analysis which ignores measurement error in the compliance measurements will provide biased estimates of compliance-response relationships. Provided that one has two or more indicators of compliance for each subject, more appropriate models can be fitted using covariance structure modelling software. If one wishes to explore interactions from repeated measures data on both compliance and response then it is also important that one recognizes that the response measures are also error-prone and that they too are dealt with appropriately.

Bias↗

PubMed1982

Sensitivity of trends in geometric mean blood levels to random measurement errors.

A statistical model is investigated that expresses observations, such as blood lead levels, as an additive function of true levels and random measurement errors. Both empirical results (obtained from a series of computer simulation experiments) and theoretical results indicate how certain summary statistics for the observations vary in response to random measurement errors. Such results are applied to a very large data base of pediatric blood lead levels collected in New York City during 1970-1976, and they indicate that the observed trends in geometric mean blood lead levels are not significantly altered by the possible presence of measurement errors.

Child, Preschool↗

PubMed2001

The effects of outcome misclassification and measurement error on the design and analysis of therapeutic equivalence trials.

In any clinical trial, the use of imperfect diagnostic procedures or laboratory techniques may lead to misclassification and measurement error in the primary outcome. Although the effects of non-differential outcome misclassification and measurement error on conventional superiority trials have been extensively investigated, less is known about the impact of these errors on the results and interpretation of therapeutic equivalence trials. In this paper we formally investigate the effects of outcome misclassification and measurement error on the estimates of treatment effects, type I error rate, and power of equivalence trials. Our results indicate that, contrary to what one may expect based on the well known attenuating effects of non-differential error in conventional studies, these errors do not always favour the goal of demonstrating equivalence. The magnitude and direction of the influence depend on a number of factors including the nature of the outcome variable, specific formulation of equivalence, size of the error rates, and assumptions regarding the true treatment effect.

Estrogen Replacement Therapy↗

PubMed1992

Proportional hazards model with covariates subject to measurement error.

When covariates of a proportional hazards model are subject to measurement error, the maximum likelihood estimates of regression coefficients based on the partial likelihood are asymptotically biased. Prentice (1982, Biometrika 69, 331-342) presents an example of such bias and suggests a modified partial likelihood. This paper applies the corrected score function method (Nakamura, 1990, Biometrika 77, 127-137) to the proportional hazards model when measurement errors are additive and normally distributed. The result allows a simple correction to the ordinary partial likelihood that yields asymptotically unbiased estimates; the validity of the correction is confirmed via a limited simulation study.

Adult↗

PubMed1997

Transformations to additivity in measurement error models.

In many problems, one wants to model the relationship between a response Y and a covariate X. Sometimes it is difficult, expensive, or even impossible to observe X directly, but one can instead observe a substitute variable W that is easier to obtain. By far, the most common model for the relationship between the actual covariate of interest X and the substitute W is W = X + U, where the variable U represents measurement error. This assumption of additive measurement error may be unreasonable for certain data sets. We propose a new model, namely h(W) = h(X) + U, where h(.) is a monotone transformation function selected from some family H of monotone functions. The idea of the new model is that, in the correct scale, measurement error is additive. We propose two possible transformation families H. One is based on selecting a transformation that makes the within-sample mean and standard deviation of replicated W's uncorrelated. The second is based on selecting the transformation so that the errors (U's) fit a prespecified distribution. Transformation families used are the parametric power transformations and a cubic spline family. Several data examples are presented to illustrate the methods.

Adult↗

PubMed1987

Computer modelling of measurement error in longitudinal lung function data.

The effect of measurement error on the accuracy of results in two epidemiological study designs involving longitudinal lung function data was assessed using computer modelling. Five realistic data sets (cohorts) were created, each of 400 subjects, half of whom were exposed to an agent whose effects approximated in magnitude that of cigarette smoking. In each cohort, FEV1 decline was modelled after 6, 4 and 2 years of observation with and without error in the measurement of level of FEV1. For each length of observation the effect of exposure on decline was estimated using a follow-up design comparing the FEV1 decline between exposure groups, and a case-control design comparing risk of exposure in subjects in the top 20th percentile of FEV1 decline (cases) to exposure in those in the bottom 20th percentile (controls). For both study designs an exposure effect at p less than or equal to 0.01 could only be consistently detected after 6 years of observation.

Computer Simulation↗

PubMed1997

Regression models for recurrent event data: parametric random effects models with measurement error.

Statistical methodology is presented for the statistical analysis of non-linear measurement error models. Our approach is to provide adjustments for the usual maximum likelihood estimators, their standard errors and associated significance tests in order to account for the presence of measurement error in some of the covariates. We illustrate the technique with a mixed effects Poisson regression model for recurrent event data applied to a randomized clinical trial for the prevention of skin tumours.

Humans↗

PubMed2004

Measurement error of two different techniques for brain atrophy assessment in multiple sclerosis.

The measurement errors associated with two techniques for brain atrophy assessment in multiple sclerosis (MS) were estimated. The measurement error of a semiautomated, nonnormalized technique with partial brain coverage was about twofold higher than that of a fully automated, normalized technique with full brain coverage (Structural Image Evaluation of Normalized Atrophy software [SIENA]). As shown by the analysis of data from a clinical trial of glatiramer acetate, the use of SIENA increases the study power to detect a treatment effect on brain volume changes in MS patients.

Atrophy↗

PubMed2003

The impact of ignoring measurement error when estimating sample size for epidemiologic studies.

The author presents two examples illustrating the bias in sample-size estimates that can result from ignoring measurement error among study variables. The first example examines the impact of ignoring misclassification of the study's outcome variable on the accuracy of sample-size estimates. In addition, the author outlines a simple yet effective means of adjusting sample-size estimates to account for outcome misclassification. In the second example, the author illustrates the potential for severe underestimation of required sample size in studies using linear regression to evaluate associations between the outcome of interest and an independent variable subject to classical measurement error. The author concludes with a discussion of pertinent literature that might be helpful to study planners interested in adjusting sample-size estimates to account for measurement errors in both outcome and predictor variables.

Epidemiologic Studies↗

PubMed2000

Iron stores and coronary artery disease: a clinical application of a method to incorporate measurement error of the exposure in a logistic regression model.

Rates of coronary artery disease (CAD) increase sharply after menopause. We examined the hypotheses that high iron stores, as measured by plasma ferritin levels, are a risk factor for CAD and that the increase in iron stores after menopause is at least in part responsible for the rise in CAD in women. We also investigated measurement error of plasma ferritin using a Bayesian conditional independence model and incorporated it into the estimation of the odds ratio (OR) for males. Cases had >/=1 coronary artery stenosis >/=70%. Controls had no visible coronary lesions on angiography. The median plasma ferritin level was 48 mg/L (interquartile range: 28 to 86) among 244 cases and 45 mg/l (24 to 85) among 140 controls. The multivariate analyses among females, males, and females and males combined did not support an association between plasma ferritin levels and CAD (OR for one unit change in log ferritin 1.01, 95% CI 0.71-1.44, OR 0.95, 95% CI 0.66-1.37 and OR 0.95, 95% CI 0.75-1.21, respectively). Accounting for the measurement error of ferritin in males slightly improved the precision of the estimate of the OR but did not unmask an association (OR: 0.94, 95% CI 0.69-1.30). We conclude that high ferritin levels before or after menopause are not associated with CAD. Measurement error might be considered in situations where a one-time measurement is assumed to be representative of long-term exposure.

Aged↗

PubMed2004

Impact of measurement error in the study of sexually transmitted infections.

Measurement is a fundamental part of all scientific research, and the introduction of errors of different sorts is an inevitable part of the measurement process in epidemiological and clinical research. Despite the ubiquity of measurement error in research, the substantial impacts which measurement error can have on data and subsequent study inferences are frequently overlooked. This review introduces the basic concepts of measurement error that are most relevant to the study of sexually transmitted infections, and demonstrates the impacts of several of the most common forms of measurement error on study results. A self assessment test and MCQs follow this paper.

Confounding Factors, Epidemiologic↗