This study examined the overall burden for 23 chronic diseases, chronic conditions, and diseases with chronic consequences in terms of QALY losses, a metric which is widely used in health economic evaluations, based on representative data captured in the BHIS. Moreover, the results of this study provide novel insights into socioeconomic inequalities in QALYs, which is useful to support policy trade-offs between improving population health and reducing unequal health distribution. In general, this study provides evidence for the need for health policies targeting chronic diseases in the most vulnerable populations, i.e. women, elderly, and low SES populations.
We stratified our findings by age, sex, SES, and time point. In 2013 and 2018, the largest QALY loss was due to dorsopathies, arthropathies, genitourinary problems, and hypertension/high cholesterol. Earlier research indicated that musculoskeletal disorders and hypertension were associated with the largest loss of QALYs in the population 1,22. Comparable with previous research, these results are mainly attributed to their high prevalence among the Belgian population 1,23. This study also recognizes the substantial impact of psychological disorders (i.e. depression as the 5th cause of QALY loss in 2018) due to greatly impaired HRQoL. Moreover, the QALY losses for different age intervals revealed that older age groups are most affected by chronic diseases as expected 24. Furthermore, women had a larger QALY loss than men which can be mainly attributed to higher disutility values in women 25,26. This study also showed that the burden is higher in 2018 compared to 2013 because of increases in disease prevalence as a result of population ageing. Chronic diseases are paramount in an ageing society and susceptibility to chronic diseases increases with age 27. More importantly, a large inequality gap in QALY losses was found between the least and the most deprived population groups, which is consistent with previous research 12,28. Indeed, SES is the main determinant of chronic disease distribution in populations 29. The largest inequality gap was seen in arthropathies and hypertension/high cholesterol, mainly due to higher prevalence rates in low SES groups. Low SES is indeed found to be associated with the risk of developing arthritis and hypertension due to higher smoking rates, body mass index (BMI), and lack of exercise compared to high SES groups 30–32. There is also strong evidence that SES is associated with worse HRQoL outcomes 25,33,34. Hence, it is expected that the inequality gap in QALY loss due to chronic diseases will continue to grow.
Several limitations related to the BHIS should be acknowledged. First, information on chronic morbidity was based on self-reports measured by a single and global question. The accuracy of self-reports depends on the participants’ knowledge and understanding of the relevant information, ability to recall it, and willingness to report it 35. This is challenging because participants are often confused to distinguish between symptoms and the actual disease, and because some diseases are very subjective (e.g. chronic fatigue). In addition, people may indicate to have several diseases (e.g. depression and chronic fatigue) because both diseases have homogeneous symptoms and common etiology. Although self-reported chronic morbidity may underestimate the prevalence of medical conditions (thus underestimating QALY losses), it is found to be a reasonably reliable instrument to measure ill health 36. Another limitation is the incomplete list of chronic diseases included in the BHIS, implying potential missing of other important chronic conditions. Besides, few mental or psychiatric conditions were included. Another limitation is potential selection bias, which may result from educational differences in survey participation and in the willingness and ability to answer the self-administered questionnaire. Accordingly, lower participation rates were found in lower educated households, especially when they have a poor health status and a risky health behaviour compared with higher educated households 37,38. Consequently, health inequalities may be underestimated in the present study. Moreover, the definition of SES is debatable as it only includes educational attainment. Indeed, income or employment status are also important indicators of SES. Nevertheless, these indicators were not used because information on these variables was less frequently available 39. However, educational attainment is found to be a relatively stable measure of SES and is usually of good quality 40,41.
Some methodological considerations should be mentioned. First, we estimated the disease prevalence in all respondents and not only in those who completed the EQ-5D. As such, the estimated prevalence corresponds better with the actual prevalence in the general population. A second methodological issue is related to the calculation of disutilities. In general, when the HRQoL score of a respondent is higher than the general population norm, the difference results in negative values (i.e. gain in HRQoL), which is methodologically irrelevant. As such, we replaced negative values by zero. Third, the possible effects of comorbidity were not taken into account when calculating QALY loss, examining the impact of combinations of conditions would provide a more dynamic and comprehensive overview, especially in older age categories. Fourth, cross-country comparisons of QALY losses are difficult due to differences in EQ-5D value sets resulting from sociocultural differences 42. It is therefore recommended to compare and interpret QALY outcomes, and cost-utility outcomes in general, from different countries with caution 43. Fifth, we did not conduct statistical testing given the descriptive nature of this study. However, additional analysis may be considered in future research.
These limitations notwithstanding, this study provides representative results at the level of the Belgian population. In addition, we used the health status of the general population as comparator when estimating HRQoL loss. Using ‘perfect health’ as comparator would have resulted in an overestimation of QALY losses. The current economic standard is to elicit and compare HRQoL estimated from the general public because economic evaluations are meant to guide social policies 44.