Improved implementation of medical tests – barriers and opportunities

The use of audit in the laboratory

Clinical audit is a well-established technique practiced in many different areas of healthcare. It consists of measuring a clinical outcome or a process against an evidence-based medicine standard [9]. Discrepancies between actual practice and the standard are addressed through a quality improvement process which should then result in an improvement of care. This process as applied to laboratory testing is very similar and indeed overlaps with the process that the C-VPLM is proposing for improved test implementation.

While clinical audit is widely used in healthcare, the application to laboratory medicine is very limited and largely confined to Europe, in particular the United Kingdom, where it is a key part of clinical governance [10]. In the UK the laboratory audit process is supported by a number of professional bodies including the Royal College of Pathologists (RCPath) and the Association for Clinical Biochemistry. The RCPath for example, as well as providing training resources, also provides a certification process to ensure that the audit process meets defined standards and outcomes [11]. The RCPath also provides a library of certified audits which provides a useful resource for laboratory professionals.

Within laboratory medicine audits have largely addressed processes occurring either inside the laboratory or are confined to determination of test request patterns, otherwise known as test utilisation. This is because such audits are much easier to perform given they examine data which resides largely within or accessible by the laboratory. Furthermore such audits are valuable in that they provide important data such as the UK Atlas of Variation in Diagnostic Practice [12]. Similarly, and equally revealing, is a rare but recent audit of test requesting conducted in Australia which showed that the Australian Federal Government’s strategy to limit requests for vitamin D was a complete failure and furthermore was likely to be adversely affecting subjects with low vitamin D levels [13].

However if audit is to assist with and/or be integral to better test implementation, then it must also address processes outside of the laboratory, and namely ones that relate to the impact of the test on the clinical pathway. A recent audit of troponin testing in the UK further highlights this deficiency. Although not conducted by laboratory professionals the authors reported on the proportion of English hospitals who were now using high-sensitive troponin testing in various algorithms to rule out acute myocardial function [14]. There was nothing in the report on the outcomes of this testing, such as whether there was an overall efficiency of patient discharge, but this is the type of implementation information which is required.

Stakeholders outside the laboratory

A key concept that is central to test implementation is the clinical pathway and the so-called stakeholders who are affected by the change in that pathway brought about by a new test. The impact of a test is often just considered in terms of the patient, generally expressed as clinical effectiveness. But in some pathways the impact of a test goes beyond the patient and it is necessary to consider a broader range of stakeholders. An example is shown in Table 1 which lists the possible stakeholders involved in a pathway where high-sensitive troponin (hsT) testing is used to rule out acute myocardial function.

The stakeholders will vary somewhat between different applications of the pathway but also shown in Table 1 is that the benefits or disbenefits will vary between the stakeholders and that underlies the importance of taking account of their contributions to the pathway [15]. The task for the laboratory professional is to liaise and collaborate with all of these stakeholders which may not be an easy task. Some of these are busy clinicians and while most of them will value aspects of the laboratory service such as assistance with result interpretation by laboratory specialists, not all will welcome those same specialists getting involved in the clinical pathways of their patients. The trend toward laboratories becoming independent entities physically remote from clinical care makes the development of professional relationships with clinical staff more difficult but not impossible and the goal is to foster and develop collaborative relationships according to local circumstances.

Collaboration with non-clinical stakeholders is also important, most particularly those involved with healthcare finance. The activities of the C-VPLM include working with finance managers in the UK’s National Health Service (NHS) as part of its research into the implementation of hsT. Those discussions have indicated that relationships between laboratory professionals and finance managers are relatively rare but when they occur, they can be immensely useful and certainly helpful in for example, presenting a business case to support the implementation of a test.

Overcoming this significant barrier will much depend on local circumstances. It will likely be easier for those laboratories that have a physical presence close to clinical care and/or have retained strong links to, and relationships with, clinical and other critical healthcare teams throughout their healthcare facility.

Information technology and the data of implementation

The process of implementation relies on gathering data and information that goes beyond the laboratory. Some of the data is in addition to those parameters usually measured in trials of clinical and cost effectiveness and they relate to what takes place when tests are implemented in local practice. They also reflect a category of data in which there is growing interest and collectively known as real world data or RWD.

The importance and attention being devoted to RWD has come about for two reasons. First is the growing view that what happens in clinical randomised trials does not reflect what happens in real clinical practice [16]. Thus there is now increased interest in observational type studies, particularly if they involve large subject numbers, and they are seen as providing necessary and complementary results to randomised controlled trials. Such studies would also recognise the contribution of all or at least a broader range of stakeholders in the pathway and not just the selected ones that might be the outcomes described in clinical trials. There are other and important differences between trials and large observational type studies providing RWD which are summarised by Ma et al., particularly as they relate to laboratory medicine [17].

The second reason for the interest in RWD is the complementary development of what is known as data mining or the ability to interrogate large data sets with modern information technology [18]. The use of RWD and data mining has already been used to generate more reliable reference intervals for various analytes with the aim of harmonising these across large numbers of laboratories [19]. The goal now is to use similar techniques to look at implementation data from the impact of tests on clinical pathways [20]. There are two important barriers or challenges to address in this area.

The first is the need for information systems which are up to the task. Better implementation requires IT systems to be in place that can be designed and programmed to capture the required information easily and without human intervention. Anecdotal evidence suggests that this is often not the case in many hospitals and healthcare systems, and instead, is often reliant on manual interventions. The latter might be satisfactory albeit expensive for a relatively short period involving a clinical trial but are not sustainable for a long term or ongoing quality improvement process like test implementation.

The second barrier is related to the privacy of patient and medical data in general. This is a highly sensitive issue in healthcare and access to such data is now much more restricted than a decade or so ago. However appropriate safeguards can be established including the de-identification of patient information and such anonymised data will be perfectly satisfactory for the types of measures required to assess whether tests are delivering the expected benefits when used on a routine basis.

Resources and training

It will hopefully be clear from the preceding text that the C-VPLM envisages a new or at least a different role for laboratory professionals and that is one where resources and time are devoted to test implementation and clinical pathway management. Below we will argue that this is a quality related issue but before that, we raise several and overlapping issues related to resources and training.

Given that there will be few in the laboratory profession who have spare time available for new activities, the key question is who will perform such implementation type activities? We would argue that this is more a question of priorities and it will be up to local decision makers to determine, for example what should be the priorities for quality related activities. Since the provision of more resources is unlikely, this means giving up certain activities in favour of others.

A second and equally important question concerns the resources of tools and training. Test implementation requires a particular skill set which will include the ability to liaise and collaborate with stakeholders outside of the laboratory, together with a knowledge and an understanding of economic and business related concepts. The latter might include the ability to prepare a business case for the implementation of a new test.

Other skills that also might also be required include the application of economic modelling; it should be emphasised that this does not need to include a detailed understanding of how to do modelling but more of knowing where such a tool can be applied [21]. Other tools of relevance are time driven activity based costing which can be applied to the optimisation of clinical pathways [22]. Finally, as we have indicated previously, knowledge of data mining and even artificial intelligence may also be skills that laboratory professionals will need in the future.

It is unlikely that any of the skills mentioned above are included in the current curricular for laboratory professionals whether scientists or those with medical qualifications. Indeed there may be a view that anything that relates to finances and economics is non-scientific. That is clearly not a sustainable argument if the profession wants to claim that they deliver value, which must in certain circumstances involve a financial dimension. So there is an opportunity here for professional bodies in the first instance to provide education and training. A promotion of the value of audit and how to perform it would be a welcome initiative particularly in those countries where audit is not currently included in clinical practice. The C-VPLM is planning a workshop on the use of economic analysis and tools but the interest of other professional bodies is needed in this area.