Discussion

This service evaluation demonstrates that patients with suspected HF and raised natriuretic peptides have a very high rate of adverse outcomes in the short term. Timely HF specialist involvement is associated with better outcomes, specifically lower rates of hospitalisation. These novel, real-world data support the implementation of HF pathways which comply with NICE guidelines.

The main strengths of this work are that this is a systematic evaluation of consecutive real-world patients with suspected HF and that those referred to specialist services received guideline-based treatment delivered by an established comprehensive HF service. It also provides valuable contemporary data from a cohort representative of a very large population of patients with poor outcomes who remain understudied. Furthermore, while NICE recommendations on referral pathways were produced because of the adverse outcomes in this population, the waiting-time targets predate relevant supporting evidence. The main limitation of the data is that they are retrospective and observational and therefore demonstrate an association rather than clear causation. There will be some unadjusted confounding factors and the possibility of referral bias influencing outcomes. However, while the non-compliant group were slightly older, the compliant group had much higher NT-pro-BNP levels, which itself is independently associated with a higher risk of adverse outcomes.11 The much lower hospitalisation rates seen in the compliant group, therefore, suggest that there is a real benefit in being seen in a specialist HF clinic.

These findings are consistent with the substantial body of evidence demonstrating that outcomes for patients with HF are improved within appropriate systems of care.6 Specialist multidisciplinary care reduces both all-cause and HF hospitalisation.7 Furthermore, in contrast to many drug and device therapies, the literature suggests that all patients with HF, including those with HFPEF, benefit from specialist care. Again, our data are consistent with this in that we included all patients regardless of left ventricular ejection fraction. The lower rates of hospitalisation observed in the compliant patients are likely to be due, at least in part, to specialist care. Expert care involves accurate diagnosis and management including appropriate use of therapies which improve outcomes. Given the short follow-up period of this study, drug therapies will have been the most important interventions and have previously been well documented to reduce hospitalisations over short follow-up periods for patients with HFREF.10 An additional benefit for compliant patients with HF and in particular those with HFREF in this study is likely to relate to ongoing access to the multidisciplinary team. Patients in the non-compliant group would have been much less likely to be managed by the multidisciplinary team who receive the vast majority of new referrals from specialists.

Since raised natriuretic peptides are known to predict an adverse outcome in patients without overt HF as well as in those with HF,11 12 it is likely that better outcomes in the compliant group are not solely due to better management of HF. For example, HF physicians are also well placed to manage other CV diseases, such as atrial fibrillation, which can mimic HF and causes elevations in natriuretic peptides. We did not individually evaluate the causes for each admission, however, it seems likely that early assessment by a physician led to changes in management for patients without HF in addition to those with HF. Similarly, it is also of interest that all-cause hospitalisations were lower and not just CV hospitalisations. This may in part be explained by the fact that HF causes multisystem adverse effects. A proportion of the observed hospitalisations may have been due to sequelae of HF which are not obviously CV, such as renal impairment. Furthermore, HF hospitalisations may not always be recorded as such. For example, while patients presenting with oedema due to HF may be recorded as an HF hospitalisation, they may also be recorded as being hospitalised due to fluid overload a non-CV cause. Misclassification of some causes of hospitalisation may be more likely as non-clinical administrative staff record hospitalisation causes in our institution, a practice which is the norm in the UK.

The low referral rate (28%) of patients with raised natriuretic peptides is noteworthy. More detailed data on those patients who were not referred are not available. However, the local healthcare provision is organised such that our centre provides the vast majority of secondary care, and thus specialist services including cardiology and HF. A small minority of patients may have been referred elsewhere, such as to the private sector. However, it can be reasonably assumed that most of the non-compliant patients were managed in primary care by their GP. As the HF pathway was established shortly before this evaluation commenced some GPs may not have known about the service despite our efforts to raise awareness and the incorporation of the referral pathway into their standard online resources. It is also likely that in many cases GPs underestimate the severity and complexity of HF and felt they could manage these patients themselves. Contemporary data suggest that GPs have relatively high levels of confidence in the diagnosis and management of HF.13 Another reason for the relatively low referral rate may relate to GPs testing patients in the non-compliant group in clinical situations outwith the pathway. Our pathway specified that natriuretic peptide levels should only be checked for patients with a potential new diagnosis of HF in keeping with NICE guidelines. However, NT-pro-BNP may, in some cases, have been used to monitor patients with known HF or even in the absence of a clinical suspicion of HF. Some patients may also have chosen against referral but again this is likely to account for a small proportion of the non-compliant patients.

The purpose of including referral-time targets in NICE guidelines is so that patients with suspected HF are seen rapidly as they are at high risk of adverse outcome. Despite offering a service compliant with these targets, we identified notable delays between a positive NT-pro-BNP result in primary care and referral to clinic. For those on the 2- week pathway, the interval between a positive NT-pro-BNP and referral to the HF clinic was longer than the interval between receipt of the referral and patient review in clinic. This seems counterproductive and unnecessary in a time-critical pathway but was not associated with worse outcomes. The absence of worse outcomes in patients who were delayed may be due to the low numbers who did not experience a delay or because rapid specialist review compensated for what were relatively short delays experienced earlier in the pathway.

The true magnitude of the benefit of early specialist review for patients with suspected HF and raised natriuretic peptides needs to be clarified. The body of evidence supporting specialist care for patients with HF means that it would be difficult to justify a randomised controlled trial that would deliberately put patients at risk. However, it may be possible to compare outcomes in regions with no established NICE compliant services to other areas with established compliant services. Our findings are in keeping with the totality of the evidence and provide further data supporting specialist care. However, since it is a single-centre study, there may be centre specific effects at work and it is important to demonstrate that these findings can be reproduced in multiple centres. Subgroup analysis has been deliberately limited to prevent analysis of small groups and reduce the likelihood of spurious findings. As a result, the benefits of specialist review for particular patient subgroups, for example, those with HFREF versus HFPEF are not known from these data. The growing HF burden means that further research to build on this work and answer these questions should be considered an urgent priority.