Agenda Item 13 ACMR17/75 Inf. doc. 4

Appendix 3
Informational submission by J. Lederberg in connection with

Engineering Models of Biohazard of Microbiological Procedures

Although many elements of risk assessment are difficult to quantitate
at the present time it is useful to outline models that help to identify
critical factors. We may view contamination as a process of hindered
diffusion. The eventual biohazard cost is, crudely C= S x T x R, where
S is the level of environmental loading in the laboratory, T is the probability
per organism of release outside the laboratory and R is the social cost per
organism released. To a first approximation these are linear functions, i.e.
the fate of each microbe is independent of others.

Each of these factors may be subdivided, and is subject to some
manipulation in the interests of minimizing C. Thus R may be drastically
reduced by biological variation: the choice of organisms which are discovered
or engineered to be relatively incapable of doing harm even if released. In
the context of molecular genetic research, these are the "disabled carriers",

Good laboratory practice is dedicated primarily to minimizing Se, |
the microbial load in the laboratory but outside the specific containers which
house the organisms. The level Se depends on the total bioburden and on the
procedures by which this is handled. 101° organisms in a quiet flask offer
less risk than 10° organisms subject to inadvertent aerosolization.

The transmission factor, T, may be divided into a physical and a
human component (Tp and Th respectively). The training of personnel is
obviously of the first importance for a) reducing the interval loads, Se,

b) for maintaining the parietal integrity (Tp) of the physical facilities,
and c) for their personal hygiene, Th. Other measures like prophylactic
immunization and clinical surveillance for early spotting of possible infection

are also appropriate for given organisms.
ACMR17/75 Inf. doc. 4

Appendix 3
page 2

We have relatively little quantitative, absolute information on these
parametersbut some guesses are possible about relative risk - attenuation
ratios, expressed in decibels (e.g. 25 db = 107°"), The numbers are rough
guides only.

Disabled organisms in the most favourable cases can afford 100-120 db
improvements, by theoretical calculation. This may be tempered by non~ideal
behaviour and other uncertainties so that a conservative 60 db is assigned to
this recourse in the figure.

We have little explicit information on the security factors of
moderate-risk oriented facilities. It is possible that substantial economy
and assurance is available through more informed choices in this area.
Aerosol reduction and internal hygiene in the laboratory have already been
discounted. Given the realities of human compliance it may be optimistic
to expect more than another 40 db from moderate-cost facilities, and an
additional 40-60 db from ultra-secure operations (involving, e.g. tele operator
handling of cultures).

Note that 160-180 db may be achievable either with the use of safe
carriers or with ultra-secure facilities.

This attemmation, 1072© is indeed an astronomical safety factor. It
is tantamount to reducing releases from, say, 10° organisms per day to a
probability of 10°*°, Another dollar's investment (which includes the fore-
going of a like benefit) would be Justified only in the remarkable case that
the expected cost (not the worst case contingency) of releasing an organism
were of the order 10? dollars. (E.G. that the hazard of one such release ex-
ceeded the annual benefits of the world's public health programmes. )

This criterion exceeds (but not by a large margin) the expected

hazard from the release of a single particle of any known infectious agent.
ACMR17/75 Inf. doc 4

Appendix 3
page 3

Thus it does leave some rationale for the interdiction of a range of high-
risk low-benefit experiments for which still higher margins are for some
reason unavoidable.

If the human population is assumed to be so vulnerable to a one-
particle risk it should give occasion to reexamine the cost-effectiveness
imperatives of other measures to reduce our vulnerability to natural sources
of infection.

Careful practice and choice of equipment can easily reduce the Se/Si
ratio by 40 db, and elementary hygiene can ameliorate T (Tp + Th) by at
least another 20 db. This 10° fold reduction in risk is achievable at low
cost by the application of the most elementary principles of laboratory
hygiene (Appendix I). This is the most cost-effective part of the safety
function.

200

180

160 —

140 a“

120 < .
: special ultra-

physical

ne oe secure
Bo carriers

ae facilities facilities
60 1 elementary | |

Ho / | bveiene

|
}

20| _/

safety ratios

 

 

cost ————_--__->
ACMR17/75 Inf. doc. 4

 

 

 

 

 

 

To Recapitulate Appendix 3
1 page 4
10 db = 10° attenuation of risk
Costs per lab safety cumul
Component . OO
increment
per worker somssea
per year approx.
$ 500 Training ho
safe equipment and
—___— int parr
$ 1000 internal barriers 20 60
1000
(amortized safe carriers 60 120
development
costs)
4000 safe lab
(sealing, ventil- ho 180
ation)
‘
20 000 ultra-secure lab 40 220
( 160
(without
safe
carrier)

 

 

 

 

 
 

 

PINAC OL.

 

   

 

2
o
&
TS
2
ly \~
PHNSIZAL _@ \ 38 |
(Ay
WY; Ui eae
Wiis 7 ¥
_ (PRocrouree x .
CONTANINATIOW CF. | Wy |

   

 

ZL 1 WCAC Wid:

Gon VV EL | PRR KIS
| ow Tad MEW Te ia

    
 

 

| TRAN SMH sy op * A / PRoBKBILITY
[vou Ce \ ac 7 Op ! Oc pox
Sit ons A KRM PE
DEN SI - f eo Renn IS. /
i ° t .
PELERSED te
Paeieras! < _
| Cc AL
Con! )
“AINEN'T EMVIRON MENT.
. __ | \
S 7 ot K
. 8 L a 32
wg Une C cau oro UMT Wise. ey
ve
Lee

[P°sur SL/L TaNOW