That has been solved by having, as I mentioned, rigid time/temperature regulations in place and these products are cooked now so that problem is solved.

Senator PRYOR. Dr. Houston, excuse me. I may have not phrased that question well. In the plant the inspectors inspect for salmonella in beef one way, I understand, and for poultry in another way. For example, how is a chicken inspected. I thought they put them in a bag of water or a bag of chemicals or something like that and then sample the fluid. Would you describe that process?

Dr. HOUSTON. In addition to the visual examination?
Senator PRYOR. Yes.

Dr. HOUSTON. Well, many poultry companies do have programs in place where they pretest birds, for example, to be sure that no chemical residues are in them. They pretest feeds, and we have cooperative agreements with those companies that share information with us. So that if they find anything we can take action ahead of time. The poultry industry is far ahead in that regard, much further ahead than the red meat industry.

Senator PRYOR. The poultry industry is ahead in this regard?

Dr. HOUSTON. Yes. The poultry industry has done a very good job in eliminating residues. By virtue of some problems they had in the 70's, they did put a number of controls into place, and this is spoken to in the Academy report as well.

Senator PRYOR. I wonder if Dr. Crawford is going to be able to stay if you have to leave. Will he have to be involved in the same ceremony you are in?

Dr. HOUSTON. No. He can certainly stay.

Senator PRYOR. You've just been appointed to stay. [Laughter.] Not to stay as a witness, but to stay if we may have to ask a technical question after a while.

Senator LEVIN. And we will keep the record open also for questions of Dr. Houston.

Senator COHEN. Save a piece of the cake for him.

Dr. HOUSTON. I will.

Senator PRYOR. Well, Mr. Chairman, I have a lot of questions, but I can sense that you're ready to move on. [Laughter.]

All I'm trying to do is just get the truth out. I'm just searching for the truth.

Senator LEVIN. You're doing a good job.

Dr. Crawford will be here in case there is a technical question which comes up later and we always can ask him.

So thank you very much, Dr. Houston.

Dr. HOUSTON. Thank you.

Senator LEVIN. Our second witness today is Brian Usilaner, the Associate Director of the General Government Division of the General Accounting Office [GAO].

At the request of the Subcommittee this division of the GAO has been reviewing the productivity of the FSIS inspection program. Mr. Usilaner, we welcome you. Am I pronouncing your name correctly?

Mr. USILANER. Yes, sir.

Senator LEVIN. Perhaps you could introduce your associates.

TESTIMONY OF BRIAN L. USILANER, ASSOCIATE DIRECTOR, NATIONAL PRODUCTIVITY GROUP, GENERAL GOVERNMENT DIVISION, U.S. GENERAL ACCOUNTING OFFICE, ACCOMPANIED BY CLIFTON E. SPRUILL AND CONNIE SAWYER, EVALUATORS, NORFOLK REGIONAL OFFICE, GAO

Mr. USILANER. Yes, I will.

As you said, my name is Brian Usilaner with the National Productivity Group of the General Government Division. And with me at the witness table is Clifton Spruill and Connie Sawyer. They are with our Norfolk Regional Office and they head up our field work on this particular assignment.

With your permission, I would like to read a summary statement that sort of summarizes our work to date.

Senator LEVIN. Your entire statement will be made part of the record, and we do appreciate your summarizing it.1

Mr. USILANER. Thank you.

We are pleased to appear today at your request to discuss preliminary results of our review requested by Senator Cohen on productivity and quality management in inspection functions of the Federal Government.

As requested, we began our review in the Department of Agriculture's Food, Safety and Inspection Service. The review is being conducted for the Subcommittee in order to assist in its oversight of management of inspection activities across the Federal Government. Overall over 40 Federal agencies perform inspections and about 76,000 Federal employees are involved.

As requested, the Department of Agriculture was selected as the starting point for our examinations. FSIS had the greatest number of inspectors within Agriculture. This first phase of our work has primarily involved poultry inspections.

For the purposes of our review we considered quality management to mean assessing first how well inspectors are performing their work and, second, how well the facility's products and work force at each plant are conforming to requirements of law and regulations

Our review is in the early stages and our observations are at this point preliminary. We believe, however, that FSIS may need to improve its data on measures of plant quality and inspector perform

ance.

With regard to plant quality, the indications are that the agency does not have objective, systematic measures of quality on a plantby-plant basis that would permit an objective assessment of quality among various plants or an assessment of plant's quality over time. Regarding quality of inspector performance, preliminary work has shown that the agency does not have the objective information to assure the quality of an inspector's performance or to assess how well inspectors are following procedures.

The agency is experimenting with various changes to its inspection procedures. We are concerned that the apparent absence of objective data could impair the agency's ability to assess the benefits of these changes.

1 See p. 102.

I would now like to discuss changing inspection policies and procedures, our preliminary observations and why we believe objective quality data on plants and inspectors are important.

Concerning inspection policies and procedures, the law requires inspection of each poultry carcass. These inspections are intended to ensure that the Nation's supply of poultry is safe for human consumption.

To achieve that end, as we have already talked about, FSIS inspectors visually inspect each bird slaughtered for consumption. Also, the inspectors inspect the slaughter and processing facilities. The agency is continuing to explore new ways to satisfy itself about product safety.

In 1983 the agency asked the National Academy of Sciences to assess its inspection program and recommend improvements. In an 1985 report the Academy recommended an optimal inspection system that would include random sampling and rapid, inexpensive screening tests to detect chemical compounds and biological agents. The Academy also recommended that new approaches not be implemented nationally until validated by objective assessments of their impact. A follow-up report just issued provides a model a for risk assessment approach to poultry inspection.

The primary reasons for considering changes in inspection procedures are, one, increasing concern that the current look-at-everybird approach is not adequate and, two, a shrinking work force that must inspect a growing number of slaughtered birds.

In addition to considering changes in slaughter inspection, the agency is also experimenting with changes in the inspection or processing operations in which the plant becomes more and more involved in quality assurance and the inspector monitors the plant's quality control system.

The desire to move from detection of problems to an approach aimed at preventing problems is quite in line with quality assurance techniques now being applied in the private sector which involve sampling of critical stages of production, statistical quality control techniques and strong participation by employees.

However, it is important to point out that such an approach requires adequate data on slaughter and processing quality and appropriate analysis of such data. Any conclusions drawn from the data are only as good as the correctness, availability and accessibility of the data.

In these new inspection procedures it is particularly important for the agency to have a means to measure the quality of the plants being inspected at key points and the quality of the work performed by inspectors.

I would now like to discuss this issue of measurement in more detail. Let me tell you what we know about the collection and use of data on plant quality levels.

First, a 1981 GAO report identified weaknesses in the collection and analysis of management data. We reported that inspection program supervisors at times did not adequately document the results of their monthly reviews.

Next, the 1985 Academy report said that there was no comprehensive statement of criteria, no systematic accumulation of data and no technical analysis of the hazards or benefits to human

health in the traditional inspection program or as a consequence of new inspection techniques.

Also, a September 1986 Agriculture IG report that Senator Cohen mentioned indicated the agency did not have adequate controls over data collection and analysis. We don't know the extent that these problems still persist since our work is in the preliminary stages.

Officials told us that the agency has requirements for the collection of various data that relate to quality. However, the desired amount of documented information is not consistently collected or analyzed.

An FSIS Future Agenda report responding to the 1985 Academy recommendations states, and I quote "In the past large amounts of data were collected for statistical purposes to report to Congress on the scope of operations. Data were not used for analytical purposes, for evaluating program effectiveness or for improving the day-today quality of inspection in the plant."

Concerning measures of quality on a plant-by-plant basis, we were told FSIS does have a great deal of data which are entered into a central computer system. Reports generated are more focused on output quantities and condemnations than on the quality of the facilities and work force.

Regarding the quality of an inspector's work, preliminary indications are that the agency does not have the information needed to objectively make assessments. We met with the Administrator, Dr. Houston, who acknowledged that his agency does not objectively track the quality of the work performed by individual inspectors. It is presently using subjective ratings to evaluate inspector performance. In fact, at one plant we visited the inspector in charge told us that the most difficult phase of his job was to rate inspector performance.

Turning to the productivity issue, the President in February of 1986 established a productivity improvement program for the Federal Government. The implementing instruction required a description of how the agency will measure quality, timeliness and efficiency and also required that baseline data be established for each. As one of the improvement initiatives Agriculture submitted to OMB recent productivity changes in meat and poultry inspection activities. We noted, however, that the submission did not include measures of quality or plans to develop such measures. Without quality measures you cannot determine whether productivity is being improved at the expense of quality.

In closing it appears that there is a critical need for objective performance measures to evaluate the effectiveness of the inspection work force and for objective systematic measures of quality on a plant-by- plant basis in order to assess the relative benefits of changes to the present inspection procedures.

Based on these observations, we plan to pursue the following matters during our continuing review:

Can FSIS objectively determine whether inspector quality is increasing or decreasing?

Is management monitoring and emphasizing inspector quality? What factors affect inspector quality performance?

How can inspector quality be improved?

How can plant quality reporting be improved?

And, finally, has the agency properly assessed the impact of productivity increases on quality?

We also plan to determine the impact of quality assurance techniques and the impact of changing the focus from detection to prevention.

We believe our work will be helpful in enabling us to identify issues for further examination when we begin our reviews of meat inspection and inspection activities in other Federal agencies.

This concludes my statement, Mr. Chairman, and we would be pleased to respond to any questions of the Subcommittee.

Senator LEVIN. Thank you, Mr. Usilaner.

You indicated that the FSIS does not have a systematic evaluation of plants on a plant-by-plant basis. Now does that mean, in your opinion, that the FSIS would not be able to tell us the compliance status of any particular company during last year say?

Mr. USILANER. We have to this date not been able to document any trend type data that's being collected and being transferred down to the operating levels. There is data, as I said, on condemnations and there is some other data such as ratings and so on, but our concern is for objective analysis on a continuing basis.

Senator LEVIN. Now won't the FSIS have to have that information in order to permit some plants to do their own inspections under a discretionary inspection program?

Mr. USILANER. Yes.

Senator LEVIN. So the information that they don't have they'll need?

Mr. USILANER. Yes. There are two parts. There is the slaughter operation and then the processing operation. Where they have this discretionary inspection in the processing operations, they have more documentation but in the slaughter operations they don't have discretionary inspection and, they do not have adequate data, or at least we haven't been able to identify it at this point.

Our concern, as was brought out in the Academy report, was that you're going to be getting a lot of questions as you move away from bird-by-bird inspection, and to assure the public advocate groups, and inspectors that there is no detriment to quality, you're going to need this objective data analysis.

Senator LEVIN. Now as I understand it, there is a discretionary test program going on in some State where 14 plants are participating.

Mr. SAWYER. Yes, that's going on in the State of Tennessee.

Senator LEVIN. Yes. Now wouldn't the information which is not available be necessary in order to permit that program to proceed in a proper way?

Mr. SAWYER. Sir, what we've been told is that this discretionary inspection will apply to the processing plant and it will be at the plants in which the agency has a concept called total quality control, and under this concept they have more documented information than they would in a slaughter operation.

Senator LEVIN. Have you looked at that documentation to see if it's adequate for this purpose?

Mr. SAWYER. Not in detail.

Senator LEVIN. You might want to do that.