The prevalence of disability in older adults with multimorbidity: a meta-analysis

Background Disability is typically correlated with lower quality of life and decreased capacity for self-care. It has been demonstrated that multimorbidity is closely linked to a variety of unfavorable events, such as disability. Researchers are still figuring out how and to what extent co-morbidities impact disability, though. In order to fill up this gap, this study examines the prevalence and contributing variables of disability in older patients who have multimorbidity. Methods We conducted a systematic search of Pubmed, Cochrane Library, Web of Science, Embase, and CINAL databases for articles from their inception until September 2023. We selected co-morbid older adults aged > 60 years and used the ADL scale or any scale that assesses disability as an assessment tool. We excluded literature that did not meet the criteria, and literature that could not be included in the data we needed. We extracted data from the included literature and calculated synthetic prevalence rates, ORs, and 95% confidence intervals. Results A total of 32 papers (71,135 older adults) were included in the study. The prevalence of disability among older patients with multimorbidity was around 34.9% (95% CI = 25.8-43.9%). Subgroup analysis showed higher rates of disability among comorbidities who were older, female, unmarried, and long-term users of health services. And the incidence of disability increased each year. Meanwhile, the regions of the United States, China, and Spain showed higher rates of disability. Conclusions Disability rates in older patients with multimorbidity are higher, thus it’s critical to focus on risk factors while fully accounting for regional variances. Supplementary Information The online version contains supplementary material available at 10.1007/s40520-024-02835-2.


Introduction
Multimorbidity, also known as multiple chronic diseases(MCC), refers to the simultaneous presence of two or more chronic diseases in an individual [1,2].Due to improvements in global healthcare, a rise in the average lifespan, and demographic shifts, multimorbidity is becoming more common.The prevalence of multimorbidity is approximately 30% in the young and middle-aged Multimorbidity and disability are strongly related, as has been shown in earlier research [7].Physical functionality may be a good indicator of the cumulative combined impact of chronic illnesses on health.For each additional disease, the patient's risk of dysfunction increases by 16%, life expectancy decreases by 1.8 years, and the risk of needing the assistance of a healthcare professional increases by 20% [8,9].These relationships are typically nonlinear, multimorbidity accelerating the weakening of bodily functions and leading to a rapid deterioration of an individual's metabolic mechanisms, and given the physiological vulnerabilities associated with aging, older adults are more likely to experience functional disability.The healthcare system is put under an additional burden because of the relationship between disabilities and adverse outcomes including falls and depression, a considerable increase in mortality, frailty, and Emergency Department visits, and a higher likelihood that patients may face social and economic squalor in later life [10][11][12][13].In order to effectively address the disability of patients, minimize eventual healthcare expenses, and enhance health, determining the disability rate of patients with multimorbidity as well as the risk factors has thus become especially crucial.
Currently, only one systematic review [14] from 2015 explored the correlation between multimorbidity and functional decline.However, the study had a large age span (> 18 years) and did not focus on older adults who were most affected by comorbidity; On the other hand, the study did not calculate the combined rate of disability, and explored and analyzed the influencing factors.Moreover, the study has a long history, and it needs to be discussed in newer literature.Therefore, this study explores and explains this part of the content, and provides a basis for future intervention to delay and control the progress of disability.

Materials and methods
This study has been registered with the International Prospective Register of Systematic Reviews (Prospero) under the registration number CRD42023425740.This article is reported according to Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA).This study was based on the Cochrane Handbook of Systematic Literature Reviews.

Search strategy
Two researchers searched the literature from the PubMed, Cochrane Library, EMBASE, CINAL, and Web of Science databases with an English-only search language from the library until September 2023.We used Medical Subject Headings (MeSH) keywords to search by title and abstract, using AND, OR to concatenate the following text: older adults (older population, aged, elderly, geriatric) AND multimorbidity (multiple chronic conditions) AND disability (activities of daily living, disabled, ADL, BADL, IADL ).Additional articles were then incorporated by way of handsearched references.

Eligibility criteria
Inclusion criteria: (1) age 60 and above; (2) multimorbidity: refers to the simultaneous presence of two or more chronic diseases in an individual; (3) studies reporting clinical outcomes including disability.Activities of Daily Living Scale (ADL) are the most common indicator of disability and includes the Basic Activities of Daily Living Scale (BADL) and Instrumental Activities of Daily Living Scale (IADL).The most common scales for BADL include the Katz Independence Index and the Barthel Index, with 6 or 10 items.Including dressing, eating, bathing, dressing, etc. while the IADL has 8 items, including shopping, cooking, etc.In addition, any literature that could define an indicator of disability was also included.(4) study design: Observational studies, including cross-sectional and cohort studies; (5) using disability as a dichotomous outcome indicator, data from the literature provided corrected ORs and 95% CI or sufficient data to calculate prevalence.
The following papers are excluded: duplicates, unrelated to the topic, documents lacking the whole text, and documents from which the necessary data cannot be retrieved.

Risk of bias assessment
Two investigators evaluated the quality of the included literature.The Agency for Healthcare Research and Quality (AHRQ) scale was utilized in the cross-sectional study [15], which has 11 entries with a yes or no answer, yes scoring 1 and no scoring 0, respectively.The final score was 0-3 for low quality literature, 4-7 for moderate quality literature, and 8-11 for high quality literature.The Newcastle Ottawa Scale (NOS) were utilized in the cohort study [16], which has 8 entries, with 0-3 being low quality literature, 4-6 being moderate quality literature, and ≥ 7 being high quality literature, with higher scores indicating better quality literature.

Data extraction
Two researchers were responsible for including, screening and extracting literature according to the criteria.When there was disagreement, we consulted a third researcher.The extracted data was as follows: (a) basic characteristics: authors, year, country, age, gender, number of events, total population, and incidence rate; and (b) study design, research instrument, sample source, sample size, and marital status.

Quality of evidence
In this study, the quality of evidence was evaluated using GRADE (Grading of Recommendations, Assessments, Development, and Evaluation), which consists of 5 evaluations: (a) risk of bias; (b) inconsistency; (c) indirectness; (d) imprecision; and (e) publication bias.The presence of one item was downgraded to a medium quality level; the presence of two items was downgraded to a low quality level; and the presence of three items was a very low quality level.

Statistical analysis
The two researchers independently extracted the data into EXCEL and then analyzed the prevalence rate and 95% confidence interval by STATA17.0 software.The magnitude of heterogeneity was assessed using I 2 and Q tests [17,18].When P < 0.05 or I 2 > 50%, it was considered that there was significant heterogeneity in the study, which was analyzed by the random-effects model (SMD); When P ≥ 0.5 or I 2 ≤ 50%, it was considered that there was no significant heterogeneity, which was analyzed by the fixed-effects model (MD)The articles differed in nadir criteria, study design, and disability criteria, so there was significant heterogeneity between studies (I 2 = 99.94%).Therefore, a random effects model was used to calculate the combined effect size.When studies had high heterogeneity, the source of heterogeneity was analyzed subgroups.The publication year (-2015, 2015-2020, 2020-), sample size (<500, > 500), study design (cross-sectional study, cohort study), data source (community, hospitals), gender (male > 50%, female > 50%), marital status (married, unmarried), and country were all taken into consideration.A sensitivity analysis was performed to assess the stability of the findings by excluding individual papers.Funnel plots were generated by Begg-Mazumdar's and Egger's tests to determine publication bias.Statistical significance was defined as P < 0.05, and 95% confidence intervals were calculated.

Study selection
In this study, 2823 documents were retrieved from Pubmed, Cochrane Library, Web of Science, EMBASE, and CINAL, and 3 documents were searched manually, totaling 2826 documents included, and 1112 documents remained after deduplication.1010 literatures were eliminated by reading titles and abstracts.Reading the remaining 102 literature, 73 of them were excluded because they did not meet the inclusion criteria.See Fig. 1.

Subgroups analysis
This study has a high heterogeneity, and subgroup analysis is used to examine the origins of heterogeneity.Divided by year of publication: pre−2015, the disability rate among comorbid patients was 27.9% [95% CI (13

Bias analysis
The results of the funnel plot showed that the graphical distribution was symmetrical, with a low likelihood of risk of bias.See Fig. 3. p = 0.03 < 0.05 in the Begg-Mazumdar's and Egger's tests, suggesting the possibility of publication bias.We evaluated the effect of publication bias on the results by trim and filling analysis.The results of trim and filling analysis (PR = 1.093, 95%CI = 0.988 to 1.208) did not affect the results of this study.

Quality of evidence
We evaluated the quality of evidence for the results of the meta-analysis, which showed that the incidence of disability was of low quality.See Table 3.

Findings of summary
This is the first meta-analysis to evaluate the rate of disability in elderly patients with multimorbidity.In all, 32 studies with moderate or high quality literature and dependable results were included in this analysis.The meta-analysis's findings demonstrated that older patients with multimorbidity had a disability rate of 34.9% (95% CI = 25.8−43.9%)).The results were higher than those of middle-aged and elderly patients in Canada (27.4%) [51] and higher than the senior disability rate globally, which is 10.2% across 194 nations and territories [52].multimorbidity may be a significant predictor of functional decline in the elderly [15,53].

Sensitivity analysis
To ascertain if the study findings were consistent and dependable, the included literature was eliminated one at a time.The omission of a single piece of literature did not significantly alter the disability rate, according to the results of the sensitivity analysis.with multimorbidity [56].To lessen and postpone the likelihood of disability, our response strategies to the aging trend and co-morbidity trend should be suggested earlier.
(2) Demographic factors: advanced age, females, and unmarried are more likely to have disabilities.Older people are less able to successfully protect themselves against a range of hostile influences, according to prior studies [55], their physiological functions weaken, their resistance declines, and their compensating systems worsen.The functional independence of the elderly is gravely challenged by the synergistic and cumulative We performed subgroup analyses to determine sources of heterogeneity.The results of the subgroup analyses showed that: (1) Year of Publication: The disability rate rose with the year of publication, we suspect it has to do with the aging trend.The elderly population currently makes up 9.3% of the world's population [54], reach 41.1% in some countries by 2050, and age is an important indication of disability [55].On the other hand, the baseline population with disabilities rises due to better medical care, longer life expectancy, and an expanding population for people who are alone or unmarried.Based on the marital resource hypothesis, marriage has a health-protective effect that delays health deterioration, and the feeling of duty and responsibility it imparts promotes positive behaviors in old age and has an influence on health status [60].We believe that having healthy interpersonal interactions and participating in social activities as appropriate may help to lessen the disability of celibacy people [61,62].(3) Sample source: Compared to those in the community, hospital patients are more likely to be disabled.This is simple to understand.People with major health problems and a high risk of becoming disabled are those who frequently get primary healthcare or live in long-term effects of many chronic disorders [57].Women had a higher likelihood of being disabled, which was in line with Yau's findings [58].In addition to biological disparities, Zhu [59] claims that due to cultural and familial obligations, women commonly face difficulties while seeking to receive healthcare treatments.Therefore, while having a greater life expectancy than men, women may have worse health outcomes.To compensate for the differences caused by aging and gender, which are objective factors that cannot be changed, the researchers suggest that social and medical resources can be appropriately tilted, and appropriate medical assistance and financial subsidies can be provided.According to Pang's study [60], the probability of disability is higher  1 3 [70][71][72].Unfortunately, a similar analysis could not be performed in this study.
In recent years, the study of disability-free survival has come to be recognized as a crucial issue in the field of gerontological research [73].With the prevalence of multimorbidity, the previous single chronic disease management model is no longer adapted to the status quo of multimorbidity.Considering the synergistic and cumulative effects of multimorbidity, countries like the United Kingdom have developed multimorbidity clinical guidelines that emphasize patientcentered healthcare services to enhance health-related quality of life [75].A multidisciplinary integrated team treatment approach, which may offer the best diagnostic and treatment plan for patients with multimorbidity and lessen the side effects of many drugs [75,76], has received attention from academics.Additionally, nutritional schedules, exercise recommendations, and other strategies for encouraging healthy living are thought to be effective approaches to prevent or postpone the onset of chronic illnesses and disabilities [74,76].Since chronic diseases require long-term care, it is envisioned that community-based staff would be incorporated as a part of the primary healthcare system to establish long-term, stable, and continuous services with the co-patient population.We urge patients to monitor their own health.Additionally, In order to effectively accompany and psychologically assist co-morbid patients, boost their adherence to rehabilitation, and decrease despair and loneliness, we expect that caregivers will be acknowledged as a crucial component of social support.

Strengths and limitations
The strength of this meta-analysis is that it synthesizes disability rates in older adults with multimorbidity, understands the influencing factors that exist, is the first of its kind to the best of our knowledge, and can inform the development of subsequent clinical guidelines.The study's shortcomings are visible.First, the non-English literature was left out of this study due to language restrictions; second, some studies were not included in the meta-analysis because it was not feasible to extract data from the literature, potentially introducing bias; third, there was a sizable amount of heterogeneity in this study, but it was not brought on by the subgroups.

Conclusion
In conclusion, around 34.9% of older co-morbid patients have varying degrees of disability.Advanced age, female gender, unmarried status, and long-term utilization of healthcare services are all likely significant factors contributing to care facilities.Elderly residents of the community are more physically and mentally well, have more energy for social contact, and are less prone to suffer negative outcomes.Consequently, the management of multiple medications, remote monitoring of their physical condition, diet and exercise advice, and standardized rehabilitation training are helpful assistance for people who have used healthcare services for a long time [63,64,76].For community residents, we support the involvement of formal and informal caregivers in assisting patients to maintain their functional independence [65].(4) Nation: We find that the disability rate in the US, China, and Spain is around 40% higher than in Norway and Sweden.A serious issue with population aging also exists in China.Due to a mismatch in the availability and demand for healthcare services, China, the world's largest developing country, has seen a steady increase in the burden of disabilities [66].Spain also has one of the fastest aging populations in the world [55].Despite being one of the best countries in the world in terms of economic strength and access to healthcare, the United States has a high rate of disability.There is a clear socioeconomic hierarchy, and those from lower socioeconomic classes have a harder time accessing better healthcare and are more likely to have health status inequalities while coping with serious sickness, according to prior research [67].The high level of economic development, high level of education, comprehensive social security system, and a substantial proportion of public health spending may all be contributing factors to the low disability rate in places like Norway and Sweden [68].
There was high heterogeneity in this study.The researchers believe that it is related to the inconsistent standards of disability.The most commonly used is the ADL scale, while EQ−5D and Specific disability measures (24 items) have also been studied.Even articles that also use ADL have different criteria for determining disability: The first is any difficulty in basic or using tools can be defined as a disability, and the second recognizes moderate or severe dependence as a disability.In addition, the physical condition of the population, the duration of multiple diseases, and the severity of the disease can all affect the course of disability.People with two or three or more chronic diseases are more likely to become disabled than those with a single chronic disease [69].This may be related to the deterioration of organ and tissue function exacerbated by multiple diseases.Certain patterns of comorbidities may accelerate and promote negative events.For example, psychosomatic comorbidities are more predictive of disability than other somatic comorbidities and have a higher degree of disability in the former disability.Through exercise regimens, nutritional advice, and other pharmacological management techniques, we may manage disability.The number and types of co-morbidities, and other factors could not be meta-analyzed due to the limited literature included, and it is anticipated that more high-quality, large-sample, multi-center studies will be carried out in the future to support the theory and serve as a foundation for the clinical development of an integrated, comprehensive, and personalized disease wellness program.

Fig. 1
Fig. 1 Flow diagram of studies selection

Fig. 2
Fig. 2 Meta-analysis of the prevalence of disability in elderly patients with multimorbidity a. Evaluate the use of different disability tools for the same population

Table 1
Basic characteristics of the included literature

Table 2
Subgroup analysis of disability prevalence among older patients with multimorbidity

Table 3
Evidence profileNo of studies Design Risk of bias inconsistency indrectness imprecision other sample effect low a: Included studies did not describe blinding, allocation concealment, etc.; b: I 2 > 50%, heterogeneity exists