Abstract:To the Editor: Disability weights (DWs) are essential factors to quantify health losses relating to non-fatal outcomes for estimates of disability-adjusted life years (DALYs). Although national and subnational sets of DWs were published recently in the Chinese mainland,[1] data from only 4925 participants who responded to population health equivalence (PHE) questions used to anchor the 0–1 DWs scale did not include participants from Fujian province. Moreover, it remains unknown whether people living in different cities share the same DWs. This study provided an alternative approach using non-parametric regression to locate the DWs scale, as used in European surveys,[2] aimed at measuring DWs at the provincial and subprovincial levels in Fujian. DWs for a set of 206 health states were assessed in this study. These health status and lay descriptions were presented in a Chinese DWs measurement study.[1] A web-based survey was conducted among permanent residents aged 18–69 years from May to July 2020 in Fujian. The questionnaire consisted of sociodemographic and 16 pairwise comparison (PC) questions that were randomly selected from 206 health states.[1] Probit regression was used to estimate choice responses to the PC questions. We performed locally weighted regression (Loess) to anchor the probit regression coefficients in the range from 0 to 1. The probit regression coefficients and logit-transformed DWs from the Global Burden of Disease 2013 (GBD 2013) study were matched and used to construct the Loess model. A total of 93,254 eligible participants (27,968 in Quanzhou, 24,384 in Fuzhou, 14,762 in Xiamen, and 26,140 in the remaining cities) were included in our analysis. The characteristics of the respondents are described in Supplementary Table 1, https://links.lww.com/CM9/B675. Sixty-three percent of the respondents were female, and the majority (85%) were below 50 years of age. In addition, 65% of the respondents had college degrees or above, and approximately half (47%) had a high income level. Compared with the general population in Fujian, the respondents in our study tended to be younger, female, and have a higher educational level and showed consistent patterns with that of GBD,[3,4] European and Japanese studies[5] [Supplementary Table 2, https://links.lww.com/CM9/B675]. The heatmap of PC response probabilities for 206 × 206 possible paired comparisons indicated a high degree of consistency among PC questions [Supplementary Figures 1 and 2, https://links.lww.com/CM9/B675]. The correlation values of the probit regression coefficients were high among different sexes (r = 0.995), age groups (r range: 0.982–0.995), education levels (r range: 0.987–0.995), income levels (r = 0.999), developed and developing cities (r = 1.000), and cities (r range: 0.997–0.999),the P values for all of the correlation coefficients above were less than 0.001.The correlation coefficients of the probit regression coefficients for different subgroups ranged from 0.982 to 1.000. The lowest correlation coefficient was between the 18–29 and 50–69 years groups. The correlation coefficient between "senior high school education and below" and "undergraduate education and above" was slightly low at 0.987 [Figure 1].Figure 1: Comparison of probit regression coefficients between different subgroups and cities in Fujian. (A–I) Correlations of probit regression coefficients between different subgroups such as men and women, high income and low income, developed and developing region, aged 18–29 and aged 30–49 years, aged 18–29 and aged 50–69 years, aged 30–49 and aged 50–69 years, senior high school and below and college graduation, senior high school and bellow and undergraduate and above, college graduation and undergraduate and above. (J–O) Correlations of probit regression coefficients between different cities such as Fuzhou and Xiamen, Fuzhou and Quanzhou, Xiamen and Quanzhou, Fuzhou and other cities, Xiamen and other cities, Quanzhou and other cities. The x-axis and the y-axis represent their probit regression coefficients. The correlation coefficient "r" is close to 1 and has statistical significant difference (P <0.001).The estimated DWs and 95% uncertainty intervals (UIs) for 206 health states in Fujian and its major cities are shown in Supplementary Table 3, https://links.lww.com/CM9/B675. The DW was the lowest for mild distance vision impairment or mild anemia (0.005, 95% UI: 0.002–0.012), and the highest DW was for severe heroin and other opioid dependence (0.657, 95% UI: 0.499–0.779) in Fujian. A total of 182 (88.3%) and 88 (42.7%) of the 206 DWs were below 0.40 and 0.10, respectively. Fuzhou, Xiamen, and Quanzhou, with high per capita GDP and sufficient sample size, were analyzed at the city level, while other cities were grouped into one group. A high correlation of DWs among cities is shown (rs >0.900, P <0.001) in Supplementary Table 4, https://links.lww.com/CM9/B675. The DW of "mild distance vision impairment" or "mild anemia" was the lowest in all cities, ranging from 0.005 (all cities) to 0.006 (Fuzhou, Xiamen, Quanzhou). Except for Xiamen, the DW of severe heroin and other opioid dependence was the highest, ranging from 0.642 (Fuzhou city) to 0.669 (Quanzhou city). In Xiamen, the DW of "major depression, severe episodes" was the highest (0.654, 95% UI: 0.502–0.761), followed by the DW of "severe heroin and other opioid dependence" (0.652, 95% UI: 0.498–0.758). Except for slight differences, the top and bottom 10 health states of each city were generally consistent with those of Fujian Province [Supplementary Figures 3 and 4, https://links.lww.com/CM9/B675]. The correlation coefficient between the DWs of Fujian anchored with GBD 2013 DWs and those anchored using the PHE method was 0.991, indicating high consistency between the two anchoring methods [Supplementary Figure 5, https://links.lww.com/CM9/B675]. There was a slight difference in DW values and rankings between Fujian and China. The differences ranged from -0.047 to 0.095, and its absolute value was less than 0.100. There were some gaps in the DWs of Fujian compared to those of GBD 2013 and Japan.There was a difference in the highest DW values (Fujian: 0.657, GBD 2013: 0.778). The differences in DWs between Fujian and GBD 2013 ranged from –0.258 (severe alcohol use disorder, GBD 2013: 0.570, Fujian: 0.312) to 0.349 (mild cannabis dependence, GBD 2013: 0.039, Fujian: 0.388). Health states with lower DW values in Fujian than in GBD 2013 included "Aquired immune deficiency syndrome (AIDS): not receiving antiretroviral (ARV) treatment," moderate dementia, various most severe musculoskeletal disorders, severe motor plus cognitive impairments, and moderate encephalopathy. Several substances use disorders with higher DW values in Fujian than in GBD 2013 included mild and severe cannabis dependence, mild amphetamine dependence, and mild cocaine dependence[Supplementary Table 5, https://links.lww.com/CM9/B675]. The gap shown above might be due to the lack of samples of the Asian population and the different anchoring methods applied. The differences between the DWs of Fujian and those of the Japanese study ranged from -0.316 (invasive device/drain, Japanese study: 0.512, Fujian: 0.196) to 0.320 (end-stage renal disease on dialysis, Japanese study: 0.278, Fujian: 0.598). In addition, the DWs of harmful alcohol use and moderate acute episode infectious disease were lower in Fujian than in Japan, while the DWs of moderate episode major depressive disorder and manic episode bipolar disorder were higher in Fujian compared to Japan. Moreover, the DWs of several stroke-related health states, moderate dementia, and severe dementia were lower in Fujian compared to both Japan and GDB 2013. Therefore, our study could not refute the influence of cultural differences on health state preferences. This study provided empirical data for countries or regions that have not yet conducted DWs surveys, suggesting that it might not be necessary for those regions with population sizes similar to that of Fujian to conduct stratified DWs studies, as they could directly use synthetic DWs results. The instinctive limitation of web-based surveys increased the chances of not deriving a random population sample from Fujian. Compared to the general population, the respondents were younger and more highly educated; thus, the DWs might not fully reflect the cognition of people in Fujian. This study built a provincial and subprovincial dataset of DWs based on 93,254 individuals and provided further evidence on the effect of cultural differences on DWs for specific health states. For the purpose of global comparability, GBD studies are encouraged to cover more countries with large demographic, cultural, and geographical distinctions to estimate a universal set of DWs values. The high consistency of DWs among cities in Fujian indicated that a set of DWs could be used when DW studies are conducted in countries or regions of the same size. Funding This work was supported by grants from the Fujian Provincial Science and Technology Project (No.2020Y0060), Fujian Provincial Health Technology Project (No.2019-CXB-15), and Fujian Provincial Health Technology Project (No.2020GGA026). Conflicts of interest None.

Eliciting National and Subnational Sets of Disability Weights in Mainland China: Findings from the Chinese Disability Weight Measurement Study.

Disability weight measurement for the severity of different diseases in Wuhan, China

Can global or national disability weights represent provincial level?

Years lived with disability of cancer in China: findings from disability weights measurement with a focus on the effect of disease burden

Estimation of Cancer Mortality Attributable to Excess Body Weight During 2006-2015 in China.

Assessment of Disability Weights at the Provincial and City Levels Based on 93,254 Respondents in Fujian, China: Findings from the Fujian Disability Weight Measurement Study

[Estimation of Disability Weights on Malignant Neoplasms in Shandong Province].

Lower BMI cutoffs to define overweight and obesity in China.

Disability weights from a household survey in a low socio-economic setting: how does it compare to the global burden of disease 2010 study?

A Cross-Sectional Survey of Disability Attributed to Mental Disorders and Service Use in China

Regional variations in and correlates of disability-free life expectancy among older adults in China

Validity of the Chinese multimorbidity-weighted index in measuring disease burden using health check-ups data in primary care

Disease burden of ageing, sex and regional disparities and health resources allocation: a longitudinal analysis of 31 provinces in Mainland China

Common values in assessing health outcomes from disease and injury: disability weights measurement study for the Global Burden of Disease Study 2010

Weight-adjusted waist index and disability: a cohort study from CHARLS

Mortality, morbidity, and risk factors in China and its provinces, 1990–2017: a systematic analysis for the Global Burden of Disease Study 2017

Mortality of People with Disabilities in China: a Prospective Cohort Study

Disability Associated with Mental Disorders in Metropolitan China: an Application of the Quantile Regression Approach

Gender disparities in the association between socio-demographics and non-communicable disease risk factors among adults with disabilities in Shanghai, China

Mortality, morbidity, and risk factors in China and its provinces 1990-2017: a systematic analysis for the Global Burden of Disease Study 2017