Effects of a nurse-led, clinic and home-based intervention on recurrent hospital use in chronic heart failure

Authors


Abstract

Background:

Few studies have examined the potential benefits of specialist nurse-led programs of care involving home and clinic-based follow-up to optimise the post-discharge management of chronic heart failure (CHF).

Objective:

To determine the effectiveness of a hybrid program of clinic plus home-based intervention (C+HBI) in reducing recurrent hospitalisation in CHF patients.

Methods:

CHF patients with evidence of left ventricular systolic dysfunction admitted to two hospitals in Northern England were assigned to a C+HBI lasting 6 months post-discharge (n=58) or to usual, post-discharge care (UC: n=48) via a cluster randomization protocol. The co-primary endpoints were death or unplanned readmission (event-free survival) and rate of recurrent, all-cause readmission within 6 months of hospital discharge.

Results:

During study follow-up, more UC patients had an unplanned readmission for any cause (44% vs. 22%: P=0.019, OR 1.95 95% CI 1.10–3.48) whilst 7 (15%) versus 5 (9%) UC and C+HBI patients, respectively, died (P=NS). Overall, 15 (26%) C+HBI versus 21 (44%) UC patients experienced a primary endpoint. C+HBI was associated with a non-significant, 45% reduction in the risk of death or readmission when adjusting for potential confounders (RR 0.55, 95% CI 0.28–1.08: P=0.08). Overall, C+HBI patients accumulated significantly fewer unplanned readmissions (15 vs. 45: P<0.01) and days of recurrent hospital stay (108 vs. 459 days: P<0.01). C+HBI was also associated with greater uptake of beta-blocker therapy (56% vs. 18%: P<0.001) and adherence to Na restrictions (P<0.05) during 6-month follow-up.

Conclusion:

This is the first randomised study to specifically examine the impact of a hybrid, C+HBI program of care on hospital utilisation in patients with CHF. Its beneficial effects on recurrent readmission and event-free survival are consistent with those applying either a home or clinic-based approach.

1. Introduction

Chronic heart failure (CHF) is a costly [1], debilitating [2 [3] 4] and deadly [5 [6] 7] syndrome that has resulted in an epidemic of old and fragile patients in whom treatment options are limited [4,8] and recurrent hospitalisation is common [9 [10] 11]. Given that this epidemic is likely to be sustained [12] there is an increasing imperative to develop and apply cost-effective programs of care to improve health outcomes in such patients.

It is within this context that predominantly nurse-led programs of care have been successfully applied to improve health outcomes in patients hospitalised with CHF. Meta-analyses of randomised studies of post-discharge programs designed to reduce recurrent hospital-use and prolong survival in patients with CHF have confirmed their benefits in this regard [13 [14] 15]. Pending more appropriately powered randomised studies of telemonitoring management strategies in CHF [16], it is clear that managing CHF patients via a nurse-led specialist outpatient clinic [17 [18] 19] or a nurse-mediated, multidisciplinary, community based program of care involving at least one home visit [20 [21] [22] [23] 24] provides the greatest benefits [14]. However, when applying such programs of care it is often difficult to follow the precise protocols applied in these research studies and many programs apply a combination of clinic and home-based follow-up.

2. Study hypothesis

Although there is obvious potential for a positive synergistic effect on health outcomes by gaining a comprehensive understanding of each patient's clinical and psycho-social status during a home visit and using clinic-based follow-up to optimise treatment thereafter [25], studies of this type of hybrid approach have been limited. The most notable exception is the brief intervention applied by Rich et al. [23] in their seminal study published in 1995. However, the major focus of that intervention was community-based management and, since publication, a number of models of care, particularly clinic and home-based interventions, have emerged. It is within this context that this study specifically examined the following null hypothesis: A nurse-led, clinic and home-based intervention (C+HBI) relative to usual post-discharge care (UC) will have no-effect on event-free survival and rate of all-cause recurrent hospital stay in patients with CHF discharged from acute hospital care.

3. Methods

3.1. Study population

Patients were recruited from both York District Hospital and Scunthorpe General Hospital in the North of England. Prior to patient recruitment the study was approved by the institution's ethics of human research committee subject to changes in study randomisation (see below) in addition to the National Health Service (NHS) Research Governance Approval. All patients signed a consent form prior to study randomisation. As such, the study conformed to the principles outlined in the Declaration of Helsinki.

3.2. Patient recruitment

Potential study patients were identified via a combination of inpatient screening and study referral. All patients recruited to this study fulfilled the following criteria:

  • Acute admission to hospital with a diagnosis of CHF.
  • Objective evidence (e.g. echocardiography or coronary angiography) of impaired left ventricular systolic function as evidenced by a left ventricular ejection fraction (LVEF) of ≤45% immediately prior to study recruitment.
  • Discharged to home.

Patients awaiting an elective cardiac procedure with the intent to reverse the cause of underlying heart failure (e.g. coronary artery bypass surgery for coronary artery stenosis) or with a terminal illness other than CHF were excluded from the study.

During the study recruitment period of 20 months, approximately 5746 patients were screened. It is important to note that despite NHS directives [26], obtaining objective evidence of left ventricular systolic dysfunction was problematic and approximately 642 otherwise eligible patients (12%) were lost to the study due to the lack of determination of underlying cardiac function. Of the 119 (2%) patients' who were eligible for active study follow-up, six died prior to discharge home and seven refused to participate. Therefore, a total of 106 patients were subject to study randomisation and follow-up.

3.3. Study intervention

The local ethics committee specified that a cluster randomisation approach be applied to avoid the confounding influence of General Practitioners (GP's) managing patients in both arms of the study. As such the Primary Care Trusts that serve the York and Scunthorpe areas were contacted and a list of all GP clinics in these areas was obtained. These clinics were then divided into two groups according to number of GP's in each clinic (a median figure was used to split the group). Then a random number allocation was used to allocate equal numbers of small and large clinics to either post discharge HBI+C or UC. Affected GP's were contacted each time a patient in their clinic was recruited to the study and informed that additional information might be forthcoming as the result of active intervention as part of the study.

Those patients allocated to the UC group (n=48) received standard care (i.e. explanation of their condition and prescribed medications by the ward nurse and referral to appropriate post-discharge support as required). Patients were given an outpatient department appointment 6–8 weeks post discharge according to individual consultant practice.

The study intervention was primarily applied by two nurses experienced in the management of heart failure and with postgraduate qualifications, using strict treatment protocols. As such, patients allocated to HBI+C (n=58) were seen by the study specialist nurses prior to their discharge and 57 (98%) received a home visit within 10 days of hospital discharge.

At the ward visit, patients received information on their condition, medications and what to expect when they went home. An appointment was arranged at a time of the patient's convenience for the home visit. In addition, the patient was given a card with a contact telephone number for the specialist nurse (office hours only) where messages could be left regarding any questions or queries the patient or their family may have. At the home visit, patients (and their family if appropriate) received education on their condition, including symptom recognition, symptom management and lifestyle issues. A thorough clinical examination, involving a review of their clinical history since hospital discharge, functional status, vital observations, heart and lung sounds and hydration status was also performed. All patients accepted an invitation to attend a monthly nurse-led outpatient heart failure clinic for at least 6 months post discharge. However, concomitant illness (e.g. influenza), early post discharge death and the logistics of travelling to the clinic meant that only 46 patients (79%) attended all their clinic sessions. During the clinic visits patients (and their family if appropriate) received a full educational package, including diagnosis, symptom recognition, symptom management, lifestyle issues and crisis management. A clinical examination (see above) was performed to assess cardiac status and fluid retention. Weight, electrolytes, renal, hepatic function and blood counts were monitored. New therapeutic agents (e.g. Spironlactone) were also commenced according to local protocol based upon the Taskforce Recommendations of the European Society of Cardiology (2000) [27] and the Scottish Inter Collegiate Guidelines (1999) [28]. Where appropriate, referral to other local health and social services was made.

3.4. Study follow-up and endpoints

Baseline clinical and demographic data were collated during the index admission. Data on recurrent hospital stay and/or death were also collated (in a blinded manner) via the local area hospital record system and death registry. A number of self-administered questionnaires were sent via mail (reminders at 14 days) to determine the impact of the study intervention relative to UC. For example, additional data on treatment adherence during study follow-up was collected using the Hill-Bone compliance scale, which comprises 23-items profiling 4-domains (diuretic medication, ACE medication, sodium reduction and appointment keeping) [29]. These data were collected at 1 (n=50) and 6 (n=46) months. Health-related quality of life (using both SF-36 [30] and the Minnesota Living with Heart Failure [31] scales) was also measured at 1 (n=46) and 6 months (n=41) post discharge (more detail on these scales are provided in the legends of Figs. Fig. 3 and Fig. 4). A survey of GP clinics and hospital records was used to determine the proportion of surviving patients prescribed an angiotensin converting enzyme (ACE) inhibitor, an angiotensin receptor blocker (ARB), loop diuretic, beta-blocker and spironolactone at the completion of 6 months follow-up.

The co-primary end-points for this study were event-free survival from either death or recurrent hospitalisation for any reason during the 6-month follow-up. Secondary endpoints included rate of recurrent hospital stay, treatment adherence and health related quality of life.

3.5. Statistical analysis

Based on an anticipated primary event rate of 60%, it was calculated that a total of 60 patients in each group were required to detect a 40% variance in the primary endpoint of death or readmission during 6 months follow-up when assuming a two-sided α of 0.05 and a β of 0.2.

Comparison of baseline and endpoint data involved χ2 analysis (with calculation of odds ratio [OR] and 95% confidence intervals [CI] where appropriate) for discrete variables, Student's t-test for normally distributed continuous variables and Mann-Whitney test for non-normally distributed variables. To adjust for differences in survival and duration of follow-up, event frequency was calculated as a mean number of events/patient/month.

Kaplan-Meier survival curves were constructed using time-dependent, all-cause survival and event-free survival data followed by analysis with both the log-rank test and the Breslow test to determine any differences in the number and/or timing of events. Data for all surviving patients were censored at 6 months following discharge from the index hospital admission.

To examine the independent effects of treatment mode and more than 40 baseline demographic and clinical variables (see Table 1) on event-free survival and all-cause mortality we used multivariate logistic regression (with initial entry and step-wise rejection of baseline variables at the 0.1 and 0.05 levels of significance, respectively). All analyses were performed on an intention-to-treat basis according to study group assignment using SPSS for Windows (11.0).

Table 1. Baseline characteristics
 HBI+C (n=58)UC (n=48)
Demographic profile  
Male (%)42 (72%)35 (73%)
Age in years73±1472±12
Live alone16 (28%)13 (27%)
Heart failure profile  
Duration of treatment for CHF (months)25±4338±42
Prior admissions for acute heart failure0.4±0.80.8±0.9
Left ventricular ejection fraction31±829±11
NYHA Class III/IV at hospital discharge44 (76%)35 (73%)
Co-morbidity  
Mean Charlson Index of Comorbidity score2.5±1.42.5±1.3
Prior acute myocardial infarction27 (47%)26 (54%)
Chronic airways limitation15 (26%)10 (21%)
Atrial fibrillation16 (28%)15 (31%)
Non-insulin/Insulin dependent diabetes8 (14%)13 (27%)
Blood profile at hospital discharge  
Sodium mmol/L135±3.8138±3.9
Potassium4.0±0.54.1±0.6
Creatinine mmol/L133±55147±85
Albumin g/L37.8±3.738.7±5.9
Haemaglobin g/dL12.3±2.511.4±3.4
Vital signs at hospital discharge  
Heart rate (beats/min)75±1579±16
Systolic blood pressure (mm Hg)114±18118±22
Diastolic blood pressure (mm Hg)67±866±11
Pharmacotherapy at hospital discharge  
Diuretic50 (86%)38 (80%)
ACE inhibitor46 (79%)28 (59%)
β-adrenoceptor blocker15 (25%)5 (10%)
Digoxin17 (29%)14 (29%)
Nitrate23 (40%)18 (38%)
Aspirin32 (55%)21 (44%)
Warfarin20 (35%)11 (23%)
Amiodarone9 (16%)9 (19%)

4. Results

4.1. Baseline characteristics

Table 1 presents the baseline characteristics of the 106 patients who were recruited to this study. As expected, given the requirement for a confirmed diagnosis of left ventricular systolic dysfunction, the majority of patients were male (72%) with an average age of 72 years. As the study involved cluster randomisation according to the treating GP, potential differences in terms of study allocation (58 vs. 48 randomised to HBI+C and UC, respectively) and baseline characteristics (HBI+C patients had fewer prior admissions for CHF and were more likely to be prescribed an ACE inhibitor) were anticipated. Overall, this was a typically old patient cohort with multiple concurrent conditions likely to complicate optimal CHF management.

4.2. Treatment profile and adherence

At 6 months, surviving patients in the HBI+C were significantly more likely to be prescribed a beta-blocker (56% vs. 18%: P<0.001) with 15 compared to 1 additional UC patient commencing such treatment during study follow-up. Whilst similar proportions of surviving patients were prescribed a loop-diuretic or ACE inhibitor at six months, more HBI+C versus UC patients also commenced adjunctive spironolactone and/or ARB therapy (16 vs. 8, respectively) although, due to the small numbers of patients, this did not reach significance.

At 1 (n=50) and 6 months (n=46) there were few differences in self-reported treatment adherence rates according to the Hill-Bone instrument. Patients assigned to the HBI+C were more likely to adhere to a prescribed Na restricted diet as determined by the Hill-Bone adherence score (mean 8.9±2.3 vs. 7.3±1.9: P=0.017) but equivalent scores were similar at 6 months. Overall, self-reported Hill-Bone adherence scores remained consistent over this time period for individual patients.

4.3. Event-free survival

During 6 month follow-up, a total of 5 (9%) vs. 7 (15%) of HBI+C and UC patients, respectively, were reported to have died, this difference did not reach statistical significance. Patients assigned to the UC group were more likely to be readmitted to hospital for any reason (21 of 48 (44%) vs. 13 of 58 (22%) in the HBI+C group: P=0.019, OR 1.95 95% CI 1.10-3.48). Overall, 15 (26%) HBI+C patients compared to 21 (44%) UC patients experienced a primary endpoint (death or readmission) during 6 months follow-up.

fig. Fig. 1 shows the event-free survival curves for the two groups with a strong trend favouring those assigned to HBI+C without reaching statistical significance (P=0.08). Cox-proportional modelling demonstrated that HBI+C was associated with an approximate 45% reduction in the risk of death or readmission during study follow-up when adjusting for all other clinical and demographic variables-although this association did not reach statistical significance (RR 0.55, 95% CI 0.28–1.08: P=0.08). Significant predictors for death or readmission were advancing age (RR 1.01, 95% 1.00–1.02 per additional year) and extent of co-morbidity as measured by the Charlson Index [32] (RR 1.15, 95% CI 1.01–1.46 per additional point score indicative of a greater co-morbid burden).

Figure Fig. 1.

Figure Fig. 1. Event-free survival.

4.4. Recurrent hospitalisation

Overall, the 13 readmitted patients assigned to HBI+C accumulated a total of 15 hospital readmissions associated with 108 days of recurrent hospital stay during 6 months follow-up. The 21 readmitted patients assigned to the UC group accumulated significantly more readmissions and associated days of recurrent hospital stay (45 and 459, respectively): P<0.01 for both comparisons when adjusting for the number of events per patient per month of follow-up. fig. Fig. 2 shows that on this adjusted basis, HBI+C was associated with an approximate 60–70% reduction in recurrent admissions and associated hospital stay.

Figure Fig. 2.

Figure Fig. 2. Comparison of recurrent hospital admission and associated stay.

4.5. Health-related quality of life

fig. Fig. 3 compares SF-36 scores at one and six months (change in mean scores) according to study assignment in the 46 patients who returned completed questionnaires. Although no statistically significant differences were detected between groups (using a Bonferroni correction for multiple comparisons significance was set at P<0.001) at either 1 or 6 months, there was a consistent improvement (as denoted by positive changes in mean scores) in each health domain measured by the SF-36 excepting “role emotional” in the HBI+C group. In the UC group such a trend was not so apparent with opposite trends (i.e. worse compared to better scores) in relation to physical and social functioning, role physical and to a lesser extent mental health.

Figure Fig. 3.

Figure Fig. 3. Comparison of 1 (n=46) and 6 month (n=41) SF-36 scores. The SF-36 measures eight dimensions of health (PF=physical functioning, RP=role functioning-physical, BP=bodily pain, GH=general health, V=vitality, SF=social functioning, RE=role functioning-emotional and MH=mental health). Scores range from 0 to 100%. Lower scores from the SF-36 indicate reduced quality of life and therefore positive changes denote improvement [29].

fig. Fig. 4 similarly compares total MLWHF and component physical and mental health scores according to study assignment at 1 and 6 months (change in mean scores). Overall patients in the HBI+C group had better quality of life scores, however, there were no statistically significant differences based on study assignment. As expected, there was an improvement in scores reported in those surviving to 6 months.

Figure Fig. 4.

Figure Fig. 4. Comparison of 1 (n=41) and 6 month (n=46) MLWHF scores. The MLWHF instrument consists of 21 questions with a score range of 0–105. Higher scores indicate reduced quality of life and therefore negative changes denote improvement. PH=physical health component score and MH=mental health component score [30].

5. Discussion

The level of evidence in favour of applying nurse-led programs of specialist CHF management involving either clinic or home-based, post-discharge follow-up is most probably so compelling [13 [14] 15] that undertaking further research involving randomising patients to a usual care group without specialist management should be considered unethical. Indeed, there is overwhelming evidence that applying these types of programs on a wide-scale basis is extremely cost-effective and should be available to all hospitalised patients with CHF [33]. However, there are a number of residual issues that need to be addressed in relation to these programs. Despite a number of limitations, this multicentre randomised controlled study undertaken in the North of England, examined the important issue of whether a “hybrid” program combining both clinic and home-based intervention will have a similar effect to those studies that have been predominantly polarised towards either a clinic [17 [18] [19] 20] or home-based [20 [21] [22] [23] 24] model of post-discharge management. Consistent with a recent meta-analysis of 29 randomised studies of this type, involving more than 4500 patients [14], the hybrid program of clinic and home-based follow-up was associated with a reduced adjusted risk (45% less without reaching statistical significance) of either death or unplanned readmission to hospital (all-cause in both cases) relative to usual care. Also, consistent with recent reports involving both home [24] and clinic [19] based programs applied in Scotland and Sweden, respectively, this program was associated with a substantive (60%) reduction in all-cause hospital readmission during 6 months. The beneficial effects of this and other interventions in CHF, hinge upon preventing recurrent and prolonged hospitalisations in a relatively small number of patients. Importantly, morbidity rates in the usual care group were similar to those reported over a similar period of follow-up in usual care cohorts enrolled in contemporary and equivalent trials in the USA [21,34] and Australia [35]. Alternatively, mortality rates in the study were relatively low compared to other studies, despite evidence that gold-standard therapies were not being applied at hospital discharge. Apart from the significant number of patients who died prior to hospital discharge and study follow-up, it is possible that the recruitment of relatively younger patients (average age 72 years), no doubt due to investigational bias and subsequent availability of a quantified left ventricular ejection fraction contributed to this. Moreover, it is possible that the local authority death registry did not capture fatal events in a timely manner. As a result of the lower than expected number of fatal events during study follow-up we were unable to properly examine the effect of the intervention on subsequent survival. Consistent with previous studies, the study intervention did not adversely affect health-related quality of life in surviving patients relative to usual care (in most cases there were non-significant improvements) and was associated with better application of gold-standard therapies (particularly beta-blockers). Overall, treatment adherence rates, as measured by the Hill-Bone compliance scale favoured the intervention group without reaching significance. This improvement in treatment, in addition to better treatments patterns at baseline, however, may have made a substantial contribution to the size of beneficial effect seen in relation to recurrent hospital stay.

Whilst it may be tempting to suggest that this type of hybrid program is likely to be more cost-effective than singular approaches (when adjusting for duration of follow-up the hybrid program reduced this important health outcome by 69% relative to usual care as opposed to 39–43% reductions observed in the Scottish [24] and Swedish studies [19]), there are insufficient data from other studies to support such a claim. However, it is clear that offering a flexible approach of at least one home visit, plus ongoing support via a CHF clinic is a feasible option in this type of patient cohort. Previous reports have highlighted the difficulty in offering a clinic-based service to old and fragile patients with CHF [36] and in this study, 21% of patients did not attend the specialist clinic. However, with the increasing need to individually titrate and monitor gold-standard CHF therapy, it is obvious that either this approach, or that adopted by the Glasgow Heart Failure Nurse Liaison Service [37] through multiple home visits be applied.

This study has a number of limitations that require comment. Firstly, due to a lack of available echocardiographic data, we were unable to recruit sufficient numbers of patients to adequately test our predefined hypotheses. The recruitment of a select group of patients (i.e. more indicative of patients recruited to large-scale clinical trials) clearly needs to be considered when examining the general applicability of our findings. This lack of available data is disturbing given the necessity to tailor treatment of the syndrome of CHF through accurate diagnosis and determination of underlying cause (i.e. impaired versus preserved left ventricular systolic function). The subsequent problem of Type II error in examining and interpreting all the pre-defined end-points is obvious and this problem was magnified in relation to mortality and quality of life data. Secondly, the imbalance between groups, both in terms of numbers, characteristics and baseline treatment was derived from the need to randomise patients according to their GP. The shortfall in participants overall, most probably exacerbated the difference between groups in this regard. To overcome this problem, we undertook careful analyses of major endpoints with adjustment for baseline differences to determine the probable underlying effect of the study intervention. It is important to note, also, that we did not differentiate endpoints on the basis of the presence/absence of CHF as a causative factor. Finally, as discussed, this patient cohort (most probably due to the widespread availability of echocardiographic data) was younger and contained a greater proportion of males than comparable studies. All of these issues need to be considered when interpreting these data.

Despite these limitations, however, this study provides evidence to suggest that a hybrid approach comprising both home and clinic-based follow-up of patients with CHF discharged from acute hospital care is equally as effective as those strategies that employ either a home or clinic-based model of care. For the present, these data provide support to the many clinicians who have adapted previously published reports of singular approaches to form such a hybrid approach. Compared to usual care, a hybrid approach clearly reduces the need for recurrent hospitalisation in the short to medium term. In the future, however, there is clearly scope to undertake the type of appropriately powered, “head-to-head” trial described by Jaarsma et al. [38] that will determine if one approach is ultimately more cost-effective than another.

Acknowledgements

SS is supported by the National Heart Foundation and the National Health and Medical Research Council of Australia. This study was supported by a research grant from Merck Pharmaceuticals UK.

Ancillary