SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. ROLE OF THE STUDY SPONSOR
  9. REFERENCES

Objective

To determine if a previously published model of the influence of self-rated health on physical, mental, and social health among patients with joint replacement surgery could be generalized to persons with symptomatic knee osteoarthritis (OA). Our second purpose was to determine if self-rated health mediated changes in physical, mental, and social health.

Methods

Persons with symptomatic knee OA (n = 1,127) who participated in the Osteoarthritis Initiative study completed the required measures at baseline and at 1-, 2-, and 3-year intervals. The key variable of interest was a single-item self-rated health measure. In addition, measures of physical, mental, and social health and a set of covariate measures over the 3-year period were analyzed. Structural equation modeling was used to test interrelationships among variables, as well as predictive and mediational relationships among self-rated health and mental, physical, and social health after adjusting for baseline covariates.

Results

The full model demonstrated good statistical fit. Prior self-rated health consistently predicted current mental health and social health. Prior social health predicted current self-rated health. Self-rated health also mediated changes in mental health and social health. Only social health changes were mediated by self-rated health over all time periods.

Conclusion

Self-rated health predicts a variety of outcomes of symptomatic knee OA. In addition, self-rated health mediates changes in social health and mental health. The use of self-rated health as a simple and efficient clinical assessment has potential for clinical utility because of its predictive capability and association with multiple health domains.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. ROLE OF THE STUDY SPONSOR
  9. REFERENCES

Self-ratings of general health are among the most commonly used assessments in epidemiologic research (1, 2). Self-rated health (SRH) is often measured using a singleitem scale: “In general, would you say your health is … ,” with possible responses of excellent, very good, good, fair, and poor. The construct of SRH appears to represent a synthesis of a patient's current physical, mental, and social health states, as well as a general enduring trait representing the individual's own views of their health along a continuum of healthy to unhealthy (3). Longitudinal studies are needed to test the utility of SRH as physical, mental, and social health conditions change over time following disease or injury.

Perruccio et al found that SRH measurements exhibited elements of changing and stable features in persons undergoing hip and knee arthroplasty over a 6-month period following surgery (4). They also examined the predictive validity of SRH for inferring current and future physical, mental, and social health constructs in a sample of persons undergoing joint arthroplasty. The authors concluded that SRH was associated with current and future health status and that clinicians should consider adoption of SRH as a clinical measurement to aid in outcome assessment and prognostic and treatment-based decisions for this particular group of patients (5).

Most arthroplasty patients experience substantial improvements in a short period of time following surgery (6–8). This is generally not the case for individuals with hip or knee osteoarthritis (OA) who are treated nonsurgically. Persons with nonsurgically treated arthritis generally report pain and disability that fluctuates, sometimes dramatically, over periods as short as a day (9) or over years (10–12). It is therefore critical to establish whether SRH assessment has potential as a prognostic or outcome measure for persons with symptomatic knee OA.

Because of their brevity, single-item scales are ideally suited for routine use in clinical practice. SRH as a single-item scale is unique among brief scales because of the volume of evidence indicating the utility of SRH responses for predicting a broad array of clinically important outcomes from physical, mental, and social health (3) to societal concerns such as health care utilization (13, 14) and mortality (15). Given that SRH is related to multiple constructs of interest to patients and clinicians and that it is simple to use, the SRH scale appears to have potential as an outcome and prognostic variable for persons with knee OA.

Our primary purpose was to test whether the model proposed and validated by Perruccio et al (5) would fit longitudinal data collected from persons with symptomatic knee OA, but who did not undergo arthroplasty. If the model proposed by Perruccio et al (5) generalized to persons with symptomatic knee OA, the model would be pertinent to a far greater number of patients and clinicians. Our secondary purpose was to determine if SRH mediated changes in health status over a 3-year period.

Significance & Innovations

  • Self-rated health (SRH) is a simple single-item clinical assessment that mediates both social and mental health changes in persons with symptomatic knee osteoarthritis (OA).

  • The use of a global SRH assessment should be considered for routine clinical assessment in patients with knee OA because it is both easy to use and mediates changes in future mental and social health, as well as predicts future physical, mental, and social health.

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. ROLE OF THE STUDY SPONSOR
  9. REFERENCES

The Osteoarthritis Initiative.

The Osteoarthritis Initiative (OAI) is a publicly and privately funded prospective longitudinal cohort study of 4,796 persons. A primary objective of the OAI study is to develop diverse cohorts of persons for the study of the natural history, risk factors, onset, and progression of knee tibiofemoral OA. The target population, those with or at risk for knee OA based on defined criteria, was recruited using focused mailings to a clinical population of persons with OA, newspaper advertisements, presentations at church, civic, and community organizations, and an educational web site about knee OA. The sample is community based to the extent that only 16.8% of persons whose data were used in the current study reported currently seeing a health care professional at baseline.

The OAI study followed 3 subcohorts: incidence, control, and progression. Each subcohort included racially and ethnically diverse patients between ages 45–79 years at baseline. The progression subcohort, which was targeted for this study, had 1,390 persons with symptomatic knee OA in one or both knees. Persons with symptomatic knee OA were defined by OAI investigators as having both of the following in at least 1 knee at baseline: 1) frequent knee symptoms in the past 12 months, defined as “pain, aching or stiffness in or around the knee on most days” for at least 1 month during the past 12 months, and 2) radiographic tibiofemoral knee OA defined as Osteoarthritis Research Society International (OARSI) atlas grades 1–3 (16), which are similar to Kellgren/Lawrence grades ≥2 (17) as measured on a standardized fixed-flexion radiograph. The study was approved by the respective institutional review boards at study sites and written informed consent was provided by all subjects. A complete study design protocol can be viewed at: http://www.oai.ucsf.edu/datarelease/docs/StudyDesignProtocol.pdf.

Study sample.

Subjects were recruited from 4 university-based recruitment centers: 1) the University of Maryland School of Medicine in Baltimore, 2) the Ohio State University in Columbus, 3) the University of Pittsburgh in Pittsburgh, Pennsylvania, and 4) Memorial Hospital of Rhode Island in Pawtucket.

Exclusion criteria were the presence of rheumatoid arthritis, bilateral knee arthroplasty or plans to undergo bilateral knee arthroplasty in the next 3 years, bilateral OARSI stage 3 knee OA, positive pregnancy test, inability to provide a blood sample, use of ambulatory aids other than a single straight cane for more than 50% of the time, comorbid conditions that might interfere with the 4-year participation, unlikely to reside in clinic area for at least 3 years, current participation in a double-blind, randomized controlled trial, and unwilling to sign informed consent. In addition, because of the need to track magnetic resonance imaging (MRI)–based changes, persons were excluded if they were unable to undergo 3.0T MRI testing. Accordingly, men weighing >130 kg and women weighing >114 kg were excluded.

Of the 1,390 persons in the OAI progression subcohort, 97 persons underwent either hip or knee arthroplasty during the 3-year followup and 166 persons had no repeated measurements. We therefore report results from 1,127 (81.1% of the sample) persons measured at least twice during the 3-year period. Despite a high rate of subject retention over the 3-year period, there was selective loss to followup in OAI, a problem common to most cohort studies (18). Persons lost to followup had greater baseline amounts of activity-related pain and physical disability, but similar levels of mental health, as compared to persons in the progression cohort who were included in our study. Persons lost to followup also tended to be male and nonwhite (Table 1).

Table 1. Characteristics of 1,127 persons in the study and 166 persons excluded due to missing data*
Baseline variablesPersons included (n = 1,127)Persons excluded due to missing data (n = 166)P, byt-test or chi-square test
  • *

    Values are the mean + SD unless indicated otherwise. CES-D = Center for Epidemiologic Studies Depression Scale; WOMAC = Western Ontario and McMaster Universities Osteoarthritis Index; KOOS = Knee Osteoarthritis Outcome Score; PASE = Physical Activity Scale for the Elderly.

  • Scores for WOMAC and CES-D decrease as symptoms or functional status improves.

  • Scores for KOOS pain, KOOS function and sports, and PASE scales increase as symptoms or functional status improves.

Age, years61.14 ± 9.1360.97 ± 9.210.83
Female sex, %55.665.60.015
White race, %72.652.7< 0.001
Body mass index, kg/m230.30 ± 5.431.45 ± 5.60.05
Comorbidity score0.46 ± 0.870.59 ± 1.160.17
Self-rated health, no. (%)   
 Excellent134 (12.0)6 (17.8) 
 Very good488 (43.7)51 (31.3)< 0.001
 Good413 (37.0)74 (45.4) 
 Fair76 (6.8)29 (17.8) 
 Poor5 (0.4)3 (1.8) 
CES-D mental health7.46 ± 7.4710.04 ± 10.05< 0.001
WOMAC physical function17.51 ± 12.6423.15 ± 15.05< 0.001
KOOS pain65.88 ± 18.958.19 ± 20.6< 0.001
KOOS function and sports54.66 ± 26.049.29 ± 27.180.08
PASE score162.29 ± 82.36135.18 ± 91.30< 0.001

OAI measurements used in the current study.

Data for the current study were obtained at the baseline clinic visit and during the 1-, 2-, and 3-year followup visits. The clinical data set release versions 0.2.2, 1.2.2, 3.2.1, and 5.2.1 were used (http://oai.epi-ucsf.org/datarelease/About.asp). At all time points, SRH was measured using the following self-report scale: “In general, would you say your health is … ,” and the respondent would select from the following options: excellent, very good, good, fair, and poor.

Activity-related pain was measured at all time points with the Knee Osteoarthritis Outcome Score (KOOS) pain scale, a 9-item scale that measures the extent of pain with common activities such as walking and stair climbing. The scale is scored from 0 (severely painful function) to 100 (no pain with function) and is highly reliable and valid (19). The OAI required subjects to complete the scale for each knee. We determined the lowest (greatest pain) score at each time period for the 2 knees of each subject and used this score for the analysis.

The Western Ontario McMaster Universities Osteoarthritis Index (WOMAC) physical function scale, a reliable and valid scale of activity limitations (20), has 17 items and is scored from 0 to 68, with higher scores indicating more severe activity limitations. Again, the OAI required subjects to complete a WOMAC scale for each knee, and we used the highest score (worst function) of the 2 scores obtained at each time point. To quantify limitation in higher-level activities, we used the validated KOOS function, sports, and recreational activities scale (21, 22), also scaled from 0 to 100, with higher scores indicating higher function. We used these 3 measures as indicators of a latent physical health variable in the structural equation model (SEM) described below.

For mental health, we used the Center for Epidemiologic Studies Depression (CES-D) Scale, a 20-item depression screener scored from 0 (no depressive symptoms) to 60 (severe depressive symptoms). Extensive reliability and validity evidence exists for the CES-D (23, 24). We used the CES-D as an observed measure of mental health.

We used the Physical Activity Scale for the Elderly (PASE) as a self-report measure of social health. The PASE contains 26 questions that quantify the extent of a person's leisure activities, household activities, and occupational and voluntary activities. The scale has high levels of reliability and validity and reflects a person's degree of involvement in socially expected roles (25, 26). The instrument was developed for older adults, but has been validated for persons as young as age 55 years (27).

For covariates we used the following variables: comorbidity (28), age, sex, level of education (less than high school, high school, some college, college degree, some graduate training, and graduate degree) (5). Comorbidity was assessed with a validated self-report instrument developed by Katz et al (28). The instrument contains 18 items (scoring range 0–32), with higher scores indicating greater numbers of comorbid conditions. In addition, we adjusted for race (white versus other) and body mass index, because these variables may influence relationships among health constructs and knee OA (29).

Statistical analyses.

The SEM was specified a priori based on work by Perruccio et al (5) and is depicted in Figure 1. Consistent with the prior research (5), each consecutive measurement of physical, mental, and social health is mediated by the concurrent SRH across 4 measurement occasions after controlling for a set of variables. In the SEM, the SRH scores ranging from 1 to 5 were treated as non-normal continuous variables in this study. Specifically, we used the restricted maximum likelihood estimator, also known as the Satorra-Bentler test statistic, to obtain the correct fit statistic and SEs (30). Physical health was represented as a latent variable with 3 indicators. Respective factor loadings were set equally across time to assure invariant measurements of physical health (31). All other variables were represented as observed variables. Restricted maximum likelihood method was used to estimate the model to account for non-normal continuous outcome variables. All available data were used during the estimation without resorting to an imputation method to replace the missing values.

thumbnail image

Figure 1. Longitudinal model from the structural equation modeling, indicating statistically significant paths (solid lines), nonsignificant paths (broken lines), and covariances (curved lines). The covariate box, along with the curved lines, indicate that both the self-rated (SR) health variable and the health status variables at each time point were adjusted for covariates in the final model. BMI = body mass index.

Download figure to PowerPoint

The model simultaneously tested SRH from several perspectives. First, it examined whether prior SRH was associated with current health status as measured by physical health, mental health, and social health. Second, the model examined whether prior health status was associated with current SRH. Third, it examined whether each of the measures in the model was associated with that same measure in the future. Fourth, it examined the association between current health status measures and current SRH. The SRH measures were adjusted for the covariates at baseline and all 3 followup periods. Finally, we tested whether changes in any of the health status measures were mediated by SRH.

We tested changes over 3 time periods: baseline to 1-year, 1-year to 2-year, and 2-year to 3-year. Mplus software (32) was used for model fitting. In addition to the significance of model parameters (e.g., path coefficient, factor loading), we tested the statistical significance of particular mediation pathways (e.g., physical2[RIGHTWARDS ARROW]self-rated health2[RIGHTWARDS ARROW]physical3).

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. ROLE OF THE STUDY SPONSOR
  9. REFERENCES

Descriptive statistics for the baseline and followup health status measures and SRH are reported in Table 2. The sample generally showed some improvement in their status from baseline to the 1-year followup, and then the sample demonstrated minor fluctuations for most measures during the subsequent 2- and 3-year followups with 1 exception: the social health measure (PASE) showed a consistent trend of worsening over the study period.

Table 2. Scores for physical, mental, and social health and self-rated health measures (n = 1,127)*
Health measureBaseline1-year followup2-year followup3-year followup
  • *

    WOMAC = Western Ontario and McMaster Universities Osteoarthritis Index; KOOS = Knee Osteoarthritis Outcome Score; CES-D = Center for Epidemiologic Studies Depression Scale; PASE = Physical Activity Scale for the Elderly.

  • Scores for WOMAC and CES-D decrease as symptoms or functional status improves.

  • Scores for KOOS pain, KOOS function and sports, and PASE scales increase as symptoms or functional status improves.

  • §

    For all variables, the proportion reporting change from the previous year is described as a percentage. The numbers in parentheses beside each percentage estimate represent the percentage of persons who demonstrated worsening followed by the percentage who demonstrated improvement, accounting for the scoring of each scale.

Physical health, mean ± SD    
 WOMAC physical function17.5 ± 12.615.4 ± 12.615.4 ± 13.015.8 ± 13.2
 KOOS pain65.9 ± 18.970.9 ± 19.670.3 ± 20.170.2 ± 20.6
 KOOS function and sports54.7 ± 26.061.9 ± 25.662.1 ± 25.860.7 ± 27.5
Mental health, mean ± SD    
 CES-D mental health score7.5 ± 7.57.6 ± 8.17.2 ± 7.77.5 ± 7.8
Social health, mean ± SD    
 PASE162.3 ± 82.4157.6 ± 83.1152.5 ± 82.8147.1 ± 83.5
Self-rated health, %    
 Excellent12.09.79.98.9
 Very good43.742.540.943.1
 Good37.039.840.339.0
 Fair6.87.28.48.4
 Poor0.40.80.50.6
Proportion reporting change from previous year§    
 In self-rated health34.9 (21.1, 13.8)30.5 (15.7, 14.8)34.5 (17.4,17.1)
 In CES-D mental health85.1 (43.6, 41.5)83.6 (39.3, 44.3)85.6 (45.7, 39.9)
 In PASE97.6 (52.3, 45.3)98.2 (53.2, 45.0)97.9 (55.7, 42.2)
 In KOOS pain92.7 (35.0, 57.7)89.8 (43.0, 46.8)88.3 (43.5, 44.8)
 In WOMAC physical function93.9 (39.7, 54.2)90.6 (44.4, 46.2)88.9 (47.9, 41.0)
 In KOOS function and sports84.3 (31.5, 52.8)82.2 (42.9, 39.3)84.4 (44.5,39.9)

The value for the chi-square test of model fit was χ2 = 1,311.87, 266df, P < 0.001. Fit indices showed a well-fitting model: Comparative Fit Index = 0.944, Tucker-Lewis Index = 0.924, root mean square error of approximation = 0.059 (90% confidence interval 0.056–0.062), and standardized root mean square residual = 0.079. Except for covariates, significant path coefficients are represented by the solid directional lines and nonsignificant paths by the broken lines in Figure 1.

As seen in Table 3, prior SRH predicted current social health during all 3 time periods and future mental health for the 1-year and 3-year time periods. Prior SRH also predicted current physical health for the 1-year and 2-year time periods. Prior social health predicted current SRH for all 3 time periods. Prior physical and mental health did not predict current SRH for any time period. Each health status measure in the model (SRH, physical, mental, and social health) predicted the same measure in the future for all time periods. Similarly, prior SRH predicted current SRH for all time periods. Current social and mental health predicted current SRH for all time periods, while current physical health predicted SRH for only the first year.

Table 3. Path coefficients for full longitudinal model*
TimeframeBaseline1-year2-year3-year
Standardized estimatePStandardized estimatePStandardized estimatePStandardized estimateP
  • *

    Higher scores in physical and social health indicate better health status, while higher scores in mental health and self-rated health indicate worse depressive symptoms or worse health.

Prior health status predicting current health status        
 Physical health  0.710< 0.0010.748< 0.0010.744< 0.001
 Mental health  0.643< 0.0010.704< 0.0010.678< 0.001
 Social health  0.601< 0.0010.608< 0.0010.647< 0.001
Prior self-rated health predicting current self-rated health        
 Self-rated health  0.510< 0.0010.540< 0.0010.523< 0.001
Prior self-rated health predicting current health status        
 Physical health  −0.084< 0.001−0.0800.002−0.0260.309
 Mental health  0.109< 0.0010.0430.0740.130< 0.001
 Social health  −0.121< 0.001−0.178< 0.001−0.095< 0.001
Prior health status predicting current self-rated health        
 Physical health  0.0060.871−0.0670.080−0.0140.716
 Mental health  0.0280.346−0.0360.223−0.0210.529
 Social health  0.137< 0.0010.153< 0.0010.175< 0.001
Current health status predicting current self-rated health        
 Physical health0.0310.3520.0070.8340.0760.0320.0360.350
 Mental health0.218< 0.0010.135< 0.0010.161< 0.0010.129< 0.001
 Social health−0.498< 0.001−0.390< 0.001−0.426< 0.001−0.417< 0.001
Effects of covariates on self-rated health        
 Female sex−0.0140.544−0.0240.236−0.0320.0980.0140.517
 Comorbidity0.103< 0.0010.0160.4320.0140.4290.0410.078
 White race−0.168< 0.001−0.079< 0.001−0.0280.204−0.089< 0.001
 Body mass index0.079< 0.0010.0210.3100.0340.0830.0090.696

Covariates showed statistically significant effects at baseline, with sex being the only covariate not demonstrating a significant effect. At least 1 covariate had a significant effect at each time point following the baseline period. After adjusting for prior physical, mental, and social health, prior SRH mediated changes in current health, most consistently through mental health and social health (Table 4). SRH mediated changes in social health at all time periods. Changes in physical health were not mediated by SRH.

Table 4. Mediation analyses for the model*
Type of mediationStandardized estimateP
  • *

    SRH = self-rated health.

Baseline physical health [RIGHTWARDS ARROW] 1-year physical health via baseline SRH−0.0030.371
1-year physical health [RIGHTWARDS ARROW] 2-year physical health via 1-year SRH−0.0010.836
2-year physical health [RIGHTWARDS ARROW] 3-year physical health via 2-year SRH−0.0020.382
Baseline mental health [RIGHTWARDS ARROW] 1-year mental health via baseline SRH0.026< 0.001
1-year mental health [RIGHTWARDS ARROW] 2-year mental health via 1-year SRH0.0060.079
2-year mental health [RIGHTWARDS ARROW] 3-year mental health via 2-year SRH0.021< 0.001
Baseline social health [RIGHTWARDS ARROW] 1-year social health via baseline SRH0.062< 0.001
1-year social health [RIGHTWARDS ARROW] 2-year social health via 1-year SRH0.068< 0.001
2-year social health [RIGHTWARDS ARROW] 3-year social health via 2-year SRH0.040< 0.001

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. ROLE OF THE STUDY SPONSOR
  9. REFERENCES

Our interest was in determining whether the model proposed by Perruccio et al (5) would generalize to the much larger population of persons with symptomatic knee OA who have not undergone joint replacement surgery. We also examined the extent to which SRH was a determinant and was influenced by a variety of health outcomes. Our model, like that of Perruccio et al, was designed to test the following key associations: 1) whether prior SRH predicted current health status, 2) whether prior health status predicted current SRH, and 3) whether SRH mediated changes in each of the health status measures. These key associations were examined simultaneously in the model while adjusting for the baseline covariates, repeated measures of the same health status and SRH measures, and concurrent associations among the measures.

Our results replicated some findings of Perruccio and colleagues, but also differed in several ways. First, our findings were similar in that prior SRH predicted current physical and mental health during 2 of 3 time periods, but unlike Perruccio et al, we found that prior SRH predicted current social health at all time periods. Perruccio and colleagues found that prior SRH most consistently predicted current mental health. Lower SRH predicted poorer subsequent health status. Second, we found that prior physical health did not predict current SRH, much like Perruccio and colleagues. However, we found that only prior social health consistently predicted current SRH, while Perruccio et al found that only prior mental health predicted current SRH.

Reasons for the differences in our findings in regard to effects of prior SRH or health status on current measures are likely due to the populations under study. Patients undergoing knee arthroplasty face a potentially stressful period of hospitalization and risks associated with a major surgery. The psychological impact of these factors likely increases a patient's general psychological distress and, following the immediate recovery period, the patient typically experiences a substantial reduction in anxiety and depression, as reported by Perruccio and colleagues. For persons in our sample with less severe arthritis, psychological stressors are much less immediate, and we observed much smaller changes in psychological status. We suspect this more stable period with no major psychological stressors as seen in our sample potentially explains the more limited role of mental health in our study compared to Perruccio et al.

Social health was the only predictor of future SRH in our study. Social health showed a consistent pattern of worsening over the study period, whereas the other measures showed only minor fluctuations after the first year. The impact of chronic knee arthritis appears to be greatest on daily interactions within a societal context, such as that measured by the PASE as compared to more proximate effects related to pain or person-level actions such as sitting, standing, and walking. Current social health also was more strongly associated with current SRH than physical health in our study. We speculate that social health, a construct that captures a person's roles and interactions with society, is more closely aligned with the SRH construct than the physical and mental health constructs in this population of persons with knee OA.

Arnadottir et al conducted a population-based cross-sectional study of 185 community-dwelling persons ages ≥65 years to determine which demographic and health-related factors predict current SRH (33); they reported that the strongest predictors of better SRH were higher scores on a subset of items from PASE and the Late Life Function and Disability Instrument (34), another scale designed to measure the extent of a person's engagement with society, in addition to age and depression status. The work of Arnadottir et al provides additional support for the influence of societal participation on ratings of SRH.

Third, we found that SRH consistently mediated changes particularly in social health and also in mental health, but not in physical health. The study by Perruccio et al found that SRH most consistently mediated changes in mental health (5). As reported earlier, we suspect that social health was consistently mediated by SRH because of the link between SRH and societal interaction, and because PASE measures showed consistent worsening over the study period. We designed our study to replicate the study by Perruccio and colleagues to the extent possible given the OAI data set limitations. Our physical health construct included very similar measures to those used by Perruccio et al (5) with the exception of a fatigue measure, which was not available in the OAI. We used the CES-D depression score to represent the mental health construct. Perruccio and colleagues used separate anxiety, depression, and fatigue measures to form their mental well-being construct. Because the CES-D measure has demonstrated moderate to strong associations with both fatigue and general psychological distress (35, 36), it appears to be a reasonable approximation of the mental well-being construct by Perruccio et al (5). For social health, Perruccio and colleagues used a variety of measures designed to capture the extent to which the patient completed socially expected roles and tasks. We used a single validated measure of task and role performance, the PASE (25, 26), which appeared to capture most of the social health dimensions assessed by Perruccio and colleagues (5). Finally, covariates showed stronger associations with baseline SRH as compared to followup, which we attribute to the more substantial changes during the first year. Race was most consistently associated with SRH over the study period.

To summarize, we contend that our statistical model is a reasonable approximation to that reported by Perruccio and colleagues (5) and that our findings demonstrate commonalities as well as differences. SRH appears to be a potentially useful outcome and prognostic measure not only for persons undergoing joint replacement surgery, but also for the much greater number of persons with symptomatic knee OA. SRH mediates changes consistently for both mental and social health.

We concur with Perruccio and colleagues (5), who argued for a more common use of SRH assessment both in epidemiologic and clinical settings. Given that SRH assessment requires only a single item and a few seconds of a patient's time, routine use of SRH assessment seems both potentially useful and practical. SRH measurements provide a broad assessment of multiple dimensions comprising overall health. For example, self-management training emphasizing behavior change and enhanced self-efficacy along with exercise for persons with symptomatic knee OA improve SRH as well as pain and function (37). Disturbed sleep, on the other hand, is associated with worse SRH in persons with symptomatic knee OA (38). Complementing these reports, our findings and those of Perruccio et al (5) suggest that SRH assessments can inform the clinician on a wide range of health-related behaviors.

Our study has important limitations that warrant discussion. We had a large sample, but relatively small numbers of persons at baseline (n = 81) with SRH scores of either fair or poor, which may reduce generalizability, particularly to persons with poor SRH. We also had persons with missing followup data generally having worse pain and physical and social health. The missing data may have diluted our findings, since persons lost to followup likely would have demonstrated larger fluctuations in symptoms over the study period given that their health was worse at baseline. The loss to followup, while common for large sample cohort studies (18), may adversely affect the generalizability of our findings.

In conclusion, our study demonstrates the potential utility of SRH assessments for persons with symptomatic knee OA in clinical settings. Our study suggests that a simple-to-use single-item SRH measure is predictive of future health status and mediates changes in both mental and social health.

AUTHOR CONTRIBUTIONS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. ROLE OF THE STUDY SPONSOR
  9. REFERENCES

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be submitted for publication. Dr. Riddle had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Riddle, Dumenci.

Acquisition of data. Riddle.

Analysis and interpretation of data. Riddle, Dumenci.

ROLE OF THE STUDY SPONSOR

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. ROLE OF THE STUDY SPONSOR
  9. REFERENCES

The Osteoarthritis Initiative study sponsors (Merck Research Laboratories, Novartis Pharmaceuticals Corporation, GlaxoSmith-Kline, and Pfizer) had no role in the study design, data collection, data analysis, or writing of this manuscript. Publication of this article was not contingent on the approval of these sponsors.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. ROLE OF THE STUDY SPONSOR
  9. REFERENCES
  • 1
    Kondo N, Sembajwe G, Kawachi I, van Dam RM, Subramanian SV, Yamagata Z. Income inequality, mortality, and self rated health: meta-analysis of multilevel studies. BMJ 2009; 339: b4471.
  • 2
    Blazer DG. How do you feel about…? Health outcomes in late life and self-perceptions of health and well-being. Gerontologist 2008; 48: 41522.
  • 3
    Bailis DS, Segall A, Chipperfield JG. Two views of self-rated general health status. Soc Sci Med 2003; 56: 20317.
  • 4
    Perruccio AV, Badley EM, Hogg-Johnson S, Davis AM. Characterizing self-rated health during a period of changing health status. Soc Sci Med 2010; 71: 163643.
  • 5
    Perruccio AV, Davis AM, Hogg-Johnson S, Badley EM. Importance of self-rated health and mental well-being in predicting health outcomes following total joint replacement surgery for osteoarthritis. Arthritis Care Res (Hoboken) 2011; 63: 97381.
  • 6
    Fitzgerald JD, Orav EJ, Lee TH, Marcantonio ER, Poss R, Goldman L, et al. Patient quality of life during the 12 months following joint replacement surgery. Arthritis Rheum 2004; 51: 1009.
  • 7
    Fortin PR, Clarke AE, Joseph L, Liang MH, Tanzer M, Ferland D, et al. Outcomes of total hip and knee replacement: preoperative functional status predicts outcomes at six months after surgery. Arthritis Rheum 1999; 42: 17228.
  • 8
    Lingard EA, Katz JN, Wright EA, Sledge CB. Predicting the outcome of total knee arthroplasty. J Bone Joint Surg Am 2004; 86A: 217986.
  • 9
    Allen KD, Coffman CJ, Golightly YM, Stechuchak KM, Keefe FJ. Daily pain variations among patients with hand, hip, and knee osteoarthritis. Osteoarthritis Cartilage 2009; 17: 127582.
  • 10
    Holla JF, Steultjens MP, Roorda LD, Heymans MW, ten Wolde S, Dekker J. Prognostic factors for the two-year course of activity limitations in early osteoarthritis of the hip and/or knee. Arthritis Care Res (Hoboken) 2010; 62: 141525.
  • 11
    Mallen CD, Peat G, Thomas E, Lacey R, Croft P. Predicting poor functional outcome in community-dwelling older adults with knee pain: prognostic value of generic indicators. Ann Rheum Dis 2007; 66: 145661.
  • 12
    Sharma L, Cahue S, Song J, Hayes K, Pai YC, Dunlop D. Physical functioning over three years in knee osteoarthritis: role of psychosocial, local mechanical, and neuromuscular factors. Arthritis Rheum 2003; 48: 335970.
  • 13
    Eisen SV, Bottonari KA, Glickman ME, Spiro A III, Schultz MR, Herz L, et al. The incremental value of self-reported mental health measures in predicting functional outcomes of veterans. J Behav Health Serv Res 2011; 38: 17090.
  • 14
    Kennedy BS, Kasl SV, Vaccarino V. Repeated hospitalizations and self-rated health among the elderly: a multivariate failure time analysis. Am J Epidemiol 2001; 153: 23241.
  • 15
    Idler EL, Benyamini Y. Self-rated health and mortality: a review of twenty-seven community studies. J Health Soc Behav 1997; 38: 2137.
  • 16
    Altman RD, Hochberg M, Murphy WA Jr, Wolfe F, Lequesne M. Atlas of individual radiographic features in osteoarthritis. Osteoarthritis Cartilage 1995; Suppl A: 370.
  • 17
    Kellgren JH, Lawrence JS. Radiological assessment of osteo-arthrosis. Ann Rheum Dis 1957; 16: 494502.
  • 18
    Hunt JR, White E. Retaining and tracking cohort study members. Epidemiol Rev 1998; 20: 5770.
  • 19
    Garratt AM, Brealey S, Gillespie WJ. Patient-assessed health instruments for the knee: a structured review. Rheumatology (Oxford) 2004; 43: 141423.
  • 20
    Beaton DE, Schemitsch E. Measures of health-related quality of life and physical function. Clin Orthop Relat Res 2003; 413: 90105.
  • 21
    Roos EM, Lohmander LS. The Knee Injury and Osteoarthritis Outcome Score (KOOS): from joint injury to osteoarthritis. Health Qual Life Outcomes 2003; 1: 64.
  • 22
    Roos EM, Bremander AB, Englund M, Lohmander LS. Change in self-reported outcomes and objective physical function over 7 years in middle-aged subjects with or at high risk of knee osteoarthritis. Ann Rheum Dis 2008; 67: 50510.
  • 23
    Ros L, Latorre JM, Aguilar MJ, Serrano JP, Navarro B, Ricarte JJ. Factor structure and psychometric properties of the Center for Epidemiologic Studies Depression Scale (CES-D) in older populations with and without cognitive impairment. Int J Aging Hum Dev 2011; 72: 83110.
  • 24
    Radloff LS. The CES-D scale: a self-report depression scale for research in the general population. Appl Psychol Meas 1977; 1: 385401.
  • 25
    Martin KA, Rejeski WJ, Miller ME, James MK, Ettinger WH Jr, Messier SP. Validation of the PASE in older adults with knee pain and physical disability. Med Sci Sports Exerc 1999; 31: 62733.
  • 26
    Dinger MK, Oman RF, Taylor EL, Vesely SK, Able J. Stability and convergent validity of the Physical Activity Scale for the Elderly (PASE). J Sports Med Phys Fitness 2004; 44: 18692.
  • 27
    Washburn RA, McAuley E, Katula J, Mihalko SL, Boileau RA. The physical activity scale for the elderly (PASE): evidence for validity. J Clin Epidemiol 1999; 52: 64351.
  • 28
    Katz JN, Chang LC, Sangha O, Fossel AH, Bates DW. Can comorbidity be measured by questionnaire rather than medical record review? Med Care 1996; 34: 7384.
  • 29
    Zhang Y, Jordan JM. Epidemiology of osteoarthritis. Rheum Dis Clin North Am 2008; 34: 51529.
  • 30
    Satorra A, Bentler PM. Corrections to test statistics and standard errors in covariance structure analysis. In: Von Eye A, Clogg CC, editors. Latent variable analysis: applications to developmental research. Thousand Oaks (CA): Sage; 1994. p. 399419.
  • 31
    Meredith W. Measurement invariance, factor analysis and factorial invariance. Psychometrika 1993; 58: 52543.
  • 32
    Muthen LK, Muthen BO. Mplus user's guide. Los Angeles: Muthen & Muthen; 2011.
  • 33
    Arnadottir SA, Gunnarsdottir ED, Stenlund H, Lundin-Olsson L. Determinants of self-rated health in old age: a population-based, cross-sectional study using the International Classification of Functioning. BMC Public Health 2011; 11: 670.
  • 34
    Jette AM, Haley SM, Coster WJ, Kooyoomjian JT, Levenson S, Heeren T, et al. Late life function and disability instrument. I. Development and evaluation of the disability component. J Gerontol A Biol Sci Med Sci 2002; 57: M20916.
  • 35
    Milette K, Hudson M, Baron M, Thombs BD. Comparison of the PHQ-9 and CES-D depression scales in systemic sclerosis: internal consistency reliability, convergent validity and clinical correlates. Rheumatology (Oxford) 2010; 49: 78996.
  • 36
    Wang B, Gladman DD, Urowitz MB. Fatigue in lupus is not correlated with disease activity. J Rheumatol 1998; 25: 8925.
  • 37
    Yip YB, Sit JW, Wong DY, Chong SY, Chung LH. A 1-year follow-up of an experimental study of a self-management arthritis programme with an added exercise component of clients with osteoarthritis of the knee. Psychol Health Med 2008; 13: 40214.
  • 38
    Allen KD, Renner JB, DeVellis B, Helmick CG, Jordan JM. Osteoarthritis and sleep: the Johnston County Osteoarthritis Project. J Rheumatol 2008; 35: 11027.