Methods and feasibility of conducting pragmatic clinical trials in small animal first opinion practice.

Doit, Hannah (2023) Methods and feasibility of conducting pragmatic clinical trials in small animal first opinion practice. PhD thesis, University of Nottingham.

PDF (PhD thesis with post-viva amendments included) (Thesis - as examined) - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
Available under Licence Creative Commons Attribution.
Download (4MB) | Preview


Chronic kidney disease (CKD) is an important cause of morbidity and mortality in cats, for which many research uncertainties remain unanswered. As for much of veterinary healthcare, the evidence base for treatment decision-making is limited. For the small number of research questions where randomised controlled trials (RCTs) are published, sample sizes are small, and the external validity of results can be limited because the patients included are not representative of the wider population who may have comorbidities. In addition, existing published research evidence does not always address outcomes of importance to treatment decision makers.

Pragmatic trials are a very new concept in veterinary healthcare where a literature search found only one pragmatic trial had been designed and carried out in full. However, they are well established in human healthcare. They are a sensible solution to many of the problems with the veterinary healthcare evidence base for several reasons. They address questions which are important to stakeholders and address outcomes of important to decision makers. They are designed to take place where everyday care happens and have less strict eligibility criteria than traditional RCTs, resulting in the inclusion of more diverse patient populations who represent the patients for whom the research will be used. The results are designed to be useful in everyday clinical decision making, in everyday clinical practice. They test real- world treatments and use flexible protocols, being designed to represent normal practice as far as possible. Their results are more widely generalisable than traditional RCTs and they are less expensive to carry out.

The aims of this PhD work were to investigate the existing published measures of treatment success in trials for cats with CKD and from there to establish the most important and most appropriate outcomes to use. This list would be designed to provide recommendations for future pragmatic treatment effectiveness trials of which treatment outcomes to assess and prioritise. In addition, this work aimed to establish the feasibility of extracting and using electronic patient records (EPRs) from first opinion veterinary practice, as a data source for clinical trials for these patients. This data source was chosen because EPR use in pragmatic trials is well established in human pragmatic trials, and because the EPR is the location where the presenting signs, diagnosis, treatment and management and outcomes of large numbers of ordinary cats with CKD is already recorded and held.

A systematic review of outcomes assessed in published CKD treatment trials was carried out. This found a broad spectrum of outcomes that were assessed in the published literature. No core set of outcomes (COS) recommended for assessment in CKD treatment trials was found and little consistency was identified in the outcomes that were assessed between publications. To address this problem, research was conducted generating a COS for feline CKD. The panel of stakeholders involved in this process included an international panel of cat owners, clinical representatives, regulatory agencies and journal editors. A proposed list of important outcomes for a future COS was brought together via a three round eDelphi and an in-person consensus meeting. The final list created contained 29 core outcomes, grouped into four key areas: the veterinary consultation, blood and urine testing, living with CKD and CKD progression. Further refinement of this COS before it is finalised for inclusion in clinical trials is recommended, to streamline the outcomes into domains, potentially reduce by consensus the size of the final COS and agree by consensus the instruments to assess each domain.

One key outcome identified by the COS generation process and known to be of importance to decision makers for cats with CKD was quality of life (QoL). A systematic review of the published literature was conducted to identify all tools used for assessing feline QoL, and the range and quality of tools available. Many of the studies found that discussed QoL either did not assess QoL at all, or assessed QoL with only unvalidated, oversimplified tools. Few publications were found that assessed quality of life in a structured way and few used validated tools to assess QoL, although a validated tool for assessing QoL for cats with CKD was found. Once the full COS for cats with CKD was established, work was conducted to examine whether some of the outcomes highlighted could be measured using data collected in practice management software systems in veterinary practices as part of routine veterinary healthcare. Data transfer from the databases of veterinary practices to those of laboratories, insurers and microchip registration companies is already well established, using XML schema. XML schemas describe the structure and content of the required data extract and present the data in a format which can be easily read by humans or computers. An XML schema was already published for the transfer of data pertaining to clinical research, ‘Clinical Evidence Schema v1.0.5’. To accommodate data from multiple PMS and multiple veterinary practices this schema was adapted, restructured and some new data fields were added. A six-month data batch in XML format was extracted by a PMS, in accordance with the data specifications of the new schema, from 282 veterinary practices. Additional data was also provided by the PMS as Excel files. The whole data batch was deidentified using bespoke script in Microsoft Visual Basic. It was then cleaned and uploaded into a bespoke database written in MySQL. This destination database was then examined and explored using scripts written in data manipulation language and run on the dataset via the SQL Command Prompt.

The usefulness of the extracted patient data for possible treatment trials for cats with CKD was then established. Cats with CKD were identified using MySQL scripts, generating a disease prevalence of 2.8%. Validation showed this method to have 83.3% sensitivity, 99.5% specificity and a 40% false positive rate. A couple of relevant outcomes from the COS were extracted for feline CKD patients including blood pressure, bodyweight and survival time. CKD treatment interventions e.g. intravenous fluid therapy, or named therapeutics could be successfully identified within patient records and the longevity of these patients followed over time.

In conclusion, EPRs are used within human healthcare for pragmatic trials, however, very few pragmatic trials exist for veterinary healthcare. This PhD thesis has demonstrated that veterinary EPRs are a valuable and feasible data source for research. Pragmatic style trials are likely to address many of the evidence gaps which currently exist in veterinary medicine. Future veterinary research should look to EPRs as a proven, feasible data source, employing the use of COSs to direct the most important outcomes to extract. The next steps in this work should explore the potential for, and practicalities of, running treatment trials within a first opinion veterinary practice environment. This will enable the profession to make real progress into filling the many evidence gaps in existence.

Item Type: Thesis (University of Nottingham only) (PhD)
Supervisors: Brennan, ML
Dean, RS
Duz, M
Keywords: cats, cats diseases, kidney disease, veterinary medicine, evidence based medicine
Subjects: S Agriculture > SF Animal culture
Faculties/Schools: UK Campuses > Faculty of Medicine and Health Sciences > School of Veterinary Medicine and Science
Item ID: 73934
Depositing User: Doit, Hannah
Date Deposited: 29 Feb 2024 13:20
Last Modified: 29 Feb 2024 13:20

Actions (Archive Staff Only)

Edit View Edit View