Heart failure is a common reason for urgent admission to hospital in the NHS and a major cause of morbidity and mortality in the UK, particularly in those over 60 years old.1^–4 The problem is likely to increase as a result of the increasing proportion of the population aged >60 years, improvements in treatment for and survival from coronary heart disease (CHD), hypertension and, paradoxically, heart failure itself.5 6

In recognition of the importance of cardiovascular disease to the UK health services the National Service Framework (NSF) for CHD was launched in England and Wales in 2000.7 The goal of this 10-year agenda was to reduce CHD and stroke-related deaths by 40% by 2010. This was followed by the National Institute for Health and Clinical Excellence (NICE) guidance on heart failure in July 2003.8 However, implementation of the NSF and NICE guidance was slow to materialise,9 with noticeable variations in diagnosis, treatment and care despite specific guidance.10

The Healthcare Commission (HC) acts as the independent NHS regulator and also as the home of the national clinical audit and patient outcome programme and acute hospital portfolio team. It has responsibility to review independently the provision of care within the NHS in England and Northern Ireland.

In 2005 the HC audited the management of emergency admissions in acute NHS trusts in England, Wales and Northern Ireland. Heart failure, as a common cause of acute emergency admissions, was used as a marker diagnosis to help understand the differences in management practice of all acute admissions between trusts and their impact on outcome for patients.

It provided an opportunity to obtain national data on acute heart failure demographics, investigation, treatment and short-term outcome. It also contributed to the HCs’ wider “improvement review” of heart failure,11 which examined the full patient pathway for people with heart failure.

METHODS

In conjunction with the British Society for Heart Failure a survey was conducted on patients admitted as an emergency with heart failure (discharge code I50). Acute trusts (defined as those with an emergency department or facility to accept emergency patients from the community) in England, Wales and Northern Ireland were asked to review the notes of patients discharged with a primary diagnosis of heart failure, starting on 30 September 2005, and retrospectively audit the previous 50 sets of eligible patient notes. These data were collated onto an online audit form and returned to the HC. It included information on diagnosis, admission, acute management, inpatient management and discharge.

Diagnosis

1. Primary discharge diagnosis of heart failure (code I50).

2. Previously documented diagnosis of heart failure.

Admission

1. Point of admission (Emergency Department (ED), Medical Admissions or Assessment unit (MAU), direct inpatient).

2. Clinical condition on admission (New York Heart Association (NYHA) classification, symptom-based assessment and cardiac rhythm).

3. Pharmacological treatment for heart failure on admission (tables 2 and 3).

4. Investigations requested and completed during the admission.

Acute management

1. Transfer to MAU/inpatient ward.

2. Change in medical team (and if specialist).

3. Further acute investigations requested and completed (on first inpatient ward).

Inpatient management

1. Further inpatient transfer (including ward type).

2. Further change in consultant (including specialist referral).

Discharge

1. Time to becoming “medically fit” for discharge, discharged or death (length of stay/delayed discharge/mortality data).

2. Clinical condition ((NYHA classification, symptom-based assessment and cardiac rhythm).

3. Pharmacological treatment for heart failure on discharge.

4. Completed investigations.

5. Follow-up arranged (rehabilitation, specialist palliative care, specialist heart failure review).

Statistical analysis

Continuous variables are summarised using median and interquartile range, except age which is presented as mean (SD). Comparisons of these variables were carried out using the Mann–Whitney test, the Kruskal–Wallis test or Student t test as appropriate. Categorical variable are presented as number (percentage) and compared using the χ² test. Percentages are based on all patients with information available for that variable, not on the entire survey population. Logistic regression was used to test whether age and gender were associated with the likelihood of patients dying or receiving medication.

RESULTS

One hundred and seventy-six acute trusts responded (of a total of 177) resulting in a cohort of 9387 patient records (median (interquartile range) 50 (50 to 51) per trust). Ninety-three per cent of patients were admitted in the 6-month period to 30 September 2005. In view of the large dataset, values of p<0.001 should be regarded as significant and p<0.05 regarded as trends, rather than true significance.

Characteristics and cardiovascular status at the time of hospital admission

Table 1 shows the overall results. Women constituted 50% of admissions and were older (80 years vs 75 years) and less likely to have had echocardiography before their index admission (52% vs 60% (p<0.001)).

Drugs prescribed for heart failure before admission

For simplicity drugs were classified into the classes of diuretics, ACE inhibitors (ACE-I) or angiotensin II receptor antagonists (AIIRA), β-blockers and aldosterone antagonists (tables 2–4).

Of those with a previous diagnosis of heart failure (ie, previous echo diagnosis before admission) men were less likely to be treated with a diuretic alone (odds ratio (OR) = 0.69 (95% CI 0.59 to 0.81), p<0.001, age adjusted OR = 0.81 (0.69 to 0.96), p = 0.013) and more likely to be treated with AIIRA or ACE-I than women (OR = 1.44 (95% CI 1.28 to 1.62), p<0.001, age adjusted OR = 1.29 (1.14 to 1.45), p<0.001).

Patients receiving β-blockers before admission, either in isolation or as part of a combination of treatments, were younger than those not receiving them by an average of 1.9 years (95% CI 1.4 to 2.4), p<0.001). Men were more likely to be receiving β-blockers on admission than women (OR = 1.22, (95% CI 1.08 to 1.39), p<0.001, age adjusted OR = 1.15 (1.01 to 1.29), p = 0.033).

Decision to admit to inpatient care

Of those patients assessed in the ED (and not assessed by a GP before admission) the decision to admit was divided equally between the ED staff (46%) and on-call medical team (45%). A cardiologist or lead consultant in heart failure took the decision in only 5% of cases, with a further 2% admitted directly by elderly care teams. Most patients (74%) were admitted under the care of the on-call medical team, with only 13% coming under the direct care of a cardiologist or the lead heart failure clinician. Some 9% of patients were admitted directly to the elderly care doctors; these patients were substantially older (7 (2) years, p<0.001) than patients admitted by other teams.

Further inpatient management and length of stay

Thirty-two per cent of all admissions were assessed by echocardiography at some point during their inpatient admission. Of these patients, half had had no previous echocardiography. Thus 68% of patients did not have an echocardiogram during or before admission. Older patients were less likely to receive an echocardiogram (OR = 0.98 (95% CI 0.97 to 0.99), p<0.001).

Of the patients who received an echocardiogram during their admission, 44% were imaged within the first 24 hours of admission, rising to over three-quarters within 72 hours and 90% with 5 days. There were no differences in time to echocardiography or likelihood of echocardiography based on gender.

Of the patients transferred from the MAU to inpatient wards, over 50% changed consultant at least once but only 14% of all admissions were referred to a specialist (cardiologist or heart failure lead clinician). There was no difference in the severity scores of patients who changed consultant in comparison with those who had a single consultant episode; however, the severity scores of those referred to specialists were significantly higher (p<0.001).

Transfer to a second non-specialist consultant added 2 days to the admission time, a third transfer added an additional 4 days and while only 2% of admissions were managed in a non-medical wards this did not significantly extend length of stay.

Table 5 summarises the factors that were related to length of stay.

Drugs prescribed for heart failure on discharge

At the time of discharge, men were still less likely to be treated with a diuretic in isolation (OR = 0.64 (95% CI 0.56 to 0.74), p<0.001, age adjusted OR = 0.79 (0.69 to 0.91), p = 0.001), were more likely to be treated with AIIRA or ACE-I than women (OR = 1.39 (95% CI 1.26 to 1.53), p<0.001, age adjusted OR = 1.21 (1.09 to 1.33), p<0.001) and were more likely to be treated with all four pharmacological agents (OR = 1.87 (95% CI 1.58 to 2.0), p<0.001, age adjusted OR = 1.54 (1.30 to 1.83), p<0.001) (table 4).

Deaths and discharge

The in-hospital mortality rate was 15%. After adjusting for age, men had a higher mortality rate (OR = 1.13 (95% CI 1.00 to 1.27), p = 0.042), with a 6% increase of death per year (OR = 1.06 (95% CI 1.05 to 1.06), p<0.001). Mortality was <3% in patients aged <55 years rising to 23% in those >85 years at presentation. As expected a higher initial NYHA classification (OR = 1.60 (95% CI 1.47 to 1.75), p<0.001) or symptom severity score (OR = 1.49 (95% CI 1.38 to 1.61), p<0.001) was associated with a greater likelihood of death. This was independent of age.

Although many patients were aged >80 years, only 7% of patients experienced delays to their discharge from the time they were declared medically fit, with 99% being discharged within 10 days, probably reflecting the time required to set up, or re-establish, the social support packages often required by this aged population. Eighty-five per cent of all patients were discharged home, 8% to care homes (residential or nursing) and the remainder to hospices or other care providers. Only one-fifth of all patients discharged were referred to follow-up heart failure services; however, <1% were referred for either rehabilitation or palliative care services.

DISCUSSION

The implementation of the NSF and NICE guidance for heart failure has been slower than other standards within the CHD NSF and significant geographical and gender variation exists in standards of diagnosis, treatment and care of this progressive and incurable condition.9 This may reflect the lack of clearly defined service for patients with this common and deadly diagnosis. Whereas myocardial infarction, angina and arrhythmia services have clearly defined targets and have been the focus for sustained investment, heart failure still appears to be regarded as a Cinderella subspecialty despite its ubiquitous nature, especially among the elderly population. It should be acknowledged that this study is predicated by correct clinical coding, but with detailed information on nearly 10 000 patients admitted with acute heart failure an accurate picture of current demographics, investigation, management and treatment of this population is possible (with significant protection from random error afforded by the large sample size). Furthermore, when taken with previous publications, such as the IMPROVEMENT-HF survey,12 EuroHeart Failure survey5 and British Heart Foundation (BHF) statistics on heart failure management,2 an assessment of progress, or lack thereof, against these current standards and guidance can be made.

Over half the patients admitted with acute heart failure had previously had an echocardiogram. This is likely to represent an overall improvement in community heart failure diagnosis, driven in part by the increasing availability of community or rapid access echocardiography clinics and likely to improve further over time as a result of the GP quality outcome framework. The survey did not assess echocardiographic findings and so it is not possible to correlate left ventricular systolic dysfunction with either diagnosis or pharmacological intervention.

The median waiting time in the emergency department falls well within the desired government 4-hour target, with one-third of patients being admitted directly to MAUs or inpatient wards. As expected, the majority of patients had evidence of pulmonary oedema on admission, but a significant minority did not, and these may represent a combination of dependent oedema and dyspnoea alone without evidence of pulmonary involvement. This may explain the large number (25%) of patients classified as NYHA I and II on admission and may represent a population that might reasonably be targeted by the proposed expansion of community-based specialist services favoured by the Department of Health in their drive to reduce admissions and use secondary care services more efficiently.

This survey suggests a substantial improvement in the use of β-blockers (28% vs 10% in the most recent BHF statistics2) and moderate improvement in the use of ACE inhibitors from previous surveys (67% vs 50%2). However, the use of these evidence-based drugs remains suboptimal, with the use of β-blockers, in particular, remaining very low despite evidence of their benefit in this patient cohort. Diuretic therapy was almost universal and the use of aldosterone antagonists shows signs of rapid uptake, being prescribed to over a quarter of patients discharged.

Women and older patients are less likely to be investigated by echocardiography and less likely to be prescribed both ACE inhibitors and β-blockers in line with previous European data.13 The lower rate of treatment in women may be a direct result of their lack of echocardiography and subsequent lack of confidence that they indeed do have systolic dysfunction in which ACE inhibitors and β-blockers have been proved to be of value.

The use of echocardiography also remains very low, despite the NSF recommendation that it should be offered to all patients. This may be because of the lack of capacity to perform echocardiography in many centres.

The overall length of stay for heart failure is similar to that reported in other recent surveys.5 Patients receiving care from cardiologists and heart failure lead clinicians did not have a reduced length of stay but were more unwell than those receiving care from general doctors. Less than 20% of patients were followed up after discharge by a specialist heart failure service—despite the evidence that this improves outcome and healthcare utilisation.14 This lack of follow-up does not reflect well on the provision of heart failure services and may ultimately increase the risk of readmission. A multidisciplinary heart failure team that took responsibility on the day of admission in conjunction with the admitting doctor might allow a faster discharge, more comprehensive heart failure management (in particular, echocardiography and systolic heart failure treatment) and increased use of appropriate follow-up.

CONCLUSION

Many patients admitted to acute hospitals in England, Wales and Northern Ireland in 2006 are not being investigated or treated fully in accordance with international evidence-based guidelines. The use of echocardiography for heart failure diagnosis and evidence-based medication for systolic heart failure remains suboptimal. Furthermore, women seem to be less well managed against recommended guidelines than men, perhaps partially owing to their increased mean age and increased likelihood of heart failure with normal systolic function. Overall mortality remains high for acute heart failure. In comparison with earlier studies from the UK, the use of echocardiography, ACE-I and β-blockers has increased, and the length of stay has reduced (but with some significant delayed discharges for non-medical reasons). Only a minority of patients are seen or followed up by a specialist service. Overall progress towards the implementation of NSF and NICE guidelines seems to have been made in the past couple of years, but significant and sustained effort is required to deal with variation in practice and gender inequalities in the provision of heart failure care.

Not required for this survey using anonymous data.