Objectives To determine the long-term trajectories of health system use by persons with dementia as they remain in the community over time.
Design Population-based cohort study using health administrative data.
Setting Ontario, Canada from 1 April 2007 to 31 March 2014.
Participants 62 622 community-dwelling adults aged 65+ years with prevalent dementia on 1 April 2007 matched 1:1 to persons without dementia based on age, sex and comorbidity.
Main outcome measures Rates of health service use, long-term care placement and mortality over time.
Results After 7 years, 49.0% of persons with dementia had spent time in long-term care (6.8% without) and 64.5% had died (30.0% without). Persons with dementia were more likely than those without to use home care (rate ratio (RR) 3.02, 95% CI 2.93 to 3.11) and experience hospitalisations with a discharge delay (RR 2.36, 95% CI 2.30 to 2.42). As they remained in the community, persons with dementia used home care at a growing rate (10.7%, 95% CI 10.0 to 11.3 increase per year vs 6.7%, 95% CI 4.3 to 9.0 per year among those without), but rates of acute care hospitalisation remained constant (0.6%, 95% CI −0.6 to 1.9 increase per year).
Conclusions While persons with dementia used more health services than those without dementia over time, the rate of change in use differed by service type. These results, particularly enumerating the increased intensity of home care service use, add value to capacity planning initiatives where limited budgets require balancing services.
- health services administration & management
- health policy
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
Strengths and limitations of this study
This population-based cohort study examined health service use over time among a large sample of community-dwelling persons with prevalent dementia compared with a matched control group of persons without dementia.
The use of multiple linked health administrative databases allowed for the comprehensive examination of healthcare use in a variety of settings, including in hospital, home and nursing home settings.
The long follow-up period of 7 years allowed for an investigation of changes in the intensity of health system use over time, which has not been described in previous studies.
Despite the use of a validated health administrative data algorithm to capture dementia, we did not have information on clinical diagnosis and there is a potential for misclassification.
Data regarding the causal reasons for health system contacts including individual, household and health system factors leading to the use of healthcare services were not available and should be explored in future research.
An estimated 564 000 Canadians were living with dementia in 2016, and this number is expected to rise to 937 000 by 2031.1 Across health systems internationally, evidence is mounting about the high health systems use and costs, and potential for poor care outcomes among older adults with dementia.2–5 Efforts to support persons with dementia in their homes, outside of acute and long-term care settings, are at the forefront of policy agendas internationally6 with a goal of promoting quality of life and limiting potentially avoidable care transitions.
These efforts coincide with concerns that current health system supports for community-dwelling older adults with dementia are inadequate and place a substantial burden on informal caregivers, including family.7–9 However, there is limited evidence describing the intensity with which older adults with dementia use health services as they remain in the community over time; particularly in relation to other older populations. While several studies2 10 11 have shown increases in utilisation at specific points in time, such as around the time of long-term care admission12 and in the year of diagnosis,13 longer time horizons and trajectories of health service use have not been well-documented.
Comprehensive information on population-level patterns of health service use for persons with dementia is required to understand current (and projected) health system impacts and to explore gaps in care. We determined trajectories of health service use, transitions to long-term care and mortality among a population-based cohort of community-dwelling older adults with prevalent dementia compared with matched controls over 7 years in Ontario, Canada.
Study design and population
We conducted a cohort study using population-based health administrative data from 1 April 2007 to 31 March 2014 in Ontario, Canada. We included all prevalent community-dwelling adults aged 65–105 years with dementia (including Alzheimer’s disease) who were alive and eligible to receive provincial health insurance coverage on 1 April 2007 (index date). These individuals were matched 1:1 to a control group with no documented history of dementia. All individuals were followed for up to 7 years to examine trends in health system use, transitions to long-term care and mortality.
Ontario provides universal healthcare coverage for its citizens. These encounters are captured in health administrative databases that were linked using unique encoded identifiers and analysed at ICES (online supplementary table 1). ICES is an independent, non-profit research institute whose legal status under Ontario’s health information privacy law allows it to collect and analyse healthcare and demographic data, without consent, for health system evaluation and improvement. This study is reported according to REporting of studies Conducted using Observational Routinely-collected Data guidelines (online supplementary table 2). The use of data in this project was authorised under section 45 of Ontario’s Personal Health Information Protection Act, which does not require review by a Research Ethics Board.
Patient and public involvement
No patients were involved at the outset of this study due to limited time and resources. We have invited patients and stakeholders to help us develop our knowledge dissemination strategy.
Persons with dementia
We identified 117 337 persons with dementia on 1 April 2007 using a previously validated algorithm based on diagnosis codes recorded in hospitalisation records and physician claims, as well as prescription drug reimbursements for cholinesterase inhibitors.14 The algorithm has been used to produce national dementia prevalence and incidence estimates by the Canadian Chronic Disease Surveillance System.15 This prevalent cohort included individuals at various stages of dementia. Individuals who resided in long-term care in the 60 days prior to and including the index date were excluded (N=54 715). This resulted in a cohort of 62 622 older adults with dementia.
Persons without dementia
We selected 1:1 matched controls (age (±1 year), sex, number of comorbidities (±2) as measured by Aggregated Diagnosis Groups (ADGs)) from the remaining population of community-dwelling older adults. ADGs were derived using the Johns Hopkins Adjusted Clinical Group ACG Case-Mix System16 V.10.0 from diagnoses in hospitalisation records and physician claims in the 2 years prior to index date.
We described age, sex, number of ADGs (0, 1–5, 6–10 and 11+), neighbourhood income quintile and rurality of residence. Neighbourhood income quintile was captured by linking to Statistics Canada census data using postal codes. Rurality was determined by the Rurality Index of Ontario, which is an index based on population factors and distance to referral centres; ranges are from 0 to 100 and communities with higher values (≥40) considered rural.17
Patterns of transition and health system use from the community
We examined long-term care placement and mortality and rates of health service use over 7 years. Mortality analyses included deaths occurring both in the community and long-term care. In long-term care placement and mortality analyses, individuals were censored at first evidence of: (a) developing dementia; (b) loss of health insurance eligibility; (c) no health system contact for 5 years or (d) study end. To be eligible for long-term care in Ontario, individuals must require access to 24 hours nursing and personal care, frequent assistance with activities of daily living and on-site supervision to ensure safety and well-being. This setting does not include retirement homes or assisted living facilities.
Health system use included home care visits, emergency department (ED) visits, acute care hospitalisations, acute care hospitalisations with any discharge delay days and visits to family physicians as well as to dementia specialists (defined as geriatricians, neurologists and psychiatrists). Publicly funded home care services may include homemaking, transportation, personal care, nursing care, end-of-life care, physiotherapy, occupational and speech-language therapy. Home care is provided on either a short-stay (ie, services provided for <60 days to aid in recovery post-surgery or injury) or long-stay (ie, clients requiring services in the home for ≥60 days in a single episode) basis; although the majority of service in our cohort is for long-stay. Discharge delays indicate a patient is medically ready for discharge, but was not able to be discharged for other reasons such as lack of suitable care in the community or lack of available long-term care beds.18 We did not examine medication use as it was not possible to track in acute care settings. For health service utilisation rates, given our focus on health system use by individuals as they remain in the community over time, individuals were censored at the first of: (a) entering long-term care; (b) death; (c) developing dementia; (d) loss of health insurance eligibility; (e) no health system contact for 5 years or (f) study end.
Descriptive statistics were used to compare the demographic characteristics of older adults with (vs without) dementia at baseline and at the start of the last year of follow-up. Sankey plots were used to visualise changes in setting of care over time. Rate ratios (RRs) were used to compare rates of health service use between persons with dementia and persons without dementia. Trends in the intensity with which individuals used health services over time as they remained alive and in the community were assessed using generalised estimating equations for serially correlated data. Differences in trends between those with and without dementia were assessed using Wald tests for interaction. All analyses were conducted using SAS Enterprise Guide V.6.1 (SAS Institute).
Characteristics of older persons with dementia
The 62 622 matched pairs had a mean age of 81.1 years (SD: 6.8 years) and 60.4% were women (table 1). The median number of ADGs was 9 and 79.4% of individuals had 6 or more ADGs, indicating a high level of comorbidity. The neighbourhood income distribution and proportion living in rural settings were similar across the matched pairs. By 1 April 2013, persons with dementia who had remained in the community were modestly younger (mean 83.5 years vs 84.3 years) and more likely to reside in urban areas (90.7% vs 88.0%) compared with persons without dementia.
Persons with dementia were followed until mortality for an average of 4.5 years (SD: 2.4 years) compared with an average of 5.0 years (SD: 2.4 years) in those without dementia. For long-term care placement, persons with dementia were followed for an average of 3.3 years (SD: 2.4 years) compared with an average of 4.8 years (SD: 2.44 years) in those without dementia.
Patterns of transition for long-term care placement and mortality
Figure 1A,B shows yearly transitions between community, long-term care and death for older adults with dementia and without dementia. Persons with dementia were more likely to be admitted to long-term care with an average of 12.7% transitioning each year (1.6% without dementia). Table 2 shows that at the end of 7 years, almost half (49.0%) of persons with dementia had been admitted to long-term care (6.8% without dementia) and 64.5% had died (30.0% without dementia).
Rates of health system use
Over 7 years, persons with dementia who continued to live in the community had higher rates of health system use than persons without dementia (table 3). Compared with persons without dementia, older adults with dementia had a fourfold higher rate of dementia specialist visits (RR=4.11, 95% CI 3.95 to 4.28) and a threefold higher rate of home care visits (RR=3.02, 95% CI 2.94 to 3.11) over the study period. The rate of hospitalisations with a discharge delay was 2.36 times higher for persons with dementia (95% CI 2.30 to 2.42) while the rate of acute care hospitalisations was 29% greater (RR=1.29, 95% CI 1.27 to 1.31). Persons with dementia had 23% more ED visits (RR=1.23, 95% CI 1.21 to 1.25) and 25% more family physician visits per person-year (RR=1.25, 95% CI 1.24 to 1.27) than persons without dementia.
Trends in health system use
Rates of home care visits increased as persons with dementia remained in the community over time; an average rate of 10.7% per year (95% CI 10.0 to 11.3) (figure 2A). This rate of increase was different from that for persons without dementia (6.7% per year (95% CI 4.3 to 9.0), p<0.001). Most home care visits were for personal support and homemaking; in 2013, 91.1% of visits for persons with dementia and 84.9% of visits for persons without dementia (data not shown). The next most common home care service was nursing visits (6.6% of visits for persons with dementia and 11.5% of visits for persons without dementia). While the ED visit rate among persons with dementia increased by an average of 1.1% per year (95% CI 0.2 to 2.0), the rate of increase was no different in persons without dementia (2.1% per year (95% CI 1.3 to 2.8, p=0.27)) (figure 2B). Neither group experienced a significant change in the rate of acute care hospitalisation over time (figure 2C). Notably, the rate of hospitalisations with discharge delay in the dementia cohort decreased by 3.2% per year (95% CI 1.8 to 4.6) while the rate for persons without dementia remained stable (figure 2D). Older adults with and without dementia experienced a decline in rates of physician visits (figure 2E,F) as they remained in the community over time. In particular, the rate of dementia specialist visits declined by 5.4% per year (95% CI 4.4 to 6.5) in persons with dementia; a significantly greater decline than the 2.5% per year experienced by the persons without dementia (p<0.001).
This study offers a novel look at population-level trajectories of health system use for individuals with prevalent dementia in Ontario, as they remained alive in community settings. We documented the substantial role of long-term care for persons with dementia, with rates of long-term care admission seven times higher than for persons without dementia. As persons with dementia remained in the community over time, they used home care services and visited the ED at increasing rates; while other rates of service use were stable or declined.
Our focus on community-based care is important to the quality of life and health status of persons with dementia, as potentially avoidable care transitions can lead to poor outcomes.19–21 Even in a health system such as Ontario’s, where provincially funded health services include home care, we found high rates of long-term care placement—with more than half of persons with dementia being placed in long-term care after 7 years of follow-up. This result is likely a function of our prevalent cohort design, where individuals are represented at various stages of dementia. It also potentially highlights the challenges associated with managing the behavioural and psychosocial symptoms of dementia as well as the care burden of worsening dementia over time. The high rate of long-term care placement and increasing intensity of home care visits over time in the dementia group indirectly suggest unmet care needs and a burden on informal caregivers. Limits on the number of publicly funded home care hours may also influence decisions to admit individuals to long-term care, particularly as persons with dementia experience disease progression and require more intensive care. Increased levels of caregiver distress have been associated with earlier long-term care placement22 and may be more pronounced at end-of-life.23 Caregiver strain can result in depression and anxiety, poorer health status and health behaviours, social isolation and financial difficulties.24 While long-term care placement can help to reduce some demands on caregivers, it may not result in an improvement in symptoms of anxiety and depression caregivers may experience.8 In order to support persons with dementia and their caregivers to live successfully in the community, additional research is needed.
We found that after 7 years, persons with dementia were almost two times as likely to die compared with persons without dementia.25 Our findings are consistent with previous work that has demonstrated persons with dementia experience poorer survival2 26 and shorter time to long-term care transition than those without dementia,11 however the magnitude of the differences we show is not as large as in some previous work. Differences observed between dementia populations and those without dementia are likely attributable to differences in dementia severity as well as other important risk factors such as frailty.27–29 While other studies have demonstrated similar rates of mortality in dementia and non-dementia populations,11 30 the differences in mortality in our study suggest unmeasured differences in baseline health characteristics beyond age, sex and comorbidities.
Persons with dementia had approximately 20%–30% higher rates of acute care hospitalisations and ED visits compared with persons without dementia. Studies have shown twofold differences between hospital use in dementia compared with non-dementia cohorts,5 11 31 32 including those who have addressed the differences in population characteristics using propensity score matching.3 Phelan and colleagues examined a cohort older adults and found a crude twofold higher rate of hospitalisation for persons with dementia; however, after adjustment for confounders, persons with dementia had a 41% increased rate of hospitalisation, a similar magnitude as in our study.31
Our cohort approach is novel and of value to capacity planning initiatives. There was notable increasing use of home care services among individuals who continued to reside in the community. This finding suggests that additional supports may be required to enable persons to remain in the community. Previous studies examining health service utilisation among persons with dementia have been inconsistent regarding the timing of healthcare utilisation increases—with some finding that utilisation increases with time following diagnosis2 and others observing increases near the time of death.2 12 13 33 International researchers will need to consider our findings in the context of their own healthcare systems. In Ontario, all residents are eligible for universal access to medically necessary physician services, hospital care, home care and prescription drugs under the publicly funded health insurance system. Most of the cost of long-term care is covered by the provincial health system and access to long-term care is centrally managed. Although we did not examine healthcare utilisation using years since diagnosis, we observed an increased rates of long-term care placement and home care visits over time and decreases in other forms of healthcare utilisation.
Strengths and limitations
Our study has several strengths. It is the largest Canadian, population-based study to examine persons with dementia matched to persons without dementia followed over a substantial period of time. The cohort design allowed an examination of trends in rates of service use among individuals who remained in the community and will help health system planners understand the long-term needs of these individuals.
Several limitations should be noted. First, our administrative data definition of dementia, although validated in a primary care population sample,14 is not a clinical diagnosis of dementia and there is the potential for misclassification. Also, we did not have access to measures of stage of dementia in the health administrative data but were able to capture time since dementia ascertainment as a proxy measure of dementia stage. Second, administrative databases allow for documentation of health system use patterns at a population-level but provide minimal information on the causal path. This poses challenges to understanding the interrelationship between individual, household and health system factors leading to the use of healthcare services among persons with dementia. Third, we were unable to obtain information on perceived quality of care for persons with dementia and their caregivers in order to better understand the individual and caregiver experience and choices made to enter long-term care. Fourth, there may have been residual confounding by unmeasured covariates such as frailty, activities of daily living, marital status, caregiver and family supports and health behaviours. Fifth, the methods chosen for the present study did not allow for the incorporation of health service utilisation following transition to dementia among persons initially without dementia in the control group. Future work will explore health service utilisation at this important period of transition.
We observed higher rates of health service utilisation, long-term care placement and mortality among community-dwelling persons with prevalent dementia compared with persons without dementia. Future work exploring the underlying reasons for patterns of health service utilisation over time may provide opportunities to improve care in the community setting to reduce the need for long-term care placement, as well as to support health system planning to meet the needs of this complex and growing population.
The authors would like to thank Matthew Kumar for his support in preparing the manuscript. A preliminary version of these findings was presented at the 2019 Alzheimer’s Association International Conference and published as a podium abstract.
Contributors SEB, JDW and CJM conceived and designed the study. JG and XW carried out the statistical analysis. SEB, LCM and JDW wrote the manuscript. JG, XW, RN, PAR, EAY, MJV and CJM contributed to the manuscript’s critical appraisal and review. All authors had full access to all the data (including statistical reports and tables) in the study and can take responsibility for the integrity of the data and the accuracy of the data analysis. All authors read and approved the final manuscript.
Funding This study was funded, in part, by an ICES Applied Health Research Question (AHRQ #2016 0900 556 001) and the Canadian Institutes of Health Research (CIHR) (operating grant MOP #136854). This research was also conducted with support from the Ontario Neurodegenerative Disease Research Initiative (ONDRI) through the Ontario Brain Institute (OBI), an independent non-profit corporation, funded partially by the Ontario government. The opinions, results and conclusions are those of the authors and no endorsement by OBI is intended or should be inferred. This study was also funded by ICES, which is funded by an annual grant from the Ontario Ministry of Health (MOH). Parts of this material are based on data and information compiled and provided by the Canadian Institute for Health Information (CIHI). However, the analyses, conclusions, opinions and statements expressed herein are those of the author, and not necessarily those of CIHI. We thank IMS Brogan for use of their Drug Information Database. The analyses, conclusions, opinions and statements expressed herein are solely those of the authors and do not reflect those of the funding or data sources; no endorsement is intended or should be inferred. JDW holds a Canada Research Chair in Indigenous Health. PAR holds the RTO/ERO Chair in Geriatric Medicine at the University of Toronto. CJM holds a University Research Chair at the University of Waterloo.
Competing interests None declared.
Patient consent for publication Not required.
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement The dataset from this study is held securely in coded form at ICES. While data sharing agreements prohibit ICES from making the dataset publicly available, access may be granted to those who meet prespecified criteria for confidential access, available at www.ices.on.ca/DAS. The full dataset creation plan and underlying analytic code are available from the authors upon request, understanding that the computer programs may rely upon coding templates or macros that are unique to ICES and are therefore either inaccessible or may require modification.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.