Communication Interventions to Improve Goal-Concordant Care of Seriously Ill Patients

An Integrative Review

Frank Bennett, MDiv, BS; Susan O'Conner-Von, PhD, RN-BC, CNE


Journal of Hospice and Palliative Nursing. 2020;22(1):40-48. 

In This Article


Study Selection

Using the aforementioned MeSH terms, a literature search was conducted using Boolean search logic tools embedded in the Ovid MEDLINE and Cochrane databases and a search of citations and the Mendeley Elsevier database. A gray literature search was not conducted because of feasibility limitations. After removing duplications, 422 abstracts were reviewed using the inclusion criteria, specifically study design methods and outcomes. Twenty-three RCT studies published between January 1, 2009, and June 1, 2018, met the PICOS inclusion criteria and were included, per the PRISMA process in Figure 1. A full text review of the 23 studies was conducted by the primary researcher. The included studies were grouped by their modality of intervention for supporting GOC communication, comprehension, and collaboration between clinicians, patients, and families: (1) Early Palliative Care for seriously ill patients, (2) Clinician Communication Training, (3) Patient Communication Support, (4) Family Communication Support, (5) Patient-Clinician Communication, and (6) Patient Decision Aids.

Appraisal of Quality

Quality appraisal was assessed using the Joanna Briggs Institute (JBI) RCT critical appraisal tool.[16] The JBI RCT quality appraisal checklist, which comprised 13 assessments of risk of bias, was used for the 23 included RCT studies. We concluded there is a low risk of bias in 21 of the 23 RCT studies, which met between 11 and 12 of the 13 JBI criteria for RCT studies. There was a moderate risk of bias in 2 studies, one in which neither participants, interventionists, nor assessors were blinded to assignment[17] and one in which it was unclear if there were statistically significant differences between the intervention and control groups.[18] The 1 component lacking in all but 3 of the included studies was blinding procedures for interventionists, which is common in educational interventions.[19–21] Thus, the overall quality of these studies did not indicate a significant risk of bias, given their intervention design limitations, and all studies were deemed to be of sufficient quality to be included.


For this review, the 23 studies were organized by their mode of communication into 6 groups: Early Palliative Care, Clinician Communication Training, Patient Communication Support, Family Communication Support, Patient-Clinician Communication, and Patient Decision Aids. The included studies generally lacked grounding in theoretical frameworks. Clinical, demographic, and geographic characteristics of the 23 included studies are summarized in Supplemental Digital Content 1 (see Supplemental Digital Content 1, There was a large overall patient, clinician, and family sample size in this review, N = 6376 with a range from N = 38 to N = 537 and an average of N = 277 subjects. Demographically, 21 of the 23 studies were located in urban or suburban areas with predominantly white, relatively well-educated patients, with average age of 63.9 years. Rural and nonwhite populations were underrepresented. For the 19 studies that included gender statistics, 46% were female. The average age of clinicians was 38.3 years, 48% of whom were female, and family decision-makers' average age was 56.2, 70% of whom were female. Geographically, the included studies emphasized American health care systems, with 19 conducted in the United States, 2 in Canada, 1 in Germany, and 1 in Australia. Clinically, 13 of the 23 studies were conducted in academic acute care facilities, 8 in outpatient care clinics, and 2 in long-term-care facilities.

The studies' interventions, measurement, outcomes, and results are summarized in Supplemental Digital Content 2 (see Supplemental Digital Content 2, Nineteen of the interventions were externally sourced, developed by the study's researchers from their proprietary designs. Most of the included studies' measurement periods were of short duration, averaging 6.7 weeks, with a range of 2 to 26 weeks. Intervention consultations were also brief, averaging 45 minutes, and 13 of 23 studies provided only 1 or 2 intervention sessions for clinicians, patients, or family to communicate. Adherence, reported in all studies except for 1 retrospective study,[22] appeared to be adequate at greater than 80% in 14 studies, with a range of 42% to 100%. Most of the studies' attrition was attributable to patient deaths or transition out of the study's facility. There were some notable elements to the different modalities of GOC communication. Early Palliative Care and Clinician Communication Training modes both used complex, multistep in-person or telehealth consultation interventions. Further, interventions for 16 of 17 studies relied on multiple in-person meetings facilitated by clinicians. By contrast, all 6 Patient Decision Aid studies' interventions used short, standardized advance care planning education videos to clarify patients' GOCs with minimal clinician involvement. All Early Palliative Care and Clinician Communication Training and Family Communication Support studies' outcomes were assessed by externally validated instruments, but only 4 of the 16 remaining studies used externally validated instruments. The studies' measurement heterogeneity and dearth of validated instruments in the majority of the studies precludes meaningful meta-analysis of their results and could impact future efforts to replicate these studies.

With respect to reporting significant results for their primary outcomes, certain modes produced consistent results. All of the studies in both Patient Decision Aids and Patient-Clinician Communication modes produced significant results for their primary outcomes. Patient Decision Aids increased patients' GOC knowledge and documentation between (P = .03) and (P < .0001).[17,21,23–26] Patient-Clinician Communication improved the frequency and quality of GOC discussions between (P = .03) and (P < .001).[27–29] Conversely, only 1 of 4 Early Palliative Care studies reported significant results for only 1 of multiple primary outcomes.[19,30–32] Other GOC communication modes have mixed results. Two of 3 Clinician Communication Training studies reported significant primary outcomes results, with P = .01[18] and P < .0001.[33] Two of 4 Patient Communication Support studies had significant primary outcomes results, increasing GOC documentation and patients' quality of life, with P < .001[34] and P = .02 and P = .04.[35] Two of the 3 Family Communication Support studies reported significant primary outcomes results, improving family satisfaction with patients' EOL care (P = .02)[36] and GOC concordance between family and patient dyads (P = .04).[37] We note that only 4 of the 8 intervention studies that used externally validated instruments achieved their primary outcomes.[32,35–37]