ABC | Volume 112, Nº6, June 2019

Original Article Fernandes et al Lifestyle and costs of medicine use Arq Bras Cardiol. 2019; 112(6):749-755 In the multivariate model considering changes over time in lifestyle behaviors (Model-C), age and body fatness were positively related to higher 12-months medicine costs. Model-C explained 13.1% of all variance in the medicine costs. In Model-D (sum of baseline and follow-up values), sleep quality had a positive relationship with medicine use (Table 3). On the other hand, alcohol consumption was negatively related to costs of medicine use. Model-D explained 21.7% of all variance in the medicine costs. Discussion This study shows that lifestyle behaviors particularly worse sleep quality leads to higher costs related to medicine use. Body fatness was also found to be an important predictor – positive effect on costs. In overall, 52.5% of adults reported any medicine use during the cohort period, while 20.5% (n = 24) of these same adults reported the use of three or more medicines. These rates are similar to Brazilian (18%) and European (20%) surveys, in which population-based samplings were carried out. 4,5 Another similarity with previous studies observed in our findings is that drugs for the treatment of cardiovascular diseases were the most reported by the participants. A study carried out in New Zealand examining the trends of medicine use in adults aged ≥ 65 years from 2005 to 2013 identified that the use of drugs to prevent cardiovascular events (aspirin and statins) increased significantly. 3 The dynamics observed for medicines to the treatment of cardiovascular diseases seems affected by aging as well (in our study, a relevant covariate in the multivariate models). Previous data have identified that consumption of aspirin and dipyridamole increased in older adults at a higher rate than observed in younger ones. 3 The increased amount paid by older adults can be supported by the natural effects that ageing exert over organs of the human body and their functions, 23 but also boosted by the reduced PA observed in older groups. 24 In the analyzed sample, although the effect of age on costs with medicine use was not mediated by PA, they were related with other in crude analyzes, denoting the relevance of actions targeting the improvement of PA practice mainly in population groups composed of older adults. 23,24 In this sample, the higher cost with medicines in adults with sleep disorders can represent not only the treatment of the sleep disorders itself, but also the use of medicines to relief its symptoms and hence to maintain the daily activities, such as work. 13,25 The findings related to alcohol consumptions were surprising because usually the higher alcohol consumption is linked to higher health care costs, 26,27 and not the opposite as observed in our study. In fact, the linkage between alcohol consumption and health care costs can be direct (e.g. diseases directly linked to alcohol consumption) and indirect (e.g. traffic car accident), but it is important to take into account that some kinds of alcoholic drinks have healthy characteristics, such as anti-inflammatory properties observed in the red wine. 28 Therefore, the explanation for our interesting finding could be due to both the type and amount of alcohol consumed. However, our study looked at only the number of days per week with alcohol consumption, and not amount and type of alcohol consumed, which characterizes a limitation in our study. Other limitations of the study are worth mentioning. The first limitation of the study is the small sample size. The current study has statistical power of 80% to detect coefficient of correlation of 0.256 or higher, while the relationship between some behaviors and costs with medicine are around r = 0.110. 16 Even considering the fact that the inclusion of covariates increases the power of multivariate models, 29 the reduced sample size may have been responsible for the absence of significant relationship between PA and costs with medicine. Another limitation related to objective measures of PA is the logging of data by the participants because every day they had to note the number of steps displayed in the pedometer. Although this method is widely used, 19 it could have led to misreporting. As above mentioned, the absence of measures of amount and kind of alcoholic drinks and sedentary behavior (by screen time on TV or computer) constitute limitations as well. Further studies could explore the impact of these. Conclusions Worse sleep quality seems to increase the costs related to medicine use in adults, while obesity and ageing play a relevant role in this phenomenon. Moreover, alcohol consumption seems a variable with relevant economic impact, but further studies are necessary to identify clearly the direction of its relationship with medicine costs. Acknowledgment To the Fundação de Amparo a Pesquisa do Estado de São Paulo (FAPESP) (process number: 2017/50026-7 and 2015/20460-1) and to the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES). Author contributions Conception and design of the research: Fernandes RA; Acquisition of data and Analysis and interpretation of the data: Mantovani AM; Statistical analysis: Mantovani AM, Anokye N; Writing of the manuscript: Codogno JS, Turi-Lynch BC, Anokye N; Critical revision of the manuscript for intellectual content: Codogno JS, Turi-Lynch BC, Pokhrel S. Potential Conflict of Interest No potential conflict of interest relevant to this article was reported. Sources of Funding There were no external funding sources for this study. Study Association This article is part of the thesis of Doctoral submitted by AlessandraMadiaMantovani, fromUniversidade Estadual Paulista. 753