- Open Access
Development of the SF-6Dv2 health utility survey: comprehensibility and patient preference
Journal of Patient-Reported Outcomes volume 6, Article number: 47 (2022)
The SF-6Dv2 classification system assesses health states in six domains—physical functioning, role function, bodily pain, vitality, social functioning, and mental health. Scores have previously been derived from the SF-36v2® Health Survey. We aimed to develop a six-item stand-alone SF-6Dv2 Health Utility Survey (SF-6Dv2 HUS) and evaluate its comprehensibility.
Two forms of a stand-alone SF-6Dv2 HUS were developed for evaluation. Form A had 6 questions with 5–6 response choices, while Form B used 6 headings and 5–6 statements describing the health levels within each domain. The two forms were evaluated by 40 participants, recruited from the general population. Participants were randomized to debrief one form of the stand-alone SF-6Dv2 HUS during a 75-min interview, using think-aloud techniques followed by an interviewer-led detailed review. Participants then reviewed the other form of SF-6Dv2 and determined which they preferred. Any issues or confusion with items was recorded, as was as overall preference. Data were analyzed using Microsoft Excel and NVivo Software (v12).
Participants were able to easily complete both forms. Participant feedback supported the comprehensibility of the SF-6Dv2 HUS. When comparing forms, 25/40 participants preferred Form A, finding it clearer and easier to answer when presented in question/response format. The numbered questions and underlining of key words in Form A fostered quick and easy comprehension and completion of the survey. However, despite an overall preference for Form A, almost half of participants (n = 19) preferred the physical functioning item in Form B, with more descriptive response choices.
The results support using Form A, with modifications to the physical functioning item, as the stand-alone SF-6Dv2 HUS. The stand-alone SF-6Dv2 HUS is brief, easy to administer, and comprehensible to the general population.
Health-related quality of life (HRQoL) is a multi-faceted concept specifically related to how one’s health affects overall quality of life as it pertains to their physical, mental, emotional, and social functioning . Measures of HRQoL can be categorized into 2 types: health profile measures and preference-based health utility measures . Profile measures provide scores for each domain of health that is measured. Examples of heath profile measures include the SF-36 Health Survey (SF-36)  and the SF-36v2® Health Survey (SF-36v2) . Alternatively, health utility measures summarize ratings of multiple health domains into a single preference-based score anchored by the values 0 and 1, where 0 = death and 1 = perfect health . These preference-based health utility measures have become increasingly valuable for calculating quality adjusted life years (QALYs), and are widely used in clinical trials and in determining the value and benefits of health care.
The SF-6D is one of the most widely used preference-based health utility measures , along with the EuroQol-5D (EQ-5D) , and the Health Utilities Index (HUI) [8, 9]. Each of these measures is unique in terms of the domains measured, the items, and the preference weights used to determine scoring. The scoring algorithm of the SF-6D is based on studies assessing the value individuals place on different health limitations. Such studies use hypothetical scenarios where individuals trade between different health states [6, 10, 11]. The SF-6D is based on 6 health domains: physical functioning, role functioning, bodily pain, vitality, social functioning, and mental health. Since its development, researchers have validated county-specific value sets of the SF-6D for populations in the United Kingdom, Brazil, China, Japan, and Portugal [12,13,14,15,16]. Additionally, improvements have been made to the scoring algorithms for the SF-6D resulting in the development of the SF-6Dv2 [12, 17]. The SF-6Dv2 score is derived from 10 items in the SF-36v2. Compared to the SF-6D, the SF-6Dv2 describes more distinct levels of health, reduces floor effects, and provides clearer wording for health state valuation scores [12, 17].
The updated scoring algorithm of the SF-6Dv2 highlighted the need for a stand-alone measure with reduced respondent burden. A stand-alone SF-6Dv2 health utility measure eliminates the need to administer all 36 items in the SF-36v2 in order derive an SF-6Dv2 score. To address this need, 2 stand-alone forms of the SF-6Dv2 Health Utility Survey (SF-6Dv2 HUS) were developed: Form A and Form B. During the initial development of the standalone measure, we wanted to test whether respondents preferred a measure that aligns with the question type format of the SF-36v2 or a measure that resembles other health utility measures (e.g., the EQ-5D). We opted to create and subsequently test two versions of the SF-6Dv2, to learn which presentation is easiest to understand and complete by respondents. Form A asks users to answer 6 questions (one per health dimension, by selecting from 5 to 6 response choices each; similar to the formatting of the SF-36v2); Form B relies on headings to identify each of the 6 health dimensions and asks users to review 5–6 descriptive statements for each health domain and select the one that best describes them (see Table 1).
While experts agree that evaluation of content validity of HRQoL patient-reported outcome (PRO) measures is advisable, preference-based measures such as the SF-6Dv2 have not been held to this standard. Evaluation of content validity includes evaluating the relevance (i.e., all items pertain to the construct of interest [generic HRQoL]), comprehensiveness (i.e., items cover all aspects of HRQoL), and comprehensibility (i.e., items are understood as intended) of PRO measures. These properties are evaluated through qualitative research methods during which individuals assess and provide direct feedback on each of these elements [18,19,20,21,22]. Although a review of the literature did not identify published studies of content validation of preference-based measures, the research team felt it was an important step in completing the development of the SF-6Dv2 HUS. Given the 8 domains measured by the SF-36 have been well established as those key to measuring HRQoL [23, 24], and recent evidence has shown the SF-6Dv2 to be conceptually equivalent to the SF-36 , this research focused on evaluating the comprehensibility of the SF-6Dv2 HUS.
This study had 2 objectives: (1) to evaluate the comprehensibility of the stand-alone SF-6Dv2 HUS (both Form A and Form B) by conducting individual cognitive debriefing interviews with adults in the general population of the United States (US); and (2) to learn which version of the stand-alone SF-6Dv2 HUS adults in the US prefer.
This was a qualitative, cross-sectional, non-interventional study consisting of one-on-one cognitive debriefing interviews. This approach to questionnaire evaluation is based in cognitive psychology and the Cognitive Aspects of Survey Methodology framework [26, 27]. Within this framework, questionnaire respondents are assumed to handle a number of cognitive tasks: (1) understanding the question(s) they are being asked; (2) retrieving their answer from memory; (3) internally evaluating their response; and (4) matching their response to the response options available in the survey. The cognitive debriefing interviews use a think-aloud approach that is designed to identify problems in comprehension that can be used to improve elements of the questionnaire.
The 75-min audio-recorded interviews were conducted by experienced qualitative researchers trained on the specific objectives of the study. All interviews were conducted by telephone or webcam; allowing for nationwide participation by a diverse geographic sample, while also alleviating health risks, given interviews took place during the COVID-19 pandemic. All study materials were approved by one central independent review board (IRB).Footnote 1
Eligible participants were age 18 and older, living in the US, and fluent in US English. Specific quotas were established to ensure a diverse and representative sample in: age (20 participants aged 18–49 and 20 aged 50+), sex (at least 5 males aged 18–49 and 5 males aged 50+, at least 5 females aged 18–49 and 5 females aged 50+), presence of chronic health conditions (at least 20 who answered yes), race/ethnicity (at least 10 identifying as non-white), and education (at least 10 participants with high school diploma or less). Participants were excluded from the study if they were unwilling or unable to participate in a single 75-min interview.
All participants were recruited from the general population via a third-party recruitment vendor’s proprietary participant panel. All potential participants completed an online screening questionnaire to assess study inclusion criteria and standard demographic information. Participants who screened into the study were then directed to a second, brief questionnaire to collect further demographic information, and then scheduled for their interview. In total, 87 people were screened to participate. Recruitment was stopped when all quotas, including the total sample size of 40, was reached.
Interviews were conducted by one of two trained qualitative researchers with experience conducting cognitive debriefing interviews. At the beginning of each interview, the interviewer reviewed the consent statement in detail, answered any questions the participant might have, and asked for each participant’s verbal consent to participate. This was documented by each interviewer.
All interviews followed a standardized, semi-structured interview guide. Participants were randomly assigned one of the two forms of the stand-alone SF-6Dv2 HUS to debrief; half of the recruited sample (n = 20) debriefed Form A and the other half debriefed Form B. Interviewers used a think-aloud approach  to learn how well participants understood each aspect of the survey. During the think-aloud approach, participants were asked to read all parts of the survey—including title, instructions, items, and response choices—out loud and to say what they were thinking as they read the survey and answered the questions. If something was confusing to them, they were asked to describe what was confusing to them, and to articulate how they ultimately decided the meaning of the instruction, item, or response choice.
Following the think-aloud, participants answered a series of semi-structured follow-up questions about the various elements of the form they just completed, including instructions, recall period, items, and response choices. Responses to these questions, and spontaneous comments made during the think-aloud, were captured and later analyzed for evidence of each form’s comprehensibility. Lastly, participants were asked to review the alternate form of the stand-alone SF-6Dv2 HUS (i.e., whichever form they did not debrief earlier in the interview) and compare it to the one they had debriefed. They were then asked which form of the stand-alone SF-6Dv2 HUS they preferred, and why.
Data coding and analysis
Data coding and analysis followed a 5-step process.
Step 1: quick code
Upon completion of each interview, the interviewer conducted a “quick code,” populating a Microsoft Excel spreadsheet with interview data solely from the interviewer’s field notes. Data included any notable issues that arose during the interview (e.g., confusing, or unclear items), suggested changes to either Form A or Form B, and overall preferences.
Step 2: cross-check transcripts
As completed transcripts were received, they were first reviewed for quality, and then cross-checked against the quick code spreadsheet to confirm all feedback had been accurately recorded during the quick coding process.
Step 3: code transcripts
Transcripts were then coded to identify additional information shared by the participants, including overall opinions on the stand-alone SF-6Dv2 HUS, and any other suggestions or insights. Coding was completed in NVivo software (QSR International Pty Ltd, 2018) and reviewed by the study PI. Coding reliability was determined through a consensus-based approach. The researchers independently coded the same first two transcripts and then met to review their coding and resolve any discrepancies through discussion. This meeting also served to allow for any initial adjustments to the codebook and code definitions. At this point, coding was consistent. The remaining transcripts were divided between the two coders and coded independently. The coders met throughout coding to ensure consistency and address any questions, and the study Principal Investigator reviewed all coding as an additional step to ensure coding reliability.
Step 4: analysis
All coded data were reviewed and analysed by the study team.
Step 5: review and consensus meetings
Determinations about potential modifications to Forms A and B were made through a consensus-based approach. The study team reviewed each of the issues identified or suggestions made by participants and noted the proportion of study participants who raised the issue/suggestion and the nature of their feedback (e.g., is the suggested edit crucial to improving comprehension or simply a matter of personal preference?). The research team evaluated each issue or suggestion—including whether it was raised spontaneously or as a response to a probe—and subsequently decided whether a modification was warranted.
All suggestions and supporting evidence for changes to either form were documented in an item tracking matrix [21, 29]. The matrix includes the original items from both forms, relevant comments suggesting a needed change, a decision on whether to change, how to change, and any new wording. The matrix also contains similar information on the instructions, recall period, and response choices.
A total of 40 individuals participated in this study. Most were white (n = 26, 65.0%), female (n = 23, 57.5%), had completed some form of post-high school education (n = 29, 72.5%), and had a chronic health condition (n = 29, 72.5%). Half of the sample was between the ages of 18–49, and the other half was age 50 or older. All participants were in the US including the Northeast (n = 12, 30.0%), West (n = 7, 17.5%), Midwest (n = 5, 12.5%), and South (n = 16, 40.0%). All participants were asked to rate their overall health; of those, fourteen (35.0%) rated their overall health as ‘very good’ or ‘excellent.’ Health satisfaction ratings were also collected and were wide-ranging across a 10-point scale, with an average of 5.8 out of 10 (min = 1, max = 9). (See Table 2).
Form A cognitive debriefing results
All participants who debriefed Form A (n = 20), found it relevant, straightforward, and easy to understand. Participants were able to easily relate the questions to aspects of their daily lives and select an answer accordingly (see Table 3 for additional data).
Instructions and recall
All participants found the instructions for Form A clear and easy to understand. One participant initially missed the instructions but was able to complete the survey with no issues.
Fourteen participants found it easy to recall how they were feeling over the past 4 weeks and to answer each question within that timeframe. Of the 6 who did not, 4 recommended shortening the recall period to 2 weeks and 2 participants felt it was difficult to recall the past 4 weeks due to monotony of the previous months (related to the COVID-19 pandemic) but did not provide an alternative recall period.
Overall, participants found the physical functioning question easy to answer (n = 17). Of those who found it difficult (n = 3), 1 participant felt it was unclear whether the response choices were mutually exclusive (i.e., if they are limited a little in moderate activities, does that mean they cannot do vigorous activities?); another did not engage in vigorous activities and could not answer whether they were limited; and another was unsure how to answer the question because their multiple chronic health conditions limited them in different ways.
Most participants found the role functioning question easy to answer (n = 16). Of the 4 participants who had difficulty answering the question: 1 struggled with recalling regular daily activities over the past 4 weeks, another felt the question was too wordy and suggested changing the wording to “felt or were less productive,” 1 felt their answer would differ depending on whether they focused on work or activities outside of work, and another suggested splitting the question into 2 separate items (1 for physical health and 1 for emotional problems). However, upon further questioning, all participants were able to understand and interpret the question accurately.
Just over half of participants found this question easy to answer (n = 13). The other 7 found it difficult for a variety of reasons. Three struggled to recall their pain over the past 4 weeks—with 2 noting their pain fluctuated requiring them to come up with an average pain level so they could answer the question. While able select a response for this item, 2 participants found it difficult to do so quickly, as they felt the question and response choices were too subjective (i.e., definitions of pain will be different and so answers cannot be accurately compared). Two participants were unsure whether the question was asking about acute or chronic pain and felt their answers would differ depending on the type of pain.
Most participants (n = 19) found the vitality question easy to answer, although 4 took a longer time to select an answer as compared to previous items. The 1 participant who had difficulty answering struggled with recalling times when they felt worn out over the past 4 weeks. Additionally, 2 participants felt the phrase “worn out” was too vague and should specify whether it includes emotional problems or just physical health, however upon further probing each person considered both physical health and emotional problems when answering the question.
Fifteen participants found the social functioning question easy to answer. The 5 participants who found it difficult to answer referred to the COVID-19 social distancing restrictions in place at the time of the interviews. Because social activities were restricted due to local ordinances, these participants experienced interference with social activities in the 4 weeks prior to the interviews. Although the interference was not due to physical health or emotional problems, it made it difficult for them to answer this item, nonetheless.
Overall, participants found the mental health question easy to answer (n = 18). Of those who found it difficult (n = 2), 1 participant felt it was hard to admit, and be vulnerable enough to answer the question, while the other felt the current state of the world (e.g., ongoing COVID-19 pandemic) made it difficult to answer the question.
Form B cognitive debriefing results
Overall, most participants who debriefed Form B (n = 18) were able to easily relate questions to aspects of their daily lives and answer accordingly, and thought it was straightforward and easy to understand, with only 2 participants finding the form confusing or difficult to answer. Of these 2 participants, 1 struggled with whether to consider their health pre-COVID-19, or if they should answer in the present day, while the other was unsure what the survey was asking overall and therefore had a difficult time selecting statements that described them (see Table 3 for additional data).
Instructions and recall
All participants found the instructions for Form B clear and easy to understand. Most participants (n = 18) found it easy to recall how they were feeling over the past 4 weeks and had no difficulty answering each question within that timeframe. Of the 2 who did report issues, 1 recommended shortening the recall period to 2 weeks, while the other suggested it would be easier to remember the past 1–2 weeks, rather than the past 4.
Overall, the physical functioning question was found to be clear and easy to answer (n = 17). Three participants (out of 20) found it difficult to answer, primarily due to general confusion over which statement best described their health and the circumstances limiting their physical functioning.
Role functioning was perceived as easy to answer (n = 17). Participants interpreted “ability to work and do regular daily activities” to mean their general responsibilities as an employee, parent, or member of society, including going to work and completing household chores. Participants who found this item difficult to answer (n = 3) found the double negative statement to be confusing (i.e., you accomplished less than you would like none of the time; n = 1) and had different answers for physical and emotional health and would have preferred to answer each separately (n = 2).
Similar to Form A, just over half of the participants (n = 11) found this item easy to answer. The 9 participants who did not reported this item was difficult to interpret and found it challenging to distinguish between the response choices mild and moderate (given the response choice of very mild), and severe and very severe. Participants also had difficulty averaging their pain over 4 weeks given daily fluctuations. One participant was unsure if the item is referring to chronic or acute pain, which made selecting a statement to describe their pain difficult.
Overall, the item on vitality was easy for participants to answer (n = 17), although 3 reported finding it difficult to select a statement to describe themselves. These participants were confused over what “worn out” was referring to (e.g., does being tired at the end of a busy day qualify?). Participants also questioned the meaning of the heading (“Vitality”) and whether the concept is easily recognizable; ultimately, it was interpreted to mean being worn out mentally, worn out physically, or both.
Similar to the results for Form A, participants found the social functioning item in Form B easy to complete (n = 15), however the ongoing COVID-19 pandemic added difficulty for some individuals (n = 5). Participants who indicated this was difficult to answer noted that all social activities were limited, regardless of their health, making it challenging to decide which statement to select. Participants who found this item easy to complete also brought up COVID-19-related social restrictions, however it did not impede their ability to select a statement, or their understanding of the item.
While 12 participants had no difficulty with the mental health item in Form B, 8 participants found it difficult to select a statement to describe their mental health. Four described their feelings of depression or anxiety as variable and found it difficult to select one statement to best describe them over the past 4 weeks. Some participants (n = 3) also had difficulty selecting a statement if they experienced only depression or only anxiety. The double-barreled nature of the item wording made it difficult to choose the most appropriate statement. Similarly, 2 participants found the word “depressed” to be triggering, articulating there is a difference between being depressed and feeling depressed and it isn’t clear which the item is referring to. Finally, 1 participant found this item difficult to answer in an interview setting with a stranger, while another wasn’t sure how to answer given how the COVID-19 pandemic has influenced all aspects of life.
Comparison of Form A and Form B
There was a general tendency to prefer the last form the respondents had seen. Of the 20 participants who debriefed Form A, 12 preferred Form B after reviewing Form B, while only 7 preferred Form A. One participant had no preference. Of the 20 who debriefed Form B, only 2 preferred Form B after reviewing Form A, while 18 preferred Form A. Taking this recency effect into account, more participants preferred Form A above Form B (See Fig. 1 and Table 4).
Overall, participants found the items in Form A clearer and easier to answer. When comparing Forms A and B, more participants preferred answering questions (Form A) over choosing from a set of statements (Form B). The numbered questions and underlining of key words in Form A fostered quick and easy comprehension and completion of the survey. Participants also felt it looked more professional and was more in line with what they were used to seeing. While participants found some of the titles in Form B to be helpful, overall, they preferred the questions in Form A. Overall preference mostly aligned with participant preferences for individual items within the two forms. Individuals whose overall preference was Form A, also tended to prefer the question/answer items in Form A over the corresponding statement items in Form B and vice-versa. However, this was not always the case.
Despite an overall preference for Form A, almost half of participants (n = 19) preferred the physical functioning question in Form B, finding it clearer and easier to answer. They found it helpful to have the descriptions of vigorous and moderate activity in the response choices (n = 11), and they found the wording easier to understand (n = 9). Eight participants found the response choices in Form A to be challenging when comparing them to Form B.
This qualitative study was designed to elicit feedback from US adults on the overall comprehensibility of 2 different Forms (A and B) of the stand-alone SF-6Dv2 HUS Form and on which form they prefer. The study provided strong evidence that both forms of the stand-alone SF-6Dv2 HUS were understandable and easy to complete. There were no difficulties with instructions or recall period on either form; however, participants expressed preference for Form A, finding it easier to complete. The only exception was the physical functioning item, for which participants preferred the format of Form B. In Form B, the definitions of vigorous and moderate activity are included in the response choices, which participants preferred over the format of Form A.
Given the overall feedback, we decided to move forward with Form A, but with revisions to the physical functioning item to make it more like Form B, ensuring it is easier to understand (see Table 5). Specifically, the definitions of vigorous and moderate activity were moved from the question stem to the response choices, as participants found having the definitions in the response choices made it easier to select an answer. Although both forms had items participants found difficult to answer (on average, 4 participants (18%) had difficulty with Form A and 5 participants (26%) had difficulty with Form B), none of these difficulties prevented them from completing the survey, nor did they warrant further changes to the items or response choices. One respondent raised the issue whether the categories describing levels of physical function were mutually exclusive. This issue was analyzed in detail during the development of the SF-6Dv2  as well as in previous analyses of these physical function items . These analyses strongly support that the health levels of the PF item forms a clear hierarchy assessing one overall construct of physical function. We believe that the revised descriptions of the levels of physical function clarifies this hierarchy.
Of particular interest during cognitive debriefing was the participant feedback on the bodily pain items. While no participants asked for clarification on the bodily pain items when completing either form, some participants reported issues with the items during think-aloud. Of those who debriefed Form A, two found the response choices “too subjective”; two noted challenges with recalling pain over the last 4 weeks; and two struggled to determine if the item was asking about acute or chronic pain. Since the chosen version of the BP item is identical to the first item of the SF-36 (and SF36v2), these results should be considered in light of the body of studies on the validity of this bodily pain item. Psychometric analyses have supported that the response choices of the BP item define separate levels of pain . The issue of length of recall has been examined by comparing different version of the BP item with an average of momentary assessments covering the same time frame . Strong correlations were found between average momentary assessment and all lengths of recall. Highest correlation was seen for 1 day recall, followed by 3 days, 4 weeks and 1 week recall. Four-week recall had higher correlation with momentary assessment than 7-day recall . On a pragmatic level, the optimal recall will depend on the population and intended use of the instrument. For conditions where pain is episodic rather than constant, a too short recall period may lead to high variability in the assessment of pain, unless the instrument is administrated very frequently. For these reasons, we decided to keep the current version of the bodily pain item in the standard version of the SF-6Dv2 HUS, but to also suggest an additional version of the SF-6Dv2 HUS, using a 1-week recall for all the items where a recall is specified.
This study had several limitations. The sample was based in the United States and the study data collection took place in October and November 2020. Participants indicated that their answers to the social functioning and mental health items were influenced by external factors including the ongoing COVID-19 pandemic, surging cases of COVID-19 in some regions of the US, COVID-19-related social distancing policies and restrictions, and stress regarding the contentiousness of 2020 US presidential election. Although these factors influenced the participants’ answers, they had no bearing on their ability to understand the items and select responses. Additionally, due to COVID-19 travel and social-distancing restrictions, all interviews were conducted by phone or webcam; it is typically preferable to conduct as least some interviews in person.
This study had unique strengths, including the characteristics of participants. Recruitment quotas ensured a representative sample, including a wide age range of participants of both sexes, participants with the equivalent of a high school education or less (27.5%), and participants with chronic health conditions (72.5%), including diabetes, hypertension, hyperlipidemia, chronic pain, HIV, arthritis, depression, and anxiety. The decision to randomize which form of the stand-alone SF-6Dv2 HUS participants debriefed was an additional strength. Randomizing the order controlled for the possibility of recency effects impacting the results. The number of interviews conducted (n = 40) is a further strength, supporting the comprehensibility of the stand-alone SF-6Dv2 HUS for use with a general population of adults. While 7–10 cognitive debriefing interviews can be sufficient to determine comprehensibility of an instrument, testing the SF-6Dv2 HUS in 40 interviews ensures it can be used with a diverse population . While identified above as a limitation, it is also a strength that participants were able to consider health-related impacts on their social function versus pandemic-induced impacts. This differentiation confirmed their understanding of the items and the concepts being measured.
Finally, testing the stand-alone SF-6Dv2 HUS with participants as part of the development process was an additional strength of this study. There is scant published literature on health utility survey measures documenting testing with participants during development, yet this testing is an important step to confirm the comprehensibility of the measure [18,19,20,21]. Furthermore, this testing ensures all aspects of the survey (including instructions, questions, and response choices) are understandable and easy for patients to complete.
The development of the stand-alone SF-6Dv2 HUS is a key addition to the field of HRQoL. Health utility measures are widely used by health regulatory agencies, and systems that review approval for payment of medication or conduct comparative effectiveness research. A brief, easy to administer, stand-alone SF-6Dv2 can be more easily implemented and interpreted across patient groups and disease areas than its predecessor, aligning it more closely with the usefulness of the EQ-5D and HUI. Current research is evaluating the psychometric properties of the stand-alone SF-6Dv2 HUS to confirm this in the general population, and further research should be done to confirm it within specific disease populations.
In conclusion, the stand-alone SF-6Dv2 HUS is an understandable and easy to use assessment of HRQoL intended for use with adults. Use of the stand-alone SF-6Dv2 HUS can contribute to the comprehensive assessment of a patient’s health status and administrators can feel confident it is measuring the intended concepts with fidelity while minimizing patient burden.
Availability of data and materials
Specific data points can be made available upon reasonable request.
New England IRB (NEIRB) Study #1293768; Given that this study posed minimal risk for study participants, NEIRB approved a waiver of signed consent.
Food and Drug Administration
Generic preference-based measure of health
Health-Related Quality of Life
Independent Review Board
Mental component summary
Physical component summary
QualityMetric Incorporated, LLC
CDC (2000) Measuring healthy days: population assessment of health-related Quality of Life, Atlanta
Khanna D, Tsevat J (2007) Health-related Quality of Life—an introduction. Am J Manag Care 13:S218–S223
Ware JE, Sherbourne CD (1992) The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection. Med Care 30:473–483
Ware J, Kosinski M, Bjorner J, Turner-Bowker D, Gandek B, Maruish M (2007) Development. User's manual for the SF-36v2® Health Survey, Lincoln, RI
Bakker CH, Rutten-van Mölken M, van Doorslaer E et al (1993) Health related utility measurement in rheumatology: an introduction. Patient Educ Couns 20:145–152. https://doi.org/10.1016/0738-3991(93)90128-J
Brazier J, Roberts J, Deverill M (2002) The estimation of a preference-based measure of health from the SF-36. J Health Econ 21:271–292. https://doi.org/10.1016/S0167-6296(01)00130-8
EuroQol Group (1990) EuroQol—a new facility for the measurement of health-related quality of life. Health Policy 16:199–208. https://doi.org/10.1016/0168-8510(90)90421-9
Feeny D, Furlong W, Boyle M et al (1995) Multi-attribute health status classification systems. Health Utilities Index. Pharmacoeconomics 7:490–502. https://doi.org/10.2165/00019053-199507060-00004
Feeny D, Furlong W, Torrance GW et al (2002) Multiattribute and single-attribute utility functions for the health utilities index mark 3 system. Med Care 40:113–128. https://doi.org/10.1097/00005650-200202000-00006
Brazier JE, Rowen D, Hanmer J (2008) Revised SF-6D scoring programmes: a summary of improvements. Patient Rep Outcomes Newsl 40(Fall):14–15
Brazier J, Usherwood T, Harper R et al (1998) Deriving a preference-based single index from the UK SF-36 Health Survey. J Clin Epidemiol 51:1115–1128. https://doi.org/10.1016/s0895-4356(98)00103-6
Mulhern BJ, Bansback N, Norman R et al (2020) Valuing the SF-6Dv2 classification system in the United Kingdom using a discrete-choice experiment with duration. Med Care 58:566–573. https://doi.org/10.1097/MLR.0000000000001324
Cruz LN, Camey SA, Hoffmann JF et al (2011) Estimating the SF-6D value set for a population-based sample of Brazilians. Value Health 14:S108–S114. https://doi.org/10.1016/j.jval.2011.05.012
Lam CLK, Brazier J, McGhee SM (2008) Valuation of the SF-6D health states is feasible, acceptable, reliable, and valid in a Chinese population. Value Health 11:295–303. https://doi.org/10.1111/j.1524-4733.2007.00233.x
Shiroiwa T, Fukuda T, Ikeda S et al (2016) Japanese population norms for preference-based measures: EQ-5D-3L, EQ-5D-5L, and SF-6D. Qual Life Res 25:707–719. https://doi.org/10.1007/s11136-015-1108-2
Ferreira LN, Ferreira PL, Pereira LN et al (2010) A Portuguese value set for the SF-6D. Value Health 13:624–630. https://doi.org/10.1111/j.1524-4733.2010.00701.x
Brazier JE, Mulhern BJ, Bjorner JB et al (2020) Developing a new version of the SF-6D health state classification system from the SF-36v2: SF-6Dv2. Med Care 58:557–565. https://doi.org/10.1097/MLR.0000000000001325
Terwee CB, Prinsen CAC, Chiarotto A et al (2018) COSMIN methodology for evaluating the content validity of patient-reported outcome measures: a Delphi study. Qual Life Res 27:1159–1170. https://doi.org/10.1007/s11136-018-1829-0
U.S. FDA (2009) Guidance for industry: patient-reported outcome measures: use in medical product development to support labeling claims. https://www.fda.gov/downloads/drugs/guidances/ucm193282.pdf. Accessed 11 May 2021
U.S. FDA (2018) Patient-focused drug development: collecting comprehensive and representative input. https://www.fda.gov/regulatory-information/search-fda-guidance-documents/patient-focused-drug-development-collecting-comprehensive-and-representative-input
Reeve BB, Wyrwich KW, Wu AW et al (2013) ISOQOL recommends minimum standards for patient-reported outcome measures used in patient-centered outcomes and comparative effectiveness research. Qual Life Res 22:1889–1905. https://doi.org/10.1007/s11136-012-0344-y
EMA (2016) Appendix 2 to the guideline on the evaluation of anticancer medicinal products in man. The use of patient-reported outcome (PRO) measures in oncology studies
Ware JE, Snow KK, Kosinski M et al. (1993) SF-36 health survey: manual and interpretation guide, Boston, MA
Ware JE (1995) The status of health assessment 1994. Annu Rev Public Health 16:327–354. https://doi.org/10.1146/annurev.pu.16.050195.001551
Poder TG, Fauteux V, He J et al (2019) Consistency between three different ways of administering the short form 6 dimension version 2. Value Health 22:837–842. https://doi.org/10.1016/j.jval.2018.12.012
Tourangeau R, Rips LJ, Rasinski KA (2000) The psychology of survey response. Cambridge University Press, Cambridge
Jobe JB (2003) Cognitive psychology and self-reports: models and methods. Qual Life Res 12:219–227. https://doi.org/10.1023/a:1023279029852
Willis GB (2005) Cognitive interviewing: a tool for improving questionnaire design/Gordon B. Willis. Sage Publications, Thousand Oaks
Patrick DL, Burke LB, Gwaltney CJ et al (2011) Content validity—establishing and reporting the evidence in newly developed patient-reported outcomes (PRO) instruments for medical product evaluation: ISPOR PRO Good Research Practices Task Force report: part 2–—assessing respondent understanding. Value Health 14:978–988. https://doi.org/10.1016/j.jval.2011.06.013
Rose M, Bjorner JB, Gandek B et al (2014) The PROMIS Physical Function item bank was calibrated to a standardized metric and shown to improve measurement efficiency. J Clin Epidemiol 67:516–526. https://doi.org/10.1016/j.jclinepi.2013.10.024
Becker J, Schwartz C, Saris-Baglama RN et al (2007) Using Item Response Theory (IRT) for developing and evaluating the Pain Impact Questionnaire (PIQ-6™). Pain Med 8:S129–S144. https://doi.org/10.1111/j.1526-4637.2007.00377.x
Broderick JE, Schwartz JE, Vikingstad G et al (2008) The accuracy of pain and fatigue items across different reporting periods. Pain 139:146–157. https://doi.org/10.1016/j.pain.2008.03.024
This study was funded by QualityMetric, which licenses the SF-6Dv2. Permission to reproduce and to use the SF-6Dv2 and the associated trademark(s) is routinely granted royalty-free to individuals and organizations that collect their own data for purposes of scholarly research. Permissions for both scholarly and commercial use can be obtained by completing a License Application Form. All other uses, commercial and noncommercial, may require payment of a license fee. Completion of the License Application Form will result in the quotation of any user fees and, upon user request and approval by QualityMetric, the issuance of a license and invoice. Any organization or individual wishing to reproduce the survey documented herein and/or any associated intellectual property (e.g., the trademarks, scoring algorithms, interpretation guidelines, and/or normative data) for any purpose must register or obtain a license from QualityMetric. For information about registering or obtaining a license, go to www. qualitymetric.com.
Ethics approval and consent to participate
All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. All participants provided consent.
Consent for publication
At the time of the original submission, LB, JBB, MLC, MKW, and MK were full-time employees of QualityMetric, BM was employed at the Centre for Health Economics Research and Evaluation, University of Technology Sydney, and JB was employed at the School of Health and Related Research (ScHARR), University of Sheffield. They have no conflicts of interest to report.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
About this article
Cite this article
Broderick, L., Bjorner, J.B., Lauher-Charest, M. et al. Development of the SF-6Dv2 health utility survey: comprehensibility and patient preference. J Patient Rep Outcomes 6, 47 (2022). https://doi.org/10.1186/s41687-022-00455-w
- Patient-centered research
- Utility score
- Preference-based measure