Comparing APTT reagents
There is an obvious mistake in Fig. 2 (Cumulative summation of differences method for comparing APTT reagents) and the interpretation of the CAP requirements for comparing APTT lots (Q&A, February 2012). One cannot compare two lots over a nine-year period using the statistics implied in this figure. You can compare two lots within a year or less using 10–30 individual paired-comparison type statistics, not means on a yearly basis. The paired comparison analysis does compare the mean differences between the individual specimens and gives a significance related to the difference. I cannot determine what clinical significance can be derived from the analysis shown in Fig 2. of this report. If there is one, please let me know.
Leon Zuckerman, PhD
Des Plaines, Ill.
• Russell A. Higgins, MD, of the University of Texas Health Science Center at San Antonio and vice chair of the CAP Coag-ulation Resource Committee, and Sandra C. Hollensead, MD, of the University of Louis-ville (Ky.) Hospital and a member of the CAP committee, reply:
Dr. Zuckerman has a valid concern about the statistical model. The cumsum model has not been evaluated formally.
To explain the method a bit further: Comparison data (APTTs on heparinized patients) are plotted with the old APTT reagent on the x-axis and the new reagent on the y-axis. Visual or regression analysis is performed to compare data and identify discrepant and outlier results. The data for each APTT reagent are summed and the mean and standard deviation determined. The difference between the means of the new and old APTT reagents are then recorded.
In the following year (or at the next APTT reagent change), the comparison data are again determined, and the difference of the means is added to the difference between the old and new reagents from the year before. In this way, drift in sensitivity of the reagent caused by multiple changes over time can be detected and controlled. If the cumulative mean is greater than seven seconds in any year of comparison data, the reagent is rejected, and another reagent performing more closely with the laboratory’s past reagents may be chosen.
Although not a formal component of the method at this time, an SD comparison year to year could be a valuable piece of information in identifying reagents with high variability.
Despite great effort on the part of laboratories, physicians may be reluctant to change practices.
The Cleveland Clinic Measures Consortium was established in 2010 with the goal of developing evidence-based quality and performance measures associated with the pre- and postanalytic stages of laboratory testing. In cooperation with the Centers for Disease Control and Prevention and a panel of national experts in laboratory testing, the consortium has identified several areas as high priorities.
We invite comments from the professional community about the relative importance of the following proposed parameters, as well as opinions regarding which quality indicators should receive priority for study. We invite recommendations for parameters not listed here. The parameters that are ultimately given the highest priority will be studied to establish essential quality and performance measures. Please address correspondence to the e-mail address at the end of this letter. The public comment period ends May 15.
Here are the high-priority areas:
Incorrect patient identification. Assessment of the baseline frequency of specimens with incorrect patient identification and the impact of incorrect patient identification (re-acquisitions, continued therapy without the benefit of the laboratory test, for example).
Unnecessary same-day duplicate orders. The frequency of unnecessary same-day duplicate orders and the financial and clinical impact of stopping such orders.
Improper handling of abnormal test results. How and when abnormal test results were acknowledged in the medical record, to determine if they received followup, to assess the impact of the failure to address these results, and to determine if missed laboratory results were likely due to interface issues.
Appropriate use of coagulation testing/hypercoagulation panel. This includes:
- The percentage of tests drawn within the appropriate time interval and the impact.
- The percentage of patients without a history of a bleeding disorder and who had routine coagulation studies performed before surgery.
- The percentage of routine coagulation studies that were performed before surgery and resulted in surgical delay.
- A determination of the inappropriate overuse of the hypercoagulation panel.
Appropriate use of HbA1c. This includes:
- The frequency of duplicate HbA1c levels drawn before the recommended re-testing date.
- Followups of patients with abnormal HbA1c results.
- An assessment of the frequency of significant differences in test results, when a repeat test was ordered within the 90-day window.
Appropriate use of constitutional molecular genetic tests. This includes:
- The frequency of repeat testing and associated rationale for the repeat request.
- The number of instances in which the clinical findings did not support the ordering of a molecular diagnostic test that cost more than $1,000 (that is, content experts deemed them inappropriate).
- How frequently requested supplemental information, which can be used to assess the appropriateness of the test order, was supplied.
- The frequency with which a consultation with a medical geneticist or genetics counselor took place after a constitutional genetic abnormality was detected by a molecular method.
Reflex molecular diagnostic assays in microbiology. An assessment of the appropriateness and timeliness of response to the results of a rapid molecular diagnostic assay, as well as the impact of the response, such as change in therapy, decreased length of stay, and diminished morbidity and/or mortality.
Testing for C. difficile by PCR. This includes:
- The percentage of cases in which a repeat C. difficile PCR was ordered within the same clinical episode, and the impact of this additional testing (that is, change in therapy or outcome).
- The frequency of “C. difficile × 3” orders.
- The frequency of repeat C. difficile testing within two days of a negative PCR and within 10 days of a positive result, and the impact of this repeat testing.
- The frequency of C. difficile orders on formed stool.
Lipid testing in association with stroke. This includes:
- The percentage of patients with stroke for whom lipid profiles had been ordered within three months.
- The frequency of followup testing on patients with an abnormal lipid profile in this population.
- If lipid abnormalities were discovered, the percentage of patients who received a lipid-lowering agent or another therapeutic intervention.
- Of the patients with abnormal lipid profiles, the percentage of those rechecked within three to six months and later.
Gary W. Procop, MD, MS
Project Director, Cleveland
Clinic Measures Consortium
Chair, Department of Molecular Pathology
Send letters to Editor, CAP TODAY, 325 Waukegan Road, Northfield, IL 60093. Fax: 847-832-8873; firstname.lastname@example.org.