Introduction to Fire Safety Management

Figure 5.11
Graph showing the initiators of fi res that resulted in more than £250 000 damage
0
10
20
30
40
50
60
70
number of fires
Deliberate ignition
Electrical
Friction, heat and sparks
Smoking materials
Naked light
Cooking
Spontaneous combustion
Oil
Fireworks
LPG cutting equipment
Spread from secondary fire
Under investigation
Unknown
source
How the £250 000 plus fires in 2004–2005 started

---

Principles of risk assessment
93
Semi-quantitative analysis
The use of semi-quantitative techniques for risk assess-
ment is widespread and it is often referred to as a quan-
titative method; however, it is easily seen that risk is 
expressed as a numerical value, the estimation of the 
magnitude of the risk is in fact subjective and therefore 
qualitative. A semi-quantitative evaluation of risk allows 
numerical values to be assigned to both severity and 
likelihood in the absence of data.
Like the qualitative approach, the semi-quantitative 
technique is essentially subjective and is based upon the 
risk assessor’s personal interpretation of the level of risk, 
but with the added assistance of a risk matrix so that a 
risk factor can be calculated numerically.
For example, the likelihood of a particular event 
occurring is assigned numerical values as shown in 
Table 5.2.
While it is easy to differentiate between the two 
opposing defi nitions at the end of each scale, there is 
often confusion with the words used to separate out the 
mid range.
There is no laid down criteria for either the words 
used to describe the value or the numerical values within 
the scale. A 5 by 5 scale is arbitrarily used here, though 
there are a wide variety of scales used from 3 through to 
8 and sometimes even 10.
Just as the likelihood rating is assigned numerical 
values, so too is the severity rating. Again using a scale of 
1 to 5 and adapting the HSE’s simple qualitative estima-
tions of severity, an example of how numerical values 
may be assigned is shown in Table 5.3.
In addition to the severity of the outcome of an 
event for an individual, the total numbers of people that 
may be affected must also be considered, as in the case 
of calculating the likelihood of the rise in fi res started 
by discarded smokers’ materials in places of public 
entertainment (if a smoking policy exists). The resultant 
effect upon persons being able to escape safely may 
also revolve around the numbers of persons present.
Having assigned numeric values to each element of 
risk, a risk matrix can then be completed to provide a 
calculation of an overall ‘risk factor’.
The resultant risk matrix can be used to provide 
and develop an action plan which may also be assigned 
numbers so that priorities can be identifi ed. Risk 
matrices are often colour coded to provide a visual 
concept of whether or not the residual risk is tolerable or 
acceptable.
An example of a semi-quantitative risk matrix, which 
incorporates the numerical scales discussed above, is 
shown in Figure 5.12.