An Intervention to Enhance Goals-of-Care Communication Between Heart Failure Patients and Heart Failure Providers

Abstract

Context

Heart failure patients contend with a markedly impaired quality of life, experiencing emotional distress and severe physical discomfort that increases in frequency in the last months of life. Improving communication between patients and providers about goals of care has the potential to improve patient-provider communication and patient outcomes.

Objectives

To determine the effects of a goals-of-care (GoC) intervention compared to usual care on the number of GoC conversations, quality of communication between patients and providers, referrals to palliative care services and completion of advance care directives.

Methods

A two-group randomized study (n = 40/group) compared a GoC intervention to usual care, conducted in an academic HF clinic. The GoC intervention was a pre-visit patient activation-education, telephone-based intervention delivered by a nurse. The primary outcome of the study was number of GoC conversations between HF patients and HF providers. Secondary outcomes were quality of communication, number of referrals to palliative care and completion of advance directives.

Results

Patients averaged 58.15±11.26 years of age, with mean LVEF= 30.31±9.72%, and SHFM scores= 95.1±1.60. There was a significant increase in goals-of-care conversations (58% vs. 2.6%, P<0.001) and quality of end-of-life communication (P=0.03) in the GoC group compared to usual care after the intervention. There were no differences between groups on the other outcomes.

Conclusion

The GoC intervention resulted in more GoC conversations and higher quality communication between HF patients and providers without increased anxiety or depression. Further studies are needed to assess impact on longer-term quality of care and patient outcomes.

Introduction

Heart failure (HF) affects more than five million people in the United States, with 550,000 new cases diagnosed annually.¹ Among older adults, advanced HF accounts for more hospitalizations, more physician visits, and greater lengths of hospital stay than any other condition,^2,3 resulting in significant financial burden. In addition, HF patients experience a markedly impaired quality of life, including both physical and emotional distress that increases in the last months of life, and many HF patients and families do not access palliative care services early enough in the trajectory of illness to derive benefit.⁴ Enhancing communication between patients and providers about goals of care (GoC) has the potential to align HF treatments received with patient and family desires and to reduce high cost care at the end of life.^4–6

Although clinical practice guidelines recommend that cardiology providers engage in GoC conversations and refer patients to palliative care,^7,8 there are no specific strategies to determine when GoC conversations should occur.⁹ Thus, HF patients and family caregivers report minimal communication with providers regarding expectations of illness trajectory, prognostic estimates, symptom management, implantable cardioverter-defibrillator deactivation, and advance care planning.^10,11 Providers often fail to initiate such discussions with patients,^12,13 and when they do, the prognostic estimates may not be well received by patients.¹⁴ Common reasons for lack of GoC conversations include poor patient and family education regarding the progressive downward trajectory of HF, lack of willingness among both patients and providers to discuss end-of-life planning, and difficulty predicting prognosis in HF.^10,15–17

Little has been written about GoC interventions in the HF population. A recent systematic review of patient-professional communication interventions for life limiting conditions found 16 published studies over a 14 year period, only 1 of which was conducted in HF patients. In the study involving HF patients, an advance directive interview resulted in improved decision making for future medical treatments.¹⁸ The review also suggested three types of interventions that are needed to effect outcomes in HF: 1) provider communication skills, 2) patient understanding of their condition, and 3) advance care planning. Another study described use of an outpatient palliative care consultation to address attitudes about and completion of advance directives in 36 symptomatic HF patients, noting that the intervention increased the completion of advance directives that were discussed with the family, but not with the health care provider.¹⁹ These studies suggest an important need for interventions that improve discussions about GoC and advance care planning between patients with advanced HF and their HF clinicians.

Many patients for whom palliative care could significantly improve quality of life do not have access to such care services.^16,20–22 Instead of palliative care, patients at end-of-life often receive intensive and costly care even when it may not contribute significantly to prolonging life.¹² Lack of GoC communication in advanced HF results in slow progression towards death with high symptom burden^23,24 and decreased quality of life,^25,26 increased length of stay in inpatient and intensive care units, and increased costs of care near the end of life.^5,27 Therefore, the purpose of the study was to determine if an intervention designed to assist patients in initiating GoC conversations with HF providers would result in increased numbers of GoC conversations, improved quality of communication with HF health care providers, referrals to palliative care services and completion of advance directives. This small trial was designed as an efficacy study to examine impact on short-term processes of care, GoC discussions, and quality of communication.

Methods

Design

The study used a randomized two-group study design (n = 40/group), testing a GoC intervention against usual care (UC). The primary outcome was number of GoC conversations between HF patients and providers during the HF clinic visit that followed the intervention. Secondarily, we described the effect of the intervention on quality of communication, referrals to palliative care services, completion of advance care directives, anxiety, and depression. Relevant data were collected from each participant at two times: baseline study entry and approximately two weeks after a regularly scheduled HF clinic visit. The study was approved by the institutional review board at the University of Washington, and all participants provided written informed consent.

Setting and Participants

The study was conducted in a HF outpatient clinic in an academic medical center in the Pacific Northwest. The HF clinic is centered within a HF mechanical circulatory support and heart transplant center. This clinic has eight attending physicians, two nurse practitioners, 24 rotating cardiology and HF fellows, and three nurses who care for approximately 600 patients each year. Eighty (N = 80) HF patients were recruited at the HF Clinic. The inclusion criteria were: 1) diagnosis of heart failure with reduced ejection fraction (HFrEF) with EF ≤40% or heart failure with preserved ejection fraction (HFpEF) with EF <50%; 2) completion of an outpatient HF visit within the past six months with a scheduled follow-up visit; and 3) ability to read, write, and speak in English. The exclusion criteria were: 1) Short BLESSED cognitive score > 10 to rule out significant cognitive impairment;²⁸ 2) diagnosis of terminal illness with life expectancy of ≤ 1 year not related to heart disease; 3) psychiatric illness that required hospitalization in the past year; and 4) age less than 18 years. Seattle Heart Failure Model (SHFM) scores were calculated using information from electronic health records (EHR).

Intervention

The patient GoC intervention, guided by the self-management for chronic conditions model²⁹ and a prior intervention³⁰ included: 1) telephone-based pre-visit coaching conducted by a nurse where the patient identified current perceived barriers and facilitators to communication with the cardiology provider,³⁰ education about standard HF therapies, patient preferences for communication about end-of-life care, and completion of the Five Wishes advance directive form; 2) a one-page patient activation outline that was constructed from pre-visit coaching call that was shared with both the patient and their HF provider, along with suggestions for addressing barriers and using facilitators; and 3) patient activation, skills enhancement, and role playing conversation openers to initiate a GoC discussion with the provider at the next upcoming HF clinic visit. At the conclusion of telephone coaching, participants were asked to rate desire and confidence to engage in a GoC conversation with their HF provider, using a 0–10 scale (0= definitely no – 10=definitely yes), and to designate their desired role in shared decision making.

The provider GoC intervention consisted of receipt of the patient activation outline that was constructed during pre-visit coaching and given to the provider on the day of the planned HF clinic visit. Additionally, providers received patient-specific mortality estimates calculated from the SHFM,^31–33 the total score on the Kansas City Cardiomyopathy Questionnaire (KCCQ),³⁴ information about the patient’s desire for involvement in shared decision making, and coaching communication tips specific to the patient’s identified barriers and/or facilitators. The provider was asked to facilitate a GoC discussion with the patient at the HF clinic visit and document this in the EHR.

Usual Care

Usual care consisted of regularly scheduled outpatient clinic visits in the HF Clinic. All patients received treatment as usual from their health care providers in addition to completing data collection questionnaires, but did not receive feedback on the questionnaires nor the intervention. Providers caring for patients randomized to usual care did not receive the GoC patient activation outline and were not told that the usual-care patients were participating in the study.

Data Collection and Measures

The primary outcome was the number of GoC conversations conducted between the patient and HF provider after the intervention. Goals-of-care conversations were determined using patient self-report as well as documentation of the conversation in the EHR.

Secondarily, we were interested in quality of communication (QOC),^35–37 number of referrals to palliative care, and number of advance care directives completed. To determine if GoC conversations effected emotional status, anxiety (GAD-7)³⁸ and depression (PHQ-9)³⁹ were also assessed. All measures were completed at study entry and again 2 weeks following the target HF clinic visit. The QOC questionnaire is a 17-item inventory that measures quality of communication with the provider. The QOC questionnaire has good reliability and validity in populations with COPD and HIV, and is sensitive to change with a communication intervention.^35–37 Referrals to palliative care services were determined from the EHR by the number of new referrals to hospice and palliative care up to 6 months after completion of the intervention. Completion of advance care directives were assessed by patient report as well as by examining the EHR for a scanned copy of the directive. Depression was assessed using the 9-item version of the Patient Health Questionnaire (PHQ), which has been validated against diagnostic interviews conducted by mental health professionals. The PHQ-9 also provides an index of depression severity; this index has correlated significantly and highly (r = 0.84) with mental health professionals’ assessments of depression symptom severity.³⁹ Anxiety symptoms were assessed using the Generalized Anxiety Disorder (GAD-7), which has been validated against diagnostic interviews conducted by mental health professionals and has reported good reliability.³⁸

Measures that facilitated completion of the intervention included the Barriers and Facilitators questionnaire,^30,40 the KCCQ,^34,41 and desired involvement in shared decision making.⁴² The Barriers and Facilitators Questionnaire asks about the patients’ preferences for communication about end-of-life care. Barriers include items such as whether patients would rather concentrate on living than talk about dying or that the patient perceives not being sick enough to talk about end-of-life care and facilitators include items such as the patient is worried about their quality of life in the future or being a burden on family.^30,40 The KCCQ evaluates distinct health status domains for heart failure: physical limitations, symptoms, social function, self-efficacy, and quality of life. A total score based on contributions from each domain quantifies the multiple domains of the KCCQ into a single summary score. The psychometric properties of the KCCQ have been tested extensively in many groups of HF patients, and the scales have demonstrated good reliability and validity.^34,41 The Degner Control of Preferences Scale allows the patient to choose one of five levels of involvement in shared decision-making ranging from (1) “I prefer to make the final decision about what treatment I will receive” through (5) “I prefer to leave all decisions regarding my treatment to my doctor.” The middle number (3) reflects sharing decisions between the patient and the provider.⁴²

Data Analysis

All of the analyses were conducted on the basis of intention to treat. Chi-square (χ2) or t-tests were used to determine baseline equivalence in demographic and outcome measures. Comparisons were made by group using analysis of covariance (ANCOVA), controlling for baseline characteristics. Statistical significance was defined as a two-tailed α level ≤0.05. SPSS version 19.0 (SPSS Inc, Chicago, IL) was used for the analysis. It was determined that a sample size of 80 patients provided 80% power to detect the between-group difference in number of GoC conversations, assuming a 28% improvement between treatment groups and a 20% dropout rate.³⁰

Results

Participants

Over a two-year period, 660 HF patients were screened for trial eligibility. Of 237 eligible, 80 (61 men and 19 women) were enrolled and randomly assigned (Fig. 1). Reasons for not entering the study included refusal with no reason given, not interested, or too busy (n=102). A smaller number (n=6) stated they did not want to engage in GoC conversations. The attrition rates in the intervention and usual care groups were comparable (GoC= 2 vs. UC=5). Patients averaged 58.2+ 11.3 years of age, with a mean LVEF= 30.3±9.7%, and NYHA FC = 2.4±0.8 and SHFM score = 95.1±1.6. A SHFM score of 95% indicates that the average SHFM predicted one-year survival for the group was 95%. Demographic characteristics were comparable between the two groups (Table 1).

Figure 1.

Table 1.

Demographic Characteristics of Sample

Demographics	Total = 80	Intervention N=41	Usual Care N=39	T or X2	p value
Age (years)	58.15 ± 11.26	60.00 ± 10.39	56.21 ± 11.93	1.52	0.13
Gender (Male)	61 (76.3%)	30 (73.2%)	31 (79.5%)	0.44	0.51
Ethnicity (Caucasian)	66 (82.5%)	33 (80.5%)	33 (84.6%)	1.76	0.62
EF %
HFrEF ≤ 40% (N,%)	69 (86.3%)	33 (41.3%)	36 (45.0%)	2.36	0.11
HFpEF > 40% (N%)	11 (13.8%)	8 (10.0%)	3 (3.8%)
Seattle HF Model Score	95.1 ± 6.0	95.0 ± 4.5	95.2 ± 7.3	0.14	0.89
NYHA Functional Class
I	9 (11.3%)	4 (5.0%)	5 (6.3%)	2.53	0.47
II	37 (46.3%)	22 (27.5%)	15 (18.8%)
III	31 (38.8%)	13 (16.3%)	18 (22.5%)
IV	3 (3.8%)	2 (2.5%)	1 (1.3%)
KCCQ-Overall summary	66.97 ± 22.43	68.66 ± 23.02	65.28 ± 21.85	0.67	0.50

On average, intervention participants reported their desire to engage in a GoC conversation as (mean+SD) 6.37±3.28, and their confidence to engage in a GoC conversation as 8.79±2.22. Using Degner et al.’s⁴² tool to assess desire for participation in medical decision making, half of intervention participants (53.7%) endorsed the statement, “I want to make decisions about my care with my provider,” while 29.3% endorsed the statement, “I want to make my own decisions about my care after considering my provider’s opinion.”

Primary Outcome

The GoC intervention was associated with a statistically significant increase in GoC conversations between HF patients and providers, 58% vs. 2.6%, P<0.001. Patients identified their top barrier to having a GoC conversation as wanting to focus on staying alive rather than talk about death (82.5%). The top facilitator to having a GoC conversation was worry about quality of life in the future (75.9%). There were no statistically significant differences between the groups in identified barriers or facilitators to having a GoC conversation with their HF provider.

Secondary Outcomes

Baseline outcome measures did not significantly differ between GoC and usual care groups. We found higher quality of end of life communication reported in GoC compared to the UC group (F=5.09, P=0.03). Importantly, the significant increase in goals-of-care conversations did not result in increased anxiety or depression (Table 3).

Table 3.

Comparisons of Quality of Communication and Psychological Reactions, Baseline to post HF clinic visit by group

Outcomes (mean±SD)		Baseline	After visit	F	p value	Effect Size Cohen’s D
QOC-General	UC	8.99 ± 1.85	9.58 ± 0.94	0.10	0.76	0.13
	IN	9.19 ± 1.10	9.46 ± 0.82
QOC-EOL	UC	3.90 ± 2.82	4.47 ± 2.78	5.09	0.03	0.43
	IN	3.74 ± 3.22	5.76 ± 3.18
PHQ-Depression	UC	7.32 ± 6.62	5.60 ± 5.80	0.41	0.52	0.02
	IN	5.41 ± 5.17	5.47 ± 5.03
GAD-Anxiety	UC	5.94 ± 6.13	4.15 ± 4.70	2.98	0.09	0.08
	IN	3.69 ± 5.21	3.72 ± 5.48

UC = usual care

IN = intervention

There were no significant differences noted between the groups on number of palliative care referrals or completion of advance directives at the conclusion of the GoC intervention; however, the study was not powered to detect these differences.

Discussion

As reported in the literature, heart failure patients must contend with a markedly impaired quality of life, often experiencing emotional distress and severe physical discomfort that increases in frequency and severity during the last months of life.⁴³ Similar to other studies, our HF patients did not access palliative care services early enough in the trajectory of illness to derive benefit.⁴⁴ There is a growing awareness of the importance of integrating palliative care in patients with heart failure; however, there is an important need for the development of strategies so that existing barriers to GoC conversations are decreased.⁴⁵ This study tested a GoC intervention that incorporated pre-visit coaching, HF education, and patient activation on number of GoC conversations between HF patients and providers, noting that the intervention significantly increased the number of GoC conversations that occurred while also increasing patient-rated quality of communication with the HF provider. The intervention was not associated with increased anxiety or depression. The total number of referrals to palliative care or completion of advance health care directives were not significantly increased after the intervention; however, this was a small study was not designed to detect these differences. This study did generate the preliminary data to support a future, large-scale randomized trial that can further tailor the GoC intervention and assess intervention effects on healthcare received and long-term patient outcomes.

The goal of having a GoC conversation is a better understanding of the patient’s values and goals as well as preferences for end-of-life care. These conversations may offer the opportunity for referral to additional services to address unmet palliative care needs and improve overall quality of life.^13,25 Our study found that a communication intervention that coached patients and reminded providers to engage in a GoC conversation increased the number of timely GoC discussions. These conversations, occurring over time and including patients and their extended support network, may allow patients to receive additional support and thereby enhance their quality of life.^18,22

We did not collect data that would tell us whether the patient or the provider initiated the GoC conversation, but the intervention was designed to encourage both to initiate the conversations. Future research could include recording of the GoC conversations to be able to better understand the multifactorial nature of the GoC conversation and who initiated the conversation. Additionally, we did not target caregivers as a part of this study and did not collect data on the percent of patients who had a caregiver. There is a substantial body of literature of the influence of caregivers in heart failure.⁴⁶ Thus, future research could include caregivers as part of the GoC intervention.

This study has several important limitations. First, it took place in a single HF clinic where providers had patients participating in both groups of the study (GoC and usual care) simultaneously. It is possible that providers may have become more aware of the need for GoC conversations in all patients, and instituted these conversations inadvertently with patients in usual care. However, given the low proportion of patients with GoC conversations in the usual care group, this appeared not to be the case. Second, patients who volunteered to participate may differ from those who declined. A large number of individuals who received care in the HF Clinic (47%) were not eligible for study participation, primarily because of the EF criteria or they had not been seen within six months of the screening date (19%). Third, we assessed GoC discussions as a binary outcome without assessing the nuances of these conversations. Future research is needed to examine the actual conversations. Finally, this study took place in one region of the U.S. at a tertiary care academic medical center and may not generalize to other regions or settings.

In conclusion, a telephone based, nurse-led GoC intervention resulted in more GoC conversations and higher quality communication between HF patients and HF providers, without significant evidence of psychological distress. Although we found improvements in short-term processes of care, further studies are needed to determine whether this intervention changes longer-term healthcare received and other patient outcomes. The intervention warrants testing in other settings with larger samples of HF patients, especially those outside the academic medical center environment.

Table 2.

Comparison of Primary and Secondary Outcomes After HF visit by group

After Visit Outcomes N (%)	Intervention=41	Usual Care=39	X2	p value
GoC Conversation	24 (58.5%)	1 (2.6%)	27.3	≤ 0.001
Palliative Care/Hospice Referral	4 (9.7%)	0 (0.0%)	2.97	0.13
Advance Directive	7 (16.1%)	3 (7.7%)	0.30	0.24