Overview and definitions

This study was conducted using a modified Delphi method. This approach generates consensus-based expert opinion using an iterative questionnaire and feedback process [29]. The Delphi method has previously been used to develop guidelines that help families care for people with dementia [30], to assess reportable and preventable events in home care [31], to prioritize quality measures of nursing home care [32–34], and to identify operational standards and guidelines for palliative care [35]. Similar to the approach used by others [30, 36], we conducted the Delphi survey independently with two expert groups (care partners and healthcare stakeholders), enabling us to compare and contrast the priorities identified by these different expert groups. The following definitions were used to guide this research.

Study setting

This study is part of a larger research program funded by the Canadian Institutes of Health Research [46] designed to compare continuing care health service policies, practices and utilization patterns between Edmonton, Alberta and Winnipeg, Manitoba. While Canada has a national insurance plan that funds medically necessary physician and hospital services in each province [47], planners have more latitude when determining the type and volume of continuing care services that are delivered provincially. Manitoba and Alberta were selected for the broader research program given their similar underlying population characteristics (e.g., household size, age structure, prevalence of caregiving) [48, 49], their long-term commitment to support AIP [50, 51], coupled with their differing continuing care health services (e.g., compared to Alberta, Manitoba has 21% more nursing home beds per capita 85+ year old [52], and one versus three levels of community-based assisted living care [37, 53]). Delphi respondents in the present study were recruited from these regions.

Framework

This research was guided by the Lau et al. (2007) Health-Related Safety Framework which shows how micro, mezzo, and macro-level factors can minimize preventable and unintended harm to community-dwelling older adults resulting from breakdowns in the societal system [54]. Micro factors include an individuals’s biological or psychological characteristics (e.g., physical, mental and functional health; attitude and knowledge; personal health behaviours). Mezzo factors include one’s social network, the structure and features of their home and/or community, and available medical and social services. Macro factors include the broader economic, societal and political forces that may affect a person’s ability to age in place (e.g., having access to funds that help people purchase assistive technology) [54]. Similar to others [55, 56], we used the Health-Related Safety Framework to (1) ensure that Delphi responses were based on a diverse array of community-based AIP factors, (2) help structure our study findings.

Selecting factors to include in the Delphi survey

This research was conducted by a team of researchers, healthcare planners, and community advisors (i.e., care partners with system experience) from each province. Two members of our research team (MC and ES) used key words from select articles and reports [38–42] to develop and apply the following search terms to select databases (Google Scholar, PubMed, CINAHL).

“older adults” AND “aging in place” AND/OR “community supports”; “older adults” AND “facilitators for aging in place”; “delay” OR “prevent” AND “nursing home admission”

Titles and abstracts were reviewed, and potential factors were extracted from relevant articles. MC, ES, HCE and MD amalgamated these factors into a comprehensive list and provided a lay description of each factor for use in the Delphi survey. As part of a full-day workshop, our entire team reviewed and refined the list of factors and their definitions.

Participant recruitment, survey development and application

This research was approved by the University of Manitoba Health Research Ethics Board (reference number HS22703 (H2019:117)), the Winnipeg Regional Health Authority Research Access & Approval Committee (#2018–027), the University of Alberta Health Research Ethics Board (MS3_Pro00089982), and Alberta Health Services. Informed consent was obtained from all participants through the online survey.

Two groups of study participants − family/friend care partners (‘care partners’) and healthcare stakeholders (i.e., planners and providers) − were recruited in June and July of 2019 with assistance from team members. To be included, care partners must have lived in Manitoba or Alberta, and have had current or previous experience providing care. These participants were recruited through personal contacts made by our community advisors, and also through Manitoba’s and Alberta’s Primary and Integrated Healthcare Innovation Networks, which help researchers to develop patient partnerships [57, 58]. Community advisor team members approached their personal networks and identified individuals who met the study criteria. MC followed up with interested individuals to confirm their eligibility, answer questions, and identify additional participants using snowball sampling.

To be included in the study, healthcare stakeholders must have had experience working in the Winnipeg or Edmonton continuum of healthcare services, with an emphasis on older adult community-based care. Those who worked exclusively in nursing homes were excluded. Healthcare planner team members identified colleagues using these criteria; MC followed up with each potential participant to determine their interest, confirm eligibility, answer questions, and to identify additional participants using snowball sampling.

A draft version of the survey was piloted with a total of four team members (two healthcare planners, two community advisors). These team members completed the survey and provided feedback about its format and content. They also completed a revised version of the survey to verify that the appropriate improvements were made. The final version of the Delphi survey was administered online using SimpleSurvey [59], separately to each participant group (S1 File contains the round 1 survey for healthcare stakeholders). The survey first explains the study goals, gives key definitions, and provides directions for completing the questionnaire, including a scoring example. Participants were then instructed to complete a consent section, a participant information section (documenting their type and duration of experiences), and the actual questionnaire. With the exception of the group-specific participant information questions, both participant groups completed the same survey.

At the end of Round 1, participants were asked to suggest additional factors to include in subsequent survey rounds. To facilitate group comparisons, the suggestions made by each group were provided to the other. In Round 2, participants from each group were provided with their Round 1 score for each factor, the median score provided by peers, and an anonymized list of (group-specific) comments used to rationalize people’s scoring choice. Participants were asked to re-score each factor taking this information into account. Those who did not complete round one were ineligible to participate in subsequent Delphi rounds.

Delphi scoring and data analysis

Care partners and healthcare stakeholder participants completed the Delphi survey independently, and hence scoring and analyses were conducted separately for each group. Participants in each group rated every factor from 1 (“not important; by providing this score, you are saying that a factor does not help older adults to live successfully the community”) to 10 (“absolutely critical; this factor is amongst the top 1 or 2 things you feel are needed to help most older adults live successfully in the community”). A score of ‘8’ defined factors that were considered ‘very important’ (“this factor is amongst the top 5 or 6 things you feel are needed to help most older adults live successfully in the community”).

Results from each Delphi round were analyzed as per the method described by van der Steen et al. (2014), separately for care partners and healthcare stakeholders [60]. Measures of central tendency (median) and variation (interquartile range, IQR) were used to define the following levels of agreement:

Very High (80⁺% of participants provided a score ≥8; IQR = 0);

High (80⁺% of participants provided a score ≥8; IQR = ≤2);

Moderate (60⁺% of participants provided a score ≥8; IQR ≤4); and

Low (all other results).

Consensus on a given Delphi factor was defined when participants reached a ‘high’ or ‘very high’ level of agreement. As per Jorm (2015), once consensus was reached for a factor, we removed it from subsequent survey rounds [61]. Care partners reached consensus on 10 factors after two Delphi rounds, at which point we decided not to conduct a third Delphi round for this group. This decision is consistent with other studies [36], supports our study goal to identify only the most important AIP factors, and aligns with our rating instructions (i.e., to give, at most, 5 or 6 factors a score of ‘8’). Group specific results were analyzed separately, enabling us to compare and contrast the priotities identified by each expert group.

Community-based aging in place factors

Twenty-three factors were extracted from the academic literature, and 4 additional factors were added by participants during the first Delphi round, for a total of 27 factors examined in this research (Table 1). These factors were categorized into the micro/person (n = 8), mezzo/community (n = 15), and macro/policy and societal (n = 4) levels.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Survey factors by source, and categorized by the Health-Related Safety Framework [54].

https://doi.org/10.1371/journal.pone.0259387.t001

Study participants

Twenty-five care partners and 36 healthcare stakeholders participated in this study with low dropout between rounds (Table 2). Care partners were predominantly female (96%) and were on average 67 years old. Forty percent of these participants were currently providing care at the time of the study, and 60% had received either formal or informal care at some point in their life.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Description of Delphi participant characteristics.

https://doi.org/10.1371/journal.pone.0259387.t002

Healthcare stakeholder respondents were also predominantly (83%) female, and all of these participants had at least some post-secondary education (100% bachelor’s degree or higher) (Table 2). At the time of the study, these participants had worked, on average, 15.8 years in the continuing care sector. Half of these participants worked as decision makers in regional programs. The remaining participants were employed by provincial governments or worked in continuing care as planners or care providers.

Delphi findings

The final Delphi results are shown in Table 3. Care partners and healthcare stakeholders reached consensus (high or very high agreement) on ten and three factors, respectively. All three factors identified by the healthcare stakeholders were also selected by the care partners. These three factors were all at the mezzo-level of the Health-Related Safety Framework [54], and were organized into social networks (having strong relationships and links with family, friends, and the community), the structure of the home and neighborhood (living in a safe home environment, e.g., with enough safety aids and equipment), and medical services (having coordinated care between all types of formal health care providers, e.g., physicians, home care workers, social workers) (Fig 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Consensus factors categorized using the Health-Related Safety Framework [54].

Bolded text: Consensus was reached by both groups; Un-bolded text: Consensus was reached by care partners only. *Factor was added after Delphi Round 1.

https://doi.org/10.1371/journal.pone.0259387.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Final consensus results by factor and participant group.

https://doi.org/10.1371/journal.pone.0259387.t003

Care partners reached consensus on an additional 7 factors at the micro- (e.g., having enough money to afford to stay successfully in the community), mezzo- (e.g., having programs that help people to cope with mental health challenges such as anxiety, depression and loneliness), and macro-levels (e.g., ensuring that community-based alternatives to nursing home use are affordable) (see Table 3 and Fig 1). The complete results of each Delphi round are provided in S1 and S2 Tables, for care partners and healthcare stakeholders, respectively.

Priority areas identified by both participant groups

Care partners and healthcare stakeholders in this study agreed that having a safe home environment, maintaining strong social networks with others, and having coordinated healthcare services are very important to help older adults successfully age in place. While researchers have used qualitative approaches [43, 45, 62] and statistical models [44, 63, 64] to identify pertinent AIP factors, to the best of our knowledge this study is the first to identify highly important community-based AIP factors prioritized across participant groups. Our results align with the findings from a UK-based scoping review that defines the kinds of care and support (e.g., social relationships and activities, help related to mobility) that older adults need [65]. These identified priorities are also in keeping with others who have reported that successful AIP relies on adequate home safety (e.g., ensuring that people have the appropriate grab rails, nonslip surfaces and ramps in place) [66], and that people’s decisions to remain in their community are based in part on their ability to effectively adapt their home [67]. Maintaining strong social networks is thought to be essential for successful AIP [68], and alternatively loneliness has been identified as an independent predictor of nursing home admission [69, 70]. Smaller scale interventions addressing these factors have been implemented, including environmental audits to help to improve home safety [71], activity and discussion groups to enhance social connection [72], and navigators to help older adults with chronic conditions transition through the healthcare system [73]. Findings from these smaller-scale interventions can help to guide the implementation and evaluation of larger scale AIP innovations including their sustainability and scale.

Differences in expert group perspectives

In addition to the aforementioned three priority areas, care partners selected seven additional factors that they felt were very important to consider. These additional factors are dispersed across the micro- (keeping mentally active, having enough money to stay in the community), mezzo- (having mental health programs, ensuring providers are aware of community-based care options), and macro- (having access to affordable housing, having funds to purchase assistive technology, having appropriate community-based alternatives to nursing homes) levels of the Health-Related Safety Framework [54]. These additional factors highlight the importance of having the finances to afford the supports and services needed to stay in the community, of ensuring that everyone has access to and is aware of these community-based programs, and the need for more mental health services. Overall, from the perspective and lived experience of care partners, successful AIP likely requires a larger number of diverse strategies to meet the varied needs of community-living older adults.

Other researchers have reported that different participant groups often have different perspectives about healthcare reform [25, 26, 36, 74, 75]. When investigating strategies to improve cancer care services in Greece, Efstathiou, Coll, Ameen and Daly (2011) reported that providers prioritized ways to better coordinate patient services, while healthcare users focused on treatments that would increase chances of survival [36]. Similarly, when investigating mental health research priorities, Owens, Ley and Aitken (2008) reported that clinicians and managers tended to prioritize research that focused on the provision of physical healthcare, while care partners and service users emphasized research that would promote independence, self-esteem and recovery [75].

Examining various perspectives and the potential reasons for their similarities and differences has value, given research showing that collaborative planning across stakeholder groups encourages different ways of thinking, greater reflexivity, and creates innovations that tend to have greater impact when implemented [27, 28, 76]. In the present study, healthcare stakeholders may have selected factors based on their scope of professional experience (e.g., selecting mezzo-level factors like coordinating formal healthcare, and arranging for safety aids and equipment). Alternatively, healthcare stakeholders in our study worked across the care continuum as policy makers, regional planners and providers, and these diverse perspectives may have limited their consensus making capacity. Conversely, care partners may have had more direct lived experiences and hence a broader understanding of the multidimensional nature of AIP, which influenced their selection of factors at the micro, mezzo, and macro-levels of the Health-Related Safety Framework [54]. Notwithstanding these potential explanations, the present study emphasizes the importance of engaging with ‘end-users’ when developing AIP research and care reform priorities, as one means to more effectively advance this agenda.

Future research directions

Two sequential research directions are proposed. First, priorities identified in the present research are based on the views of experts from two Canadian healthcare regions. While our study findings have AIP reform implications for these local regions, further research is needed to define the extent to which they can be generalized to other jurisdictions. While there is some evidence to support the generalized nature of our study results (e.g., UK-based research identifies similar supports needed to age in place) [65], additional and ongoing research in this area is required.

Second, research is needed to examine the extent to which these key supports and services are provided successfully and equitably to various groups of community-dwelling older adults. Since Canada is not alone in its desire to enhance AIP, this analysis should ideally occur using an international comparative lens, in particular engaging with Denmark and Norway that are known to have well-developed home and community-based care programs [5]. The lessons learned from these comparisons can help to guide the development, feasibility testing, and eventual larger-scale implementation of innovations designed to better help older adults successfully age in place.

Limitations

Strengths of this research include our rigorous use of the Delphi method, and our ability to create comparative results across participant groups. These strengths are offset by three potential limitations. First, as previously mentioned, our results are based on expert opinion from two Canadian regions, and generalizations to other national or international jurisdictions, may be limited. Second, while participants were asked to rank the importance of each factor individually, it is likely that multiple factors interact to affect one’s ability to AIP. We also did not ask participants to identity which of the listed supports and services were already available; it is feasible that participants, purposefully or unconsciously, rated factors based on perceived importance and availability. Third, we did not conduct an exhaustive literature review to identify all potential AIP supports, which may have influenced our study results. We minimized this third limitation by asking participants to provide additional factors during the first survey round.