Introduction

Discussion of prognosis may be viewed as one of the most important elements of the practitioner- client relationship in veterinary practice. Although this may be considered to be a 'routine', subjective component of most veterinary consultations, prognosis is firmly underpinned by the science and art of risk assessment and communication. Although, until recently, this subject has received little formal attention in undergraduate curricula in the medical sciences, assessment and communication of risks is a fundamental part of the practice of evidence-based medicine (EBM). As such, there exist a number of formal quantitative scientific approaches for the estimation and expression of prognostic information and a small but growing literature describing their application in veterinary clinical settings. At the same time, there is also a growing body of research concerned with understanding people's perception of medical risks and an increasing awareness of the need to communicate this clearly to avoid confusion and misunderstanding. It should be incumbent on the medical professions to be familiar with these measures and issues so as to be best able to use the most up-to-date information to provide prognostic advice. This argument is redundant in the absence of objective data on which to base such advice, and this has, until recently, been the case for the majority of animal diseases. However, clinical epidemiological studies, designed with explicit EBM objectives, are starting to appear in veterinary literature and these can only have utility if their findings are used properly in practice.

Risk Assessment

The type of prognostic information available is dependent on the type of study undertaken and standard epidemiological and EBM texts describe the derivation of a number of basic measures to estimate (absolute risk) and compare risk (relative risk, risk difference, odds ratio) between groups of animals exposed to different treatments. From these, more complex measures, such as number needed to treat (NNT--the number of animals that need to be treated to avoid one bad outcome) or relative risk reduction, can be derived and may, nevertheless, be more intuitive from an owner's perspective. Provided data are available, such calculations can readily be performed by hand or by using a calculator or even a spreadsheet. More specialist studies, recording time to an event, such as death or implant failure (survival studies), may be able to estimate parameters such as median survival time (for example, the time taken for 50% of patients to die or experience implant failure) or the proportion of treated animals still alive after a fixed period of time or the hazard ratio (the relative likelihood of survival in treated and untreated animals to a fixed point in time). In practice, all of these measures are probabilistic expressions of risk and veterinarians need to be comfortable with distinguishing between them and interpreting them correctly. It is of paramount importance to comprehend that all these measures are estimates and describe the average risk experienced by the population: they cannot be interpreted literally when considering an individual patient.

Risk Communication

Most of the research in this area comes from human medicine and relates to a person's perception and interpretation of his/her own risk, and the situation described below may be further complicated in veterinary medicine by issues some owners may have with making decisions for their animal. Effective communication of risk is difficult because the owner generally understands none of the various quantitative measures of risk. However, research has shown that qualitative statements of risk such as 'rare' or 'more often than not' mean different things to different people and may result in a false impression of risk. Therefore, there is a need to 'translate' quantitative measures into a language that can be correctly interpreted by owners. Studies have shown that patients prefer risks expressed as relative risk (i.e., how much is my risk reduced by having the treatment relative to not having the treatment), whereas some experts believe that NNT is the most intuitive measure. Furthermore, there is evidence that people make irrational judgments about risk, i.e., a 5% chance of dying is considered by some to be worse than a 95% chance of survival. Of course, different owners' perspectives of the same risk can be variable and are influenced by many factors such as age, culture, previous experience in the same and other species and cost. This is further complicated when important issues such as quality of life come into consideration, with some owners valuing a 'short-term survival, near-normal quality of life' scenario more highly than a 'longer-term survival, reduced quality of life' scenario and vice versa. Therefore, it is most important to assess the owner, individualise the risk message and achieve a balance between providing the owner with sufficient information and overloading them with too much.

In summary, the application of epidemiological tools has the ability to provide better, objective information on which to base prognoses.

However, assessment of animal owners and delivery of clear, individually tailored risk communication messages remain of paramount importance to the clinician in making best use of this information.

References

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