February 1998

A small group of individuals with technical expertise from the US pharmaceutical industry and the Federation of American Scientists (FAS) has met periodically to discuss issues and measures related to strengthening the BWC with a compliance regime. The discussions have resulted in greater understanding on both sides, as well as some new ideas. It would be valuable for governments to promote and participate in give-and-take of this kind with industry and other non-governmental representatives in order to elicit the constructive involvement that is needed on all sides in developing an effective and workable protocol.

This paper reflects the private views of the participants in the meetings and not necessarily those of their organizations or any other group. It contains points of agreement reached at sessions held during the first half of 1997. Additional topics are currently under discussion.

Every effort should be made to minimize or eliminate requirements for confidential information in declarations. Organizations should be willing to give any information that is already in the public domain, in order to establish credibility and cooperation. This includes a considerable amount of relevant information of interest to a compliance regime. The industry participants are concerned, however, that if many types of information are requested, it will be impossible for them to be sure that their declarations are always accurate and complete.

Some companies are as big as small countries. Even with the best intent, it is likely that declarations will contain inaccuracies. This causes concern that a company may be accused of non-compliance even when it is making its best effort to provide the required information. Thus, US industry wants declaration requirements that minimize the possibility of error. Industry participants want to make sure that their declarations are 100% accurate and complete; FAS participants emphasize the need for correction mechanisms. Implementing legislation that provides for active interaction between declared facilities and government in preparing and approving declarations would help allay fears.

The simpler the declaration requirements, the smaller the risk of inaccuracies. Certainly, simplicity of execution is desirable in declarations. For facilities with minimum capabilities, providing only name, address, phone number, facility purpose and products (marketed products and those that are under development and publicly declared, eg, in an annual report), sources(s) of funding, and (specified) containment capabilities should be acceptable. Minimal declarations would at least indicate which facilities a State Party is willing to acknowledge, and would provide information on their ostensible purposes.

Computer-based declaration forms consisting mainly of checklists would help meet the simplicity criterion. Another way to decrease the probability of inaccurate declarations is to require submission of quantitative data only in broad ranges, such as those used in reports to the U.S. Environmental Protection Agency. This would make it easier to avoid errors. Ranges would also provide some protection for confidential information. For confidentiality purposes, expressing quantitative data simply as above or below a single threshold would be even better, and in many instances would likely be adequate for the purposes of the compliance regime.

Triggers for Declarations

Triggering criteria will be needed to specify the types of facilities that must file declarations under a BWC compliance regime. Criteria for facilities that must declare should be objective, clear, easy to demonstrate and not open to multiple interpretations. The triggers should be politically neutral and should not be based on commercially sensitive information. Capabilities that could raise concerns if concealed should be triggered for declaration. Types of facilities under consideration for mandatory declarations include those in the current voluntary Confidence Building Measures (which include human vaccine and BL4 facilities), as well as other types of dual-capability facilities (eg, antibiotic, animal vaccine, and other microbiological and biotech manufacturing facilities). If these should be required to declare, then we favor simple, general triggers rather than very specific ones, which could provide loopholes. General triggers will likely lead to a greater number of declared facilities that will include those with the greatest potential to produce BW agents.

The following facility declaration triggers are reasonable to consider: 1. Participation in a biodefense program.

2. Possession of a pathogen that has the potential for BW use (exemption: exclusively diagnostic and/or therapeutic facilities, provided that isolates of such agents are destroyed immediately following diagnosis and no live samples are retained). These agents may appear on an agreed-upon list or set of criteria.

3. Microbial production capability above a to-be-agreed-upon threshold (including fermentation, tissue culture, eggs, animals, plants). If this is the only trigger for a particular facility, the declaration should be very simple and short.

4. Transfers of to-be-agreed-upon agents and equipment. (These could be included in the declaration forms for facilities triggered by agents, production facilities, etc, if possible.)

The following are NOT desirable triggers for declarations:

· Vaccine/plant innoculant/biocontrol agent production. These would be caught under the microbial production trigger and do not require a separate trigger. It is unfair to single these out, when many other types of production could also be of concern.

· Genetic modification. This would be meaningless as a trigger because it is too widespread and there is no identifying equipment.

· BL4 or other containment. Such facilities would either be caught under other triggers (containment should be a question on the declaration forms), or would be irrelevant. However, there are no objections to BL4 as a trigger.

A central element of a BW compliance regime is the deterrent effect of short-notice challenge inspections. Short-notice inspections could also raise questions and highlight the need for future surveillance of a facility. The notice should be made as short as is practical. Although a modern pharmaceutical facility can be cleaned up in a few hours and most other facilities can likely be cleaned up in eight hours, the shorter the notice the greater the likelihood that a mistake in cleanup will be made. Evidence suggesting non-compliance could take many forms: for example, production records that were not destroyed, equipment and supplies that do not fit the declared use of the facility, or, in a very few instances, evidence of undeclared potential BW agents.

Procedures for Launching a Challenge Inspection

a) Weighing the Evidence: A Standing Committee of Technical Experts

A request by a State Party for a challenge inspection should be based on evidence suggesting non-compliance at the facility. The evidence should be submitted with the request to the Director-General. A mechanism is needed to discourage challenges based on flimsy evidence. It is not possible to specify in advance, however, the elements of evidence required or to assign weights to different kinds of evidence according to their seriousness, since every case will be different.

Evidence will generally be of two kinds: technical and political. Because of the dual-use nature of most biological capabilities, technical evidence of a BWC violation may not be clear-cut, so will require expert judgement. Since a decision should be reached within 24 hours, it would be important to have a standing Committee of Experts on call to provide technical help to the Director General in examining the evidence, before he gives his opinion to the Executive Council. The Committee could form a part of the decision-making process.

Alternatively, the Committee could be called on by, and report to, the Executive Council. There would be a certain advantage, however, in having the Committee report to the Director General before the Executive Council is notified of the inspection request, assuming that confidentiality could be maintained. Depending on the Committee's report, the Director General would then have the opportunity to suggest withdrawal of the request. Stopping the proceedings at this point rather than later would prevent any aspersions on the reputation of the facility in question.

The Committee could provide its input without delaying a challenge inspection. The clock should start when the request for an inspection is made, and the inspectors should prepare and head for their destination. They could be recalled, if necessary. The Committee's deliberations and the Executive Council's decision could be made while the inspection team is en route, with a deadline of 24 hours or less. It is desirable to have the inspectors on site as quickly as is physically possible, in order to optimize the potential for detecting illicit activity.

b) The Role of the Executive Council

Two different types of procedures are under discussion: a "red light" procedure, in which the States Parties on the Executive Council need respond to a request only if they wish to stop an inspection; and a "green light" procedure, in which the Executive Council must vote within a specified short time to approve an inspection.

Any attempt to reduce the speed and certainty of a challenge inspection will seriously degrade the effectiveness of the regime. The red light procedure (with a large vote required, as in the CWC) is fast and sure, and inaction due to apathy will not prevent an inspection from going forward. The green light procedure can also be fast, but apathy would tend to prevent rather than permit the inspection, thus making it less certain. In the green light procedure, Executive Council members who vote to approve take on a visible responsibility for their action. Consequently, since there may be political reluctance to vote in some cases, a green light procedure will tend to discourage challenge inspections. A red light procedure will tend to be more permissive, particularly if a large number of votes is required to deny a request.

A key concern with the red light procedure, if the number of Executive Council votes needed to prevent an inspection is high, is the possibility that an inspection will proceed even if its basis is unfounded or even frivolous. In terms of public relations and cost to a pharmaceutical or biotechnology company (eg, stock market reaction to the inspection), an unwarranted inspection could substantially damage the company. Challenge inspections should be used to investigate only well-founded suspicions and should avoid inadvertently punishing institutions that are in full compliance.

The green light procedure better answers the industry concern about unwarranted inspections, because more than one Party would have to demonstrate a belief that an inspection was needed. The question is, how large a vote should be required? Industry would consider a large vote to be safer, but it might be difficult to achieve a large vote to approve because of either apathy or irrelevant political considerations, thereby preventing a requesting Party from satisfying its legitimate concerns.

Perhaps a combination of red and green light procedures could satisfy the different concerns. In its simplest form, a mixed red-green procedure would require a vote by at least X-number of Executive Countil members to approve and fewer than Y-number of members to deny, in order for the inspection to proceed. The implications of the procedure depend on the values chosen for X and Y, and are not easy to predict. Nonetheless, it might be possible to select values that would satisfy all concerned.

Another possibility would be to set a quorum and let the majority of those voting to approve or to deny decide whether or not to proceed with a requested inspection. Since either "red" or "green" votes could be cast, this is also a mixed red-green procedure. The problem here is the default decision: if the quorum is not met, what should happen? The industry participants feel that all the Executive Council members should weigh the evidence and vote; if they are not willing to commit themselves and meet the quorum, the inspection should not go forward. On the other hand, defaulting to approve the inspection would preserve the rights of the requesting Party and would serve as an incentive for Parties opposing the request to cast votes; if they do not do so, the inspection should proceed. These two points of view have not yet been reconciled, but some form of mixed red-green procedure holds promise for reaching consensus.

Mandate and Access During the Conduct of Challenge Inspections

The CWC guideline on the mandate for an inspection (Verification Annex, Part X, paragraph 44) should be adopted: "In carrying out the challenge inspection in accordance with the inspection request, the inspection team shall only use those methods necessary to provide sufficient relevant facts to clarify the concern about possible non-compliance with the provisions of this Convention." Further, if the inspection team uncovers areas of suspicion outside its mandate during an inspection, the observations should be reported but not pursued without a new mandate.

The inspected Party has the right to refuse access to proprietary information. It may be possible in some cases, however, for industry to allow the inspection team to see the confidential information requested, but not to copy or remove it without permission. The CWC guideline (Verification Annex, Part X, paragraph 42) is appropriate: "If the inspected State Party provides less than full access to places, activities, or information, it shall be under the obligation to make every reasonable effort to provide alternative means to clarify the possible non-compliance concern that generated the challenge inspection."

Consultation and Reporting After a Challenge Inspection

It is important to have in place a procedure in which the inspected facility and its parent institution have ample opportunity to respond to the inspection team's findings. Before leaving the site, the inspection team should prepare a preliminary factual report, discuss it with the facility personnel on site, and allow them an opportunity to respond. The preliminary report could then be modified, or the response appended. The inspected facility should also have an opportunity to respond to the final report before it is issued by the Director General. These steps should be carried out confidentially.

The inspection team should not make a formal judgment on compliance or non-compliance; its report should be factual, but it may contain comments on the degree of cooperation, on whether or not the inspectors were able to satisfy their mandate, and on other factors affecting confidence.

Many of industry's concerns could be dealt with in national implementing legislation rather than in the protocol itself. Others should be covered in annexes to the protocol, so that revision would be possible after experience is gained. In developing and carrying out procedures for treaty implementation, the views of industry should be heard and accommodated by their governments to the extent possible, in order to protect industry from unnecessary cost and adverse publicity that could lead to financial damage. Implementing legislation should contain explicit procedures for accurate compilation of declarations, prompt notification to a target facility of an impending inspection, and assistance in preparing for and conducting inspections. The legislation should provide for active participation of industry in the preparation and approval of their declarations, and should make it clear that the government, as the responsible party, will consult the facility in preparing responses to inspection reports; the inspected facility and the government should confer on the final response. US implementing legislation should be prepared and agreed upon for consideration by Congress at the same time as ratification of the protocol.

All groups that will bear the burdens of the compliance regime should be consulted by the negotiators and their views considered. Industry believes that it will incur real costs, eg, the cost of training to prepare for inspections, whether or not the facility is ever inspected. Industry may also incur political costs, eg, a false accusation of non-compliance. A carefully constructed compliance regime should avoid this. The Inspectorate should not be permitted to make accusations, and the inspected facility and government should be given opportunity to counter any misapprehensions.

Nonetheless, challenge inspections could carry a high political cost to the inspected facility, since it would be assumed that the inspections had legitimate reasons behind them. If they were not based on legitimate reasons, they would carry a high political cost for the BWC as well. For this reason, it is expected that challenge inspections will be rare, and it is important that there be provisions and disincentives to prevent frivolous challenges.

As for non-challenge visits (which are now under discussion by the group), it would be interesting if a mechanism could be found to allow facilities to volunteer for them, without destroying their usefulness for validating declarations. This would remove any stigma that might be attached to an involuntary visit.

Further elaboration of these ideas and others, through collaborative efforts among government and private organizations, would likely contribute to the development of an effective compliance regime that does not impose undue burdens.