Home
« Previous | TOC | Next »

Sampling Plan for Hematocrit Quality Control

Introduction

In conjunction with the percent recovery of red cells in frozen blood, a system for providing quality control on the hematocrit is necessary to assure that the samples provided have sufficient hemoglobin. The method described below provides a mechanism to provide assurance that the hematocrit of blood stored frozen, thawed, and washed has a minimum acceptable level.

Statistical Method

The sampling plan is derived from ANSI/ASQC Z1.9 (formerly MIL-STD 414), Inspection by Variables for Percent Defective (1993). This document presents sampling plans to assure an acceptable quality level (AQL) from lots of product based on the results of a sample meeting a tabulated criterion. The size of the sample and the requirement to be met are functions of the size of the lot. For the purposes of this evaluation, 90 units of frozen blood stored up to the approved number of years will be considered a lot.

Data from the Navy Blood Laboratory based on 52 units stored frozen 7-17 years indicated that the percent hematocrit was not significantly reduced over this time period (regression P=0.12). Thus the available data indicate that the hematocrit over a similar time period can be considered from a single population identified solely by its mean and standard deviation. For the sample of 52, the mean hematocrit was 35.94 and the standard deviation was 4.58. The Shapiro-Wilk test of normality had a P-value of 0.96, indicating that a hypothesis of normality cannot be rejected.

The plan based on a 90 unit lot size under General Inspection Level II (Table A-2, ANSI/ASQC Z1.9) calls for the sampling plan under letter E. The procedure for a Single Specification Limit - Variability Unknown - Standard Deviation Method is to be used. This method is justified because the hematocrit data in the sample of 52 units discussed above was consistent with normality. Under this method, a sample is selected of a size stated in Table B-1 (ANSI/ASQC Z1.9), the mean and standard deviation are calculated and the quality limit QL is computed. The formula for the quality limit is provided below.

where L is the minimum hematocrit allowed from frozen blood. If QL is greater than or equal to the tabulated value for the size corresponding to the letter of the sampling plan, the lot has acceptable quality at the stated AQL.

For example, if L=28, the sample size for letter E in table B-1 is 7, and the mean is 35.5 with a standard deviation of 5.0, QL is 1.5. This value is compared to the tabulated value for the chosen acceptable quality level. If we choose 2.5% as the AQL, the tabulated value is 1.33. Since the computed quality limit exceeds the tabulated value, the hematocrit of the lot is acceptable at that AQL.

Thus for each 90 units of frozen stored blood, seven units need to have hematocrit done. The mean and standard deviation are computed from the sample, the quality limit is computed and compared to 1.33. If it exceeds 1.33, the lot is accepted as having acceptable hematocrit at an AQL of 2.5%.

By Richard P. Chiacchierini, Ph.D.
C.L. McIntosh and Associates

 

« Previous | TOC | Next »
Home