Why is N of 1 a problem? Simply because the fact that something happened in a particular way once does not provide strong evidence for believing that it will do so again, and, generally speaking, each additional observation of the same happening adds greater confidence that a true relationship exists. Of course, in most instances, N of 1 in terms of a veterinary surgeon does not mean N of 1 in terms of cases that vet has treated; it is the cases that are the more important and it makes a great deal of sense to use this resource. Although it is natural and intuitive to use previous experience to guide management of present and future cases, it is possible that this approach, limited to subjective and anecdotal 'memory', may not optimise outcomes for each individual patient treated. It would be far preferable to be able to access summary information about subsets of animals seen previously by the vet or practice that had a similar presentation or lab result or procedure to the case in hand, to ensure an objective choice between further case management options. Taken as a continuous process, as new cases are added, summary information is updated and the weight of evidence to support or refute particular case management options grows. This is the crux of so-called case-based reasoning, and on the face of it, it is a very attractive proposition and it could be argued that everyone should be doing it. Confined to the caseload of a single veterinary practice, what this approach may lack in terms of case numbers, it makes up for in terms of population relevance, in that, for example, regional differences in risk factors for disease are unlikely to be relevant. Extending beyond the caseload of a single practice, it is also very attractive to consider the power of pulling together information from large numbers of cases presenting in a variety of settings and being able to use constantly updated summary information to optimise patient management. However, a number of considerations need to be borne in mind.

Fundamental to being able to use information about cases is the need to store that information in a suitably structured format, thus the electronic patient record takes on great importance. Imposing the necessary structure on case information to allow meaningful interrogation and retrieval comes at a cost: that of needing to categorise clinical and diagnostic information. Where the system involves more than one veterinary surgeon, it is also of vital importance that such categorisation is standardised for all and adhered to, i.e., strict case definitions are needed, and this has proved very difficult to achieve. However, the advent of electronic patient recording systems for veterinary practice offers a ray of hope that such an idea may indeed become possible.

Extending this vision further, the natural progression is to set up sentinel practice networks. It has long been recognised that first-opinion (primary care) practitioners act as the 'eyes and ears' of the veterinary profession. First-opinion veterinary practitioners encounter disease as it occurs in the field and, importantly, in assessing the relative significance of animal diseases, can provide an acceptably accurate estimate of the population at risk. A number of studies, in which a network of first-opinion veterinary practitioners have been recruited to collect prospective data relating to specific animal disease problems for a fixed period of time, has served to demonstrate the success of this approach. However, many improvements are still needed if sentinel veterinary practice-based research is to fulfill its potential. In the first instance, there is a need to recruit whole practices, rather than enthusiastic individual veterinary surgeons as practice representatives. This will enable more complete disease reporting and better estimations of the population at risk. Networks of sentinel practices with more long-term or permanent affiliations to research institutions will be better able to address different animal health problems over time, and will be much more productive than those created for a finite time period to address a single problem. The prospect of a coordinated network of practices, with a long-term remit to record data on disease occurrence in a well defined population of animals seems almost irresistible. Nevertheless, in the past, several factors have combined to forestall the full realisation of this appealing arrangement, although recent developments in information technology may offer some solutions. Advances in Palm-Top computing have the potential to reduce to a minimum the additional clerical requirement of disease recording, which has previously been cited as a serious disincentive for practitioner participation. This, in combination with the availability of powerful computers with versatile software, will also lead to more rapid collation and analysis of data. In turn, this will facilitate timely and meaningful feedback of research findings to practitioners. Further incentives for practitioners could be provided if national veterinary regulatory bodies were prepared to offer continuing professional development/ continuing education credit for participation in sentinel practice research.

In the past, questions have been raised as to the validity of data derived from sentinel practice-based research because of a perceived inaccuracy of diagnosis in first opinion practice. However, it is to be hoped that a better understanding of the scope of sentinel practice-based studies, coupled with developments in diagnostic technology and communications, will convince doubters of the compelling merits of this approach as a vital component of the management of animal disease and welfare in the future.

References

1.  Cockroft P, Holmes M. (eds) Handbook of Evidence-Based Veterinary Medicine. Oxford: Blackwell Publishing, 2003.

2.  Thrusfield MV. Veterinary epidemiology (3rd edition). Oxford: Blackwell Publishing, 2005.