Identifying Patients in Need of Goals of Care Conversations: Reliability,Acceptability,and Prognostic Significance of the Preference-Aligned Communication and Treatment Conversation Trigger Tool for Patients with Cancer

Abstract

Background:

Simple methods to help teams identify patients with goals of care (GOC) conversation needs are lacking.

Objectives:

To develop a tool to identify hospitalized patients who may benefit from GOC conversations.

Methods:

The Preference-Aligned Communication and Treatment (PACT) Conversation Trigger Tool was implemented as part of a quality improvement initiative in 10 Illinois hospitals and validated in a cohort of patients admitted to the coordinating site's oncology unit (n = 135).

Results:

The tool was reliable and acceptable to clinicians using it across sites. Thirty percent (n = 40) of patients screened at the coordinating site's oncology unit triggered positive. These patients were more likely to have a do-not-resuscitate order (43% vs. 11%) and palliative care consult (53% vs. 20%) and had lower mean survival time (125 vs. 248 days) than those who did not trigger (p < 0.001).

Conclusions:

The tool is reliable, acceptable, and can identify hospitalized oncology patients who may benefit from GOC conversations.

Introduction

For patients with serious illnesses, optimal communication with clinicians is a key component of high-quality care.^1,2 Goals of care (GOC) conversations between clinicians and seriously ill patients have been associated with improvements in patient experience,³ reductions in anxiety and depression,⁴ and possible reductions in health care spending.^5,6 Frequently cited barriers to discussing GOC include prognostic uncertainty and limited physician time.^7,8

Systematic interventions focused on stratifying those who have a short prognosis and may benefit most from GOC conversations are needed. A variety of tools are available for prognostication^9–12 and identifying patients with palliative care needs.^13,14 Many of these, however, are difficult to apply at the time of hospital admission due to need for laboratory results or physician input,^15–17 creating missed opportunities for nurses or social workers to facilitate these conversations earlier in the hospitalization.

The Preference-Aligned Communication and Treatment (PACT) project was a quality improvement intervention designed to promote and standardize GOC conversations among a broad spectrum of clinicians in the inpatient setting. The PACT Conversation Trigger Tool was developed as part of this project to screen for GOC conversation needs on hospital admission. In this study, we describe the reliability and acceptability of the PACT Conversation Trigger Tool across sites, as well as its validity within a subpopulation of patients with advanced cancer at the coordinating site.

Methods

The PACT faculty (10 experienced palliative medicine physicians) reviewed publicly available screening criteria, with a goal of developing a tool that could be quickly and accurately completed by any admitting clinician without extensive chart review or specialized medical knowledge. The PACT tool was adapted from consensus criteria for consideration of palliative care consultation, with a focus on variables suggesting unmet communication needs, rather than symptoms.¹⁸

Accordingly, the tool includes both medical comorbidities and markers of declining function. The faculty reached consensus about assigning points for each factor; two for each marker of functional decline and one for each major comorbidity. A total of three points was the threshold for triggering a GOC conversation because that would indicate at least one chronic illness with clinical decline or a combination of three or more chronic illnesses that would increase the risk of complications.

The tool was implemented at 10 sites as part of the PACT Project. For more information on the PACT process, please see Supplementary Appendix SA1. Teams at each site were provided education on the rationale and use of the tool through centralized training at an initial conference and web-based modules. The tool was completed within two days of admission, and patients who triggered positive were subsequently referred for a GOC discussion. The study was deemed exempt from review by the Northwestern University Institutional Review Board.

Reliability

To evaluate inter-rater reliability of the tool, 10 patients were double-scored using the tool at each site (n = 100). Each trigger criterion was evaluated for reliability using Cohen's kappa.

Acceptability

Acceptability of the tool was evaluated using an investigator developed 10-item survey derived in part from two published acceptability instruments.^13,14 The survey was completed by clinicians who used the tool at each site. Participants used a 5-point Likert scale (1 = strongly disagree to 5 = strongly agree) to rate the extent to which they agreed with each statement (e.g., “The tool is easy to understand.”).

Validity

As an initial assessment of validity of the tool, chart reviews were conducted on patients with solid tumors admitted to an oncology unit and screened with the tool at one site, an academic tertiary medical center, during the first six months of project implementation (October 1, 2015 through March 31, 2016), before initiation of GOC conversations. This single site was chosen because it was the coordinating site, and investigators had direct access to medical records.

Chart reviews were conducted between March and April of 2017. Data were collected on the index admission, subsequent readmissions (through December 2016), and patient mortality. Patient demographics were summarized using descriptive statistics and comparisons between groups were based on t tests and chi-square tests. Survival was evaluated using the Kaplan–Meier method.¹⁹ Statistical analyses were performed using SPSS, version 25.

Results

Reliability

The PACT Trigger Tool demonstrated moderate to high (0.73–1) reliability scores across the risk indicators. “Frequent admissions” showed the lowest reliability at 0.73 (Table 1).

Table 1.

Preference-Aligned Communication and Treatment Trigger Tool Reliability

PACT Conversation Trigger Tool	Points	k
Serious illness
Metastatic cancer	1	0.90
Heart failure, symptoms at rest, and/or requiring home oxygen	1	0.91
COPD or advanced lung disease, on home oxygen	1	0.89
End-stage kidney disease or failure, on dialysis or refusing dialysis	1	1.00
End-stage liver disease with ascites or encephalopathy	1	0.80
Dementia requiring any assistance with basic activities of daily living (e.g., dressing, bathing, and eating)	1	1.00
Frequent admissions related to one of the aforementioned illnesses (i.e., ≥2 admissions within three months for the same condition/disease)	2	0.73
Decline in function, feeding intolerance, or unintended decline in weight (i.e., failure to thrive)	2	0.91
Resident of long-term care facility or complex home care	2	0.89

Scores of ≥3 points were considered positive. Reliability was assessed using Cohen's kappa (k).

COPD, chronic obstructive pulmonary disease; PACT, Preference-Aligned Communication and Treatment.

Acceptability

One hundred eighty-five providers were surveyed (94.6% RNs); the completion rate was 77% (range 60–100%). The location where screening occurred (one per site) included general medical/surgical floors, cardiac telemetry, solid-tumor oncology, and intensive care units. Exploratory factor analysis with varimax rotation of the survey results yielded two factors reflecting the ease of completion and usefulness of the tool (Cronbach's α = 0.82 and 0.90, respectively). These factors demonstrated that providers found the tool useful (M = 4.07, standard deviation [SD] = 0.73) and easy to complete (M = 4.24, SD = 0.67).

Validity

One hundred thirty-five patients were screened at the coordinating site during the first six months of implementation, of which 40 patients (30%) triggered positive. Patient demographics are shown in Table 2.

Table 2.

Patient Demographics at Index Admission

	Triggered positive (n = 40)	Did not trigger positive (n = 95)
Age, mean (SD)	59.9 (15.3)	60.0 (12.4)
Gender
Male	14 (35%)	59 (62%)
Female	26 (65%)	36 (38%)
Race
White	10 (25%)	41 (43.2%)
Black	8 (20%)	19 (20%)
Other	22 (55%)	35 (36.8%)
Ethnicity
Hispanic or Latino/a	2 (5%)	6 (6.3%)
Non-Hispanic or Latino/a	35 (87.5%)	81 (85.3%)
Other	3 (7.5%)	8 (8.4%)
Marital status
Single	14 (35%)	28 (29.5%)
Married/partner	18 (45%)	53 (55.8%)
Separated/divorced	3 (7.5%)	7 (7.4%)
Widowed	4 (10%)	3 (3.2%)
Unknown	1 (2.5%)	4 (4.2%)
Insurance
Private	15 (37.5%)	40 (42.1%)
Public	25 (62.5%)	54 (56.8%)
Other	0	1 (1.1%)
Trigger Tool Score
0	0 (0%)	3 (3.2%)
1	0 (0%)	79 (83.2%)
2	0 (0%)	13 (13.7%)
3	25 (62.5%)	0 (0%)
4	7 (17.5%)	0 (0%)
≥5	8 (20%)	0 (0%)
Cancer diagnosis
Breast	10 (25%)	14 (14.7%)
Noncolorectal gastrointestinal	9 (22.5%)	14 (14.7%)
Colorectal	6 (15%)	9 (9.5%)
Lung	5 (12.5%)	15 (15.8%)
Other	4 (10%)	12 (12.7%)
Urogenital	3 (7.5%)	15 (15.8%)
Head and neck	3 (7.5%)	9 (9.5%)
Skin	0 (0%)	7 (7.4%)
Metastatic cancer	38 (95%)	86 (90.5%)

SD, standard deviation.

On the index admission, patients who triggered positive were more likely to have a do-not-resuscitate order (43% vs. 11%) and palliative care consult (53% vs. 20%) than those who did not (p < 0.001). Of those who triggered positive, 35% died or were discharged with hospice support (6/40 died during hospitalization; 8/40 discharged to hospice), compared with 6% (1/95 died; 5/95 discharged to hospice) in the group that did not trigger positive (p < 0.001). No significant difference was found between groups in length of hospitalization or rate of ICU transfers. Patients who triggered positive had a significantly lower mean and median survival time (M = 125 days, Mdn = 69 days,) than those who did not (M = 248 days, Mdn = 240 days, p < 0.001; Fig. 1).

FIG. 1.

Overall survival of patients who triggered positive versus who did not trigger at the index admission. Shown are the Kaplan–Meier curves for overall survival. The tick marks indicate censored data.

Discussion

The PACT Conversation Trigger Tool was shown to be reliable and acceptable to clinicians of multiple disciplines across 10 hospital sites and showed promising validity in a single site analysis. In a population of hospitalized patients with metastatic cancer, the PACT tool identified patients in the later stages of illness who may benefit most from a GOC conversation. A variety of prognostic tools have been developed previously, particularly for patients with cancer. However, these tools were developed for patients already known to be at high risk of short-term (30-day) mortality.^9–11

Other tools have been developed to identify patients who may benefit from expert palliative care consultation, often with a focus on symptom management or end-of-life care.^{13,14,18,20,21} Therefore, the utility of these tools in stratifying hospitalized patients who may benefit from a GOC conversation is unclear. Even the simplest of screening tools—“The Surprise Question” (“Would I be surprised if this patient died in the next 12 months?”)—has potential drawbacks,⁴ including its variable performance as a predictive tool for death, variable interobserver reliability, need for specialist input, and reluctance of some clinicians to answer.²²

The PACT tool may represent a promising addition to available tools, as it has several unique features. It is designed to be used for patients with an array of serious illnesses without the need for expert knowledge or familiarity with external guidelines.^23,24 It can also be completed without the need for physician input and integrated into the electronic medical record (EMR) allowing for incorporation into admission workflow without laboratory testing or in-depth chart review. Using such a tool on admission can help identify patients who may benefit from targeted communication interventions during the hospitalization.

Although GOC conversations should be introduced earlier, they often are not,^25,26 and hospitalization represents an opportunity to prompt discussion when clinicians and patients are available and a clinical change may prompt discussion.²⁷ Although one may argue that both groups in the validity analysis had a relatively poor prognosis, the tool identified patients with a significantly shorter prognosis who would benefit most from the limited resources to hold these discussions.

This study represents the practical experience with this tool and has several limitations. The cutoff for triggering positive was determined by expert discussion rather than empirical study. Reliability of one item (frequent admissions related to one of the aforementioned illnesses: ≥2 admissions within three months for the same condition/disease) was poor perhaps due to use of the “≥” symbol. This was changed on a subsequent version of the tool not included in this analysis.

This study only assessed validity using mortality and outcomes determined by chart review within one site; therefore, generalizability to other patient populations and settings is limited. Validation of the tool in patients hospitalized with noncancer diagnoses, as well as in the outpatient setting, will be necessary to understand its utility in these settings. Future research will assess the impact of these GOC conversations, with the hope of empowering patients and clinician champions to better align care with values at the end of life.

Conclusion

The PACT Conversation Trigger Tool was found to be reliable and acceptable across diverse hospital sites, and able to identify hospitalized oncology patients in the later stages of illness who may benefit from a targeted communication intervention. A larger validity study including other patient populations and evaluation of the impact of communication interventions on patient outcomes are needed.

Footnotes

Authors' Contributions

All authors made substantial contributions to the design of the project, data acquisition and analysis, drafting of the article and revisions. All authors have approved the final version.

Research Ethics and Patient Consent

The Northwestern University Institutional Review Board waived approval for this study as it was not considered research involving human subjects.

Funding Information

This study was supported by Blue Cross Blue Shield of Illinois. The funding source was not involved in the study design, in collection, analysis, or interpretation of data, in the writing of the report, or in the decision to submit the article for publication.

Author Disclosure Statement

No competing financial interests exist.

Supplementary Material

References

Sanders

, Paladino

, Reaves

, et al.: Quality measurement of serious illness communication: Recommendations for health systems based on findings from a symposium of national experts. J Palliat Med, 2020; 23:13–21.

Bernacki

, Block

; American College of Physicians High Value Care Task Force: Communication about serious illness care goals: A review and synthesis of best practices. JAMA Intern Med, 2014; 174:1994–2003.

Paladino

, Koritsanszky

, Nisotel

, et al.: Patient and clinician experience of a serious illness conversation guide in oncology: A descriptive analysis. Cancer Med, 2020; 9:4550–4560.

Bernacki

, Paladino

, Neville

, et al.: Effect of the serious illness care program in outpatient oncology: A cluster randomized clinical trial. JAMA Intern Med, 2019; 179:751–759.

Lakin

, Neal

, Maloney

, et al.: A systematic intervention to improve serious illness communication in primary care: Effect on expenses at the end of life. Healthc (Amst), 2020; 8:100431.

Patel

, Sundaram

, Desai

, et al.: Effect of a lay health worker intervention on goals-of-care documentation and on health care use, costs, and satisfaction among patients with cancer: A randomized clinical trial. JAMA Oncol, 2018; 4:1359–1366.

You

, Downar

, Fowler

, et al.: Barriers to goals of care discussions with seriously ill hospitalized patients and their families: A multicenter survey of clinicians. JAMA Intern Med, 2015; 175:549–556.

You

, Aleksova

, Ducharme

, et al.: Barriers to goals of care discussions with patients who have advanced heart failure: Results of a multicenter survey of hospital-based cardiology clinicians. J Card Fail, 2017; 23:786–793.

Maltoni

, Nanni

, Pirovano

, et al.: Successful validation of the palliative prognostic score in terminally ill cancer patients. Italian Multicenter Study Group on Palliative Care. J Pain Symptom Manage, 1999; 17:240–247.

10.

Maltoni

, Scarpi

, Pittureri

, et al.: Prospective comparison of prognostic scores in palliative care cancer populations. Oncologist, 2012; 17:446–454.

11.

Hung

, Wang

, Kao

, et al.: Magnitude of score change for the palliative prognostic index for survival prediction in patients with poor prognostic terminal cancer. Support Care Cancer, 2014; 22:2725–2731.

12.

Bestvina

, Wroblewski

, Daly

, et al.: A rules-based algorithm to prioritize poor prognosis cancer patients in need of advance care planning. J Palliat Med, 2018; 21:846–849.

13.

Waller

, Girgis

, Davidson

, et al.: Facilitating needs-based support and palliative care for people with chronic heart failure: Preliminary evidence for the acceptability, inter-rater reliability, and validity of a needs assessment tool. J Pain Symptom Manage, 2013; 45:912–925.

14.

Bowman

, George

, Barrett

, et al.: Acceptability and reliability of a novel palliative care screening tool among emergency department providers. Acad Emerg Med, 2016; 23:694–702.

15.

Moss

, Ganjoo

, Sharma

, et al.: Utility of the “surprise” question to identify dialysis patients with high mortality. Clin J Am Soc Nephrol, 2008; 3:1379–1384.

16.

Moroni

, Zocchi

, Bolognesi

, et al.: The ‘surprise’ question in advanced cancer patients: A prospective study among general practitioners. Palliat Med, 2014; 28:959–964.

17.

Walter

, Brand

, Counsell

, et al.: Development and validation of a prognostic index for 1-year mortality in older adults after hospitalization. JAMA, 2001; 285:2987–2994.

18.

Weissman

, and Meier

: Identifying patients in need of a palliative care assessment in the hospital setting: A consensus report from the Center to Advance Palliative Care. J Palliat Med, 2011; 14:17–23.

19.

Bland

, and Altman

: Survival probabilities (the Kaplan-Meier method). BMJ, 1998; 317:1572.

20.

ElMokhallalati

, Bradley

, Chapman

, et al.: Identification of patients with potential palliative care needs: A systematic review of screening tools in primary care. Palliat Med, 2020; 34:989–1005.

21.

Wang

, Huang

, Feng

, et al.: Validation of a concise screening tool for the identification of palliative care needs among inpatients: A prospective study in hospital setting. J Formos Med Assoc, 2019; 118:883–890.

22.

Downar

, Goldman

, Pinto

, et al.: The “surprise question” for predicting death in seriously ill patients: A systematic review and meta-analysis. CMAJ, 2017; 189:E484-E493.

23.

Fischer

, Gozansky

, Sauaia

, et al.: A practical tool to identify patients who may benefit from a palliative approach: The CARING criteria. J Pain Symptom Manage, 2006; 31:285–292.

24.

Youngwerth

, Min

, Statland

, et al.: Caring about prognosis: A validation study of the caring criteria to identify hospitalized patients at high risk for death at 1 year. J Hosp Med, 2013; 8:696–701.

25.

Mack

, Cronin

, Keating

, et al.: Associations between end-of-life discussion characteristics and care received near death: A prospective cohort study. J Clin Oncol, 2012; 30:4387–4395.

26.

Davison

: End-of-life care preferences and needs: Perceptions of patients with chronic kidney disease. Clin J Am Soc Nephrol, 2010; 5:195–204.

27.

Rocque

, Barnett

, Illig

, et al.: Inpatient hospitalization of oncology patients: Are we missing an opportunity for end-of-life care?. J Oncol Pract, 2013; 9:51–54.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.19 MB