Pauline Wu, BSc1, Zahra Goodarzi, MD, MSc1-4, Jacqueline McMillan, MD, MSc1-3
1Department of Medicine, University of Calgary
2Department of Community Health Sciences, University of Calgary
3O’Brien Institute of Public Health, University of Calgary
4Hotchkiss Brain Institute, University of Calgary, Calgary, ABDOI: https://doi.org/10.5770/cgj.28.826
ABSTRACT
Older adults may require longer recovery periods prior to being discharged from the hospital after an acute care stay. For some, returning to their previous living arrangement may no longer be safe or feasible after an acute care admission, and they may require alternate levels of care. It can be challenging to evaluate which patients may benefit most from inpatient rehabilitation versus those for whom alternate levels of care are more suitable. Using a prioritization procedure, this study identified and ranked predictive factors for successful inpatient rehabilitation (defined as discharge to previous living arrangement) from most to least important. The final round of the prioritization procedure resulted in a list of the top 20 predictive factors, ranked by health-care providers in the field, from most to least important. Predictive factors included demographic information, past medical history factors, acute care illness factors, and results of investigations performed during the index hospitalization. The top ranked predictive factors related to patients’ previous living arrangements, level of independence before hospitalization, and presence or absence of cognitive impairment. The bottom ranked predictive factors related to physical measures and results of inpatient investigations at the time of transfer. These findings highlight the importance of considering patients’ lived experiences prior to hospitalization when determining who may obtain the greatest benefit from further, intensive inpatient rehabilitation following an acute care hospitalization.
Key words: rehabilitation, hospitals, rehabilitation, clinical decision rules, frail elderly, geriatrics
In 2022–2023, there were nearly 2.96 million acute inpatient hospitalizations in Canada, with an age-adjusted average length of hospital stay of 7.3 days.(1) Older adults experience greater comorbidity(2) and frailty(3) that often require longer recovery periods prior to discharge from hospital.(4) Moreover, those with more comorbidities are at higher risk for hospital readmission.(5) For some older adults, returning to their previous living arrangement may no longer be safe or feasible after acute care admission, requiring alternate levels of care.(6) During the time of acute illness, trauma, or perioperative state, it can be challenging to evaluate which patients may benefit most from inpatient rehabilitation versus those for whom alternate levels of care are more appropriate.
Informed by the needs of the end-user, we have taken a pragmatic approach to understanding the factors that impact inpatient rehabilitation outcomes for older adults. This is a first step in an integrated knowledge translation approach to the creation of a prediction tool. Our approach was informed by priority-setting partnership methods, with a focus on ensuring all participants have an equal voice in sharing their lived experience. Prediction models enable clinicians to better advise on lifestyle or therapeutic interventions and incorporate multiple factors (e.g., demographics, medical history, test results) to predict outcomes.(7,8) Predicting outcomes based on single predictors would be insufficient to calculate accurate estimates.(8) While some factors may be associated with successful rehabilitation or the need for alternative living arrangements, it is not clear how they work in combination and predict specific outcomes. With a multivariable prediction model, multiple predictors for a single individual are considered for the diagnosis or prognosis of health outcomes.(8) Once the model is developed, it is validated for use in a clinical setting.(9) This represents a knowledge gap in who may benefit most from further inpatient rehabilitation after acute care admission. Given the increasing older adult population and system-wide pressures, this work is aimed at helping patients understand their options, and to optimize bed flow in a system experiencing significant challenges with overcapacity.
We aim to help patients understand their options, and to optimize bed flow in a system experiencing significant challenges with overcapacity, by identifying factors that predict individual patients’ likelihood of successful rehabilitation after an acute care hospital admission (defined as discharge to their previous living arrangement).
We completed a prioritization exercise to analyze which predictive factors accurately identify patients who will benefit most from inpatient rehabilitation after an acute care admission. We planned to conduct three rounds of surveys. Surveys were hosted online by QualtricsXM (Seattle, WA; www.qualtrics.com).
We identified and created an initial list of predictive factors separated into four categories based on the clinical experience and expertise of the study team investigators. Chosen predictive factors needed to be accessible and clearly defined.(10)
Round One participants were asked to review the initial list of predictive factors and suggest additional factors not already included. In Round Two, additional factors provided by Round One participants were reviewed for appropriateness, repetition, feasibility, objectivity, and reliability of data collection by two authors (J.M. and Z.G.). Once the additional factors were reviewed and finalized, they were added to the initial list of predictive factors, resulting in the Round Two survey. Participants were then asked to rank from most (1) to least important (n) in each category. In the final round, the top predictive factors across all categories were identified, and participants were asked to rank them from most (1) to least (20) important.
The study population included health-care providers (e.g., geriatricians, care of the elderly physicians, nursing staff, physical therapists, occupational therapists) in Alberta involved in the clinical care of older adults who receive inpatient rehabilitation. Care providers not providing care to older adults and/or not involved in the pathway from acute care hospitalization through to inpatient rehabilitation were excluded. All participants provided consent prior to completing each survey. This study was approved by the University of Calgary Conjoint Health Research Ethics Board (CHREB23-0745).
Statistical analyses were conducted using Microsoft Excel (version 2402) (Microsoft Corporation, Redmond, WA; www.microsoft.com). We calculated median, interquartile range (IQR), mean, standard deviation, and mode. Medians (IQR) were used to rank participant responses. If two or more predictive factors ranked the same, the predictive factor with the smaller IQR was ranked higher.
Twenty individuals participated in Round One of the prioritization procedure. Demographic information was not collected. The initial list of predictive factors included four categories: patient demographic factors (n= 9), patient illness factors (n= 18), patient medical history factors (n= 11), and results of investigations performed during the index hospitalization (n= 9) (see Table 1). A total of 19 additional factors were added to the original list based on participants’ suggestions, forming the Round Two survey. Additional factors were evaluated for appropriateness, repetition, feasibility, objectivity, and reliability of data collection.
TABLE 1 Round One categorized predictive factors identified by the study team
Round Two included 15 participants who ranked factors across each section including the factors added after round one. The top predictive factor from the patient demographic category was “Independence with basic activities of daily living (immediately prior to hospitalization)” (Median = 2, IQR = 2.5). The top predictive factor from the patient illness category was “Number of days from admission to the time of initiating physical or occupational therapy” (Median = 5, IQR = 7). The “Presence or absence of a diagnosis of any active malignancy (i.e., not a previously treated cancer or a cancer in remission)” (Median = 3, IQR = 10.75) was rated highest from the patient past medical history category. Finally, “Change in weight from time of index admission to hospital to time of transfer to the acute geriatric unit” (Median = 2.5, IQR = 3) ranked highest in the results of investigations performed during the index hospitalization category. See Table 2 for Round Two rankings.
TABLE 2 Ranked predictive factors from Round One, including those added by participants, using median and interquartile range (IQR) according to factor category
Finally, Round Three included 13 individuals who ranked the 20 top predictive factors from Round Two from most to least important. Of the 20 predictive factors, four were from the patient demographic category, six from patient illness category, seven from patient past medical history category, and three from results of investigations performed during the index hospitalization category. Two predictive factors not in the top 20 from Round Two were included over higher ranked factors. These two factors are seen in Table 2 (Rank 8 under the patient illness category, and Rank 4 under the results of investigations performed during the index hospitalization category). In these cases, the decision was made to include data that are readily and accurately available in the health electronic records and which had greater face validity.(10) Participants ranked patient demographics highest, prioritizing patients’ living arrangements and level of independence immediately prior to the index hospital admission. Participants ranked results of investigations performed during the index hospitalization in the bottom three. The 20 ranked, predictive factors are presented in Table 3.
TABLE 3 Top 20 ranked predictive factors from round two using median and interquartile range (IQR)
This prioritization exercise identified 20 predictive factors that could help identify patients where inpatient rehabilitation after acute illness may be most beneficial. This work will help to inform further work on prediction modelling.
Prior function and living arrangements are key when predicting rehabilitation outcomes. The consensus of this prioritization procedure indicated that patient independence with activities of daily living and their living arrangements prior to admission were important predictive factors. These key factors are directly related to patients’ available resources after acute care and their baseline function vs. need for support, thus making key factors in recovery.
As older Canadians experience higher degrees of frailty and comorbidity, there are more acute care admissions and thus inpatient rehabilitation.(11) Recent Canadian research validated a mortality risk model for older adults in the home care setting and found, similar to our study, that measures of function (i.e., basic and instrumental activities of daily living) were the most important in predicting deterioration and death.(12) Additionally, older adults who live with others may be more likely to have family and friends who are able to provide support, whereas those who live in a supportive living environment prior to admission may be more frail and need additional support.(5)
Severity of illness while in acute care will impact rehabilitation. Requiring intensive care unit admission or experiencing delirium were also key factors when predicting rehab outcomes. Both represent more complicated and/or prolonged hospital stay and are associated with worsening function. This is especially the case given that we know delirium and intensive care unit stays during hospitalization are independently associated with increased risk of death, institutionalization, and incident dementia.(13)
Baseline comorbidities that limit function and exercise capacity will affect rehabilitation trajectories. Certain comorbidities, such as severe chronic obstructive pulmonary disease, end-stage heart failure, and any cancer were associated with a high risk of mortality, and poor baseline function and recovery post-acute exacerbation were also key factors.(14–16) For example, the hazard ratio for mortality increases with each increment in the Global Initiative for Chronic Obstructive Lung Disease (GOLD) classifications.(17) Similarly, in patients with severe asymptomatic aortic stenosis, the risk of death is 5% at one year(18) and rises to 50% for those with severe symptomatic aortic stenosis.(19) Presence of comorbidities can greatly impact patient outcomes in a clinical setting.
This study recruited participants from a single, urban setting. Additionally, we did not include patients or referring providers. These two considerations limit the generalizability of our results. Future research could expand on these results by collecting data across a broader provincial or national region and including patients and other providers in their recruitment strategies.
The factors ranked demonstrate a consensus of health-care providers’ experiences working in geriatric medicine and acute rehabilitation regarding which factors are most predictive of acute care rehabilitation outcomes. These results are the first step toward more evidence-based decision-making to provide equitable access to a scarce resource.
We would like to acknowledge the study participants for their time and expertise, as well as management and administrative staff for help with planning and dissemination of our study surveys.
The authors have read and understood the Canadian Geriatrics Journal’s policy on conflicts of interest disclosure and declare no conflicts of interest.
This research did not receive external funding.
1. CIHI. Hospital stays in Canada, 2022–2023 [Internet]. 2024 [cited 2024 Jul 25]. Available from: https://www.cihi.ca/en/hospital-stays-in-canada-2022-2023
2. Salanitro AH, Hovater M, Hearld KR, Roth DL, Sawyer P, Locher JL, et al. Symptom burden predicts hospitalization independent of comorbidity in community-dwelling older adults. J Am Geriatr Soc. 2012 Sep;60(9):1632–37.
Crossref PubMed PMC
3. Hsieh TJ, Su SC, Chen CW, Kang YW, Hu MH, Hsu LL, et al. Individualized home-based exercise and nutrition interventions improve frailty in older adults: a randomized controlled trial. Int J Behav Nutr Phys Act. 2019 Dec;16(1):119.
Crossref PubMed PMC
4. Terbraak M, Kolk D, Vroomen JLM, Twisk JWR, Buurman BM, Van Der Schaaf M. Post-discharge light physical activity indicates recovery in acutely hospitalized older adults—the Hospital-ADL study. BMC Geriatr. 2023 May 19;23(1):311.
Crossref PubMed PMC
5. Glans M, Kragh Ekstam A, Jakobsson U, Bondesson Å, Midlöv P. Risk factors for hospital readmission in older adults within 30 days of discharge—a comparative retrospective study. BMC Geriatr. 2020 Dec;20(1):467.
Crossref PubMed PMC
6. Werner RM, Coe NB, Qi M, Konetzka RT. Patient outcomes after hospital discharge to home with home health care vs. to a skilled nursing facility. JAMA Intern Med. 2019 May 1; 179(5):617–23.
Crossref PubMed PMC
7. Collins GS, Reitsma JB, Altman DG, Moons KGM. Transparent reporting of a multivariable prediction model for individual prognosis or diagnosis (TRIPOD): the TRIPOD Statement. Ann Intern Med. 2015 Jan 6;162(1):55–63.
Crossref PubMed
8. Moons KGM, Altman DG, Reitsma JB, Ioannidis JPA, Macaskill P, Steyerberg EW, et al. Transparent reporting of a multivariable prediction model for Individual Prognosis Or Diagnosis (TRIPOD): explanation and elaboration. Ann Intern Med. 2015 Jan 6;162(1):W1–W73.
Crossref PubMed
9. Su TL, Jaki T, Hickey GL, Buchan I, Sperrin M. A review of statistical updating methods for clinical prediction models [Internet]. [cited 2024 Jul 25]. Available from: https://journals-sagepub-com.ezproxy.lib.ucalgary.ca/doi/epub/10.1177/0962280215626466
10. Grant SW, Collins GS, Nashef SAM. Statistical primer: developing and validating a risk prediction model. Eur J Cardiothorac Surg. 2018 Aug 1;54(2):203–08.
Crossref PubMed
11. CIHI. Frailty among hospitalized seniors [Internet]. 2023 Nov 30 [cited 2024 Sep 11]. Available from: https://www.cihi.ca/en/frailty-among-hospitalized-seniors
12. Hsu AT, Manuel DG, Spruin S, Bennett C, Taljaard M, Beach S, et al. Predicting death in home care users: derivation and validation of the Risk Evaluation for Support: Predictions for Elder-Life in the Community Tool (RESPECT). Can Med Assoc J. 2021 Jul 5;193(26):E997–E1005.
Crossref
13. Witlox J, Eurelings LSM, de Jonghe JFM, Kalisvaart KJ, Eikelenboom P, van Gool WA. Delirium in elderly patients and the risk of postdischarge mortality, institutionalization, and dementia: a meta-analysis. JAMA. 2010 Jul 28;304(4):443–51.
Crossref PubMed
14. Hutchinson RN, Gutheil C, Wessler BS, Prevatt H, Sawyer DB, Han PK. What is quality end-of-life care for patients with heart failure? a qualitative study with physicians. J Am Heart Assoc. 2020 Sep 15;9(18):e016505.
Crossref PubMed PMC
15. Iyer AS, Sullivan DR, Lindell KO, Reinke LF. The role of palliative care in COPD. Chest. 2022 May 1;161(5):1250–62.
Crossref PMC
16. Tafuri A, Sandri M, Martini A, Capitanio U, Mantica G, Terrone C, et al. External validation of the Palacios’ equation: a simple and accurate tool to estimate the new baseline renal function after renal cancer surgery. World J Urol. 2022 Feb 1; 40(2):467–73.
Crossref
17. Guo C, Yu T, Chang LY, Bo Y, Yu Z, Wong MC, et al. Mortality risk attributable to classification of chronic obstructive pulmonary disease and reduced lung function: a 21-year longitudinal cohort study. Respir Med. 2021 Aug 1;184:106471.
Crossref PubMed
18. Campo J, Tsoris A, Kruse J, Karim A, Andrei AC, Liu M, et al. Prognosis of severe asymptomatic aortic stenosis with and without surgery. Ann Thorac Surg. 2019 Jul 1;108(1):74–79.
Crossref PubMed
19. Leon MB, Smith CR, Mack M, Miller DC, Moses JW, Svensson LG, et al. Transcatheter aortic-valve implantation for aortic stenosis in patients who cannot undergo surgery. N Engl J Med. 2010 Oct 21;363(17):1597–607.
Crossref PubMed
Correspondence to: Jacqueline McMillan, MD, MSc, Calgary Zone and Cumming School of Medicine, University of Calgary, Room 1104, South Tower, Foothills Medical Center, 1403-29,th Street NW, Calgary, AB T2N 2T9, Email: Jacqueline.McMillan@albertahealthservices.ca
COPYRIGHT
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial No-Derivative license (https://creativecommons.org/licenses/by-nc-nd/4.0/), which permits unrestricted non-commercial use and distribution, provided the original work is properly cited.
Canadian Geriatrics Journal, Vol. 28, No. 2, JUNE 2025