Safe and effective person- and family-centered care practices during transitions from hospital to home-A web-based Delphi technique.

BACKGROUND
Research has shown that adverse events during care transitions from hospital to home can have a significant impact on patients' outcomes, leading to readmission, delayed healing or even death. Gaps exist in the ways of monitoring care during transition periods and there is a need to help organizations better implement and monitor safe person-and family-centered care. Value statements are a way to obtain narratives in lay terms about how well care, treatment and support is organized to meet the needs and preferences of patients/families. The purpose of this study was to identify the value statements that are perceived by decision-makers and patients/families to best signify safe person- and family-centered care during transitions from hospital to home.


METHODS
Between January and September 2017, a web-based Delphi was used to survey key stakeholders in acute care and home care organizations across Canada.


RESULTS
Decision-makers (n = 22) and patients/families (n = 24) from five provinces participated in the Delphi. Following Round 1, 45 perceived value statements were identified. In Round 2, consensus was received on 33/45 (73.3%) by decision-makers, and 30/45 (66.7%) by patients/families. In Round 3, additional value statements reached consensus in the decision-makers' survey (3) and in the patients/families' survey (2). A total of 30 high priority value statements achieved consensus derived from both the decision-makers' and patients/families' perspectives.


CONCLUSION
This study was an important first step in identifying key consensus-based priority value statements for monitoring care transitions from the perspective of both decision-makers and patients/families. Future research is needed to test their usability and to determine whether these value statements are actually suggestive of safe person-and family-centered care transition interventions from hospital to home.

value statements achieved consensus derived from both the decision-makers' and patients/ families' perspectives.

Conclusion
This study was an important first step in identifying key consensus-based priority value statements for monitoring care transitions from the perspective of both decision-makers and patients/families. Future research is needed to test their usability and to determine whether these value statements are actually suggestive of safe person-and family-centered care transition interventions from hospital to home.

Background
Person-and-family-centered care (PFCC) allows the planning, delivery, and evaluation of health care to be grounded in mutually beneficial partnerships among health care providers, patients, and families [1]. Providers who believe in PFCC will typically develop relationships, communicate, collaborate, share information and engage with patients and their families with regards to health care [2]. Patient and family engagement is fundamental to both a PFCC approach to health delivery as well as to improving overall patient safety in our healthcare system [2][3][4]. This engagement process is defined as "patients, families, their representatives, and health professionals working in active partnership at various levels across the healthcare system-direct care, organizational design and governance, and policy making-to improve health and health care" [3], p. 224. It may encompass the complete spectrum of engagement, from understanding patients' experiences and perspectives with the health system to fully involving patients and families in improving the health of their communities [3,5]. Research shows that patients who are more involved in the decision-making process related to their care are better able to manage complex chronic conditions, [6][7][8] have reduced anxiety and stress [9] and have shorter lengths of stay in hospital [10]. There is growing consensus that engaging patients and families can improve the quality of their care, particularly when transitioning from hospital to home [11].
PFCC practices during transitions in care are defined as: "a set of actions designed to ensure the coordination and continuation of health care as patients transfer between different locations or between levels of care within the same clinical setting" [12], p. 533. The challenges associated with care transitions are complex, and, thus require a multifaceted approach [13]. Existing care transition models such as the Care Transitions Interventions (CTI) [14] and the Transitional Care Model (TCM) [15] can provide insight into interventions that can best support patients during care transitions. Despite, the many studies that have trialed multifaceted care transition interventions to improve the patient experience and to reduce readmissions to hospital [16][17][18][19][20], the results have been inconsistent [21][22][23][24][25].
Thus, care transitions, particularly from hospital to home, continue to be poorly managed and pose a high risk for harm [26][27][28][29]. A recent systematic review of transitional care interventions (n = 12) for older people with chronic illnesses has shown the need for more involvement from patients and from their informal caregivers [30]. There is a need for additional research to focus on engaging patients and their families in identifying aspects of care transitions that are most important to them and that are linked to safe person-and family-centered. In this study, we used value statements as a way to obtain narratives in lay terms about how well care, treatment and support was organized to meet the needs and the preferences of patients and families [31]. Value statements were chosen over the more usual performance measures because they were easier for lay participants to understand [32]. The purpose of this study was to identify the value statements that are perceived by health care decision-makers, patients and families to best signify safe person-and family-centered care during transitions from hospital to home.

Methods
We conducted a three-round, web-based Delphi technique [33] to obtain consensus on potential value statements that best reflect safe person-and family-centered care transitions. We obtained ethics approval for our study from the University of Ottawa Research Ethics Board (uOttawa REB), and the Ottawa Health Science Network Research Ethics Board (OHSN REB).

Sampling strategies and recruitment
Key stakeholders including health care decision-makers from organizations across Canada, and patients/families chosen from a pan-Canadian patient advisory group were invited to participate. Decision-makers included: quality and/or performance measurement (decision support) leads, medicine and/or surgery directors and managers, and physician and nursing leadership from hospitals, regional health authorities or home care agencies who were involved in the oversight of initiatives related to improving care transitions from hospital to home. Decision-makers were either previously known to the research team for their expertise, or were identified through an Internet-based search of their organization website. Patient and family representatives were recruited through the primary contact within each of the participating organizations. Eligible patient and family representatives were defined as those 18 years of age or older with experience as a patient or as a primary caregiver or who had previously experienced a care transition for an acute trauma or an illness.
Email invitations were sent using the modified Dillman approach [34]. To recruit decisionmakers, the email invitations consisted of an initial standardized recruitment message with a link to the survey. The message consisted of an invitation for the decision-makers to complete the survey or to share it with other decision-makers within their organization, as appropriate. To recruit patient and family representatives, a separate email invitation was sent via the decision-makers to specifically invite their patient and family advisory council members to complete the survey. At two and four weeks intervals after the initial invitation email, reminder emails were sent.

Data collection
In Round 1, decision-makers were asked to participate in a web-based survey to identify the interventions and related measures in place in their organization, to monitor transitions from hospital to home. The specific questions are listed in Table 1. These results were gathered to develop the Round 2 survey. The research team also evaluated best practices based on previously published literature to identify other potential value statements.
In Round 2, all participants (decision-makers and patients/family members) were asked to rate on a scale of 1 to 5, each of the value statements according to their importance to overall quality and safety during care transitions from hospital to home. The scale corresponded to: (1) Not important, (2) Slightly important, (3) Moderately important, (4) Important, and (5) Very Important. Participants were also asked to suggest other value statements that they felt were missing from the list provided.
In Round 3, all participants (decision-makers and patients/family members) were asked to reconsider their answer regarding the value statements for which no consensus had been reached regarding their importance. Beside each statement, we provided the participants with the group response (decision-makers vs. patient/family members), and their own individual response to the value statement. We also requested that participants comment on their answers.

Data analysis
The lead author (CB) led the data analyses. In Round 1, a trained research assistant and CB independently grouped the responses together in semantically similar themes, and removed any duplicate statements. This was done by grouping similar response statements, and then theming them. To identify any missing themes, the research assistant reviewed the themes against the existing relevant literature found in our recent systematic review on the effectiveness of PFCC transition interventions from hospital to home [35]. Missing themes were then added to the primary data collected during Round 1. From the common themes identified, the research assistant and CB developed value statements which were reviewed by the research team. This created a list of value statements for Round 2. In Round 2 and Round 3, data were reported using descriptive statistics. The consensus scores looked at the number of respondents within a group (decision-makers or patient/family members) that rated the importance of a specific value statement at a value of 4.0 or higher. The consensus score was thus defined as follows: Consensus = # of respondents scoring a statement above 4.0/total number of respondents. For the purpose of this study, consensus on each item was pre-defined to be at least 90% agreement among survey participants. The means for each of the value statements were calculated and used to rank the statements in order from the most important to the least important. The mean of the value statement was simply the numerical mean calculated as follows: Mean = Sum of importance scores / # of respondents. The additional themes suggested by the participants in Round 2 were also developed as value statements by the research team.

Results
Twenty-two decision-makers (n = 22) and twenty-four patients/families (n = 24) from five Canadian provinces participated in the Delphi survey. The demographic characteristics of the participants for the three Delphi rounds can be found in Table 2.

Round 1
Following the first round, a total of 45 value statements were suggested by participants. Duplicates were removed, providing a list of 36 possible value statements. With the addition of 9 value statements from the literature, a total of 45 value statements were available for Round 2. The complete list of value statements for decision-makers and for patients/families can be found in the [S1 Supporting Information].

Round 2
We identified a high level of consensus (�90%) by decision-makers for 33/45 (73.3%) of the value statements. The consensus score looked at the number of respondents within a group that rated the importance of a specific value statement at a value of 4.0 or higher. We also identified high levels of consensus (�90%) by patients/families for 30/45 (66.7%) of the value statements. The value statements that received high level of consensus and the new value statements suggested by participants in Round 2 are available in the S1 Supporting Information.

Round 3
In Round 3, the value statements that did not receive consensus (those statements that did not reach 90% consensus or higher) for decision-makers (n = 12) and for patients/families (n = 15) were surveyed again. In this round, we also asked the participants to rate the additional value statements suggested by the participants in Round 2. In Round 3, a total of 4 decision-makers and 12 patients/families participated and reconsidered their previous response in the context of the group response to each benchmark. The consensus results of the Round 3 survey are summarized in Table 3 for decision-makers and Table 4 for patients/families. In the decision-makers' Round 3, three additional value statements reached consensus. In the patient/family Round 3, two additional value statements reached consensus.
The value statements that received high level of consensus in Round 3 are available in the S1 Supporting Information.
Decision-makers and the patients/families' means and consensus levels were then compared to identify which value statements were deemed important (mean>4.0) and obtained consensus (>90%) in each of the groups. The value statements that were important to both group include the following:

Discussion
This study resulted in a consensus-based list of value statements associated with safe personand family-centered care transitions from hospital to home, derived from both the decisionmakers and patients/families' perspectives. The high priority value statements which achieved consensus across both patients/families and decision-makers' groups (n = 30) represent key elements of care transitions that could enable patients as well as decision-makers to determine if their transition was successful or not. These statements are foundational to building a learning system that can grow and improve with each patient care episode. The value statements that reached consensus (>90%) and were important (mean>4.0) in both groups spanned the full transition experience from hospital admission to post-discharge follow up. We acknowledge the relatively limited attention span and 'improvement fatigue' in health care delivery. Presenting a list of 30 items for care improvement will likely be met with considerable push back. One strategy to counter this might be to examine the current list through the lens of what metrics/performance indicators may already address some of the recommended/ high concordance approaches. This would help with prioritization of the recommendations.
Additionally, several value statements that were identified as a high priority for decisionmakers were not priorities for patients/families, specifically the patient-reported outcome measures (PROMs) including: health status, activities of daily living/social activities, missed work, quality of life, symptoms (fatigue), and distress (anxiety or depression). This appears to be consistent with the work of others [36] where patients have shown less interest in measures that are dependent on their own subjective factors rather than system/provider factors. Furthermore, the 'post-discharge phone call' value statement did not reach consensus and it was not identified as being important by both groups. With the increased focus on patients and families' involvement as active partners on the healthcare team, it is important that they are part of the priority setting for care transitions as their involvement will influence the improvement efforts [37].
To our knowledge, there has been no previous research that has focused on identifying value statements to monitor safe person-and family-centered care transition interventions.
Our study was also focused on engaging patients and their families, and specifically used instruments-value statements-which were more accessible to people that were not in the health care field, as this engagement is an integral aspect of person-and family-centered care. This allowed us to gain rich insights into the perspectives of both the decision-makers and the patients and their families. Moreover, these experts were gathered from both home care organizations and acute care settings across five provinces across the country. Another strength of this study was that it was done in multiple rounds over a lengthy period of time. The passage of time and potentially changing levels of health can alter an individual's perception of which statements are most important. This allowed patients/families time to truly reflect on the most valuable statements and change their responses.
The Delphi technique proved helpful in systematically identifying and gaining consensus, on a core set of appropriate priorties from patients/families and decision makers where no consensus had previously existed. Careful consideration was necessary in relation to understanding the Delphi process, identification of 'experts', questionnaire design, agreement on an appropriate level of consensus and the number of rounds to conduct.
The 30 value statements on which consensus was gained will enable patients/families and decision makers to identify gaps in their practice against their peers, encourage improvement in care transitions and establish 'standards' of what types of care are feasible. The value statements may also be used by health care providers to monitor, evaluate, and improve the quality of care provided during care transitions. It is important that these priorities are linked with other research initiatives aimed at addressing the quality of care transitions as a whole. Although this study is an important step towards routinely measuring the quality of care provided to patients during care transitions, it is important that it is incorporated into a process of continuous quality improvement.
Because the respondents were based across Canada, it was not possible to meet face to face in a consensus conference or to take part in nominal groups. Therefore, the web-based Delphi approach was selected as the most appropriate and the most relevant research approach. Although we sought representation across Canada, not all experts responded to our invitations to participate in the survey. Direct care providers (i.e, nurses, physicians, or social workers) who are involved in discharge and care transition planning were not involved in the study. Furthermore, it is possible that patients/families who participated had greater interest in their health and in promoting change due to their involvement as patient and family advisors. Thus, their rating of the level of importance of each value statement could be higher. The patients/ families could also have been a healthcare professional in the past.
The web-based Delphi worked well in this study. The reason for this may be because most of the experts identified in the Delphi sample had easy access to e-mail and they use it as the main form of communication. The advantages of the web-based Delphi are obvious; not only is it an environmentally friendly way to carry out research, it also lead to more rapid responses from participants and it sped up the analysis as the electronic responses could be directly uploaded into SPSS. In addition, reminder e-mails can be sent out automatically, and there is no cost in terms of postage or printing. It is also possible that an electronic questionnaire where the busy respondent sees one page at a time is perceived as being easier to commence than a full printed questionnaire. The disadvantages include the possibility that not all Delphi experts would have an e-mail account-although this is getting less common. Furthermore, as with all questionnaires it is possible that busy people will complete the web-based Delphi in a casual fashion or may decide not to participate. For some managers who may be potential Delphi experts, it is often the case that their secretaries or personal assistants have access to their e-mail accounts and this may threaten response anonymity. It is important that all e-mails are labelled as strictly private and confidential. Finally, the sensitivity of computer firewalls in some organisations may block web-based Delphi questionnaires or direct them into a junk folder.
Next steps in this research will be to work alongside other organizations to help inform the ongoing development of value statements, interventions and standards that are meaningful and reflect the perspectives of both decision-makers and patients and their families. The consensus-based list of value statements found in this study can also inform the development of safe person-and family-centered care transition interventions as many health care organizations are seeking to develop and prioritize improved care transitions as part of their quality rubric. Further research should also focus on the barriers that different organizations and systems may face in operationalizing these value statements.

Conclusion
The present study was an important first step in identifying key consensus-based priority value statements for monitoring care transitions from hospital to home from the perspective of both the decision-makers and the patients/families. Further research is needed to test their usability and to determine whether these value statements are actually suggestive of safe and effective person-and family-centered care transition interventions from hospital to home.