Person-centred care for patients with chronic heart

0 downloads 0 Views 1005KB Size Report
Jan 24, 2015 - same for all patients, and an individual-related charge cost, obtained and .... computer program SAS 9.2 (SAS Institute, Cary, NC). (Figures 2 and .... 0.708. QALY. Mean (SD). 0.149 (0.035). 0.150 (0.034). 0.843. 0.155 (0.031).
567035

research-article2015

CNU0010.1177/1474515114567035European Journal of Cardiovascular NursingHansson et al.

EUROPEAN SOCIETY OF CARDIOLOGY ®

Original Article

Person-centred care for patients with chronic heart failure – a cost–utility analysis

European Journal of Cardiovascular Nursing 1­–9 © The European Society of Cardiology 2015 Reprints and permissions: sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/1474515114567035 cnu.sagepub.com

Elisabeth Hansson1,2, Inger Ekman1,2, Karl Swedberg2,3,4, Axel Wolf1,2, Kerstin Dudas1,2, Lars Ehlers5 and Lars-Eric Olsson1,2

Abstract Background: Costs of care for patients with chronic heart failure have been estimated at between 1% and 2% of the total health care expenditure in Europe and North America. Two-thirds are for inpatient care. Person-centred care (PCC) asserts that patients are persons and should not be reduced to their diseases alone. Aims: The aim of this study was to estimate the cost–utility of PCC when compared with conventional care in patients hospitalized for worsening chronic heart failure. Methods and results: Data for the cost–utility analysis were collected alongside a prospective clinical intervention study with a controlled before and after design from 2008 to 2010. Patient-specific resources used and preference-based health status data were collected at an individual level. Only 63% received PCC as intended illustrating the difficulties of introducing new methods in established organizations. The group intended to have PCC yielded higher costs in comparison with the conventional care group. The incremental cost was estimated as €98. The costs for those who actually received PCC, per protocol (PP) (63%) were significantly (p=0.026) lower than for those in the conventional care group, with an incremental cost-saving of €863. For the first three months, patients in the conventional care group showed decreasing health-related quality of life, with a corresponding improvement in the PCC(PP) group. Conclusion: It must be emphasized, however, that these positive effects, both cheaper and somewhat better, were obtained only among those receiving the PCC intervention in its intended form, PCC(PP). Keywords Person-centred care, chronic heart failure, cost–utility analysis, QALY, organizational culture Date received 9 September 2014; revised 9 December 2014; accepted 13 December 2014

Introduction Costs of care for patients with chronic heart failure (CHF) are among the highest for chronic diseases in the Western world.1-3 CHF is characterized by frequent hospitalizations and high mortality.4,5 The costs of treating patients with CHF have been estimated at between 1% and 2% of the total health care expenditure in Europe and North America. Two-thirds of the costs are for inpatient care.6,7 Among patients with CHF, care can be experienced as fragmented and confusing even though guidelines recommend active patient involvement as a requirement for effective care.4 Person-centred care (PCC) asserts that patients are persons and should not be reduced to their

1Institute

of Health and Care Sciences, Sahlgrenska Academy, University of Gothenburg, Sweden 2Centre for Person-Centred Care (GPCC), University of Gothenburg, Sweden 3Department of Molecular and Clinical Medicine, Sahlgrenska Academy, University of Gothenburg, Sweden 4Institute of Heart and Lung Diseases, Imperial College, London, UK 5Centre for Health Improvement, Aalborg University, Denmark Corresponding author: Elisabeth Hansson, Centre for Person-Centred Care (GPCC), Institute of Health and Care Sciences, Sahlgrenska Academy, University of Gothenburg, Arvid Wallgrens Backe, House 1, Gouthenburg, Sweden. Email: [email protected]

Downloaded from cnu.sagepub.com at Gothenburg University Library on January 24, 2015

2

European Journal of Cardiovascular Nursing

diseases alone. Their context, experiences, goals and wishes should be taken into account in deciding on care.8 A person-centred approach has been advocated as a key indicator for quality of care by the World Health Organization9 and by the Institute of Medicine at the US National Academy of Science.10 PCC has also been shown to improve outcomes such as length of hospital stay and ability to perform activities of daily living in patients with CHF.11 To introduce new ideas or changes such as PCC in publicly funded organizations, like the Swedish health care systems, with deep-rooted traditions is known to be difficult.12,13 Aij et al. for example found that implementation of new systems in highly professionalized hospital settings often is associated with challenges, because of the ambiguous and complex environment.14 To our knowledge, no study has evaluated the cost– utility of PCC in patients with CHF. The aim of this study was to estimate the cost–utility of PCC when compared with conventional care in patients hospitalized for CHF.

Methods Study design and data collection The data for this economic evaluation were collected alongside a prospective controlled intervention study conducted in one setting between 2008 and 2010 that used a controlled before and after design.11 A complete description of design and results of the clinical intervention have been published previously.11 In the present study all patient related resources used during the hospital stay, as well as the preference-based health status data, were collected on an individual basis. The cost–utility analysis (CUA) was based on comparison of the two modes of care, conventional care and PCC. The cost difference was calculated by subtracting the mean costs of PCC from the mean costs of conventional care. Cost per quality-adjusted life years (QALYs) was used as primary outcome. The difference in QALYs was calculated by subtracting the mean QALYs for PCC from the mean QALYs for conventional care. The economic evaluation was performed from a healthcare perspective, which in this study meant that only the costs for inpatient care were included. In this case, the effect of the PCC intervention was assumed to last longer than just the inpatient episode, and we therefore used data at three months as well as baseline. Only patients with complete data on costs and effects (utility) (baseline and at three months) were included in the CUA (Figure 1). The study population included patients who had been previously diagnosed with CHF according to the European Society of Cardiology (ESC) guidelines.15 Patients were admitted to a university hospital if they came to the emergency room of the hospital because of increased symptoms and/or signs indicating worsening heart failure. Before including any patient in the study, a study physician

confirmed the diagnosis of worsening CHF. All heart failure diagnoses were subsequently adjudicated by a committee before discharge.

Ethics statement Eligible patients were informed of the study and provided signed informed consent. The Regional Ethical Review Board approved the study (2008-03-17) Dnr: 046-08 and permission was also obtained from the head of the clinic. The investigation conforms to the principles outlined in the Declaration of Helsinki.

Conventional care (control group) A control group was recruited from February 2008 until April 2009, in order to carefully map and assess outcomes of conventional care. After the signed, informed consent, eligible patients were informed that they would be treated according to the usual care routines for heart failure at the department. They were also asked to complete questionnaires at baseline and follow-up (three months). The patients in the conventional care group were followed for three months to allow for comparisons with the patients in the PCC group. During the admission, the patients in the conventional care group were treated according to a common procedure where most of the attention is focused on the medical investigation and treatment strategies. Information from the patients was taken but in most part limited to their illness and their symptoms. Patients’ personal history regarding their living circumstances and ability was not taken during admission and no health-plan was developed in concert with the patient or relative. The patients were not given a view of the expected care or an expected date of discharge.11

PCC intervention Patients included in the PCC group were enrolled from April 2009 until April 2010 using the same inclusion and exclusion criteria.11 The intervention aimed to give the care in a structured and systematical way, combining evidence-based guidelines and clinical knowledge with the patients’ individual prerequisites, and form a partnership. The PCC was specifically designed to identify each patient’s resources and barriers and to guide the care. At admission, a comprehensive narrative was obtained from each patient, covering their everyday life, resources, motivation and goals. The patients rated themselves using both general and specific quality of life measures, as well as descriptions of their activity of daily living. After the examination, a tentative detailed healthcare plan was drawn up by a nurse built on the narrative, medical examination

Downloaded from cnu.sagepub.com at Gothenburg University Library on January 24, 2015

3

Hansson et al.

Figure 1.  Flow diagram of the participants in the conventional care and person-centred care groups.

and self-rated surveys. This specified the patient’s shortand long-term goals, resources and special needs after discharge. Based on the patient’s prerequisites, a prognosis for the hospital lengths of stay was also included in the healthcare plan. A follow-up on the health-plan was done after 72 h and there after every 48 h. If nothing new was found that could influence the care and the prognosis the health-plan was considered valid; if not, appropriate measures had to be taken and an update of the health-plan was done.11 All the patients intended to receive the intervention belonged to the PCC group. In spite of the detailed introduction provided to all personnel on the new concept of PCC some of the personnel did not provide the patients with PCC as intended. For that reason a per protocol (PP) group was defined consisting of those patients who received the complete PCC intervention during their hospital stay. Costs and effects were thus analysed in both the PCC and PCC(PP) groups. The cost for development and introduction was covered by the clinic’s ordinary policy for quality and competence development. The researchers who led the development were funded by independent research funds.

Costs All costs were collected from the hospital’s administrative database allowing use of comprehensive information relating to each patient. The total cost consists of a fixed cost, the same for all patients, and an individual-related charge cost, obtained and analysed retrospectively (consisting of the total costs for hospitalization from admission to discharge). Daily care costs included all fixed costs, that is, staff salaries, accommodation, administration and competence development of the staff. Some routine treatment costs like common drugs were also included in the daily care costs. Those costs varied somewhat depending on the type of ward (in one of the wards a few beds were used for more intensive care, which made the average daily care costs higher). Individualrelated costs consisted of patient-specific expenses such as laboratory tests or imaging. The costs are presented in euros, using 2010 values (average 2010 exchange rate).

Quality of life For this analysis, assessment of health-related quality of life used the EQ-5D 3L instrument at baseline and at three

Downloaded from cnu.sagepub.com at Gothenburg University Library on January 24, 2015

4

European Journal of Cardiovascular Nursing

months after discharge for the conventional care, PCC and PCC(PP) groups (Figure 1). EQ-5D 3L is a multi-attribute utility instrument where patients’ self-reported health states are valued in terms of utility using the preferences of a sample of the public.16 The instrument consists of five dimensions (mobility, self-care, usual activities, pain/discomfort, and anxiety/depression). Each dimension is divided into three levels (none, moderate and severe problems). A total of 243 possible health states are therefore defined. Each state is referred to in terms of a five-digit code. Each combination is connected to a quality of life weight, by using the so-called time trade-off method. A summary index between 1 (full health) and 0 (death) has been calculated. Values below 0 correspond to worse than death.17,18 Social tariff values (elicited health state valuations from a representative sample of the UK health population) for these health states have been previously published19 and were used in our study.

the costs nor the effects were discounted. The incremental cost-effectiveness ratio (ICER) in the PCC and PCC(PP) was calculated as:

QALY

Sensitivity analysis

The quality of life weight was then used to calculate QALYs. This measure combines years of life with quality of life so that the QALY, as a result of a treatment, can consist in increasing life expectancy and/or increased quality of life. QALY calculations were made on an individual level, reflecting the change from baseline to three months, assuming a linear increase in quality of life (QoL) between the two measurements.

A single trial with a single estimate of the ICER may be a ‘bad’ representation of the ‘true’ value due to, for example, sampling uncertainties. To examine the robustness of the results, the bootstrapping technique was used. This sensitivity analysis estimates an empirical sampling of ICERs as if a number of studies were performed, each with its own ICER.

  EQ5D baseline +   QALY =    / 2  * ( 0.25 years )   EQ5D 3 months   A regression-based adjustment was used for the calculation of QALYs, to manage the imbalance between groups in mean baseline utilities. This is important since a patient with a low EQ-5D score at baseline is more likely to experience a QALY gain over time.20,21 QALYi = β0 + β1 * t i + β2 * Q b β0 = intercept; β1 = beta-coefficient for the treatment arm; β2 = beta-coefficient for EQ-5D value at baseline; i = patient identifier; ti = treatment arm (0= conventional care group, 1=PCC group); Qb = EQ-5D value at baseline.

CUA The CUA was performed from a healthcare perspective. Since the follow-up period was only three months, neither

ICER= Δ COST(Cost CC – Cost PCC or PCC (PP))/ Δ Utility (QALY CC – QALY PCC or PCC (PP)) ICER can be interpreted as the extra cost of obtaining one extra QALY, and allows comparisons between interventions in all areas of healthcare. The results of the analysis can demonstrate whether the PCC yields higher or reduced costs, decreased or improved utility in comparison with usual care, and are presented in a so called cost-effectiveness plane. If the intervention is both cheaper and better, it can be regarded as dominant.16

Statistics The statistical analyses were performed using IBM SPSS statistics version 19. Between-group differences were tested using the chi-squared test and Fisher’s exact test for categorical variables, and Student’s t-test and the Mann–Whitney U test for continuous variables. The bootstrapping technique with unrestricted random sampling was used for the estimation of the 95% confidence interval (CI) for the ICER, in order to provide bootstrapped estimates of costs and QALYs, to be used to plot the cost-effectiveness plane (Figures 2 and 3). This technique is recommended to assess uncertainty in trial based (primary data) cost-effectiveness analyses.22,23 The number of units is the same as the number of patients in the analysis and the CI is based on 1000 such samplings. The bootstrapping analysis was done in computer program SAS 9.2 (SAS Institute, Cary, NC) (Figures 2 and 3).

Results A total of 248 patients, 123 in the conventional care and 125 in the PCC group, were enrolled (see Figure 1). No statistically significant differences between the groups were found in baseline characteristics except for New York

Downloaded from cnu.sagepub.com at Gothenburg University Library on January 24, 2015

5

Hansson et al.

Figure 2.  Bootstrapped (n= 1000) and observed cost versus regression-based QALY. PCC group versus control group. Diff: difference; PCC: person-centred care; QALY: quality-adjusted life year

Figure 3.  Bootstrapped (n=1000) and observed cost versus regression-based QALY. PCC(PP) group versus control group. Diff: difference; PCC(PP): person-centred care according to protocol; QALY: quality-adjusted life year

Heart Association (NYHA) score and EQ-5D. The patients in the PCC group had worse symptoms and lower quality of life than the conventional care group11 (Table 1). Complete data for costs and effects were registered at three months for 95 patients in the conventional care group and 97 patients in the PCC group. Almost a quarter (23%) of the conventional care group and 22% of the PCC group had incomplete data, 8% died in each group, 2% of the conventional care group had no baseline EQ-5D and 3% in the conventional care group and 2% in the PCC group had no economic data. A drop-out analysis showed that there were no statistically significant differences between those with complete and incomplete data in either group in terms of EQ-5D

(conventional care p=0.245, PCC p=0.066) or NYHA class (conventional care p=0.928, PCC p= 0.262). The majority (63%) of patients in the PCC group received the complete intervention and so were included in the PCC(PP) group (Figure 1). The remaining patients in the PCC group received a mix of usual care and PCC. The average hospital stay in the three groups were: in conventional care group 7.3, PCC 6.6 and PCC(PP) 5.3 days. The PCC(PP) group differed in a statistically significant way from the conventional care group (p=0.036).

Costs Table 2 shows unit costs for daily care and for individual costs.

Downloaded from cnu.sagepub.com at Gothenburg University Library on January 24, 2015

6

European Journal of Cardiovascular Nursing

Table 1.  Baseline characteristics for the conventional care group (CC) and person-centred care group (PCC). Variable

CC n=123

PCC n=125

p-value

Age, mean (SD) Gender, female, n (%) EQ-5D score, mean (SD) NYHAb Class 1, n (%) Class 2, n (%) Class 3, n (%) Class 4, n (%)

80.3 (9.14) 50 (41) 0.629 (0.311)

77.5 (10.60) 52 (42) 0.536 (0.318)

5 (4) 50 (41) 64 (52) 4 (3)

0 (0) 37 (31) 69 (58) 13 (11)

0.08 0.98 0.02a   0.002c      

aA

regression-based adjustment was used in the quality-adjusted life year calculations. York Heart Association (NYHA) class level. cChi-squared test and Fisher’s exact test. bNew

Table 2.  Costs per unit (€). Daily care costsa (€) Ward, 1 (including an ICU with four beds) Wards, 2–4 Individual costsb per unit Imaging: lung/heart examinations Clinical physiology: ultra cardiography Laboratory tests: routine blood testc Microbiology: Urine C/S

659 554 60 280 61 20

aDaily

care costs included all fixed costs, that is, staff salaries, accommodation, administration and even some routine treatment costs like common drugs. bIndividual costs such as imaging, laboratory tests, ultrasound and urine C/S. Urine C/S was frequent. cNa–, K+, creatinine, haemoglobin (HB), B-type natriuretic protein (PRO BNP). ICU: intensive care unit; C/S: culture/sensitivity

Table 3 shows the costs for the patients in the three analysed groups, conventional care, PCC and PCC(PP). There were no statistically significant differences in costs between the PCC and the conventional care groups, although the individual costs tended to be somewhat lower in the PCC group. When the conventional care and PCC(PP) groups were compared, on the other hand, the daily care costs, the individual costs, as well as the total costs were all lower in the PCC(PP) group (p