Download stats - Streck

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts
no text concepts found
Transcript
STATS
®
I N T ER L A B O R ATO RY Q UA LI T Y
C O N T RO L PRO G R A M
Q.C. Concepts
Accuracy
Refers to how close or similar data is compared to the true
value.
Assay
Instrument specific values provided by the control
manufacturer.
Assay Range
Low and High Limits are determined by applying 2x the
range to the assay mean.
Bias
The positive or negative difference from the peer group
mean or true value.
Mean
Mathematical average of all data points. Any change in
accuracy – known as a shift or trend – will be reflected as
a change in the mean.
Outlier
A value that is removed from calculations if it falls outside
the assay range, or is removed during a Grubbs Test
calculation.
Peer Group
A grouping of similar instrument types for comparison.
PI – Precision Index
Measure of relative precision. A value less than 1 indicates
good lab precision.
CV – Coefficient of Variation
Measures a method’s performance over multiple runs
(data dispersion compared to the mean). A small CV
equals good precision.
PI is calculated as follows and is represented as a
percentage:
CV is calculated as follows and is represented as a
percentage:
Precision
An indication of the amount of random error that exists in
the data or process
%CV = (SD ÷ Mean) x 100%
Note – The CV may be large on some parameters due to
the small numeric value, i.e. differential parameters, and/
or the small number of significant digits.
Iterative Grubbs Test
Streck uses a two-tailed Grubbs Test applied iteratively to
customer data to determine outliers. The first step in the
calculation is to determine whether or not the maximum
value is an outlier (either > or < the mean).
G = I Outlier Value - Mean I
Standard Deviation
Look up the critical value by using the number of included
data points in the dataset. If G > the critical value then the
value is an outlier. The process is then applied again until
no more outliers are found.
Levey-Jennings Graphs
The graphs provide a visual plot of the data compared to
the assay value, peer group mean and 2 SD range from
the peer group mean. Each mark on the graph indicates
how far off the actual result was from the mean. This
allows at-a-glance review of the data and helps identify
trends, shifts and outliers. When the control values fall
within the expected distribution, the values are classified
as being “in-control” and results can be accepted.
800.898.9563 | streck.com
%PI = Your %CV ÷ Group %CV
Real-Time Reports
A real-time report can be run before data is processed
at the report date. A submission must be submitted for
the current report date before a real-time report option is
available.
SD – Standard Deviation
Refers to the precision within a set of data. The distance
from the mean is measured in standard deviations. A
high SD indicates a wide distribution of values and poor
precision in the results.
S.D.I. – Standard Deviation Index
The S.D.I. is a calculated value for each parameter that
indicates the degree of separation between the peer
group mean and an individual lab’s mean. This allows
laboratories to quickly compare their results on each level
of control to results from peers. The closer the S.D.I. is to
zero, the closer that result is to the target mean. As a rule
of thumb, an S.D.I. greater than 2.0 (+2 or -2), should be
reviewed.
S.D.I. is calculated as follows:
S.D.I. = (Your Mean – Peer Group Mean)
Peer Group SD
880018-4
STATS
®
I N T ER L A B O R ATO RY Q UA LI T Y
C O N T RO L PRO G R A M
Q.C. Concepts
Shift
An abrupt change in the pattern of data points on a plot, graph
or quality control file.
TE – Total Error
Refers to how far results vary in total from the peer group
mean or true value.
TE is calculated as follows:
TE = |Bias| + 2SD
Trend
A gradual change in the pattern of data points on a plot, graph
or quality control file.
Special Considerations for Linearity Products
R Value
The R-Value measures the correlation between X and Y. The
desired value is as close to 1 as possible.
Best Fit Line
Linearity products are graphed utilizing a best fit line. The
standard formula for best fit line is utilized and is y=mx + b.
The x value is the CVA Multipliers and the y value represents
means.
800.898.9563 | streck.com
880018-4