VOLUNTARY SPLIT-SAMPLE PROGRAM FOR ENVIRONMENTAL LABORATORIES IMPROVES BACTERIOLOGICAL DATA

By Barney Asato, Program Coordinator, and Anita Fernandez, Laboratory Committee Co-Chair, HWEA

INTRODUCTION
The inter-laboratory performance evaluation program was established in 1998 to improve analytical proficiency at environmental testing labs in Hawaii. Labs participating in the program analyze sub-sample replicates of a single wastewater grab. By comparing results from many labs, individual labs can determine how they are performing in reference to their peer group. The program is administered twice per year, free-of-charge, under the direction of laboratory professionals of the Hawaii Water Environment Association.

Labs benefit by participating because their analytical proficiency is documented. This information can be used to support marketing of laboratory services to prospective clients. Also, involvement in these programs is recommended by the Environmental Protection Agency (1) and is required to meet certification under the National Environmental Laboratory Accreditation Conference. (2) 

The public benefits because the quality of laboratory data is improved. The quality of this data is important because it is routinely used to make decisions affecting public health and the environment; examples include beach closures, monitoring of sewage discharges, and water pollution remediation activities. 

Until now, variability of bacteriological data between laboratories has not been properly addressed. Past inter-laboratory studies have been sporadic or used methods designed to detect rather than quantify target microorganisms. These qualitative methods indicate presence or absence but fail to provide sufficient data to uncover any analytical differences, if present, between laboratories.

The membrane filter method used in this program is a quantitative procedure. It provides a direct count of bacteria in water based on the development of colonies on the surface of a membrane filter when placed on a selective nutrient medium. The density of organisms is reported as colony forming units per one hundred milliliters (CFU per 100 mL) of a water sample. Four membrane filter procedures for bacterial indicators of water pollution commonly used in Hawaii are approved for use in this program; they are two methods for enterococci (3), and a single method each for fecal coliform (4) and Clostridium perfringens. (5)

STUDY DESIGN
Nine split-sample studies have been conducted since 1998 involving a total of 12 private and government laboratories. For each study, participating labs analyzed sub-sample replicates of a single wastewater grab. Laboratory results, sample transport conditions, and analytical quality control were recorded on standardized worksheets and submitted to the program coordinator. Data from all labs were compiled using numeric identification codes to maintain confidentiality. The reports sent to participants included information on group precision, bias, and suggestions for improvement. 

RESULTS AND DISCUSSION
The first performance evaluation (PE) study was conducted in the summer of 1998. Fecal coliform counts ranged from >160,000 to 21,000,000 colony forming units per 100 milliliters (CFU per 100 mL) and Clostridium perfringens ranged from 1,500 to 42,000 CFU per 100 mL. Enterococci counts were somewhat closer ranging from 250,000 to 460,000 CFU per 100 mL. Labs were asked to review their analytical methodology and internal quality control practices. A few sought advice and training from experienced microbiologists at other labs in the program. 

Precision improved for five of the six labs participating in the second PE study. Fecal coliform ranged from 12,000,000 to 14,000,000 CFU per 100 mL. Clostridium perfringens for five labs ranged from 24,000 to 39,000, but Lab 3 reported 8,000 CFU per 100 mL. Enterococci for five labs ranged from 280,000 to 550,000 CFU per 100 mL, but Lab 3 was again low at 33,000 CFU per 100 mL. Labs reported sample holding times ranging from 1 to 7 hours after collection and transport temperatures from 2 to 14°C. Agreement between laboratory data continued to improve with subsequent PE studies. Refer to Results of Microbiology PE Studies 1 to 9.

Group precision
Precision of laboratory data was tracked using the relative standard deviation (RSD) statistic. Also called the coefficient of variation (COV), RSD of a data set was calculated by dividing the sample standard deviation by its mean and the result multiplied by 100. The RSD values decreased as precision between labs improved. For each study, RSDs were calculated separately for enterococcus, fecal coliform and Clostridium perfringens. 

Table of RSD Values for Inter-Laboratory Split-Sample Program 

Note: A number of labs began reporting results separately for Enterococcus Method 1600 starting from Performance Evaluation Study 7.

The RSD values were plotted against the corresponding PE study (Graph 1. Precision Between Laboratories, Bacterial Indicators on Split Samples). The graph clearly shows that fecal coliform precision quickly stabilized below 40 RSD after PE Study 1. However, enterococci and C. perfringens did not stabilize until PE Study 4 to 5. Lab precision for enterococcus rose to 60 RSD for PE Study 9, but this may be due to poor sample preservation as four of nine labs reported transport temperatures at or above 10.0°C. PE Study 9 samples were not cooled prior to distribution as in previous studies in which samples were held on ice for about 30 minutes after collection.

Go to page 2 >>

(1) U.S. Environmental Protection Agency, Microbiological Methods for Monitoring the Environment, EPA 600/8-78-017, Dec 1978, p.244-5.
(2) National Environmental Laboratory Accreditation Conference, Proficiency Testing Program, Revision 12, July 1, 1999.
(3) U.S. Environmental Protection Agency, Improved Enumeration Methods for the Recreational Water Quality Indicators: Enterococci and Escherichia coli, EPA/821/R-97/004, March 2000, Membrane Filter Method 1600 for Enterococci, and [Original] Method 1106.1 for Enterococci.
(4) APHA, AWWA, WEF, Standard Methods for the Examination of Water and Wastewater, 18th Edition, 1992, Membrane Filter Method for Fecal Coliform.
(5) Bisson, J.W., Cabelli, V.J., Membrane Filter Enumeration Method for Clostridium perfringens, Applied and Environmental Microbiology, Jan 1978; Armon, R., Payment, P., "A Modified mCP Medium for Enumerating C. perfringens from Water Samples," Canadian Journal of Microbiology, 1988.