Palliative care in the home: a scoping review of study quality, primary outcomes, and thematic component analysis

A scoping review was completed [13]. Quality of included studies was assessed to increase the utility of this scoping review and to add the needed quality lens to the literature. In August 2016, the following electronic databases were searched from inception: PubMed, Embase, Cumulative Index to Nursing and Allied Health Literature (CINAHL), Web of Science, Cochrane Library, EconLit, PsycINFO, Centre for Reviews and Dissemination, Database of Abstracts of Reviews of Effects, and National Health Service Economic Evaluation Database. Palliative care in the home experts were contacted to identify additional papers. Search results were limited to those published after the year 2000, to ensure that included studies were representative of modern palliative home care. Only published studies were reviewed thus ethics approval was not required. Grey literature was not included. Only English search results with human subjects were included in abstract review. Preferred reporting items for systematic reviews and meta-analyses (PRISMA) guidelines were followed to ensure methodological best practices [14].

The search strategy consisted of three concepts. First, terms for palliative care such as “palliative care,” “terminally ill,” and “end of life care” were searched. Second, terms for home care such as “home care services/trends,” “health system pathway,” and “component” were searched. Third, terms for outcomes such as “health care quality, access, and evaluation,” “quality,” and “patient satisfaction” were searched. These three concepts were combined using the Boolean operator “and.” The PubMed search strategy is included in Additional file 1: Appendix 1.

Titles and abstracts were reviewed for eligibility in duplicate. Inclusion criteria were: 1) must assess at least one component of palliative care; 2) must report location of care as “home of patient” irrespective of if care was delivered in an individual home, hospice, or continuing care; 3) report on any outcome; and 4) report primary data. No study designs were excluded. Abstracts included by either reviewer proceeded to full text review. At full text review, a third reviewer was consulted to resolve inclusion disagreement.

Included studies were classified by primary outcome. Data extracted, independently and verified by a second reviewer, included objective, methods, primary outcomes, and themes addressed. Study quality of randomized controlled trials was assessed using the Cochrane Risk of Bias (ROB) tool [15]. Controlled before and after studies was assessed using the same criteria as randomized controlled trials, but reported as high risk of bias on random sequence generation and allocation sequence concealment [16]. Quality of cohort studies was assessed with Newcastle-Ottawa criteria [17]. Quality of qualitative studies was assessed using the Critical Appraisal Skills Programme Qualitative Checklist, as recommended by Cochrane [18, 19]. Cross-sectional study quality was assessed using the Cochrane suggested “Quality Assessment Tool for Quantitative Studies” [20, 21]. Cost studies did not include outcomes in analysis, and were therefore not appropriate for any existing validated quality assessment tools.

Thematic component analysis was used to describe palliative home care interventions [22]. Face validity of intervention themes was established with expert opinion. The five themes categorizing palliative home care interventions were: accessibility of healthcare, family and caregiver support, individualized patient centered care, multidisciplinary care provision, and quality improvement.

Searches of electronic databases returned 1993 records (Fig. 1). Following removal of duplicates, 986 records remained for title/abstract screening. Five hundred four articles were excluded at the title/abstract screening stage. The full text of 436 articles were assessed for inclusion, and 53 articles were included in the final data set (Fig. 1). Qualitative study designs were most numerous with 10 studies identified [8, 23‐31]. This was followed by retrospective cohort study designs, with eight studies included [32‐39]. The country in which the greatest number of studies was conducted was the United States, with 12 studies [32, 33, 40‐49]. This was followed by Canada with nine studies [6, 30, 35, 37, 38, 50‐53]. Included study characteristics are presented in Additional file 2: Table S1.

Included studies were assessed with the scales indicated in Fig. 2, and a category for overall quality was added. When using the Cochrane Risk of Bias (ROB) tool, categories assigned low risk of bias were given a value of one, categories assigned moderate risk of bias were given a value of two, and categories assessed as high risk of bias were given a value of three [15]. The sum of all ROB values was calculated. Values from 6 to 10 were categorized as low risk of bias or high quality, values from 11 to 14 were categorized as unclear/moderate quality, and values from 15 to 18 were categorized as low quality with the Cochrane ROB tool. When using the Newcastle-Ottawa scale, from one to three stars was categorized as low quality, four to six stars was categorized as unclear/moderate quality, and seven to nine stars was categorized as high quality [17]. In qualitative studies assessed using the Critical Appraisal Skills Programme Qualitative Checklist, the sum of the number of times “yes” was used as the answer to questions was calculated [18]. From one to three “yes” answers was categorized as low quality, four to six “yes” answers was categorized as unclear/moderate quality, and seven to nine “yes” answers was categorized as high quality.

Most studies had an unclear risk of bias or quality (Fig. 2). The only study tool used that did not identify any studies with high quality was the quality assessment for quantitative studies tool.

Multiple outcomes described studies in which the objective statement identified a combination of resource use, symptom burden, quality of life, satisfaction, caregiver distress, or place of death as the primary outcome (Fig. 3). The most commonly reported outcome was descriptive in nature with the objective of the study being to describe experiences with services offered.

Intervention components were analyzed thematically, and five themes were identified (Table 1). Study interventions were described with one or more themes. Theme one described interventions in which a major component was an increase in the ability of participants to access healthcare as required. Theme two, caregiver support, was identified in 12 studies and reported effects on caregivers or provided services to caregivers. Theme three, individualized patient centered care, was identified in 35 studies. Individualized patient centered care described interventions in which tailored care to participant needs. Theme four, multidisciplinary care provision, was identified in 24 studies and described interventions with care provided by disciplines other than, or in addition to, physicians and nurses. To be categorized as addressing the theme of quality improvement, the publication had to describe the palliative care in the home intervention as a change to previously existing palliative care in the home service with the goal of improvement.