Examining the incremental Impact of long-standing health conditions on subjective well-being- is there anything missing from the EQ-5D?
0 likes162 views
This study examined whether the EQ-5D adequately captures the impact of long-standing health conditions on subjective well-being (SWB). Regression models were used to analyze data from the South Yorkshire Cohort. The models showed that anxiety/depression had the largest negative impact on life satisfaction, followed by self-care and pain. Insomnia also negatively impacted SWB and could be an additional dimension for the EQ-5D. Overall, the EQ-5D captured the effects of health on SWB, though insomnia provided some incremental information beyond the EQ-5D dimensions.
Examining the incremental Impact of long-standing health conditions on subjective well-being- is there anything missing from the EQ-5D?
- 1. Examining the incremental impact of long-standing health conditions on subjective well-being – is there anything missing from the EQ-5D? Mengjun Wu1, John E. Brazier1, Joanna Blackburn2, Cindy L. Cooper1, Clare Relton1 and Christine Smith2 1School of Health and Related Research, University of Sheffield, Sheffield, UK., 2Research & Development Department, Barnsley Hospital NHS Foundation Trust, Barnsley, UK. Generic preference-based measures (GPBMs) such as the EQ-5D and SF-6D are commonly used to obtain health utility scores for economic evaluations. Previous studies examining the way that subjective well-being (SWB) might value health using GPBMs did not explore whether aspects of health important to patients in terms of the impact on SWB were not being picked up by the EQ-5D. Our intention is to examine whether the EQ-5D is adequate in reflecting the impact of health conditions on SWB by examining whether long-standing health conditions have an incremental impact on SWB alongside the EQ-5D. Introduction Data from the South Yorkshire Cohort (SYC) were used to undertake all the analyses. Two regression models were used to examine the impact of additional input of long-standing health conditions on life satisfaction. • Ordered logit model was used when the proportional odds assumption held. • Partial proportional odds ordered logit model namely generalised model was employed when the assumption did not hold. • Regression models estimate: I. Odds ratios – the coefficients of the independent variables in terms of the sign and the magnitude II. Model performance: AIC, BIC and McKelvey & Zavoina's R2 III. Threshold cuts: the estimates on life satisfaction to differentiate respondents choosing from one level to the next Methodology • White people were likely to report higher levels of life satisfaction but currently employed people were likely to report lower levels of life satisfaction. Generalised ordered logit: • People with more severe anxiety/depression were more likely to report lower levels of life satisfaction and people with the most severe anxiety/depression were likely to report the lowest level of life satisfaction. • People with insomnia were likely to become completely dissatisfied or satisfied with their lives. • Males were likely to report higher levels of life satisfaction than females. People educated at GCSEs/A levels were likely to report higher levels of life satisfaction. • The difference between people who reported lower levels and those who reported the highest levels of life satisfaction was significant for being white and educated at a degree level. Results Continued • Similar findings about anxiety/depression as in the literature were confirmed – the largest negative association with SWB among five dimensions. Surprisingly, self-care tended to have the second largest negative impact. • The coefficient of insomnia was significant and inclusion of insomnia had some impact on the EQ-5D dimensions particularly anxiety/depression and self-care. Therefore, insomnia could be considered as an additional ‘bolt-on’ dimension to the EQ-5D. • An interesting finding was that all other health conditions had little or no impact on SWB. • For further research, longitudinal data sets should be used for analysis to observe whether the association remains over time and establish a causality relationship. Discussion * This study was funded by the NIHR. This poster presents independent research. The views expressed in this poster are those of the authors and not necessarily those of the NHS, the NIHR or the Department of Health. Contact: Mengjun Wu Postal address: School of Health and Related Research, Regents Court, 30 Regent Street, Sheffield, S1 4DA, UK. Email: m.wu@sheffield.ac.uk Website: www.shef.ac.uk/heds Contact Explanatory variables Model 1 EQ-5D dimensions + health conditions Model 1 EQ-5D score + health conditions Odds ratio SE Odds ratio SE Mobility 2 0.927 0.053 Self-care 2 0.793*** 0.068 Usual activities 2 1.056 0.062 Usual activities 3 1.027 0.162 Pain/discomfort 2 0.925** 0.035 Pain/discomfort 3 0.860 0.088 Anxiety/depression 2 0.728*** 0.029 Anxiety/depression 3 0.603*** 0.077 EQ-5D score 1.971*** 0.162 Tiredness/fatigue 0.949 0.049 0.914* 0.046 Insomnia 0.877* 0.061 0.840** 0.058 Diabetes 0.973 0.067 0.976 0.067 Breathing problems 0.917* 0.048 0.915* 0.048 High blood pressure 1.028 0.044 1.038 0.044 Heart disease 1.098 0.077 1.114 0.078 Osteoarthritis 0.970 0.058 1.003 0.058 Stroke 0.873 0.105 0.868 0.104 Cancer 0.960 0.093 0.953 0.090 Control variables Age 0.952*** 0.005 0.952*** 0.005 Age2 1.001*** 5E-05 1.001*** 5E-05 Male 1.039 0.032 1.041 0.032 White 1.152* 0.097 1.155* 0.097 GCSEs 1.054 0.056 1.043 0.055 A levels 1.029 0.024 1.023 0.023 Degree 1.016 0.016 1.013 0.016 White collar 1.001 0.035 0.998 0.035 Currently employed 0.933** 0.030 0.932** 0.030 Threshold 1 -3.285*** 0.155 -2.592*** 0.169 Threshold 2 -1.853*** 0.153 -1.163*** 0.167 Threshold 3 -0.947*** 0.152 -0.260 0.167 Threshold 4 0.220 0.152 0.904*** 0.167 Threshold 5 1.215*** 0.153 1.899*** 0.168 Observations 13591 13591 Likelihood ratio 𝝌 𝟐 642.88 596.00 McKelvey & Zavoina's 𝑹 𝟐 0.045 0.042 AIC 46288 46320 BIC 46521 46501 * p < 0.1, ** p < 0.05, *** p < 0.01. Table 1 Ordered logit regression: EQ-5D & long-standing health conditions Life satisfaction level I II III IV V Odds ratio Odds ratio Odds ratio Odds ratio Odds ratio Mobility 2 0.865 0.893 0.912 0.960 0.993 Self-care 2 0.881 0.791** 0.828* 0.779** 0.761* Usual activities 2 1.092 1.053 1.111 1.009 0.990 Usual activities 3 0.847 1.038 1.072 0.944 1.215 Pain/discomfort 2 0.997 0.952 0.911** 0.916* 0.939 Pain/discomfort 3 0.869 0.966 0.832 0.858 0.827 Anxiety/depression 2 0.646*** 0.656*** 0.700*** 0.791*** 0.912 Anxiety/depression 3 0.253*** 0.524*** 0.694*** 0.889 0.890 Tiredness/fatigue 1.060 0.931 0.948 1.039 0.931 Insomnia 0.812* 0.908 0.906 0.881 0.780** Diabetes 0.721** 0.938 0.982 1.039 1.061 Breathing problems 0.864 0.900 0.905* 0.973 0.937 High blood pressure 1.005 1.023 1.021 1.058 0.984 Heart disease 1.199 1.218** 1.095 1.014 1.007 Osteoarthritis 0.929 0.940 0.981 1.041 0.875 Stroke 0.833 0.943 0.800 0.901 0.931 Cancer 1.158 0.957 0.956 0.925 0.988 Control variables Age 0.946*** 0.963*** 0.960*** 0.951*** 0.947*** Age2 1.001*** 1.000*** 1.001*** 1.001*** 1.001*** Male 0.995 1.079* 1.030 1.028 1.024 White 1.212 1.463*** 1.312*** 0.927 0.613*** GCSEs 1.145 1.166** 1.103 0.990 0.892 A levels 1.052 1.070** 1.051* 1.002 0.968 Degree 1.092** 1.076*** 1.046** 0.976 0.917*** White collar 0.957 1.000 0.964 1.047 0.983 Currently employed 0.917 0.935 0.900*** 0.951 0.942 Observations 13591 Likelihood ratio 𝝌 𝟐 923.94 P value 0.000 * p < 0.1, ** p < 0.05, *** p < 0.01. Life satisfaction level I II III IV V Odds ratio Odds ratio Odds ratio Odds ratio Odds ratio EQ-5D score 3.143*** 2.080*** 1.911*** 1.762*** 1.477*** Tiredness/fatigue 1.022 0.880** 0.915 0.912 0.917 Insomnia 0.768** 0.851* 0.866* 0.862* 0.777** Diabetes 0.733** 0.934 0.987 1.041 1.065 Breathing problems 0.893 0.897 0.904* 0.966 0.933 High blood pressure 1.043 1.040 1.029 1.060 0.983 Heart disease 1.308* 1.251** 1.116 1.009 1.007 Osteoarthritis 1.080 0.989 1.011 1.047 0.880 Stroke 0.859 0.939 0.801 0.887 0.911 Cancer 1.143 0.941 0.949 0.912 0.975 Control variables Age 0.945*** 0.962*** 0.960*** 0.951*** 0.946*** Age2 1.001*** 1.000*** 1.001*** 1.001*** 1.001*** Male 1.007 1.084* 1.031 1.027 1.027 White 1.214 1.467*** 1.319*** 0.928 0.617*** GCSEs 1.114 1.148** 1.086 0.983 0.894 A levels 1.038 1.061** 1.043 0.998 0.966 Degree 1.073** 1.070*** 1.043** 0.975 0.917*** White collar 0.951 1.001 0.963 1.043 0.984 Currently employed 0.922 0.938 0.900*** 0.950 0.944 Observations 13591 Likelihood ratio 𝝌 𝟐 823.42 P value 0.000 * p < 0.1, ** p < 0.05, *** p < 0.01. Table 2 Generalised logit regression: EQ-5D dimensions & long-standing health conditions Table 3 Generalised logit regression: EQ-5D score & long-standing health conditions Ordered logit: • Anxiety/depression had the largest negative impact on life satisfaction, followed by self-care and pain/discomfort. • People with better health-related quality of life were likely to report higher levels of life satisfaction. • People with insomnia/breathing problems were likely to report lower levels of life satisfaction. • Age had a negative association with SWB but age squared had a positive correlation with SWB. Results