Abstract
Purpose:
To evaluate a general question about ability to meet monthly bills as an alternative to direct questions about income and assets in health utility studies.
Methods:
We used data from the National Health Measurement Study–a U.S. nationally representative telephone survey collected in 2005-2006. It included health utility measures (EuroQol-5D-3L, Health Utilities Index Mark 3, Short Form-6D, and Quality of Well-being Index) and household income, assets, and financial ability to meet monthly bills questions. Each utility score was regressed on: income and assets (MModel 1); difficulty paying bills (DPB) (MModel 2); income, assets, and DPB (MModel 3). All models used survey weights and adjusted for demographics and education.
Results:
Among 3666 respondents, as income and assets increased, DPB decreased. The DPB question had fewer missing values (n=30) than income (n=311) or assets (n=373). Model 2 (DPB only) explained more variance in health utility than Model 1 (income and assets only). Including all measures (Model 3) had very modest improvement in R-squared: e.g., values were 0.112 (Model 1), 0.166 (Model 2), and 0.175 (Model 3) for EuroQol-5D-3L.
Conclusions:
The single question on DPB yields more information and has less missing values than the traditionally used income and assets questions.
Keywords: health-related quality of life, health utility, socioeconomic status, income, assets, financial ability
Introduction
Health-related quality of life (HRQoL) measures in both population and clinical studies are often adjusted for socioeconomic status (SES). SES has been shown to be associated with HRQoL in both cross-sectional [1-4] and longitudinal analyses [5] after controlling for health conditions and demographic information (age, sex, race/ethnicity). However, SES questions about income and assets often have relatively high rates of missing data and there are concerns about inaccuracies in the data that are collected because of the invasiveness of these questions [6-9]. Obtaining income and asset information in interviewer administered surveys can also require long and complicated skip patterns.
From a sociological perspective, accurate and direct measures of income and assets are important. From the perspective of health researchers who would like to adjust for socioeconomic factors while examining other relationships, there may be simpler and less invasive ways to capture important SES information for adjustment purposes. In this study, we compare a single question about difficulty meeting monthly bills to direct questions about assets and income. This question has previously been used to evaluate the association between financial strain and psychological well-being and chronic conditions, but not HRQoL [10-12].
Methods
Subjects
We used data from the National Health Measurement Study, a random-digit–dial telephone survey of 3844 noninstitutionalized U.S. adults in the 48 contiguous states, aged 35 to89 years [1]. Data were collected between June 2005 and August 2006, with a simple response rate of 46%. We excluded 181 respondents (4.7%) reporting their race as neither white nor black to build off of previously published results [2].
The study over-sampled black Americans and persons over age 65 using standard survey sampling methods. Sampling and post-stratification weights were available to adjust results to the U.S. Census 2000 population, [1] so results are generalizable to the U.S. noninstitutionalized adult population aged 35–89 years in the 48 contiguous states for people reporting their race as white or black.
Dependent Variables
We used health utility scores as our dependent variables, which are constructed such that 0 is “dead” and 1.0 is “full health.” We used four of the most commonly used health utility scores: 1) The EuroQol-5D-3L (EQ-5D-3L) [13] with U.S. scoring [14]; 2) Health Utilities Index Mark 3 (HUI3) [15]; 3) SF-6D calculated from the SF-36v2tm [16]; and 4) Quality of Well-being Scale (QWB-SA) [17,18].
Independent Variables
The primary variables of interest were household income, household assets, and difficulty paying monthly bills. Household income for the previous year was grouped into four categories: <$20,000, $20,000–34,999, $35,000–74,999, and $75,000 or more. Household assets were measured by asking: “Suppose you needed money quickly, and you (and your spouse) cashed in all of your checking and savings accounts, any stocks and bonds, and real estate other than your home. If you added up what you got, would this be … (provided dollar ranges)?” We coded household assets as <$25,000, $25,000–99,999, and $100,000 or more. Difficulty paying monthly bills (DPB) was measured with a single question, “How difficult is it for you to meet the monthly payments on your bills? Extremely difficult, very difficult, somewhat difficult, slightly difficult, or not difficult at all?” These were collapsed into three categories: Not difficult at all, slightly or somewhat difficult, and very or extremely difficult.
Other self-reported independent variables include sex, race (Caucasian/white [white] or black/African American [black]), age, and education. Eleven self-reported chronic health conditions were used in sensitivity analyses [19].
Analyses
We conducted multivariate weighted least squares (WLS) regressions to examine the associations between health utility and household finances while adjusted for age, sex, race, and education. Household finances were assessed as household assets and income (Model 1), DPB (Model 2), or household assets and income and DPB (Model 3). Analyses were performed using SAS System for Windows (Version 9.3) procedures that incorporate survey weights, PROC SURVEYFREQ, and PROC SURVEYREG (SAS Institute Inc.).
Sensitivity analyses also included eleven separate chronic health conditions in models 1, 2, and 3.
Results
We used 3666 respondents who reported black or white race. Weighted distributions within this sample were 53.6% female and 11.5% black. Age groups were 35-44 (30.8%), 45-54 (23.8%), 55-64 (20.4%), 65-74 (14.4%), and ≥75 (10.7%). Highest educational attainment was either less than high school (8.0%), high school graduate (28.7%), some college (22.3%), or college graduate (41.0%).
The majority of people reported no difficulty paying their bills (59%) while a small but significant proportion reported meeting their monthly bills was very difficult (4.2%) or extremely difficult (3.1%). As household income and assets increased, reported difficulty paying bills decreased (Tables 1-2; P<0.0001 for both). For example, those in the lowest income category were evenly distributed across reported difficulty paying bills while 75% of those in the highest income category reported no difficulty paying monthly bills.
Table 1:
Weighteda cross tabulation of household income category and difficulty meeting monthly bills in black and white adults ≥35 years old from the 2005/2006 National Health Measurement Study (n = 3666).a
| Household Income Category |
Difficulty Meeting Monthly Bills | |||||
|---|---|---|---|---|---|---|
| Extremely Difficult n (column %) |
Very Difficult n (column %) |
Somewhat Difficult n (column %) |
Slightly Difficult n (column %) |
Not Difficult At All n (column %) |
Total N |
|
| <$10,000 | 21 (21) | 23 (16) | 25 (5) | 15 (2) | 27 (1) | 111 |
| $10,000-$14,999 | 19 (19) | 18 (13) | 29 (6) | 14 (2) | 21 (1) | 101 |
| $15,000-$19,000 | 17 (17) | 13 (9) | 47 (10) | 27 (4) | 52 (3) | 157 |
| $20,000-$24,999 | 14 (14) | 20 (14) | 24 (5) | 47 (7) | 63 (3) | 168 |
| $25,000-$34,999 | 20 (20) | 19 (14) | 87 (19) | 65 (10) | 145 (7) | 337 |
| $35,000-$49,999 | 3 (3) | 14 (10) | 81 (17) | 103 (16) | 291 (15) | 493 |
| $50,000-$74,999 | 7 (7) | 26 (19) | 93 (20) | 165 (25) | 435 (22) | 726 |
| ≥$75,000 | 2 (2) | 7 (5) | 82 (18) | 224 (34) | 939 (48) | 1252 |
| Total | 102 | 140 | 468 | 661 | 1973 | 3344 |
sampling and post-stratification weights to adjust results to the U.S. adult population.
all weighted counts were rounded.
Table 2:
Weighteda cross tabulation of asset category and difficulty meeting monthly bills in black and white adults ≥35 years old from the 2005/2006 National Health Measurement Study (n = 3666).b
| Asset Category | Difficulty Meeting Monthly Bills | |||||
|---|---|---|---|---|---|---|
| Extremely Difficult (column %) |
Very Difficult (column %) |
Somewhat Difficult (column %) |
Slightly Difficult (column %) |
Not Difficult at All (column %) |
Total | |
| <$10,000 | 87 (80) | 94 (67) | 177 (39) | 131 (20) | 129 (7) | 618 |
| $10,000-$24,999 | 9 (8) | 19 (13) | 60 (13) | 126 (20) | 228 (12) | 442 |
| $25,000-$49,999 | 8 (7) | 17 (12) | 89 (20) | 119 (19) | 240 (13) | 474 |
| $50,000-$74,999 | 5 (5) | 3 (2) | 31 (7) | 75 (12) | 171 (9) | 285 |
| $75,000-$99,000 | 0 (0) | 3 (2) | 27 (6) | 44 (7) | 186 (10) | 260 |
| ≥$100,000 | 0 (0) | 5 (4) | 64 (14) | 148 (23) | 944 (50) | 1161 |
| Total | 109 | 141 | 449 | 642 | 1899 | 3239 |
sampling and post-stratification weights to adjust results to the U.S. adult population.
all weighted counts were rounded.
All respondents had age and sex information as this information was necessary for respondent selection. Education had rare missing information (n=21). The number of respondents with missing health utility values ranged from 29 (EQ-5D-3L) to 253 (HUI3). The DPB question had fewer missing values (n=30) than income categories (n=311) or asset categories (n=373).
Table 3 includes results from the multivariate analyses. Models with DPB (Model 2) explained more variance in health utility than models with absolute measures (Model 1) in all utility measures. Including all variables (Model 3) had very modest increases in R-square over the model with only DPB (Model 2). For example, the EQ-5D-3L R-squared values were 0.112 (Model 1), 0.166 (Model 2), and 0.175 (Model 3). In most of the models 1, the lowest income categories were significant predictors of health utility; these variables became insignificant or the parameter estimates became smaller with the inclusion of DPB in Model 3. In contrast, DPB was a significant predictor of health utility in all measures alone (Model 2) and with assets and income (Model 3).
Table 3.
Health utility measures regressed (weighted least squares) on three models in black and white adults ≥35 years old from the 2005/2006 National Health Measurement Study (n = 3666). Reference categories are “male,” “age 35-44,” “white,” “college graduate,” “income ≥$75,000,” “assets ≥$100,000,” and “bill payments somewhat or slightly difficult.”
| EQ-5D-3L | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Model 1 | Model 2 | Model 3 | |||||||
| β | SE | P-value | β | SE | P-value | β | SE | P-value | |
| Intercept | 0.954 | 0.010 | <.0001 | 0.926 | 0.010 | <.0001 | 0.932 | 0.012 | <.0001 |
| Female | −0.009 | 0.008 | 0.32 | −0.012 | 0.007 | 0.097 | −0.011 | 0.007 | 0.14 |
| Age 45-54 | −0.025 | 0.011 | 0.73 | −0.024 | 0.010 | 0.020 | −0.025 | 0.010 | 0.016 |
| Age 55-64 | −0.053 | 0.012 | 0.0005 | −0.060 | 0.011 | <.0001 | −0.057 | 0.011 | <.0001 |
| Age 65-74 | −0.022 | 0.013 | 0.09 | −0.041 | 0.011 | <.0001 | −0.032 | 0.012 | 0.0068 |
| Age 75-89 | −0.036 | 0.014 | <.0001 | −0.070 | 0.010 | <.0001 | −0.056 | 0.012 | <.0001 |
| Black | −0.031 | 0.012 | 0.043 | −0.028 | 0.011 | 0.011 | −0.023 | 0.011 | 0.042 |
| Less than high school education | −0.063 | 0.018 | <.0001 | −0.069 | 0.016 | <.0001 | −0.056 | 0.017 | 0.0008 |
| High school educations | −0.026 | 0.009 | 0.036 | −0.032 | 0.008 | 0.0001 | −0.025 | 0.009 | 0.0038 |
| Some college education | −0.031 | 0.011 | 0.0092 | −0.032 | 0.010 | 0.0014 | −0.029 | 0.010 | 0.0038 |
| Income <$20,000 | −0.087 | 0.018 | <.0001 | −0.053 | 0.016 | 0.0007 | |||
| Income $20,000 to $34,999 | −0.043 | 0.015 | 0.0045 | −0.017 | 0.013 | 0.19 | |||
| Income $35,000 to $74,999 | −0.018 | 0.009 | 0.039 | −0.012 | 0.009 | 0.18 | |||
| Assets <$25,000 | 0.016 | 0.018 | 0.38 | 0.010 | 0.010 | 0.29 | |||
| Assets $25,000 to $99,999 | 0.008 | 0.015 | 0.60 | −0.009 | 0.008 | 0.27 | |||
| Bill Payments not difficult | 0.031 | 0.007 | <.0001 | 0.028 | 0.008 | 0.0002 | |||
| Bill payments very difficult | −0.140 | 0.022 | <.0001 | −0.131 | 0.021 | <.0001 | |||
| R-sq | 0.112 | 0.166 | 0.175 | ||||||
| HUI3 | |||||||||
| Model 1 | Model 2 | Model 3 | |||||||
| β | SE | P-value | β | SE | P-value | β | SE | P-value | |
| Intercept | 0.896 | 0.018 | <.0001 | 0.857 | 0.016 | <.0001 | 0.859 | 0.019 | <.0001 |
| Female | −0.006 | 0.013 | 0.62 | −0.012 | 0.012 | 0.32 | −0.008 | 0.012 | 0.54 |
| Age 45-54 | −0.004 | 0.016 | 0.80 | −0.006 | 0.016 | 0.73 | −0.006 | 0.016 | 0.72 |
| Age 55-64 | −0.048 | 0.021 | 0.022 | −0.066 | 0.019 | 0.0005 | −0.056 | 0.020 | 0.006 |
| Age 65-74 | 0.009 | 0.021 | 0.67 | −0.030 | 0.018 | 0.087 | −0.007 | 0.020 | 0.72 |
| Age 75-89 | −0.031 | 0.024 | 0.21 | −0.092 | 0.019 | <.0001 | −0.061 | 0.023 | 0.008 |
| Black | −0.042 | 0.020 | 0.035 | −0.039 | 0.019 | 0.043 | −0.029 | 0.020 | 0.14 |
| Less than high school education | −0.123 | 0.031 | <.0001 | −0.139 | 0.029 | <.0001 | −0.109 | 0.029 | 0.0002 |
| High school educations | −0.014 | 0.014 | 0.31 | −0.026 | 0.013 | 0.036 | −0.014 | 0.014 | 0.32 |
| Some college education | −0.040 | 0.017 | 0.022 | −0.043 | 0.016 | 0.0092 | −0.037 | 0.016 | 0.022 |
| Income <$20,000 | −0.151 | 0.030 | <.0001 | −0.102 | 0.027 | 0.0001 | |||
| Income $20,000 to $34,999 | −0.076 | 0.025 | 0.0021 | −0.042 | 0.024 | 0.085 | |||
| Income $35,000 to $74,999 | −0.025 | 0.015 | 0.10 | −0.016 | 0.016 | 0.32 | |||
| Assets <$25,000 | −0.010 | 0.018 | 0.58 | 0.016 | 0.018 | 0.38 | |||
| Assets $25,000 to $99,999 | 0.000 | 0.015 | 0.99 | 0.008 | 0.015 | 0.60 | |||
| Bill Payments not difficult | 0.055 | 0.013 | <.0001 | 0.049 | 0.013 | 0.0002 | |||
| Bill payments very difficult | −0.214 | 0.036 | <.0001 | −0.189 | 0.034 | <.0001 | |||
| R-sq | 0.109 | 0.149 | 0.161 | ||||||
| SF-6D | |||||||||
| Model 1 | Model 2 | Model 3 | |||||||
| β | SE | P-value | β | SE | P-value | β | SE | P-value | |
| Intercept | 0.932 | 0.012 | <.0001 | 0.816 | 0.009 | <.0001 | 0.829 | 0.010 | <.0001 |
| Female | −0.011 | 0.007 | 0.14 | −0.013 | 0.006 | 0.052 | −0.010 | 0.007 | 0.12 |
| Age 45-54 | −0.025 | 0.010 | 0.016 | −0.008 | 0.010 | 0.40 | −0.010 | 0.009 | 0.31 |
| Age 55-64 | −0.057 | 0.011 | <.0001 | −0.038 | 0.010 | 0.0002 | −0.038 | 0.010 | 0.0001 |
| Age 65-74 | −0.032 | 0.012 | 0.0068 | −0.026 | 0.010 | 0.0076 | −0.021 | 0.011 | 0.053 |
| Age 75-89 | −0.056 | 0.012 | <.0001 | −0.060 | 0.011 | <.0001 | −0.049 | 0.012 | <.0001 |
| Black | −0.023 | 0.011 | 0.042 | −0.017 | 0.010 | 0.071 | −0.010 | 0.010 | 0.29 |
| Less than high school education | −0.056 | 0.017 | 0.0008 | −0.064 | 0.014 | <.0001 | −0.046 | 0.015 | 0.0019 |
| High school educations | −0.025 | 0.009 | 0.0038 | −0.018 | 0.009 | 0.038 | −0.011 | 0.009 | 0.23 |
| Some college education | −0.029 | 0.010 | 0.0038 | −0.021 | 0.008 | 0.013 | −0.016 | 0.009 | 0.062 |
| Income <$20,000 | −0.053 | 0.016 | 0.0007 | −0.043 | 0.015 | 0.0033 | |||
| Income $20,000 to $34,999 | −0.017 | 0.013 | 0.19 | −0.014 | 0.012 | 0.24 | |||
| Income $35,000 to $74,999 | −0.012 | 0.009 | 0.18 | 0.000 | 0.009 | 0.99 | |||
| Assets <$25,000 | 0.010 | 0.010 | 0.29 | −0.011 | 0.010 | 0.27 | |||
| Assets $25,000 to $99,999 | −0.009 | 0.008 | 0.27 | −0.017 | 0.008 | 0.033 | |||
| Bill Payments not difficult | 0.043 | 0.007 | <.0001 | 0.036 | 0.008 | <.0001 | |||
| Bill payments very difficult | −0.102 | 0.015 | <.0001 | −0.089 | 0.014 | <.0001 | |||
| R-sq | 0.116 | 0.157 | 0.171 | ||||||
| QWB-SA | |||||||||
| Model 1 | Model 2 | Model 3 | |||||||
| β | SE | P-value | β | SE | P-value | β | SE | P-value | |
| Intercept | 0.750 | 0.012 | <.0001 | 0.712 | 0.011 | <.0001 | 0.717 | 0.014 | <.0001 |
| Female | −0.022 | 0.008 | 0.0082 | −0.023 | 0.008 | 0.0033 | −0.022 | 0.008 | 0.0049 |
| Age 45-54 | −0.028 | 0.012 | 0.017 | −0.027 | 0.011 | 0.0162 | −0.028 | 0.011 | 0.014 |
| Age 55-64 | −0.061 | 0.013 | <.0001 | −0.067 | 0.013 | <.0001 | −0.066 | 0.013 | <.0001 |
| Age 65-74 | −0.048 | 0.013 | 0.0002 | −0.065 | 0.011 | <.0001 | −0.059 | 0.012 | <.0001 |
| Age 75-89 | −0.078 | 0.015 | <.0001 | −0.107 | 0.012 | <.0001 | −0.099 | 0.014 | <.0001 |
| Black | −0.004 | 0.012 | 0.71 | −0.001 | 0.011 | 0.96 | 0.002 | 0.011 | 0.83 |
| Less than high school education | −0.062 | 0.017 | 0.0003 | −0.063 | 0.013 | <.0001 | −0.054 | 0.015 | 0.0003 |
| High school educations | −0.009 | 0.012 | 0.43 | −0.013 | 0.010 | 0.17 | −0.009 | 0.011 | 0.44 |
| Some college education | −0.028 | 0.012 | 0.016 | −0.028 | 0.010 | 0.0059 | −0.026 | 0.011 | 0.019 |
| Income <$20,000 | −0.059 | 0.019 | 0.0015 | −0.027 | 0.017 | 0.12 | |||
| Income $20,000 to $34,999 | −0.041 | 0.015 | 0.0056 | −0.015 | 0.014 | 0.28 | |||
| Income $35,000 to $74,999 | −0.015 | 0.011 | 0.19 | −0.008 | 0.011 | 0.48 | |||
| Assets <$25,000 | −0.016 | 0.012 | 0.19 | 0.002 | 0.013 | 0.87 | |||
| Assets $25,000 to $99,999 | −0.009 | 0.011 | 0.40 | −0.003 | 0.011 | 0.80 | |||
| Bill Payments not difficult | 0.044 | 0.009 | <.0001 | 0.042 | 0.010 | <.0001 | |||
| Bill payments very difficult | −0.112 | 0.015 | <.0001 | −0.106 | 0.015 | <.0001 | |||
| R-sq | 0.107 | 0.159 | 0.162 | ||||||
β: parameter estimate.
Age, sex, and education parameter estimates were in the directions expected and generally remained significant in all models [2,3]. Results were similar for the sensitivity analyses which included chronic conditions. Chronic condition estimates were consistent in magnitude and significance across models 1, 2, and 3 (data not shown).
Discussion
In this study, we compare two measures to adjust for SES in studies of health utility: a single DPB question and directly-assed assets and income questions. The single DPB question was correlated with both income and assets, but explained more variance when modeling health utility outcomes in a general U.S. population sample aged 35 to 89. The DPB question also had a higher response rate compared to either income or assets questions.
Due to complexity and sensitively of self-reported income questions, nonresponse is a significant issue in survey research [6,7,20, 21]. To circumvent this issue, Clemens and Dibben (2014) estimated a synthetic measure from occupational classification and labor force surveys. They concluded “that occupation based synthetic estimates of wage are as effective in capturing the underlying relationship between income and health as survey reported income.” Fukuoka et al (2007) found that higher education and greater social and economic satisfaction were significantly related to refusal to self-report income after controlling for age and sex in 247 cardiac patients (≥65 years old). A comprehensive review of government-sponsored survey research on the topic of income measurement error [6,7] highlighted several possible contributors to inaccurate reporting of income, including cognitive problems (e.g., lack of knowledge, misunderstanding of definitions/questions, recall problems) and respondents’ sensitivity related to revealing their income.
There are important debates about the structure of the relationship between SES and health. Different sociological approaches include fundamental cause theory, life course theory, age-as-level hypotheses, and cumulative advantage/disadvantage hypotheses [2]. A possible limitation of the DPB question is that, unlike incomes or assets questions that are direct quantitative measurement of one’s financial status, the DPB question may be considered a relative and/or subjective assessment. Although careful direct measurement of SES may be important to address fundamental sociological questions, many health utility studies are primarily concerned with measuring the effect of a health intervention while adjusting for SES factors.
Since our study only included respondents self-reporting as white or African American of ages 35 years old or older, future research will need to examine the relationship between DPB and health utility in other subgroups of the population. Other limitations typical of survey data are detailed by Fryback et al. (2007).
Despite these limitations, our study contributes novel and important information about an alternative assessment of SES in survey studies that measure health utility. The comprehensive database of the NHMS allowed us to demonstrate results that were consistent across multiple health utility indexes. The robustness of our findings was further demonstrated by including chronic health conditions in all models in sensitivity analyses.
Conclusion
A single question about difficulty paying bills has less missing values and may yield more information than the traditionally used income and assets questions. This simple item about financial security should be considered as an alternative SES measure in future studies.
Acknowledgments:
Janel Hanmer was supported by the National Institutes of Health through Grant Number KL2 TR000146. The funding agreements ensured the authors’ independence in designing the study, interpreting the data, writing, and publishing the report.
Funding: This study was funded by the National Institutes of Health through Grant Number KL2 TR000146.
Footnotes
Conflict of Interest: Janel Hanmer, MD, PhD declares that she has no conflict of interest. Dasha Cherepanov, PhD declares that she has no conflict of interest.
Ethical approval: This study was performed using deidentified, publicly available data which are except from IRB review. This article does not contain any studies with animals performed by any of the authors.
References
- 1.Fryback DG, Dunham NC, Palta M, Hanmer J, Buechner J, Cherepanov D, Herrington SA, Hays RD, Kaplan RM, Ganiats TG, Feeny D, Kind P. US norms for six generic health-related quality-of-life indexes from the National Health Measurement study. Med Care. 2007;45(12):1162–70. PubMed PMID: 18007166 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Robert SA, Cherepanov D, Palta M, Dunham NC, Feeny D, Fryback DG. Socioeconomic status and age variations in health-related quality of life: results from the national health measurement study. J Gerontol B Psychol Sci Soc Sci. 2009;64(3):378–89. PubMed PMID: 19307286 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Cherepanov D, Palta M, Fryback DG, Robert SA. Gender differences in health-related quality-of-life are partly explained by sociodemographic and socioeconomic variation between adult men and women in the US: evidence from four US nationally representative data sets. Qual Life Res. 2010;19(8):1115–24. Erratum in: Qual Life Res. 2010;19(8):1125. PubMed PMID: 20496168 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Luo N, Johnson JA, Shaw JW, Feeny D, Coons SJ. Self-reported health status of the general adult US population as assessed by the EQ-5D and Health Utilities Index. Medical care. 2005; 43(11); 1078–1086. [DOI] [PubMed] [Google Scholar]
- 5.Hemingway H, Stafford M, Stansfeld S, Shipley M, Marmot M. Is the SF-36 a valid measure of change in population health? Results from the Whitehall II study. BMJ. 1997; 315(7118): 1273–1279. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Moore JC, Stinson LL, Welniak EJ (1999) Income reporting in surveys: Cognitive issues and measurement error Pp. 155–174 in Sirken MG, Herrmann DJ, Schechter S, Schwarz N, Tanur JM and Tourangeau R (Eds.) Cognition and Survey Measurement. NY: Wiley; Accessed 7/19/2015 at: https://www.census.gov/srd/papers/pdf/sm97-05.pdf [Google Scholar]
- 7.Moore JC, Stinson LL, and Welniak EJ. Income measurement error in surveys: A review. Journal of Official Statistics 2000; 16: 331–361. [Google Scholar]
- 8.Locander WB and Burton JP. The Effect of Question Form on Gathering Income Data by Telephone. Journal of Marketing Research. 1976; 13(2): 189–192. [Google Scholar]
- 9.Shih WP. Nonresponses to Income Questions in Telephone Surveys. Proceedings of the Survey Research Methods Section, American Statistical Association 1983; 283. [Google Scholar]
- 10.Krause N, Jay G, Liang J. Financial strain and psychological well-being among the American and Japanese elderly. Psychology and aging, 1991; 6(2): 170. [DOI] [PubMed] [Google Scholar]
- 11.Boardman JD. Stress and physical health: the role of neighborhoods as mediating and moderating mechanisms. Social Science & Medicine, 2004; 58(12): 2473–2483. [DOI] [PubMed] [Google Scholar]
- 12.Ellison CG, Boardman JD, Williams DR, Jackson JS. Religious involvement, stress, and mental health: Findings from the 1995 Detroit Area Study. Social forces, 2001; 80(1): 215–249. [Google Scholar]
- 13.Kind P Size matters: EQ-5D in transition. Medical Care 2007; 45: 809–810. [DOI] [PubMed] [Google Scholar]
- 14.Shaw JW, Johnson JA, Coons SJ. US valuation of the EQ-5D health states: Development and testing of the D1 valuation model. Med Care 2005;43:203–220. [PubMed: 15725977] [DOI] [PubMed] [Google Scholar]
- 15.Feeny D, Furlong W, Torrance GW, et al. Multiattribute and single-attribute utility functions for the health utilities index mark 3 system. Med Care 2002;40:113–28. [PubMed: 11802084] [DOI] [PubMed] [Google Scholar]
- 16.Brazier J, Roberts J, Deverill M. The estimation of a preference-based measure of health from the SF-36. J Health Econ 2002;21:271–92. [PubMed: 11939242] [DOI] [PubMed] [Google Scholar]
- 17.Andresen EM, Rothenberg BM, Kaplan RM. Performance of a self-administered mailed version of the Quality of Well-Being (QWB-SA) questionnaire among older adults. Med Care 1998;36:1349–60. [PubMed: 9749658] [DOI] [PubMed] [Google Scholar]
- 18.Kaplan RM, Sieber WJ, Ganiats TG. The Quality of Well-being Scale: Comparison of the intervieweradministered version with a self-administered questionnaire. Psychol & Health 1997;12:783–791. [Google Scholar]
- 19.Hanmer J, Cherepanov D, Palta M, Kaplan RM, Feeny D, Fryback DG. Health Condition Impacts in a National Representative Cross-Sectional Survey Vary Substantially by Preference-Based Health Index. In press at Medical Decision Making. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Clemens T, Dibben C. A method for estimating wage, using standardized occupational classifications, for use in medical research in the place of self-reported income. BMC Med Res Methodol. 2014; 28(14):59. PubMed PMID: 24779534. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Fukuoka Y, Rankin SH, Carroll DL. Systematic bias in self-reported annual household incomes among unpartnered elderly cardiac patients. Appl Nurs Res. 2007;20(4):205–9. PubMed PMID: 17996808. [DOI] [PubMed] [Google Scholar]