Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “Sampling Errors”

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

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

At least 253 records · Page 14Linked to original sources
PubMed1994

[Quality control of intraoperative diagnosis. Annual review of 1490 frozen sections].

A retrospective evaluation of the accuracy of frozen section diagnoses over 1-year period has been carried out. Of 20998 surgical pathology specimens accessioned during 1991, 1265 (6.0%) had intraoperative pathologic consultation; 1490 were the frozen section examinations effectively carried out, with an average of 1.2 frozen section diagnoses per case. Of all frozen sections performed, 4.6% were deferred; in the remaining cases, the concordance between frozen sections and the final histologic diagnoses was 97%. The reasons for diagnostic discordances were gross sampling error (51.2%), misinterpretation (44.2%), microscopic sampling (9.3%), technical problems in sectioning (2.3%). We believe that a similar continuous monitoring should be performed in every pathology department, to recognize the reasons of errors and, if possible, to reduce them.

Academies and Institutes↗
PubMed2003

Modelling trends in woody vegetation structure in semi-arid Australia as determined from aerial photography.

Accounting of carbon stocks in woody vegetation for greenhouse purposes requires definition of medium term trends with accurate error assessment. Tree and shrub cover was sampled through time at randomly located sites over a large area of central Queensland, Australia using aerial photography from 1945 to 1999. Calibration models developed from field data for the same land types as those represented within the study area allowed for the extrapolation of overstorey and understorey cover, basal area and biomass values and these were modelled as trends over the latter half of the 20th century. These structural attributes have declined over the region because of land clearing with values for biomass changing from a mean of 58.0(+/-1.2)t/ha in 1953 to 41.1(+/-1.0)t/ha in 1991. The biomass of Acacia on clay and Eucalypt on texture contrast soils land types has declined most dramatically. Within uncleared vegetation there was an overall trend of increase from 56.1(+/-1.2)t/ha in 1951 to 67.6(+/-1.3)t/ha in 1995. The increase in structural attributes within uncleared vegetation was most pronounced for the Eucalypt on texture contrast soils and Eucalypt on clay land types. It was demonstrated that the sites sampled were representative of their land types and that spatial bias of the photography, undetected tree-killing, sampling error, inherent variability of structural attributes and measurement error should not have impacted greatly on bias or precision of trend estimates for well-sampled land types. Certainly the errors are not likely to be substantial for trends averaged over all land types and they provide an accurate assessment of the magnitude and direction of change. The technique presented here would appear to be a robust means of accounting for the above-ground woody component of woodlands and open forests and will also contribute to a broader understanding of savanna dynamics.

Australia↗
PubMed2006

Confidence intervals for the true classification error conditioned on the estimated error.

Bias and variance for small-sample error estimation are typically posed in terms of statistics for the distributions of the true and estimated errors. On the other hand, a salient practical issue asks, given an error estimate, what can be said about the true error? This question relates to the joint distribution of the true and estimated errors, specifically, the conditional expectation of the true error given the error estimate. A critical issue is that of confidence bounds for the true error given the estimate. We consider the joint distribution of the true error and the estimated error, assuming a random feature-label distribution. From it, we derive the marginal distributions, the conditional expectation of the estimated error given the true error, the conditional expectation of the true error given the estimated error, the conditional variance of the true error given the estimated error, and the 95% upper confidence bound for the true error given the estimated error. Numerous classification and estimation rules are considered across a number of models. Massive simulation is used for continuous models and analytic results are derived for discrete classification. We also consider a breast-cancer study to illustrate how the theory might be applied in practice. Although specific results depend on the classification rule, error-estimation rule, and model, some general trends are seen: (I) if the true error is small (large), then the conditional estimated error is generally high (low)-biased; (II) the conditional expected true error tends to be larger (smaller) than the estimated error for small (large) estimated errors; and (III) the confidence bounds tend to be well above the estimated error for low error estimates, becoming much less so for large estimates.

Algorithms↗
PubMed1986

Parameters that bias the measurement of airborne concentration within a respirator.

This paper describes the theoretical basis upon which a test system has been set up to evaluate the sampling error associated with in-facepiece sampling on half-mask respirators. The in-facepiece sampling technique evaluated in this study is the one currently used in the U.S. to conduct quantitative facepiece fit testing. An experimental design was developed to study the sampling bias associated with in-facepiece sampling when selected parameters of the man/respirator system were varied. The results indicated that significant errors can be made in estimating concentration within a respirator when the current in-facepiece sampling technique is employed. Sampling bias was determined when in-facepiece samples were collected only during the inhalation phase of the respiratory cycles. They were found to range from greater than -99% to greater than +98%. The mean sampling bias was -17 +/- 38%. When measured in-facepiece concentrations were used to calculate a fit factor the resulting range was 44 to 4728 even though the actual fit factor was only 87. Based upon the data presented, it was hypothesized that faceseal leakage was streamlining within the respirator cavity. As a result, quantitative facepiece fit data on half-mask respirators may be biased by the large measurement error.

Acetone↗
PubMed1997

Motor unit number estimation: sample size considerations.

A computer model of the motor unit number estimation procedure was developed to evaluate the sampling error associated with estimates of the number of motor units in muscles. Two different distributions were used to model the motor unit amplitude distribution and were chosen in such a manner that they qualitatively matched the distributions observed under both normal and neurogenic conditions. As expected, the results indicated that estimation error decreases as a function of sample size. However, the relationship between these two variables was nonlinear in the sense that successive increases in sample size lead to progressively smaller decreases in estimation error. The results also indicated that the shape of the motor unit amplitude distribution plays an important role. Specifically, estimates obtained using the distribution modeling normal muscle were generally higher than the actual number of motor units in the muscle, which was not the case for the distribution modeling neurogenic muscle. In addition, the neurogenic distribution was associated with much smaller estimation error, suggesting that motor unit number estimation is well suited to the analysis of neurogenic disease processes.

Cell Count↗
PubMed1991

Protonation of interacting residues in a protein by a Monte Carlo method: application to lysozyme and the photosynthetic reaction center of Rhodobacter sphaeroides.

We used Monte Carlo methods to treat statistical problem of electrostatic interactions among many titrating amino acids and applied these methods to lysozyme and the photosynthetic reaction center of Rhodobacter sphaeroides, including all titrating sites. We computed the average protonation of residues as a function of pH from an equilibrium distribution of states generated by random sampling. Electrostatic energies were calculated from a finite difference solution to the linearized Poisson-Boltzmann equation using the coordinates from solved protein structures. For most calculations we used the Metropolis algorithm to sample protonation states; for strongly coupled sites, we substantially reduced sampling errors by using a modified algorithm that allows multiple site transitions. The Monte Carlo method agreed with calculations for a small test system, lysozyme, for which the complete partition function was calculated. We also calculated the pH dependence of the free energy change associated with electron transfer from the primary to the secondary quinone in the photosynthetic reaction center. The shape of the resulting curve agreed fairly well with experiment, but the proton uptake from which the free energy was calculated agreed only to within a factor of two with the observed values. We believe that this discrepancy resulted from errors in the individual electrostatic energy calculations rather than from errors in the Monte Carlo sampling.

Electron Transport↗
PubMed1991

Total error in PES estimates of population.

"We describe a methodology for estimating the accuracy of dual systems estimates (DSE's) of population, census estimates of population, and estimates of undercount in the census. The DSE's are based on the census and a post-enumeration survey (PES). We apply the methodology to the 1988 dress rehearsal census of St. Louis and east-central Missouri and we discuss its applicability to the 1990 [U.S.] census and PES. The methodology is based on decompositions of the total (or net) error into components, such as sampling error, matching error, and other nonsampling errors. Limited information about the accuracy of certain components of error, notably failure of assumptions in the 'capture-recapture' model, but others as well, lead us to offer tentative estimates of the errors of the census, DSE, and undercount estimates for 1988. Improved estimates are anticipated for 1990." Comments are included by Eugene P. Ericksen and Joseph B. Kadane (pp. 855-7) and Kenneth W. Wachter and Terence P. Speed (pp. 858-61), as well as a rejoinder by Mulry and Spencer (pp. 861-3).

Americas↗
PubMed1996

Doppler grid surface scanning applications for pulmonary subsurface parenchymal perfusion assessment.

Subsurface perfusion to lung parenchyma underlying the pleura is difficult to assess in live ventilated animals. The purpose of this study was to assess applicability of a newly developed laser Doppler grid scanning imaging technology that measures perfusion of pleural subsurface lung regions in intact normal and abnormal animal lungs. Eighty-six Doppler grid perfusion measurements were performed in five New Zealand White Rabbits (3-5 kg); four with unilateral bullous lung disease, one normal control. Left upper lobe lung surface was exposed to 10 1-sec spot Nd:YAG exposures (70 W/cm2). One week following laser exposure, all rabbits underwent sequential bilateral open thoracotomy. Unaffected left lower lobes in these animals and all four lobes of a previously untreated rabbit were used as controls. Pleural subsurface perfusion measurements were recorded over a contiguous 900-pixel square surface grid using quantitative noncontact laser Doppler imaging during open thoracotomy procedures. Scans were obtained in a normal volume ventilation mode, at 30 cm of inspiratory hold airway pressure, and postinflation. A perfusion-pressure response curve was obtained in normal lung at 10-, 20-, and 30-cm static airway pressure. Post mortem measurements were used as 0 flow controls. Normal lung tissue was found to have relatively high pleural subsurface perfusion (1362 +/- 328 corrected units on a scale of 0-4095). Areas of atelectasis had decreased perfusion (659 +/- 512 U., 48.4 +/- 12.5% compared to normal lung, p < 0.02), but returned to normal levels after inflation of the lung (1253 +/- 363 U., p = 0.21 compared to normal). Pleural subsurface perfusion decreased uniformly and progressively as lung inflation pressure increased (p < 0.0001). Perfusion increased immediately to supranormal values following release of high inspiratory inflation pressure holds (1603 +/- 626 U., 117 +/- 18% compared to normal lung, p = 0.03). Bullae had markedly decreased perfusion (541 +/- 68 U.) that was not further reduced by increased inflation pressures. Noncontact laser Doppler grid perfusion imaging appears to provide a new tool for measuring pleural subsurface perfusion over a large area of lung surface in clinical experimental settings. Results are rapid, reproducible, and consistent. Sampling errors inherent in current point sampling Doppler flow techniques are reduced by the multiple contiguous measurements. We have used this technique to demonstrate inspiratory pressure-related reduction in pleural subsurface perfusion in normal lung, reversible decreased perfusion in atelectatic regions, and reduced perfusion in bullous and laser-treated lung regions.

Analysis of Variance↗
PubMed2003

Fixed precision sequential sampling plans for the greenbug and bird cherry-oat aphid (Homoptera: Aphididae) in winter wheat.

The numbers of greenbugs, Schizaphis graminum (Rondani), and bird cherry-oat aphids, Rhopalosiphum padi L., per wheat tiller (stem) were estimated in 189 production winter wheat (Triticum aestivum L.) fields located throughout Oklahoma. Taylor's power law regressions were calculated from these data and used to construct fixed precision sequential sampling schemes for each species. An evaluation data set was constructed from 240 samples taken during three growing seasons from winter wheat fields at four locations in Oklahoma. Wheat cultivar and growth stage were recorded for each field on the day of sampling. Taylor's power law parameters for evaluation fields differed significantly for both species among growing seasons, locations, and plant growth stages. Median precision achieved using the fixed precision sequential sampling schemes for each species departed <20% from expected precision over the range population intensity in the evaluation data. For the 10% of samples with greatest deviation between observed and expected precision, observed precision was 13.8-81.8% greater than that expected precision depending on aphid species and population intensity. For the greenbug, the distribution of the percentage deviation between observed and expected precision was positively skewed, so that the sampling scheme tended to over-predict precision. For the bird cherry-oat aphid, the distribution was more symmetric. Even though precision observed using the sampling schemes frequently varied from expected precision, because of the inevitable consequence of sampling error and environmental variation, the sampling schemes yielded median observed precision levels close to expected precision levels over a broad range of population intensity.

Agriculture↗
PubMed2004

An application of meta-analysis in food safety consumer research to evaluate consumer behaviors and practices.

Meta-analysis provides a structured method for combining results from several studies and accounting for and differentiating between study variables. Numerous food safety consumer research studies often focus on specific behaviors among different subpopulations but fail to provide a holistic picture of consumer behavior. Combining information from several studies provides a broader understanding of differences and trends among demographic subpopulations, and thus, helps in developing effective risk communication messages. In the illustrated example, raw/undercooked ground beef consumption and hygienic practices were evaluated according to gender, ethnicity, and age. Percentages of people engaging in each of the above behaviors (referred to as effect sizes) were combined using weighted averages of these percentages. Several measures, including sampling errors, random variance between studies, sample sizes of studies, and homogeneity of findings across studies, were used in the meta-analysis. The statistical significance of differences in behaviors across demographic segments was evaluated using analysis of variance. The meta-analysis identified considerable variability in effect sizes for raw/undercooked ground beef consumption and poor hygienic practices. More males, African Americans, and adults between 30 and 54 years (mid-age) consumed raw/undercooked ground beef than other demographic segments. Males, Caucasians, and Hispanics and young adults between 18 and 29 years were more likely to engage in poor hygienic practices. Compared to traditional qualitative review methods, meta-analysis quantitatively accounts for interstudy differences, allows greater consideration of data from studies with smaller sample sizes, and offers ease of analysis as newer data become available, and thus, merits consideration for its application in food safety consumer research.

Adult↗
PubMed1985

Time sampling and measurement error: the effect of interval length and sampling pattern.

The accuracy of momentary time-sampling was evaluated for behaviors occurring for different proportions of time, for different momentary time-samples and for sessions of different lengths. For 1-hr sessions, momentary time-samples of 10, 20 or 30 see intervals between observations yielded 90% of estimates within 10% of the comparison standard, for behaviours of 25%, 50% and 70%. Behaviours occurring for 2% of the session were not accurately recorded even at 10-sec momentary time-samples. For a 5-min momentary time-sample, this criterion was only approached when 8 hr data was averaged to give a single score for each individual, for behaviours of 25% or more. The 5-min momentary time-sample was also compared with a 1 in 5-min cluster sample; although less accurate, cluster-sampling might be a useful strategy in some settings.

Behavior↗
PubMed1983

Adaptive control of theophylline therapy: importance of blood sampling times.

A two-observation protocol for estimating theophylline clearance during a constant-rate intravenous infusion is used to examine the importance of blood sampling schedules with regard to the information content of resulting concentration data. Guided by a theory for calculating maximally informative sample times, population simulations are used to assess the effect of specific sampling times on the precision of resulting clearance estimates and subsequent predictions of theophylline plasma concentrations. The simulations incorporated noise terms for intersubject variability, dosing errors, sample collection errors, and assay error. Clearance was estimated using Chiou's method, least squares, and a Bayesian estimation procedure. The results of these simulations suggest that clinically significant estimation and prediction errors may result when using the above two-point protocol for estimating theophylline clearance if the time separating the two blood samples is less than one population mean elimination half-life.

Blood Specimen Collection↗
PubMed1998

Detection of occult breast cancer micrometastases in axillary lymph nodes using a multimarker reverse transcriptase-polymerase chain reaction panel.

BACKGROUND: Axillary lymph node status in breast cancer patients remains the single most important predictor of outcomes. Current methods of histopathologic analysis may be inadequate because 30% of node-negative patients recur. The purpose of this study was to test the hypothesis that a multigene reverse transcriptase-polymerase chain reaction (RT-PCR) panel provides a more sensitive method to detect axillary lymph node metastases than routine pathologic examination. STUDY DESIGN: Sixty-one consecutive breast cancer patients were evaluated, with nine normal control patients. Nodes > 1 cm were bisected for histopathologic and RT-PCR analysis. Nodal tissue was homogenized, and total RNA was converted into cDNA with reverse transcriptase. Reverse transcriptase-polymerase chain reaction analysis was performed with primers specific for keratin-19, c-myc, prolactin inducible protein (PIP), and beta-actin using ethidium bromide gel electrophoresis. Reverse transcriptase-polymerase chain reaction positive/ pathology negative axillary lymph nodes were reevaluated using step sectioning and immunohistochemical staining. RESULTS: Thirty-seven patients had pathologically negative axillary lymph nodes, of which 15 (40%) were positive by RT-PCR analysis. Two RT-PCR negative results (one probably from tissue processing error and the other secondary to sampling error) among the 24 histologically positive specimens were detected (8%). The number of patients in each pathologic stage was 26 patients in stage I; 18, stage IIA; 7, stage IIB; 7, stage IIIA; 3, stage IIIB; and 0 patients in stage IV. By RT-PCR staging, 8 of 26 patients went from stage I to IIA (30%), and 7 of 18 from stage IIA to IIB (39%). Of the RT-PCR positive individuals who were stage I by pathologic analysis, 100% were found to be c-myc positive, 0% keratin-19 positive, and 0% PIP positive; for stage IIIB patients these markers were 50%, 100%, and 100% respectively. Additionally, an increasing number of positive markers per specimen appeared to correlate with larger primary tumor size (p < 0.01) and decreased predicted 5-year survival (r = 0.950, p < 0.002). CONCLUSIONS: Multimarker RT-PCR analysis appears to be a readily available and highly sensitive method for the detection of axillary lymph node micrometastases. Longterm followup of RT-PCR positive patients will be required to determine its clinical relevance. If validated as a predictor of disease recurrence, this method would provide a powerful complement to routine histopathologic analysis of axillary lymph nodes.

Axilla↗
PubMed2006

Combined influence of maternal and paternal quality on sex allocation in red-capped robins.

Sex allocation theory predicts females will adaptively manipulate sex ratios to maximize their progeny's reproductive value. Recently, the generality of biased sex allocation in birds has been questioned by meta-analytic reviews, which demonstrate that many previously reported significant results may simply reflect sampling error. Here, we utilize a robust sample size and powerful statistical approach to determine whether parental quality is correlated with biased sex allocation in red-capped robins. Indices of maternal quality (including interactive effects of age and condition) were strongly related to sex allocation. These relationships were in the predicted directions, with larger effect sizes than those of previous studies in this field. There were also paternal correlates, involving age and the source of paternity. We propose that biased sex allocation occurs in this species, and is maintained by differing production costs of each sex and genetic benefits to females of producing sons when fertilized by high-quality males.

Animals↗
PubMed2001

Tissue viral load variability in chronic hepatitis C.

OBJECTIVE: Liver biopsy is regarded as the gold standard for assessing disease activity in chronic hepatitis C, but sampling error is a potential limitation. Whether sampling variability applies equally to viral load assessment as it does to histology is uncertain. To examine this, we compared viral load between right- and left-lobe biopsy specimens from patients infected with hepatitis C virus (HCV). METHODS: Bilobe biopsies were taken from 16 patients who were serum positive for HCV RNA by reverse transcription-polymerase chain reaction. Genotype was identified by reverse line probe hybridization. There was an absence of competing risk factors for infectious and other liver diseases in this patient group. Histology and hepatic viral load were assessed blindly. None of the patients had received antiviral therapy at the time of study. RESULTS: Detection of HCV in right and left lobes was concordant with serum positivity in all cases. The viral load between lobes was highly correlated (p = 0.0003, r = 0.79). In contrast, the histological activity indices of inflammation and fibrosis/cirrhosis were poorly correlated between lobes (p = 0.038, r = 0.60, and p = 0.098, r = 0.50, respectively). CONCLUSION: Hepatic viral load variability does not suffer from the same degree of heterogeneity of sampling variability as does histology.

Adult↗
PubMed2005

Anatomic pathology databases and patient safety.

CONTEXT: The utility of anatomic pathology discrepancies has not been rigorously studied. OBJECTIVE: To outline how databases may be used to study anatomic pathology patient safety. DESIGN: The Agency for Healthcare Research and Quality funded the creation of a national anatomic pathology errors database to establish benchmarks for error frequency. The database is used to track more frequent errors and errors that result in more serious harm, in order to design quality improvement interventions intended to reduce these types of errors. In the first year of funding, 4 institutions (University of Pittsburgh, Henry Ford Hospital, University of Iowa, and Western Pennsylvania Hospital) reported cytologic-histologic correlation error data after standardizing correlation methods. Root cause analysis was performed to determine sources of error, and error reduction plans were implemented. PARTICIPANTS: Four institutions self-reported anatomic pathology error data. MAIN OUTCOME MEASURES: Frequency of cytologic-histologic correlation error, case type, cause of error (sampling or interpretation), and effect of error on patient outcome (ie, no harm, near miss, and harm). RESULTS: The institutional gynecologic cytologic-histologic correlation error frequency ranged from 0.17% to 0.63%, using the denominator of all Papanicolaou tests. Based on the nongynecologic cytologic-histologic correlation data, the specimen sites with the highest discrepancy frequency (by project site) were lung (ranging from 16.5% to 62.3% of all errors) and urinary bladder (ranging from 4.4% to 25.0%). Most errors detected by the gynecologic cytologic-histologic correlation process were no-harm events (ranging from 10.7% to 43.2% by project site). Root cause analysis identified sources of error on both the clinical and pathology sides of the process, and error intervention programs are currently being implemented to improve patient safety. CONCLUSIONS: A multi-institutional anatomic pathology error database may be used to benchmark practices and target specific high-frequency errors or errors with high clinical impact. These error reduction programs have national import.

Benchmarking↗
PubMed1997

The accuracy of Papanicolaou smear predictions: cytohistological correlation of 283 cases.

The Papanicolaou smear is a highly effective screening test for the detection of cervical neoplastic changes. The success of the test has resulted in unrealistic expectations of the accuracy of the test by both referring medical practitioners and the public. However, as with any pathological test, it has irreducible false negative and positive rates. This report is a comparison between interpretations based on cytological and histological tests and was undertaken to estimate the sensitivity of the Papanicolaou test as practised in Hong Kong. The overall absolute concordance rate for the study was 51.2%. The concordance rates within one diagnostic category were 63.9% and 74.6% for low- and high-grade squamous intraepithelial lesions respectively. The overall sensitivity of the test was 91.7% with a positive predictive value of 93.5%. Ten percent of the error rate was attributed to laboratory error; the remainder was attributed to sampling error and poor smear preparation. Forty-five percent of cases of atypical squamous cells of undetermined significance showed evidence of cervical intraepithelial neoplasia on subsequent biopsy. Follow-up biopsies of low-grade squamous intraepithelial lesions also showed as many lesions from cervical intraepithelial neoplasia grade I as from grades II and III. These findings suggest that colposcopies and biopsies should be performed as soon as possible rather than to repeat the smears in 3 to 6 months. The results of the study may provide guidelines for formulating follow-up recommendations.

Journal Article↗
PubMedDate not supplied

[P value and confidence intervals: reporting and interpreting the result of a clinical study].

The main purpose of statistics in the analysis of clinical and epidemiological studies is to summarize data and information, as well as assess variability, trying to distinguish between chance findings and results that may be replicated upon repetition. Statistical analyses only convey the effect of chance element in data (random error). Statistics cannot control non-sampling errors concerning study design, conduct and methods adopted. At the end of the study, a result is defined statistically significant if the observed difference in the outcome variable is too large to be attributed to chance. A small P value provides evidence against the null hypothesis (of no effect), since data have been observed that would be unlikely if the null hypothesis was true. However, confidence intervals estimate separate the two data dimensions (strength of the relation between exposure and disease, and precision with which the relation is measured), and add to the hypothesis testing useful information for finding interpretation and further research.

Biomedical Research↗