Burden and Associated Factors of Unmet Health Care Needs in Individuals With Osteoarthritis: A Systematic Review

Abstract

Objective

This systematic review aimed to describe the prevalence, incidence, and associated factors of unmet health care needs among adults with osteoarthritis.

Methods

We searched Medline (Ovid), Embase (Ovid), CINAHL (EBSCO), and PsycINFO (Ovid) from inception through May 15, 2024. Eligible studies were cross‐sectional, cohort, and case‐control studies investigating the prevalence, incidence, associated factors, or risk factors of unmet health care needs in adults with osteoarthritis. We restricted to articles published in English, French, Italian, and Chinese for feasibility. Reviewers independently screened articles, assessed risk of bias using the Joanna Briggs Institute Checklists, and extracted data. We descriptively synthesized results from low/moderate risk‐of‐bias studies, stratifying results by age (<60 vs ≥60 years).

Results

Of 3,589 citations screened, 7 cross‐sectional studies with low/moderate risk‐of‐bias were included in the synthesis (3 from South Korea, 4 from the United States). In South Korea, the 12‐month prevalence of unmet health care needs was 31.6% (95% confidence interval [CI] 29.9%–33.3%) among adults aged ≥50 years with osteoarthritis and 31% (95% CI 30.9%–31.1%) among those aged ≥65 years with arthritis. In the United States, the 12‐month prevalence of unmet needs in the general population due to unaffordability ranged from 15% to 30% in adults with osteoarthritis or arthritis. Prevalence was higher among those who exclusively used complementary and alternative medicine and varied during the COVID‐19 pandemic, peaking in the summer of 2020. Evidence suggests that unmet needs are associated with lower income, no insurance, and activity limitations.

Conclusion

Unmet health care needs are common in adults with osteoarthritis, particularly those facing socioeconomic disadvantages or functional limitations. Given the paucity of high‐quality studies, additional research is needed.

INTRODUCTION

Osteoarthritis affects approximately 528 million individuals worldwide and is a major contributor to years lived with disability among musculoskeletal conditions. ¹ Since 1990, the prevalence of osteoarthritis has increased by 113% globally and is projected to grow due to population aging and increasing rates of obesity and injury. ² Osteoarthritis presents substantial challenges with mobility, daily activities, and participation. Individuals with osteoarthritis are five times more likely to experience mobility limitations compared to those without osteoarthritis. ³ About two‐thirds of individuals with osteoarthritis (344 million) experience severity levels that would benefit from rehabilitation. ¹

SIGNIFICANCE & INNOVATIONS.

• This is among the first systematic reviews to synthesize evidence on the prevalence and associated factors of unmet health care needs among adults with osteoarthritis.

• Unmet health care needs are common in adults with osteoarthritis, particularly those facing socioeconomic disadvantage or functional limitations.

• Findings suggest that the prevalence of unmet needs is higher among those who exclusively use complementary and alternative medicine. Unmet needs were also impacted by the COVID‐19 pandemic, particularly during its early stages.

• Our review underscores the need for strategies to improve access to care, and future research focused on the incidence and associated factors of unmet needs in adults with osteoarthritis.

Despite this need, a substantial proportion of affected individuals do not receive appropriate health care, which represents unmet needs, primarily from barriers concerning availability, affordability, and accessibility. ⁴ Unmet health care needs refer to the inability to access appropriate care when required. ⁴ In a systematic review and meta‐analysis by Rahman et al, unmet health care and long‐term care needs among older adults in the general population were examined. ⁴ This review reported a pooled prevalence of unmet health care needs of 10% (95% confidence interval [CI] 9.2%–10.9%) among older adults; however, there are limitations to this meta‐analysis due to heterogeneity of studies (differences in populations and methods used to measure unmet needs). The review reported unaffordability as a common reason for unmet needs, and other barriers included a lack of available health facilities, mistrust or fear of providers, and time constraints. ⁴ However, that review focused on older adults in the general population rather than individuals with osteoarthritis. To our knowledge, no systematic reviews have examined the burden and associated factors of unmet health care needs in this population, representing a critical knowledge gap.

Given the prevalence and associated disability of osteoarthritis, it is important to provide insights into the extent of unmet health care needs in individuals with osteoarthritis. By identifying the burden and discrepancies in needs and access to health services, we can better inform the development of policies and resource allocation strategies. Gaining insights into unmet health care needs will enable us to better understand how to tailor the delivery of more accessible, affordable, and equitable services for individuals living with osteoarthritis.

The objectives of this systematic review were to determine (1) the incidence and risk factors of unmet health care needs among adults with osteoarthritis (including all joints) and (2) the prevalence and associated factors of unmet health care needs among adults with osteoarthritis (including all joints).

METHODS

Protocol and reporting

We developed our systematic review protocol using the Preferred Reporting Items for Systematic Review and Meta‐Analysis Protocols (PRISMA‐P) checklist ⁵ and preregistered our protocol with the Open Science Framework (https://doi.org/10.17605/OSF.IO/AK2NP). We reported our systematic review using the PRISMA statement. ⁶

Eligibility criteria

Population

Our systematic review targeted studies involving adults aged ≥18 years with osteoarthritis (any joint). Although osteoarthritis is most commonly observed in individuals aged ≥60, the condition often begins to manifest in middle age. ¹ Given variations in defining middle age across the literature, we adopted a comprehensive approach by including all adults with osteoarthritis. Where possible, results were stratified by age (<60 and ≥60 years) to facilitate a nuanced understanding of unmet health care needs in older adults as defined by the World Health Organization. ⁷

Osteoarthritis is defined as a degenerative joint condition that affects the entire joint and surrounding tissues. ¹ Diagnostic criteria for osteoarthritis have evolved over time, transitioning from primarily radiographic in the 1960s ⁸ to encompass both clinical and radiographic assessments since the mid‐1980s. ⁹ To capture this transition, we considered studies reporting clinical or radiographic diagnostic criteria for osteoarthritis. Results were analyzed separately, where applicable, to account for variability in diagnostic methods.

Studies of joint pain due to other types of arthritis were included if they incorporated osteoarthritis within their sample; if those with osteoarthritis represented less than the majority of the sample (<50%), those studies were excluded. These results were reported and synthesized separately. We also included studies targeting arthritis in the general population (but that did not report on the type of arthritis), since osteoarthritis is the most prevalent type among adults. ¹ These results were reported and synthesized separately.

Outcomes

Unmet health care needs are defined as a person's inability to access appropriate health care when required. ⁴ We focused on self‐reported unmet health care needs to capture perceived unmet needs. Unmet health care needs include unmet rehabilitation needs, where rehabilitation is defined as a set of interventions designed to optimize functioning in individuals with health conditions when interacting with their environment. ¹⁰ We included unmet needs in primary, secondary, and tertiary health care settings. We included self‐reported unmet health care needs for any reason, as well as for specific reasons (eg, reasons regarding availability, affordability, accessibility, and access) when available. Eligible services included, but were not limited to, health care provided by physicians, orthopedic surgeons, rheumatologists, nurses, physiotherapists, chiropractors, occupational therapists, psychiatrists, psychologists, counselors, and social workers. Based on the World Health Organization package of interventions for rehabilitation of osteoarthritis, examples of interventions included, but were not limited to, health care targeting cognitive function, pain management, motor functions and mobility, exercise and fitness, activities of daily living, education and vocation, lifestyle modifications, and self‐management. ¹¹

Risk factors and associated factors

We included studies that investigated risk factors or factors associated with unmet health care needs. Potential categories included, but were not limited to, sociodemographic, health‐related, lifestyle, and occupational factors.

Study designs and characteristics

Eligible studies had to meet the following inclusion criteria: (1) published in English, French, Italian, and Chinese languages (included these languages due to feasibility); (2) published from database inception to May 15, 2024, in a peer‐reviewed journal; (3) investigated adults with osteoarthritis (all joints); (4) investigated risk factors or factors associated with unmet health care needs, or the prevalence or incidence of unmet health care needs; and (5) used an eligible study design: cross‐sectional studies, cohort studies, case‐control studies, and quantitative components of a mixed‐methods study if that component was a cohort, case‐control, or cross‐sectional design.

The following were excluded: (1) letters, editorials, commentaries, guidelines, books, conference proceedings, meeting abstracts, lectures and addresses, and consensus development statements; (2) case reports, case series, randomized controlled trials, qualitative studies, systematic and nonsystematic reviews, biomechanical studies, laboratory studies, and studies not reporting on methodology; and (3) cadaveric or animal studies.

Data sources and search strategy

The following databases were searched from database inception to May 15, 2024: Medline (Ovid), Embase (Ovid), CINAHL (EBSCO), and PsycINFO (Ovid) (Supplementary Material 1). The search strategy was developed in consultation with an experienced health sciences librarian and reviewed by a second librarian using the Peer Review of Electronic Search Strategies Checklist. ¹² The search strategy was developed in Medline and adapted for use in other bibliographic databases. Search terms included subject headings (eg, medical subject headings in Medline) for each database and free‐text words for the concepts of osteoarthritis, unmet health care needs, and study design. Duplicates were identified electronically in EndNote and removed.

Data collection and analysis

Study selection

We completed the screening process using EPPI‐Reviewer software. Before beginning this process, we conducted a training phase. For training in phase 1 screening, reviewers independently screened a random sample of 50 titles and abstracts for eligibility and categorized the articles as possibly relevant or irrelevant using the predefined eligibility criteria. For training in phase 2 screening, reviewers independently screened a random sample of 25 full‐text articles for eligibility and categorized them as relevant or irrelevant. Paired reviewers discussed disagreements to reach consensus for both screening phases, with involvement of a third reviewer if necessary. We calculated the agreement between reviewers for title and abstract and full‐text screening. If the agreement was <80%, the team addressed any discrepancies and clarified eligibility before beginning screening.

The screening process involved two phases. In phase 1 screening, pairs of independent reviewers (KR, SdS‐O, HY, and SB) assessed titles and abstracts to determine study eligibility, categorizing them as possibly relevant or irrelevant. In phase 2 screening, pairs of independent reviewers (KR, SdS‐O, HY, and SB) evaluated possibly relevant studies in full text, documenting eligibility and reasons for exclusion. Reviewers met to resolve disagreements and establish consensus on study eligibility in both phases, with a third reviewer consulted if needed. Study authors were contacted if additional information was required for clarification.

Methodologic quality and risk‐of‐bias assessment

The JBI (formerly the Joanna Briggs Institute) risk‐of‐bias assessment tools were used to evaluate the methodologic quality and risk of bias of studies, and reviewers used these criteria to inform their judgment on internal validity. Paired independent reviewers (KR, SdS‐O, and HY) appraised studies and discussed disagreements to reach consensus, with involvement of a third reviewer if necessary. The following checklists were used depending on the study design of the relevant article: (1) descriptive cross‐sectional designs focused on prevalence—JBI Checklist for Prevalence Studies ¹³ ; (2) analytical cross‐sectional designs assessing associated factors—JBI Checklists for Analytical Cross‐Sectional Studies and for Prevalence Studies ¹⁴ ; (3) cohort designs—JBI Checklist for Cohort Studies ¹⁴ ; and (4) case‐control designs—JBI Checklist for Case Control Studies. ¹⁴ For the overall risk‐of‐bias assessment, rating of “include” on the JBI checklist represented low‐to‐moderate risk‐of‐bias studies, and “exclude” represented high risk‐of‐bias studies.

Reviewers received training in the use of the JBI risk‐of‐bias tools and independently appraised five sample articles using the appropriate JBI risk‐of‐bias assessment tools and summarized their judgments as low, moderate, or high risk of bias. Percent agreement on overall bias judgment (low/moderate vs high) was calculated, with clarifications provided if agreement was <80%. Following training, two reviewers independently assessed the risk of bias for all included studies. Disagreements were resolved through discussion or, if needed, adjudicated by a third reviewer.

Data extraction

We developed a data extraction table in Microsoft Word. After a training exercise, two reviewers (KR and JJW) independently extracted data from all low/moderate risk‐of‐bias studies. All disagreements were resolved through discussion.

Our systematic review sought to determine the incidence and prevalence of unmet health care needs and the risk factors and associated factors related to unmet needs. Incidence was defined as the proportion of a group of people initially free of the outcome of interest who developed the condition over a given period of time. ¹⁵ Prevalence was defined as the proportion of a group of people possessing a clinical condition or outcome at a set point or period in time. ¹⁵

To summarize findings, we created evidence table templates. For each study, reviewers extracted the following data: authors, year of publication, study location or setting, study design, methodology, participant demographics, definition of osteoarthritis, definitions of unmet health care needs, definition of risk factors and associated factors, effect estimates (eg, odds ratios, relative risk, hazard ratios, and prevalence ratios), and prevalence or incidence rates of unmet health care needs, along with the 95% CI for each respective statistic. Crude and adjusted associations between unmet health care needs and associated factors were also extracted. In cases of missing data, study authors were contacted.

Evidence synthesis

We conducted a narrative synthesis of included studies following the Synthesis Without Meta‐Analysis (SWiM) for systematic reviews checklist. ¹⁶ Results were stratified by incidence, risk factors, prevalence, and associated factors. Where applicable, we further stratified results by age group (<60 vs ≥60 years) and arthritis type (ie, osteoarthritis vs osteoarthritis mixed with other types of arthritis). When synthesizing results, we emphasized studies with low/moderate risk of bias and excluded high risk‐of‐bias studies from the synthesis. When available, we reported the prevalence of unmet health care needs and incidence rates as a proportion (or rate) with 95% CI. The associations between risk factors and unmet needs were reported as odds ratios, hazard rate ratios, and relative risk (or other effect estimates), and associated factors were reported as prevalence ratios or odds ratios when available, depending on how the included studies reported the statistic.

Statistical analysis

Percent agreement was calculated for all the training exercises. We computed the percent agreement and Cohen's kappa statistic with 95% CI to describe the interrater agreement for all phases of screening, data extraction, and risk‐of‐bias assessment.

RESULTS

Study selection

Our literature search yielded 4,091 articles. We removed 502 duplicates and screened 3,589 articles for eligibility (Figure 1). After screening, 3,575 articles did not meet our eligibility criteria, and 14 relevant studies were critically appraised. The main reasons for exclusion in full‐text screening were irrelevant population, outcome measure, study design, or language or failure to assess prevalence, incidence, associated factors, or risk factors. After critical appraisal, seven studies were determined to have a low/moderate risk of bias and were included in the synthesis ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ , ²¹ , ²² , ²³ (Table 1). The remaining seven studies were deemed high risk of bias. ²⁴ , ²⁵ , ²⁶ , ²⁷ , ²⁸ , ²⁹ , ³⁰ All seven low/moderate risk‐of‐bias studies reported prevalence estimates, ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ , ²¹ , ²² , ²³ of which one study also investigated factors associated with unmet health care needs using regression modeling. ²³ We did not identify any relevant studies that assessed incidence or risk factors of unmet needs in this population.

Figure 1.

Preferred reporting items for systematic reviews and meta‐analyses flow diagram.

Table 1.

Risk of bias of relevant cross‐sectional (prevalence) studies based on JBI (formerly the Joanna Briggs Institute) checklist

Author (year)	Representative sample	Sampled in appropriate way	Adequate sample size	Study participants and setting described in detail	Sufficient coverage of entire sample in data analysis	Valid methods for identification of condition	Condition measured in standard and reliable way	Appropriate statistical analysis	Adequate response rate	Overall risk of bias
Ahn et al (2013) 17	Yes	Yes	Yes	Yes	Yes	Unclear	Yes	Yes	Yes	Low or moderate
Bala et al (2020) 24	No	No	Unclear	Yes	Unclear	Unclear	Unclear	Yes	No	High
Boddu et al (2024) 25	Unclear	No	Yes	Yes	Unclear	Unclear	Yes	Yes	Unclear	High
Gwynne‐Jones and Iosua (2016) 26	No	No	Unclear	No	Yes	Unclear	Yes	Yes	Yes	High
Hagglund et al (2005) 27	No	No	Unclear	Yes	Unclear	Unclear	Yes	Yes	Yes	High
Hoerster et al (2011) 18	Yes	Yes	Yes	Yes	Yes	Unclear	Yes	Unclear	Yes	Low or moderate
Jo et al (2020) 19	Yes	Yes	Yes	Yes	Yes	Unclear	Yes	Yes	Yes	Low or moderate
Jung et al (2022) 20	Yes	Yes	Unclear	Yes	Yes	Unclear	Yes	Yes	Unclear	Low or moderate
Lopez et al (2005) 28	No	Unclear	Unclear	Yes	Unclear	Unclear	Yes	No	Unclear	High
MacDonald et al (2021) 29	No	Yes	Unclear	Unclear	Yes	Unclear	Yes	Yes	Yes	High
Schwartz et al (2019) 21	Yes	Yes	Yes	Unclear	Yes	Unclear	Yes	Yes	Unclear	Low or moderate
Tsuzaki and Taira (2022) 22	Yes	Yes	Yes	Yes	Yes	Unclear	Yes	Unclear	Yes	Low or moderate
Weinberger et al (1989) 30	No	Yes	Unclear	Yes	Yes	Unclear	Yes	Unclear	Yes	High
Youn et al (2023) 23	Yes	Yes	Yes	Yes	Yes	Unclear	Yes	Yes	Unclear	Low or moderate

The interrater agreement for screening was 93.8% for phase 1 screening and 74.1% for phase 2 screening (Supplementary Material 2). Disagreements during both phases of independent screening were discussed to reach a consensus. The interrater agreement for the overall methodologic quality of studies based on independent risk‐of‐bias assessment was 100%. The interrater agreement for independent data extraction was 91.8% (45/49 data items). All disagreements from independent data extraction were discussed to reach a consensus.

Risk‐of‐bias assessment

Among the 14 relevant studies, seven were rated as having low/moderate risk of bias, ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ , ²¹ , ²² , ²³ whereas the remaining seven were high risk of bias ²⁴ , ²⁵ , ²⁶ , ²⁷ , ²⁸ , ²⁹ , ³⁰ (see Table 1). As all relevant studies were cross‐sectional studies, we appraised the 14 studies using the JBI Checklist for Prevalence Studies. Of these studies, two had an analytic cross‐sectional component so they were appraised using the JBI Checklist for Analytical Cross‐Sectional Studies. The seven studies classified as low/moderate risk of bias adequately addressed the main methodologic components, including a well‐defined sampling frame, appropriate sampling methods, clear descriptions of study participants and settings, sufficient sample coverage, and appropriate statistical analyses. ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ , ²¹ , ²² , ²³ The primary methodologic weaknesses of the seven high risk‐of‐bias studies included a nonrepresentative sample in six of the seven studies, ²⁴ , ²⁶ , ²⁷ , ²⁸ , ²⁹ , ³⁰ limitations in participant recruitment methods in four studies (eg, nonprobabilistic methods), ²⁴ , ²⁵ , ²⁶ , ²⁷ and low or unclear response rates in three studies. ²⁴ , ²⁵ , ²⁸

Two relevant studies also examined associated factors using multivariable regression modeling and were appraised using the JBI Checklist for Analytical Cross‐Sectional Studies (see Table 2). The study by Youn et al ²³ was deemed low/moderate risk of bias, as it adequately addressed the following: had clear inclusion criteria, participants and setting described in detail, a representative sample, objective standard criteria for measurement of osteoarthritis, covariates identified and dealt with appropriately, and appropriate statistical analysis. The study by Boddu et al ²⁵ was classified as high risk of bias due to a nonrepresentative sample with risk of selection bias, and limitations with methods used for statistical analysis and reporting of results from regression modeling (including potential Table 2 fallacy). ³¹

Table 2.

Risk of bias for relevant analytical cross‐sectional studies based on JBI (formerly the Joanna Briggs Institute) checklist^*

Author (year)	Criteria for inclusion in the sample clearly defined	Study participants and setting described in detail	Exposure measured in valid and reliable way	Objective, standard criteria used for measurement of condition	Confounding factors identified	Strategies to deal with confounding factors stated	Outcomes measured in valid and reliable way	Appropriate statistical analysis used	Overall risk of bias
Boddu et al (2024) 25	Yes	Yes	Yes	Yes	N/A	N/A	Unclear	No	High
Youn et al (2023) 23	Yes	Yes	Yes	Yes	Yes	Yes	Unclear	Yes	Low or moderate

^*N/A, no data given.

Study characteristics

The seven low/moderate risk‐of‐bias studies were all cross‐sectional and examined the prevalence of unmet health care needs in individuals with osteoarthritis (two studies) ¹⁹ , ²³ and with arthritis in the general population (five studies). ¹⁷ , ¹⁸ , ²⁰ , ²¹ , ²² Three of these studies also reported prevalence stratified by certain characteristics. ¹⁸ , ¹⁹ , ²¹ Among these studies, one study also investigated factors associated with unmet needs in individuals with osteoarthritis using multivariable regression modeling. ²³ Four studies were conducted in the United States, ¹⁸ , ²¹ , ²² , ²³ and three studies were conducted in South Korea. ¹⁷ , ¹⁹ , ²⁰ Four studies focusing on arthritis in the general population did not specify the subtypes of arthritis included; we contacted the authors for clarification but received no response. ¹⁷ , ²⁰ , ²¹ , ²²

Osteoarthritis

Two studies focused on individuals with osteoarthritis. ¹⁹ , ²³ The first study included participants aged ≥50 years (73.3% female) with knee, hip, or lumbar spine osteoarthritis. ¹⁹ The second study focused on adults aged ≥18 years with hip osteoarthritis, with 51% of the sample aged ≥60 years (64.2% female). ²³ Among these studies, one focused on unmet health care needs, which were assessed based on participants’ reported difficulties in accessing care, ¹⁹ whereas the other study examined self‐reported unmet needs due to cost‐related barriers. ²³

Arthritis in the general population

Five studies focused on individuals with arthritis in the general population. Three studies included individuals ≥18 years old with arthritis, ¹⁸ , ²⁰ , ²² whereas two studies focused on adults aged ≥60 years. ¹⁷ , ²¹ The proportion of females in these studies ranged from 50.3% ²⁰ to 62.8%. ¹⁸

Three of these studies defined unmet health care needs as the self‐reported inability or delay in accessing care when needed. ¹⁷ , ¹⁸ , ²¹ The other two studies ²⁰ , ²² assessed the impact of the COVID‐19 pandemic by asking participants whether they were unable to access medical care for reasons unrelated to COVID‐19 due to pandemic‐related disruptions.

Factors related to unmet health care needs

Three studies ¹⁸ , ¹⁹ , ²¹ explored factors related to unmet health care needs by stratifying the prevalence of unmet needs by activity limitation status ¹⁹ (self‐reported limitations in daily or social activities due to health or disability), types of health care use (eg, conventional vs complementary and alternative medicine [CAM]), ¹⁸ and types of health insurance. ²¹ One study investigated socioeconomic factors (ie, race, income, and insurance status) associated with unmet needs through regression modeling, ²³ accounting for other covariates (age, sex, sociodemographic factors, and comorbidities).

Summary of evidence

Prevalence of unmet health care needs: Osteoarthritis

Among adults aged ≥50 years with knee, hip, or lumbar spine osteoarthritis in South Korea, the 12‐month prevalence of unmet medical needs for any reason was 31.6% (95% CI 29.9%–33.3%) (Table 3), whereas the 12‐month prevalence due to specific reasons was as follows: accessibility 14.8% (95% CI 13.5%–16.1%), financial barriers 12.2% (95% CI 11.1%–13.5%), acceptability 6.7% (95% CI 5.8%–7.7%), availability 6.2% (95% CI 5.3%–7.1%), and transportation difficulties 2.6% (95% CI 2.0%–3.2%).

Table 3.

Prevalence and associated factors of unmet health care needs in adults with osteoarthritis^*

Author(s), year, study design	Setting, participants, number of participants, n	Definition of osteoarthritis or arthritis	Unmet health care needs definition	Other factors assessed	Prevalence	Associated factors
Jo et al (2020) 19 Design: cross‐sectional study Study setting: Seoul, South Korea	Used data from the fifth and sixth KHANES from January 2010–December 2013. The study was limited to participants 50 years and older (mean age: 67.1 ± 8.6). In total 2,782 participants (73.3% women) were diagnosed with osteoarthritis in this sample and included in the study.	Osteoarthritis was defined as a KL grade ≥2 in the knee, hip and lumbar spine joints with pain reported to have lasted for 3 months or more based on criteria from Korea Centers for Disease Control and Prevention. For radiographic diagnosis, the knee, hip, and lumbar spine joints were evaluated by radiography. Lumbar osteoarthritis was diagnosed in participants aged 50 years or older who experienced lumbar pain within the past 3 months for ≥30 days and showed KL grade ≥2 in radiography.	Unmet medical needs were assessed based on self‐reported unmet health care needs. This was determined from the following questions: “Have you been unable to go to the hospital/clinic (excluding dental clinic) when you wanted to within the past year?”—those who responded “yes” were said to have unmet health care needs. For those who responded yes to the above question, a following question was asked: “If so, what was the reason you could not go to the hospital/clinic when you wanted to among the following?” Hospital or clinic not open, did not want to wait for long time at hospital or clinic; it was difficult to make a reservation at the hospital or clinic; symptoms were mild; transportation was inconvenient; due to financial reasons.	Limitations in activities—participants who responded, “yes” to the question: “Are you experiencing limitations in daily and social activities due to health problems or physical/psychological disability?” were classified as having limitations in activities.	All individuals with OA who reported unmet needs: 31.56% (95% CI 29.86%–33.31%) a Specific reasons for unmet needs Availability: 6.15% (95% CI 5.31%–7.10%) a Accessibility: 14.77% (95% CI 13.50%–16.14%) a Financial accessibility: 12.22% (95% CI 11.06%–13.49%) a Transportation accessibility: 2.55% (95% CI 2.03%–3.21%) a Acceptability: 6.69% (95% CI 5.82%–7.68%) a Prevalence stratified by activity limitation status (see Supplementary Material 4 for extracted relevant data)
Youn et al (2023) 23 Design: cross‐sectional study Study setting: San Francisco, CA, United States	Noninstitutionalized adult civilian population based on data from 2016 NHIS. They limited their study to participants who reported hip osteoarthritis (weighted n = 9,456,820). 35.8% were male, 6.38% were between 18 and 29 years, 7.9% were between 30 and 39 years, 11.6% were between 40 and 40 years 23% were between 50 and 59 years, 26.6% were between 60 and 69 years and 24.4% were 70 or older.	Among adults who had been informed by a doctor or health professional that they had osteoarthritis or a related condition, those with symptomatic hip osteoarthritis were identified by their responses to two consecutive questions: the first question, “During the past 30 days, have you had any symptoms of pain, aching or stiffness in or around the joint?,” was used to identify whether an individual was experiencing joint pain currently. An affirmative response to this question was followed by the second question, “Which joints are affected?” which included the identification of pain in either hip.	Impediments to health care use were evaluated using the following variables present in the 2016 NHIS: whether patients did not have access to medical care or had to delay needed medical care owing to cost within the last 12 months; and whether patients needed but did not receive follow‐up care, specialist care, or prescription medications owing to cost within the last 12 months. Response option of yes was considered to have impediment to health care.	The role of socioeconomic factors in access to health care were assessed in the following ways: Race: race or ethnicity was categorized into non‐Hispanic White, Hispanic, non‐Hispanic Black or African American, non‐Hispanic Asian and non‐Hispanic other by combining two variables: “Hispanic ethnicity, dichotomous” and “self‐reported race” Income: total combined family income was obtained from the NHIS 2016 survey and stratified as follows: less than $35,000, $35,000–74,999, $75,000–99,999, and $100,000 or greater Insurance: insurance status was categorized as private, public, uninsured, and other. Public insurance refers to Medicaid or Medicare, and individuals with “other insurance” comprise those who were unaware of their insurance type and those who had other types of insurance coverage.	Those with hip OA reporting: Inaccessible medical care owing to cost: 12.53% a (95% CI 12.50–12.55) a Delayed care owing to cost: 16.53% a (95% CI 16.51–16.56) a Unmet follow‐up care due to cost: 5.90% a (95% CI 5.89–5.92) a Unmet specialist care due to cost: 8.67% a (95% CI 8.65–8.69) a Unable to afford prescription medications due to cost 13.14% a (95% CI 13.12–13.16) a	See Supplementary Material 4 for relevant data related to race, income, and insurance.

^*CI, confidence interval; KHANES, Korean National Health and Nutrition Examination Survey; KL, Kellgren‐Lawrence; NHIS, National Health Interview Survey; OA, osteoarthritis.

^a Calculated by authors of this systematic review based on data reported in the study.

Prevalence of unmet needs stratified by activity limitations

Adults with knee, hip, or lumbar spine osteoarthritis and activity limitations in South Korea had 1.55 times the 12‐month prevalence of unmet medical needs (41.6%; 95% CI 38.4%–44.9%) compared to those without limitations (26.9%; 95% CI 25.0%–29.0%) (Table 3). ¹⁹ The most commonly cited reason for unmet medical needs in both groups was accessibility, with 23.5% (95% CI 20.8%–26.4%) of those with activity limitations and 10.8% (95% CI 9.5%–12.2%) of those without limitations identifying it as a barrier. ¹⁹

In the United States, individuals with hip osteoarthritis reported unmet medical needs (ie, delayed care or could not access care) due to unaffordability. ²³ The 12‐month prevalence of delayed medical care due to unaffordability was 16.5% (95% CI 16.5%–16.6%), and the 12‐month prevalence of inaccessible medical care due to unaffordability was 12.5% (95% CI 12.5%–12.5%). When examining specific types of care, 13.1% (95% CI 13.1%–13.2%) needed but could not afford prescription medications, 8.7% (95% CI 8.7%–8.7%) could not afford specialist care, and 5.9% (95% CI 5.9%–5.9%) could not afford follow‐up care.

Arthritis in the general population

In adults ≥18 years with arthritis in the United States, the 12‐month prevalence of unmet medical needs due to unaffordability was 14.6% (95% CI 13.5%–15.8%), and the 12‐month prevalence of unmet medical needs due to an access‐related barrier was 15.5% (95% CI 14.4%–16.7%) (Table 4). ¹⁸ Among adults aged ≥65 years with arthritis in South Korea, the 12‐month prevalence of unmet medical needs was 31% (95% CI 30.9%–31.1%). ¹⁷ Among older adults with arthritis in the United States, the 12‐month prevalence of unmet medical needs was 4.9% (95% CI 4.9%–4.9%) in those aged 60 to 64 years and 2.3% (95% CI 2.3%–2.4%) in those aged 66 to 70 years. ²¹

Table 4.

Prevalence and associated factors of unmet health care needs in adults with arthritis^*

Author(s), year, study design	Setting, participants, number of participants, n	Definition of osteoarthritis or arthritis	Unmet health care needs definition	Other factors assessed	Prevalence	Associated factors, 95% CI
Ahn et al (2013) 17 Design: cross‐sectional study Study setting: South Korea	Used data from the KHANES IV conducted in 2007–2009. The study included 3,943 participants older than 64 in their analysis (64.5% were 65–74 years and 35.5% ≥75 years and 40.5% identified as male). In total 41% of participants reported having arthritis (n = 1,617).	Arthritis was defined as self‐reported arthritis as a chronic disease diagnosed by a physician.	Unmet health care needs was measured using the question “during the past 12 months, have you ever not visited clinics or hospitals when you needed to obtain health care services?” Those who answered yes (1) were considered to have unmet health care needs.	N/A	All individuals who reported arthritis with unmet needs: 31.03% (95% CI 30.96%–31.09%) a	N/A
Hoerster et al (2012) 18 Design: cross‐sectional study Study setting: the United States	Adult, civilian, noninstitutionalized US population based on data from the 2007 NHIS as well as the 2007 supplemental questionnaire which assessed CAM use. Those with provider‐diagnosed arthritis were included (n = 3,850); 62.8% were female, 14.8% were between 18 and 44 years, 18.8% 45 and 54 years, 26.3% 55–64 years, 15.8% 75–84 years, and 5.2% were 85 or older.	Arthritis was defined as provider‐diagnosed arthritis—excluding those with only fibromyalgia, gout, lupus or rheumatoid arthritis.	Unmet medical needs were measured use two delay‐of‐care variables. If a respondent answered yes to “during the past 12 months, have you delayed seeking medical care because of worry about the cost?” or “during the past 12 months was there any time when you needed medical care, but did not get it because you couldn't afford it?” they were considered to have delayed care due to cost. Respondents had an access‐related unmet need if they answered yes to any of the following reasons for delaying care: couldn't make telephone contact or get a prompt appointment; office not open during convenient hours; and lack of transportation.	The dependent variable characterized health care used in the previous 12 months: conventional and CAM (both), only conventional (conventional), only CAM (CAM), and no health care (neither). Conventional services included care from a primary care provider; a foot, women's health, or specialist physician; a physical, speech, respiratory, or occupational therapist; an audiologist; or a nurse practitioner, physician assistant, or midwife. CAM therapies included 36 therapies, including 10 provider‐based therapies (eg, acupuncture) and 26 therapies for which providers are not needed (eg, special diets).	Full sample Delayed care due to cost 14.60% (95% CI 13.52%–15.75%) a Delayed care due to access: 15.51% (95% CI 14.40%–16.68%) a Used both Delayed care due to cost 15.19% (95% CI 13.58%–16.95%) a Delayed care due to access: 18.68% (95% CI 16.92%–20.58%) a Conventional only Delayed care due to cost 12.81% (95% CI 11.41%–14.36%) a Delayed care due to access: 12.51% (95% CI 11.12%–14.04%) a CAM only Delayed care due to cost 34.80% (95% CI 23.11%–48.65%) a Delayed care due to access: 10.50% (95% CI 4.66%–21.98%) a Neither Delayed care due to cost 28.86% (95% CI 20.44%–39.06%) a Delayed care due to access: 16.10% (95% CI 9.87%–25.17%) a	N/A
Schwartz et al (2019) 21 Study setting: cross‐sectional (as relevant to our review) Study design: the United States	Used data 2015 MEPS‐HC. They limited their study population to those who were between the ages of 60 and 70 and who had a self‐reported diagnosis of arthritis (weighted n = 17,457,166). There were 50.3% between 60–64 years and 49.7% between 66–70 years. Participants between 60 and 64 reported primary insurance coverage as 19.8% public, 74.7% private, and 5.5% uninsured, and those between 66 and 70 reported insurance coverage as 42.0% public, 57.7% private, and 0.3% uninsured.	Arthritis was defined as self‐reported diagnosis of arthritis.	Unmet medical needs were measured using access to care variables, which gathers information on family members’ abilities to receive treatment and without delay. The question inquires whether the person was unable to receive treatment or was delayed in receiving treatment (within last 12 months). Answering yes indicates that the person needed treatment but was unable to receive it or was delayed in receiving it.	Data on types of health insurance were collected by the MEPS‐HC and were separated into public, private, and no insurance.	Ages 60–64 with arthritis reporting unmet medical needs Among those with public insurance 10.55% (95% CI 10.50–10.59) a Among those with private insurance 3.54% (95% CI 3.52–3.55) a Among those with no insurance 3.59% (95% CI 3.54–3.64) a Among those with all insurance types 4.93% (95% CI 4.91–4.94) a Public insurance: 2.14 Ages 66–70 with arthritis reporting unmet medical needs Among those with public insurance 2.48% (95% CI 2.47–2.50) a Among those with private insurance 2.07% (95% CI 2.06–2.08) a Among those with no insurance 42.77% (95% CI 42.14–43.40) a Among those with all insurance types 2.35% (95% CI 2.34–2.36) a

^*CAM, complementary and alternative medicine; CI, confidence interval; KHANES, Korean National Health and Nutrition Examination Survey; MEPS‐HC, Medicare Expenditure Panel Survey Household Component; N/A, no data given; NHIS, National Health Interview Survey.

^a Calculated by authors of this systematic review based on data reported in the study.

Prevalence of unmet needs stratified by health care use or insurance status

When stratified by types of health care use in the United States (Table 4), ¹⁸ the 12‐month prevalence of unmet medical needs due to unaffordability ranged from 12.8% (95% CI 11.4%–14.4%) among those relying solely on conventional health care to 34.8% (95% CI 23.1%–48.7%) among those who exclusively used CAM. The prevalence of unmet medical needs also varied when stratified by insurance status in the United States. ²¹ Among individuals aged 60 to 64 years with arthritis, the 12‐month prevalence of unmet medical needs ranged from 3.5% (95% CI 3.5%–3.6%) among those with private insurance to 10.6% (95% CI 10.5%–10.6%) among those with public insurance. In adults aged 66 to 70 with arthritis, the 12‐month prevalence of unmet medical needs ranged from 2.1% (95% CI 2.1%–2.1%) among those with private insurance to 42.8% (95% CI 42.1%–43.4%) among those who were uninsured. ²¹

Prevalence of unmet needs during the COVID‐19 pandemic

The prevalence of unmet health care needs varied during the COVID‐19 pandemic among individuals with arthritis in the United States, peaking in summer 2020 at 22%, followed by 9.4% in fall 2020 and 6.2% in winter 2021 (95% CI not reported) ²² (Table 5). In South Korea, the prevalence of unmet medical needs was 38.2% among individuals with arthritis in the summer of 2020 during the COVID‐19 pandemic. ²⁰

Table 5.

Prevalence and associated factors of unmet health care needs in adults with arthritis during the COVID‐19 pandemic^*

Author(s), year, study design	Setting, participants, number of participants, n	Definition of osteoarthritis or arthritis	Unmet health care needs definition	Other factors assessed	Prevalence	Associated factors
Tsuzaki and Taira (2022) 22 Design: cross‐sectional study Study setting: the United States	Data were analyzed from the MCBS COVID‐19 supplemental public use files for summer 2020, fall 2020 and winter 2021. The sample consisted of Medicare beneficiaries aged 65 years or older and people younger than 65 years with a Medicare‐qualifying disability (eg, end‐stage renal disease); 11,114 beneficiaries were interviewed for the summer survey, 9,686 for the fall survey, and 11,107 for the winter survey. All participants were living in the community. See Supplementary Material 4 for demographic characteristics.	Data on arthritis were self‐reported by the beneficiary as a condition diagnosed by a doctor or health professional.	Unmet health care needs were referred to as forgone health care because of the pandemic. This information was self‐reported in response to the question “Since [reference date], did you need medical care for something other than coronavirus, but not get it because of the coronavirus pandemic?” Forgone care included the need for urgent care, surgery, diagnostic tests, regular checkups, treatment of previous conditions, prescription drugs, and dental, vision, and hearing care. Answering yes to the above question was considered as a forgone health care because of the pandemic.	N/A	Beneficiaries with arthritis who reported forgone health care in: Summer 2020 22.5% (95% CI not reported) Fall 2020 9.4% (95% CI not reported) Winter 2021 6.2% (95% CI not reported)
Jung et al (2022) 20 Design: cross‐sectional study Study setting: Daejeon, South Korea	Used data from the Daejeon Metropolitan City Health Survey conducted from August 19 to September 29, 2020. A total of 1,494 adults (aged ≥18 years) were included, of whom 50.3% were female, 85.2% were aged <65 years. A total of 30 individuals reported having arthritis.	Authors selected five chronic diseases associated with health service use, including arthritis. To analyze the participants’ chronic diseases, they used the following question: “Are you currently receiving treatment for chronic diseases?”	Unmet medical needs were measured based on the following question: “Since the outbreak of COVID‐19, have you ever experienced unmet medical needs cause by COVID‐19?” Answers were coded as yes or no and scored as 1/0. The outbreak of the COVID‐19 pandemic is considered to be March 2020.	N/A	Those with arthritis reporting unmet medical needs: 38.2% (95% CI could not be calculated)	N/A

^*CI, confidence interval; MCBS, Medicare Current Beneficiary Survey; N/A, no data given.

Factors associated with unmet health care needs in individuals with osteoarthritis

One study ²³ examined whether socioeconomic factors were associated with unmet medical needs in individuals with osteoarthritis (Table 3). Compared to individuals with annual household income ≥$100,000, individuals with lower income were more likely to be unable to access needed medical care due to unaffordability, ranging from adjusted odds ratio (aOR)_{$75,000–$99,999} = 1.3 (95% CI 0.5–3.2) to aOR_<$35,000 = 7.7 (95% CI 3.7–16.0). Similarly, compared to those with a household income ≥$100,000, individuals with lower income were more likely to have delayed medical care due to unaffordability aOR_{$75,000–$99,999} = 1.3 (95% CI 0.6–2.2) to aOR_<$35,000 = 5.2 (95% CI 2.9–9.3).

Compared with individuals with private insurance, those who were uninsured or had other forms of insurance were more likely to be unable to access needed medical care due to unaffordability (aOR_uninsured = 4.8 [95% CI 3.0–7.7]; aOR_{other insurance} = 1.2 [95% CI 0.8–1.7]), whereas those with public insurance were less likely (aOR_{public insurance} = 0.7 [95% CI 0.5–1.1]). A similar pattern was observed for delayed care due to unaffordability (aOR_uninsured = 5.0 [95% CI 3.1–8.0]; aOR_{other insurance} = 1.0 [95% CI 0.7–1.5]; aOR_{public insurance} = 0.5 [95% CI 0.3–0.8]). Race was not associated with unmet medical needs in this study.

DISCUSSION

Evidence from our systematic review suggests that (1) nearly one‐third of adults ≥50 years with osteoarthritis or arthritis in South Korea experience unmet health care needs in the past 12 months, (2) unmet health care needs due to unaffordability ranged 15% to 30% in adults with osteoarthritis or arthritis in the general population in the United States, and (3) unmet health care needs vary by type of health care use (eg, CAM vs conventional health care) and is higher among those who are uninsured or without private insurance, have lower income, or have activity limitations. Findings suggest that the prevalence of unmet needs is higher among those who exclusively use CAM in the United States. This may reflect limited access to conventional health care services. This review finding is based on the study by Hoerster et al; in this study, CAM interventions included both provider‐based and self‐administered therapies, whereby self‐administered therapies did not require a visit to a provider. Moreover, this study suggests that uninsured individuals were more likely to use CAM alone or no care at all. ¹⁸ Taken together, it is possible that unmet health care needs are higher in those who use CAM alone due to issues with accessing health care. Future research in this area is needed to investigate this further. Evidence suggests that unmet health care needs due to disruptions from the COVID‐19 pandemic were common in South Korea and the United States and that unmet needs varied during the COVID‐19 pandemic in the United States (highest in summer 2020).

Our findings generally align with previous literature examining health care needs among adults aged ≥65 years in the general population. ⁴ In a systematic review focused on unmet health care needs among older adults in the general population, common barriers to accessing care included unaffordability and inaccessibility, ²⁶ which is reflected in our findings. Rahman et al also reported that unmet health care needs were more common among those who were uninsured and had a lower socioeconomic status, which also aligns with our findings. ⁴

Our findings inform future research directions on the burden and associated factors among individuals with osteoarthritis—a leading cause of disability worldwide. We identified only seven high‐quality studies, primarily focused on the prevalence of unmet needs in this population. Given the global burden of osteoarthritis, further research is needed in this area, particularly on the incidence and associated factors of unmet needs. Notably, the studies that focused specifically on osteoarthritis only included individuals with lumbar, knee, or hip osteoarthritis, conditions that are linked to mobility and activity limitations; these limitations may then negatively affect access to care. ³² Future studies should further explore how unmet needs may be driven by socioeconomic factors, insurance, and types of health care use to inform more targeted and equitable health care access and delivery, as well as resource planning. It is important to recognize that all included studies were conducted in the United States and South Korea and that contexts may differ from those in other health care systems worldwide. Differences in health care delivery models, insurance systems, and cultural attitudes toward seeking care could influence the burden and factors of unmet health care needs. Therefore, the results of this review may not be generalizable to other countries. Future research on unmet needs in other countries is warranted, particularly in low‐ and middle‐income countries, as well as in high‐income countries outside the United States and South Korea.

Our findings carry important policy implications. Given the substantial proportion of individuals with osteoarthritis reporting unmet health care needs, our review underscores the critical need for system‐level interventions that improve affordability, accessibility, and enhance transportation access for this population. Exploring opportunities to improve access to care such as potentially expanding insurance coverage and implementing targeted policies that address the unique needs of individuals may help reduce unmet care needs and improve health outcomes. Further research is needed to guide such policy initiatives.

Our review has strengths. Our protocol was established a priori and preregistered on the Open Science Framework. Our review used the reporting guidelines from PRISMA ⁶ and the SWiM checklist for reporting narrative syntheses, ¹⁶ and we used the JBI standardized critical appraisal tools to assess the methodologic quality of the evidence. We also developed a comprehensive search strategy across multiple databases, which was peer‐reviewed by a second librarian using the PRESS checklist. ¹² Additionally, we only included studies with low or moderate risk of bias in our synthesis, thereby improving the certainty of our findings.

There are limitations. First, we limited relevant studies to English, French, Chinese, and Italian languages for feasibility reasons, which may have excluded relevant articles in our synthesis. However, we identified only two possibly relevant articles in other languages, which are listed in Supplementary Material 3 for transparency. Second, we did not search the gray literature, so we could not assess for potential publication bias. Third, although our review included studies examining arthritis in the general population, we attempted to contact study authors to clarify the arthritis subtypes included but received no responses. As a result, we stratified findings from studies that specifically identified osteoarthritis from those reporting on arthritis more generally. Fourth, we did not include qualitative studies that were focused on the lived experiences of adults with osteoarthritis and unmet health care needs. Although such studies would not have addressed our research questions around prevalence and incidence of unmet needs, they could have provided valuable insights into factors contributing to unmet health care needs. This area warrants future research. Lastly, the interrater agreement (kappa statistic) was low for phase 1 screening despite high percent agreement (93.8%). This likely reflects the kappa paradox, wherein a high percentage of agreement can coincide with a low kappa value due to an imbalance in the distribution of ratings (eg, skewed prevalence). ³³ However, independent reviewers discussed disagreements to reach consensus in both screening phases.

Findings highlight that unmet health care needs are common among individuals with osteoarthritis. Evidence suggests that unmet needs are more common among those with lower income, lack of insurance, and activity limitations, and that prevalence of unmet needs varied with type of health care use (conventional vs CAM) and during the COVID‐19 pandemic. However, these findings are based on studies from only two high‐income countries (South Korea and the United States); therefore, additional research in this area is needed, particularly on the incidence and associated factors of unmet needs in adults with osteoarthritis in low‐ and middle‐income countries and other high‐income countries.

AUTHOR CONTRIBUTIONS

All authors contributed to at least one of the following manuscript preparation roles: conceptualization AND/OR methodology, software, investigation, formal analysis, data curation, visualization, and validation AND drafting or reviewing/editing the final draft. As corresponding author, Dr Ragagnin confirms that all authors have provided the final approval of the version to be published and takes responsibility for the affirmations regarding article submission (eg, not under consideration by another journal), the integrity of the data presented, and the statements regarding compliance with institutional review board/Declaration of Helsinki requirements.

Supporting information

Disclosure Form:

Data S1 Supporting Information

Data S2 PRISMA 2020 Checklist