Global and regional prevalence of multimorbidity in the adult population in community settings: a systematic review and meta-analysis

1We searched PubMed, ScienceDirect, and Google Scholar for peer-reviewed papers and research reports on the prevalence of multimorbidity, using the search words 'prevalence' and 'multimorbidity' and similar terms published between January 1, 2000 and December 31, 2021. One meta-analysis combined 68 studies from 1992 to 2017 and showed that the global pooled prevalence of multimorbidity in community settings was 33.1%. In 2021, another meta-study focused on articles that investigated people in community settings from Latin America and the Caribbean.This research used studies until 2021 to analyze multimorbidity prevalence in community settings worldwide. South America has the highest prevalence of multimorbidity when comparing prevalence estimates across geographic regions. The prevalence difference was obtained across age groups, gender, country and income level, and study periods. For the first time in a subgroup study, we stratified the number of conditions to estimate the prevalence of multimorbidity. Studies that included mental health in the definition of multimorbidity resulted in a high pooled prevalence. Our research also uses statistical techniques to estimate the pooled prevalence of multimorbidity in adults while capturing heterogeneity in the estimates. This study summarizes the available evidence and encourages policymakers to use more standardized methods to reduce the burden of multimorbidity, which is a critical step toward meeting the sustainable development goal (SDG) goal of reducing premature mortality from non-communicable diseases by one-third through prevention and treatment by 2030.Our findings show that the landscape of multimorbidity prevalence has increased in the last two decades though it has remained relatively unchanged since 2010, implying a slow reduction in the burden of multimorbidity. About half of the South American adult population had multimorbidity, and thus these countries should take it as a priority agenda to develop more sustainable and integrated models of care. Research like this is crucial as the world tries to balance lowering the expense of multimorbidity on society and improving healthcare outcomes.

2We searched PubMed, ScienceDirect, and Google Scholar for peer-reviewed papers and research reports on the prevalence of multimorbidity, using the search words 'prevalence' and 'multimorbidity' and similar terms published between January 1, 2000 and December 31, 2021. One meta-analysis combined 68 studies from 1992 to 2017 and showed that the global pooled prevalence of multimorbidity in community settings was 33.1%. In 2021, another meta-study focused on articles that investigated people in community settings from Latin America and the Caribbean.

3This research used studies until 2021 to analyze multimorbidity prevalence in community settings worldwide. South America has the highest prevalence of multimorbidity when comparing prevalence estimates across geographic regions. The prevalence difference was obtained across age groups, gender, country and income level, and study periods. For the first time in a subgroup study, we stratified the number of conditions to estimate the prevalence of multimorbidity. Studies that included mental health in the definition of multimorbidity resulted in a high pooled prevalence. Our research also uses statistical techniques to estimate the pooled prevalence of multimorbidity in adults while capturing heterogeneity in the estimates. This study summarizes the available evidence and encourages policymakers to use more standardized methods to reduce the burden of multimorbidity, which is a critical step toward meeting the sustainable development goal (SDG) goal of reducing premature mortality from non-communicable diseases by one-third through prevention and treatment by 2030.

4Our findings show that the landscape of multimorbidity prevalence has increased in the last two decades though it has remained relatively unchanged since 2010, implying a slow reduction in the burden of multimorbidity. About half of the South American adult population had multimorbidity, and thus these countries should take it as a priority agenda to develop more sustainable and integrated models of care. Research like this is crucial as the world tries to balance lowering the expense of multimorbidity on society and improving healthcare outcomes.

5Multimorbidity has emerged as a significant public health issue in the world. It is typically defined as the presence of two or more chronic conditions at the same time in one individual. Multimorbidity has increased in various population groups due to population aging, lifestyle changes, improved socioeconomic conditions, and improved diagnostic capabilities by health services.,, Due to a lack of data from low-income countries and the use of different definitions of multimorbidity, a recent systematic review highlighted the need to estimate the prevalence of multimorbidity and patterns of multimorbidity.

6The high prevalence of multimorbidity has several negative consequences, including a high mortality rate, increased healthcare utilization, and increased healthcare expenses, influencing overall functioning and quality of life.,,,, According to a recent review and meta-analysis, those with at least two morbidities have a 1.73 times higher risk of death than people without multimorbidity. Moreover, healthcare demands and costs of multimorbidity continue to rise as populations age.

7Although few systematic reviews and meta-analyses on multimorbidity in community settings have been published in recent years, these included fewer studies or are restricted to a specific geographic region.,,, According to a systematic review and meta-analysis of studies with data collected between 1992 and 2017, the global pooled prevalence of multimorbidity in community settings was 33.1% (95% confidence interval: 30.0–36.3%). This prior study, however, did not look at how multimorbidity patterns changed over time or gave insight into multimorbidity definitions based on the number of conditions.

8In recent years, many studies have been conducted to identify the clinical patterns of chronic conditions.,,,, Two systematic reviews on multimorbidity identified depression, hypertension, and diabetes as the most prevalent co-occurring chronic diseases., Another study of multimorbidity identified cardiovascular and metabolic diseases as the most common diseases, followed by mental health disorders and musculoskeletal conditions. In a multi-national cross-sectional study of non-institutionalized adults aged 50 and over in Finland, Poland, Spain, China, Ghana, India, Mexico, Russia, and South Africa, hypertension, cataract, and arthritis were the most prevalent comorbid conditions. A study conducted in Germany among health-insured individuals aged 65 and older identified three broad multimorbidity patterns–cardiovascular/metabolic disorders, anxiety/depression disorders, and pain/neuropsychiatric disorders. It indicates that mental health disorders were prevalent in the studies, so we examined the prevalence of multimorbidity with and without mental health disorders.

9These findings provide an explanation for the clinical patterns as well as the burden of multimorbidity that was observed among the studied people. An accurate and up-to-date prevalence estimation is critical to assess the impact of multimorbidity on public health and project effective and integrative interventions to reduce premature death due to multimorbidity. It is challenging to conduct a meta-analysis to estimate a global prevalence as the different studies used a different number of diseases and disease combinations. There is no gold standard for quantifying multimorbidity; definitions of multimorbidity and statistical approaches for evaluating prevalence differ greatly.,,,, But the trade-off of generating pooled estimate of multimorbidity exceed the drawbacks of the variability in the data. However, the prevalence of multimorbidity was not thoroughly assessed based on geographic regions, country's economic level, age, study periods, and the number of diseases considered for defining multimorbidity.

10Given the growing concern about the rising burden of chronic diseases, understanding the prevalence of multimorbidity in the adult population is critical for developing preventive strategies. As a result, we conducted a systematic review and meta-analysis to examine the global and regional prevalence of multimorbidity and changes in multimorbidity prevalence over time among the adult population in community settings.

11We searched PubMed, Google Scholar, Embase and ScienceDirect online databases to select peer-reviewed papers for our systematic review and meta-analysis. We screened observational studies (cross-sectional and baseline in a cohort) to determine the global prevalence of multimorbidity in the adult population in community settings. Our search included articles published in any language between January 2000 and December 2021, which would help minimize data heterogeneity and provide a more precise estimate of global multimorbidity prevalence. The screening was conducted primarily in English, but we also utilized the Google translation tool for article selection. A description of search terms is given in Appendix A. The search results were compiled using Mendeley citation management software. In addition to the database search, we explored references of selected studies and previously published systematic reviews on similar topics to incorporate all potential pertinent articles to construct our summary estimates. The Preferred Reporting Items for Systematic Review and Meta-analysis (PRISMA) checklist was followed in this study. The protocol was registered in the PROSPERO database (CRD42020150945).

12Our systematic review included studies that (1) defined multimorbidity as having more than one underlying chronic conditions; (2) documented multimorbidity as the outcome of interest; (3) provided the number of participants in the study, with at least 200; (4) defined multimorbidity in the article, with at least five chronic conditions; (5) were observational studies, either cross-sectional or cohort, including adults 18 years and above; (6) published in years 2000–2021; and (7) were conducted in a community setting. Furthermore, only the recent study was considered if more than one study studied the same population. Only prevalence at baseline was included when the design was a cohort. Studies were excluded if they (1) focused only on comorbidity, (2) defined multimorbidity as more than two diseases (3) studied only inpatients or outpatients in hospital and primary care settings, (4) studied institutional population, i.e., people in nursing home, old home etc., (5) included acute conditions in the list of conditions, (6) used less than 5 conditions to define multimorbidity, or (7) were qualitative, interventional studies, opinion articles, conference presentations, books, letters, editorials, reviews, dissertations/theses, or abstracts.

13Using Covidence, two independent reviewers (S.R.C. and D.C.D.) screened the articles. The reviewers examined successively the titles, abstracts, and full texts of all possibly relevant articles identified by our searches. The differences in article selection and data extraction were handled by consensus and, if necessary, discussion with another reviewer (A.H.). Two independent reviewers (S.R.C. and T.C.S.) created a data-extraction form to establish the type of information to be extracted. The reviewers (S.R.C. and T.C.S.) recorded pertinent data on the name of the first author, study settings (e.g., country, year of publication, study period (start-end year), region), and study conduct (e.g., study design, population age and male percentage, number of study participants, data sources, method of ascertainment of morbidity, and minimum number of conditions included in multimorbidity), prevalence of multimorbidity, and number of participants with multimorbidity from the published article only. We further stratified the articles based on the country's income level (World Bank classification by income, GNI per capita). Moreover, the study participants were cross tabulated by age group and gender, and multimorbidity was documented whenever possible. If the prevalence of multimorbidity was not directly given, it was manually computed from the data supplied in the articles. In studies providing longitudinal prevalence estimates over a period, we utilized baseline prevalence. After settling any differences, the two reviewers (S.R.C. and T.C.S.) independently extracted the data, discussed the inputs, and revised the extracted data. Unresolved issues were resolved by involving a third reviewer (J.B.).

14The Newcastle-Ottawa Scale (NOS), the tool for assessing the quality of non-randomized research, was used to determine the risk of bias for individual studies. The eight items of NOS are categorized into three domains of potential bias, namely “selection (representativeness of the sample, sample size, non-respondents, ascertainment of the exposure),” “comparability (the subjects in different outcome groups are comparable, based on the study design or analysis; and confounding factors are controlled),” and “outcome (assessment of the outcome and statistical test)”.,, A few points on the NOS were modified to be relevant to our research question (Supplementary File 1). The articles' methodological stringency, lucidity, and clarity are reflected in the subjective scores. However, we did not eliminate any articles based on their quality scoring. A study can be given one star for each item within the selection and outcome categories. For comparability, a maximum of two stars can be awarded. Thus, a cross-sectional study can be awarded a maximum of 10 stars (10 points), and a cohort study can be awarded a maximum of 9 stars (9 points). Overall, the studies were categorized as “low risk of bias (8–10 stars)”, “moderate risk of bias (6–7 stars)”, and “high risk of bias (0–5 stars)”. Two independent reviewers (S.R.C. and D.C.D.) assessed the quality of the included studies, and the discrepancies were resolved with discussion with the third reviewer (A.H.). The PRISMA statement consists of a 27-item checklist given in Supplementary File 2.

15The statistical analysis was performed using meta and metafor packages in the R statistical software (version 4.1.1). Multimorbidity prevalence was estimated as the ratio of the number of people with multimorbidity (numerator) and sample size (denominator). The numerator was derived from the percentage of people with multimorbidity when the numerator was not available. We obtained the pooled prevalence (with 95% CIs) of multimorbidity among the overall population from all studies and subgroups. The pooled prevalence was estimated using a random-effects model that allows the actual effect size to vary from study to study. The calculated proportion from each study and the combined effect estimate with 95% CI were represented graphically using forest plots. We assessed potential publication bias by visually observing the symmetry of funnel plots and using Egger's test. The I2 statistic was used to quantify heterogeneity across the selected studies. The I2 statistic indicates the proportion of overall variation across studies due to heterogeneity rather than chance. Subgroup analysis was carried out to determine the pooled prevalence for each group and look for potential explanations for the heterogeneity. Geographical region (Africa, Asia, Europe, North America, Oceania, and South America); WB/WHO income region (High, Upper-middle, Low- and Lower-middle); Study design (Cross-sectional, Cohort); Multimorbidity (5–9 conditions, 10–19 conditions, ≥20 conditions); Mental health included in the multimorbidity definition (Yes or No); Age groups of study participants (≥30 years, ≥40 years, ≥50 years, ≥60 years) and Gender (male and female) were considered for sub-group analysis. We conducted a trend analysis to see the global multimorbidity prevalence over time (2000–2021). We also conducted sensitivity analyses to assess the findings' robustness in consideration of sample size, multimorbidity prevalence, multimorbidity definitions based on the number of conditions studied, and NOS overall quality of the studies. Two-sided P <.05 was considered statistically significant.

16There was no funding available for this study. All of the study's data was accessible to all of the authors, and the corresponding author had responsibility for publication.

17A flowchart of the literature search to select the relevant articles is summarized in the PRISMA format and is presented in Fig. 1. The initial search retrieved 8003 studies from the three pre-specified databases. After excluding the duplicates, the titles and abstracts were screened for a further selection of probable articles. Subsequently, the investigators selected 376 articles based on eligibility criteria for full-text review. By manual searching through the included papers’ reference lists and reference lists of previous systematic reviews on similar topics, 12 studies were considered for scrutiny, resulting in the total number of potential articles being 388. After excluding 262 studies in full-text review, finally, 126 studies with a total of 15,400,421 (approximately 15.4 million) people were included in the systematic review and meta-analysis. Sample sizes in the studies range from 264 to 3,759,836.,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,

18Table 1 shows the characteristics of the included studies. The 126 population-based studies were conducted across 54 countries. Six of the 126 research included were carried out in multiple countries. The majority of the studies (n = 47) were conducted in Asia, followed by Europe (n = 27), South America (n = 19), Africa (n = 10), North America (n = 14), Oceania (n = 6), and various continents (n = 3). Between 2000 and 2021, 53 studies were carried out in high-income countries (HICs), 48 in upper middle-income countries (UMICs), and 24 in low- and lower-middle-income countries (Low- and LMICs). Most of the studies (121 studies) were cross-sectional in design, and the remaining five had a cohort design, from which we used data from the baseline assessment. When defining multimorbidity, 37 studies looked at 5–9 diseases, 64 studies at 10–19 diseases, and 24 studies at more than 20 diseases.

19The prevalence of multimorbidity among the adult population ranged from 4.0% to 92.8% in the studies. Prevalence estimates along with confidence intervals for multimorbidity are shown in Fig. 2 by using a forest plot. The random-effects overall pooled estimated (126 studies) prevalence of multimorbidity was 37.2% (95% CI = 34.9%–39.4%, I2 = 99.7%). The pooled proportion of multimorbidity was the highest in South America with 45.7% (95% CI = 39.0%–52.5%, I2 = 99.0%). On the other hand, the pooled prevalence of multimorbidity was the lowest in Africa with 28.2% (95% CI = 15.6%–40.8%, I2 = 99.0%). However, studies from Asia, Europe, North America, and Oceania were calculated to have the pooled prevalence of multimorbidity 35% (95% CI = 31.4%–38.5%, I2 = 99.3%), 39.2% (95% CI = 33.2%–45.2%), 43.1% (95% CI = 32.3%–53.8%), and 32.5% (95% CI = 26.8%–38.2%, I2 = 98.9%), respectively.

20The subgroup analysis of the prevalence of multimorbidity by continents, study design, number of diseases included in multimorbidity, age, and gender is shown in Table 2. The forest plots are given in the Supplementary File 3. Of note, 85 studies reported the prevalence of multimorbidity in males and females. According to the table, the pooled prevalence of multimorbidity was higher among female participants (39.4%, 95% CI = 36.4–42.4%, I2 = 99.6%) than male participants (32.8%, 95% CI = 30.0–35.6%, I2 = 99.6%). The Fig. 3 shows the gender segregation of pooled prevalence of multimorbidity by geographic regions. Female participants from South America (prevalence 50.1% and 95% CI = 39.7–60.4%) appeared to have the most multimorbid conditions in the world. Multimorbid illnesses were notably more prevalent in European and North American women than in male participants.

21Based on the continents of the studies, the estimated pooled prevalence of multimorbidity was found 38.6% (95% CI = 35.3%–41.9%, I2 = 99.2%) in high-income countries, 38.7% (95% CI = 35.5–41.9%, I2 = 99.2%) in upper middle-income countries (UMICs), and 32.1% (95% CI = 24.3–40.0%, I2 = 99.5%) in Low- and LMICs. In the case of the number of diseases included in the multimorbidity, the prevalence was found 44.7% (95% CI = 39.5%–50.0%, I2 = 99.3%) among the studies that considered ≥20 diseases. The prevalence of multimorbidity was 25.0% (95% CI = 22.3–27.8%, I2 = 99.0%) for studies with 5–9 diseases, and 41.3% (95% CI = 37.6%–45.0%, I2 = 99.0%) for studies with 10–19 diseases. When mental health is included in the multimorbidity definition, the prevalence (38.4%, 95% CI = 35.9–41.0%, I2 = 99.0%) was higher than without inclusion of mental health (33.2%, 95% CI = 27.1–39.2%, I2 = 99.1%).

22Among the different age groups of the study participants, the highest prevalence was found in the studies that included the respondents more than 60 years with 51.0% (95% CI = 44.1%–58.0%). The pooled prevalence was 44.4% (95% CI = 39.3%–49.4%, I2 = 99.1%) among the participants with 30 years and above. When the study participants were ≥40 years and ≥50 years, the pooled proportion of multimorbidity was 45.7% (95% CI = 40.2%–51.2%, I2 = 99.0%) and 47.2% (95% CI = 42.0%–52.5%, I2 = 99.1%), respectively.

23There was a difference in the prevalence of multimorbidity by study design among the studies. The pooled prevalence of multimorbidity was 37.4% (95% CI = 35.1%–39.6%, I2 = 99.0%) for cross-sectional studies, and 32.4% (95% CI = 27.9%–36.9%, I2 = 96.7%) for cohort studies.

24The global prevalence of multimorbidity by 5-year interval is displayed in Fig. 4, considering studies that contains 10 or more diseases. The five-year span was categorized based on the year in which investigations were done. If a study was completed between 2013 and 2016, we assumed it was conducted between 2011 and 2015 because the majority of years fell within the interval. The study was removed from the analysis if it did not belong to any of the groups. We excluded papers that reported a multimorbidity prevalence of less than 10% or greater than 80% in order to minimize variability in trend analysis. The prevalence of multimorbidity has been on the rise globally since 2000, but it has remained rather stable since 2011. The trend analysis with the studies that considered ten or more illnesses in multimorbidity classifications, showed that the global prevalence of multimorbidity remained high, exceeding 40%.

25We conducted sensitivity analyses including studies with more than 1000 participants, removing studies from Africa, and removing studies that showed prevalence of less than 20% and more than 80%. The reasons for removing studies with less than 1000 participants are to increase estimate reliability and precision of the estimate with the studies with a larger sample size. Furthermore, we excluded papers with extreme prevalence estimates of less than 20% and more than 80% because these values could lead to heterogeneity in predicting worldwide prevalence. Forest plots are reported in Supplementary File 4. When considering studies of more than 1000 participants, the global prevalence among participants tends to be 36.1% (95% CI = 33.7–38.4%, I2 = 98.8%), which is in line with the findings of the meta-analysis with 126 studies. After excluding African studies, the prevalence was 37.9% (95% CI = 35.4%–40.2%), which is comparable to the meta-analysis with 126 studies. We also found the global prevalence was higher than the overall pooled prevalence after removing studies with extreme prevalence. The results showed the prevalence 42.3% (95% CI = 39.8–44.7%, I2 = 98.8%) after excluding studies with extreme prevalence. The findings excluding studies with extreme prevalence are, therefore, higher than the meta-analysis of 126 studies. With high-quality papers (minimal bias according to NOS), we found the prevalence to be 36.6% (95% CI = 33.6–39.5%, I2 = 99.8), which imply a similar result that we analyzed in the meta-analysis of 126 studies. Moreover, the studies using self-reported multimorbid data indicate a prevalence of 38.3% (95% CI = 35.1–41.5%), but the studies with data from medical records indicate a prevalence of 34.3% [95% CI = 30.3–38.2%].

26The Egger test found that there was no statistically significant publication bias (P >.05) among the 83 population-based studies evaluating the relationship between gender and multimorbidity status. However, the Egger test revealed a statistically significant publication bias among the 126 population-based studies for proportion (Supplementary File 5). We also have applied trim-and-fill method to adjust for this publication bias in the analysis. We see that the procedure identified and trimmed 42 added studies. The overall effect estimated by the trim-and-fill is 26.71% (95% CI = 0.2350–0.2799). Our initial estimate with 126 studies was 37.1%, which is substantially larger than the bias-corrected effect. If we assume that publication bias affected our findings, the trim-and-fill method allows us to hypothesize that our initial results were overstated because of publication bias, and the global estimate when controlling for selective publication might be 26.71%. Moreover, considering the odds ratio in a funnel plot we found a high existence of publication bias in our study. Consequently, publication bias may be a cause of heterogeneity in investigating overall proportion.

27This study analyzed data from 126 studies that involved nearly 15.4 million people from 54 countries, providing an up-to-date global multimorbidity prevalence of 37.2% (95% CI = 34.9–39.4%). A previous meta-analysis with studies until 2017 found that 33.1% had multimorbidity in the adult population aged 18 and older living in the community. In comparison to that meta-analysis including studies in community settings, we found a higher prevalence of multimorbidity. Another meta-analysis that included studies from both community and healthcare settings estimated the overall prevalence of multimorbidity was 42.4% (95% CI = 38.9–46.0%) among adults. The inclusion of studies from primary care and health care settings in the meta-analysis resulted in a higher pooled prevalence than ours.

28The sub-group analysis by region showed significant differences in the pooled prevalence of multimorbidity. Our analysis showed that the prevalence of multimorbidity was highest in South America. The result is consistent with a meta-analysis that found that the pooled proportion of multimorbidity in Latin America and the Caribbean was as high as 43% (95% CI: 35–51%). Africa had the lowest prevalence of multimorbidity, according to our analysis. The result could be attributable to the low age group of participants in the African studies compared to other geographic regions. The lowest rate of multimorbidity in Africa should be interpreted with caution because it raises the possibility that many people living with chronic illnesses in Africa are going undiagnosed.

29In subgroup analysis, the prevalence of multimorbidity was lower in Low- and LMICs than in UMICs and HICs. The prevalence of multimorbidity was highest in UMICs. This difference is consistent with another study's findings, where a meta-analysis in community settings found that the pooled multimorbidity prevalence was higher in HICs than LMICs., The majority of the survey included in the meta-analysis were from HICs and UMICs, with a few studies conducted in Low-income countries. It may reflect the differences in diagnostic and data management systems among HICs, UMICs, and Low- and LMICs. According to a study, the disparity in prevalence estimates between HICs and LMICs could be due to the fewer publications on multimorbidity prevalence in LMICs because of limited understanding and importance of multimorbidity in LMICs compared to HICs. People in low-income countries may be less likely to seek treatment for diseases than those in high-income countries. Therefore, the prevalence in low-income countries may be underestimated if diseases are defined using medical records.

30The pooled prevalence of multimorbidity was higher for the cross-sectional study design than for the cohort study type in this meta-analysis. This disparity in multimorbidity prevalence could be due to study designs with varying levels of methodological differences, such as various study populations, sampling procedures, sample coverage, sample sizes, data collection, and so on. Besides, we considered the baseline sample for a cohort study design that might contribute to the lower prevalence.

31For included studies, the more the number of diseases evaluated for multimorbidity, the higher the prevalence. When examining 20 or more conditions for multimorbidity, the prevalence was 44.7%, but it was lowered to 41.3% for 10–19 diseases and 25.0% for 5–9 diseases to define multimorbidity. According to a study, the different combinations of illnesses may cause the prevalence of multimorbidity to differ significantly., A range of different combinations of multimorbidity definitions has been proposed in the literature, ranging from a list of 16 chronic diseases to 291 diseases.,,,, Furthermore, the pooled estimate of multimorbidity prevalence with the studies those included mental health in the definition of multimorbidity was greater. Previous studies identified a correlation between multimorbidity and mental health.,, Our findings, the higher prevalence of multimorbidity with the studies that included mental health, reveal consistency with the findings of previous research.

32Our study showed that prevalence estimates varied substantially according to age and gender. Our research showed that females had a higher pooled prevalence of multimorbidity than males. It indicates an association between gender and multimorbidity (evidence of which was provided in multiple studies).,, According to our findings, multimorbidity increases with age. While the prevalence estimates varied between and within age groups, our meta-analysis indicated that a large proportion of individuals over 60 had multimorbidity. It is well established that the prevalence of multimorbidity increases in very old persons.,

33The calculation of the global prevalence of multimorbidity based on the study's publication interval of 5-year is one of the most important findings of our research. According to our findings, the prevalence of multimorbidity has changed considerably over the previous two decades but has remained relatively consistent since 2011. This suggests a gradual decline in the global burden of multimorbidity. The plateau observed in multimorbidity prevalence since 2011 may be attributable to a handful of the 19 studies that showed low prevalence in 2016–2021. Therefore, this conclusion should be studied further. Over the years, the global prevalence of multimorbidity among adults has exceeded 40 percent, indicating a high burden of multimorbidity exists over years.

34One of the study's strengths was its strong study selection and screening protocols. Because of our rigorous search approach and inclusion criteria, we were able to conduct the largest systematic review of multimorbidity prevalence in community settings to date. The majority of the papers included in the review were of high quality. The comprehensive subgroup analyses demonstrate that our findings are applicable to a wide range of contexts. One important finding of our study is the estimation of the global prevalence of multimorbidity by year of publication. This review did, however, have several limitations. To report multimorbidity prevalence, the majority of the studies in our sample used self-reported data. As a result, such research findings were prone to response bias. High heterogeneity between studies in our meta-analysis implies that the prevalence of multimorbidity varies between studies. To overcome this constraint, we used a random-effects model and performed subgroup analyses. Furthermore, considerable heterogeneity may indicate that the prevalence of multimorbidity varies significantly by geographical region, country income classification, gender, age group, number of diseases considered for multimorbidity, or study methodology.

35The high prevalence of multimorbidity highlights the need for healthcare reforms and improvements in several continents. Policymakers should commit to increasing multimorbidity awareness, particularly in relation to mental health management, supporting innovation, maximizing the use of existing resources, and coordinating the efforts of multiple countries to reduce the burden and fatal effects of multimorbidity. More than half of the global adult population over the age of 60 has multimorbid illnesses, and female adults are more prone to develop multimorbidity than male adults. Therefore, management should incorporate these findings into healthcare policies, and countries, particularly in South America, should aim to increase their preventative efforts and build more integrated care models to reduce the burden.

36A.H., S.R.C., D.C.D., and T.C.S. contributed to the study concept, literature search, and design. A.H., S.R.C., D.C.D., T.C.S. and J.B. contributed to the data acquisition. A.H. and S.R.C. accessed the data and contributed to the data analysis. A.H., S.R.C., and J.B. contributed to the data interpretation. A.H., S.R.C. and D.C.D. drafted the manuscript. All authors contributed to the critical revision of the manuscript.

37Because this meta-analysis was based on data extracted from previously published research, most of the data and study materials are available in the public domain. For further discussions, we invite interested parties to contact the corresponding author.

38All other authors declare no competing interests.