Health Coverage and its Relation to the Prevalence and Intensity of Symptomatic Knee Osteoarthritis

Abstract

Objective

To determine if health coverage among older adults is associated with 1) the prevalence of symptomatic knee osteoarthritis (OA) and 2) disablement in those with symptomatic knee OA.

Methods

Data were collected from the Osteoarthritis Initiative (OAI) dataset 5.2.1, a cohort study of subjects with or at risk for knee OA. Prevalence of symptomatic knee OA (knee symptoms on most of the last 30 days and Kellgren-Lawrence grade ≥2 on knee radiograph) was compared between those with and without health coverage amongst subjects age 45–65, adjusted for age and BMI. For those with symptomatic knee OA, Physical Activity for the Elderly (PASE) scores, and Knee Osteoarthritis Outcomes Survey (KOOS) function, pain, and quality of life scores were compared between those with and without health coverage before and after adjustment for age and BMI.

Results

Among subjects with health coverage, 27.8% had symptomatic knee osteoarthritis compared with 36.1% of those without health coverage (p=0.0204 before and >0.24 after adjustment). Among subjects with symptomatic knee OA with and without health coverage physical activity differed significantly (p=0.048), as did pain (p<0.0001), function (p=0.0001), and quality of life (p<0.0001).

Conclusion

Lack of health coverage was not associated with the prevalence of symptomatic knee OA after adjustment. However, those with symptomatic knee OA without health coverage reported reduced physical activity, greater pain, worse functional limitations, and decreased quality of life.

Introduction

Osteoarthritis (OA) is the most common musculoskeletal disorder.¹ An estimated 27 million US adults have clinical OA.² Estimates vary as to the economic impact of OA, although in all studies the impact is substantial. Mean out-of-pocket expenditures attributed to OA are estimated to be $1,379 per annum for women with OA and $694 per annum for men with OA. In addition, third-party payors spend an average of $4,833 per woman and $4,036 per man each year.³ The total cost is even higher if productivity losses are included.^4–5 Aggregate annual medical care expenditures in the United States on OA are estimated at $185.5 billion.³ The economic costs alone of osteoarthritis make it an important health topic.

The knee is the most common weight-bearing joint affected by OA,⁶ with a lifetime risk of 44.7% (95% CI: 40.0–49.3%.).⁷ Symptomatic knee OA is the combination of radiographic features of OA with daily symptoms. Since symptoms limit physical function, prompting patients to seek health care, symptomatic knee OA has great relevance to public health. Knee OA has been shown to significantly limit mobility and activities of daily living.⁸ Further, symptomatic knee OA can have a significant negative impact on quality of life.

The societal impact of knee OA is expected to grow for several reasons. The prevalence is expected to increase over the next 2 decades due in large part to the increasing prevalence of obesity.⁹ Concurrent with the increasing prevalence of knee OA, the number of Americans without health coverage has also increased. Between the years 2000 and 2006, the number of uninsured Americans increased from 38.4 million to 47.0 million.¹⁰ In the United States, health coverage has been tightly connected with employment and the unemployment rate also has risen over the last several years, potentially contributing to an increase in the number of Americans without health coverage. Further, an association has been shown between income level and obesity¹¹, so the decrease in salaries associated with a loss of employment may also increase the rate of obesity.

Adults without health coverage experience worse health outcomes than those with coverage.¹² In addition, the loss of health coverage is associated with a decrease in access to a usual source of care and an increased rate of reporting no physician visits in the past 12 months.¹³ In one study, patients with arthritis related conditions (arthritis, rheumatoid arthritis, gout, lupus, or fibromyalgia) without health coverage are 4.7 times as likely to report no visits to healthcare professionsals in the past 12 months than those with coverage. (AJ Davidoff, 2005 Uninsured Americans with Chronic Health Conditions: Key Findings from the National Health Interview Survey) A reduced number of visits to primary care providers by those without health coverage may indicate that those with symptomatic knee OA who lack health coverage are not diagnosed as quickly, and are less likely to receive appropriate treatment. People with arthritis and without coverage were reported to have significantly more healthcare visits related to their arthritis after becoming eligible for Medicare than those who had been continuously insured.¹³ This likely indicates that those without health coverage may be failing to receive appropriate medical care and as a result when they do gain health coverage they require more care to deal with a condition that is worse than it otherwise might have been. For this reason, there is legitimate concern that a lack of health coverage might impact the prevalence and/or severity of symptomatic knee OA.

To our knowledge, no study has yet investigated whether 1) the prevalence of symptomatic knee OA differs between those with and without health coverage and 2) whether activity level, knee pain, knee function, and quality of life are worse in those with symptomatic knee OA who lack health coverage. Determining if a gap in impairment and physical functional limitations exists between adults with and without health coverage may facilitate decision-making regarding the allocation of limited health care resources. Therefore, to better inform public health policies we assessed these relationships.

Materials and Methods

Subjects and Data Collection

This study was a cross-sectional analysis of baseline data from subjects in the Osteoarthritis Initiative (OAI), a publicly available, multi-center, longitudinal, prospective observational study of adults.¹⁴ Subjects in OAI included men and women between the ages of 45 and 79 years, who were either overweight (defined with separate cutoffs dependent on age and sex), had a previous knee injury or surgery, had “pain, aching, or stiffness, in or around the knee on most days for at least one month within the past year,” or had a parent or a sibling who had had a knee replacement. OAI also included a control group without knee symptoms or radiographic findings of knee OA. Exclusion criteria included rheumatoid arthritis, bilateral knee joint replacement, an inability to walk without assistance, or a contraindication to undergoing a knee MRI. All OAI subjects were enrolled between February 2004 and May 2006. All subjects completed an informed consent process approved by the investigators’ institutional review boards.

After a telephone interview confirmed initial eligibility subjects attended an initial assessment visit where height (mm) and weight (kg) without shoes, jewelry, wallets and in light weight clothing was collected for calculation of BMI. Additionally, subjects underwent bilateral fixed flexion knee radiographs for assessment of radiographic knee OA, defined as Kellgren-Lawrence (KL) grade 2 or higher. ¹⁵

At a second study visit, self-administered questionnaires including the Knee Osteoarthritis Outcome Scale (KOOS),¹⁶ The Physical Activity for the Elderly (PASE) scale,¹⁷ and the Center for Epidemiologic Studies Depression (CES-D) Scale,¹⁸ were administered. The KOOS is a validated self-report of knee-related impairments and functional limitations, The PASE is a brief, easily scored, instrument for the assessment of physical activity in epidemiologic studies of older people. The CES-D is a questionnaire developed for use in studies of the epidemiology of depressive symptoms in the general population that can be used as a screening tool for depression. CES-D was measured to investigate the possibility of depression affecting the outcome scores on the above surveys. The time elapsed between the telephone interview, initial assessment, and second visit was variable accommodating the schedules of each participant but in all cases no more than six weeks were allowed to elapse in total between the initial telephone interview and the second visit.

Much of the demographic information from our subjects comes from a self-administered take-home questionnaire that was completed in the time between their first and second visits, this included information such as marital status, education history, yearly income, health coverage status, the Charlson Comorbidity Index¹⁹, the presence of frequent knee symptoms, and physician-diagnosed arthritis.

Central to our study was determining whether or not a patient had a form of heath coverage. Health coverage status was determined with the question, “Do you currently have any kind of health care coverage? This would include private health coverage (such as Blue Cross), prepaid plans (such as health maintenance organizations),PPO’s, or any government-sponsored plans, such as Medicare, Medicaid, or VA coverage.”

To determine whether a patient could be considered to have symptomatic knee OA, it had to be determined if the patient had frequent knee OA symptoms. The following questions were used to determine the presence of frequent knee symptoms: “During the past 30 days have you had any pain, aching, or stiffness in your (right or left) knee?” Subjects who answered yes were then asked, “During the past 30 days have you had this pain, aching, or stiffness in your (right or left) knee on most days.” Subjects who answered this second question affirmatively were considered to have frequent knee symptoms.

In addition, we wanted to see whether a difference existed in the rates of previous diagnosis of knee OA between those with and without health coverage amongst those we determined to have symptomatic knee OA. As discussed in the introduction, decreased access to medical care may delay the diagnosis of knee OA for those without health coverage. The following question was used to determine the presence of physician-diagnosed arthritis: “Has a doctor ever told you that you have some other (besides rheumatoid) arthritis?” Subjects were then asked “What kind of arthritis did the doctor say it was? Did the doctor say you had osteoarthritis or degenerative arthritis of the knee?”

Subjects included in our study met the OAI study criteria and were between the ages of 45 – <65. This age range was chosen because OAI included Medicare as health coverage, and we found only 3 subjects over age 65 reported not having health coverage. Medicare is a form of health coverage provided by the Federal Government of the United States of America that is accessible for all citizens over the age of 65. We felt that including subjects over 65 would unnecessarily skew the difference in mean ages between those with and without health coverage and so only included subjects between ages 45 and 65. Additionally, only subjects without missing data regarding their health coverage status and symptomatic knee OA status were included.

Symptomatic knee OA was defined as the combination of the presence of knee symptoms on most of the past thirty days (frequent knee symptoms) and radiographic knee OA (KL ≥2). For our study, physical activity level was assessed using PASE¹⁷, while pain intensity, lower limb function, and quality of life were assessed by the KOOS¹⁶. Data from the CES-D questionnaire was used to characterize whether there was a difference in prevalence of depressive symptoms among subjects with and without health coverage. ¹⁸

Statistical Analyses

Differences in the prevalence of symptomatic knee OA among those with and without health coverage were analyzed first using the chi square test and then using logistic regression to adjust for age and BMI. In the subset of those with symptomatic knee osteoarthritis, generalized linear models, adjusting for age and BMI, were used to assess whether physical activity level, pain, functional limitation, and quality of life differed between those with and without health coverage. All analyses were conducted using SAS version 9.1, (SAS Institute, Inc. Cary, NC) P values <0.05 were considered to be significant.

Results

Characterization of Subjects

Out of 4,796 OAI subjects, 2,937 had complete data on health coverage and symptomatic knee OA status, were between ages 45 and 65, and thus were eligible to be included in our study. Significant differences between those with and without health coverage were found in age, BMI, marital status, income, CES-D score, and education. Differences in Charlson comorbidity index scores, and sex distribution were not found to be significant. The results for these can be seen in table 1 except for income and educational level, which are reported here.

Table 1.

Characterization of Subjects

	Subjects With Health Coverage (N=2771)	Subjects Without Health Coverage (N=166)	p value
Age (Mean ±SD)	54.9±0.1	53.5±0.4	0.05
BMI (Mean ±SD)	28.8± 5.01	30.9± 6.12	<0.001
Percent of Subjects Married	64.9%	27.1%	<.0001
CES-D Score (Mean ±SD)	8.01 ± 8.4	13.3 ± 9.0	<0.0001
Charlson Comorbidity Index (Mean ±SD)	0.3183± 0.7686	0.3478 ± 0.7767	0.6401

Our results show the median income for persons with health coverage was reported as being between $50,000 and $100,000 with an interquartile range from the $25,000–$50,000 to $100,000 or greater. Persons without health coverage on reported median income between $10,000 and $25,000 with an interquartile range from less than $10,000 to between $25,000–$50,000 (p<0.0001).

For education the median for persons with health coverage was some graduate schooling with an interquartile range of having a college degree versus a graduate degree, whereas the median for persons without health coverage having completed some college with an interquartile range of a high school degree versus a college degree (p<0.0001).

Our demographic results indicate that persons with health coverage were older, had lower BMI, were more likely to be married, had a higher income, had less depression symptoms, and had higher education than those without health coverage.

Relationship Between Health Coverage and Prevalence of Symptomatic Knee OA

Subjects with health coverage had a prevalence of symptomatic knee osteoarthritis of 27.8% while subjects without health coverage had a prevalence of symptomatic knee osteoarthritis of 36.1%. When unadjusted for age and BMI, this is statistically significant (p=0.0204). However, after adjustment for age and BMI this difference was no longer significant (p=0.25).

In those with symptomatic knee OA, a significant difference (p=0.0135) was found between those with health coverage and those without health coverage in the percentage of individuals who reported physician diagnosed knee OA (45.6% vs 29.5%, respectively.) While actual rates of symptomatic knee OA were not different to a statistically significant degree, rates of diagnosis prior to entry in the study were, correlating with past studies showing decreased access to health care for those without health coverage. ¹³

Relationship Between Health Coverage and Disablement in Those with Symptomatic Knee OA

Table 2 presents the differences in physical activity level, pain, function, and quality of life between subjects with and without health coverage who have symptomatic knee osteoarthritis, after adjustment for age and BMI. These results show significantly decreased activity level, higher pain, lower function, and lower quality of life amongst those without healthcare coverage compared to those with coverage. It was also examined whether KL grade differed between persons with symptomatic knee OA with and without health coverage. It was found that both before and after adjustment for age and BMI no significant difference could be found.

Table 2.

Knee OA Outcome Measures in Subjects with Symptomatic Knee OA

Outcome	With Health Coverage (n=770)	Without Health Coverage (n=60)	Unadjusted p value	Adjusted* p-value
Knee pain severity score (KOOS) **	98.6±6.7	88.3±7.1	<.0001	<0.0001
Function (KOOS)	53.2±0.9	43.4±4.2	<.0001	<0.0001
Daily Activity Score (PASE)	179.3±87.3	149.95±84.9	0.0122	0.0282
Quality of Life (KOOS)	49.6±20.3	40.3±19.5	<.0001	<0.001

^*All variables adjusted for age and BMI.

^**Note that a lower Knee Pain score indicates greater pain.

Discussion

This study addressed two novel questions: 1) Whether an association exists between lack of health coverage and the prevalence of symptomatic knee osteoarthritis, and 2) whether among individuals with symptomatic knee OA there is a significant difference in knee pain, function, activity level and/or quality of life between those with and without healthcare coverage. We hypothesized that those without healthcare coverage would have a greater prevalence of symptomatic knee OA, but did not find a significant association after adjustment for age and BMI. Specifically, body mass index was found to substantially account for the difference in presence of symptomatic knee OA between the groups. We found a significant association between low income levels and a higher BMI in concordance with an earlier report.¹¹ It has been shown that BMI is an independent risk factor for knee OA and symptomatic knee OA. ²⁰ It seems unlikely that income is mediating the relationship between BMI and symptomatic knee OA and that instead income and healthcare coverage are likely intricately linked variables.

The second main question of this study was whether, among people with symptomatic knee OA, there was an association between having health coverage and the level of physical activity level, knee pain, function, and quality of life. The results of this study suggest that lack of health coverage was associated with less physical activity, higher pain intensity, lower knee function, and lower quality of life. This finding remained significant after adjustment for radiographic severity. Due to the cross-sectional design of this study, longitudinal studies are indicated to determine the temporal relationship between lack of health coverage and the outcome differences found between the two groups. The lower rate of physician diagnosis amongst subjects with symptomatic knee OA without health coverage does indicate a decreased amount of health care services received. Other evidence supporting this association is found in a prior study, which demonstrated that patients with severe OA with lower income are offered the option of arthroplasty at a lower rate than those with higher income.²¹ However, this evidence may not be directly applicable to our patient population.

Other possibilities to explain the differences in outcome between those with and without health coverage include impaired access to adequate analgesics or supportive braces/orthoses among those without healthcare coverage. In addition, decreased access to healthcare might also reduce the psychological benefits that such healthcare could provide to lower pain. These include the certainty of a diagnosis, education regarding prognosis, and the placebo effect that can occur with any intervention, all of which could possibly influence a person’s perception of pain. If the better outcomes seen in those with health coverage can be linked to specific treatments and interventions, greater efficacy and efficiency in treatment might be possible, in turn leading to less patient disablement and optimization of the use of limited healthcare resources.

The fact that the mean CES-D score of subjects without health coverage was above the screening cut off for depression, even after adjustment for age and BMI, is an interesting finding, and is also worthy of additional study. We looked at this measure out of concern that increased depression among those without health coverage may have affected their survey results negatively. It is an important finding because treating depression in patients with arthritis has been associated with increased physical function, decreased pain and increased quality of life. ²²

Our search for reports of other studies that demonstrated a similar difference in depression scores between those with osteoarthritis with and without health coverage did not reveal any that were specific to osteoarthritis distinct from arthritis in general. As the current study is cross-sectional, discussion of causation for the difference in CES-D scores between those with and without health coverage would be purely speculative. Previous studies have shown that serious psychological distress among patients with arthritis was significantly associated with younger age, lower socioeconomic status, divorce/separation, recurrent pain, physical inactivity, having functional or social limitations, and having comorbid medical conditions.²³ In our study, persons with health coverage were found to be better educated, have higher income, and were more likely to be currently married. It should be noted that the CES-D was developed for use in studies of the epidemiology of depressive symptoms in the general population, and is not intended for a clinical diagnosis of depression.¹⁸

The main strengths of this study include the large community-acquired cohort and the high quality data collection of the OAI. A limitation of the cross-sectional design of the study is that results can only demonstrate associations but cannot establish temporal relationships nor causation between variables. Another limitation was that primary outcome measures within the symptomatic knee OA population were survey based and thus somewhat subjective. This allowed the outcome measure to be more easily influenced by factors such as participant depression levels.

Although this study did not establish the reason for a relationship between a lack of health coverage and worsened outcomes amongst persons with symptomatic knee OA, the results of this study indicate a need for resources amongst individuals without healthcare coverage. It also indicates a compelling reason for future studies that might develop new therapeutic methods for symptomatic knee OA that are accessible to those without health coverage.

Conclusions

This study demonstrated that adults without health coverage have a higher prevalence of symptomatic knee OA in comparison to those with health coverage, but that significance is lost once BMI is adjusted for. In those with symptomatic knee OA there was a significant difference between those with and without health coverage in having been diagnosed by a physician prior to entry in the study. Lower activity levels, increased pain, lower knee-specific function, greater and decreased quality of life was found among subjects without health coverage with symptomatic knee OA in comparison to subjects with symptomatic knee OA who had health coverage. Future research should be directed at establishing temporal causality for these outcome gaps, as well as discovering new therapeutic measures that can be easily accessed by those without health coverage.