Physical self perceptions of women with rheumatoid arthritis




To assess the reliability and validity of the Physical Self-Perception Profile (PSPP) and the Perceived Importance Profile (PIP) and to assess relationships between these scales and disease, function, and negative affect in rheumatoid arthritis (RA) patients.


Fifty-two women (mean ± SD age 48.4 ± 10.4 years) completed the PSPP, PIP, and other measurements: the core measures of European League Against Rheumatism; the Hospital Anxiety and Depression Scale; distance walked; and peak and extrapolated maximal oxygen consumption during a 10-meter shuttle walk test.


PSPP subscales showed high internal consistency (Chronbach's alpha 0.73–0.81) and factor structure and strong relationships with physical self-worth (PSW; r = 0.40–0.63). Multiple regression analysis showed that all subscales (except sport) significantly contributed to PSW variance (R2 = 59.1%). Very low PSPP scores, particularly for strength and sport competence, and PIP scores were observed in RA patients (significantly lower than US college-aged and obese women), which were reflected in low PSW scores. Aspects of PSPP were related to depression and swollen joint count but not functional fitness. Discrepancy scores were associated with lower PSW scores (r = 0.48), substantiating that subjects were unable to meet their perceptual needs concerning their physical selves.


The PSPP and PIP are both reliable and valid and are sensitive to significant constructs in the mental health of women with RA. The PSPP appears to measure distinct mental properties not represented in other common RA measures; hence it may be useful in measuring an important aspect of RA patients' psychology.


Patients' perceptions are an important outcome of clinical treatment; satisfaction, confidence, mental wellbeing, and life quality have all become important targets. Specifically, self esteem has been identified as a key component of mental health as well as a determinant of a range of health behaviors (1), and so warrants further attention in the therapeutic setting. In recent years, significant advances in our understanding of the nature and construction of the self have taken place. In particular, it has become clear that the self is multidimensional and possibly hierarchically organized (2).

Individuals make global judgments of their self worth, or esteem, based upon the quality of their interactions in several life domains, such as work, family, and social life. A number of instruments have been developed to simultaneously assess adult's self perceptions in these domains, such as Harter's Self-Perceptions Profile for Adults (3) and Marsh's Self-Description Questionnaire III (4). One dimension that has consistently emerged as being closely related to global ratings of self esteem is perception of the physical self (5). Specific interest in the detailed construction of the physical self began in the late 1980s with the development of the Physical Self-Perception Profile (PSPP) (6) and later by Marsh and colleagues with the Physical Self-Description Questionnaire (7). The PSPP, a multidimensional instrument, extended Harter's work and was designed to assess 5 critical components of physical self perceptions: body image, sport competence, physical strength, physical condition, and physical self-worth—the latter being seen as a superordinate construct in a hierarchical structure (8). Physical self-worth (PSW) has been shown to be directly associated with indicators of emotional adjustment, regardless of self esteem levels and socially desirable responding, having mental wellbeing properties in its own right (9). Previous research in a range of populations, not including those with rheumatic diseases, indicates that these subdomains contribute to the overall judgment of physical self worth, which in turn contributes, substantially in most populations (r = 0.6–0.7), to global self esteem (10).

Several theorists and researchers have also suggested that to fully understand the influence on self esteem of experience in different life domains, it is also useful to consider the importance that the individual places on those domains. For example, performance at work may not be as important and as critical to self esteem as a harmonious family life for some individuals, and vice versa. Harter (11, 12) suggested that to experience higher self esteem, individuals need to match their aspirations with their perceived level of competence, or success, in each domain. Where this does not occur, discrepancies between the desired and perceived self may result and this may contribute to lower self esteem. Harter has shown that individuals may use the self-serving process of “discounting” the importance of domains in which they do not perceive they are particularly competent and attach higher importance to areas in which they experience regular success. To address these tenets, the Perceived Importance Profile (PIP) (8) was devised for use in conjunction with the PSPP. The PIP allows the individual to rate the importance of success in different subdomains in terms of self- esteem enhancement. The concepts evaluated in the PSPP and PIP and their postulated relationship with self esteem is illustrated in Figure 1. The hierarchical structure depicted has been established in several samples using structural equation modeling (10).

Figure 1.

Hierarchical structure of physical self perceptions, importance, and self esteem.

The functioning of this importance filter between physical self perception and self esteem can be quantified through calculation of a discrepancy score. These assess the degree of mismatch between levels of competence and the importance to the individual of being competent in each element. Discrepancy scores have been negatively related to both physical worth and global self esteem. Perceived competence and importance scores have also been predictive of physical activity participation (13). Therefore, the PSPP and PIP can potentially provide a rich source of information about people's states of mind regarding their body and their physical capabilities. Clinical groups have already been studied (14, 15); however, this is restricted primarily to those with psychiatric disorders and there are no previous studies of people with rheumatoid arthritis (RA). Given the profound impact of RA on physical function and the potential value of physical self perceptions in both rehabilitation and clinical management, these concepts are worthy of investigation. This study therefore set out to provide 1) initial reliability and validity data on the PSPP and PIP; and 2) descriptive data on physical self perceptions and their relationship with measures of disease status, function, and negative affect in women with RA.


Location and ethics.

The study was conducted in the Physiotherapy Department of North Glasgow University Hospitals, Glasgow Royal Infirmary NHS Trust (GRI). The ethics committees of both Glasgow Caledonian University and GRI granted ethical clearance.


Fifty-two female patients with a confirmed diagnosis of classical or definite RA (diagnostic index criteria of the Centre for Rheumatic Diseases GRI) with consideration of defined selection criteria were recruited through the outpatient clinic. Subjects were recruited for a prospective study of an aerobic exercise intervention (16). Inclusion criteria were age between 30 and 70 years and Steinbrocker functional class I or II (17). Exclusion criteria were unstable RA medication regimens or medications affecting cardiac function, comorbidity contraindicating exercise, or cognitive impairment preventing informed consent.

To evaluate the PSPP and PIP scores in the context of disease status and function of women with RA, a range of physical and psychological measurements were taken.

Physical measures.

Disease status.

RA status was assessed with the core measures developed by the American College of Rheumatology and recommended in the European League Against Rheumatism (EULAR) Handbook of Standard Methods (18). These appear frequently in studies of RA and are erythrocyte sedimentation rate (ESR) and plasma C-reactive protein levels (CRP); Ritchie Articular Index (count and rating of joint tenderness in response to standardized palpation); total swollen joint count; patient assessment of disease activity (patients complete a visual analog scale [VAS] anchored at “extremely active” and “not active at all”); physician assessment of disease activity (5-point scale from “asymptomatic” to “very severe”); duration of early morning stiffness (patient self report); and joint pain (patients complete a VAS rating their pain).

Walking ability.

This was assessed as performance on a symptom-limited maximal 10-meter incremental shuttle walk test (SWT), developed by the Glenfield Hospital, Leicester and Loughborough University in 1992 (19, 20). The number of shuttles completed, heart rate, and oxygen uptake calculated by Cosmed K4b2 software were recorded (a Cosmed K4b2 portable respiratory gas analyzer [Cosmed, Rome, Italy] with associated software measuring oxygen uptake breath by breath was employed).

Functional capacity.

This was calculated as peak oxygen consumption rate (VO2peak) achieved during the SWT.

Aerobic power.

This was calculated as a linear extrapolation of the heart rate/oxygen consumption data from the SWT to age-predicted maximum (VO2max). The method has been described in full elsewhere (21, 22).

Psychological measures.

Physical self perceptions.

The PSPP (6, 8) is an instrument with 30 questions broken into 5 subscales that is designed to assess perceptions in 4 subdomains of the physical self: perceived sports competence (sport), perceived body attractiveness (body), perceived physical strength (strength), and perceived physical condition (condition). In addition, a separate subscale, PSW, assesses higher-order perception of physical self worth, an integration of the combined effect of the other 4 self perception subdomains. This instrument has been widely reported in the social, psychological, and sport psychology literature and has been translated into at least 11 languages (5, 6, 8, 14, 15).

Perceived importance.

In an attempt to incorporate personal weightings of importance of subdomain self perceptions to more global constructs, such as self esteem, the PIP was also administered (8). This allows a statement of importance, using a 2-question subscale for each of the 4 subdomains, to overall physical self worth and self esteem. This measure allowed a test of the impact of competence-importance discrepancies on self esteem and the presence of discounting of subdomains yielding low competence (11).

Depression and anxiety.

The Hospital Anxiety and Depression Scale (HADS) has been widely used in clinical settings to assess current level of depression and anxiety (23).


Participants were welcomed, issued an information sheet, and given a consent form to sign. Height and weight were recorded. All measures of the core EULAR set were taken (with exception of ESR and CRP) followed by the administration of the PSPP, PIP, and HADS. Participants then listened to the prerecorded instructions for the walking test and were asked if they understood and were willing to proceed with the test. The Cosmed K4b2 was calibrated to ambient temperature and humidity (as manufacturer's instructions), and then fitted to the participant. Participants were free to stop whenever they wished, but standardized instructions were given to keep going as long as possible. The Cosmed K4b2 was removed and the participants rested for a period if required. Blood samples were then drawn for analysis of ESR and CRP.

Statistical analyses.

Summary statistics were calculated for each of the subscales in the PSPP, PIP, and HADS. Because there is an absence of published data from a comparable RA group, the PSPP and PIP scores for the sample were compared with data from samples of American college-age and overweight or obese women (8) using one-way analysis of variance and Bonferroni post-hoc testing (where appropriate). The internal consistency of different items within each subdomain was quantified by Cronbach's alpha (24). Further checks of content validity were performed: principal components factor analysis (using oblique rotation) of the PSPP subdomain subscales and correlation and stepwise multiple regression of PSPP subdomain scores against the PSW score. Competence-importance discrepancy scores were calculated with the method advocated by Messer and Harter (3) and Fox (8). Correlation analysis was employed to test the strength and significance of any linear relationships between PSPP subscale scores and measures of disease status, aerobic fitness, walking ability, and HADS scores.


The final sample comprised 52 women with a mean ± SD age of 48.4 ± 10.4 years. Mean body mass index (BMI) was 27.3 ± 5.3, with 38.5% classed as overweight and 25.0% as obese. Mean disease duration was 10.3 ± 8.1 years. Although all participants completed the PSPP and PIP, 10 did not complete the core EULAR measures or the HADS (due to logistic difficulties), reducing the sample for related analyses to 42 participants.

Physical self perceptions and perceived importance.

High internal consistency and factor structure were demonstrated: Chronbach's alpha for PSPP subscales ranged from 0.73 to 0.81. This supports the assertion that the previously demonstrated reliability of the PSPP was reproduced in a sample of women with RA. Principal component factor analysis confirmed the 4-factor structure of the subdomain subscales with each item loading on its intended factor. Even with this relatively small sample size, the instrument seems to produce clearly defined and internally consistent subscales with this clinical population.

Summary statistics for the PSPP subdomains, PIP, and HADS are presented in Table 1. The PSPP and PIP scores for the RA subjects were compared with data from samples of American college-age women (n = 431) and overweight or obese women (BMI >28, n = 422) with a mean BMI of 36 and mean age of 41 years (8). For all scores, there were significant differences between the 3 groups (P < 0.001) and the significant differences between the RA sample and the comparator groups are detailed in Table 1.

Table 1. PSPP and PIP scores for women with RA, US college-age women, and obese women*
 RAUS collegeObese
  • *

    Data presented as mean ± SD. PSPP = Physical Self-Perception Profile; PIP = Perceived Importance Profile; RA = rheumatoid arthritis; PSW = physical self-worth.

  • RA significantly different from college-age women at P < 0.05.

  • RA significantly different from obese women at P < 0.05.

 Sport9.3 ± 2.614.2 ± 4.311.3 ± 3.7
 Condition10.6 ± 3.014.4 ± 4.110.3 ± 3.1
 Body10.6 ± 2.813.3 ± 4.38.6 ± 2.6
 Strength9.8 ± 2.914.7 ± 3.713.5 ± 3.9
 PSW10.9 ± 2.714.2 ± 3.810.6 ± 3.1
 Sport3.5 ± 1.44.6 ± 1.64.0 ± 1.4
 Condition4.6 ± 1.35.4 ± 1.55.0 ± 1.4
 Body4.4 ± 1.15.8 ± 1.35.8 ± 1.3
 Strength4.3 ± 1.35.1 ± 1.34.9 ± 1.4

It can be seen that the women with RA (as did the obese women, with the exception of the body and strength PIP scores) scored significantly lower in all PSPP and PIP subscales than college-age women. The RA subjects' perceived sport competence and strength scores were also significantly lower than those for obese women and close to the minimum possible; whereas physical condition and overall PSW were similar to those for obese women, and body attractiveness was significantly higher that that for obese women. The importance of attractive body and physical strength was significantly lower for RA patients than for obese women, whereas the importance of sport and condition was similar.

The relationships between the PSPP subscales and PSW were evaluated to assess their contribution to this higher-order construct. The resultant correlation coefficients were all significant: condition r = 0.55, body r = 0.63, and strength r = 0.58 at P < 0.001; and sport r = 0.40 at P < 0.01. Stepwise multiple regression revealed that with the exception of sport, all subscales significantly and independently contributed to the R2 value of 59.1% of explained variance in PSW (condition and body P < 0.001; strength P < 0.05; and sport P = 0.642). This is a slightly lower explained variance than has been observed with other populations, suggesting that other factors that were not measured here may also contribute to physical self worth in patients with RA. The low contribution of sport was further supported by the low PIP sport score; the combination of low competence and importance scores suggest that perceptions in this domain are unlikely to influence higher-order constructs of self worth.

Discrepancy scores represent the mismatch between an individual's aspirations and their perceived competence. In this sample, the correlation between total discrepancy in the 4 PSPP subdomains and scores on the PSW subscale was r = 0.48 (P < 0.01). The scatter plot featured in Figure 2 demonstrates the spread of scores and substantiates that a portion of PSW may be determined by RA patients' feelings that they are unable to meet their perceptual needs.

Figure 2.

Scatter plot of physical self-worth score versus discrepancy/adequacy score, with line of best fit (least squares method).

The correlation coefficients between the PSPP subscales and the other measures are presented in Table 2; only those relationships that were significant are presented. Physical self perceptions were significantly related to only 2 of the measures of disease status and not to any of the measures of fitness; indicated by the negative weak to moderate correlations between the swollen joint count and levels of perceived strength, condition, and sport competence and between sport competence and ESR. The only other significant correlations lay between the depression aspect of negative affect (the depression subscale of the HADS) and the condition and PSW subscales.

Table 2. Correlation coefficients for significant linear relationships between the PSPP subscales and other measures for women with RA, n = 42*
  • *

    PSPP = Physical Self-Perception Profile; RA = rheumatoid arthritis; SWJ = swollen joint count; ESR = erythrocyte sedimentation rate (mm/hour); HDep = Hospital Anxiety and Depression Scale depression score; PSW = physical self-worth.

  • Significant at P < 0.05.

  • Significant at P < 0.01.

PSW  −0.307
Condition−0.306 −0.466


The aims of this study were twofold: first to establish the PSPP and PIP as internally reliable and valid instruments with an RA sample and second to analyze the levels of physical self perceptions and their relationships with measures of disease, function, and negative affect in women with RA.

Administration of the PSPP and PIP in this sample was found to be straightforward with no participants reporting difficulties in completing or understanding the questions. The reliability coefficients, factor structure, and relationships among PSPP and PIP subscales observed here suggest that the instrumentation functions equally well with RA subjects as it does in other populations (5, 8, 14, 15). Similarly, the pattern of mean subscale scores was logical when compared with that seen in other populations and provides further evidence of construct validity. The content of the 4 subdomains appeared salient to RA subjects in that all 4 correlated significantly with PSW and, as a set of constructs, explained almost 60% of variance in PSW. This is a slightly lower explained variance than has been observed with other populations, suggesting that other factors not measured here may also contribute to physical self worth in patients with RA; additional study would be necessary to confirm this.

On the basis of the data presented here, the physical self perceptions of women with RA are extremely low when compared with healthy college-age women and also clinically obese women—the latter being a similar age to the RA patients studied (8). The RA subjects scored significantly lower on all PSPP subscales than the college-age women and significantly lower in the sport and strength domains than the obese group. The RA subjects' extremely low mean score on perceived strength was particularly noticeable. In the context of the consistently low scores observed in all aspects of physical self perception, this may well reflect a lack of confidence caused by a perceived loss in functional aspects of fitness, in addition to any actual loss of strength experienced. This lack of confidence may also be reflected in the similarly low sport competence score, although there is some evidence of discounting the importance of sport competence, reducing its impact on self worth (8, 12). Overall PSW, the global view of the physical self, was also very low and close to that of the obese group. This may again reflect an overall lack of confidence in the physical aspects of the self as a result of disease. This theory is supported by the significant negative correlations between the HADS depression score and both PSW and condition. Furthermore, low PSPP scores have previously been seen to be related to low self esteem and other aspects of mental wellbeing in a range of populations (9, 14). Interestingly, in the body domain, RA subjects did score significantly higher than the obese women. This, however, reflects the profound depth of the obese subjects' negative responses in this particular domain rather than any positive finding for the RA subjects.

Again, on the basis of the present evidence, RA patients attach lower importance to aspects of the physical domain than is typically seen in other populations; this is evident in comparison with college women and to a lesser extent with obese women. The RA subjects' ratings of importance were significantly lower than the college-age women in all domains and significantly lower than the obese women in the body and strength domains. Although this is preliminary data from a relatively small sample, it may be that this is an indicator that women with RA are discounting the importance of physical competence and appearance (8, 12). To substantiate this finding, additional research with measures of self perceptions in other domains is recommended. The importance scores were not sufficiently low, however, to avoid discrepancies between the desired and perceived self in some domains for some of these women. These discrepancies were negatively associated with PSW. Although there was some evidence of discounting low competence, this only occurred in the sports subdomain. Low perceived strength and condition, in particular, combined with (relatively) high importance ratings, appeared to contribute to low PSW. It may be the case that clinical populations who suffer from physical difficulties, such as those with RA, cannot discount their need and desire for higher competence in such aspects as physical strength and condition. This may well reflect their desire to reach normal levels and become healthy again. Given that PSW has been associated with indices of emotional and mental wellbeing, independent of self esteem, and socially desirable responding (9), this may be an important therapeutic finding. Strategies for improving confidence and competence may help increase physical self worth and subsequently mental wellbeing and aid with coping with RA.

In the analysis of the relationships between how women with RA perceive themselves in physical terms and measures of their disease, function, and negative affect, it was seen that 3 of the PSPP subscales (strength, sport, and condition) were significantly related to the clinical measure of swollen joint count. This may indicate that patients utilize this variable as a salient statement of their condition and that it is significantly related to their physical self perceptions. Nevertheless it was surprising that swollen joint count, a very outward physical manifestation of the disease, did not correlate with body score and the results show stronger relationships between swollen joint count and the self perception domains related to functioning rather than appearance. Perhaps this arose from either the relatively low importance the subjects allocated to body in the PIP results (the RA subjects rated this factor significantly lower in importance than did either of the 2 comparator groups); or the fact that not all the joints that are scored in the EULAR swollen joint count are visible to others. Measures of depression were also significantly related to 2 subscales: PSW and condition. It therefore appears that physical self perceptions are more closely related to certain direct markers of the disease (swollen joint count and ESR) and negative affect (HADS depression score) rather than measures of functional fitness. This supports previous literature showing that positive changes in physical self perceptions, which often accompany exercise (25), are not usually related to measures of functional fitness. Although no data was collected here to allow a direct comparison of the fitness of RA and healthy subjects, several previous studies have reported both similar low values for the measures of fitness reported here and markedly reduced fitness in RA subjects over both their age-matched healthy and osteoarthritic peers (26–29). Lack of fitness and exercise and the accompanying absence of pride and confidence in the body may have its impact on reduced mood experienced in depression in RA subjects.

The conclusions are made with the caveat that the present sample is relatively small and this is the first published study of these instruments applied in an RA sample. The evidence presented here indicates, however, that the PSPP and PIP are reliable and valid instruments that are sensitive to important constructs in the mental health of women with RA. The PSPP appears to measure distinct mental properties not represented in other measures previously employed in evaluating RA patients. This study also provides evidence of reduced physical self perceptions in RA subjects. Physical self perceptions have been shown to be related to adherence and persistence in physical activity as well as aspects of mental wellbeing; evidence that they are important in both the rehabilitative process and clinical management and therefore should not be overlooked in the care of patients with RA.