To evaluate community-based values for avoiding pandemic influenza (A) H1N1 (pH1N1) illness and vaccination-related adverse events in adults and children.
Adult community members were randomly selected from a nationally representative research panel to complete an internet survey (response rate = 65%; n = 718). Respondents answered a series of time trade-off questions to value four hypothetical health state scenarios for varying ages (1, 8, 35, or 70 years): uncomplicated pH1N1 illness, pH1N1 illness-related hospitalization, severe allergic reaction to the pH1N1 vaccine, and Guillain-Barré syndrome. We calculated descriptive statistics for time trade-off amounts and derived quality adjusted life year losses for these events. Multivariate regression analyses evaluated the effect of scenario age, as well as respondent socio-demographic and health characteristics on time trade-off amounts.
Respondents were willing to trade more time to avoid the more severe outcomes, hospitalization and Guillain-Barré syndrome. In our adjusted and unadjusted analyses, age of the patient in the scenario was significantly associated with time trade-off amounts (p-value<0.05), with respondents willing to trade more time to prevent outcomes in children versus adults. Persons who had received the pH1N1 vaccination were willing to trade significantly more time to avoid hospitalization, severe allergic reaction, and Guillain-Barré syndrome, controlling for other variables in adjusted analyses.(p-value<0.05)
Community members placed the highest value on preventing outcomes in children, compared with adults, and the time trade-off values reported were consistent with the severity of the outcomes presented. Considering these public values along with other decision-making factors may help policy makers improve the allocation of pandemic vaccine resources.
Citation: Lavelle TA, Meltzer MI, Gebremariam A, Lamarand K, Fiore AE, Prosser LA (2011) Community-Based Values for 2009 Pandemic Influenza A H1N1 Illnesses and Vaccination-Related Adverse Events. PLoS ONE 6(12): e27777. https://doi.org/10.1371/journal.pone.0027777
Editor: Vittoria Colizza, INSERM and Université Pierre et Marie Curie, France
Received: July 15, 2011; Accepted: October 25, 2011; Published: December 19, 2011
This is an open-access article, free of all copyright, and may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose. The work is made available under the Creative Commons CC0 public domain dedication.
Funding: This study was funded by the Centers for Disease Control and Prevention. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
In April 2009, the first influenza pandemic in over forty years began in North America; the causative virus was 2009 pandemic influenza (A) H1N1 (pH1N1). Under guidance from the Advisory Committee for Immunization Practices (ACIP), the Centers for Disease Control and Prevention recommended target groups for vaccination . A vaccine became available during October 2009, and a program was implemented on an emergency basis to reduce the impact of the expanding pandemic.
Vaccination programs, such as the one implemented for pH1N1, involve an inherent trade-off of risks. Vaccinating for a particular disease reduces the risk of infectious illness, but introduces new risks of vaccine-related adverse events. The acceptability of a vaccination program depends in part on how the public values the potential risks and benefits of vaccination. By examining the likelihood of these risks and benefits, as well the value of prevention, decision makers can determine the potential value of a public vaccination program. When pH1N1 vaccine recommendations were made in the U.S. the only studies reporting community values associated with influenza illness and vaccination were based on data from seasonal influenza , , . Outcomes related to pH1N1 illness and vaccination may be valued differently, however. We present in this study estimates of community-based values for avoiding adult and pediatric health events related to pH1N1 illness and vaccination.
This study was reviewed and provided with exempt status by the University of Michigan institutional review board. All study data were de-identified; no informed consent was required by the board in order for individuals to participate in the study.
We used the time trade off (TTO) approach to evaluate community-based values for avoiding pH1N1 illness and vaccination-related adverse events in adults and children. The TTO method estimates the value each respondent puts on avoiding a particular health outcome by estimating their willingness to trade quantity of life for quality of life. For example, a TTO question may value diabetes prevention by measuring the amount of time a person would be willing to give up from her life span to avoid living with diabetes (living instead a reduced number of years without diabetes). The resulting TTO values can be interpreted as subjective measures of quality of life, and are the basis for constructing quality adjusted life years (QALYs). QALYs are created by weighting a segment of time spent in a specific health state by the quality of life value associated with that health state. QALYs have been used to measure the morbidity associated with chronic illness over an extended time period . In our study, to value the morbidity associated with the health states of pH1N1 illness and vaccination-related adverse events, we used TTO responses from our survey to calculate short-term QALY losses.
We randomly sampled adult community members to complete an internet survey from a research panel designed to be statistically representative of the U.S. general adult population. The survey was administered by Knowledge Networks (Menlo Park, CA), which currently recruits new research panel members by mail from a published address-based sample frame that covers approximately 98% of U.S. households . Non-internet households who choose to join the panel are provided with internet access and a laptop computer. Households who use their own computer and internet service to answer online surveys administered by Knowledge Networks receive small monthly stipends in exchange for their participation . Demographic information collected for all new panel members includes gender, age, ages of their household members, race/ethnicity, income, and education level.
Participation in the study required completion of a 15-minute survey during January 2010. Respondents answered a series of TTO questions to value hypothetical health state scenarios describing: uncomplicated pH1N1 illness, pH1N1 illness-related hospitalization, severe allergic reaction to the pH1N1 vaccine, and Guillain-Barré syndrome, a potential vaccine-related adverse event. Each of the 4 health state scenarios had 4 versions; each referencing a hypothetical person aged 1, 8, 35, and 70 years. Respondents were randomly assigned to value 2 different ages for each of the 4 scenarios, for a total of 8 TTO questions. The different age versions of each health state scenario were identical except for the description of usual activities, which included school/daycare for children and work/household responsibilities for adults. We instructed respondents to imagine a family member or friend that closely matched the age description in the scenario at hand. Respondents were also asked whether they had been vaccinated for pH1N1 or seasonal influenza, and whether they or anyone else in their family had ever experienced pH1N1 or seasonal influenza illness or an influenza vaccination-related adverse event.
We used a modified bidding algorithm, combining binary and open ended response questions, to measure TTO amounts. This method is less prone to non-response problems compared to a single open ended question . After presenting one age-specific health event related to pH1N1 illness or vaccination, we first asked respondents whether they would trade a fixed amount of time from the end of their life in exchange for avoiding the health event. (Figure 1) The amount of time that the respondents were asked to trade was randomized to reduce anchoring bias, with initial TTO amounts ranging from 2 days to 2 months for uncomplicated flu and severe allergic reaction outcomes, and 2 weeks to 1 year for hospitalization and Guillain-Barré syndrome outcomes. A follow up binary question offered a higher TTO amount if the initial response was “yes,” and a lower TTO amount if the initial response was “no.” These two binary questions were followed by an open-ended question which asked respondents for the maximum amount of time they would trade from the end of their life (in days, weeks, months, and years) to avoid the health state in question; this maximum TTO value was used for all analyses.
We calculated descriptive statistics for TTO data, including means, medians, 5th and 95th percentiles, minimums and maximums. Confidence intervals around mean values were estimated using bootstrapping with replacement procedures . We used the Kruskal-Wallis non-parametric test in unadjusted analyses to evaluate whether median values differed by scenario age. All summary statistics used unweighted data, due to the similarity between unweighted and weighted summary statistics. In our primary analysis, TTO amounts greater than life expectancy were reset to equal the respondent's life expectancy, and we evaluated the effect of this in sensitivity analyses.
As respondents were asked their willingness to trade time from the end of their life, we adjusted for the potential impact of time preference by using a 3% discount rate to calculate discounted TTO values , . Dividing the respondents' discounted TTO amount by their discounted life expectancy allowed us to calculate a short term QALY loss associated with the temporary health state in question.
To evaluate the association between TTO amounts and respondent/scenario characteristics, we used a generalized estimating equation negative binomial regression model. This type of regression model is bounded at 0 to account for the lower limit of TTO responses and adjusts for the correlations associated with multiple evaluations per respondent . Using the undiscounted TTO amounts reported for the four different health states as the dependent variables, the four final regression models each included as independent variables: scenario age, gender, respondent age, education, race/ethnicity, having a child under the age of 18, vaccination status, and experience with the health state in question. The goodness of fit of each model was measured using a test of concordance between the observed and predicted TTO values .
The survey was sent to 1,110 members of the survey panel. Of those invited by email to participate in the online survey, 65% completed the survey (n = 718); 9% of respondents were eliminated from the primary analysis due to missing or invalid responses, leaving a final analysis sample size of n = 659. Observations were excluded from the analysis if TTO amounts were missing for more than half (4 or more) of the scenarios (n = 56), the responses in all four time metrics were equivalent (n = 2), or the TTO amount was nonsensical (e.g., 999999 months) (n = 1).
Demographic characteristics among those who responded to the survey were statistically different from those who did not respond to the survey for all demographic characteristics except location (country region and metropolitan status). Compared to non-responders, responders were more likely to be male, white, married, aged 45 or older, college educated, and earn more than $35,000 annually; respondents were also less likely to have a child under the age of 18 years. (p<0.05 for all)
Respondent characteristics included in the primary analysis are summarized in Tables 1, and 2. Without survey weights, 50% of respondents were male, 56% were married, 78% were white, non-Hispanic, 33% had a child under the age of 18 living at home, and 84% rated themselves in excellent/very good or good overall health. Forty two percent of all respondents had received the seasonal flu vaccine in the previous 12 months, and 21% had received the pH1N1 vaccine in this time period. Thirty one percent of respondents reported that they had experienced seasonal influenza themselves, and 18% had a family member who had experienced this illness at some point in the past. Three percent of respondents had experienced pH1N1 illness themselves, and 5% had a family member who had experienced this illness. Only a small fraction of respondents (<1–2%, depending on question) reported that they had experienced a hospitalization related to influenza, or a side effect from an influenza vaccine, either personally or through a family member.
Respondents were willing to trade a median of 7 undiscounted days to avoid a hospitalization related to pH1N1 influenza and 30 days to avoid Guillain-Barré syndrome, compared to a median of 2 and 4 undiscounted days to avoid uncomplicated pH1N1 illness and severe allergic reaction, respectively (Table 3). Due to the right skewed distribution for TTO amounts in all 4 health states (unsymmetrical, with the greatest proportion of respondents willing to trade 0 days), mean values were substantially higher and more variable than median values. Respondents were willing to trade a mean of 291 and 376 undiscounted days to avoid a hospitalization and Guillain-Barré syndrome, and a mean of 226 and 222 undiscounted days to avoid uncomplicated pH1N1 illness and a severe allergic reaction, respectively. (Table 3)
When stratified by scenario age within each health state, median TTO amounts differed significantly by age (p-value<0.05 for all health states). (Table 3) On average, respondents were willing to trade more time to avoid pH1N1-related illnesses and vaccination-related adverse events in children, compared to adults. Respondents were willing to trade a median of 3 and 14 undiscounted days to avoid pH1N1 illness and hospitalization in a 1 year old child, but were only willing to trade a median of 2 and 7 days to avoid these same outcomes in a 70 year old adult. Likewise, respondents were willing to trade a median of 7 and 60 undiscounted days to prevent a severe vaccine allergic reaction and Guillain-Barré syndrome in a 1 year old child, but were only willing to trade a median of 2 and 28 days, to avoid these same outcomes in a 70 year old adult. (Table 3)
The median values for the loss in QALYs from pH1N1 illness and vaccination-related adverse events also exhibited a significant difference by scenario age (p-value<0.05 for all health states). (Table 4) For example, pH1N1-related hospitalization was associated with a 0.0007 median QALY loss for a 1 year old and a 0.0003 median QALY loss for a 70 year old. Likewise, Guillain-Barré syndrome was associated with a 0.0039 median QALY loss for a 1 year old and a 0.0012 median QALY loss for a 70 year old. Mean values were consistently higher and more variable than median values. (Table 4)
After adjusting for respondent characteristics, the 1 year and 8 year old scenario ages were significantly associated with greater TTO amounts (compared with the 35 year old scenario age) in all four final regression models. (p-values<0.05, Table 5) Seventy year old scenario age was significantly associated with lower TTO amounts (compared to the 35 year scenario age) in the final regression models for hospitalization and Guillain-Barré syndrome outcomes.
For all four health states, having less than a college degree was significantly associated with greater TTO amounts. (Table 5) Other demographic characteristic associations were not consistent across outcomes, however. Compared with a white, non-Hispanic reference group, being Hispanic or black, non-Hispanic, was significantly associated with greater TTO amounts for uncomplicated pH1N1 illness and allergic reaction only. Being over the age of 30 was significantly associated with greater TTO amounts for uncomplicated pH1N1 illness and Guillain-Barré syndrome only.
Respondent health characteristic associations were also inconsistent predictors of TTO amounts. Experience with uncomplicated pH1N1 illness and Guillain-Barré syndrome was significantly associated with greater TTO amounts for those respective health states, but experience with pH1N1-related hospitalization and severe allergic reaction was not significantly associated with the TTO amounts for these outcomes. Compared to respondents that had not been vaccinated for pH1N1, those that had been vaccinated were willing to trade significantly more time to avoid a pH1N1-related hospitalization (p-value = 0.03) but were also willing to trade more time to avoid both vaccination related adverse events. (p-value<0.05 for both)
Respondent's gender, and having a child under 18, did not significantly impact TTO responses. Concordance coefficients, used to measure the goodness of fit of our models, ranged from 0.071 to 0.129. All coefficients were significantly greater than zero, indicating that there was a significant and positive correlation between our observed and predicted TTO values. (Table 5) Sensitivity analyses which excluded respondents who traded amounts larger than life expectancy yielded very similar results to the primary analysis, which included these respondents with their TTO amounts reset to their life expectancy (results not shown).
This study reports community values for avoiding pH1N1 illness-related outcomes and vaccination-related adverse events in the U.S. On average, respondents' values for avoiding pH1N1-related health events and vaccination-related adverse events were aligned with the portrayed severity of these events in our survey. Compared to pH1N1 illness-related hospitalization, respondents were willing to trade less time to avoid uncomplicated pH1N1 illness and a severe allergic reaction from vaccination, across all scenario ages. Respondents were willing to trade the greatest amount of time to avoid the most severe outcome, Guillain-Barré syndrome. This relative ranking of these TTO values across outcomes is consistent with previous findings for outcomes associated with seasonal influenza illness and vaccine related adverse events . In regression analyses, 1 year and 8 year old scenario ages were consistently associated with greater TTO amounts, indicating that the public may give preference to preventing pH1N1 illness and vaccine- related health outcomes in children compared with adults. These data are consistent with earlier findings that indicate that community members may prefer to prioritize child health , , .
These findings are also consistent with the ACIP's recommendations in July of 2009 which stated that children and young adults aged 6 months–25 years should be among those prioritized for pH1N1 vaccination, and that children 6 month–4 years should be one of the groups prioritized under a scenario of limited vaccine supply . These recommendations were made based on data of disease prevalence and risk of complications, and some limited data from community engagement exercises performed as part of pandemic preparedness . Also considering these new preference data obtained from community members after the recent pH1N1 influenza pandemic may help policy makers better define key target groups to prioritize for vaccination during the next influenza pandemic.
Our analysis also indicates that certain characteristics of community members may be significant predictors of health state valuations. In adjusted analyses, we found that respondents with less than a bachelor's degree were willing to trade significantly more time than those with a higher level of education to avoid all four health states, controlling for other variables in a multivariate regression. This finding is not consistent with values elicited for seasonal influenza, and may represent a finding that is important to note in light of the novel nature of pH1N1 compared to seasonal influenza . Hispanic and black, non-Hispanic respondents were also willing to trade significantly more time than white respondents to avoid uncomplicated pH1N1 illness and severe allergic reaction. This statistical association between respondent race and health state valuation is consistent with values elicited from community members for seasonal influenza and other health states , . Although no consensus exists regarding the cause of the association, one possible explanation is that respondents without a usual source of care may demonstrate a stronger preference to avoid illness. Previous research has shown that compared to white individuals, Hispanic and black individuals are less likely to have a doctor's office as their usual source of care, regardless of insurance coverage, family income and geographic region , . As we did not measure usual source of care, it is possible that this variable confounded the race association found in our analysis. Future research should assess respondents' usual source of care and parse out its contribution, along with race and other factors, to health state preferences.
An important limitation of this study is that we used a stated preference approach to value health states. These stated preferences may not reflect the actual choices that these respondents may make when faced with a choice between accepting or rejecting vaccination. In addition, we used the TTO approach for valuing health states, but other methods may have produced different results . As with most vignettes used to estimate preferences, the scenarios used in our survey were concise descriptions of complex health events; adding additional dimensions of health to these vignettes may have influenced respondents' valuations .
We also do not know the generalizability of these results. Our measurement of public values for health states related only to this influenza pandemic, and may not relate to more severe influenza pandemics. Another limitation is that both the timing of our survey and the representativeness of the sample may not have been optimal for determining truly representative public values. The survey was fielded after the fall epidemic had passed and the vaccination program had been initiated, and so may not reflect the important public values that were relevant during the time that vaccination program decisions were being made. Data have shown that the public's concern about getting sick from pH1N1 as well as their concern about the safety risks associated with vaccination declined over the duration of the epidemic . Also, compared to non-respondents, our respondents were more likely to be college educated, married, white, older males, and thus may have reported values different from a more population representative sample.
In this study we measured values for health outcomes related to pH1N1 illness and vaccination from the general U.S. public, and not specifically from those that have experienced pH1N1 illness. Previous studies have found that compared to a sample of persons who have not experienced an ill health state, those who have experienced it are typically willing to trade less time to avoid the illness , , . Many of these studies, however, have focused on chronic illnesses, and there is limited evidence as to how experience or familiarity with a short term health state may influence preferences for avoiding these health outcomes. Van Hoek, et al. estimated a 0.008 QALY loss attributable to pH1N1 in a sample of confirmed pH1N1 cases using the EQ-5D questionnaire. This QALY loss among those who have experienced pH1N1 illness is difficult to compare to our results, however, because it averages over a sample of confirmed cases with and without complications . In our adjusted analyses, we found that those who experienced uncomplicated pH1N1 illness or Guillain-Barré syndrome were willing to trade significantly more time to avoid these health states compared with those without experience.(p-value<0.05 for both) More research is needed to determine if such differences can be measured among other experienced temporary health states.
Our findings suggest that the community-based values for avoiding health events related to pH1N1 illness and vaccination are consistent with the severity of the outcomes. These data also suggest that the public places a greater value on preventing outcomes in children, compared to adults, consistent with previous findings from seasonal influenza. The valuations derived from these data can be used along with other decision-making factors during the development of pandemic influenza vaccination programs in the U.S. and the allocation of future pandemic vaccine supplies.
The views expressed are those of the authors and do not necessarily reflect the official position of the Centers for Disease Control and Prevention.
Conceived and designed the experiments: TAL LAP MIM AEF. Analyzed the data: AG. Wrote the paper: TAL LAP MIM AEF KL.
- 1. National Center for Immunization and Respiratory Diseases (2009) Use of influenza A (H1N1) 2009 monovalent vaccine: recommendations of the Advisory Committee on Immunization Practices (ACIP), 2009. MMWR Recomm Rep 58: 1–8.National Center for Immunization and Respiratory Diseases2009Use of influenza A (H1N1) 2009 monovalent vaccine: recommendations of the Advisory Committee on Immunization Practices (ACIP), 2009.MMWR Recomm Rep5818
- 2. Prosser LA, Bridges CB, Uyeki TM, Rego VH, Ray GT, et al. (2005) Values for preventing influenza-related morbidity and vaccine adverse events in children. Health Qual Life Outcomes 3: 18.LA ProsserCB BridgesTM UyekiVH RegoGT Ray2005Values for preventing influenza-related morbidity and vaccine adverse events in children.Health Qual Life Outcomes318
- 3. Prosser LA, Payne K, Rusinak D, Shi P, Uyeki T, et al. (2011) Valuing health across the lifespan: health state preferences for seasonal influenza illnesses in patients of different ages. Value Health 14: 135–143.LA ProsserK. PayneD. RusinakP. ShiT. Uyeki2011Valuing health across the lifespan: health state preferences for seasonal influenza illnesses in patients of different ages.Value Health14135143
- 4. Johnston SS, Rousculp MD, Palmer LA, Chu BC, Mahadevia PJ, et al. (2010) Employees' willingness to pay to prevent influenza. Am J Manag Care 16: e205–214.SS JohnstonMD RousculpLA PalmerBC ChuPJ Mahadevia2010Employees' willingness to pay to prevent influenza.Am J Manag Care16e205214
- 5. Gold MR (1996) Cost-effectiveness in health and medicine. New York: Oxford University Press. MR Gold1996Cost-effectiveness in health and medicineNew YorkOxford University Pressxxiii425
- 6. Dennis JM (2010) KnowledgePanel® Design Summary: KNOWLEDGEPANEL® OVERVIEW. Knowledge Networks. JM Dennis2010KnowledgePanel® Design Summary: KNOWLEDGEPANEL® OVERVIEW. Knowledge Networks.[cited 2011 February 8]; Available: http://www.knowledgenetworks.com/knpanel/docs/KnowledgePanel(R)-Design-Summary-Description.pdf. [cited 2011 February 8]; Available: http://www.knowledgenetworks.com/knpanel/docs/KnowledgePanel(R)-Design-Summary-Description.pdf.
- 7. (2009) RESPONDENT INCENTIVES FOR KNOWLEDGEPANEL®. 2009RESPONDENT INCENTIVES FOR KNOWLEDGEPANEL®.[cited 2011 February 8]; Available: http://www.knowledgenetworks.com/ganp/irbsupport/docs/KN%20IRB%20Doc%20-%20Section%204%20-%20Respondent%20Incentives.doc. [cited 2011 February 8]; Available: http://www.knowledgenetworks.com/ganp/irbsupport/docs/KN%20IRB%20Doc%20-%20Section%204%20-%20Respondent%20Incentives.doc.
- 8. Johannesson M (1996) Theory and methods of economic evaluation of health care. Dordrecht; Boston: Kluwer. M. Johannesson1996Theory and methods of economic evaluation of health careDordrecht; BostonKluwerx245
- 9. Efron B (1987) Better Bootstrap Confidence-Intervals. Journal of the American Statistical Association 82: 171–185.B. Efron1987Better Bootstrap Confidence-Intervals.Journal of the American Statistical Association82171185
- 10. Attema AE, Brouwer WB (2010) The value of correcting values: influence and importance of correcting TTO scores for time preference. Value Health 13: 879–884.AE AttemaWB Brouwer2010The value of correcting values: influence and importance of correcting TTO scores for time preference.Value Health13879884
- 11. Hanley JA, Negassa A, Edwardes MD, Forrester JE (2003) Statistical analysis of correlated data using generalized estimating equations: an orientation. Am J Epidemiol 157: 364–375.JA HanleyA. NegassaMD EdwardesJE Forrester2003Statistical analysis of correlated data using generalized estimating equations: an orientation.Am J Epidemiol157364375
- 12. Lin LI (1989) A concordance correlation coefficient to evaluate reproducibility. Biometrics 45: 255–268.LI Lin1989A concordance correlation coefficient to evaluate reproducibility.Biometrics45255268
- 13. Eisenberg D, Freed GL (2007) Reassessing how society prioritizes the health of young people. Health Aff (Millwood) 26: 345–354.D. EisenbergGL Freed2007Reassessing how society prioritizes the health of young people.Health Aff (Millwood)26345354
- 14. Eisenberg D, Freed GL, Davis MM, Singer D, Prosser LA (2011) Valuing health at different ages: evidence from a nationally representative survey in the US. Appl Health Econ Health Policy 9: 149–156.D. EisenbergGL FreedMM DavisD. SingerLA Prosser2011Valuing health at different ages: evidence from a nationally representative survey in the US.Appl Health Econ Health Policy9149156
- 15. U.S. Department of Health and Human Services USDoHS, editor (2008) Guidance on Allocating and Targeting Pandemic Influenza Vaccine. U.S. Department of Health and Human Services USDoHS, editor2008Guidance on Allocating and Targeting Pandemic Influenza Vaccine.[cited 2011 March 17]; Available: http://www.flu.gov/individualfamily/vaccination/allocationguidance.pdf. [cited 2011 March 17]; Available: http://www.flu.gov/individualfamily/vaccination/allocationguidance.pdf.
- 16. Wittenberg E, Halpern E, Divi N, Prosser LA, Araki SS, et al. (2006) The effect of age, race and gender on preference scores for hypothetical health states. Qual Life Res 15: 645–653.E. WittenbergE. HalpernN. DiviLA ProsserSS Araki2006The effect of age, race and gender on preference scores for hypothetical health states.Qual Life Res15645653
- 17. Gaskin DJ, Arbelaez JJ, Brown JR, Petras H, Wagner FA, et al. (2007) Examining racial and ethnic disparities in site of usual source of care. J Natl Med Assoc 99: 22–30.DJ GaskinJJ ArbelaezJR BrownH. PetrasFA Wagner2007Examining racial and ethnic disparities in site of usual source of care.J Natl Med Assoc992230
- 18. Lillie-Blanton M, Martinez RM, Salganicoff A (2001) Site of medical care: do racial and ethnic differences persist? Yale J Health Policy Law Ethics 1: 15–32.M. Lillie-BlantonRM MartinezA. Salganicoff2001Site of medical care: do racial and ethnic differences persist?Yale J Health Policy Law Ethics11532
- 19. Weinstein MC, Torrance G, McGuire A (2009) QALYs: the basics. Value Health 12: Suppl 1S5–9.MC WeinsteinG. TorranceA. McGuire2009QALYs: the basics.Value Health12Suppl 1S59
- 20. Brazier J, Rowen D, Tsuchiya A, Yang Y, Young TA (2011) The impact of adding an extra dimension to a preference-based measure. Soc Sci Med 73: 245–253.J. BrazierD. RowenA. TsuchiyaY. YangTA Young2011The impact of adding an extra dimension to a preference-based measure.Soc Sci Med73245253
- 21. SteelFisher GK, Blendon RJ, Bekheit MM, Lubell K (2010) The public's response to the 2009 H1N1 influenza pandemic. N Engl J Med 362: e65.GK SteelFisherRJ BlendonMM BekheitK. Lubell2010The public's response to the 2009 H1N1 influenza pandemic.N Engl J Med362e65
- 22. Prosser LA, Kuntz KM, Bar-Or A, Weinstein MC (2003) Patient and community preferences for treatments and health states in multiple sclerosis. Mult Scler 9: 311–319.LA ProsserKM KuntzA. Bar-OrMC Weinstein2003Patient and community preferences for treatments and health states in multiple sclerosis.Mult Scler9311319
- 23. Ubel PA, Loewenstein G, Jepson C (2003) Whose quality of life? A commentary exploring discrepancies between health state evaluations of patients and the general public. Qual Life Res 12: 599–607.PA UbelG. LoewensteinC. Jepson2003Whose quality of life? A commentary exploring discrepancies between health state evaluations of patients and the general public.Qual Life Res12599607
- 24. Baron J, Asch DA, Fagerlin A, Jepson C, Loewenstein G, et al. (2003) Effect of assessment method on the discrepancy between judgments of health disorders people have and do not have: a web study. Med Decis Making 23: 422–434.J. BaronDA AschA. FagerlinC. JepsonG. Loewenstein2003Effect of assessment method on the discrepancy between judgments of health disorders people have and do not have: a web study.Med Decis Making23422434
- 25. van Hoek AJ, Underwood A, Jit M, Miller E, Edmunds WJ (2011) The Impact of Pandemic Influenza H1N1 on Health-Related Quality of Life: A Prospective Population-Based Study. PLoS One 6: e17030.AJ van HoekA. UnderwoodM. JitE. MillerWJ Edmunds2011The Impact of Pandemic Influenza H1N1 on Health-Related Quality of Life: A Prospective Population-Based Study.PLoS One6e17030