The main results of the present study show lower OOP expenses in medicines among the participants speaking Quechua or other languages, the gap being greater in women, among persons over 60 years old, and among those who belong to the first level of care. The reasons for these findings were explained only 43% by the observable differences of the variables considered.
In recent years, coverage of access to health insurance has grown significantly in Peru, facilitating access to more excellent protection for people of diverse socioeconomic and ethnic conditions 2. Health insurance is relevant in a country where regions with ethnic minorities still have disadvantages related to health. For example, women from ethnic minorities such as Quechua and Aymara present disadvantages in different indicators of maternal and child health [17, 18]. Similarly, mistreatment when using health services has been documented in Quechua, Asháninca and Nomatsiguengas patients and patients from the Mantaro Valley in the central Andes Mountains of Peru [19, 20]. Health insurance in Peru increased from 60.5% in 2008 to 76.4% in 2017. In this way, the Universal Health Insurance (SIS) coverage, which is the insurance for the population with fewer resources, increased from 34% in 2008 to 47% in 2017 2. Being a woman increased the possibility of affiliation to the SIS, while being between 18 and 39 years old, residing in Metropolitan Lima and being non-poor reduced this possibility 2. This increase in coverage was not homogeneous; in the departments where many ethnic minorities live, this coverage was less than 50% in 2017 2.
This evidence perhaps explains that OOP payment in health is still an issue in purchasing drugs in Peru. One study estimated that, on average, OOP expenses in drugs were US$ 8.14 in 2007, compared to US$ 9.68 in 2016 13. This apparent discrepancy between more excellent insurance coverage in Peru and an increase in OOP payments may be because a greater demand for services due to the extension of insurance to the SIS highlighted the weakness of the service offer and probably that this was deteriorating 1.
However, according to our results, there are fewer OOP expenses for ethnic minorities. Likewise, there has been reported ineffective access to medicines in the vast majority of health establishments, despite having accessed outpatient care 21. Therefore, access to the physician is achieved but not to the necessary medicines leading to an increase in OOP payment.
Unlike our study, other authors have found higher OOP costs in some ethnic minorities. For example, a study in the United States showed that Mexicans had a higher proportion of OOP payments than Whites after adjusting for socioeconomic and demographic factors 22. Another study in the same country showed that being African-American lowered costs for Medicare-managed diabetes care, although it increased the disparity between total healthcare costs and OOP expenses23. In contrast, another study showed that although racial and ethnic disparities were initially observed between non-Hispanic Whites and Hispanics and between non-Hispanic Whites and non-Hispanic Blacks in the OOP expenses ratio for older adults in the United States. These disparities disappeared after controlling by utilization or health needs 24.
On the other hand, these studies were carried out in a specific health system and considered different ethnic minorities from those evaluated in our research; however, some aspects were found in these studies and may explain our findings. Probably, as observed in the study of older adults, the use of health services is also a variable to consider 24; in that sense, this apparent decrease in OOP expenses in the Peruvian ethnic minorities could be due to a lower use of health services. Indeed, it has been reported that, in the rural population, which is the area where a large part of these minorities lives, more than 60% do not seek medical attention 25.
Likewise, another study that evaluated the non-use of formal health services found that 62% in the Highlands and 56% in the Amazonian region do not use health services despite needing them, territories where ethnic minorities frequently live in Peru26. Although various reasons explain it, including aspects such as distance and difficulties in acceding health centers in rural areas, discrimination, among other reasons, language may also be a factor to consider. A study observed that Latinos in the United States were less likely to use drugs compared to Whites; it was found that one of the most critical factors for this disparity was the limited English proficiency 22. These differences in language proficiency could also explain that people over 60 years of age had less OOP expenses as they were probably less prone to proficiency in Spanish and therefore made less use of health services. On the contrary, it is also probable that the greater use of health services and the greater probability of obtaining care decreases the OOP payment for women 27.
However, our findings are explained to warrant a more in-depth analysis of unobservable characteristics. Our results showed that less than 40% of the gap was explained by our variables for women, while it was less than 25% for older adults. The unexplained component of the gap in the OOP expenses for drug purchase was higher than the component explained by observable characteristics. These results contrast with a study that evaluated the use of different medications by comparing groups by the different ethnic conditions, including Hispanics, non-Hispanic Afro-descendants, and non-Hispanic whites 28. This study showed that observable characteristics mainly explained the differences in the use of therapeutic drugs for Afro-descendants and Whites. Although our study used a nationally representative survey, our analysis was subject to the variables pre-established, limiting the possibility of analyzing other unobservable factors that explain our findings. Although there are studies that have evaluated the issue in other ethnic minorities 11,12, to our knowledge, there are no studies that have evaluated the ethnic differences of OOP expense in medicines for ethnic minorities present in Peru 14. This point is relevant because, for many ethnic minorities, the concept of health is linked to the community beyond the individual and includes an essential spiritual element firmly related to the well-being of the ecosystem and the planet 29. These elements, associated with their worldview, have not been evaluated and perhaps are part of the unobservable characteristics that explain our findings.
According to the level of care, the component explained with the variables studied was about 18% in the second level and 40% in the first level. According to the technical standard of the categories of health establishments of the MINSA, the first level of care includes health centers attended by non-medical or medical professionals, including hospitalization services, to care for less complex problems. In the case of second-level establishments, general or specialized care hospitals are included 30. Although we cannot establish which aspects the unexplained component includes according to the level of care, they may include variables related to some of the characteristics of the health care of Quechua and Aymara patients. Although, according to the 2019 National Household Survey, 82.2% and 68.3% of the Quechua and Aymara population, respectively 31, had health insurance, there are still problems in the use of health services, which would increase the probability of not using them and increase the OOP. A study showed that, although patients belonging to Amazonian peoples used MINSA's unpaid dental services, Aymara patients paid for them, probably in private services 32. Various structural and cultural barriers prevent the use of health services, such as distance from establishments or lack of health personnel of the same ethnic group 33; an additional factor is mistreatment34. Another study showed that being a Quechua patient or speaking a native language doubled, and being of Aymara ethnicity tripled the probability of not going to health services due to abuse 34. Although these studies evaluated all health services, regardless of the level of care, it is likely that these barriers are similar at all levels of health and perhaps more so in second-level services, where problems such as obstetric-gynecology are treated. Indeed, for example, delivery care in the Aymara population of Puno includes beliefs 35 that make them more likely not to use the free health services of MINSA and therefore increase OOP.
The present study has certain limitations. First, as secondary data analysis, it does not have a specific design for analyzing the research question. Second, the analysis is based on a single health indication through the OOP payment in medicine. Third, the analysis uses a measure of the person's ethnic status based on language, with the possibility that we have underestimated the proportion of people from these ethnic minorities. Fourth, many of the variables have been self-reported so that they could be misclassified.
In conclusion, there was less out-of-pocket expenditure on medicines in those ethnic minorities, the gap being greater among women and those over 60 years old. The observable differences explain only 40% of these gaps.
We recommend following the intercultural approach that adapts documents to native languages or disseminated by trained people from their communities to identify unobservable elements that explain the health aspects of ethnic minorities in our country 36.