Introduction

Chronic condition is described as a health problem that persists across time and requires some degree of health care management [1]. Globally, chronic conditions are on the rise, leading to diminished health-related quality of life (HRQoL), reduced physical functioning, and a higher risk of depression and anxiety [2–4]. Multimorbidity is commonly defined as the presence of two or more chronic conditions in an individual [5]. Given that most healthcare systems worldwide were designed to focus on single-disease management, multimorbidity poses major challenges for healthcare providers [6]. There is a consensus that current disease-specific clinical guidelines might be inappropriate for patients with more than one chronic conditions. If we follow the current guidelines, and each condition is considered in isolation, patients will have to visit healthcare providers more frequently, implement life-style changes and take more medicines [7]. In view of the complexity of efficiently caring for patients with chronic condition and multimorbidity, there is now a global consensus that a key component of patient-centered healthcare should be support specifically designed to manage multimorbidity. Healthcare services should transfer the focus from the deterioration of physically functioning caused by a chronic condition or associated multimorbidity to the proactive improvement of patients’ HRQoL.

Chronic conditions, especially multimorbidity, have the potential to induce profound negative effects on a person’s HRQoL or wellbeing, which has been the subject of studies for decades [5,8,9]. Recently, in Asian-Pacific countries, a growing number of studies has indicated an inverse relationship between multimorbidity and HRQoL in Australians [10] and Korean adults, especially elderly women with a lower socioeconomic status [11]. In addition, Pati et al. found a significantly high impairment of physical and mental HRQoL in patients with multimorbidity above the age of 50 years and different combinations of chronic conditions appears to impact on HRQoL differently [12]. However, although an increasing numbers of people in Hong Kong (HK) live with multimorbidity, few studies explored the effect of multimorbidity on the HRQoL using generic preference-based measure in that population. A study by Chung et al. of the HK general population demonstrated the following significant risk factors for chronic conditions and multimorbidity: being female and 25 years or older, having a low education level, having a low household income, and being unemployed or retired [13]. Another recent study conducted by Chin et al., which used the SF-12 questionnaire with patients from primary care clinics in HK, found a positive relationship between multimorbidity and depressive symptoms [14]. Although these studies are valuable and heuristic, did not explore the impact of multimorbidity on HRQoL in the local population. Understanding this impact could enable policymakers to design effective and patient-centered healthcare system to address the growing challenge of multimorbidity.

An increasing number of studies conducted worldwide have explored new approaches to evaluate HRQoL [15], in recent decades, several instruments have been introduced in this field. For example, the EQ-5D, developed by the EuroQol group, is a simple and generic measurement that is widely applied for HRQoL analysis [16], which is increasingly used as a guideline to inform the evaluation of patient-reported outcomes in many jurisdictions [15]. Recently, there has been an increased interest in defining and calculating the minimally important difference (MID) of utility scores using generic measurement. MID could be used to quantify the minimal variation in an index score, such as the EQ-5D, and represents a meaningful improvement for the patient in order to help professionals assess patient-reported outcomes based on different diseases or conditions. Around the world, several countries, including Canada, China, Spain, Japan, England, and Uruguay, have established the MID for their general populations [17]. However, no MID estimates have been published using the EQ-5D-5L scoring algorithms for the HK general population. Therefore, the aims of this study were to 1) evaluate the impact of eight common chronic conditions and multimorbidity on HRQoL, and 2) estimate the MID in HK’s general population.

Method

Results

A total of 1,033 HK residents participated in the study. The results of 19 respondents were discarded after either declining to be interviewed due to unavailability or after providing incomplete responses, leaving 1,014 responses for data analysis. Table 1 shows the background characteristics of the sampled respondents. The percentage of women in the study was 59.2%. An education level below primary school was reported by 19.8% of respondents, and nearly 70% of respondents were between the ages of 25 and 65 years. The prevalence of chronic conditions and multimorbidity was 30.3% and 12.3%, respectively. No statistically significant differences were found between our sample and the general population in terms of background characteristics and chronic conditions.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Background characteristics of respondents.

https://doi.org/10.1371/journal.pone.0224970.t001

Fig 1 indicates that nearly 51% of respondents with no chronic conditions reported full health (utility = 1.0), 39.4% with one chronic condition, 36.9% with two chronic conditions, and 14.7% with three or more chronic conditions also reported full health. The health utility began to fluctuate when there were two or more chronic conditions. Fig 2 shows the distribution of EQ-5D utility by sex and age groups based on the respondents’ chronic conditions (with/without multimorbidity). The distribution of the EQ-5D utility scores for both men and women was similar and more respondents reported no multimorbidity scored a higher EQ-5D utility than respondents with multimorbidity. Fewer respondents (3.8% and 3.0% respectively) with multimorbidity in younger groups (18–24 years and 25–44 years) reported having full health (utility = 1.0). Lower utility scores were reported for respondents with multimorbidity than those without multimorbidity in all age groups, except for the oldest group (aged 65 and above). 65.4% of respondents with multimorbidity aged 65 years and above reported full health. Fig 3 depicts the mean utility score of EQ-5D for each chronic condition. Respondents with diabetes had the highest utility score of 0.89, followed by, hypertension, heart disease, and mental problems. Physical disability scored the lowest score with 0.66.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. The distribution of EQ-5D utility of multimorbidity.

https://doi.org/10.1371/journal.pone.0224970.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. The distribution of EQ-5D utility of gender and age group based on people’s chronic conditions.

https://doi.org/10.1371/journal.pone.0224970.g002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. The EQ-5D mean utility for each chronic condition.

https://doi.org/10.1371/journal.pone.0224970.g003

Table 2 shows the percentage of respondents with health problems on EQ-5D reporting multimorbidity by sex, age and educational level. Findings showed that elderly respondents were more likely to have multimorbidity. 75% of all respondents aged 65 years or older with three or more chronic conditions reported having problems with self-care, and follow by anxiety/depression (66.7%), mobility (62.5%), pain/discomfort (61.1%), and usual activities (55.6%). Moreover, up to 76.9% of the respondents with two chronic conditions between 45 and 64 years of age reported having some problems with regular daily activities. Respondents with a low educational level tended to report having multimorbidity on all five dimensions of EQ-5D, remarkably, 100% of respondents with ≥ 3 chronic conditions reported having problem on self-care were not received any education or only completed primary school. No statistically significant difference was identified between male and female respondents.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Percentage of respondents with health problem on EQ-5D reporting multimorbidity by sex, age group and educational level.

https://doi.org/10.1371/journal.pone.0224970.t002

Table 3 presents the relationships between chronic conditions and EQ-5D health states in each dimension, with adjustments for background characteristics (sex, age, and educational attainment). Findings showed that respondents with physical disabilities were most likely to report having health problems on all five dimensions of EQ-5D, especially for self-care (OR = 30.78, 95% C.I [5.15, 94.51], p < 0.001). The Tobit model revealed that respondents with physical disabilities (beta = -0.28), mental problems (beta = -0.14), hypertension (beta = -0.05), or cancer (beta = -0.18) were significantly associated with having lower EQ-5D utility scores. Moreover, people living with multimorbidity were more likely to report health problem on all five dimensions of EQ-5D. For example, respondents living with three or more chronic conditions were 19 times more likely to have problems with self-care than those living without chronic conditions, and more likely to have problems with mobility (OR = 12.09), usual activities (OR = 7.59) and pain/discomfort (OR = 3.59). The Tobit model also revealed that respondents living with an increasing number of chronic conditions (one condition [beta = -0.045], two [beta = -0.046], and ≥three [beta = -0.187]) tend to have significantly lower utility scores.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. The chronic conditions and multimorbidity model on EQ-5D utility and its five dimensions.

https://doi.org/10.1371/journal.pone.0224970.t003

Table 4 shows the results of our MID estimates among HK general population. Both the mean and median MID estimate was 0.093 (S.D. of mean = 0.011, interquartile range IQR of median: 0.085 ~ 0.101). The maximum-valued scoring difference was 0.189, referring to the transition between level 3 (moderate) and level 4 (severe) of mobility. The minimum-valued scoring difference was 0.026, referring to the transition between level 2 (minor) and level 3 (moderate) of self-care dimension. Moreover, the adjusted mean MID was 0.071 (S.D. = 0.008). For respondents living with chronic conditions, the mean MID estimate was 0.083 (S.D. = 0.006) while the adjusted MID estimate was 0.079 (S.D. = 0.006).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. MID estimate of EQ-5D-5L utility score using HK scoring algorithm.

https://doi.org/10.1371/journal.pone.0224970.t004

Discussion

This study demonstrated the impact of chronic conditions and associated multimorbidity on HRQoL in HK’s adult general population. Overall, people reported having physical disability had lower EQ-5D utility scores than those with other chronic conditions. There was a significant association between multimorbidity and decreasing HRQoL, when controlling for socio-demographic characteristics. Our results are generally consistent with findings from previous studies in UK, France, Australia, the US [8,10,24,25] and other Asian countries, such as Singapore and Japan [26,27]. Moreover, our mean MID estimate of EQ-5D-5L for HK’s general population was 0.093 (adjusted MID = 0.071).

Our findings showed that HRQoL was significantly and negatively affected by the number of chronic conditions and the type of condition, particularly, involving physical disability. Previous studies have indicated similar findings in terms of the associations between physical functioning and HRQoL. for example, in a longitudinal study in Norway, Tollisen et al. found a strong correlation between physical disability and HRQoL in juvenile idiopathic arthritis patients [28]. Further, in a French study, Mouthon et al, found that patients with systemic sclerosis with reduced wrist and hand mobility had lower mental HRQoL [29]. It is important for policymakers to consider improving HRQoL by incorporating additional assistance measures in public spaces and providing barrier-free access throughout communities and natural environments, empower people with physical disabilities to live well in communities, which can improve their HRQoL. However, these findings should be interpreted with some caution. Physical functional impairments may impact respondents’ psychological status, and psychological instability may lead to physical discomfort, both of which could potentially lead to fluctuations of HRQoL [30–32]. However, our study was unable to distinguish between the fluctuations of HRQoL due to physical activity engagement and those due to psychological health improvement or deterioration. It should be further explored in the future.

Our study indicated that people living with multimorbidity were more likely to report worse HRQoL. The management of multimorbidity is complex, caused primarily by the challenges of specialized healthcare, which results in fragmented care, polypharmacy, multiple treatment burdens, mental health problems, and increased healthcare utilization that strains available resources [33–35], which was described by O’Brien as an “endless struggle” for comorbidity patients [36]. When dealing with increased threats from multimorbidity, two priorities need to be addressed by government and professional entities. First, chronic condition management should target functions and abilities necessary to manage daily household and social activities, not only the disease itself [4]. Second, revisions to clinical guidelines and research protocols are required because they currently focus only on care programs for single disease. The traditional clinical guidelines might be appropriate for patients living with a single condition, but the sum of clinical recommendations from various sources suggest different approaches for multimorbidity [37,38]. Further studies should explore the influence of different combinations of chronic conditions on HRQoL. Those findings could provide added value when designing guidelines for different chronic conditions which might be grouped or separately assessed. These predictions could be extremely useful for front-line medical staff seeking optimal approaches to care for patients with multimorbidity.

The study also found that educational level has a significant impact on the relationship between the EQ-5D utility score and multimorbidity. Some previous studies have reported that people with low educational level tend to have low HRQoL. For example, a study in Germany found that adults in low educational groups experienced increased levels of health impairments as well as lower HRQoL once health is impaired [39]. Another study that investigated esophageal cancer patients’ HRQoL in Sweden found that low education was associated with poorer functioning in certain HRQoL domains for women [40]. Similarly, a recent study in Malaysia that included 347 respondents in a low household income group found that the quality of life (QoL) was negatively associated with educational level and chronic conditions. However, few studies have explored the relationship between HRQoL and levels of education in the context of people with multimorbidity. Considering similar findings our previous study in specialist out-patient clinics in HK [41], we believe that in HK, people with a low educational level and multimorbidity have lower EQ-5D utility. One reason for this finding is that education can provide skills and knowledge that are essential for navigating the complex healthcare system. As educational inequity is intertwined with the expanding epidemic of multimorbidity [38], policymakers should pay close attention to these linked phenomena.

Additionally, age is not necessary inversely associated with multimorbidity patients’ HRQoL. In fact, elderly respondents, in our sample were less likely to report depression or anxiety than young respondents. Previous studies have similarly found that elderly patients are more likely to develop better resilience and adaptability to living with multimorbidity through their life experiences and wisdom, thus resulting in higher self-reported HRQoL [42]. However, an essential question must be asked: Does high utility score really reflect a better life for this population? Further, does better health equate to higher HRQoL? Apparently, the answer is not simply yes or no. Discussing the limitations of EQ-5D is beyond the scope of this paper (Brazier has discussed this in depth [43]). However, when we discuss people’s HRQoL, specifically those living with either a single chronic condition or multimorbidity, this study suggests HRQoL is impacted by things other than chronic conditions or multimorbidity. Otherwise, young people should have higher HRQoL than elderly people. Instead, factors such as equity, fairness, and public willingness might also have strong influences on well-being [44]. More effort and resources should be invested into developing valid and reliable instruments to evaluate people’s overall QoL, rather than just HRQoL [43].

Although the evidence is limited, the estimation of MID using generic preference-based measure is highly recommended to provide insight on the evaluation of clinical interventions from the perspective of HRQoL [45]. Crosby et al. indicated that for patients, the change in HRQoL may reflect the reduction in symptoms or improvement in function, whereas for doctors, the change may reveal the effectiveness of the treatment or in the prognosis of the illness [46]. The MID estimate could help us to understand what change in index score is clinically meaningful from a comprehensive perspective. Our study found the MID estimate for our sample of HK’s general population is higher than the MID of other jurisdictions [45], which may illustrate that a higher price our social care system has to pay for improving the local people’s health. The MID could be considered as a threshold effect providing information that a clinical treatment could not be indicated and offers the most direct considerations for both benefits and harms from the patient perspective [47]. Given these values have clinical and social implications in defining and comparing the people’ health states, MIDs for different conditions, especially chronic conditions, should be estimated. Thus, we will report our MID estimation using EQ-5D index values for different physical and mental conditions in the following studies.

One strength of using EQ-5D is that the specific HK EQ-5D-5L value set is well-established, thereby overcoming some cultural biases that might arise when using other jurisdiction’s values. Another strength is that our results could be used as a baseline to conduct the economic evaluation in HK and ensure consistency with evidence in the future studies. In addition to the strengths already mentioned above, there are some limitations worth noting. First, the ceiling effect was more than 45% in our study, which could influence the precision of calculating quality-adjusted life years in certain economic evaluations. Another one is based on cultural considerations, in that, we did not include the actual income as a variable into our survey, and instead used employment and living status as surrogate variables; actual income may be an important factor influencing QoL. Also, the sample size of patients in our study who reported having multimorbidity was not big, and thus a large population-based multimorbidity survey is needed in future studies. The last limitation is that chronic conditions in our study may not have had adequate precision, given that they were self-reported by individuals and possibly influenced by recall bias. In future studies, we suggest using the ICD-10 codes of the International Classification Diseases to define health conditions.

Conclusion

This study explored the burden of chronic conditions and multimorbidity on HRQoL and primarily defined the MID estimate using EQ-5D-5L in HK’s general population. The findings suggest that HRQoL was negatively affected by the chronic conditions of physical disabilities, mental problems, and hypertension, as well as multimorbidity, all of which were strongly associated with low HRQoL. Reforming the healthcare system address foreseeable challenges arising as more patients live with chronic conditions and multimorbidity could improve service efficiency and the effectiveness of healthcare interventions and policies, and, ultimately, improve the HRQoL of HK citizens.