• Heart failure
• audit
• inpatient
• mortality

The recent publication of the 2009/10 national heart failure audit report on over 21000 patients hospitalised with heart failure1 has not been lost on either the tabloids or broadsheets,2 3 where cries of ‘unnecessary deaths’ and a ‘lottery’ of care are not without justification. What we learn from the more measured terms of Cleland et al4 (see page 876), based on the 2008/09 audit, and data from just over six thousand patients hospitalised with decompensated heart failure, is a not dissimilar and equally salutary message. The overall mortality during an index admission with heart failure is 12%, but whereas in those patients who are looked after on a specialist cardiovascular ward this figure drops, it increases for those whose care continues on general medical wards. This translates into a twofold increase in the risk of death for those whose care does not include management on cardiovascular wards. Even when adjustments were made for age and other predictors of poor outcome this relationship held.

The impact of this index inpatient care also determined longer term outcomes: patients who went home from general medical wards being less likely to survive than those who had been cared for on specialised cardiovascular wards, with a median follow-up of 158 days. With the modelling of 12 month outcomes, where adjustments were made for identified predictors of outcome, the relationship between index ward of care and likelihood of death persisted. This influence only disappeared when the model adjusted for the influence of follow-up by cardiology/heart failure services.

Age is identified as a powerful predictor of mortality during the index admission, so it is of interest that patients managed on cardiovascular wards tended to be younger and more often male. They were also more likely to have an ejection fraction below 40% which did not predict inpatient mortality. But the elderly, who are often among the most complex patients and at greatest risk, appear to have been deprived of specialist cardiovascular and heart failure care during the index admission, with inpatient mortality rates of 15% and above. These patients were then less likely to have had a robust diagnosis (only 65% had undergone echocardiography compared with 88% on cardiology wards), were less likely to be discharged on angiotensin converting enzyme inhibitors (ACEI), β-blockers or aldosterone antagonists, and were less likely to have cardiovascular/specialist heart failure follow-up.

The model, adopted by the authors to explore longer term outcomes, and which excluded the ward of care, identified important predictors of an adverse outcome as greater age, severity of oedema, lack of use of ACEI (or angiotensin-II receptor blockers where ACEI intolerant) and lack of use of β-blockers. The paper describes an expected 12 month mortality of 50% in those aged above 75 as compared with a value of 25% for those aged below 75. Both figures are excessive, and both figures increased when patients were managed on general medical wards. The 2008/09 audit report has already shown the beneficial impact of prescribing these drugs during the index admission on mortality over the ensuing 12 months—the addition of each of these drugs confers substantial reduction in the risk of death.5

So why do we see this selection bias towards younger and male patients? The paper does not speculate on this but it is an observation with resonance elsewhere. The recently published report from the British Cardiovascular Society (BCS) on Access to Cardiac Care in the UK6 did not initially address this issue but further analyses suggest that there may be an under-representation of women in the older age group (BCS, private communication, with permission).

Why has the cardiovascular community allowed this situation to develop, with only about 50% of patients admitted to hospital getting access to appropriate cardiovascular care? The comparator of recent guidance for the management of people admitted with a non-ST elevation myocardial infarction,7 which suggests that patients with a calculated 6 month mortality risk of 3% or above should be offered coronary angiography (with follow on per-cutaneous intervention (PCI) where indicated) within 96 h of first admission, serves to emphasise the need to fundamentally reorganise the care of the vulnerable heart failure population within the acute hospital setting. The trigger to the care of patients with acute coronary syndrome is often an elevated troponin, and the suggestion from Cleland et al4 that natriuretic peptides should be readily available to facilitate earlier diagnosis is well made, though will require sensible guidance alongside its introduction within the hospital sector. This strategy alone will not drive improved care; a radical rethinking in our approach to heart failure services is also required. We have long known of the vulnerability of this patient group, but research has for the most part addressed strategies for reducing readmissions and does not describe inpatient care.8 9 Only two studies have prespecified inpatient optimisation as part of a protocol, and both report low mortality rates at follow-up.10 11 Elsewhere inferred high quality baseline inpatient care also translates into improved outcomes.12 Improved baseline care may also be a contributing factor in the reduced mortalities reported at 1, 3 and 6 months by Mehta et al,13 for a 2004–05 cohort of patients with newly diagnosed heart failure, when compared with outcomes for a similar cohort from the earlier London heart failure study of incident heart failure for the period 1995–97. Nonetheless, the more usual absence of comprehensive good inpatient care during an index heart failure admission, and any readmission, has resulted in the audit data now presented in this paper by Cleland et al.4 It is of note that the inpatient deaths did not occur early, but after a median of 11 days, and there was little evidence of palliative care—both suggest missed opportunities to transform outcomes.

To date there has been no NICE guidance on acute hospital care for patients admitted with heart failure, but recognising the importance of these acute episodes, the 2010 NICE guidance on chronic heart failure14 now advocates that the heart failure specialist lead should be involved in the care of all patients admitted to hospital with heart failure. The heart failure specialist lead will usually (though not invariably) be a consultant cardiologist with an interest in heart failure and responsible for a multidisciplinary team working across primary and secondary care, and working with heart failure nurse specialists, general practitioners (both with and without special expertise in heart failure), consultants and other colleagues from diverse disciplines, including palliative care, rehabilitation and others. If commissioners and providers can be persuaded to implement this guidance as intended, it will translate into improved hospital care with better outcomes, and ensure continuity of care across primary and secondary care, and more effective resource utilisation. But a more sophisticated approach to the needs of this patient group will be necessary, including the recognition that a sometimes longer index admission, with high quality care, will reduce mortality rates and the likelihood of early heart failure related readmissions.

Many hospitals will need to identify and/or appoint the consultants to develop and lead these services, while ensuring adequate resource to support them. As early as 2005, the British Cardiovascular Society with the British Society for Heart Failure, recognised the need for increased provision for cardiovascular care at consultant level within heart failure, with a suggested workforce of between 5.9 and 8.7 full time equivalent consultant cardiologists with expertise in heart failure, per million population.15 These estimates may now seem rather conservative given the complexities of some of the inpatient care and increasing need to deliver more comprehensively across primary and secondary care. Yet these imperatives may also make such appointments achievable at a time when expansion of the consultant grade is likely to become more restrained than it has been in recent years. There is a growing body of enthusiastic specialist registrars selecting heart failure as their chosen area in which to undergo advanced training, often alongside imaging, or complex pacing and device implantation. The UK is unusual, and arguably forward thinking, in recognising heart failure as one of five areas of advanced training within cardiovascular medicine, and in this respect is unlike its European counterparts. Further support for this discipline will likely come with revalidation requirements.

Recognising the impact of coronary care units on outcomes, Cleland et al propose the development of heart failure units. This merits further exploration, but will need adequate capacity to ensure all patients have access to specialist care rather than a select subset. National heart failure audit data from the Whittington Trust may be relevant in this context. Hospital-wide guidance and provision of a very proactive heart failure service involving both shared care with general physicians, and care of the most complex patients on a cardiovascular ward, has been delivered since 2002. We aim to identify all heart failure patients and undertake echocardiography within 48 h of admission where heart failure is suspected, but have not had access to inpatient natriuretic peptide use. The 2009/10 audit suggests we are identifying 65% of a possible heart failure population of the 251 patients given a hospital episode statistics (HES) discharge code of heart failure. Inpatient care includes optimising diuretics, ACEI and aldosterone antagonists, initiating β-blockers, stability prior to discharge, with early cardiology, and/or heart failure nurse specialist, follow-up to ensure completion of β-blocker up-titration and to assess the need for additional interventions. Of note, the main referring primary care trust has a year on year excess, though falling, cardiovascular mortality when comparators include local, regional and national populations. Interrogation of the national heart failure database, for the Whittington Trust, suggests that this model of care results in an inpatient mortality of less than 4.3%. If we accept incomplete data entry (using HES figures as the comparator), the inpatient mortality rises to 7.6% for the HES heart failure coded cohort, which compares with 12.2% for the London strategic health authority and 14.6% nationally. This suggests that it is realistic to improve outcomes for heart failure patients by ensuring better access to and delivery of specialist inpatient heart failure care, as advocated by Cleland et al, and emphasises another role of the audit, namely as an invaluable tool by which individual clinicians and acute trusts can monitor, understand, modify and demonstrate improvement in their patient care.

So what can we learn from this notable paper on the national heart failure audit from Cleland et al? That we need to transform our hospital services, and that it is essential to identify all people admitted to hospital with heart failure, with the urgency that has come to be associated with acute ischaemic episodes, optimise their care and treatment, and thereafter ensure continued access to specialist heart failure care. An ongoing ‘lottery of care’ is unnecessary and unacceptable. If the same energy is now given to the quality of inpatient care and reducing mortality that has hitherto focused on length of stay, this will be readily achieved, and improved bed usage will likely result.