Development of a safety risk matrix
for laboratory work with Avian
influenza
Jill Banks
How is work with Avian influenza (AI) viruses in the laboratory
setting managed to ensure occupational safety?
• At the laboratory level we have developed a risk matrix - better metrics for
assessing the risk to staff working with AI viruses.
– This couples a virus risk category (based on knowledge of human infections,
literature, known mutations of interest) with a Boston square matrix for the activity
being carried out e.g. egg inoculation.
•
APHA has an occupational health protocol for staff that work in the
laboratory with AI or who may come in to contact with infected animals.
• This includes prophylaxis with Tamiflu if required.
• Analysis of the any outbreak virus includes full genome sequencing followed
by phylogenetic analysis and screening for known humanising mutations
(e.g.as defined in CDC mutation inventory)
Overall probability:
Likelihood of MSC class I
breach occurring
Almost certain,
might well be
expected
> 100
Quiet possible
> 10-1
Unusual but
possible
> 10-2
Only remotely
possible
> 10-3
Conceivable, but
highly unlikely
> 10-4
Practically
impossible
> 10-5
impossible unless
aided
> 10-6
(virtually)
impossible
< 10-6
Infection risk/impact:
Virus group
A1
A2
B/C
Determining the risk
Classify viruses that we work with into groups of different zoonotic potential :
Virus classification document.
Some of the factors used to determine risk included the:
• pathogenicity of the agent,
• agent’s route of transmission,
• stability of the agent,
• infectious dose or concentration of the agent,
• origin of the agent,
• availability of effective prophylaxis against the agent availability of the
use of antiviral agents for treatment of human infections with
influenza A.
Premise is that all infection is considered unacceptable. The
consequence/impact of infection is captured through the virus categorisation
and this feeds into the matrix as high, medium or low infection risk/impact.
Text in footer
3
Steps in the process
• Consider the virus
• We produced a document that helps to categorise viruses for potential
zoonotic risk.
• Consider the procedure/technique to be performed
• In particular the likelihood of MSC class I breach occurring
• This relates the to the type of process being conducted
• Takes into account the probability of airflow being disrupted or material being
expelled through the front opening of the class I MSC.
• Exposure risk in event of breach.
• This considers the titre of the virus in the material that could potentially be
expelled from the MSC and the form that it is in e.g. the aerosol producing
potential of tissue v.s liquid, volume etc.
Text in footer
4
Virus categories
A1
A2
B
C
Virus categories
A1
A2
B
C
• A1. Those that are considered to be the highest risk and
most work with any live virus will be done in class III
microbiological safety cabinets (MSC).
• A2. Those that are classed ‘high risk’, but less so than A1
While some procedures must be carried out in a class III
MSC others can be carried out in a class I MSC.
Category A1
• 1. Viruses classified as sub-category A1:
• 1.1
Novel[1] human isolates (or isolates derived from them) that have been
associated with severe disease in humans (i.e influenza-like-illness other than
conjunctivitis), human-to-human transmission, or lethality in people that has been
confirmed in laboratory tests. Typically this would not include Type I and Type II
submissions/samples (see end of doc).
• 1.2
Isolates that have been laboratory manipulated to alter their biological
properties in a way that may result in increasing transmissibility in humans or
increased ability to replicate in mammalian species as determined in vivo, in vitro or ex
vivo.
• 1.3
Any H5 virus derived from the progenitor strain A/goose/Guandong/1/96 [i.e.
any virus shown to be or thought to be genetically linked to the H5N1 viruses that
arose in SE Asia in 1996/97 and subsequently spread across Asia and into Europe and
Africa].
• 1.4
Any GM04 virus for which there is no current derogation in place.
• 1.5
All submissions from external laboratories unless there is definitive evidence
that the submitting country is free of Category A1 viruses
• 1.6
Uncharacterised[2] avian influenza viruses for which there is concern that they
may represent a risk over and above normal uncharacterised viruses.
• 1.7
A1.
Any virus that the Virus Categorisation Group consider should be classified as
Use of the Boston Square Matrix – Steps in the process
Boston square matrix considered as potential model for Risk
Assessment (RA). System can take account of different elements that
interact and become a potential hazard.
Overall probability:
Likelihood of MSC class I
breach occurring
Almost certain,
might well be
expected
> 100
Quiet possible
> 10-1
Unusual but
possible
> 10-2
Only remotely
possible
> 10-3
Conceivable, but
highly unlikely
> 10-4
Practically
impossible
> 10-5
impossible unless
aided
> 10-6
(virtually)
impossible
< 10-6
Infection risk/impact:
Virus group
A1
A2
B/C
• All members of the APHA avian influenza virus team were asked to
provide a list of the techniques/procedures that they use in their work.
• Team Leaders then considered the list of procedures and sample
types to assign probabilities for a cabinet breach and the exposure
risk
Text in footer
8
The basis of the assessment is a probability rating that is assigned to each hazard.
Please use the table below to guide your assessments.
Probability of hazard to occur in a 5 year time period:
Description of the probability
Indicative statistical value of the probability
Almost certain, might well be expected
> 50%
100
Quite possible
> 1/10
10-1
Unusual but possible
> 1/100
10-2
Only remotely possible
> 1/1000
10-3
Conceivable, but highly unlikely
> 1/10,000
10-4
Practically impossible
> 1/100,000
10-5
impossible unless aided
> 1/1000,000
10-6
(virtually) impossible
< 1/1000,000
10-7
Examples: Probability of infection to occur via aerosol i.e. Breach of cabinet integrity
whilst working in class 1 in a 5 year time period:
Probability of occurring
scenarios
1
Swab flicks out – aerosol
10 -3
2
Tissue flicks out – aerosol
10 -3
3
Liquid spill outside cabinet - aerosol
10 -2
4
Aerosol – homogeniser
10 0
5
Aerosol - HA plate
10 -2
6
Aerosol – VNT plate
10 -3
7
Risk of becoming infected through exposure to low
titre (Not amplified) virus
10 -3
8
Risk of becoming infected by respiratory route by
exposure to Amplified virus
10 -1
BOSTON SQUARE MATRIX
Overall probability:
Likelihood of MSC class I
breach occurring
Almost certain, might
well be expected
> 100
Quiet possible
> 10-1
Unusual but possible
> 10-2
Only remotely
possible
> 10-3
Conceivable, but
highly unlikely
> 10-4
Practically impossible
> 10-5
impossible unless
aided
> 10-6
(virtually) impossible
< 10-6
Infection risk/impact:
Virus group
A1
A2
B/C
Red = work in class III
Green = work in class I
Orange = use additional parameters to assess risk meanwhile work in class III
Risk Matrix probability scoresheet
1
Procedure1
2
3
4
5
6
Probability
of MSC
class I
breach
Virus titre
Exposure
risk2
Overall
probability/
risk
= column 2
score x
column 4
score
Class I/III
cabinet
operational
mode
(Determine this
from the Boston
square matrix)
A1
A2
B
RNA extraction
Swabs
10-2
Diagnostic
non
amplified
10-4
10-6
I
I
I
Egg work
Inoculation
10-2
Non
amplified
10-3
10-5
III
I
I
Tissue
Homogenisation
Machine
100
Very high
10-1
10-1
III
III
III
BOSTON SQUARE MATRIX
Overall probability:
Likelihood of MSC class I
breach occurring
Almost certain, might
well be expected
> 100
Quiet possible
> 10-1
Unusual but possible
> 10-2
Only remotely
possible
> 10-3
Conceivable, but
highly unlikely
> 10-4
Practically impossible
> 10-5
impossible unless
aided
> 10-6
(virtually) impossible
< 10-6
Infection risk/impact:
Virus group
A1
A2
B/C
Red = work in class III
Green = work in class I
Orange = use additional parameters to assess risk meanwhile work in class III
Thank you for your attention