• Systemic sclerosis;
  • Scleroderma;
  • Health values;
  • Health utilities;
  • Health status


  1. Top of page
  2. Abstract
  8. Acknowledgements


To assess health values in subjects with systemic sclerosis (SSc) and determine variability explained by demographics, clinical factors, health status, and disease severity.


We interviewed 107 individuals with SSc who attended national and local Scleroderma Foundation meetings in 2005. Health status was measured using the Short Form 36 (SF-36) Physical Component Summary (PCS; range 0–100) and Mental Component Summary (MCS; range 0–100), the Center for Epidemiologic Studies Depression Scale (CES-D; range 0–60), and the Health Assessment Questionnaire (HAQ) disability index (DI; range 0–3). Disease severity was assessed using a visual analog scale (VAS; range 0–150). Health value measures included the 0–100 health rating scale (RS), standard gamble (SG; range 0.0–1.0), and time trade-off (TTO; range 0.0–1.0). We performed univariate analyses to compare scores between participants with limited cutaneous SSc (lcSSc) and diffuse cutaneous SSc (dcSSc), and multivariable analyses for 3 outcome measures: RS, SG, and TTO, controlling for demographics, type of SSc, health status, and disease severity.


Of the 107 participants, 48 had dcSSc and 59 had lcSSc. Ninety-seven were women and 83 were white. The median scores for the PCS, MCS, and HAQ DI were 36.9, 45.5, and 0.9, respectively. Fifty-five subjects had significant depressive symptoms (CES-D score ≥16). The median RS, SG, and TTO scores were 62, 0.83 (indicating a willingness to accept up to a 17% risk of immediate death in exchange for perfect health), and 0.88 (indicating a willingness to give up a median of 12% of life expectancy in exchange for perfect health), respectively. Subjects with dcSSc had lower RS scores but higher SG scores (corresponding to a willingness to accept only a smaller risk of death) than subjects with lcSSc. TTO scores were similar in the 2 groups. Health values were variably related to factors such as demographics, VAS score, disease classification, and SF-36 PCS and MCS scores (R2 = 0.22, 0.23, and 0.66 for the SG, TTO, and RS models, respectively).


Individuals with dcSSc have lower health ratings but higher SG health values than individuals with lcSSc. These findings have implications for decision analysis and cost-effectiveness analysis.


  1. Top of page
  2. Abstract
  8. Acknowledgements

Chronic diseases often have a relapsing and remitting course with substantial impact on function and quality of life. Scleroderma (systemic sclerosis [SSc]), a multisystem autoimmune disorder with no effective treatment or cure, is an example of a disease in which individuals must cope with pain, disfigurement, disability, and feelings of helplessness, each of which can affect quality of life and outlook on the present and the future (1, 2).

The ultimate goal of health care is to improve, restore, or preserve functioning and well-being, i.e., health-related quality of life (HRQOL) (3). There are 2 standard ap proaches to assessing HRQOL: 1) the health status approach, which describes functioning and the impact of illness on specific domains of health (e.g., physical functioning and mental health [4]), and 2) the value/preference/utility approach, which assesses the value or desirability of health states and summarizes HRQOL in a single number. Health values serve as quality-adjustment factors for calculating quality-adjusted life years (QALYs) in decision analysis and cost-effectiveness analysis.

If the goal of health care is to maximize health and longevity, then the goal of cost-effectiveness analysis is to maximize QALYs under constrained budgets. The philosophical underpinning of economic evaluation in health care is utilitarianism, or variations thereof (5). Utility, under classical utilitarianism, is considered a measure of “good stuff” (6). In comparison, under von Neumann and Morgenstern utility theory, utility as applied to health care is a measure of strength of preference for health states under conditions of uncertainty (7).

A variety of methods exist for rating and valuing health states. Measures such as the category rating scale have origins in experimental psychology. The simplest is the rating scale (RS), which asks the respondent to rate health states along a continuum from 0 (often representing “dead”) to 100 (often representing “perfect health”). Unlike the RS, 2 other methods, the standard gamble (SG) and time trade-off (TTO), value health as opposed to an external metric such as time (TTO) or risk (SG), exemplifying notions of sacrifice and exchange (8, 9). The SG and TTO are generally considered to be appropriate for use in calculating QALYs, whereas the RS is not. (For simplicity, we use the term health value to discuss the RS, TTO, and SG.)

Several recent studies have measured health status in patients with SSc (10, 11), but none have assessed health values. Our study had 2 objectives: to assess health values in individuals with SSc, and to determine the variability explained in health values by demographics, clinical factors, health status, and disease severity.


  1. Top of page
  2. Abstract
  8. Acknowledgements

Study subjects.

We recruited 110 participants age ≥18 years with SSc, as defined by the American College of Rheumatology (formerly the American Rheumatism Association) classification criteria (12), from the 2005 national Scleroderma Foundation annual meeting in Boston, Massachusetts (n = 60) and from 3 local Scleroderma Foundation meetings in Cincinnati and Columbus, Ohio and Los Angeles, California (n = 50). Interested participants were recruited from attendees of educational seminars at the 4 chapter meetings. For the national meeting, we advertised the study on the Internet ( For the 3 local chapter meetings, we sent flyers to the registered attendees describing the study. Participants were classified as having either limited cutaneous SSc (lcSSc) or diffuse cutaneous SSc (dcSSc) based on the distribution of skin thickness on skin examination performed by one of the authors (DK, MA, or DEF); a formal modified Rodnan skin score was not assessed (13). As per accepted criteria, lcSSc is characterized by skin thickening distal to the elbows and knees and proximal to the clavicles (including the face), whereas dcSSc is characterized by skin thickening proximal and distal to the elbows and knees, including the trunk and face (14). The University of Cincinnati Institutional Review Board approved the study protocol, and all participants provided informed consent before study participation.


We administered questionnaires in structured interview sessions. Participants first answered questions about their health status to facilitate reflecting upon their health prior to valuing health. Health status was operationalized as physical functioning and emotional well-being and was assessed using 3 self-administered instruments: the Short Form 36 (SF-36) (4), the Health Assessment Questionnaire (HAQ) disability index (DI) (15), and the Center for Epidemiologic Studies Depression Scale (CES-D) (16).

The SF-36 is a generic health status measure consisting of 36 items assessing 8 domains (4, 17). Each of the SF-36 subscales is scored from 0 to 100, with a higher score representing better health. The 8 SF-36 subscales can be summarized into a Physical Component Summary (PCS) and a Mental Component Summary (MCS) score. The summary scores are normed to the US general population, where the mean ± SD score is 50 ± 10. We used version 2 of the SF-36 and a standard (4-week) recall period (18).

The HAQ DI (19) is a self-administered 20-question arthritis-specific instrument that assesses an individual's level of upper and lower extremity functioning. The HAQ DI has been used extensively in persons with SSc (11, 20, 21). The overall HAQ DI score is determined by summing the highest item score in each of the 8 domains and dividing the sum by 8, yielding a score ranging from 0 (no disability) to 3 (severe disability). In the original HAQ DI, an additional grade of difficulty was added for persons using assistive/adaptive devices (such as a cane or walker); however, as in more recent studies (21, 22), we did not modify subjects' responses for use of assistive/adaptive devices.

The CES-D is a 20-item instrument that captures the frequency of a particular mood or symptom in the prior week on a 4-point scale ranging from 0 (none of the time) to 3 (most of the time). After reversing the positive mood items, scores on the items are summed such that they range from 0 (best) to 60 (worst); scores ≥16 represent significant depressive symptoms (16, 23, 24).

Disease severity was assessed using a 150-mm visual analog scale (VAS). The question read: “In the past week, considering pain, discomfort, limitations in your daily life, and other changes in your body and life, how severe would you rate your disease today?” (15). The scales ranged from “no disease” to “very severe limitation” (15), with higher scores representing more severe disease.

Health values were elicited using UMaker, a computer-assisted utility assessment software package (25). First, participants rated their current health on an RS, which was presented as a “feeling thermometer” with scores ranging from 0 (dead) to 100 (perfect health). Participants were given the following directions concerning current health: “Current health is defined by you: how you yourself feel about how you are doing, not what others may tell you about your health (doctors, nurses, friends).” Next, participants completed a TTO exercise, which assessed the individual's willingness to live a shorter but healthier life. The TTO was represented graphically as a choice between 2 horizontal bars, one representing the subject's life expectancy in current health (followed by death) and the other representing a given number of years (less than or equal to the life expectancy) in perfect health followed by death (25). Based on the age of the subject, UMaker utilized the life expectancy reported in US life tables (26), rounding the life expectancy to the nearest 5 years. For example, for a subject who was 50 years old, the life expectancy table estimated a life expectancy of 30 years; to start, therefore, a 50-year-old participant was asked to choose between living 30 years in their current health followed by death versus living 30 years in perfect health followed by death. If the participant preferred 30 years in perfect health, he or she was offered a choice between 30 more years in their current health or 0 years in perfect health (i.e., immediate death). If 30 years in current health were preferred, the number of years in perfect health was varied in a bisection fashion until the subject did not have a clear preference between living 30 more years in current health or living the given amount of time in perfect health (27). The TTO score was calculated by dividing the number of years of perfect health at the indifference point by the subject's life expectancy.

The final utility task was the SG, which assessed the individual's willingness to risk a bad outcome (death) in exchange for a chance of perfect health. Participants were shown 2 circles: one was labeled “current health with scleroderma” and remained the same on all of the screens; the second circle represented perfect health. Participants were offered a choice between living the remainder of their life in their current state of health versus taking a gamble in which the 2 outcomes were perfect health for the remainder of life or immediate death (28). Initially, the second circle was displayed as a pie chart with a 100% probability of perfect health. Assuming the subject preferred perfect health in that scenario, the probabilities of perfect health and death in the second circle were then varied systematically using bisection until the subject had no preference between the certainty of life in his or her current state of health or the gamble. The SG score was calculated by the following formula: 1 minus the maximum acceptable probability of death. In all 3 health value tasks, the definition of perfect health was not provided; it was self defined by each participant.

Statistical analysis.

Descriptive statistics for continuous variables are presented as means and SDs for normally distributed variables and medians and 25th–75th percentiles for non-normally distributed variables. Normality of health value measures was assessed using the Shapiro-Wilk test; RS scores were normally distributed whereas TTO and SG scores were not (negatively skewed). Categorical variables are presented as frequencies and proportions in a contingency table format. Unadjusted comparisons for categorical outcomes were made using chi-square and Fisher's exact tests. Unadjusted and adjusted (for disease type) strengths of association between health values, health status, and disease severity scores were assessed using Spearman's nonparametric correlation coefficients and were interpreted as proposed by Franzblau (29): 0.0–0.20 indicating no correlation, 0.21–0.40 indicating a low degree of correlation, 0.41–0.60 indicating a moderate degree of correlation, 0.61–0.80 indicating a marked degree of correlation, and 0.81–1.00 indicating a high degree of correlation.

We then performed separate multivariable analyses for each of the 3 dependent variables: RS, TTO, and SG. Candidate variables for multiple regression models were selected using bivariate regression at a threshold of P less than 0.15 for entry into the final multivariable model. Independent variables included age (reference group: men), sex, income (reference group: <$25,000/year), ethnicity, insurance status, and year of diagnosis; type of SSc (reference group: dcSSc); disease severity (VAS scores); and health status (SF-36 PCS and MCS, HAQ DI, and CES-D) scores. In addition, for the TTO, we also assessed a model with an interaction term between age and type of disease because participants with lcSSc tended to be older. Preliminary analysis showed marked correlations between the HAQ DI and the SF-36 PCS (ρ = −0.62) and between the CES-D and the SF-36 MCS (ρ = −0.71). The multivariable test statistics were similar when entering either the CES-D and HAQ DI or the SF-36 PCS and MCS. Therefore, we present multivariable models using the SF-36 PCS and MCS instead of the CES-D and HAQ DI. We also performed regression diagnostics to assess the validity of the models. Because the residuals for the TTO and SG were not normally distributed, we also estimated linear regression models after log transforming the dependent variables, but the residuals for the TTO and SG remained non-normally distributed. To improve interpretability, we present the raw TTO and SG instead of the log-transformed TTO and SG. Although tests of statistical significance may be slightly inaccurate, we relied on the relatively large sample size to ensure that the test statistics were approximately normal. In the final multivariable models, P values less than 0.05 were considered statistically significant. All analyses were performed using STATA software, version 8.2 (StataCorp, College Station, TX).


  1. Top of page
  2. Abstract
  8. Acknowledgements

Subjects' characteristics.

Of the 110 subjects, all but 3 (2.7%) completed the health values exercises; the remaining 107 participants formed our study sample (Table 1). Most of the participants were women and most were white. Almost all subjects (98.8%) graduated from high school and 58 (56.9%) had incomes exceeding $50,000 per year. Patients with lcSSc tended to be older than patients with dcSSc (P < 0.05).

Table 1. Demographics of study participants, overall and by disease type*
VariablesTotal sample (n = 107)Diffuse SSc (n = 48)Limited SSc (n = 59)
  • *

    Values are the number (percentage) unless otherwise indicated. SSc = systemic sclerosis.

  • P < 0.05 by Fisher's exact test for 3 categories.

  • Data not available for 5 patients.

  • §

    Data not available for 1 patient.

Age, years   
 18–346 (5.6)1 (2.0)5 (8.6)
 35–5437 (34.6)24 (49.0)13 (22.4)
 ≥5564 (61.8)24 (49.0)40 (70.0)
Female sex97 (90.6)45 (91.8)52 (86.6)
Disease duration, median (25th–75th percentile) years7.0 (4.0–12.0)8.0 (4.0–12.0)6.5 (4.0–13.0)
 White83 (77.6)40 (81.6)43 (74.1)
 African American8 (7.5)5 (10.2)5 (5.2)
 Hispanic9 (8.4)4 (8.2)5 (8.6)
 Other4 (6.5)0 (0.0)7 (12.1)
Annual income   
 <$12,0007 (6.9)2 (4.3)5 (9.1)
 $12,000–25,00015 (14.7)7 (14.9)8 (14.6)
 >$25,000–50,00019 (18.6)7 (14.9)12 (21.8)
 >$50,000–75,00025 (24.5)14 (29.8)11 (20.0)
 >$75,00033 (32.4)15 (31.9)18 (32.7)
 Unknown3 (2.9)2 (4.3)1 (1.8)
 <High school2 (2.0)1 (2.0)1 (1.8)
 High school graduate53 (50.0)27 (55.1)26 (45.6)
 College graduate31 (29.2)11 (22.5)20 (35.1)
 Graduate degree20 (18.8)10 (20.4)10 (17.5)

Health status.

The median (25th–75th percentile) SF-36 PCS and MCS scores were 36.9 (29.4–45.1) and 45.5 (34.5–53.4), respectively, which are 1.3 and 0.5 SDs below unadjusted US general population mean scores, respectively (Table 2). The median (25th–75th percentile) HAQ DI score was 0.9 (0.5–1.4), indicating mild disability. A total of 55 (54.5%) subjects had CES-D scores ≥16, indicating significant depressive symptoms (Table 2). The median (25th–75th percentile) disease severity score as assessed by the VAS was 50 (29–79); disease severity tended to be worse in participants with dcSSc than in those with lcSSc (58.0 versus 47.0; P = 0.2).

Table 2. Health status, disease severity, and health value scores*
VariablesTotal sample (n = 107)Diffuse SSc (n = 48)Limited SSc (n = 59)
  • *

    SSc = systemic sclerosis; SF-36 = Short Form 36; PCS = Physical Component Summary; MCS = Mental Component Summary; HAQ = Health Assessment Questionnaire; DI = Disability Index; VAS = visual analog scale; CES-D = Center for Epidemiologic Studies Depression Scale. For the SF-36 and health values, a higher score indicates better health and greater desirability of one's current health state, respectively. For the HAQ DI, VAS severity, and CES-D, a higher score indicates greater functional disability, more severe disease, and more significant depressive symptoms, respectively.

  • P < 0.05 for comparison of limited versus diffuse SSc.

Health status   
 SF-36 PCS (possible range 0–100)   
  Median (25th–75th percentile)36.9 (29.4–45.1)38.1 (28.5–45.0)36.8 (30.0–46.0)
 SF-36 MCS (possible range 0–100)   
  Median (25th–75th percentile)45.5 (34.5–53.4)44.5 (33.5–53.3)46.3 (35.8–53.3)
 HAQ DI (possible range 0–3)   
  Median (25th–75th percentile)0.9 (0.5–1.4)1.0 (0.6–1.5)0.88 (0.4–1.4)
 VAS severity index (possible range 0–150)   
  Median (25th–75th percentile)50.0 (29.0–79.0)58.0 (37.0–85.0)47.0 (15.0–79.0)
 CES-D score ≥16   
  Number (%)55 (54.5)24 (51.1)31 (57.4)
Health values   
 Rating scale (possible range 0–100)   
  Median (25th–75th percentile)62 (50–80)60 (50–70)70 (59–85)
  Mean ± SD64.3 ± 19.359.9 ± 18.768.0 ± 19.1
 Time trade-off (possible range 0–1.0)   
  Median (25th–75th percentile)0.88 (0.63–0.94)0.88 (0.63–0.97)0.88 (0.60–0.94)
  Mean ± SD0.76 ± 0.250.76 ± 0.280.77 ± 0.23
 Standard gamble (possible range 0–1.0)   
  Median (25th–75th percentile)0.83 (0.63–0.94)0.88 (0.69–0.94)0.70 (0.52–0.94)
  Mean ± SD0.74 ± 0.260.79 ± 0.240.69 ± 0.26

Health values.

The mean ± SD RS score was 64.3 ± 19.3; in other words, on average, participants rated their current health at 64.3% of perfect heath (Table 2; Figure 1). On average, participants with lcSSc rated their health better (mean ± SD 68.0 ± 19.1) than did those with dcSSc (59.9 ± 18.7; P = 0.03).

thumbnail image

Figure 1. Histograms of health value scores. A, Rating scale. B, Time trade-off. C, Standard gamble.

Download figure to PowerPoint

The median (25th–75th percentile) TTO score was 0.88 (0.63–0.94), indicating a willingness to give up 12% ([1 − 0.88] × 100%) of one's life expectancy in exchange for perfect health. For example, a prototypic subject with a 15-year life expectancy was willing to live a median of 13.2 years, or trade a median of 1.8 years of life ([1.0 − 0.88] × 15 years = 1.8 years), in perfect health in exchange for living 15 years with SSc. There was no statistical difference in TTO values between subjects with lcSSc and those with dcSSc.

The median (25th–75th percentile) SG score was 0.83 (0.63–0.94). Thus, participants were willing to accept up to a median 17% risk of death ([1.0 − 0.83] × 100% = 17%) for a chance at perfect health. Participants with dcSSc were willing to take a median 12% risk of death (median score 0.88) compared with a 30% risk for participants with lcSSc (median score 0.70; P = 0.03).

In bivariate analyses, RS scores correlated moderately with health status (ρ = 0.46–0.54) and markedly with VAS disease severity (ρ = −0.69) (Table 3). TTO and SG values correlated less well with health status and VAS severity (ρ = 0.16–0.34). When the groups were stratified by extent of SSc (lcSSc versus dcSSc), the RS and SG had stronger correlations with health status in the lcSSc group whereas the TTO had a stronger correlation with health status in the dcSSc group (Table 3).

Table 3. Univariate relationships between health values, health status, and disease severity measures, overall and by disease type*
 GenericDisease specificDisease severity
  • *

    Values are Spearman's rho. See Table 2 for definitions.

Rating scale0.460.49−0.54−0.50−0.69
 Diffuse SSc0.390.39−0.48−0.48−0.75
 Limited SSc0.510.57−0.63−0.510.65
Time trade-off0.250.16−0.19−0.21−0.32
 Diffuse SSc0.240.26−0.31−0.290.45
 Limited SSc0.250.06−0.08−0.17−0.24
Standard gamble0.330.23−0.34−0.32−0.24
 Diffuse SSc0.180.12−0.16−0.15−0.07
 Limited SSc0.440.36−0.45−0.52−0.42

Multivariable correlates of health values.

In multivariable linear regression models, the RS was positively associated with younger age, SF-36 PCS and MCS scores, having lcSSc, less disease severity, and higher incomes (model R2 = 0.66) (Table 4). Higher TTO scores (a higher score implies willingness to trade less time) were positively associated with less disease severity and with higher income (R2 = 0.23). The interaction between age and type of SSc was not statistically significant (P = 0.3). Higher SG scores (a higher score implies less willingness to gamble for perfect health) were positively associated with female sex and with having dcSSc (R2 = 0.22).

Table 4. Multivariable determinants of health values*
VariablesRating scaleTime trade-offStandard gamble
  • *

    See Table 2 for definitions.

  • Reference group: male sex.

  • Reference group: participants with diffuse SSc.

  • §

    Reference group: annual income <$25,000.

Female sex−2.60.5−9.30.2515.30.05
Limited SSc6.30.020.350.95−8.90.07
SF-36 PCS0.340.06−0.220.540.490.13
SF-36 MCS0.52< 0.0010.050.840.370.12
VAS severity−0.24< 0.001−0.340.002−0.050.61
Income ($25,000–75,000)§
Income (>$75,000)§9.70.00912.
Model R20.66 0.23 0.22 
F value< 0.001 0.005 0.007 


  1. Top of page
  2. Abstract
  8. Acknowledgements

SSc is a chronic, often disabling disease. Unfortunately, current treatments do little to alter the course of SSc and the disease is associated with pain, disability, and a feeling of helplessness. Given the importance of HRQOL in SSc, it is surprising that relative to other rheumatic diseases, HRQOL in SSc has received very little attention (30).

Of the few HRQOL studies performed in persons with SSc to date, all have assessed health status (10, 11, 31) and none have assessed health values. Our results confirm previous findings that individuals with SSc have poor physical functioning (mean PCS score 36.9) and low-normal mental health (mean MCS score 45.5) (10, 11).

With regards to health values, participants with dcSSc rated their health, on average, worse than did participants with lcSSc. These results are intuitive in that dcSSc is associated with greater disease burden, disfigurement, functional disability, and mortality, all of which can lead to feeling helpless (14). With regards to the TTO, however, subjects with lcSSc or dcSSc were ready to give up similar amounts of time, a median of 12% of their life expectancy, in exchange for perfect health. Although not formally assessed, when participants with SSc were asked why they would not trade more time for better health, they often stated that they wanted to see their children and grandchildren grow up.

One of the surprise findings was that when it came to taking a gamble in hopes of achieving perfect health, participants with lcSSc were willing to accept a median 30% risk of death, which was substantially greater than the 12% median risk among participants with dcSSc. One potential explanation is selection bias (i.e., the sickest subjects with dcSSc may not have been interviewed because they were unable to attend the meetings). Alternatively, participants with dcSSc might have adapted better to their chronic disease, or subjects with lcSSc may have been anxious about their disease potentially progressing to dcSSc. In any case, our TTO and SG findings have implications for decision analysis and cost-effectiveness analysis of treatments of SSc in that holding life expectancies constant, subjects with dcSSc would be credited with the same number of QALYs (based on the TTO results) or more QALYs (based on the SG results) than subjects with lcSSc. In other words, if the life expectancy is 10 years, then living with dcSSc for 10 years equates to 8.8 QALYs, whereas living with lcSSc for 10 years equates to only 7.0 QALYs.

As previously described in the literature, TTO and SG scores were higher than RS scores (8, 32). One possible explanation is that the concepts of probability and trade-off are poorly understood among many subjects. In addition, individuals are generally risk averse, which tends to increase SG scores.

In addition to ascertaining health values, we sought to assess whether the value or desirability of a health state was related to the health state per se. Although one would expect that the healthier a person is, as assessed by a health status measure, the less time they would trade or the smaller the risk they would take (as determined by health utility measures) in exchange for perfect health, studies have shown that the association between individuals' current health status and their value for that health state is at best modest (33, 34). Of the 3 value measures, the RS had the strongest correlation with health status and VAS disease severity; demographics, type of SSc, and health status explained 66% of the variability in RS scores. Overall, the RS is a good indicator of HRQOL, but in the strictest sense it is not a utility measure because it does not involve trade-offs against external metrics such as time or risk. For the SG and TTO, demographics, health status, type of SSc, and VAS severity only accounted for 22–23% of variability in health values. This finding suggests that there are other as yet unmeasured determinants of individuals' values for their health, potentially including social and psychological factors (33, 34).

Our study had some limitations. First, the study was performed at Scleroderma Foundation meetings; participants might not reflect the true mix of individuals with SSc because sick and poor persons might not have been able to travel to the meetings. Therefore, individuals who did not participate may have had different (perhaps lower) health values than those reported here. Second, most participants were white and most were well educated. Some studies have shown that sociodemographic status and ethnicity may have an impact on health value scores (35). Third, we did not adjust for multiple statistical comparisons between participants with lcSSc and dcSSc; therefore, we cannot rule out a Type I statistical error.

These limitations notwithstanding, we conclude that persons with dcSSc have lower health ratings, but similar TTO and higher SG health values as compared with persons with lcSSc. Demographics, health status, type of SSc, and disease severity explained only a small proportion of variance in health value scores. Further research is needed to address factors that influence health value scores.


  1. Top of page
  2. Abstract
  8. Acknowledgements

Dr. Khanna 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 design. Drs. Khanna, Furst, and Tsevat.

Acquisition of data. Drs. Khanna, Ahmed, and Furst, and Ms Ginsburg.

Analysis and interpretation of data. Drs. Khanna, Furst, Park, and Tsevat.

Manuscript preparation. Drs. Khanna, Furst, and Tsevat.

Statistical analysis. Drs. Khanna, Park, and Hornung.


  1. Top of page
  2. Abstract
  8. Acknowledgements

We thank Carolyn Weller, RN, and other members of the National Scleroderma Foundation; Amelia Yaussy, Kathy Haas, and Elaine Rosen, and other members of the Columbus, Cincinnati, and Southern California Foundation Chapters; and participants who volunteered for this study.


  1. Top of page
  2. Abstract
  8. Acknowledgements
  • 1
    Haythornthwaite JA, Heinberg LJ, McGuire L. Psychologic factors in scleroderma. Rheum Dis Clin North Am 2003; 29: 42739.
  • 2
    Malcarne VL. Psychological adjustment in systemic sclerosis. In: ClementsPJ, FurstDE, editors. Systemic sclerosis. Philadelphia: Lippincott Williams & Wilkins; 2004. p. 33150.
  • 3
    Osoba D, King M. Meaningful differences. In: FayersP, HaysRD, editors. Assessing quality of life in clinical trials. New York: Oxford; 2005. p. 24457.
  • 4
    Ware JE Jr, Sherbourne CD. The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection. Med Care 1992; 30: 47383.
  • 5
    Garber AM, Weinstein MC, Torrance GW, Kamlet MS. Theoretical foundations of cost-effectiveness analysis. In: GoldMR, SiegelJE, RussellLB, WeinsteinMC, editors. Cost-effectiveness in health and medicine. New York: Oxford University Press; 1996. p. 2553.
  • 6
    Conditional expected utility. In: KrantzD, LuceR, SuppesP, TverskyA, editors. Foundations of measurement. New York: Academic Press; 1971. p. 369421.
  • 7
    Von Neumann J, Morgenstern O. Theory of games and economic behavior. New York: Wiley; 1953.
  • 8
    Kind P. Values and valuation in the measurement of HRQOL. In: FayersP, HaysRD, editors. Assessing quality of life in clinical trials. New York: Oxford; 2005. p. 391404.
  • 9
    Torrance GW. Measurement of health state utilities for economic appraisal. J Health Econ 1986; 5: 130.
  • 10
    Khanna D, Clements PJ, Furst DE, Chon Y, Elashoff R, Roth MD, et al. Correlation of the degree of dyspnea with health-related quality of life, functional abilities, and diffusing capacity for carbon monoxide in patients with systemic sclerosis and active alveolitis: results from the Scleroderma Lung Study. Arthritis Rheum 2005; 52: 592600.
  • 11
    Khanna D, Furst DE, Clements PJ, Park GS, Hays RD, Yoon J, et al. Responsiveness of the SF-36 and the Health Assessment Questionnaire Disability Index in a systemic sclerosis clinical trial. J Rheumatol 2005; 32: 83240.
  • 12
    Subcommittee for Scleroderma Criteria of the American Rheumatism Association Diagnostic and Therapeutic Criteria Committee. Preliminary criteria for the classification of systemic sclerosis (scleroderma). Arthritis Rheum 1980; 23: 58190.
  • 13
    Clements P, Lachenbruch P, Siebold J, White B, Weiner S, Martin R, et al. Inter and intraobserver variability of total skin thickness score (modified Rodnan TSS) in systemic sclerosis. J Rheumatol 1995; 22: 12815.
  • 14
    Clements PJ. Systemic sclerosis (scleroderma) and related disorders: clinical aspects. Baillieres Best Pract Res Clin Rheumatol 2000; 14: 116.
  • 15
    Steen VD, Medsger TA Jr. The value of the Health Assessment Questionnaire and special patient- generated scales to demonstrate change in systemic sclerosis patients over time. Arthritis Rheum 1997; 40: 198491.
  • 16
    Radloff L. The CES-D scale: a self-report depression scale for research in the general population. Appl Psychol Meas 1977; 1: 384401.
  • 17
    Ware JE Jr, Kosinki M, Keller S. SF-36 Physical and mental health summary scales: a user's manual. Boston: New England Medical Center Health Institute; 1994.
  • 18
    Ware J, Kosinski M, Dewey J. How to score version two of the SF-36 Health Survey. Lincoln (RI): QualityMetric Inc.; 2000.
  • 19
    Fries JF, Spitz P, Kraines RG, Holman HR. Measurement of patient outcome in arthritis. Arthritis Rheum 1980; 23: 13745.
  • 20
    Khanna D, Furst DE, Hays RD, Park GS, Wong WK, Seibold JR, et al. Minimally important difference in diffuse systemic sclerosis: results from the D-Penicillamine Study. Ann Rheum Dis 2006; 65: 13259.
  • 21
    Clements PJ, Wong WK, Hurwitz EL, Furst DE, Mayes M, White B, et al. The Disability Index of the Health Assessment Questionnaire is a predictor and correlate of outcome in the high-dose versus low-dose penicillamine in systemic sclerosis trial. Arthritis Rheum 2001; 44: 65361.
  • 22
    Clements PJ, Wong WK, Hurwitz EL, Furst DE, Mayes M, White B, et al. Correlates of the Disability Index of the Health Assessment Questionnaire: a measure of functional impairment in systemic sclerosis. Arthritis Rheum 1999; 42: 237280.
  • 23
    Schulberg HC, Saul M, McClelland M, Ganguli M, Christy W, Frank R. Assessing depression in primary medical and psychiatric practices. Arch Gen Psychiatry 1985; 42: 116470.
  • 24
    Coulehan JL, Schulberg HC, Block MR. The efficiency of depression questionnaires for case finding in primary medical care. J Gen Intern Med 1989; 4: 5417.
  • 25
    Clinical Informatics Research Group. UMaker User's Manual, UMDNJ Technical Report. Picastaway (NJ): UMDNJ; 1999.
  • 26
    Arias E. United States life tables, 2002. Natl Vital Stat Rep 2004; 53: 138.
  • 27
    Stalmeier PF, Goldstein MK, Holmes AM, Lenert L, Miyamoto J, Stiggelbout AM, et al. What should be reported in a methods section on utility assessment? Med Decis Making 2001; 21: 2007.
  • 28
    Stalmeier PF. Discrepancies between chained and classic utilities induced by anchoring with occasional adjustments. Med Decis Making 2002; 22: 5364.
  • 29
    Franzblau AN. A primer of statistics for non-statisticians. New York: Harcourt, Brace & World; 1958.
  • 30
    Malcarne VL, Greenbergs HL. Psychological adjustment to systemic sclerosis. Arthritis Care Res 1996; 9: 519.
  • 31
    Cossutta R, Zeni S, Soldi A, Colombelli P, Belotti Masserini A, Fantini F. Evaluation of quality of life in patients with systemic sclerosis by administering the SF-36 questionnaire. Reumatismo 2002; 54: 1227. In Italian.
  • 32
    Hellinger FJ. Expected utility theory and risky choices with health outcomes. Med Care 1989; 27: 2739.
  • 33
    Tsevat J. What do utilities measure? Med Care 2000; 38 Suppl 9: II1604.
  • 34
    King JT Jr, Tsevat J, Moossy JJ, Roberts MS. Preference-based quality of life measurement in patients with cervical spondylotic myelopathy. Spine 2004; 29: 127180.
  • 35
    Souchek J, Byrne MM, Kelly PA, O' Malley K, Richardson M, Pak C, et al. Valuation of arthritis health states across ethnic groups and between patients and community members. Med Care 2005; 43: 9218.