Association between chronic conditions and health-related quality of life: differences by level of urbanization in Peru

Depressive mood and chronic bronchitis, but not HTN nor T2D, were associated with lower EQ-VAS scores. Site was not an effect modifier for the association between EQ-VAS score and having HTN, chronic bronchitis, and T2D. However, site was an effect modifier of the association between EQ-VAS and depressive mood: the EQ-VAS score was lower in people with depressive mood than in those without depressive mood, but this difference was larger in Lima than in rural Puno with a clear pattern of a gradient in the magnitudes of difference in EQ-VAS scores observed across the range of urbanization profiles. Regarding the number of chronic conditions, it was inversely associated with EQ-VAS scores. The association between number of chronic conditions and EQ-VAS scores was similar across all study sites.

Depressive mood was the chronic condition that seemed to have the larger impact in QOL, similar to studies in Finland [8], Canada [11], Spain [27], and the United States [28]. However, since these studies were also cross-sectional, it is possible that, at least in some cases, depressive mood could be the cause rather than the consequence of low QOL. Thus, longitudinal approaches are needed to identify the real impact of depressive mood in QOL.

Chronic bronchitis also had a high impact on QOL, as found in studies that assessed chronic pulmonary conditions (chronic bronchitis, asthma, COPD) [7, 29, 30], which is attributed to the impact of chronic cough, shortness of breath, and phlegm production [7, 31‐34]. This suggests that chronic bronchitis should be prioritized in primary care programs in our sites.

EQ-VAS score in persons with T2D was on average 1.3 points lower than in those without T2D. Conversely, in a population study in Korea, EQ-VAS score was 4.6 points lower in persons who self-reported T2D than those who did not [35]. This variation may be due to our inclusion of not only self-reported T2D cases, but also of those with abnormal point-of-care fasting blood glucose. As expected, the very nature of having T2D or hypertension, be it a short duration with the condition or having them without complications, could explain the little impact of these conditions on QOL found in our study [36, 37]. To evaluate this hypothesis, we made a post hoc linear regression in our population, and found that those receiving a prescription for T2D medications (i.e., those with a T2D diagnosis) scored 2.6 points lower on the EQ-VAS than those not on hypoglycemic agents (p = 0.10). This suggests that QOL was diminished among participants who were aware of their diagnosis, although other variables such as aware of the gravity and the complications of T2D could also play a role in subjective QOL.

For each additional chronic condition, our participants showed a 3.4-unit reduction in their EQ-VAS score. This reduction is substantial, since data from a 5-year cohort population study of elderly Italians found that one unit decrease in EQ-VAS was associated to a 1% increase in mortality and hospitalization rates [38]. As such, strategies to improve QOL in persons with chronic conditions are needed. Increasing healthcare access for physician visits, incorporating disease self-management skills [39], and designing individualized patient regimen with consideration of treatment side effects [40] could potentially decrease treatment burden and therefore improve QOL.

EQ-VAS score was 3.4 points lower for each additional chronic condition. In comparison, a previous study in the United Kingdom found that EQ-VAS score was 5.9 points lower for each additional chronic condition [9]. This study evaluated six chronic conditions: HTN, T2D, cerebrovascular disease, ischemic heart disease, COPD, and asthma, but it did not evaluated depressive mood, which had the highest impact in QOL in our study. Thus, it is possible that the EQ-VAS score per chronic condition would be even lower in this UK study if depressive mood had been included. Then, there could be a great difference between this study and ours in terms of QOL affectation by chronic conditions. This could be explained by the severity of the United Kingdom subjects, who were registered with general practitioners who probably see patients with more advanced complications of their chronic diseases.

Average EQ-VAS score was higher in Lima and semi-urban Tumbes, intermediate in urban Puno, and lower in rural Puno. This linearly reflects the degree of urbanization, with the exception of semi-urban Tumbes, with values similar to urban Lima. It is possible that semi-urban Tumbes dwellers have strong social or familiar support, similar to those in rural areas as described elsewhere [41], while at the same time having better healthcare access, similar to urban areas. This could also explain the low depressive mood prevalence in this semi-urban site.

Lower QOL values in the rural site could be due to less education, poor economic status, and limited access to health services; all increasing their risk of developing severe health complications [14‐16]. In addition, people living in rural areas are more likely to be involved in manual labor and may require more physical skills to perform their usual activities; thus, the decrease of physical capabilities in rural dwellers could have a greater impact on subjective wellbeing than in urban dwellers. Such regional observations would suggest that a higher number of chronic diseases would more dramatically affect QOL in rural than urban sites. However, our results show that EQ-VAS scores per number of chronic conditions were similar across studied sites.

This is probably because the average EQ-VAS score in those without chronic conditions was higher in Lima than in rural Puno (75 and 62 points, respectively). Due to the subjective nature of the EQ-VAS assessment, the same condition could cause a larger decrease in QOL in Lima than in rural Puno [42]. If this is true, it would be incorrect to compare linear regression coefficients with EQ-VAS; therefore, future studies should evaluate this hypothesis comparing EQ-VAS with other more objective QOL questionnaires. Another potential explanation is that rural dwellers may have a higher emotional resilience, which could diminish the subjective effect of chronic conditions on their QOL, as seen in the lower impact of depressive mood on EQ-VAS in rural sites. This should be more thoroughly studied to better understand what factors predict and define QOL for people living in rural versus urban sites.

However, few studies have evaluated the differences of social support, familial support, healthcare access, education, economic status, and labor characteristics, across Peruvian sites, and especially in sites with a heterogeneity of urbanization degree patterns. Moreover, there is a paucity of information about the impact of these variables in objective and subjective QOL. Understanding what levels or profiles of urbanization are directly linked with be deeply studied in order to understand QOL across different sites warrants further evaluation.

The most prevalent EQ-5D dimension was pain and the least prevalent was self-care, similar to previous studies [9, 11, 43]. All dimensions but anxiety/depression had a higher prevalence in the rural site, while the lowest prevalence across all EQ-5D dimensions was observed in semi-urban Tumbes. This could relate to the rural contextual features and the Tumbes familiar/social support described above.

This study compared the association of multiple and specific chronic conditions and QOL within urban and rural sites. Its results might help us to understand how health-related QOL varies among intra-national sites.

Limitations of our study are as follows: (1) not all chronic conditions with chief impact on QOL were included, such as cancer, chronic pain, and arthritis [10, 11], so that the inclusion of these conditions could increase the impact of number of chronic conditions on EQ-VAS. (2) Some of our diagnostic criteria for chronic conditions are not the current gold standard, such as blood pressure measure in a single day for HTN, and the evaluation of a single fasting blood glucose sample for T2D. In addition, our chronic bronchitis diagnosis was based on symptoms presented during the last 12 months while other studies required symptom persistence for at least 2 years to consider a diagnosis of chronic bronchitis [24]. These evaluations could be over-estimating the prevalence of some chronic conditions, although they could be more accurate than the self-report used in similar previous studies [35‐37]. (3) Diminished sample representativeness, especially in older and poor participants, could be underestimating the association, since these subjects could have a greater QOL decrease due to insufficient care of their chronic conditions. (4) Low sample size, especially in rural Puno, which could affect the extrapolation of our results to this site.