Hypothetical acne-related health outcomes questions. TTO indicates time trade-off; WTP, willingness to pay.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Chen CL, Kuppermann M, Caughey AB, Zane LT. A Community-Based Study of Acne-Related Health Preferences in Adolescents. Arch Dermatol. 2008;144(8):988–994. doi:10.1001/archderm.144.8.988
To examine preferences for acne-related outcomes in adolescents.
Community-based, cross-sectional survey study.
Four public high schools in San Francisco, California.
Volunteer sample of 266 adolescents with acne.
Main Outcome Measures
Health utilities for current acne state and 3 hypothetical acne-related states (100% clearance, 50% clearance, and 100% clearance but with residual scarring) using time trade-off (TTO) and willingness to pay metrics. A self-administered written survey was used.
The mean (median) utilities for current acne state using the TTO metric was 0.961 (0.985). One hundred percent acne clearance received a higher utility score (mean [median] score, 0.978 [0.994]) than 50% clearance (0.967 [0.992]; P < .001 by Wilcoxon signed rank test) and 100% clearance with scarring (0.965 [0.992]; P < .001). Although current acne state utility was not correlated with physician-rated severity (P = .23), a significant association with subject-rated severity was observed in both univariate analysis (P = .01) and after adjustment for duration of acne, having seen a physician for acne, and income (P = .05). Adolescents were willing to pay a median of $275 to have never had acne in their lifetime and were willing to pay significantly more for 100% clearance than for 50% clearance or 100% clearance but with scarring (P < .001 for both comparisons).
We describe adolescents' acne-related health state utilities. Compared with current acne state utility, both partial and total clearance with scarring are substantially less preferable than total clearance. Subjects' self-rated disease severity correlates with current acne state utility, whereas physician-rated severity does not. A self-administered paper instrument can effectively assess adolescents' acne-related preferences in community-based samples.
Acne vulgaris affects almost all adolescents,1,2 and its psychological impact on this age group has been documented in the medical literature for more than 50 years.3,4 Recent studies5-9 have characterized the relationship between acne and outcomes such as anxiety, depression, embarrassment, lack of self-confidence, social dysfunction, and unemployment. Reducing the psychosocial impact of acne is considered one of the guiding principles for its clinical management,10,11 and it is important to measure and evaluate this impact.
Despite adolescents' vulnerability to acne's psychosocial effects, surprisingly little research attention has been paid to these effects specifically in this age group.12 Much of the characterization of acne-related disability has been largely drawn from convenience samples of patients under dermatologic care, with no particular focus on adolescents.13-15 There have been few community-based studies of the psychosocial impact of acne on adolescents,16-18 and we are not aware of any attempts to quantify adolescents' preferences for various acne-related states.
Health state utilities are quantitative measures of health outcome preferences usually measured on a scale of 0 (death) to 1 (ideal health state or outcome).19 They have distinct advantages over disease-specific quality-of-life (QOL) instruments in that they (1) allow comparison across different disease states and (2) can be used to generate quality-adjusted life years (QALYs), which are then utilized in decision and cost-effectiveness analyses. The time trade-off (TTO)20 and willingness to pay (WTP) metrics measure how patients value various disease-related outcomes by assessing how much they are willing to give up of their life expectancy or money, respectively, to be free of their disease or achieve some improved outcome. We sought to measure acne-related health state utilities among adolescents by assessing health outcome preferences in a community-based sample of San Francisco, California, public high school students with acne.
The protocol for this study was approved by the San Francisco Unified School District (SFUSD) and the University of California, San Francisco (UCSF), Committee on Human Research. Participation by all subjects was voluntary. Student participation required minor assent and parent or guardian (hereinafter referred to as “parent”) consent. Parents indicated their verbal (implied) consent through the return of their completed survey.
A convenience sample of adolescent subjects was recruited from 4 public high schools in the SFUSD using the following on-campus recruitment methods: (1) tabling, wherein students voluntarily approach a manned display; (2) in-class information sessions (usually in health or physical education classes); and (3) daily bulletin announcements. Male and female high school students of any race, aged 14 to 18 years, were considered for inclusion if they were graded with a score of 1 or higher on the Investigator's Static Global Assessment (ISGA) scale for acne21 (ie, at or above a severity level described as “almost clear: rare noninflammatory lesions present, with rare noninflamed papules”) and could complete a written survey in English. A completed parent survey was also required for enrollment. Parents or guardians were included if they identified themselves as the student's caregiver and completed a written survey in English, Spanish, or Chinese at home. During on-campus after-school appointments, adolescents completed written surveys and had their acne graded by a trained investigator using the ISGA scale.
A self-administered written survey completed by the students assessed TTO and WTP values for the following hypothetical acne-related outcomes: 100% clearance or cure (being free of acne from this day forward), 100% clearance but with permanent residual scarring, 50% clearance, and never having acne in their lifetime (ie, “never had it, never will”). Question stems for TTO elicitation were: “How much of your lifetime would you be willing to give up . . . ,” whereas the question stem for WTP elicitation was “How much money would you be willing to pay today . . . ” (Figure). Parents were asked to complete a written survey assessing their WTP for the same 4 scenarios for their children because parents typically bear the expense of their child's acne treatment.
Responses to the TTO scenarios for never having acne were used to calculate subjects' current acne state utility score (primary outcome). This calculation was performed by dividing the participant's reduced life expectancy (in years) without acne by his or her full life expectancy (in years) with acne. This calculated utility value for current acne state then served as the anchor for chained calculations of TTO scores for the remaining 3 scenarios. Chained calculations were performed by multiplying the participant's current acne state utility by the ratio of full life expectancy to reduced life expectancy based on their answer for the given question. The chaining technique provides utility scores for hypothetical acne-related outcomes using the value they place on their current acne state as a reference. Calculations did not discount end-of-life years or specify acne duration time frames. The 2002 United States Life Tables22 provided life expectancy data for the TTO utility calculations.
Logistical considerations precluded the implementation of traditional interview-based or electronic format titration methods for preference elicitation. We opted for a self-administered survey to assess preferences for hypothetical acne-related outcomes with an investigator (C.L.C.) present to “identify and explain the apparent inconsistency to the subject but not to insist on its rectification.”23 For each of the 4 acne-related health outcomes, subjects were instructed to choose 1 best answer from a list of possible TTO or WTP choices with the additional option to write-in any value of their choice (Figure).
Pilot-testing of the survey instruments in 13 adolescent patients and their accompanying parents during visits for care at the UCSF dermatology clinics led to the revised health state utility questions that were used for primary data collection. In addition to health state utility preferences, information about acne history and demographics was collected from both adolescents and parents. Race information was collected by subject self-report and was used for 2 primary purposes: (1) to aid in the calculation of TTO utilities from race-adjusted life expectancy tables and (2) to explore the potential influence of race on acne-related TTO or WTP utilities. Adolescents and parents were each asked to indicate which racial or ethnic group best described them from a list on the survey instrument. Parents were also asked to estimate current annual pretax income for their household from a list of income categories on the survey.
Adolescents were asked about the duration of their acne history, how they would rate their current acne severity on an 11-point scale (0 [clear] to 10 [worst imaginable]), whether they had ever seen a physician for their acne, and whether they had ever used medications to treat their acne. Parents were asked if they themselves had a history of acne. After survey completion, adolescents also completed the Acne-QOL, a validated self-administered questionnaire designed to measure acne-specific QOL.24
A sample size calculation was performed to approximate the number of subjects required to achieve a 95% confidence interval total width of 0.10 around the utility estimate. Based on the variation of utility scores reported in previous studies25,26 assessing dermatologic conditions or adolescent populations, a standard deviation of 0.30 was used. The calculated requisite sample size was 138.27
Health state utility scores are reported using both mean and median values given their commonly skewed distribution. The Wilcoxon signed rank test was used to compare within-subject differences in utility scores for the multiple hypothetical outcomes. Medians were independently compared using the Mann-Whitney rank sum test and across single covariates using the Kruskal-Wallis test. The relationship between the current acne state utility score (and WTP for never having acne) and the hypothesized primary predictor, a subject's self-rated acne severity, was examined using univariate and multivariate linear regression. The nonparametric bootstrap method (using 1000 replicates) was used in regression analyses to cope with nonnormality.28 The multivariate models were constructed using backward selection wherein all predictors of interest (self-rated severity, age, sex, race, income, duration of acne, having seen a physician for acne, having used acne medications, and ISGA score) were put into the model, and those covariates that did not meet the retention criterion of P < .20 were sequentially removed from the model until the final model contained only those meeting that criterion. For multivariate analysis of WTP measures, income was retained for face validity. Data were managed using Access 2000 and Excel 2000 software (Microsoft, Redmond, Washington). All statistical analyses were performed using Intercooled Stata software (version 9.2; Stata Corp LP, College Station, Texas). Responses to TTO or WTP questions that were incomplete or could not be accurately encoded were not included in the analysis.
A total of 266 students were recruited from 4 San Francisco public high schools. Enrolled students ranged in age from 14 to 18 years, 59% were female, and most (65%) identified themselves as Asian (Table 1). Girls had clinically less severe acne by ISGA rating (P < .001) but reported having a longer history of acne (P < .001) and were more likely to have tried medication for treatment of their acne (P = .001). Parent respondents' mean (SD) age was 47.7 (7.1) years, and 73% were female. The parents' racial and ethnic distribution was similar to that of adolescent subjects, except that there was a smaller multiethnic proportion. Most of the parents (65%) reported having a history of acne themselves.
The mean (SD) TTO utility score for current acne state was 0.961 (0.092), and the median TTO utility score was 0.985 (interquartile range [IQR], 0.968-0.999) (Table 2). Compared with current acne health state, all 3 hypothetical health states (100% acne clearance, 50% clearance, and 100% clearance but with residual scarring) were associated with higher utility scores (P < .001 for all comparisons). The TTO utility scores were significantly higher for 100% clearance than both 50% clearance and 100% clearance with scarring (P < .001 for both comparisons). The TTO utilities did not differ significantly across age, sex, or race in univariate analyses (see Table 3 for P values).
Significant predictors of lower TTO utility scores for current acne state by univariate analysis included higher self-rated severity, having seen a physician for acne, longer duration of acne history, and lower income. When stratified by having seen a physician for acne, the mean and median utility scores for current acne state were 0.932 and 0.983, respectively, for those who had seen a physician and 0.972 and 0.992 for those who had not (P = .002). No correlation was found between clinical acne severity (ISGA score) and utility for current acne state (P = .23).
We hypothesized a priori that self-rated acne severity would correlate with TTO utility for current acne state. The statistical significance of this inverse relationship seen in univariate analysis using bootstrap linear regression (P = .01) was maintained after adjustment for income, duration of acne, and having seen a physician for acne (P = .046) (Table 3). Income also maintained its positive correlation with TTO for present acne state in the multivariate model (P = .01). The total Acne-QOL score correlated well with TTO utility for current acne state (P < .001), with all 4 domains positively associated in individual univariate analysis.
Adolescents were willing to pay a median of $275 (IQR, $100-$1000) to have never had acne in their lifetime. Similarly, parents were willing to pay a median of $250 (IQR, $100-$1000) (Table 4) for their child to have never had acne in their lifetime. Adolescents and parents were also willing to pay comparable amounts for the other 3 outcomes: 100% clearance, 50% clearance, and 100% clearance but with scarring. Like the relationships observed between TTO utilities, both parents and adolescents were willing to pay significantly more for 100% clearance than for 50% clearance or 100% clearance but with scarring (P < .001 for all comparisons). Self-rated severity was a significant predictor of WTP for never having acne in univariate analysis (P = .02) and maintained significance after adjusting for income, race, and having seen a physician for acne (P = .048) (Table 5). Income (despite being retained in the WTP model for face validity), age, and sex were not found to be significant predictors of this WTP outcome. The total Acne-QOL score correlated well with WTP for never having acne (P = .02).
Our study has several important findings that highlight both clinical and methodological considerations in adolescents' acne-related QOL. We have described preferences for acne-related outcomes among community adolescents using TTO and WTP metrics. We also measured parents' WTP for these outcomes with regard to their child's acne.
Prior studies reporting acne-related health state utilities differ from the current study in several methodological ways. One study26 cataloging health state utilities for numerous dermatologic outcomes included a small clinic-based sample of patients with acne that was not restricted to adolescents. Another study,29 using a different utility measurement method, reported population-based utility scores for a variety of medical conditions and included a small sample of patients with acne. This sample included only those patients requiring medication for their acne and was not restricted to adolescents.29 The fact that these studies were limited to those patients seen in a clinic or those requiring medication for their acne might explain the slightly lower mean utility values compared with those seen in our sample (Table 6). Similarly restricted subsets in our sample yielded comparable results (Table 6). The disability caused by acne has also been measured through the use of a number of general, dermatologic, and acne-specific QOL instruments.13,18,34-37 To our knowledge, this is the first community-based study of acne-related TTO and WTP utilities among adolescents.
A benefit of health state utilities is their ability to be compared with utilities for other disease states. A comparison between the present study's results and those from other studies for such disease states as epilepsy, myopia, stable angina, skin neoplasm of uncertain behavior, and procedure-related miscarriage is given in Table 6.
Although the adolescents in our study seemed to prefer all of the hypothetical outcomes to their current state, they strongly preferred complete clearance over partial clearance or clearance with scarring. Based on their respective utility scores, it seems that adolescents consider 50% clearance and total clearance with scarring to be not much of an improvement over their current level of acne.
Knowledge of these patient preferences may help dermatologists balance clinical trial results with patients' expectations of therapy. Randomized, blinded, placebo-controlled trials have shown that 3 to 4 months of conventional acne therapy, including topical benzoyl peroxide, topical retinoids, and oral antibiotics, typically produces reductions in lesion counts in the 40% to 60% range.38,39 It has also been suggested that the incidence of scarring from facial acne approaches 95%.40 Thus, adolescents' marked preference for total clearance over partial (50%) clearance or clearance with scarring suggests that physicians must weigh high patient expectations against these clinical data regarding efficacy and risk of sequelae.
We observed a close relationship between adolescents' utility score and subject-rated acne severity. However, this relationship was not seen with physician-rated severity when using the ISGA scale. These findings support those of other studies41,42 in which acne-related disease burden correlates with self-rated disease severity and not physician-rated severity. Based on these data, physicians would do well to incorporate patients' individual acne-related burden and preferences for outcomes into their medical decision making by considering patients' self-rated acne severity along with their own clinical grading when formulating treatment decisions.
Interestingly, income was a significant predictor of TTO utility for current acne state in both the univariate analysis and multivariate linear regression model. The positive association (the higher the income group, the higher the utility score) suggests that adolescents from families of lower income categories may report a greater burden of disease from acne, even when adjusting for duration of acne history, self-rated acne severity, and whether they had seen a doctor for their acne. Income also explains the level of significance of the effect of Caucasian race on TTO for present acne state in univariate analysis (P = .06) (Table 3) because the Caucasian group in our cohort had a significantly higher reported income than the Asian reference race group (P < .001). In the presence of the income variable in the multivariate model, the categorical race variables failed to meet the retention criterion of the backward selection procedure and were consequently dropped from the model.
Our study findings have methodological implications as well. First, validity of our TTO utility measures for current acne state as well as WTP for never having acne is supported by their strong correlation with scores from the total Acne-QOL, a validated acne-specific QOL instrument.24 Second, our self-administered paper instrument is a practical method for assessing TTO and WTP for current health state as well as hypothetical outcomes. Third, adolescents in the community are capable of understanding the preference scenarios and making utility valuations in this manner.
Our study presents several unique advantages: (1) The larger sample size provides more precise estimates of acne-related utility scores, which can present meaningful differences at the level of the second or third decimal; (2) to our knowledge, no previous study has quantified acne-related preferences specifically in adolescents; (3) we introduced a self-administered paper instrument for assessing acne-related utilities, including both TTO and WTP metrics; and (4) our community-based design captured the disability among adolescents with acne who may not otherwise have the means or knowledge to seek professional dermatologic care.
By examining a community-based sample we were able to assess acne-related utilities among both those adolescents who had previously sought professional care for their acne and those who had not. Utilities measured in those who had seen a physician for their acne likely reflect values that would be found in a clinic-based sample. That subset demonstrated lower TTO utility scores for current acne state (mean [median], 0.932 [0.983]) than the overall cohort, comparable to those seen in clinic-based samples of patients with: acne (0.938 [0.990]),26 epilepsy (0.92 [0.98]),31 and CCS class II angina (median, 0.997)30 (Table 6).
Our study has limitations that deserve mention. Owing to logistical considerations at the high schools, we did not conduct interviewer-administered or computerized utility assessments to compare with the values we obtained using our self-administered paper instrument. This precluded us from comparing the values we obtained with those which would have been obtained using the more traditional tools. However, 2 validation studies43,44 using self-administered paper instruments to assess utilities in adults have both concluded that paper instruments are reliable tools for assessing utilities.
Logistical considerations also precluded administration of a “test scenario” asking participants to valuate a different health state to ensure their understanding of the exercise. Suspecting poor understanding in 25 medical records, in which preferences may have been illogically ordered, we conducted a sensitivity analysis excluding those medical records. This analysis produced a TTO utility for present acne state with an identical median, a mean that differed by only 0.002, and no changes to the overall conclusions of the study.
Generalizability of our findings may be limited by the high proportion of Asian students in our cohort and the voluntary nature of participation. Our cohort's Asian predominance is reflective of the high representation of this ethnic group in SFUSD high schools.45 Nearly half of the cohort was recruited from a large, urban college preparatory school with a competitive admissions process and higher relative proportions of Asian and Caucasian students. Although this might raise concerns about generalizability, TTO utility scores were not found to differ significantly across racial categories in our cohort as discussed in the “Preferences for Acne-Related Outcomes” subsection in the “Results” section. The high proportion of Asians in our study, however, may also highlight an ethnic group that has traditionally received little attention in the dermatology literature.
The voluntary nature of participation for both adolescents and their parents may have also introduced selection bias that favored recruitment of a convenience sample of students with more severe acne. However, although adolescents who believed their acne to be severe may have been more apt to participate because of an interest in the topic, this is likely offset by a potential reluctance to participate by those with severe acne who may have been embarrassed by their skin condition and wished to avoid drawing attention to it. Furthermore, although nearly half of our volunteers came from a competitive college preparatory high school, our judgment that adolescents are able to reliably report health-related preferences may be limited by the high proportion of motivated and academically oriented students in our cohort.
Despite potential limitations, our findings contribute novel information to our understanding of the burden of acne in the community. The TTO utilities reported herein can be used to calculate QALYs for future cost-effectiveness analyses of management and therapies and in the medical decision-making process for adolescents with acne. Future studies can utilize the paper TTO instrument described herein to sample large populations feasibly and would do well to compare health state utilities from paper and interview elicitation methods. Adolescents in the community have demonstrated their ability to make acne-related TTO and WTP utility valuations and should not be overlooked in future assessments of health state preferences. Their inclusion in these studies may aid in the management of acne in this age group by providing information for cost-effectiveness analyses of medications, aiding in therapeutic decision making, and increasing physicians' understanding of the psychosocial impact of this disease.
Correspondence: Lee T. Zane, MD, MAS, Anacor Pharmaceuticals Inc, 1020 E Meadow Circle, Palo Alto, CA 94303 (firstname.lastname@example.org).
Accepted for Publication: June 19, 2007.
Author Contributions: Dr Zane 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 concept and design: Chen, Kuppermann, Caughey, and Zane. Acquisition of data: Chen. Analysis and interpretation of data: Chen, Kuppermann, Caughey, and Zane. Drafting of the manuscript: Chen, Kuppermann, and Zane. Critical revision of the manuscript for important intellectual content: Chen, Kuppermann, Caughey, and Zane. Statistical analysis: Kuppermann, Caughey, and Zane. Obtained funding: Chen and Zane. Administrative, technical, and material support: Chen and Zane. Study supervision: Kuppermann and Zane.
Financial Disclosure: Dr Zane has participated on advisory boards for Connetics Corp, Stiefel Laboratories Inc, Medicis Pharmaceutical Corp, and QLT Inc, and is now an employee of Anacor Pharmaceuticals Inc.
Funding/Support: This study was supported in part by the Mount Zion Health Fund of the Jewish Community Endowment Fund.
Role of the Sponsors: The sponsors had no role in the design and conduct of the study; in the collection, analysis, and interpretation of data; or in the preparation, review, or approval of the manuscript.
This article was corrected online for typographical errors on 8/18/2008.