Flowchart ("Roadmap") for risk assessment: the necessary steps to conduct risk assessment according to Annex III of the Protocol

In a first discussion paper posted in this forum (‘Flowchart or ‘Roadmap’ for risk assessment and risk management under the Cartagena Protocol on Biosafety’) we have done a proposal for a ‘roadmap’ for risk assessment, that lists the steps of LMO risk assessment, and the points to consider that are relevant in each of these steps.
In a second contribution (‘‘Roadmap’ for risk assessment and risk management: information relevant to the points to consider’) I have gone into the information that may be relevant at each of the points to consider.

The roadmap is presented as a linear process of steps to be taken, and points to consider that belong to each step. The points to consider at the roadmap may also be seen as crossroad from which information that has been identified as relevant can flow into the risk assessment process. In this contribution I will show one example of how this would work at a crossroad on the roadmap:

Step 1: An identification of any novel genotypic and phenotypic characteristics associated with the LMO that may have adverse effects on biological diversity in the likely potential receiving environment, taking also into account risks to human health.
The first point to consider is:
(a)  The biological characteristics of the recipient organism relevant for its interaction with the likely receiving environment.

‘Biology documents’ for specific organisms are relevant here, as is discussed in the second contribution. Examples can be found in <http://www.oecd.org/document/51/0,3343,en_2649_34387_1889395_1_1_1_1,00.html>. By adding other available relevant information documents, we would have a comprehensive set of documents, encompassing all organisms that are currently used as receiving organisms for the construction of LMO.
This would be a first ‘peg’ on the crossroad, where we can ‘hang’ relevant information.

Clearly, filling the set of relevant biology documents will be quite some task that will probably take a lot of time and effort. Once information has been identified and added to the ‘peg’ on the crossroad, we should take care that this effort does not need to be repeated.

The following example shows how this may be accomplished. Information that is available on the Biosafety Information Resource Centre (BIRC) of the BCH could be analyzed for its usefulness as biology documents. When a document is found to be useful, a ‘flag’ could be added to its entry into the database, for instance as a keyword that indicates that it is useful at this crossroad, Step 1(a).
If all information that is relevant to step 1(a) is marked in this way, this would make it possible to retrieve, in one step, all information that is useful at this crossroad. If wanted, further keywords can be used to allow further filtering of the information, e.g. add the name of the species, so that different biology documents for one organism can be found in one step.
The same information may of course be useful at different crossroads; that would mean that it gets multiples ‘flags’; this is very well possible in practice.

In the contribution on information for the points to consider I have also indicated that another type of document would be useful at the crossroad Step 1(a): documents that discuss which specific information on the biology of an organism is relevant and has to be taken into account for LMO risk assessment. This is a type of general discussion document, whereas the first type of document described above is more factual.
It would probably be useful to discriminate between these to types of documents at this crossroad: the factual documents and the discussion documents.
This could be done by creating two ‘pegs’, one for each type, with two ‘flags’ or keywords for the different types of documents. The factual documents could for instance get the keyword ‘Step 1(a) biology documents’, the other documents could get ‘Step 1(a) discussion documents’.

In this way the information in a database like the BIRC could be very easily accessed for information that is useful for risk assessment. The main advantage would be that information that has already been found by one person and classified as useful for a certain step and crossroad on the roadmap, can be found by one push on a button by all other interested persons. In future the keywords referring to the crossroad where a document is useful can be added at the moment when a document is added to the BIRC. This would not require much extra effort. Only the addition of keywords to the already existing entries in the BIRC would be time consuming, but the result would be sufficiently rewarding.

Dear colleagues,

my name is Helmut Gaugitsch and I work for the Austrian Federal Environment Agency. First of all I would like to stress that I regard this discussion on the flowchart/roadmap concerning the necessary steps to conduct an LMO risk assessment according to Annex III of the Protocol as very useful and interesting. This is a topic which is of a very practical nature, it affects all the risk assessors dealing with LMO risk assessment on a daily basis worldwide. We can and should learn from each other a lot!

From my point of view it is important to note that on the one hand we have accumulated some knowledge on how to do risk assessment in practice, quite some amount of guidance is available at the different levels. At the same time we still struggle with the task on how to do it in practice in a scientifically valid way. How to use the different approaches and guidance documents available? There are many open questions on how to do that and there is much room for improving, standardizing andf harmonizing risk assessment.

A roadmap/flow chart may be very helpful in that respect and  the discussions in this online forum and the AHTEG to be convened are crucial to make progress towards this goal.

I would like to support the detailed and thoughtful comments and suggestions by Hans Bergmans on this topic, I can agree with most of the suggestions. I would just like to add the following additional considerations:

- performing an LMO risk assessment is not an easy task and therefore I fully understand and support the complexity of the  suggestions on how to integrate the points to consider (para 9 of Annex III) into the methodolog of the risk assessment (para 8 of Annex III).
- at the same time a roadmap/flowchart must be practicable and give guidance for a step-by-step procedure useful also for colleagues/institutions in the process of developing their capacity in LMO risk assessment.
- a practical instrument on how to do risk assessment in practice in my point of view has been the "matrix-approach" used by some colleagues worldwide. In this approach the 5 steps in risk assessment ("hazard identification", evaluation of likelihood, evaluation of consequences, evaluation of overall risk, risk management, summary of risk assessment including uncertainty) are performed for each introduced gene and each potential adverse effect identified in a systematic manner. Careful and precautionary analysis in a step-wise manner allows a risk assessment which dis not disregard any potential risks just because they seem to be unlikely at a first glance.
- an important element in LMO risk assessment is the completeness check of a notification. Is the notified dossier complete in order to be able to perform a risk assessment? A raodmap/flowchart should also give guidance on how to perform this basic check!
- an important addition to methodology and points to consider is also the consideration of the quality of the data provided as well as the suitability of the methods with which the date have been generated. Especially in this area I see a lot of need for improvement!
- Finally another challenge is the question: Do we need criteria which are adepted to the regionally different situations? In which cases and how should they look like?

Dear colleagues,
My name is Marja Ruohonen-Lehto and I work in the Finnish Environment Institute in Helsinki, Finland. Due to traveling and other obligations I am only now able to join the discussions. Again, I must state that the discussions are of high quality and very comprehensive. The description of the way forward by Dr Hans Bergmans is a good start for this important work on a flowchart/roadmap. Further, I very much appreciate the comments and additions made by Dr Helmut Gaugitch. He raises very pertinent issues in risk assessment that we still have to work on: we must emphasize and set scientifically sound criteria for the quality of data (see e.g. publications by Michelle Marvier and Peter Kareiva) and suitability of methods used and to be used for e.g. data collection must be further evaluated. Statistical power of the gathered/existing data is a very important issues.

Moreover, the 5 step approach of risk assessment that Dr Gaugitch refers to in his comment is very much in line with the Protocol approach to risk assessment and risk management

We certainly have a lot of experience on some aspects and applications of biotechnology but there is clear lack of data and information on certain, more recent applications. This must be kept in mind and lack of data must be approached in a scientifically sound way.