Clinical Surveillance Tool to Screen for Unmet Palliative Needs Among Patients in the Final Year of Life

NCT ID: NCT04171830

Last Updated: 2023-11-02

Basic Information

• Recruitment Status: UNKNOWN
• Clinical Phase: NA
• Total Enrollment: 3536 participants
• Study Classification: INTERVENTIONAL
• Study Start Date: 2019-06-17
• Study Completion Date: 2024-03-31

Brief Summary

One of the most important obstacles to improving end-of-life care is the inability of clinicians to reliably identify those who are approaching the end-of-life. Every aspect of a palliative approach to care - screening for unmet needs, treating symptoms, discussing goals of care, and developing a palliative management plan - depends on the reliable and accurate identification of patients with palliative needs. The investigators developed an accurate and reliable mortality prediction tool that automatically identifies patients in hospital at elevated risk of death in the coming year. In previous studies it has been shown that these patients also frequently have unmet palliative care needs at the time they are identified by the tool. This tool has been demonstrated feasible, acceptable to patients and providers, and effective for changing physician behaviour in an inpatient clinical context.

In this project, this tool is implemented as part of an integrated knowledge translation project to facilitate reliable and timely identification of unmet palliative needs across multiple hospitals with different clinical settings and contexts. The investigators have partnered with 12 hospitals to improve the quality of palliative and end-of-life care provided to patients and families. With each partner site the investigators will develop a comprehensive implementation plan, including stakeholder engagement, education, and feedback. Process measures will be collected at each site to determine whether the tool was effective for promoting the identification and documentation of unmet palliative needs. Patients who were identified by the tool will also be followed over time to collect outcome and impact measures to see if their end-of-life care was affected by the intervention compared to control groups.

Detailed Description

Recently, van Walraven et al described the Hospital One-year Mortality Risk (HOMR) score for predicting 1-year mortality for patients admitted to hospital. HOMR is based on 12 administrative data points routinely coded by hospitals at the time of discharge and available in the CIHI Discharge Abstract Database. The model has been externally validated with excellent discrimination and calibration. Among HOMR's 12 data fields, nine are routinely available in the Electronic Health Record (EHR) at the time of admission in Ontario. Using a method similar to that used to derive HOMR, the investigators developed a "modified" HOMR (mHOMR) model based on the nine data fields available at the time of admission. mHOMR had comparable accuracy to HOMR (C-statistic .89 vs .92, respectively). Additionally, an updated version of mHOMR has recently been developed and validated, called HOMR Now!, which has the same c-statistic as the original HOMR (.92) but is calculated using ten data fields and an interaction available in many hospital admissions data, similar to mHOMR. Using either mHOMR or HOMR Now!, hospitals are able to retrieve admissions data from the EMR and calculate each patient's mortality risk on admission. If any patient's mortality risk exceeds a predefined threshold, the application would send a message to their clinical team prompting them to assess and address unmet palliative needs.

mHOMR has been implemented in four hospitals in Ontario to date and has been adapted to work with different EHRs. The mHOMR application identified a gender-balanced cohort of generally elderly patients (mean age of 83 years) who were admitted for several days (median length of stay of 5 days) and discharged alive (89%), meaning they were not in their final days of life and there would be an opportunity to screen for unmet needs and participate in care planning. A second pilot study found >90% of patients identified by the application had an unmet palliative need- either a severe symptom or a desire to discuss ACP with a physician or both-and that patients with higher mHOMR scores had more severe symptoms. The application preferentially identified patients with non-cancer illnesses-most were admitted with a frailty-related condition (56.8%), followed by end-stage organ failure (23.5%), and cancer (20%)-meaning that the tool did not show a bias towards cancer but instead identified patients who reflected the actual population of dying Canadians. These results are similar to findings from HOMR Now! validation work. Furthermore, investigators found <50% of those identified by mHOMR had a documented palliative care consultation or Goals of Care discussion, but after the integration of mHOMR notifications into existing workflow, the incidence of early Goals of Care discussions and palliative care consultation increased significantly. Additionally, qualitative results show the application is acceptable to patients and clinicians alike.

Both the mHOMR and HOMR Now! applications are intended to be a reliable and accurate "trigger" to improve the effectiveness of any palliative intervention by focusing attention on a small group of patients with a high risk of death and unmet palliative needs. Both applications can also be versatile depending on the situation-it produces a numerical risk output rather than a binary yes/no like the Surprise Question, Gold Standards Framework or NECPAL tools, so the user can decide what threshold to use for identifying "high risk" patients. Thus, organizations concerned with the efficient use of limited resources could set a higher mortality threshold, while organizations using more scalable interventions could lower the mortality threshold.

Given the initial success of the mHOMR and HOMR Now! applications in identifying unmet palliative care needs in an acceptable way among patients nearing the end-of-life, the next step is to implement and rigorously evaluate the immediate long-term effects of this highly scalable intervention in a large population to determine whether it improves screening and documentation processes and ultimately leads to better outcomes for patients, family members, and the healthcare system as a whole. To achieve this aim, investigators have partnered with twelve acute care hospitals from across Ontario to implement the mHOMR/HOMR-Now! intervention.

Objective

To determine whether implementation of an mHOMR or HOMR Now! application to identify patients at increased risk of death and trigger screening for unmet palliative needs improves (1) identification and documentation of those needs and (2) the end-of-life care provided to patients.

Intervention-Implementation Procedures

Every inpatient at each site will automatically be given the intervention (an mHOMR or HOMR Now! Score, depending on which application can be most easily integrated into each site's existing EMR system) upon admission to an implementing unit at a participating hospital and considered for secondary interventions (i.e. palliative care) based on their score.

At a minimum, each individual identified by the chosen HOMR tool should receive two additional assessments to screen for severe symptoms and the patient's desire to engage in advanced care planning (ACP):

1. Edmonton Symptom Assessment System Revised (ESAS-R): scores of >6 will be flagged as 'severe'. Individual clinical teams can then choose to address the symptoms as appropriate for the patient, or consult a PC team.

2. 4-item Advanced Care Planning Engagement Survey: Scores of 3-4 indicate a patient is ready to discuss ACP with a member of the clinical team. Clinical teams may choose to discuss ACP and goals of care (GoC) themselves, activate a local ACP/GoC intervention, or distribute ACP documentation (e.g. SpeakUp resources), as applicable.

Implementation of the mHOMR/HOMR-Now! intervention in each site will follow 4 phases informed by the Quality Implementation Framework:

1. Site-Specific Considerations

To facilitate successful implementation of the mHOMR/HOMR-Now! application, three strategies will be used to tailor implementation to the site-specific context, including needs, resources, fit, capacity, and readiness. Firstly, members of the coordinating and implementation research team will virtually conduct a detailed readiness assessment to determine the best way to implement the application given the local context of each site. Secondly, semi-structured focus group interviews will be held virtually at each site with the Implementation team (i.e. an executive champion, implementation lead, clinical lead, and information technology lead), as well as staff who will interact with the HOMR application output (i.e. either recipient of or actors on the notification). Lastly, select members of the implementation team will be virtually interviewed individually to determine site-specific implementation barriers and facilitators.

Considering the unique site context, investigators will then work with stakeholders to determine the additional interventions that each site will implement once a patient has been identified as being at elevated risk for mortality and unmet palliative needs by the application. Notably, determination of secondary interventions will consider the normal workflow and resources available at each site. Details of these interventions will be provided in each site's individual protocol.

2. Establishment of Information Technology Infrastructure for Implementation at each Site

Logistics of the mHOMR or HOMR-Now! application will be discussed with the IT lead at each site to determine the technical approach for implementation in the electronic health record (EHR), including which application would be most appropriate to implement given each site's existing EHR. Some EHR platforms are used by more than one site; thus, solutions derived for one EHR will be shared among other partners as appropriate.

Once the specific technical implementation process has been defined for each site, electronic and print educational material will be developed to teach staff at each site about mHOMR/HOMR-Now!, how the application works, and steps to take when they receive a notification.

3. Development, Deployment, and Ongoing Support for mHOMR/HOMR-Now! Implementation

Integration of the application into each site's EHR will be managed by the site IT lead and tested to ensure the application is correctly calculating the mortality risk scores and notifying the appropriate members of the care team. Once this is complete, each site will host a "go-live" kick-off event to help generate awareness, enthusiasm, and uptake of the intervention.

In the following six to nine months (depending on the site's funding source), the application will be 'live' at each site, actively identifying newly admitted patients and notifying care teams to conduct the ESAS-R and 4-item ACP Engagement Survey as appropriate. Any additional site-specific interventions will also be implemented.

Process and outcomes evaluations will also occur during phase 3. In addition to phase 1 interviews and focus groups, these evaluations will involve: (1) a second set of semi-structured interviews with members of the implementation team at each site to examine determinant factors associated with successful implementation of the mHOMR application; (2) a chart review to examine clinical and implementation outcomes, and; (3) analysis of linked health administrative data held at ICES to evaluate long-term clinical outcomes.

4. Continuous Improvement of mHOMR/HOMR-Now! Implementation (Concurrent with Phases 1-3)

Each site will be regularly updated of study progress through teleconferences and newsletters. These communications will also share learnings across sites to improve implementation through establishment of best practices and identification of strategies to overcome implementation barriers. Additionally, this process will inform implementation of the application in other hospitals in the future.

All clinical secondary outcomes will be measured for HOMR positive patients for six to nine months pre and post implementation, and followed until end of study follow-up (up to one year after hospital discharge) or death. While each of these outcomes will be measured, aggregated, and linked to databases held at ICES, results from individual site data will be used to improve patient care by driving existing clinical best practices for palliative care, symptom management, and advanced care planning. In this sense, the clinical secondary outcome measures will also serve as indicators for continuous quality improvement at each hospital.

Conditions

Palliative Therapy; Implementation Science; Screening; Terminal Care

Study Design

• Allocation Method: NA
• Intervention Model: SINGLE_GROUP
• Primary Study Purpose: SCREENING
• Blinding Strategy: NONE

Interventions

modified Hospital One-year Mortality Risk (mHOMR)

Every inpatient will automatically be given the intervention (an mHOMR/HOMR Now! score) upon admission to hospital and considered for secondary interventions based on their score. Baseline threshold will be set as >0.21 (59% sensitivity and 90% specificity for 12-month mortality with mHOMR).

At minimum each individual identified by the tool will receive two assessments to screen for severe symptoms and desire to engage in advance care planning (ACP):

1. Edmonton Symptom Assessment System Revised: scores >6 will be flagged as severe. Clinical teams will address the symptoms as appropriate for the patient.

2. 4-item Advanced Care Planning Engagement Survey: Scores of 3-4 indicate readiness to discuss ACP. Clinical teams may choose to discuss ACP and goals of care themselves, activate a local ACP intervention, or distribute ACP documentation.

Both of these assessments will be done by a member of the treating team within 72 hours of the patient's hospital admission.

Intervention Type: OTHER

Other Intervention Names

Hospital One-year Mortality Risk-Now! (HOMR-Now!)

Eligibility Criteria

Inclusion Criteria

• All newly admitted patients to selected medical units in participating sites during the 6-9-month intervention implementation period
• [To be assessed for unmet palliative needs] the patient must be competent and have the ability to participate in assessments (i.e. answer assessment questions and understand and speak sufficient English to participate).

Exclusion Criteria

• N/A for mHOMR/HOMR-Now! intervention
• For palliative needs assessments: incapability of completing the ESAS and 4-Item ACP tools, either because of capacity/cognitive impairment or English-language ability.

• Eligible Sex: ALL
• Accepts Healthy Volunteers: No

Sponsors

Canadian Frailty Network

OTHER

Sponsor Role: collaborator

Centre for Aging and Brain Health Innovation

OTHER

Sponsor Role: collaborator

Canadian Foundation for Healthcare Improvement

UNKNOWN

Sponsor Role: collaborator

Bruyère Health Research Institute.

OTHER

Sponsor Role: collaborator

William Osler Health System

OTHER

Sponsor Role: collaborator

Hopital Montfort

OTHER

Sponsor Role: collaborator

Queen's University

OTHER

Sponsor Role: collaborator

Pembroke Regional Hospital

UNKNOWN

Sponsor Role: collaborator

Unity Health Toronto

OTHER

Sponsor Role: collaborator

Cambridge Memorial Hospital

OTHER

Sponsor Role: collaborator

The Ottawa Hospital

OTHER

Sponsor Role: collaborator

Queensway Carleton Hospital

OTHER

Sponsor Role: collaborator

University of Ottawa

OTHER

Sponsor Role: collaborator

The Hospital for Sick Children

OTHER

Sponsor Role: collaborator

ICES

INDUSTRY

Sponsor Role: collaborator

London Health Sciences Centre

OTHER

Sponsor Role: collaborator

Windsor Regional Hospital

OTHER

Sponsor Role: collaborator

Humber River Hospital

OTHER

Sponsor Role: collaborator

North York General Hospital

OTHER

Sponsor Role: collaborator

Ontario Health - Quality

UNKNOWN

Sponsor Role: collaborator

Healthcare Excellence Canada

UNKNOWN

Sponsor Role: collaborator

Headwaters Health Care Centre

UNKNOWN

Sponsor Role: collaborator

Ottawa Hospital Research Institute

OTHER

Sponsor Role: lead

Responsible Party

James Downar

Principal Investigator

Responsibility Role: PRINCIPAL_INVESTIGATOR

Locations

William Osler Health System

Brampton, Ontario, Canada

Cambridge Memorial Hospital

Cambridge, Ontario, Canada

Kingston Health Sciences Centre

Kingston, Ontario, Canada

London Health Sciences Centre

London, Ontario, Canada

Headwaters Health Care Centre

Orangeville, Ontario, Canada

The Ottawa Hospital

Ottawa, Ontario, Canada

Montfort Hospital

Ottawa, Ontario, Canada

Queensway Carleton Hospital

Ottawa, Ontario, Canada

Pembroke Regional Hospital

Pembroke, Ontario, Canada

North York General Hospital

Toronto, Ontario, Canada

Humber River Hospital

Toronto, Ontario, Canada

Saint Michael's Hospital

Toronto, Ontario, Canada

Windsor Regional Hospital

Windsor, Ontario, Canada

Countries

Canada

References

van Walraven C. The Hospital-patient One-year Mortality Risk score accurately predicted long-term death risk in hospitalized patients. J Clin Epidemiol. 2014 Sep;67(9):1025-34. doi: 10.1016/j.jclinepi.2014.05.003. Epub 2014 Jun 25.

Reference Type: BACKGROUND

PMID: 24973823 (View on PubMed)

van Walraven C, McAlister FA, Bakal JA, Hawken S, Donze J. External validation of the Hospital-patient One-year Mortality Risk (HOMR) model for predicting death within 1 year after hospital admission. CMAJ. 2015 Jul 14;187(10):725-733. doi: 10.1503/cmaj.150209. Epub 2015 Jun 8.

Reference Type: BACKGROUND

PMID: 26054605 (View on PubMed)

van Walraven C, Forster AJ. The HOMR-Now! Model Accurately Predicts 1-Year Death Risk for Hospitalized Patients on Admission. Am J Med. 2017 Aug;130(8):991.e9-991.e16. doi: 10.1016/j.amjmed.2017.03.008. Epub 2017 Mar 31.

Reference Type: BACKGROUND

PMID: 28366426 (View on PubMed)

Wegier P, Koo E, Ansari S, et al. mHOMR: A pilot study of automated prospective clinical surveillance for inpatients having an elevated risk of one-year mortality. Under Review.

Reference Type: BACKGROUND

Watanabe SM, Nekolaichuk C, Beaumont C, Johnson L, Myers J, Strasser F. A multicenter study comparing two numerical versions of the Edmonton Symptom Assessment System in palliative care patients. J Pain Symptom Manage. 2011 Feb;41(2):456-68. doi: 10.1016/j.jpainsymman.2010.04.020. Epub 2010 Sep 15.

Reference Type: BACKGROUND

PMID: 20832987 (View on PubMed)

Sudore RL, Heyland DK, Barnes DE, Howard M, Fassbender K, Robinson CA, Boscardin J, You JJ. Measuring Advance Care Planning: Optimizing the Advance Care Planning Engagement Survey. J Pain Symptom Manage. 2017 Apr;53(4):669-681.e8. doi: 10.1016/j.jpainsymman.2016.10.367. Epub 2016 Dec 29.

Reference Type: BACKGROUND

PMID: 28042072 (View on PubMed)

Meyers DC, Durlak JA, Wandersman A. The quality implementation framework: a synthesis of critical steps in the implementation process. Am J Community Psychol. 2012 Dec;50(3-4):462-80. doi: 10.1007/s10464-012-9522-x.

Reference Type: BACKGROUND

PMID: 22644083 (View on PubMed)

Metz A, Louison L. The Hexagon Tool: Exploring Content. Chapel Hill, NC: National Implementation Research Network, Frank Porter Graham Child Development Institute, University of North Carolina at Chapel Hill; 2018:1-5.

Reference Type: BACKGROUND

Other Identifiers

1590

Identifier Type: -

Identifier Source: org_study_id