Statistics assessment bias

Historical data on the indicator. Existing information on the statistical variation, bias, and other interpretational attributes of potential biological indicators should be examined and considered in the design of a sampling program for assessing trends in mercury bioaccumulation. 

The typical user of an SRM performs a measurement of the certified property using the apparatus in his laboratory. He compares his result with the certificate value and, if his measurement fails to reproduce the certified value within acceptable limits of precision he investigates his measurement system to remove sources of bias. There are numerous examples of measurement systems in which the level of agreement among measured results by employing SRM s in this way, and the statistical assessment of the performance of measurement systems employing SRM s is a well-established science (3). 

The calculation of a statistical test and the obtaining of a value for P addresses only chance as the possible explanation for the difference. Being statistically significant is therefore not a sufficient basis for a conclusion that the difference is real - allocation and assessment bias must also be considered. 

Inspection of a set of replicate measurements or results may reveal that one or more is considerably higher or lower than the remainder and appears to be outside the range expected from the inherent effects of indeterminate (random) errors alone. Such values are termed outliers, or suspect values, because it is possible that they may have a bias due to a determinate error. On occasions, the source of error may already be known or it is discovered on investigation, and the outlier(s) can be rejected without recourse to a statistical test. Frequently, however, this is not the case, and a test of significance such as the Q-test should be applied to a suspect value to determine whether it should be rejected and therefore not included in any further computations and statistical assessments of the data. 

In a performance-based approach to quality assurance, a laboratory is free to use its experience to determine the best way to gather and monitor quality assessment data. The quality assessment methods remain the same (duplicate samples, blanks, standards, and spike recoveries) since they provide the necessary information about precision and bias. What the laboratory can control, however, is the frequency with which quality assessment samples are analyzed, and the conditions indicating when an analytical system is no longer in a state of statistical control. Furthermore, a performance-based approach to quality assessment allows a laboratory to determine if an analytical system is in danger of drifting out of statistical control. Corrective measures are then taken before further problems develop. 

Definition and Uses of Standards. In the context of this paper, the term "standard" denotes a well-characterized material for which a physical parameter or concentration of chemical constituent has been determined with a known precision and accuracy. These standards can be used to check or determine (a) instrumental parameters such as wavelength accuracy, detection-system spectral responsivity, and stability (b) the instrument response to specific fluorescent species and (c) the accuracy of measurements made by specific Instruments or measurement procedures (assess whether the analytical measurement process is in statistical control and whether it exhibits bias). Once the luminescence instrumentation has been calibrated, it can be used to measure the luminescence characteristics of chemical systems, including corrected excitation and emission spectra, quantum yields, decay times, emission anisotropies, energy transfer, and, with appropriate standards, the concentrations of chemical constituents in complex S2unples. 

Cochrane Library. The Cochrane Library [44] includes The Cochrane Database of Systematic Reviews, a collection of regularly updated, systematic reviews of the effects of health care. It is maintained by contributors to the Cochrane Collaboration. Cochrane reviews are reviews mainly of randomized controlled trials. To minimize bias, evidence is included or excluded on the basis of explicit quality criteria. Data are often combined statistically, with meta-analysis, to increase the power of the findings of numerous studies, each too small to produce reliable results individually. Database of Abstracts of Reviews of Effectiveness is also included. It consists of critical assessments and structured abstracts of good systematic reviews published elsewhere. The Cochrane Controlled Trials Register with bibliographic information on controlled trials and other sources of information on the science of reviewing research and evidence-based health care are part of the Cochrane Library. It is commercially available on CD-ROM or the Internet. 

II the difference approach, which typically utilises 2-sided statistical tests (Hartmann et al., 1998), using either the null hypothesis (H0) or the alternative hypothesis (Hi). The evaluation of the method s bias (trueness) is determined by assessing the 95% confidence intervals (Cl) of the overall average bias compared to the 0% relative bias value (or 100% recovery). If the Cl brackets the 0% bias then the trueness that the method generates acceptable data is accepted, otherwise it is rejected. For precision measurements, if the Cl brackets the maximum RSDp at each concentration level of the validation standards then the method is acceptable. Typically, RSDn> is set at <3% (Bouabidi et al., 2010),... 

III the equivalence approach, which typically compares a statistical parameters confidence interval versus pre-defined acceptance limits (Schuirmann, 1987 Hartmann et al., 1995 Kringle et al., 2001 Hartmann et al., 1994). This approach assesses whether the true value of the parameter(s) are included in their respective acceptance limits, at each concentration level of the validation standards. The 90% 2-sided Cl of the relative bias is determined at each concentration level and compared to the 2% acceptance limits. For precision measurements, if the upper limit of the 95% Cl of the RSDn> is <3% then the method is acceptable (Bouabidi et al., 2010) or,... 

Human data and their relevance have to be assessed carefully on a case-by-case basis due to limitations of the techniques available. In particular, attention should be paid to the adequacy of the exposure information, confounding factors, and to sources of bias in the smdy design. The statistical power of the test may also be considered. 

In addition to random errors derived from samplings, systematic errors are peculiar to each particular method or system. They cannot be assessed statistically. A major effect of systematic error known as bias is a shift in the position of the mean of a set of readings relative to the original mean. 

The suspicion that Schrodinger s interpretation of wave mechanics was suppressed and rejected by quantum physicists for non-scientific reasons, is inescapable. Because of this inherent bias the form of wave mechanics which became established as the basis of theoretical chemistry has, understandably, never been assessed independently for this purpose. The point electron that jumps between quantum states with statistical probability fails to explain chemical behaviour with the same authority that it enjoys in physics. Nevertheless, the Schrodinger alternative is dismissed out of hand by chemists. A typical expert on quantum chemistry declares [33] ... 

Well-designed proficiency studies provide a good estimate of the method bias. Several protocols and statistical methodologies have been developed for assessing this bias, for example, ISO 5725-4 [28], International Harmonized Protocol for Proficiency studies [29], and Youden plot [30],... 

Random error can over- or underestimate risk and is generally not as severe as bias. Moreover, the magnitude of error can be estimated with statistical techniques. Assessment of confounding, synergism, or effect modification can be accomplished in the analytical phase (by stratification or multivariate modeling), providing sufficient data have been collected on those factors. Restriction or randomization procedures also can be used in the design phase to minimize confounders. 

It should be emphasized that the approach presented in this section is part of an overall assessment of measurement errors. The measurement model is used as a filter for lack of replicacy to obtain a quantitative value for the standard deviation of the measurement as a fvmction of frequency. The mean error identified in this way is equal to zero thus, the standard deviation of the measurement does not incorporate the bias errors. In contrast, the standard deviation of repeated impedance measurements t)q)ically includes a significant contribution from bias errors because perfectly replicate measurements can rarely be made for electrochemical systems. Since the line-shapes of the measurement model satisfy the Kramers-Kronig relations, the Kramers-Kronig relations then can be used as a statistical observer to assess the bias error in the measurement. 

Systematic reviews are recent additions to the medical literature. In contrast to traditional narrative reviews, these reviews aim to answer a precisely defined clinical question and to do so in a way that is transparent and designed to minimize bias. Some of the defining features of systematic reviews are (1) a clear definition of the clinical question to be addressed, (2) an extensive and explicit strategy to find ail studies (published or unpublished) that may be eligible for inclusion in the review, (3) criteria by which studies are included and excluded, (4) a mechanism to assess the quahty of each study and, in some cases, (5) synthesis of results by use of statistical techniques of meta-analysis. By contrast, traditional reviews are subjective, are rarely well focused on a clinical question, lack explicit criteria for selection of studies to be reviewed, do not indicate criteria to assess the quality of included studies, and rarely can use meta-analysis. 

---