Modeling changes in assessments to predict needs and guide care planning in home care

This study developed a predictive frailty measure for seniors receiving home care by applying advanced analytic methods to data from the Resident Assessment Instrument – Homecare (RAI-HC). Because the new frailty measure was based on the RAI-HC, it can be used in multiple Canadian jurisdictions as well as other countries.

Possible Research Results

Findings: A wide variety of measurements are available in the RAI-HC for predicting risk for residential care or mortality (e.g., health conditions, physical/cognitive impairment, health instability). Although recent measures are valuable, our approach identifies changes occurring in the older adult over time that allows prediction of future care needs.

Impact of findings: Armed with better information, clinicians will be able to identify seniors at highest risk for transitions and direct care resources where they can most effectively make a difference. In additions, some aspects of how we make predictions are more important than others and identifying these factors will be helpful at informing clinical practice and forming policy. Ultimately, this work improves the ability to predict a senior’s home care needs and also identifies what aspects of their health is most important in that regard. With greater attention to the care needs that most predict moving from home to residential care, the senior will have an opportunity to stay at home longer with appropriate support and at a lower cost to government.

About the Project

Although RAI-HC data for tracking change in client status over time is readily available, the predictive value has been largely untapped. RAI-HC application has generally been limited to evaluating outcomes at one point in time.

This project mathematically modeled an individual’s changes in RAI-HC assessments to predict likely changes in the level of care required for an individual. Further, it translated this model into a measure, derived from RAI-HC assessments, for use by home care clinicians, physicians and other care providers. The measure was piloted by working with case managers in Island Health to identify relevant cases and providing results from the new predictive frailty measure to help guide decisions.

Once validated, this new frailty measure will provide a new level of guidance for key decisions, especially the timing of transition from the home to alternative living arrangements. The intent of this eHome (electronic Home Monitoring to Empower)-iCare (inform Caregivers And Robotic Effectors) project was to reduce unnecessary or premature transitions.

Project Team

Principal Investigators:

Debra Sheets, PhD, MSN, RN, FAAN — University of Victoria

Stuart MacDonald, PhD — University of Victoria

Co-Investigators:

Carl Asche, PhD — University of Illinois, College of Medicine at Peoria

Cheryl Beach, PhD — Vancouver Island Health Authority

Paul Brewster, PhD — University of Victoria

Sandra Hundza, PhD — University of Victoria

Andrew Mitz, PhD — National Institutes of Health

Jeff Poss, PhD — University of Waterloo

Knowledge Users and Partners:

Marilyn Malone, MD, FRCPC — Vancouver Island Health Authority

Office of Seniors Advocate, B.C. Ministry of Health

University of Victoria – Faculty of Human and Social Development

University of Waterloo – School of Public Health and Health Systems

Project Contact: Debra Sheets — dsheets@uvic.ca

Rationale, Hypothesis, Objectives & Research Plan

Rationale: The Resident Assessment Instrument-Home Care (RAI-HC) is a routine assessment widely used in Canada for making care decisions for older adults living in the community. Typically, clinicians make decisions using only the most recent assessment. This study shows how a mathematical approach applied to several assessments for a client can make sense out of changes and inform better decisionmaking.

Hypothesis: We propose that changes in the health assessment of an individual done at multiple times is useful in predicting risk for transition.

Objectives: We are applying advanced mathematical tools to RAI-HC data to show how multiple assessments can more accurately predict the
client’s future than using a single assessment. In addition, we are getting feedback from clinical experts on the usefulness of the new frailty tool in their decision making.

Research plan: Our research plan has two stages: (1) Use the huge national database of RAI-HC assessments provided by the Canadian Institute for Health Information to guide the development of equations that can be used to help predict an individual’s future needs, and (2) present results to clinical users and educators of the RAI-HC to receive feedback on the feasibility, practicality and utility of our findings.

Communication to Policy Makers

Key Findings:

  • The frailty change tool based on homecare assessments can help us predict risk for transitions to higher levels of care (e.g. home to residential care).
  • The frailty change tool offers clinicians a useful process for decision-making that is based on meaningful changes in the individual’s function and health rather than comparison with population norms.
  • Analyzing assessment data from multiple points in time can help avoid premature placement in residential care by identifying increasing needs for home care services that allow earlier intervention.
  • The frailty change tool based on homecare assessments can help us predict risk for transitions to higher levels of care (e.g. home to residential care).
  • The frailty change tool offers clinicians a useful process for decision-making that is based on meaningful changes in the individual’s function and health rather than comparison with population norms.
  • Analyzing assessment data from multiple points in time can help avoid premature placement in residential care by identifying increasing needs for home care services that allow earlier intervention.

Why was this study needed?

Our study developed a frailty change tool using assessments from multiple time points to predict transition risk in level of care for home care clients. The frailty change tool may be useful for guiding care planning and clinical decision-making. It may prevent premature transition from home to subsequent levels of care.

Suggestions on how administrators or policy maker could use the findings:

  • Understanding the importance of changes from an individual’s baseline (rather than using population norms) is important clinical information that can makes healthcare more person-centered.
  • Change trajectories using assessment scores at multiple times offer more insight to inform clinical decision making than using single assessments.
  • Frailty change scores can be useful in care planning and may delay placement for clients living with frailty by allowing earlier interventions.

Brief comment on type of study:

  • We aimed to create a frailty change tool that can predict when a client leaves home care (e.g. transition to residential care or death).
  • The sample was 314,851 home care clients, age 65 and over, living in Ontario in the community at the time of the first assessment.
  • A total of 1,328,722 assessments from the Home Care Reporting Services (HCRS) database from the Canadian Institute of Health Information (CIHI).
  • Our main outcome was predicting risk of transition to subsequent levels of care (i.e. from home care to residential care).
Communication to Researchers

Key Findings:

  • The frailty change tool based on homecare assessments can help us predict risk for transitions to higher levels of care (e.g. home to residential care)
  • The frailty change tool offers clinicians a useful process for decision-making which is based on meaningful changes in the individual’s function and health rather than comparison with population norms.
  • Analyzing assessment data from multiple points in time can help avoid premature placement in residential care by identifying increasing needs for home care services that allow earlier intervention.
  • We found 413 articles that identified participants as frail without measuring frailty, and 204 articles that measured frailty.
  • The majority (81%) of research that measured frailty were observational studies.
  • Most (60%) of the reported frailty measures were developed to measure frailty; however, 27% of frailty measures used operational definitions for the purpose of that study and 2% defined frailty using clinical judgement.
  • The Clinical Frailty Scale, Frailty Index, and Frailty Phenotype were the most commonly reported scales.
  • 44% of studies used frailty to predict adverse health outcomes, 20% used frailty as an inclusion and exclusion criterion, and 4% used frailty as an outcome measure. One third of studies used frailty for descriptive purposes only.
  • Most studies showed that frailty tools are predictive of adverse health outcomes, specifically mortality (84% of cases) and institutionalization (93% of cases).

Why was this study needed?

Our study developed a frailty change tool using assessments from multiple time points to predict transition risk in level of care for home care clients. The frailty change tool may be useful for guiding care planning and clinical decision-making.  It may prevent premature transition from home to subsequent levels of care.

Brief overview of the methodology:

  • This study created a frailty change tool to predict risk of transition from home care to other levels of care (e.g. residential).
  • The sample included 314,851 home care clients living in Ontario in the community.
  • Analyses were based on 1,328,722 assessments using the Home Care Reporting Services (HCRS) database from the Canadian Institute of Health Information (CIHI).
  • Exclusion criteria included: non-Ontario residents, less than 65 years of age, less than two home care assessments, missing data, or if the first home care assessment was not in the community.
  • Clients had significant chronic conditions—33% had dementia, 52% had decline in daily living activities, 89% had difficulty with IADL’s, and 29% lived alone.
  • Our key outcome was predicting risk of transition to subsequent levels of care (e.g., from home care to residential care).

Potential impact of findings on clinical practice/patient care and how this impact might be measured:

  • Understanding the importance of changes from an individual’s baseline (rather than using population norms) is important clinical information that can makes healthcare more person-centered.
  • Change trajectories using assessment scores at multiple times offer more insight to inform clinical decision making than using single assessments.
  • Frailty change scores can be useful in care planning and may delay placement for clients living with frailty by allowing earlier interventions.

Remaining knowledge/research gaps:                                                                              

  • The frailty change tool needs to be tested widely by healthcare professionals working with home-care clients.
  • The impact of the frailty change tool on decision-making, services utilization and health outcomes needs to be demonstrated.