Examining Interventions Designed to Support Shared Decision Making and Subsequent Patient Outcomes in Palliative Care: A Systematic Review of the Literature

Abstract

Background:

Shared decision making (SDM) is a key attribute of patient-centered care, which empowers palliative care patients to be able to make optimal medical decisions about end-of-life treatments based on their own values and preferences.

Aim:

The aim of this systematic literature review is to detail and compare interventions supporting SDM over the last 10 years (January 2008 to December 2017) and to analyze patient/caregiver outcomes at the end of life.

Methods:

This review followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. PubMed, CINAHL, Embase, and Cochrane Library were searched with key search terms: SDM, decision aid, decision support, palliative care, and hospice care. The Mixed Methods Appraisal Tool was used to assess the quality of the included studies.

Results:

The initial search yielded 2705 articles, and 12 studies were included in the final review. The quality of the studies was modest and technology-enabled delivery modes (e.g., video, DVD, web-based tool) were most commonly used. Patient/caregiver knowledge of end-of-life care was the most common primary outcome across studies. The strength of the association between the SDM interventions and patient/caregiver outcomes varied.

Conclusion:

The findings from the studies that examined the effects of the SDM intervention on patient outcomes were inconsistent, highlighting the need for further SDM intervention studies among diverse patient populations using consistent measures. Given the availability of health technologies, future studies should focus on developing individual-tailored, technology-enabled interventions to support patient-centered medical decision making.

Introduction

Patient engagement in health care is essential for patient-centered care planning and medical decision making. Shared decision making (SDM) is a collaborative clinical decision-making process in which patients/caregivers and healthcare professionals create healthcare plans together by discussing the goals of care and medical treatments.¹ Charles and colleagues² illustrated the process of SDM in health care as follows: patients/caregivers and healthcare professionals are invited to engage in the process. Both parties exchange information; healthcare professionals provide care treatment options and explain their benefits and risks; and the patients/caregivers express their values and preferences about care plans and treatments. An optimal decision can be reached through mutual agreement.² The importance of SDM was highlighted as a vital component of patient-centered care by the Institute of Medicine report Crossing the Quality Chasm.³

Decision making at the end-of-life is affected by individual patient values, cultures, and preferences; thus, SDM is critical for palliative care patients in managing end-of-life care.^4,5 Palliative care patients and their caregivers experience complex medical care options and high emotional distress in choosing between life-prolonging treatments with side effects and improving the quality of life with less aggressive treatments.⁴ To support informed decision making about end-of-life options, decision aids or interventions that support SDM can help patients and caregivers have more knowledge about hospice care and medical treatments and effectively communicate with healthcare professionals.^1,6

Recent literature reviews have synthesized research on the effects of SDM interventions on patient outcomes of specific diagnosed population samples (e.g., type 2 diabetes⁶ and cancer⁷) and on health inequalities.⁸ Belanger and colleagues⁴ synthesized knowledge about the SDM process in palliative care using thematic analysis, and Hajizadeh and colleagues⁵ compared SDM with other forms of decision making processes including 7 articles for the final review. However, little research has been conducted to analyze interventions designed to support SDM and synthesize the relationship between the SDM interventions and the patient/caregiver outcomes at the end of life. Specifically, there is a lack of understanding of how information is being delivered and shared with palliative patients and caregivers. The aims of this systematic literature review are (1) to detail and compare SDM interventions at the end of life over the last 10 years (January 2008 to December, 2017) and (2) to examine patient/caregiver outcomes following the SDM interventions. We selected this time interval to capture the most recent trends in SDM intervention studies.

Methods

This systematic literature review followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.⁹

Search Strategy and Selection Criteria

A search of the literature was conducted using 4 electronic databases: PubMed, CINAHL, Embase, and Cochrane Library. Additional searches were run using reference lists obtained from literature relevant to this review. The search strategy was developed though a literature review and consultation with an informationist in the health sciences library at Columbia University. The searches used MeSH terms for PubMed, CINAHL headings for CINAHL, Emtree terms for Embase, and free-text terms. Boolean searching techniques using “AND” and “OR” included the following search terms: “shared decision making,” “decision aid,” “decision support,” “palliative care,” and “hospice care.” More detailed search terms used for each database are outlined in Appendix A.

Studies were included if they met the following criteria: (1) published between January 2008 and December 2017, (2) written in English, (3) original research study with data, (4) studies in which all patients were receiving palliative care, and (5) studies where an intervention to support SDM was conducted. Studies were excluded if they (1) provided discussion, opinion, commentary, review, editorial, or protocol only, (2) were a published conference abstract only or presentation slides, (3) involved patients who were less than 18 years old, or (4) examined the SDM intervention for only healthcare professionals without patients/caregivers.

Data Extraction and Analysis

Titles and abstracts were independently reviewed by 2 reviewers (D.B. and H.C.) to identify eligible studies. The 2 reviewers met and compared opinions on included and excluded articles, and consensus was reached through discussion when there were different opinions. Full-text articles were independently reviewed (D.B. and H.C.). To facilitate the systematic review process, the web-based literature review tool, Covidence,¹⁰ was used. The following data were extracted in the final review and documented in an Excel spreadsheet: (1) author, (2) year of publication, (3) years of data collection, (4) country, (5) study design, (6) study setting, (7) patient characteristics (number of participants, age, gender, race/ethnicity, diagnosis), (8) interventions to support SDM, and (9) patient/caregiver outcomes measured. The extracted data were analyzed and synthesized using descriptive statistics.

Quality Assessment

The Mixed Methods Appraisal Tool (MMAT)¹¹ was used to assess the quality of the included studies because it is validated and designed to systematically assess the rigor of various types of study designs in the included studies. The MMAT provides scores ranging from 0% to 100% in 25 percent-point intervals (0%, 25%, 75%, and 100% scores). The higher the percentage score, the higher the study quality. When the included studies did not report specific types of study designs, these studies were deemed a quantitative nonrandomized design based on the guideline of the MMAT.¹¹ This is because this review included only intervention studies that examined the effects of SDM interventions on psychological and behavioral outcomes of patients/caregivers and/or healthcare professionals. Two reviewers (D.B. and H.C.) independently assessed the quality of the included studies. Consensus was reached through discussion when there were discrepancies.

Results

Search Results

A flow diagram illustrating the review process is shown in Figure 1. The initial search yielded 2705 articles from 4 electronic databases and additional searches using reference lists. After screening titles and abstracts and reviewing full-text articles, 12 studies were included for the final review.

Figure 1.

PRISMA flow diagram.

Study Characteristics

The characteristics of the included studies are presented in Table 1. Of the 12 included studies, 7 were experimental studies (6 randomized controlled trials [RCTs] and 1 quasiexperimental study), 3 were observational studies, and 2 studies did not report a specific type of study design. Studies were conducted in the United States (n = 11) and Malaysia (n = 1). Research participants were recruited at hospitals (e.g., inpatient, outpatient, or both; not report specifically) (n = 9), nursing homes (n = 2), and home-based care settings (n = 1).

Table 1.

Characteristics of Included Studies.

Author, Year	Study Design, Setting, Years of Data Collection	Participants (Age, Gender, Race, Primary Diagnosis)	N	Intervention	Outcome Measures	Quality Score
Benton  et al20	Retrospective cohort study,  Medical Center of Central  Georgia in the United States;  January, 2008 ~ December,  2010	Age: 18-65: 934 (32.9%), > 65: 1786 (62.8%),  Unknown: 123 (4.3%); Gender: female:  1,473 (51.8%), male: 1,362 (47.9%),  Unknown: 8 (0.3%); Race: White: 1,432  (50.4%), Black: 996 (35.0%), Unknown: 415  (14.6%)  Diagnosis: Cancer: 1,958 (68.9%), Non-  cancer: 609 (21.4%), Unknown: 276 (9.7%)	2843	Interprofessional Team  Palliative Care Consultation	End-of-life treatment  preferences (Full code, DNR,  and care and comfort only)	50%
Einterz  et al12	One group pretest–posttest  design  2 nursing homes in the United  States  May, 2011 ~ April, 2012	Surrogate decision maker (n = 18); Mean age:  67; Gender: female (10, 56%); Race: White  (18, 100%); Relationship to Resident: spouse  (3, 17%), daughter (6, 33%), son (6, 33%),  other relatives (3, 17%); Nursing home  residents (n = 18); Mean age: 90; Gender:  female (15, 83%); Race: White (18, 100%);  Diagnosis: Dementia	36	Video decision aid, print  material, and a structured  care plan meeting between  surrogate and  interdisciplinary care team	Surrogate knowledge, quality of  communication, surrogate-  provider concordance on  primary goals of care, number  of palliative care domains in  the care plan, frequency of  communication regarding goals of care, satisfaction with  care, resident quality-of-life,  symptom management	75%
El-Jawahri  et al13	RCT; Outpatient oncology  clinics at Massachusetts  General Hospital in the  United States; July, 2008 ~  March, 2009	Intervention group (n = 23); Mean age: 56;  Gender: female (9, 39.1%); Race: White (22,  95.7%); Control group (n = 27)  Mean age: 51; Gender: female (13, 88.9%);  Race: White (24, 96%); Diagnosis: Cancer  (malignant glioma)	50	Video decision aid (after  verbal narrative)  vs.  Verbal narrative alone	Knowledge, goals-of-care  preferences (life-prolong-  care, basic medical care,  comfort care), preferences  for cardiopulmonary  resuscitation (CPR), comfort  with video decision aid	100%
Green  et al14	RCT; The Penn State Milton S.;  Hershey Medical Center in  the United States; 2007 ~  2012	Intervention group (n = 99); Mean age: 60.6;  Gender: female (41, 41.1%); Race: White  (94, 95.9%); Control group (n = 101); Mean  age: 59.6; Gender: female (40, 39.6%); Race:  White (96, 96%); Diagnosis; Advanced  cancer (including hematology/oncology,  radiation oncology, surgical oncology,  otolaryngology, neurosurgery, and breast  cancer clinic)	200	Online decision aid, Making your  wishes known: planning your  medical Future  vs.  Educational materials	Hope, hopelessness, anxiety,  knowledge of advance care  planning, self-determination,  satisfaction with advance care  planning processes	75%
Hanson  et al15	RCT; 22 nursing homes in  North Carolina in the United  States; April, 2012 ~  September, 2014	Residents (n = 302); Intervention group (n =  151): Age: 85.7; Gender: female (121,  80.1%) Race: White (134, 88.7%); Control  group (n = 151): Age: 87.3; Gender: female  (125, 82.8%); Race: White (123, 82.0%);  Diagnosis: Advance dementia; Family  decision makers (n = 302); Intervention  group (n = 151): age: 62.4; Gender: female  (104, 68.9%); Race: White (136, 90.1%);  Relationship to patient: daughters or  daughters-in-law (76, 50.3%)  Control group (n = 151): age: 63.4; Gender:  female (100, 66.2%)  Race: White (125, 83.3%) Relationship to  patient: daughters or daughters-in-law (85,  56.3%)	302 Resident-  family  dyads	Video decision aid, print  material, and discussion with  the nursing home care team  vs. Informational video on  interaction with someone  with dementia and a usual  care plan meeting with staff	Quality of communication,  concordance with clinicians  on goals of care, use of  resident preferences to guide  treatment, symptom management, satisfaction  with care, goals of care  documentation, treatment  preferences	100%
Hing  Wong  et al19	Cross-sectional study  Regular hemodialysis unit at  University Malaya Medical  Center in Malaysia  August, 2014 ~ November,  2014	Mean age: 59.5; Gender: male (32, 51.7%),  female (24, 42.9%); Race: Malays (39, 69.6%);  Diagnosis: End-stage renal disease	58	Educational brochure on  advance care planning  decisions and CPR	Knowledge, attitude toward  advance care planning and  end-of-life issues	25%
Machare  Delgado  et al23	Not reported  Medical ICU of Thomas  Jefferson University Hospital  in Philadelphia, Pennsylvania  in the United States; April  2008 ~ May 2008	Mean age: 54; Gender: male (17, 59%); Race:  not reported  Diagnoses: pneumonia (18, 62%), septicemia  (4, 14%), pulmonary embolism (3, 10%),  upper gastrointestinal bleeding (2, 7%), drug  overdose (1, 3%), angioedema (1, 3%)	29	Structured multidisciplinary  family meeting (MDFM)  program	Number of MDFM, rate of  presence of palliative care  team, quality of  communication, knowledge	25%
Matlock  et al16	RCT; Inpatient palliative care;  service at University of  Colorado Hospital in the  United States; December  2009 ~ May 2010	Patients: Intervention group (n = 9): mean age:  54; Gender: female (7, 78%); Race: White (7,  78%); Diagnoses: not reported; Control  group (n = 11): mean age: 56; gender:  female (6, 55%); Race: White (11, 100%);  Diagnoses: not reported; Decision makers;  Intervention group n = 16): mean age: 57;  gender: female (7, 44%); Race: White (11,  69%); Relationship to patient: not reported;  Control group (n = 15): mean age: 56;  Gender: female (7, 47%); Race: White (14,  93%); Relationship to patient: not reported	51 (Patient:  20; Decision  maker:  31)	Looking ahead: choices for  medical care when you’re  seriously ill decision aid  (booklet and DVD)  vs.  Usual care from palliative care  service	Knowledge, decision conflict	50%
Nakagawa  et al21	Prospective descriptive study  Single medical center in the  United States  January, 2014 ~ September,  2016	Mean age: 57.5; Gender: male (90, 80.3%)  Race: not reported  Diagnosis: End-stage heart failure	112	Palliative care consultation	Knowledge, family awareness of  patients’ unacceptable health  state	75%
Radwany  et al22	RCT; Home-based care settings  in Ohio’s community-based  long-term care Medicaid  waiver program in the United  States; Not reported	Intervention group (n = 40); Mean age: 69.5;  Gender: female (29, 72.5%); Race: White  (34, 85%); Control group (n = 40); Mean  age: 68.8; Gender: female (31, 77.5%); Race:  White (34, 85%); Diagnosis; Congestive  heart failure (American Heart Association  Stage C), chronic obstructive pulmonary  disease (COPD), diabetes with renal  disease, neuropathy, visual problems, or  coronary artery disease, end-stage liver  disease or cirrhosis; cancer except skin  cancer; renal disease and actively receiving  dialysis; amyotrophic lateral sclerosis with  history of aspiration; Parkinson’s disease  stages 3 and 4; or pulmonary hypertension	80	In-home palliative care  interdisciplinary consultation  vs.  Usual care	Symptom management, quality  of life, anxiety,  depression, decision making/care  planning, spirituality,  healthcare utilization (emergency room visits,  hospital visits)	50%
Van Scony  et al17	Not reported; Cardiology and  pulmonology  subspecialty clinics of a  university hospital in the  United States  2009 ~ 2011	Mean age: 66; Gender: male (36, 73.5%); Race:  White (48, 98%); Diagnosis: COPD (25,  51%), heart failure (24, 49%)	49	Online decision aid: Making your  wishes known: planning your  medical future	Knowledge of advance care  planning, satisfaction with  advance care planning,  satisfaction with decision,  decisional conflict	50%
Vogel  et al18	RCT; University of Minnesota’s  Gynecologic Oncology clinic  in the United States; Not  reported	Intervention group (n = 20); Mean age: 59.6;  Gender: female (100%); Race: White (19,  95%); Control group (n = 15); Mean age:  55.5; Gender: female (100%); Race: White  (11, 78.6%); Diagnosis; Ovarian cancer	35	Web-based tool vs. general  website including usual care  information	Completion of advance  healthcare directive,  decisional conflict	50%

Abbreviations: CPR, cardiopulmonary resuscitation; MDFM, multidisciplinary family meeting.

Participant Characteristics

The participant characteristics of the included studies are illustrated in Table 1. The number of participants ranged from 29 to 2843, totaling 3845 participants. Research participants were generally older adults (mean age: 54-90 years old), the majority were caucasian (50%-100%), and the proportion of females ranged from 39% to 100%. Of the 11 studies that reported the primary diagnosis of patients, 4 recruited patients with mixed diagnoses (e.g., cancer, pneumonia, heart failure, and chronic obstructive pulmonary disease), followed by studies that recruited patients with cancer (n = 3), dementia (n = 2), heart failure (n = 1), and end-stage renal disease (n = 1). The 2 studies that recruited patients with dementia also included family decision makers as study participants who were generally female (56%-69%) and caucasian (83%-100%).

Study Interventions to Support SDM

Table 1 also lists the study interventions. We classified the interventions facilitating SDM in palliative care into 4 main categories: technology-enabled delivery modes (e.g., video, DVD, web-based tool; n = 7),^12-18 print materials (n = 4),^12,15,16,19 palliative care consultation (n = 3),^20-22 and structured meeting (n = 3).^12,15,23 Nine studies used a single delivery mode, while 3 conducted mixed delivery modes; of the latter, 2 conducted the same interventions including the Goals of Care video decision aid, print materials, and a structured meeting between patients/caregivers and interprofessional care team. The remaining study used a booklet and a DVD to support SDM for palliative care patients. We briefly describe the interventions with the 4 main categories in the following section.

Technology-enabled delivery modes: video, DVD, and web-based tool.

The 7 studies that used the technology-enabled delivery modes included a video (n = 3), web-based tool (n = 3), and DVD (n = 1). Among the studies that employed a video decision aid to show patients visual images about the goals of care,^12,13,15 2 used an 18-minute video including content about treatment goals of advanced dementia, treatment options, and role of the surrogate decision maker.^12,15 The other study used a 6-minute video to help palliative care patients understand 3 levels of medical care including life-prolonging care, basic medical care, and comfort care.¹³ Among studies that used a web-based tool to provide more accessible and user-friendly decision aid,^14,17,18 2 implemented the same online decision aid, Making Your Wishes Known: Planning Your Medical Future.^14,17 This tool provided information about common medical conditions and medical treatments using audio and video to help palliative care patients prioritize their own values and goals. The third study that used a website with social medial features included components about distress monitoring, frequently asked questions to healthcare professionals, goals of care options, and an information library to provide tailored and comprehensive information based on ovarian cancer stage.¹⁸ One study that used a DVD focused on global issues about the importance of advance care planning and clarifying patients’ goals and values.¹⁶

Print materials.

Of the 4 studies that used print materials,^12,15,16,19 3 implemented mixed delivery modes such as video, DVD, and structured meeting.^12,15,16 They also used the print materials to reinforce palliative care content provided in video or DVD. The one study that employed a single delivery mode used an educational brochure that provided information on advanced care planning plans and cardiopulmonary resuscitation decisions.¹⁹

Palliative care consultation.

Among 3 studies that used palliative care consultation,^20-22 2 were conducted in the hospital^20,21 and 1 was implemented in a home-based setting.²² Benton and colleagues conducted a counselor-led consultation for patients and their families.²⁰ The counselors facilitated patients and their families to make medical decisions on advanced care planning by providing communication strategies and helped them document advance directives by visiting each patient and family 2 to 4 times and spending 30 minutes per visit. The counselors met with an interprofessional care team (nurse practitioner, physician assistant, clinical nurse specialist, and palliative physician) every weekday during a huddle to discuss the patient’s care plans and preferences. Nakagawa and colleagues conducted palliative care consultation for all patients with heart failure before left ventricular assist device implantation performed by a palliative care physician or a palliative care nurse practitioner.²¹ The palliative care team consisted of palliative care physicians, nurse practitioners, social workers, and chaplains and was involved in assisting in decision making and providing psychosocial support. The palliative care consultation was conducted with individual-tailored information after pre-evaluation including the content about patient comfort, patient and family understanding of left ventricular assist device therapy, patient goals and expectations, spiritual needs, possible complications and unacceptable conditions, and decision making and information sharing preferences. Radwany and colleagues conducted an in-home, interdisciplinary, care management delivered by care managers who coordinate the patient’s care with a palliative interprofessional care team and the patient’s primary care physician.²² The core interprofessional team consisted of a palliative medicine specialist, a geriatrician, a care manager, a palliative care nurse specialist, a social worker, a spiritual advisor, and a pharmacist. After the team develops individual-tailored care plans based on standardized protocols, the care manager discussed the care plans with the patient and their family during a home visit. Once they reach agreement about the end-of-life care plan, the care manager implemented the care plans by providing information about symptom management, access to community resources, psychological support, assistance with spiritual needs, facilitating communication with healthcare professionals, and completing legal documents.

Structured meeting.

Among the 3 that conducted a structured meeting with patient/caregivers,^12,15,23 2 were conducted in nursing homes using the same structured meeting format for patients with dementia, their family, and interprofessional care teams (physicians, nurse practitioners, physician assistants, nurses, social workers, therapists, and nutritionists).^12,15 In the structured meeting, healthcare professionals discussed goals of care with patients and their caregivers using the VALUE framework (Value family comments, Address emotions, Listen, Understand the patient as a person, and Elicit family questions). Before the discussion, healthcare professionals received a 1-hour training session that provided the VALUE principles and a short role-play about goals of care discussion.¹⁵ Machare Delgado and colleagues conducted a structured multidisciplinary family meeting program in the medical ICU.²³ The multidisciplinary team consisted of a critical care attending physician, fellow and resident, ICU nurse, clinical nurse specialist, members of the palliative care team, social worker, and pastoral care provider. The team members first asked the patient’s family about patient diagnosis, prognosis, and goals of care to assess the level of understanding; then healthcare professionals provided information about treatment options, symptom management, and spiritual and emotional support.

Patient/Caregiver Outcomes

Patient/caregiver knowledge.

Patient and/or caregiver knowledge was examined in 8 studies using quantitative methods^{12-14,16,17,19,21} and qualitative comments.²³ Seven studies used quantitative methods that utilized heterogeneous measurements of knowledge and assessed various knowledge domains related to the end-of-life decision making such as goals of care, advanced care planning, and treatment options of each diagnosis. Among the 7 studies that used quantitative methods, 5 reported significant improvements in the level of patient/caregiver knowledge after SDM interventions,^12-14,17,21 while 2 reported insignificant improvements in knowledge score between the pre- and post-test.^16,19 Specifically, the 5 studies that reported significant impact used video decision aids, web-based tool, palliative care consultation, and mixed delivery modes (video decision aid, print material, and structured meeting). The one study that analyzed qualitative comments from the structured multidisciplinary family meeting reported the family caregiver’s improved understanding of medical conditions and treatment plans.²³

Patient/caregiver satisfaction.

Patient/caregiver satisfaction was assessed in 4 studies^12,14,15,17 which investigated a wide range of dimension of satisfaction including surrogate satisfaction with care^12,15 and satisfaction with advanced care planning process.^14,17 Two studies that examined satisfaction with the advanced care planning process used the same instrument consisting of 12 questions with 4 domains: (1)quality of information provided, (2) helpfulness in values clarification, (3) helpfulness in decision making, and (4) helpfulness in communicating wishes.^14,17 Of the 4 studies, one that used a web-based tool reported a significant change in satisfaction with the advanced care planning process in the intervention group,¹⁴ while 3 that used 3 mixed delivery modes (video decision aid, print material, and structured meeting),^12,15 and a web-based tool¹⁷ found no significant improvement in satisfaction.

Decision quality.

Three studies examined decision quality using the Decisional Conflict Sale^16-18 that measures decision quality employing a validated scale that assesses patients’ decision confidence and satisfaction regarding the end-of-life decision making. Three studies that conducted mixed delivery modes (booklet and DVD)¹⁶ and a web-based tool^17,18 reported that decision confidence in a post-test¹⁷ or among an intervention group^16,18 was higher than a pre-test or a control group; however, there was no statistically significant improvement in decision quality.

Quality of communication.

The quality of communication with healthcare professionals was examined by 3 studies^12,15,23 using quantitative methods^12,15 and qualitative comments.²³ Two studies used the Quality of Communication scores as a valid and reliable tool,^12,15 one of which also measured the frequency of communication regarding the Quality of Communication scores to assess additional quality of communication.¹² The 2 studies conducted the same interventions (combination of video decision aid, print material, and structured meeting) and reported that patients/caregivers had significant improvements in the quality of communication with nursing home healthcare professionals.^12,15 The remaining study analyzed qualitative comments from multidisciplinary family meetings and reported improved communication between an interprofessional team and family caregivers following the meetings.²³

Quality of life.

Quality of life was examined in 1 pre-and post-test study¹² and 1 RCT study.²² The pre-and post-test study conducted mixed delivery modes (video decision aid, print material, and structured meeting) and used the Quality of Life in Late-Stage Dementia scale that measures surrogates’ perceptions regarding the quality of life of patients with advanced dementia.¹² The RCT study was conducted via in-home palliative care consultation and used the Quality at the End of Life Scale that measures self-reported quality of life for patients with mixed diagnoses.²² The 2 studies reported that there was no change in the quality of life of patients following the SDM interventions.

Behavioral outcomes.

Patients’ behavioral outcomes following the SDM interventions were assessed in 3 studies^15,18,22 using electronic medical records. Hanson and colleagues conducted mixed delivery modes (video decision aid, print material, and structured meeting) and examined the level of completion of goals of care documentation reporting a significant increase in documenting goals of care. Vogel and colleagues¹⁸ examined the effects of a web-based tool on the completion of an advanced healthcare directive but no significant difference in completion of an advanced healthcare directive between intervention and control groups. Radwany and colleagues²² examined the level of healthcare utilization at 12 months following the in-home palliative care interdisciplinary consultation. The authors reported that the intervention group had fewer hospital visits and fewer nursing facility admissions; however, the differences between the 2 groups were not statistically significant.

Quality Appraisal

The quality of the included studies was moderate with an average score of 60% of 100% (Table 1). The detailed quality assessment of each study is represented in Appendix B. The major methodological limitation in the RCT studies was that the allocation concealment or blinding was not clearly reported.^14,16,18 The remaining 6 studies had weaknesses related to risk for selection bias,^{12,17,19-21,23} low response rates and high missing data,^17,19,20,23 and unclear description of measurements used.^19,23

Discussion

This review provides an overview of interventions to support SDM and their effects on specific outcomes of patients and their caregivers. Our study findings confirm that technology-enabled delivery modes were used more frequently to support SDM and that patient/caregiver knowledge was measured most often as an outcome.

There is relatively limited evidence on the use of technology-enabled delivery modes in palliative care compared to other fields of health care.²⁴ Although more than half of the studies included in this review used technology-enabled delivery modes to facilitate SDM, the interventions of the studies were limited to use of video, DVD, and web-based tools. One weakness of the studies was that global, rather than individualized, technology-based tools were used for setting goals of care and clarifying patients’ preferences.

Although healthcare professionals play an important role in facilitating the SDM process by building trusting relationships and interacting effectively with patients/caregivers, a strategic approach to improving the skills and competencies of healthcare professionals in palliative care SDM has not received attention. Only 1 study added the interventions for healthcare professionals to help facilitate the implementation of the interventions to support SDM.¹⁵ In addition, the important role of social workers in the interprofessional palliative care team is underrepresented in the studies included in this review.

Across studies, there were inconsistent findings about whether the SDM intervention improved patient/caregiver outcomes. There could be several possible reasons for this. First, the included studies did not verify whether the duration and frequencies of the SDM interventions were sufficient to achieve psychological or behavioral changes nor did the studies provide the rationale for the dosage or frequency of SDM interventions for patients/caregivers across studies. Moreover, the studies did not explore whether the content provided in the SDM interventions was effectively delivered and evaluated using appropriate measurements. Finally, although diagnosis-based care trajectories and individual care plans/treatments are different, the majority of studies included in this review recruited patients with mixed diagnoses.

Limitations

This review is limited by only including studies written in English. By nature of the included study population, this review is also limited by the lack of diversity in the patient populations represented in the studies. There were methodological limitations of the included studies, including a lack of reporting reliability and validity of study instruments among the sample studied. Moreover, this review was unable to evaluate the effects of SDM interventions across the included studies using a meta-analysis due to the heterogeneity of study designs, SDM interventions, statistical methods, and patient/caregiver outcomes measured.

Future Directions

From this review, we have identified a number of areas for further research, including evaluating the impact of technology-enabled delivery modes available to improve SDM across both rural and urban geographic locations. Future studies should develop and test various technology-enabled delivery modes to determine which are most effective within specific patient populations based on disease trajectories and patient/caregiver preferences. Further research should also examine SDM interventions for both patients/caregivers and healthcare professionals and investigate psychological and behavioral outcomes of both groups to make optimal decisions about end-of-life treatments. Finally, SDM intervention studies also need to employ consistent measures across diverse patient populations.

Conclusion

Optimal decision making and end-of-life planning can be enhanced by SDM between patients and healthcare professionals. Given the potential of health technology to improve patient care, future studies are needed to explore and develop individual-tailored technology-enabled interventions to support patient-centered medical decision making.

Appendix A

Search Terms and Structure

PubMed	CINAHL	Embase	Cochrane Library
#1 “shared decision making”[tiab] OR  “shared decision making”[tiab]  OR “shared decision  support”[tiab] OR “decision  making”[mesh] OR “decision  making”[tiab] OR “decision  support techniques” [mesh] OR  “decision making”[tiab] OR  “decision aid*”[tiab] OR  “decision support”[tiab] OR  “counseling”[mesh] OR  “counseling”[tiab] OR  “coaching”[tiab] OR “patient  participation” [mesh] OR  “patient participation” [tiab] OR  “patient centered care” [mesh]  OR “patient centered care”  [tiab] OR “patient engagement”  [tiab] OR “patient-provider  communication” [tiab] OR  “patient preference”[mesh] OR  “patient preference*”[tiab]	MH “decision making, patient”  OR MH “decision making,  family” OR MH “decision  making+” OR MH “decision  support techniques” OR MH  “patient centered care” OR TI  “shared decision making” OR  AB “shared decision making”  OR TI “decision making” OR  AB “decision making” OR MH “counseling” OR TI  “counseling” OR AB  “counseling” OR TI  “coaching” OR AB “coaching”  OR TI “patient-provider  communication” OR AB  “patient-provider  communication” OR TI  “decision aid*” OR AB  “decision aid*” OR TI  “decision support” OR AB  “decision support” OR TI  “patient participation” OR AB	“shared decision making”/exp  OR “shared decision making”:  ti, ab OR “shared decision making”:  ti, ab OR “decision  making”/exp OR “decision  making”: ti, ab OR “decision  making”: ti, ab OR “decision  aid”/exp OR “decision aid*”:  ti, ab OR “decision support  system”/exp OR “decision  support system”: ti, ab OR  “decision support”: ti, ab OR  “counseling”/exp OR  “counseling”: ti, ab OR  “coaching”: ti, ab OR “patient  centered care”: ti, ab OR  “patient participation”/exp  OR “patient participation”: ti,  ab OR “patient engagement”/  exp OR “patient engagement”: ti, ab OR  “patient preference”/exp OR  “patient preference”: ti, ab	“shared decision making”: ti, ab  OR “shared decision making”:  ti, ab OR “decision making”: ti,  ab OR “decision making”: ti,  ab OR “patient-provider  communication”: ti, ab OR  “decision aid*”: ti, ab OR  “decision support”: ti, ab OR  “counseling”: ti, ab OR  “coaching”: ti, ab OR  “patient centered care”: ti, ab  OR “patient participation”: ti,  ab OR “patient engagement”: ti,  ab OR “patient preference*”:  ti, ab
	“patient participation” OR TI  “patient centered care” OR  AB “patient centered care”  OR TI “patient engagement”  OR AB “patient engagement”  OR TI “patient preference”  OR AB “patient preference”	OR “patient-provider communication”: ti, ab
#2 “palliative care”[mesh] OR  “palliative care”[tiab] OR  “hospice care”[mesh] OR  “hospice*”[tiab] OR “end of  life”[tiab] OR “end-of-life”[tiab]  OR “incurable”[tiab] OR  “terminally ill”[mesh] OR terminally III[tiab] OR “terminal  illness*” [tiab]	MH “palliative care” OR MH  “hospice care” OR MH  “terminally ill patients” OR TI  “palliative care” OR AB  “palliative care” OR TI  “hospice*” OR AB “hospice*”  OR TI “end of life” OR AB  “end of life” OR TI “incurable”  OR AB “incurable” OR TI  “terminally ill patient*” OR AB  “terminally ill patient*” OR TI  “terminally ill” OR AB “terminally ill” OR TI  “terminal illness*” OR AB  “terminal illness*”	“hospice care”/exp OR  “hospice”/exp OR “hospice  care”: ti, ab OR “end of life”/  exp OR “end of life”: ti, ab OR  “terminally ill patient”/exp OR  “terminally ill patient”: ti, ab  OR “palliative care”: ti, ab OR  “end of life”: ti, ab OR  “incurable”: ti, ab OR  “terminally ill”: ti, ab OR  “terminal illness*”: ti, ab	“palliative care”: ti, ab OR  “hospice care”: ti, ab OR  “hospices”: ti, ab OR “end of  life”: ti, ab OR “incurable”: ti,  ab OR “terminally ill”: ti, ab  OR “terminal illness*”: ti, ab
#3. #1 AND #2 #4. Limit #3 to English language #5. Limit #4 to 2008-2017

Appendix B

Quality Appraisal of All Included Studies

Author  Year	Components	Quality Criteria	Yes	No	Cannot  tell
Benton et al20	Quantitative non-  randomized	3.1. Are participants (organizations) recruited in a way that minimizes selection  bias?			x
		3.2. Are measurements appropriate (clear origin, or validity known, or standard  instrument; and absence of contamination between groups when appropriate)  regarding the exposure/intervention and outcomes?	x
		3.3. In the groups being compared (exposed vs. non-exposed; with intervention  vs. without; cases vs. controls), are the participants comparable, or do  researchers take into account (control for) the difference between these  groups?	x
		3.4. Are there complete outcome data (80% or above), and, when applicable, an  acceptable response rate (60% or above), or an acceptable follow-up rate for  cohort studies (depending on the duration of follow-up)?			x
Einterz et al12	Quantitative non-  randomized	3.1. Are participants (organizations) recruited in a way that minimizes selection  bias?			x
		3.2. Are measurements appropriate (clear origin, or validity known, or standard  instrument; and absence of contamination between groups when appropriate)  regarding the exposure/intervention and outcomes?	x
		3.3. In the groups being compared (exposed vs. non-exposed; with intervention  vs. without; cases vs. controls), are the participants comparable, or do  researchers take into account (control for) the difference between these  groups?	x
		3.4. Are there complete outcome data (80% or above), and, when applicable, an  acceptable response rate (60% or above), or an acceptable follow-up rate for  cohort studies (depending on the duration of follow-up)?	x
El-Jawahri  et al13	Quantitative  randomized  controlled (trials)	2.1. Is there a clear description of the randomization (or an appropriate sequence  generation)?	x
		2.2. Is there a clear description of the allocation concealment (or blinding when  applicable)?	x
		2.3. Are there complete outcome data (80% or above)?	x
		2.4. Is there low withdrawal/drop-out (below 20%)?	x
Green et al14	Quantitative  randomized  controlled (trials)	2.1. Is there a clear description of the randomization (or an appropriate sequence  generation)?	x
		2.2. Is there a clear description of the allocation concealment (or blinding when  applicable)?		x
		2.3. Are there complete outcome data (80% or above)?	x
		2.4. Is there low withdrawal/drop-out (below 20%)?	x
Hanson et al15	Quantitative  randomized  controlled (trials)	2.1. Is there a clear description of the randomization (or an appropriate sequence  generation)?	x
		2.2. Is there a clear description of the allocation concealment (or blinding when  applicable)?		x
		2.3. Are there complete outcome data (80% or above)?	x
		2.4. Is there low withdrawal/drop-out (below 20%)?	x
Hing Wong  et al19	Quantitative non-  randomized	3.1. Are participants (organizations) recruited in a way that minimizes selection  bias?		x
		3.2. Are measurements appropriate (clear origin, or validity known, or standard  instrument; and absence of contamination between groups when appropriate)  regarding the exposure/intervention and outcomes?			x
		3.3. In the groups being compared (exposed vs. non-exposed; with intervention  vs. without; cases vs. controls), are the participants comparable, or do  researchers take into account (control for) the difference between these  groups?	x
		3.4. Are there complete outcome data (80% or above), and, when applicable, an  acceptable response rate (60% or above), or an acceptable follow-up rate for  cohort studies (depending on the duration of follow-up)?			x
Machare Delgado  et al23	Quantitative non-  randomized	3.1. Are participants (organizations) recruited in a way that minimizes selection  bias?			x
		3.2. Are measurements appropriate (clear origin, or validity known, or standard  instrument; and absence of contamination between groups when appropriate)  regarding the exposure/intervention and outcomes?			x
		3.3. In the groups being compared (exposed vs. non-exposed; with intervention  vs. without; cases vs. controls), are the participants comparable, or do  researchers take into account (control for) the difference between these  groups?	x
		3.4. Are there complete outcome data (80% or above), and, when applicable, an  acceptable response rate (60% or above), or an acceptable follow-up rate for  cohort studies (depending on the duration of follow-up)?		x
Matlock  et al16	Qualitative	1.1. Are the sources of qualitative data (archives, documents, informants,  observations) relevant to address the research question (objective)?	x
		1.2. Is the process for analyzing qualitative data relevant to address the research  question (objective)?	x
		1.3. Is appropriate consideration given to how findings relate to the context, e.g.  the setting, in which the data were collected?		x
		1.4. Is appropriate consideration given to how findings relate to researchers’  influence, e.g. through their interactions with participants?		x
	Quantitative randomized controlled (trials)	2.1. Is there a clear description of the randomization (or an appropriate sequence  generation)?	x
		2.2. Is there a clear description of the allocation concealment (or blinding when  applicable)?		x
		2.3. Are there complete outcome data (80% or above)?			x
		2.4. Is there low withdrawal/drop-out (below 20%)?	x
	Mixed methods	5.1. Is the mixed methods research design relevant to address the qualitative and  quantitative research questions (or objectives), or the qualitative and  quantitative aspects of the mixed methods question (or objective)?	x
		5.2. Is the integration of qualitative and quantitative data (or results) relevant to  address the research question (objective)?	x
		5.3. Is appropriate consideration given to the limitations associated with this  integration, e.g. the divergence of qualitative and quantitative data (or results)?		x
Nakagawa  et al21	Quantitative non-randomized	3.1. Are participants (organizations) recruited in a way that minimizes selection  bias?			x
		3.2. Are measurements appropriate (clear origin, or validity known, or standard  instrument; and absence of contamination between groups when appropriate)  regarding the exposure/intervention and outcomes?	x
		3.3. In the groups being compared (exposed vs. non-exposed; with intervention  vs. without; cases vs. controls), are the participants comparable, or do  researchers take into account (control for) the difference between these groups?	x
		3.4. Are there complete outcome data (80% or above), and, when applicable, an  acceptable response rate (60% or above), or an acceptable follow-up rate for  cohort studies (depending on the duration of follow-up)?	x
Radwany  et al22	Quantitative  randomized  controlled (trials)	2.1. Is there a clear description of the randomization (or an appropriate sequence  generation)?		x
		2.2. Is there a clear description of the allocation concealment (or blinding when  applicable)?	x
		2.3. Are there complete outcome data (80% or above)?			x
		2.4. Is there low withdrawal/drop-out (below 20%)?	x
Van Scony  et al17	Quantitative non-  randomized	3.1. Are participants (organizations) recruited in a way that minimizes selection  bias?			x
		3.2. Are measurements appropriate (clear origin, or validity known, or standard  instrument; and absence of contamination between groups when appropriate)  regarding the exposure/intervention and outcomes?	x
		3.3. In the groups being compared (exposed vs. non-exposed; with intervention  vs. without; cases vs. controls), are the participants comparable, or do  researchers take into account (control for) the difference between these  groups?	x
		3.4. Are there complete outcome data (80% or above), and, when applicable, an  acceptable response rate (60% or above), or an acceptable follow-up rate for  cohort studies (depending on the duration of follow-up)?			x
Vogel et al18	Quantitative  randomized  controlled (trials)	2.1. Is there a clear description of the randomization (or an appropriate sequence  generation)?	x
		2.2. Is there a clear description of the allocation concealment (or blinding when  applicable)?	x
		2.3. Are there complete outcome data (80% or above)?	x
		2.4. Is there low withdrawal/drop-out (below 20%)?			x