5 January 2006

Mr Philip Bishop
Head of Water Branch
Environment Division
Welsh Assembly Government
Cathays Park
Cardiff

CF10 3NQ

Dear Phil,

Cryptosporidiosis in North Wales

You asked for my technical advice on the question “why is it difficult to confirm whether, or not, the water supply is the source of the outbreak of cryptosporidiosis”.  I understand that this question has arisen as a result of dialogue with the Acting Deputy Chief Medical Officer, Welsh Assembly Government.

In addressing this question it is necessary to first recognise that water is not the only source of Cryptosporidium infection in humans.  The organism can be acquired from other sources such as swimming pools, food, and milk, as well as through contact with farm and domestic animals and person-to-person transmission.  That said there are several unique and complicating factors involved with the investigation of outbreaks of disease thought to be related to a supply of drinking water. These include:

• Water treatment works tend to produce relatively large volumes of water which supply a large geographic area
• The water supply to an area can be a blend from more than one treatment works, not all of which may be under suspicion
• By the time an outbreak is recognised, any contamination event has usually passed therefore pathogen testing of responsively collected samples rarely, if ever, informs as to the causative events, I will return to such testing later.
• The mere presence of low numbers of Cryptosporidium oocysts in drinking water does not imply that water is the source of an outbreak; there are many situations where such findings occur in the absence of disease.
• The widespread and general nature of exposure of the population to mains water can obscure the actual source of contamination.
• Gastroenteritis, the illness most often caused by waterborne outbreaks, is not uncommon in the population and has many causes. Many cases, even in the midst of an outbreak, turn out not to be due to water exposure or cryptosporidium.

Descriptive epidemiology can lead to the relatively early finding of a statistically significant association between being ill and residence in a particular water supply zone but this alone does not necessarily mean that illness was caused by the water.  However, in conjunction with the finding of oocysts in mains water or the occurrence, in an appropriate timescale, of an event potentially impacting on the water supply (source, treatment or distribution network), such epidemiological findings provide strong support for a hypothesis that mains water is the vehicle of infection. It is this hypothesis that then generates the impetus for a public health decision to be taken about the need to immediately implement control measures. 

The next step to be taken is to test the hypothesis by carrying out an individual based analytical study however before making the decision to proceed in this way, two factors, known to reduce the power of such a study, must be considered very carefully.  These factors are

• If the population has had a previous exposure this can lead to a higher level of immunity

§          If advice to boil notice has been issued this will influence (bias) the responses of cases and controls 

When an analytical study is done and where the population at risk is large, as is the case in nearly all outbreaks where mains water is a possible vehicle of infection, a case control study is the most appropriate methodology.  Here it is relevant to explain that the methodological approach that would be taken for a private water supply outbreak is not the same as for a public supply because the exposure is typically restricted to a small easily identifiable at risk population. In such situations a cohort study can be used.  The statistical power of a cohort study is generally far greater than can ever be obtained with a case control study and furthermore, a cohort study can usually be completed in a shorter timescale. 

Since, as stated above, an analytical study of an outbreak where mains water is under suspicion requires a case control study approach, it is further necessary to understand that such studies will only generally convey whether or not there was a dose response relationship between tap water consumption and infection rates. It can also provide evidence for associations with other factors and infection rates and, a probability based assessment of the most likely vehicle by which infection was spread.  Thus on its own, the science of a case control study, however well it is carried out cannot “confirm that mains water is the cause of an outbreak” and it is necessary for the findings of such studies to be verified by other information and facts.

It should be appreciated that the gathering of these other relevant facts will not be confined solely to matters relating to the operation of the water treatment works or the results of water tests, they are far more wide reaching and will involve, for example, evidence provided by consumers and other relevant persons through the taking of formal statements using PACE procedures. All this takes time and effort once the epidemiological results become known.  There exists precedence that means the nature of epidemiological studies is such that alone they cannot be relied upon to attribute cause in respect of mains water and an outbreak of cryptosporidiosis. 

I would draw your attention to the tendency for the “results of water tests” to become the focus of public attention during an outbreak and to thus attract a significance far beyond that which can ever be scientifically justified.  There is an old but very useful adage that best summarises the limitations of water testing - the result of a single set of water samples tells you only what was in the sample bottle/device at the time when it was analysed in the laboratory – on its own it tells you nothing about the quality of the water source/supply from where the sample was drawn.   Another common misapplication of the science of water testing is the comparison of a set of results of samples of raw and treated water and the presumption that what is measured going in, is the same as, or can be related to, that measured coming out of, for example, a filter.   There are many excellent peer reviewed studies and PhD theses that demonstrate why such simplistic comparisons are invalid but despite this knowledge, many “experts” as well as lay people (and the media) fall into the trap of making conclusions of this nature.  

As you know, my inspectors are carrying out their investigation of the incident in North Wales and in drawing their conclusions they will take into account the epidemiological findings of the Outbreak Control Team together with all other relevant information gathered in the course of their formal investigation.  The water company is fully co-operating with my inspectors and we will advise the company, the OCT, the Consumer Council for Water and you of our findings, including whether there are grounds for us to initiate either prosecution or enforcement (Water Act 1991).  Obviously it would be prudent for all interested parties to exercise caution about making statements as to “cause” whilst our investigation is ongoing, not least because this can increase the length of time of our investigation considerably.  You will appreciate that anyone who makes such a statement could have material information and we thus have to consider if we need to interview them to establish if that is indeed the case.  You may wish to brief officials and ministers on this particular point, as nobody is exempt.

You can rest assured that the Inspectorate and the Health Protection Agency will be working together to understand what if any, further guidance we can jointly give to health authorities, local authorities and water companies to control and prevent outbreaks of cryptosporidiosis. I am so far unconvinced that the outbreaks this autumn in Wales and England are due to unique circumstances requiring additional research, rather it is probably a matter of refocusing attention and refreshing guidance on the way risk assessment, risk management and risk communication is carried out by all the various agencies.

However, at the present point in time it can be safely concluded that there is evidence to support the hypothesis of an association between exposure to water in Cwellyn reservoir and infection in the community.  What has not yet been established is the origin of the source of contamination of the reservoir and the extent to which, if any, the treatment at the works either contributed to, or mitigated, the extent of the outbreak – both are equally plausible – it is often forgotten in the heat of an outbreak situation that the treatment of the water will have had some beneficial effect irrespective of whether the treatment could be improved upon and thus provide a more robust barrier in the future.   It can be stated with some certainty that not all the cases of infection will have been due to consumption or contact with contaminated water.    It should be appreciated that the actual source of the outbreak will have been one or more infected persons living in or visiting the community combined with the specific conditions that subsequently encouraged both person-to-person transmission and further spread of the organism by contamination of the water environment with their excreta.

In their wish to quickly establish and attribute a cause and a cure, it is quite natural and easy for commentators to focus on water treatment (and the water company) but it behoves all the responsible agencies to address with equal energy the human source and the community dimension (in other words how was it that the reservoir water was contaminated with excreta from infected persons and what can be done to prevent a future occurrence of such circumstances). Time is obviously required for all concerned to identify and evaluate the most, sustainable and cost effective long-term solution(s).  I advise that any interim measures taken by the water company, for example, the installation of additional treatment such as UV, whilst helpful as regards offering reassurance to consumers and the OCT and enabling cessation of short term precautions (boil water notice) should not be seen as an end point (or a proof of cause) by any of the agencies – investigations need to continue.  

I hope this advice proves helpful but please do not hesitate to contact me again if I can be of further assistance.

Yours sincerely,

Prof. Jeni Colbourne MBE
Chief Inspector of Drinking Water

Cc     Jim O'Sullivan, Dwr Cymru Welsh Water
Diane McCrea, Consumer Council for Water Wales