Introduction
Frailty is an age-related syndrome of vulnerability to stressors (Andrew et al., Reference Andrew, Searle, McElhaney, McNeil, Clarke, Rockwood and Kelvin2020; Clegg et al., Reference Clegg, Young, Iliffe, Rikkert and Rockwood2013; Gilmour & Ramage-Morin, Reference Gilmour and Ramage-Morin2021; Harwood & Enguell, Reference Harwood and Enguell2022; Kojima, Reference Kojima2015; Lorbergs et al., Reference Lorbergs, Prorok, Holroyd-Leduc, Bouchard, Giguere, Gramlich, Keller, Tang, Racey, Ali, Fitzpatrick-Lewis, Sherifali, Kim and Muscedere2021) present in 52.3% of long-term care (LTC) residents (Kojima, Reference Kojima2015). It is associated with increased dependence (Kojima, Reference Kojima2015) and disability (Aranha et al., Reference Aranha, Smitherman, Patel and Patel2020; Clegg et al., Reference Clegg, Young, Iliffe, Rikkert and Rockwood2013; Kojima, Reference Kojima2015) as well as functional decline (Aranha et al., Reference Aranha, Smitherman, Patel and Patel2020; Hamaker et al., Reference Hamaker, van den Bos and Rostoft2020; Harasym et al., Reference Harasym, Brisbin, Afzaal, Sinnarajah, Venturato, Quail, Kaasalainen, Straus, Sussman, Virk and Holroyd-Leduc2020; Harwood & Enguell, Reference Harwood and Enguell2022; Kojima, Reference Kojima2015; Lorbergs et al., Reference Lorbergs, Prorok, Holroyd-Leduc, Bouchard, Giguere, Gramlich, Keller, Tang, Racey, Ali, Fitzpatrick-Lewis, Sherifali, Kim and Muscedere2021), hospital admissions (Aranha et al., Reference Aranha, Smitherman, Patel and Patel2020; Harwood & Enguell, Reference Harwood and Enguell2022; Kojima, Reference Kojima2015), and greater risk of mortality (Aranha et al., Reference Aranha, Smitherman, Patel and Patel2020; Harasym et al., Reference Harasym, Brisbin, Afzaal, Sinnarajah, Venturato, Quail, Kaasalainen, Straus, Sussman, Virk and Holroyd-Leduc2020; Kojima, Reference Kojima2015), among other negative health outcomes (Aranha et al., Reference Aranha, Smitherman, Patel and Patel2020; Clegg et al., Reference Clegg, Young, Iliffe, Rikkert and Rockwood2013; Gilmour & Ramage-Morin, Reference Gilmour and Ramage-Morin2021; Harasym et al., Reference Harasym, Brisbin, Afzaal, Sinnarajah, Venturato, Quail, Kaasalainen, Straus, Sussman, Virk and Holroyd-Leduc2020; Harwood & Enguell, Reference Harwood and Enguell2022). Frailty is on a continuum, with those at higher degrees of frailty being at greatest risk (Aranha et al., Reference Aranha, Smitherman, Patel and Patel2020; Gilmour & Ramage-Morin, Reference Gilmour and Ramage-Morin2021).
However, there is no standard approach to the assessment and management of frailty in Canadian LTC homes (Muscedere et al., Reference Muscedere, Andrew, Bagshaw, Estabrooks, Hogan, Holroyd-Leduc, Howlett, Lahey, Maxwell, McNally, Moorhouse, Rockwood, Rolfson, Sinha and Tholl2016). This lack of standardized approach results in suffering, as LTC residents must wait to access services to alleviate symptoms associated with frailty and reduced quality of life (Buckinx, Reference Buckinx2017). To reduce this suffering, there is a need to develop a pathway that identifies frailty and delivers resident-centred care to the needs of residents (Boscart et al., Reference Boscart, Heckman, Davey, Heyer and Hirdes2018; Muscedere et al., Reference Muscedere, Andrew, Bagshaw, Estabrooks, Hogan, Holroyd-Leduc, Howlett, Lahey, Maxwell, McNally, Moorhouse, Rockwood, Rolfson, Sinha and Tholl2016; Sussman et al., Reference Sussman, Kaasalainen, Mintzberg, Sinclair, Young, Ploeg, Bourgeois-Guérin, Thompson, Venturato, Earl, Strachan, You, Bonifas and McKee2017). Resident-centred care advocates for a resident’s care and considers their and their caregivers’ desires and wishes for care, including goals of care (Muscedere et al., Reference Muscedere, Andrew, Bagshaw, Estabrooks, Hogan, Holroyd-Leduc, Howlett, Lahey, Maxwell, McNally, Moorhouse, Rockwood, Rolfson, Sinha and Tholl2016).
This pathway should start with a systematic approach to the detection of frailty. Once identified, the many effective interventions to address frailty can be selected (Avgerinou et al., Reference Avgerinou, Bhanu, Walters, Croker, Tuijt, Rea, Hopkins, Kirby-Barr and Kharicha2020; Clegg et al., Reference Clegg, Young, Iliffe, Rikkert and Rockwood2013; Hubbard et al., Reference Hubbard, O’Mahony and Woodhouse2013; Lorbergs et al., Reference Lorbergs, Prorok, Holroyd-Leduc, Bouchard, Giguere, Gramlich, Keller, Tang, Racey, Ali, Fitzpatrick-Lewis, Sherifali, Kim and Muscedere2021). For example, a comprehensive geriatric assessment (Rubenstein et al., Reference Rubenstein, Stuck, Siu and Wieland1991), promotion of physical activity/exercise (Freiberger et al., Reference Freiberger, Kemmler, Siegrist and Sieber2016), nutritional interventions (Lorbergs et al., Reference Lorbergs, Prorok, Holroyd-Leduc, Bouchard, Giguere, Gramlich, Keller, Tang, Racey, Ali, Fitzpatrick-Lewis, Sherifali, Kim and Muscedere2021), addressing polypharmacy (Hubbard et al., Reference Hubbard, O’Mahony and Woodhouse2013), and/or palliative care may be appropriate (Kaasalainen et al., Reference Kaasalainen, Sussman, Thompson, McCleary, Hunter, Venturato, Wickson-Griffiths, Ploeg, Parker, Sinclair, Dal Bello-Haas, Earl and You2020).
Any person living with an incurable condition, such as frailty, could benefit from a palliative approach that aims to provide comfort and improve quality of life through person-centred care. Palliative care improves the quality of life of both older adults and their caregivers by relieving pain and other distressing symptoms (Carter et al., Reference Carter, Leanza, Mohammed, Upshur and Kontos2021; Harwood & Enguell, Reference Harwood and Enguell2022). Unlike illnesses such as cancers, where the terminal phase is more easily identified, it can be difficult to determine when palliative care should be initiated in frailty (Hamaker et al., Reference Hamaker, van den Bos and Rostoft2020).
Our objective is to develop a care pathway to improve the detection, management, and eventual palliation of frailty for LTC residents.
Methods
A modified Delphi process was conducted from March 2021 to October 2021. This method was selected to support the achievement of consensus on issues that are complex and/or controversial (Boulkedid et al., Reference Boulkedid, Abdoul, Loustau, Sibony and Alberti2011; Hasson et al., Reference Hasson, Keeney and McKenna2000; Jones & Hunter, Reference Jones and Hunter1995). The Delphi was modified by scheduling a voluntary meeting between rounds one and two, allowing the study team to report preliminary findings and meeting participants to provide further feedback that could be incorporated into round two (Boulkedid et al., Reference Boulkedid, Abdoul, Loustau, Sibony and Alberti2011). The meeting consisted of the study team and interested panelists (including caregivers). Ethics approval was obtained from the Conjoint Health Research Ethics Board (CHREB20–2212), and informed consent was obtained from all panelists.
Initial pathway development and review
The pathway utilized resources already developed and evaluated (Ahmed et al., Reference Ahmed, Naqvi, Sinnarajah, McGhan, Simon and Santana2023; Biondo et al., Reference Biondo, King, Minhas, Fassbender and Simon2019; Kaasalainen et al., Reference Kaasalainen, Sussman, Thompson, McCleary, Hunter, Venturato, Wickson-Griffiths, Ploeg, Parker, Sinclair, Dal Bello-Haas, Earl and You2020). These resources focused on palliative care and advanced care planning within LTC. They also identified the barriers and facilitators to end-of-life care in LTC homes and what interventions promoted uptake were also considered (Harasym et al., Reference Harasym, Brisbin, Afzaal, Sinnarajah, Venturato, Quail, Kaasalainen, Straus, Sussman, Virk and Holroyd-Leduc2020; Harasym et al., Reference Harasym, Afzaal, Brisbin, Sinnarajah, Venturato, Quail, Kaasalainen, Straus, Sussman, Virk and Holroyd-Leduc2021). We utilized these resources by adapting them to frailty to better ensure their relevancy. These data, combined with pooled input from international experts and resident-family advisory councils as well as research conducted by the study team, resulted in a series of initial statements (n = 24) (Ahmed et al., Reference Ahmed, Naqvi, Sinnarajah, McGhan, Simon and Santana2023; Biondo et al., Reference Biondo, King, Minhas, Fassbender and Simon2019; de Villiers et al., Reference de Villiers, de Villiers and Kent2005; Harasym et al., Reference Harasym, Brisbin, Afzaal, Sinnarajah, Venturato, Quail, Kaasalainen, Straus, Sussman, Virk and Holroyd-Leduc2020; Harasym et al., Reference Harasym, Afzaal, Brisbin, Sinnarajah, Venturato, Quail, Kaasalainen, Straus, Sussman, Virk and Holroyd-Leduc2021; Kaasalainen et al., Reference Kaasalainen, Sussman, Thompson, McCleary, Hunter, Venturato, Wickson-Griffiths, Ploeg, Parker, Sinclair, Dal Bello-Haas, Earl and You2020). Initial statements were drafted and organized into five sections to follow the typical order in which organizations approach care and how a pathway should be applied: (i) detection of frailty, (ii) identify resident needs and contributors to frailty, (iii) illness understanding and communicate prognosis, (iv) coordinate care, and (v) manage resident needs and symptoms. Statements were refined through three rounds of feedback from the expert members of the study team (including geriatricians, care of the elderly physicians, LTC researchers, and nurses) prior to inclusion in the draft questionnaire. The draft was piloted with the co-investigators of the study team, and feedback obtained was incorporated into the questionnaire finally used for round one of our modified Delphi process.
Inclusion criteria and recruitment
Using purposive and convenience sampling, eligible individuals were invited to take part in the modified Delphi process. Eligible participants were residents of LTC homes experiencing frailty or persons who knew or cared for LTC home residents living with frailty (Hasson et al., Reference Hasson, Keeney and McKenna2000). Specifically, health care providers (e.g., doctors, nurses, health care aides, and allied health care workers), health care leaders (such as managers), policy makers, family/friend caregivers, and resident partners were targeted for recruitment from across Canada. LTC organizations aided in panelist recruitment to promote those with more experience within LTC homes (About Us – Bethany Seniors, n.d.; The AgeCare Difference, n.d.; The Brenda Strafford Foundation Ltd. | Overview, n.d.). Inclusion criteria were broad to maximize the perspectives, experiences, and expertise of potential panelists.
Recruitment of panelists, nationally, consisted of three methods: emailing respective networks via organizations’ administration, the study team posting on their Twitter accounts, and snowball sampling (i.e., invited individuals could share with others) (Emerson, Reference Emerson2015; Savard & Kilpatrick, Reference Savard and Kilpatrick2022). Posters were sent to potential participants with a direct link or QR code to the survey, allowing for ease of access.
Delphi rounds
We conducted two Delphi rounds, with an additional third round if further consensus was needed. This is based on recommendations from the literature (Boulkedid et al., Reference Boulkedid, Abdoul, Loustau, Sibony and Alberti2011). In the first round, the developed questionnaire was shared with the enrolled panel. After the first round, panelists received the analysis (i.e., median, range) of the first round, their individual response, and a summary of comments received (Boulkedid et al., Reference Boulkedid, Abdoul, Loustau, Sibony and Alberti2011).
Our modified Delphi process consisted of two rounds using the electronic platform, Qualtrics (2023) (Qualtrics XM – Experience Management Software, n.d.), for dissemination and response collection. Panelists received a link to the survey through an introductory email. Consent was obtained before panelists could proceed to the survey. Next, panelists filled out demographic information, including sex, gender, age group, role, years in role, language (s) most spoken at home, and place of birth. For round one, they then rated statements in the proposed pathway on either a 7-point Likert scale or by choosing “Yes/No” (McMillan et al., Reference McMillan, King and Tully2016). Panelists could choose not to answer a statement and to provide feedback after each statement if they felt it was necessary to add any comments or had other suggestions.
Following the completion of round one, all panelists’ feedback on the questionnaire was compiled into a single document and reviewed by another author. All relevant comments and feedback from panelists were highlighted by two authors, reviewed by the core study team, and used to refine questionnaire statements. Feedback was grouped by themes to allow the study team to better understand barriers and facilitators to implementing the pathway. However, no specific analysis method was used to group. The amalgamated questionnaire was then sent to the core study team for editing, revisions, and approval. The core study team feedback was reviewed and finalized. The edited questionnaire provided the statements used for round two. Round one panelists who submitted complete responses were invited to take part in round two.
Invitation to round two included a panelist’s signed consent form, their responses to round one, and the link to round two. Round two allowed the panelists to compare the original and revised statement, the median and range of the original statement, and narrative comments from round one. This is in line with the common Delphi procedure methodology, which informs panelists of how their responses compare to the group’s (Boulkedid et al., Reference Boulkedid, Abdoul, Loustau, Sibony and Alberti2011). Panelists first confirmed their consent to participate before rating the round two statements on a 7-point Likert scale or by choosing “Yes/No” to the respective statements. The same process for reviewing panelist feedback was repeated, and the finalized pathway consists of the edited statements. A potential round three would follow the same process as round two.
After each round, group results were reviewed and analyzed using Microsoft Excel (2023), version 2208 (Microsoft Excel Spreadsheet Software | Microsoft 365, n.d.). All communication and surveys were conducted in English.
Consensus process synthesis and pathway creation
To generate the care pathway from round two, we compiled all the feedback and incorporated it into the statements. In addition, median, interquartile ranges (IQR), mean, standard deviation, and range were calculated. Agreement was determined if the median was ≥6 using the 7-point Likert scale and ≥75% for “Yes/No” questions. A draft of the final pathway was sent to the study team for review. The study team recommended edits as per panelist feedback and ratings. Statements with high agreement would be carried into the final pathway, while lower-rated statements would be refined according to the study team recommendations and, finally, added into the final pathway. Once all statements had been reviewed and refined, the final care pathway for frailty was created.
Results
Process overview and panelists
Initial development of the Delphi process included drafting the round one questionnaire, refinement, and piloting of round one before sending it off for panelist recruitment. Rounds one and two followed similar processes of recruitment, followed by analysis and statement refinement and/or clarification. Twenty-three participants were lost between round one and round two for an attrition rate of 44.23%. Supplementary Figure F1 presents a flow diagram of our Delphi process.
Fifty-two panelists completed round one of the modified Delphi process. Among the initial panelists were 14 physicians, 6 family/friend caregivers, and a combination of 32 nurses (including registered nurses, licensed practical nurses, and nurse practitioners), health care aides, managers, and other experts (e.g., social workers, support staff). No LTC home residents took part. The panelists were mostly female (n = 45, 86.5%) and born in Canada (n = 41, 78.8%). Less than half of the panelists were between the ages of 35 and 49 years (n = 22, 42.3%), and nearly a third had over 15 years of experience in their roles (n = 19, 28.8%) (Table 1).
Demographics of frailty Delphi process panelists collected at round one
a RN = registered nurse; LPN = licensed practical nurse; NP = nurse practitioner.
b Includes family physicians, geriatricians, psychiatrists, and other physician sub-specialties.
Round one
Round one consisted of 26 statements and was open from March 24, 2021, to June 25, 2021. It had high agreement with all statements using a Likert scale, receiving a median score of 7. Most statements had an IQR = 1. “Yes/No” statements also had high agreement levels. Supplementary Table S1 presents the analysis of round one. Written feedback from panelists and scoring range were used to reframe, rewrite, or clarify statements. A summary of written feedback is provided in Supplementary Table S2.
The highest-rated statements dealt with the resident, family/friend caregivers, and inter-professional team in care planning. The four lowest-rated statements (statements 1, 2, 4, and 5) fell within the first section of the pathway, Detection of Frailty. However, all of them still received high levels of agreement (median = 7, IQR = 1).
Round two
Round two was sent to panelists on September 7, 2021, and closed on October 25, 2021. It was completed by 29 of the 52 potential panelists (55.8%). In the second round, mid to high levels of agreement to the statements were seen (see Table 2). No statements were dropped from round two. Fifteen statements were modified/updated after analysis and incorporating feedback. Some statements included in round two (n = 28) were edited as required to reflect feedback from round one. Two statements were merged in the Detection of Frailty section. Three new statements were added based on the feedback obtained from round one. In the Illness Understanding and Communicating Prognosis section, two new statements were added regarding communicating frailty diagnosis and conversations that should occur among health care providers, residents, caregivers, and loved ones. One new statement in the Manage Resident Needs and Symptoms section focused on end-of-life care. Supplementary Table S3 presents the analysis of round two.
Median and interquartile range for each section of the finalized frailty care pathway
Pathway development and final results synthesis
The final pathway includes 28 statements under the 5 main sections (see Figure 1 for an overview of pathway and Supplementary Table S4 for the detailed pathway). The detection of frailty section (i) focused on how and when to identify frailty, who should be involved, and when reassessment should take place (n = 6 statements). The identification of resident needs and contributors to frailty section (ii) highlighted the importance of contributing factors to frailty, and how addressing them may reduce frailty severity (n = 2 statements). Section (iii), on communicating accurate information about frailty and its prognosis (n = 6 statements), centred around training and educating staff, residents, and caregivers about frailty and its trajectory. Section (iv) on coordination of care (n = 4 statements) covered when to consult and promote inter-professional involvement. The final section (v) focused on managing resident needs and symptoms (n = 10 statements). These statements included tailoring care plans and addressing symptoms while honoring the beliefs, wishes, and preferences of the resident.
Summary of the frailty care pathway’s five sections: detect frailty, identify resident needs and contributors to frailty, illness understanding and communicate prognosis, coordinate care, and manage resident needs and symptoms.
Discussion
A clinical care pathway for frailty in older adults living in LTC homes was developed that spans from detection, identification of resident needs and contributors to frailty, communicating information about frailty and its prognosis, coordination of care, and management of resident needs and symptoms.
Detection and diagnosis
Our study showed a high perceived need for frailty screening and diagnosis (Buckinx, Reference Buckinx2017). Panelists agreed that frailty is not currently being systematically detected and diagnosed in a timely manner. This represents a knowledge-to-practice gap in care for those living with frailty, which was evident in the comments received during the modified Delphi process. Using a validated, feasible tool such as the Clinical Frailty Scale to measure frailty would reduce subjectivity and improve reproducibility (Andrew et al., Reference Andrew, Searle, McElhaney, McNeil, Clarke, Rockwood and Kelvin2020; Clegg et al., Reference Clegg, Young, Iliffe, Rikkert and Rockwood2013; Rockwood, Reference Rockwood2005). The detection of frailty could help communicate frailty to other disciplines and then lead to further action, such as addressing and/or mitigating its contributing factors (Muscedere, Reference Muscedere2020). Systemic barriers to the detection and diagnosis of frailty exist (Boscart et al., Reference Boscart, Heckman, Davey, Heyer and Hirdes2018). Moreover, 60–80% of LTC residents have dementia, which complicates prognosis further, as care staff tend to overestimate prognosis in residents with advanced dementia (Payne et al., Reference Payne, Sheppard, Steinberg, Warren, Baker, Steele, Brandt and Lyketsos2002; Sampson, Reference Sampson2010). Our care pathway recommends the use of an inter-professional care team, working closely with patient caregivers, to help improve the quality of care provided for LTC home residents living with frailty.
Illness understanding and early use of palliative care approaches
We highlight how crucial a shared understanding of frailty is for health care providers, LTC home residents, and family/friend caregivers. Consistent with previous research, we found that education around frailty is needed (Carter et al., Reference Carter, Leanza, Mohammed, Upshur and Kontos2021; Warren et al., Reference Warren, Gordon, Pearson, Siskind, Hilmer, Etherton-Beer, Hanjani, Young, Reid and Hubbard2022). Providing education and training to staff enables them to provide high-quality care for older adults with frailty and helps support implementation of interventions most likely to be effective (Kojima et al., Reference Kojima, Liljas and Iliffe2019). Greater residents and family understanding of frailty can also support more informed decision-making (Harasym et al., Reference Harasym, Afzaal, Brisbin, Sinnarajah, Venturato, Quail, Kaasalainen, Straus, Sussman, Virk and Holroyd-Leduc2021) and better care plans developed by the inter-professional team, LTC home resident (as able), and family/friend caregivers.
Given the persistent and life-limiting nature of frailty, early discussion of a palliative approach to care is often relevant, though inconsistently undertaken. These discussions about palliative care would, we believe, lead to more informed decisions about care and the importance of both advanced care planning and developing goals of care (Harasym et al., Reference Harasym, Afzaal, Brisbin, Sinnarajah, Venturato, Quail, Kaasalainen, Straus, Sussman, Virk and Holroyd-Leduc2021). Panelists agreed that meetings to discuss care should occur at regular intervals and if there is a significant change in a resident’s health status.
Management
Once frailty has been diagnosed and inter-professional conversations with residents and caregivers initiated, priorities should focus on adapting care plans, managing resident needs, and addressing symptoms. Interventions specific to the required level of care can take place to support optimal aging and decrease the risk of mortality (Gilmour & Ramage-Morin, Reference Gilmour and Ramage-Morin2021). Appropriate interventions could include physical exercise, addressing dietary needs, and consultation with other medical professionals uniquely suited to address specific care needs (Gilmour & Ramage-Morin, Reference Gilmour and Ramage-Morin2021; Lorbergs et al., Reference Lorbergs, Prorok, Holroyd-Leduc, Bouchard, Giguere, Gramlich, Keller, Tang, Racey, Ali, Fitzpatrick-Lewis, Sherifali, Kim and Muscedere2021). Geriatric medicine/care of the elderly, psychiatry, or palliative care teams should be consulted when there is a need and the resources are available. For example, a comprehensive geriatric assessment delivered by an inter-professional team can assist in identifying contributors to frailty in a given resident, management options for them, and assist in establishing care priorities that can help create a care plan that considers frailty (Clegg et al., Reference Clegg, Young, Iliffe, Rikkert and Rockwood2013).
Limitations
A range of LTC personnel was included in the Delphi to reflect various perspectives, including care partners and health care providers. Residents of LTC homes did not participate. We attempted to share the survey with residents via posters and staff engagement. However, the lack of recruitment was in large part due to virtual recruitment and participation being due to the pandemic. Future studies can mitigate this barrier by providing paper surveys or having a member of the study team administer the surveys. Our recruitment methods tried to reach representative groups through social media and email to cover a diverse population. While we have created a clinical care pathway for frailty in older adults living in LTC homes consisting of a series of statements that have face validity, this pathway has not been validated nor have we addressed the issues that might impede its implementation.
Conclusion and implications
The final pathway promotes a resident-centred approach that highlights caregiver involvement and inter-professional teamwork to identify and manage frailty, as well as initiate palliative care earlier. Implementing this pathway will allow health care providers to adopt screening measures and adapt care to a resident’s frailty severity. Panelists in our study agreed on the importance of developing a standardized approach to the detection and management of frailty in LTC homes to improve the quality of life of both residents and caregivers. Our study results also emphasize the need to consider a tailored approach for this population on how to plan for, manage, and attempt to decrease the severity of frailty. Incorporating this frailty pathway into already-existing pathways could help to further target and personalize care and provide a more holistic approach to care.
Further work is required to validate the utility of the care pathway, such as peer review and piloting first to understand the feasibility and identify barriers to implementation. Following that, implementation can be expanded and tailored for sites. If results are positive, we propose that this pathway should be integrated into the care offered to LTC home residents with frailty. Adoption of this care pathway will create a baseline that LTC homes and other care providers can refer to when caring for an LTC home resident.
Supplementary material
The supplementary material for this article can be found at http://doi.org/10.1017/S0714980825100123.
Acknowledgements
The authors would like to thank Dr. Barbara Liu, Dr. Peter Tanuseputro, Dr. Mary Jane Shankel, Navjot Sidhu, Beth Gorchynski, Wayne Morishita, Glenn McKinley, Lisa Poole, the Brenda Strafford Foundation, AgeCare, and Bethany Care Society for their expertise and assistance with recruitment.
Financial support
Funding was provided by the Canadian Institutes of Health Research, Canadian Foundation for Healthcare Improvement, and Canadian Patient Safety Institute.
Sponsor’s role
This study was supported by a grant from the Canadian Institutes of Health Research, Canadian Foundation for Healthcare Improvement, and Canadian Patient Safety Institute. The three organizations had no role in the design, methods, subject recruitment, data collection, analysis, and preparation or submission of the paper.
Open access
