For the Physician Office Lab
Data Analysis FAQs:
Questions:
How do I report a non-numeric result on the data submission form?
How can I send you my data for your data reduction service?
Where can I find a detailed explanation of the Data Reduction Report?
Can you provide guidance on interpreting the results?
Can you provide guidance on setting acceptable performance criteria?
What should I do if I get results that fall outside of the applied limits?
Answers:
How do I report a non-numeric result on the data submission form?
Enter your results on the data submission form exactly as the instrument system provides. A non-numeric result does not always imply a ‘zero’ recovery. Reporting a non-numeric result as zero (0) on your data submission form can effect your data analysis.
How can I send you my data for your data reduction service?
Data can be submitted to us via email, fax, mail or using our MSDRx® software. A Data Submission form, included with each shipment of product, is provided for the fax or mail method of submitting data. To submit data online, download our Electronic Data Submission Form, fill it out, and submit it online. We will return your completed reports within five days.
Where can I find a detailed explanation of the Data Reduction Report?
Use the following link to download a copy of our Data Reduction Report Cover Letter.
Or use this link for a more detailed explanation of the Data Reduction Reports.
Can you provide guidance on interpreting the results?
Prior to interpretation, your laboratory should establish acceptable performance criteria for calibration verification / linearity experiments for each analyte. Interpretation should always include consideration from a ‘clinical use’ perspective for each analyte.
Care should be observed when interpreting results that fall outside of the applied limits on the extreme low end of a calibration verification / linearity evaluation. The results are generated solely on the statistical evaluation of the data. Often, the analytical differences are small and do not impact clinical use of the test. Clinical significance of differences needs to be considered in the interpretation of any results.
Can you provide guidance on setting acceptable performance criteria?
For each test performed, the laboratory is responsible for establishing performance specifications for calibration verification and for reportable range verification. When setting limits for non-linearity, the error allowed can be based on either analytical goals or clinical goals. Ideally, the amount of error allowed due to non-linearity should be independent of error that results from bias and other analytical error (imprecision). Bias is typically that error that is the result of calibration and the element of accuracy imparted by calibration set points. Imprecision error can be influenced by a number of variables such as system maintenance and sample integrity. The CLIA limits represent the entire error allowed, which includes all three sources of error outlined above. Each lab should establish what portion of the total error budget is allowed for non-linearity.
Acceptance criteria should always include consideration from a ‘clinical use’ perspective for each analyte. For example, ask the question: with a "Target" of 4.30 mg/dL glucose, is a recovered value of 4.5 mg/dL glucose an acceptable value? This type of logic can be applied to the entire reportable range to help set acceptance criteria.
The CLIA '88 'criteria for acceptable performance' can be found on the internet in two locations:
- For routine chemistry
- For toxicology
What should I do if I get results that fall outside of the applied limits?
Results need to be compared to the acceptance criteria you established for the analyte. Points that fall outside of the applied limits may or may not affect the linear range, depending upon your acceptance criteria.
Care should be observed when interpreting results that fall outside of the applied limits on the extreme low or high end of a calibration verification / linearity evaluation. The results shown in our output table are generated solely on the statistical evaluation of the data – clinical significance is not taken into account. Often, at the low or high end of the evaluation, the statistical differences may not impact clinical use of the test. Clinical significance of differences needs to be considered in the interpretation of any the results. More information on statistical flags on your data reduction reports is provided in this letter.
It is also useful to look at how your results compare to other labs performing the same method of testing. Peer Group Analysis provides statistical evaluation of your results versus your peers.
When I use the data analysis method described in your product insert, I get different target values than those from your data reduction report. Why?
The methods described in the data analysis in our product insert are methods that can easily be done without the use of complicated statistics. Our Data Reduction method uses more complex statistical analysis to calculate target values. Both methods are acceptable methods to analyze the data.
One method of analysis in the package insert uses two consecutive levels to calculate the target values. The difference between the chosen consecutive levels is first calculated. Then, this difference is added or subtracted as needed to calculate the five target values based on the chosen levels.
The second method in the package insert uses Levels 1 and 5 to calculate the target values. The difference between Level 5 and Level 1 is divided by four to determine the delta for each level. Then, this difference is consecutively added to Level 1 three times to determine the target values for Levels 2 through 4.
Our Data Reduction method uses linear regression with specific data points as part of the target value calculation. This method is more complex than the one outlined in the product insert and most likely will not result in exactly same target values. For a detailed explanation of this process, please refer to the method explained page.
Downloads:
Free Data Reduction
Software Download
© 2001-2010 Maine Standards Company, LLC. VALIDATE is a registered trademark of Maine Standards Company, LLC.