Article Text

Download PDFPDF
Impacts of person-centred integrated chronic heart failure and palliative home care on pharmacological heart failure treatment: a substudy of a randomised trial
  1. Rickard Markgren1,
  2. Margareta Brännström2,
  3. Claes Lundgren3 and
  4. Kurt Boman4
  1. 1 Internal medicine clinic, St Göran hospital, Stockholm, Sweden. At the time of writing the article at Internal medicine clinic, Skellefteå hospital, Skellefteå, Sweden
  2. 2 Department of Nursing, Strategic Research Program in Health Care Sciences (SFO-V), “Bridging Research and Practice for Better Health”, Umeå University, Umeå, Sweden
  3. 3 Palliative Care Unit AHS, Skellefteå Hospital, Skellefteå, Sweden
  4. 4 Research unit, Department of Medicine, Skellefteå, Institution of Public Health and Clinical, Medicine, Umeå University, Umeå, Sweden
  1. Correspondence to Professor Kurt Boman, Department of Medicine, Research unit, Skellefteå Hospital, Skellefteå 93186, Sweden; kurt.boman{at}


Objective Patients with chronic heart failure (CHF) may be insufficiently treated pharmacologically. Recently, we presented a person-centred integrated Palliative advanced homecaRE and heart FailurE caRe (PREFER) strategy and compared it with usual care (control). Patients managed according to PREFER had improved health-related quality of life and markedly reduced hospitalisations compared with the control group. We hypothesised that these improvements may have been partly due to better drug treatments within the PREFER strategy. Thus, our aim in this study was to explore the management of drug treatments in the PREFER group compared with the control group.

Methods Doses and numbers of drugs and the number of patients receiving the target doses based on current guidelines were measured and compared between the groups at the start and finish of the study.

Results The percentages of ACE inhibitors (ACEIs) or mineralocorticoid receptor antagonists (MRAs) increased, while loop diuretics decreased in the PREFER arm during the study, although the differences were not significant. Beta-receptor blockers (BBs) decreased somewhat in both groups. The number of patients treated with MRAs differed the most between groups, and increased from 10 (28%) to 15 (48%) in the PREFER arm compared with 13 (35%) vs 13 (39%) in the control group. The change in patients receiving full target doses (+8 vs. +1) of the ACEIs/angiotensin receptor blockers, BBs and MRAs were significantly higher (p=0009) in the PREFER arm than in the control arm.

Conclusions Person-centred integrated care of patients with severe CHF was associated with increased evidence-based drug treatments, especially MRAs.

Clinical trial number NCT01304381.

  • Person-centred
  • Pharmacology
  • Chronic heart failure
  • Palliative care

Statistics from


Chronic heart failure (CHF) has a very poor 1-year survival rate among those severely affected, and the prognosis is worse than for many common forms of cancer.1 ACEIs/angiotensin receptor blockers (ARBs), β-receptor blockers (BBs) and mineralocorticoid receptor antagonists (MRAs) have improved management with survival benefit and reduced hospitalisations and are recommended in guidelines.1–6 We found great differences in end-of-life care in the Swedish Palliative Registry, suggesting that patients with heart disease and cancer receive unequal care.7 The European Society of Cardiology (ESC) states that selected patients require palliative care services in addition to HF management and that there is a great need for new models of integrated care, for example, HF and palliative, to facilitate equality of care for dying patients, regardless of the diagnosis.1

The course of CHF is unpredictable8 and pharmacological treatment has to be closely followed and often changed due to cardiovascular and renal instability in severe cases.9 The current structure and drug management of these severely ill patients seem to be suboptimal due to complicating comorbidities.10 A clinical study was conducted from 2011to 2013 with the purpose of integrating Palliative advanced home caRE and heart FailurE clinical caRe (PREFER) for patients with severe CHF.11 The main findings of the PREFER trial were significant improvements in health-related quality of life, functional classes, total symptom burden, reduced hospitalisations and fewer days spent in the hospital.11 Explanations for these improvements are likely to be multifactorial, and may include changes in pharmacological treatment. Many studies have revealed that pharmacological treatment of patients with CHF is often insufficient, based on current guidelines.12 ,13 Our hypothesis was that the person-centred structure within the PREFER concept led to improved pharmacological treatment, according to evidence-based guidelines.


We aimed to study: (1) the number of patients receiving target doses of ACEIs/ARBs, BBs and mineralocorticoid receptor antagonists (MRAs) and (2) the number of patients not reaching the target dose due to contraindications or side effects that prevented further increases. We also examined unexplained reasons for not reaching the doses.

Materials and methods

Study design

The present substudy is a post hoc observational explanatory study. The original PREFER trial was a prospective randomised study with open evaluation.

The PREFER model

The PREFER model was developed during the process of providing a person-centred integrated palliative care (PCC) and HF care for patients with CHF as recommended by ESC1 ,14 ,15 These steps were: examine patient features, confirm diagnosis, patient education, establishment of an advanced care plan, organisation of services, symptom management, identifying end-stage HF, breaking bad news to the patient and family and establishing new goals for care. The starting point is the patient's narrative, which is recorded in a structured manner and from which a mutual care plan is created that incorporates goals and strategies for implementation and follow-up.16 The PCC model is called the six S and consist of the six S key words; self-image, self-determination, social relationships, symptom control, synthesis and surrender.17 Patients in the intervention group (IG) were offered a multidisciplinary approach involving collaboration between specialists in palliative and HF care, that is, specialised nurses, palliative care nurses, cardiologist, palliative care physician, physiotherapist and occupational therapist. Rounds were scheduled every 2 weeks with all the team members (n=7), where evaluation and optimisation of drug therapy was performed. Advanced care, such as the administration of injections/infusions of diuretics and transmission of the ECG to the hospital, was performed in the patient's home. In addition, a personal care/health plan was developed using resources identified in each patient's illness narratives and by defining potential barriers16 such as problems using prescribed drugs. The PREFER model also included support/bereavement to family members support to team members and ethical rounds.15


Patients with a confirmed diagnosis registered at the HF clinic or primary healthcare centres and those who met the criteria mainly proposed by ESC1 were invited to participate in the study. Patients were included in the study if they had a verified left ventricular systolic HF with New York Heart Association (NYHA) III−IV symptoms and were considered optimally treated by their physician. Inclusion and exclusion criteria have been described earlier in detail.11 ,15

Randomisation and intervention

Patients (n=72) were randomised into either the PREFER intervention (n=36) or the control group (n=36). Randomisation was made in blocks of 10, blinded to participating doctor and nurses. The intervention was done as follows; (1) after identifying a patient who fulfilled the inclusion and had no exclusion criteria a responsible physician and nurse were appointed for each patient and; (2) the patient was then called for a thorough medical examination with identification of comorbidities and assessment of physiological, social and spiritual needs followed by; (3) a meeting with nurses who used a model for person-centred palliative care. Patients in the control group were managed by their responsible physicians primarily at the primary healthcare centre or at the hospital as appropriate.


Drug use

All drugs used during the study by each patient were noted at baseline (BL) and, for surviving patients, at the end of the study.

Patient data

Systolic and diastolic blood pressure (SBP/DBP), heart rate (HR), serum potassium (K) and serum creatinine (Cr) were measured at BL and at the end of the study.

Analysis plan

Disease modifying drugs

Doses of drugs termed ‘disease-modifying’ for CHF by ESC (ie, ACEIs/ARBs, BBs and MRAs) were calculated as the total daily dose (eg, 10 mg of enalapril twice daily becomes a total daily dose of 20 mg) and also as a percentage of the total daily target dose according to ESC guidelines for treatment of CHF (eg, metoprolol 50 mg twice daily, 100 mg/day, equals 50% of the recommended total daily target dose of 200 mg). When the target dose was an interval (eg, enalapril 20–40 mg daily), the lowest number in the interval was used. Since ACEIs and ARBs are interchangeable according to the guidelines, the aggregated dose of either one of these drugs was also calculated.

Regarding loop diuretics, only the regular daily doses were counted, not as needed (prn) additional doses. Those patients not taking loop diuretics were included in the average and assumed to have 0 mg daily. Almost all patients taking loop diuretics were treated with furosemide. Bumetanide was recalculated to an equivalent dose of furosemide, to a ratio of 1 mg bumetanide=40 mg furosemide.

Other drugs of secondary importance

They were collectively termed ‘other cardiovascular drugs’—a group that included non-loop diuretics, digoxin, statins, long-acting nitrates, platelet inhibitors, and anticoagulants—and ‘other cardiovascular drugs’, literally every other drug used by the patients. Unlike for disease-modifying drugs and loop diuretics, only the numbers of these drugs were noted in the study, not which drugs or doses. In cases with incomplete data, these were excluded when calculating the average number of ‘other drugs’ used for the different groups.

Analysis of patients not receiving target doses

We first calculated the number of patients who reached 100% of the target dose in the different drug classes and compared intervention and control groups at BL and at the end of the study. The reasons for not reaching target doses given by the treating physician were noted in the study and were: (A) side effects, (B) contraindications or (C) unknown. We summed up the number of patients for each reason at BL and at the end of the study in both groups. We analysed the data for patients with unknown reasons. In each of these patients, we checked vital signs and blood samples to find contraindications against further increases in the dose. When doing this, strict limits for contraindications were used regarding ACEIs/ARBs, BBs and MRAs.18 The contraindications were as follows:

  • For ACEIs/ARBs, estimated glomerular filtration rate (eGFR) <30, K >5.0 or SBP <90.

  • For BBs, HR <50 or SBP <90.

  • For MRAs, K >5.0.

eGFR was calculated using modification of diet in renal disease (MDRD) based on Cr. All limits for contraindication are from ESC guidelines for treatment of CHF.

We calculated the number of patients who did not receive the target dose of a certain drug and who had no obvious contraindication or side effect. This was used as an estimation of the possible room for improvement at BL. Then we evaluated whether the numbers improved in the IG compared with the usual care control group during the course of the study.

Deceased patients.

Twelve patients, eight in the IG and four in the control group, did not survive the full 6 months of the study. Of these, eight patients (five from the study group and three from the control group) only had BL data available. For the other four patients (three from the study group and one from the control group), data were recorded shortly before their death, at either 1 or 3 months into the study. These data were used for these four patients and the last noted findings were used as end of the study values, even though 6 months had not passed. This left us with 31 patients in the IG and 33 in the control group with drug use data from both BL and the end of the study.


Results are presented as the mean± SD, medians or frequencies expressed as percentages. Comparisons between the intervention and control groups were made using the χ2 test for proportions and Student’s t test for normally distributed continuous variables. All analyses were performed on an intention-to-treat basis. A p value <0.05 was considered significant. The data were analysed using PASW.


Oral and written informed consents were obtained. The study complies with the Helsinki Declaration. The Regional Ethics Committee for Human Research at Umeå University (reference number 2010-294-31M) approved the study, which was registered at (registration number NCT01304381).


From 24 January 2011 through 30 April 2013, we recruited 72 (36 in each group) patients, mainly at the medicine-geriatric clinic, Skellefteå County Hospital or from primary health care. The two groups were similar with respect to BL characteristics except for the mean age (PREFER 81.9±7.2 years vs control 76.6±10.2 years, p=0.012, table 1).11 There was no difference between the groups for use of devices or surgical procedures.

Table 1

Baseline characteristics

Drug use

Drug use in the PREFER and control groups at BL and the end of the study is presented in table 2.

Table 2

Prescribed drugs for CHF at BL and the end of the study

In the PREFER arm, the percentages of ACEIs and MRAs increased at the end of the study from BL, while loop diuretics decreased, although the differences were not significant (table 2). In the PREFER group, five patients died of whom two had ACEIs. Of the surviving patients, two were started on ACEIs. In the control group, three patients died, two of whom had ACEIs. One patient in each group had both an ACEI and an ARB. β Blockers decreased somewhat in both groups. The mean number of other cardiovascular drugs decreased from 1.7 to 1.5 in the PREFER group and from 1.7 to 1.6 in the control group. The mean number of other non-cardiovascular drugs increased from 5.6 to 6.3 in the PREFER arm and from 6.0 to 7.3 in the control group. The largest difference, although not significant, between the groups was found for MRAs, which increased from 10 (28%) to 15 (48%) in the PREFER arm compared with a change from 13 (35%) to 13 (39%) in the control group.

The average percentage of ESC target doses reached for separate classes of drugs was analysed at baseline and at the end of the study. For ACEIs, this percentage increased from 84% to 92% in the PREFER group compared to a decrease from 79% to 71% in the control group, For ARBs, there was an increase from 47% to 50% and from 62% to 67% in the PREFER and control groups, respectively. Corresponding values for the combination ACEI/ARB were an increase from 70% to 78 and a decrease from 70% to 69%, and for BB the increase was from 45% to 48% and from 48% to 51%, respectively. None of these changes were statistically significant either between or within group analysis.

Patients receiving the target dose of evidence-based pharmacological treatment for CHF are presented in table 3; while, the reasons given during the study why patients did not receive the target dose of disease-modifying drugs are given in table 4, the total number and percentages of patients receiving the target dose increased more in the PREFER arm than in the control arm, except for BBs (table 4). The change in patients receiving full target doses (+8 vs. +1) of the ACEIs/angiotensin receptor blockers, BBs and MRAs were significantly higher (p=0009) in the PREFER arm than in the control arm. The average dose of loop diuretics decreased from 58 to 48 mg in the PREFER arm, but increased from 54 to 58 mg in the control arm.

Table 3

Number and percentages of patients receiving the evidence-based target dose of pharmacological treatments for CHF

Table 4

Reasons given during the study for not reaching the target dose

Patients who did not receive the target dose for unknown reasons were examined to determine contraindications against increasing doses or starting treatment. Few contraindications were found among patients with unknown reasons for not reaching the target dose in both groups.


The main finding was that the guideline recommended pharmacological treatment for CHF could be improved. The total number of subjects receiving the target dose of ACEIs/ARBs and MRAs was significantly higher in the PREFER arm than in the control arm. It should be noted that each patient was supposed to be optimally treated by the responsible physician before randomisation.

Some patients received suboptimal treatment, even though the renin-angiotensin-aldosterone blockade was rather high at BL. A major increase in MRA usage was seen in the IG. It should be noted that doses were not recorded for MRA drugs in the study, making the data less reliable. Although the increased usage of MRAs was not significant, this finding suggests that MRA treatment can be increased in these patients. There was no obvious reason for the relatively low usage of MRAs at BL. The treating physician may have hesitated in using MRAs, possibly due to earlier studies showing a risk of hyperkalaemia and unstable renal function in patients at a relatively high age.19 ,20

The ACEI dosage was high at BL in groups and the proportion of patients on an ACEI increased somewhat in the IG. These findings indicate that the dosage of the ACEI can be increased, although the room for it was less than for MRAs. For ARBs, the picture was somewhat different. In both the PREFER and control groups, the average percentage of target dose reached among patients at BL was lower than for ACEIs. Moreover, usage did not increase significantly in either group during the course of the study. There was no obvious reason for this since ARBs have very similar contraindications and side effects as ACEIs. One explanation may be that doctors are not as comfortable with these drugs compared with ACEIs, or are unsure of the target doses.

The proportion of patients receiving BBs at entry was high and similar to that of ACEIs/ARBs and hardly changed during the study. The average target dose reached was lower than for ACEIs/ARBs and changed less than ACEIs/ARBs during the study period in both groups. The reason for this is unknown. For bradycardia or hypotension as possible contraindications for BBs, few patients (at most one third) had these contraindications, suggesting that the dose can be increased. Further efforts should be made to clarify why a higher BB target dose was not achieved.

For loop diuretics, there were some differences between the groups. Although they were not significant, these findings may still be important. There was a decrease of usage in the PREFER group and a small increase in the control group. One might speculate that better adherence or increased doses according to guideline recommendations of target doses may have contributed to a lower need for diuretics in the PREFER group, although this study cannot confirm this. For the quality of life of patients, it is important not to overtreat with diuretics, which can diminish thirst and lead to other side effects.

A prerequisite was that all patients should be optimally treated by the responsible physician before they could be randomised in the main PREFER study11 Although no statistical difference was found for each class of drugs, the summary of evidence-based treatment was improved, there was room to improve the pharmacological treatments in patients on new drug therapies. There is still a lack of good drug management among physicians.21 The management of CHF is not easy and the reasons for undertreating patients are probably likely to be multifactorial. One explanation may be that the structure of care for patients with CHF is inefficient due to the number of different caregivers. The structure and design of the new PREFER study may also have facilitated an improvement of drug treatment. The use of MRAs to control renal function and hyperkalemia may have increased so that patients could easily receive treatment during home visits by the nurses. The strategy used in the PREFER model goes beyond traditional HF management at hospitals by including person-centred and specialised palliative care.

One of the main differences from standard HF management22 is that the PREFER concept involves more comprehensive and person-centred care of the patient, taking into consideration symptoms and signs related to accompanying comorbidities, which were common in the present study. The person-centred approach could have contributed to optimisation of drug use by identifying barriers and facilitators through the patient's illness narratives.23 The support and interaction with family members in the person-centred strategy may also have contributed to the optimisation of drug treatment. We have found no other report of drug therapy management based on the concept of a person-centred integrated care. Our findings are reassuring, but further studies in patients with severe HF or other chronic conditions are mandatory.

Future studies

Larger randomised studies using the PREFER concept should be performed to determine whether it is possible to further optimise drug treatment for patients with severe CHF. Even if drug treatments were improved, the target dose of BBs was still relatively low. A possible secondary aim could be to find correlations between achieved target doses and the well-being and symptom relief.

Clinical implications

In patients with severe CHF and concomitant comorbidities, it is possible to optimise the drug treatments for HF provided that:

  • There is a good continuity of the caregivers to ensure a close monitoring of cardiac deterioration.

  • The caregiver team regularly questions if drug dosages should be increased/decreased or withdrawn.

  • Close control is maintained of renal function and electrolytes after dose adjustments.

  • Side effects are closely monitored.

  • There is an easy access for caregivers to report on the healthcare situation.

  • Patients and their family members are invited to work in partnership with the team members.

  • There is adequate documentation of the optimum tolerated dose and the reasons why target doses could not be achieved.


Ideally, the methodology and reasons for changing pharmacological treatments, especially in the IG, should have been noted more carefully during the study. Doses of MRA drugs were missing along with reasons for not achieving target doses. This makes data on this class of drugs less reliable. A major limitation was also that the study was not double blinded and with limited power due to the number of included patients and few events. Patients and care professionals were well aware of any pharmacological changes and were actively trying to improve treatment in a number of ways, not just by changing drug doses. There was a great discrepancy between the studied groups in the proportion of patients who did not reach the target doses. This was probably due to the information provided in the case records from the control group. In the PREFER group, there was regular discussion of optimal dosage of drugs, but it was more difficult to find this information for the control group. This may be one reason to discern if changes in pharmacological treatment had any effect, or whether it was any of the other improvements made to the patients’ treatments that caused changes. Finally, ‘other drugs’ should have been clarified in more detail, at least to determine whether possibly harmful drugs, such as non-steroidal anti-inflammatory agents, were used, and if they changed in the IG.


A new concept involving person-centred, integrated care of patients with severe CHF was associated with increased usage of evidence-based drug treatments for CHF, especially the aldosterone blockade. Future studies are needed to establish how much this change in drug treatment could explain the improvements in quality of life and morbidity.



  • Contributors See attached Contributor statement.

  • Funding This work was supported by the Swedish Association of Local Authorities and Regions, the Swedish Heart and Lung Association, the Ronnbaret Foundation. Skelleftea Municipality and King Gustaf V and Queen Victoria's Foundation.

  • Competing interests None declared.

  • Patient consent Obtained.

  • Ethics approval The Regional Ethics Committee for Human Research at Umeå University (reference number 2010-294-31M).

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.