Haddon Matrix and ‘Black Swans’

The Haddon Matrix is an extremely useful way to express the factors associated with a hazardous event and the changes that need to be affected in the host, the equipment and the environment (both social and physical). I have covered the Haddon Matrix in a previous post, in fact to date the most popular post on this blog. I am not denigrating the Haddon Matrix and its usefulness but recent publications Nassim Nicolas Taleb such a The Black Swan: The Impact of the Highly Improbable (2007, second edition 2010) highlight the potential of unexpected, rare events in systems. Taleb does not believe that effort such be wasted trying to predict these rare events but rather than robust systems should be devised to avoid the negative impacts of these events. So does the Haddon Matrix help to prevent hazards or accidents when a Black Swan strikes?

The Haddon Matrix tends to focus on specific events and their immediate impact. The ‘classic’ example often seen on the Web is a car accident where there is a clearly defined agent or host, a clearly defined piece of equipment and a fuzzy but often clearly defined environment at least in the mind of the person who constructs the matrix. The matrix is focused on a particular event usually one that is well known to the person constructing the matrix. The event is singular and derived from thinking about common scenarios of ‘what ifs’. Importantly, the event is divorced and isolated from its complex context. The event is treated as an individual example of an oft-repeated set, as an individual example of a particular kind of hazard or accident. This means that the contours of the event are relatively well know, the limited impact and the limited range of changes that need to be made to the host or equipment clearly demarcated. The event is somewhat simplified by removing it from its context.

Rare events can also be considered within the Haddon Matrix and planned for but events that have never happened or are not within the experience of the constructor of the matrix can not be considered. A series of events could be dealt with by interlinking matrices or even by using Reason’s Swiss cheese model of accidents but each matrix or cheese slice will deal only with a single event not the interconnected system as a whole not the complex and potentially unique relations that these rarities activate within the whole system of which the event identified is only a part. In this case, however, the accident or hazard itself is actually a chain or web of events operating in unison under the influence of the rare event. The exact connections in the system will give the rare event its character. Given the rarity of the event can you be sure that when it happens again the system will be connected, or rather interconnected, in exactly the same manner and so will the precautions that you take have to be exactly the same? As the complexity of the system behind the hazard or accident you are dealing with increases then the possibility that impacts will occur via different connections or pathways is likely to increase. A static Haddon Matrix may not be able to cope with such dynamism that a Black Swan generates within a system.

Black Swan events may also imply that there are two classes of hazards or accidents that need to be considered. The first is the hazard that is known about, one for which have occurred and reoccurred again and again with sufficient regularity that their characteristics can be well defined and clearly defined steps taken to prevent their escalation. The second class of hazards or accidents are those that occur so rarely that each instant is a novel and unusual case with its own set of peculiar characteristics. These events are so infrequent that no reasonable plans can be made to prevent them. It is only after they have happened that we can understand why they happened, what aspects of the system were compromised and then take steps to ensure that the same pathways to failure do not happen again, although the next Black Swan event may be so different as to circumvent our efforts.

If the Black Swan, almost by definition, falls outside the experience of the matrix constructor then is the matrix of any use in these cases? Black Swans may not be predictable but that should not stop attempts to build a robust system to manage impacts. A densely connected system is likely to transmit impacts rapidly from one part to another, maybe along channels or by connections that can be predicted as weak links or pinch points.  Ensuring that there are ‘firebreaks’ in the system, potential break-points in its connectivity, could help prevent a systemic failure even if the exact nature of the rare event is unclear and unpredictable.

The Two-Tier Haddon Matrix

An interesting extension and alternative to the Haddon Matrix is suggested by Mazumdar et al. (2007) (http://www.ciop.pl/21107). They are concerned with aiding the understanding and prevention of operational hazards at a large construction site. The standard Haddon Matrix below could be used firstly for analysing a hazard or disaster and, secondly, for identifying how to prevent it – a two-tier structure. In the first matrix, the pre-event consist of risk build-up, the event itself and then the consequences, whilst in the second matrix there would be pre-event risk reduction, event prevention and consequence minimization.

To help understand both these matrices, they also suggest that ‘fish-bone’ diagrams might help to identify and put into context specific actions and behaviours to help understand both how the event happens and how it might be prevented or at least its impact minimized. In some ways this is similar to following a scenario through the Swiss-cheese model outlined in an earlier blog. The higher up the main arrow an action, the earlier on it occurs in the build-up to an event or in the event and post event sequence of actions. Early prevention stops the sequence of events occurring in the first place.

Each of the points made in the fish-bone diagrams and in the matrices can be assigned a reference code that relates that point to a specific event or action. So A1, for example, could be the initial decision of a person to not follow a particular minor safety procedure, A2 is then the event that results because of this, whilst C1 could be supervisory environment that permits such lax practices. This breakdown of events and actions for pre- during and post-event can be carried out along with the associated preventative measures in the second tier of the matrix that would stop these events occurring.

Using this reference code they then build up a cybernetic analysis of the problem (see their paper for the worked example). Leaving aside the mathematical analysis of the relationships the linking together of the events/actions involves, they do provide an alternative way to look at an accident or hazard. The important point is that they identify positive and negative feedback loops in the accident or hazard, the nodes, and are able to link these loops together to form the overall accident or hazard and its outcomes. Using this sort of diagram it is possible to identify how interconnected certain events or actions are; which events or actions provide bridges between feedback loops and which nodes in the network it would most effective to tackle in terms of disrupting or easiest to control the occurrence of the event or hazard.

Haddon Matrix: Getting The Message Across

The Haddon Matrix is a tried and trusted tool for thinking about managing risk, particularly in public health. The basic matrix contains 3 key elements: the host, the agent (or vehicle) and the environment (subdivided into physical and social). These form the columns of the matrix. The rows refer to pre-event, the event and post-event activities.

	HOST	AGENT	ENVIRONMENT
PRE-EVENT
EVENT
POST-EVENT

In a previous post I illustrated this with a road traffic accident example as below.

This is useful but does it get across two important aspect when trying to manage risk: which of the cells in the matrix is the most significant AND which of the cells is it most effective to try to influence?

Not all the cells and their contents affect the risks of a hazard equally. From public health take the example of smoking - it could be argued that the social and cultural norms that an individual grows up in have a huge impact on their propensity to smoke. The host or individual smokes but what chance did they have given their environment. In other words the environment cells, particularly pre-event (the event being the cancer caused by smoking) has a massive impact on the risk. Likewise, survival of the event depends greatly on the level of health care available including catching the cancer early on, so both the event and post-event are greatly influenced by the environment. Visually the Haddon Matrix might look as below.

	HOST	AGENT	ENVIRONMENT
PRE-EVENT
EVENT
POST-EVENT

This alteration of cell size to match the perceived level of influence of host, agent and environment can help to get across the message as to which of the three is most important.

A second aspect, however, is which of the different elements to try to influence. It could be argued that changing a social or cultural environment is a long-term and difficult process but the one that has the greatest imapct on smoking levels. Effectiveness cna be defined in many ways and it maybe that in the short term targeting the host to change their behaviour is much easier (and cheaper) to produce and places responsibility firmly in the lap of the person smoking. As well as being cheap for health authorities and potentially politically more palatable as it highlights individual responsibility (depending on your political persuasion) altering cells sizes to reflect this, as below, does highlight to the individual that they do have potential control over their fate (whether this is an illusion or not is another question).

	HOST	AGENT	ENVIRONMENT
PRE-EVENT
EVENT
POST-EVENT

So for risk analysis and management maybe it is worthwhile changing the sizes of the cells when discussing both degree of influence and effectiveness of potential actions however this is defined. This may help in targeting resources effectively for the ends in mind.

Haddon Matrix and Hazardous Events

Looking at hazards in different ways, through different conceptual frameworks is always useful as it tends to make you think about things, however slightly, in a different way. A framework often used in injury prevention, in road accident research and public health is the Haddon Matrix. This was devised by William Haddon back in the 1970s for use in road traffic accidents. The basic matrix is divided into 12 cells. The rows are defined by the temporal aspect of the event; pre, during and post, whilst the columns are defined as ‘host’ (you could rethink this as ‘the individual’), ‘equipment’ and two for environment; one for ‘physical’, one for ‘social’. The idea is to fill in each of the cells with key aspects that will influence or did the hazardous event. Effectively you are playing out different scenarios and filling in the cells depending on what factors you see as significant in each scenario. The framework forces you to deal systematically with the nature of the hazard and how it might play out in reality.


The example provided is for road traffic accidents but the basis can be translated to other types of hazard. In the crash, the condition of the individual before the crash may be important for the reasons in the matrix. Each individual will have different characteristics that could be important and each can be included as appropriate. Simiarly, different aspects of the equipment will be important depending on the nature of the crash and so these factors may not be clear until after the event. The environmental factors, seem to be more diffuse and provide a context, that for certain types of individual behaviour and certain equipment failings produce an environment conducive to a hazardous event. Importantly, despite the descirption and divsion of the event into these spearate cells, the contents of each cell depends upon the relationships between the host, equipment and environment. Fro eample, the scoial norms that permit DUI, would not be improtant had not the host not had a seatblet and been drinking. The poorly designed fuel tanks only become significant when the drunk driver crahse and so on.



This framework does have its limitations. The recognition of important factors can be so wide ranging as to be useless in planning if extreme scenarios, with infinitesimal probabilities of occurring are considered. On the other hand, it may not be until the event happens that it becomes clear what factors are important. The matrix will probably be of most use when similar hazardous events are being considered, as similar events would be expected to have roughly similar important factors. The matrix can also be used to identify where particular factors are not relevant. In a pile-up on a foggy motorway, for example, the detailed life history of the individual in the second car in the crash may not have any significance to their survival, it is the general physical conditions that are of over-riding significance. Equipment factors, such as airbag installation, age of car, may have an impact however. In other words, the matrix might be useful to explore the topographies of different hazards or disaster; in exploring the nature or shape of the hazard and what factors dominate that landscape and which are incidental ‘bumps’ on the terrain (please excuse the landscape metaphor, but I am a physical geographer!)

Something useful might be gained by overlaying the matrix with the Swiss cheese model of Reason outlined in an earlier blog. The matrix framework helps to identify the factors that might be important at each stage; the Swiss cheese identifies if a particular trajectory of factors lines up to produce a disaster. The matrix helps identify the possibles, the Swiss cheese, whether these possibles are important in combination. In the case of the BP oil spill, for example, the Haddon matrix could be used to identify key pre, during and post disaster factors, such as the alleged failure in safety procedures and lack of disaster planning. The trajectory arrow of the Swiss cheese model can then be used to assess if this one failure affects the next layer, if one failure or factor then lines up with another to produce the cascade of errors that result in a disater.

Some potentially useful books for assessment of hazards of injury are:

Injury Prevention in Children by David Stone (2011)



Injury Control: A Guide to Research and Program Evaluation by Rivara et al. (editors) (2009)




Injury Epidemology: Research and control strategies by Leon Robertson (2007)