We are scientists and, as such, we seek accurate data upon which we can form solid treatment recommendations with regard to our patients' health, comfort, lifestyle and well-being. Such is the intent of evidence-based medicine (EBM), which has been defined as 'the conscientious, explicit and judicious use of current best evidence in making decisions about the care of individual patients'. For the purpose of training clinicians to critically analyse available information, study designs have been stratified into classes ranging from strongest (Class I) to weakest (Class IV) evidence (Figure 1).

Figure 1. Study designs.

In addition to seeking strong (class I and II) evidence through clinical trials, veterinary surgeons should seek meaningful, relevant evidence that helps them to answer the questions that their pet-owning clients are asking. Indeed, one of the essential steps of effective utilisation of EBM is training clinicians and clinical researchers to ask the right questions. The mnemonic 'PICO' is helpful in guiding clinicians to ask the right questions and to evaluate the relevance, usefulness and meaningfulness of clinical trial results to their patient:

 P: Patient/problem--do the results of the study pertain to my patient?

 I: Intervention--are the interventions evaluated in the study a consideration for my patient?

 C: Comparison--to what are the results of the intervention compared? A control group?

 O: Outcome--what outcome is important to the patient (pet owner)?

Unfortunately, there is often a misunderstanding between ourselves and our clients with regard to what outcomes from orthopaedic surgery are important. We can often get a sense of what outcome is of importance to our individual clients by listening to their questions. Why did the pet owner seek our services for this condition? What is their desired outcome? What will they use to define 'success' or 'failure' of treatment? Our relationship with our patients and our patients' families is not unlike that of a paediatrician. Together we make decisions for our patients on their behalf, 'by proxy' if you will.

Orthopaedic care is unique in that our treatments seldom impact mortality. Instead, our clients often present their pets to us in hopes that we can improve their health-related quality of life (HRQL) as compared to natural progression of the disease/injury. As such, a well designed study that collects objective data points that are of little direct concern to our clients may be of little real help in our answering their questions. 'Will he be able to climb the stairs'? 'Will she still be able to go on 10 mile hikes'? 'Can I keep her in competition'? 'Will I be able to reduce her reliance on anti-inflammatory medications'? These are the questions asked by our clients. With regard to juvenile-stage canine hip dysplasia (jCHD), our clinical studies often measure parameters such as dorsal acetabular rim slope, acetabular angle, radiographic osteoarthritis scores and force-plate data as surrogate measures for what we really want to know. It takes a relatively large leap to get from any of these measures to accurate answers to our clients' questions. This is often the case in medical studies where we study what has been called disease-oriented evidence (DOE). In contrast, patient-oriented evidence that matters (POEM) measures patient morbidity or mortality, the patients' ability to perform some meaningful and measurable lifestyle task or health-related quality of life.

Because executing large-scale, prospective, blinded and placebo or sham-controlled randomised clinical trials (class I) is time-consuming and costly, we ideally seek to design studies that will provide meaningful answers to relevant questions regarding applicable interventions that pertain to a large group of patients. HRQL questionnaires are being used with increasing frequency in human orthopaedic studies for this purpose. In essence, before we leap into randomised clinical trials in search of objective answers, we should first evaluate our questions to be sure that the answers yielded by the study are likely to be meaningful and relevant to our clients. What do we really want to know? What do our pet-owning clients really want to know?

HRQL questionnaires are one instrument being used in randomised clinical trials to arrive at patient-orientated evidence that matters. Development of such an instrument is a rigorous process. The questionnaire then requires validation to be certain that the questionnaire measures that which it intends to measure across a demographic spectrum of pet owners and across a spectrum of time. The process of developing a HRQL questionnaire requires researchers to identify patient behaviours that a demographic spectrum of pet owners defines as measurable, describable and meaningful with regard to their pets' quality of life as it is related to the health condition in question. One such method currently being employed at Colorado State University is described herein. The process is initiated by a series of client interviews in order to identify these meaningful, measurable patient behaviours. A list of positive and negative terms frequently used by pet owners to describe each of these behaviours is developed in order to facilitate their measurement in the eventual HRQL questionnaire. These descriptors are elicited from pet owners by first asking them several general 'open-ended' questions encouraging them to describe their pet's current health status and their pet's normal health status prior to developing the medical condition. Next, a semi-structured interview is conducted with each pet owner. The semi-structured interview contains a set of questions regarding observable pet behaviours under consideration. For each question there is a set of 'floating prompts' to facilitate pet owners' ability to provide positive and negative descriptors if needed. The goal of this process is to a develop a list of terms that are consistently used across a demographic spectrum of pet owners to describe positive or negative outcomes with regard to each observable behaviour that they regard as meaningful in their pet's HRQL. Once the HRQL questionnaire is developed, there are several ways to test the validity of the instrument, one of which is called 'test-retest' validation. It is important to realise that some HRQL questionnaires may be valid for one pet owner demographic, but not for another due to variation in regional lexicon and/or what they regard as meaningful and important for their pets' lifestyles.

In our experience, it rapidly becomes evident during this HRQL questionnaire development process whether or not previously published clinical trials of therapeutic interventions for a given medical condition have answered the questions that are meaningful to one's client pet-owners. As an example, in one of our HRQL studies currently in progress, it quickly became apparent that the ability to return dogs to an unrestricted level of activity (absence of physical dependence upon the owners for patient confinement) without ensuing lameness was the desired therapeutic outcome. Previously published studies of radiographic osteoarthritis scores, veterinarian-measured lameness scores, and even force plate data at a walking gait are not ideal measurement parameters for this condition because relatively large assumptions or inferences are required to assume that a given pet will benefit in a meaningful way from any of the interventions evaluated. Our client surveys suggest that return of each of these parameters to normal 'pre-injury' levels would be meaningless if the owners observed lameness whenever owner-enforced patient confinement was not instituted.

References

1.  Berg AO. Dimensions of evidence, in Geyman JP, Deyo RA, Ramsay SD (eds): Evidence based clinical practice,Boston: ButterworthHeinemann,2000, pp 21-27.

2.  Cockroft P, Holmes M. (eds) Handbook of EvidenceBased Veterinary Medicine. Oxford: Blackwell Publishing, 2003

3.  Rosenberger PH, Jokel P, et al. Shared decision making, preoperative expectations, and postoperative reality: differences in physician and patient predictions and ratings of knee surgery outcomes. Journal of Arthroscopy & Related Surgery 2005; 21: 562-569.