Article Text

Download PDFPDF

Original research
SARS-CoV-2 self-testing in Peru: a cross-sectional survey of values and attitudes of the general population
  1. Guillermo Z Martínez-Pérez1,
  2. Sonjelle Shilton1,
  3. Patricia Silvia Mallma Salazar2,
  4. Paola Pflucker Oses2,
  5. Paola Alejandra Torres-Slimming2,
  6. Deepshikha Batheja3,
  7. Abhik Banerji3,
  8. Amber Mallery1,
  9. Elena Ivanova Reipold1,
  10. Cesar Carcamo2
  1. 1 FIND, the global alliance for diagnostics, Geneva, Switzerland
  2. 2 Facultad de Salud Pública y Administración, Universidad Peruana Cayetano Heredia, Lima, Peru
  3. 3 Center for Disease Dynamics Economics & Policy, Delhi, India
  1. Correspondence to Dr Guillermo Z Martínez-Pérez; gmartinezgabas{at}


Objectives To assess the values of and attitudes towards the use of rapid SARS-CoV-2 antigen-detection tests for self-testing in a rural and an urban area in Peru.

Design Cross-sectional, street-based population survey.

Setting A series of over 400 randomly selected street points in Valle del Mantaro and in Lima.

Participants 438 respondents (203 female) participated. They were all older than 17 years and provided informed consent for participation.

Intervention All respondents answered on the spot, a 35-item questionnaire developed in KoboToolbox.

Primary and secondary outcome measures Primary outcomes of interest were: likelihood to use a SARS-CoV-2 self-test; willingness to pay for a SARS-CoV-2 self-test and likelihood to comply with recommended actions following a positive SARS-CoV-2 self-test result. Bivariate analyses and Poisson regression (PR) analyses were performed to identify significant associations between dependent variables and independent variables pertaining to respondents’ characteristics, risk perception and previous experiences with conventional COVID-19 testing.

Results Of the 438 respondents, 51.49% had previous experience with conventional COVID-19 testing; 20.37% had COVID-19 disease; 86.96% accepted the idea of SARS-CoV-2 self-testing; and, 78.95% would be likely to use it if needed. Almost all (94.75%) would pay for a self-testing device (mean acceptable payment: US$10.4) if it was not provided free of charge by health authorities. Overall, 93.12%, 86.93% and 85.32% would self-isolate, report the results and warn their contacts, respectively. Being a female (adjusted PR 1.05, 95% CI 1.00 to 1.09, p<0.018), having completed secondary education (adjusted PR 1.18, 95% CI 1.02 to 1.37, p<0.024) and expressing likelihood to use self-testing (adjusted PR 1.08, 95% CI 1.01 to 1.16, p<0.0.24) could be predictors of willingness to pay for a self-test.

Conclusions Self-testing is perceived as an acceptable approach. Health authorities in Peru should facilitate access to this approach to complement healthcare facilities-led testing efforts for COVID-19. Future research is necessary to understand the impact of self-testing in case detection and pandemic control.

  • COVID-19

Data availability statement

Data are available on reasonable request. Data will be available on reasonable request by contacting the corresponding author in this article.

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See:

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


  • The survey sampling method was designed to ensure diversity in terms of location of residence, proximity to health facilities and socioprofessional activity.

  • This study was conducted in accordance with a standardised research methodology, which was harmonised for its operationalisation in eight middle-income countries.

  • Social desirability might have influenced how survey respondents indicated how they would react if they received a positive self-test result.

  • Findings might have been different if the survey methodology had combined street-based and household-based recruitment strategies.

  • Collection of information around respondents’ attitudes towards SARS-CoV-2 self-testing did not consider if attitudes could be affected by changes in the COVID-19 incidence rates in the two survey geographies.


The first case of the COVID-19 caused by the SARS-CoV-2 was confirmed in Wuhan in December 2019.1 The WHO declared the COVID-19 a pandemic just 5 days after the first case of COVID-19 in Peru was reported on 6 March 2020.2 In November 2021, the highly contagious Omicron variant of concern appeared in South Africa.3 The public health impact of the Omicron has been of such magnitude that, even in industrialised countries such as China,4 the health authorities have not been able to keep up to pace with their societies’ testing needs and have opted to approve new approaches to scale up testing such as the use of rapid SARS-CoV-2 antigen-detection tests for self-testing.

Find and test, trace and isolate (FTTI) for both symptomatic and asymptomatic individuals was a crucial strategy to halt the spread of the SARS-CoV-2 at the peaks of the pandemic.5 To ensure FTTI has an impact, community-grounded interventions to accelerate the public’s access to COVID-19 diagnostics are needed to complement health authorities’ testing efforts. At the start of the pandemic, professional rapid antigen tests (RATs) in the form of lateral flow assays were deployed for healthcare practitioners to test symptomatic patients and their contacts in a ubiquitous mode. However, in the Pan American Health Organization (PAHO) region, various countries have expressed, even before the upsurge of the Omicron, their limitations to scale up professional-led mass screening for COVID-19.6–8 In parts of the PAHO region, while having primary healthcare settings dedicated to screening for COVID-19 has been helpful to decrease COVID-19 incidence rates, this approach has also deviated attention and resources away from other conditions such as HIV9 or rheumatic diseases.10

Various community-based FTTI approaches have been implemented in Peru.11 12 As of February 2023, over 4,5 million cases have been confirmed in the country.13 To better assist Peru’s Ministry of Health (MINSA, as per its Spanish acronym), which has a vast territory to cover, new diagnostic paradigms to empower lay people even in the hardest-to-reach Amazonia and Andes areas are necessary. Rapid SARS-CoV-2 antigen-detection self-testing (hereafter, just self-testing) is among such new paradigms. Same as professional RATs, self-testing devices have the form of lateral flow assays,14 and have been already recommended by health authorities of the European Union,14 India,15 Malaysia16 and Brazil,17 among other jurisdictions.

In Peru, research has shown how self-diagnostics for HIV are acceptable among at-risk populations.18 19 In this context, self-testing for SARS-CoV-2 could be an additional tool in the hands of a population that has undergone strict movement restrictions.20 To assist the population in need of testing and to relieve the burden on public testing facilities, the Peru’s Minister of Health announced that self-testing was approved for commercialisation in pharmacies at the start of 2022.21 22 The announcement caught stakeholders in the health sector by surprise as there was no locally produced evidence on how the Peruvian public might accept and use self-testing. To tackle this knowledge need, this article reports the findings of a survey which was conducted, in a rural and in an urban area in Peru, and prior to the approval of self-testing by the Ministry of Health, with the aim to assess what could potentially be the public’s values and attitudes towards self-testing.


Design, population and sites

This was a street-based survey conducted in September 2021. The survey population was the general public of two geographies: Lima, Peru’s capital and Valle del Mantaro. These geographies were selected to engage survey respondents whose perspectives on self-testing could be considered as proxies for the perspectives of those who live in urban and rural areas, respectively.

The sample size was calculated separately for both geographies with the aim to understand if geography-based differences might influence the respondents’ values towards self-testing. It was estimated that 196 or more respondents in each geography had to be recruited to have a confidence level of 95% that the real value (likelihood of using self-testing) is within±7% of the measured value. This sample size will have a statistical power of 82% to detect urban/rural differences of 15% or more in the likelihood of using self-testing, for values around 50%.

Sampling and recruitment

A multistaged sampling process to identify potential survey respondents was applied. The process was proposed with the purpose to avoid sampling by convenience or proximity.

First, the boundaries of the two geographies were defined in Google MyMaps. Once defined, the maps were divided into 40 areas of similar width with the intention that persons from areas which could differ in terms of proximity to health facilities or of engagement in socioprofessional possibilities could have a chance of being approached as survey respondents. Second, using RANDOM.ORG, the two lists of these 40 areas were reordered. The first 14 areas in the newly arranged lists were chosen as recruitment areas. Third, the two lists of the 14 areas were randomly reordered again to determine the sequence that the study site would follow in visiting each area during the survey conduct. Fourthly, in each of the 14 areas, a series of randomly selected street points were manually marked. When the process ended, the maps were uploaded into the ViewRanger app.

As an additional step, the two lists of the 14 areas were randomly reordered. The areas were then consecutively assigned to a geography-specific survey schedule which comprised a Monday to Sunday morning and an afternoon shift. Thus, during the survey conduct, which lasted a full week in each geography, the surveyors arrived in the area assigned in their respective schedules and walked from one street-point to the next. They had to recruit only one respondent in each street point. When arriving at each street point, they stopped the first passer-by they would see, and would invite them to participate. Approached street passers-by had to be older than 17 years old, willing to give informed consent, and without symptoms of COVID-19 to be considered eligible for participation.

If the individual met the inclusion criteria but declined participation, the surveyors waited 3 min until they could stop a new passer-by. If the individual met the inclusion criteria and agreed, they were asked to sign two copies of the informed consent and to keep one signed copy. Respondents received no incentives for their participation other than a pack of face masks.

Data collection and analysis

A 35-question structured questionnaire was used. This was a questionnaire that included questions on respondents’ sociodemographics (country, age, gender, ethnicity, education, employment status); previous experiences with conventional COVID-19 testing; knowledge of other self-test kits; likelihood-to-use and willingness-to-pay for a self-test; barriers to using self-testing; and likely actions on self-testing reactive to SARS-CoV-2.23 This questionnaire was designed in English, translated into Spanish and prepiloted in Lima. When a final version was defined, the questionnaire was developed in KoboToolbox.

Responses were collected on-the-spot where privacy could be guaranteed. The surveyors used the KoboCollect app to capture the responses, and to submit them to the KoboToolbox server. Submissions were monitored daily. Once the survey week ended, the questionnaire was archived to impede further submissions, and all data were exported for analysis.

Univariate descriptive and bivariate and multivariate analyses were performed in software STATA V.14. Missing data were imputed with mean response to the question observed in the overall data. The dependent variables of interest were: the likelihood of using a self-test; willingness to pay for a self-test; potential to comply with recommended actions following a positive self-test result (i.e., self-isolating, requesting counselling, warning contacts, reporting the result). Significant associations, at a p<0.05, were sought between respondents’ characteristics, dependent variables and other independent variables of interest, such as: experiences with conventional COVID-19 testing; willingness-to-weekly-use self-testing; awareness of other self-testing devices and COVID-19 risk perception.

Bivariate analyses were performed to explore significant associations between each dependent variable and other independent variables. The variables significantly associated with the dependent variables in the bivariate analyses were entered into a multivariate regression model. A Poisson regression model was used to identify associations between likelihood to use a self-test, willingness to pay for a self-test, and potential characteristics of the respondents. The results are interpreted as prevalence ratios. Ordinary least squares (OLS) regression was used for the perception of utility through the construction of a four-item index to identify potential drivers for compliance with recommended actions after self-testing positive.

Patient and public involvement

No patient involved in this study.


Participants’ characteristics

Of the 499 approached passers-by, 438 consented to participate (table 1). The majority of passers-by declining participation expressed that they lacked time. Among the consenting participants, 203 (46.34%), 233 (53.19%) and 2 (0.45%) respondents self-identified as female, male and non-binary, respectively. A total of 174 (39.73%), 174 (39.73%) and 90 (20.55%) were in the 18–35, 36–55 and 56 years and above age groupings. Education levels were high across all gender and geography groups, with less than 1/10th (9.66%) of respondents having completed none or primary education only. Regarding occupation, there were gender-based disparities to note: overall, rates of unemployment (34.16% females, 5.60% males), full-time employment (18.32 females, 33.19 males) and of self-employment full time (13.86% females, 27.16% males) also varied across genders. Among the just 103 (23.51%) respondents who reported their ethnic identity, 2 (0.48%), 2 (0.48%), 1 (0.24%), 96 (22.97%) and 2 (0.48%) self-identified as Apurimac, Aymara, Jewish, Quechua and Wanka, respectively (data not presented).

Table 1

Participants’ characteristics and experiences with COVID-19

Experience with COVID-19 testing

Just 59 (13.47%) respondents felt no risk of COVID-19 (table 1). Half of the sample felt at either moderate risk (32.88%) or at high risk (12.10%) of COVID-19. Almost half (47.94%) of respondents felt that they were not living with anybody at risk of severe COVID-19 disease. Overall, 20.37% had COVID-19 disease confirmed by a professional-delivered test. Among the respondents who had COVID-19, almost all (n=83/89, 93.25%) self-isolated.

Of the sample, 225 (51.49%) had tested at least once for COVID-19 (table 1). On average, the respondents highly rated the convenience of the last COVID-19 test they received (i.e., 3.69/5, SD 1.03). Almost half of the individuals who had tested (n=101/225, 44.88%) reported that they had felt at least once that they could not access testing when they needed it. Another 50 (11.41%) respondents who had never tested also reported having felt at least once that they could not access COVID-19 testing when they needed it.

Acceptability of self-testing

There was high agreement with the idea of self-testing for any infectious disease (79.80% females, 95% CI 74.20% to 85.30%; 84.48% males, 95% CI 79.39% to 88.84%) (table 2). Agreement with the concept of self-testing at home for COVID-19 specifically was even higher (87.19% females, 95% CI 82.15% to 91.82%; 86.64% males, 95% CI 82.22% to 91.04%). Overall, 51.98% (95% CI 44.79% to 58.65%) and 27.72% (95% CI 21.38% to 33.78%) of females answered that they would be ‘likely’ or ‘very likely’, respectively, to use self-testing; and 44.21% (95% CI 37.78% to 50.6%) and 34.33% (95% CI 28.19% to 40.47%) of males answered that they would be ‘likely’ or ‘very likely’, respectively, to use self-testing.

Table 2

Acceptability of self-testing for COVID-19

Univariate analyses on this first dependent variable of interest suggested that feeling at moderate risk (p<0.048) and at high risk (p<0.006) of severe COVID-19, as well as considering convenient their last COVID-19 testing experience (p<0.018) were significantly associated with stating that they would be ‘very likely’ or ‘likely’ to use self-testing. Bivariate analyses confirmed that those feeling at moderate risk or high risk of COVID-19 (Poisson regression (PR): 1.13, 95% CI 1.03 to 1.24, p<0.010) could be more likely to use self-testing than those who felt at no risk or at low risk of COVID-19 (table 3). The multivariate regression analyses did not confirm as significant any of these associations.

Table 3

Bivariate and multivariate associations

Willingness to pay for a self-test device

Respondents were asked how much they would be willing to pay for a self-test device. Almost all respondents (94.75%, 95% CI 92.65% to 96.84%) would pay a mean of US$10.4 (SD 19.055) (table 2). For both female and male groups, the mean acceptable payment was more than twofold for respondents in Lima in comparison with respondents in Valle del Mantaro.

The bivariate analyses suggested significant associations between willingness to pay and being female (p<0.012); being aged 56 and above (p<0.027); having completed secondary (p<0.030) and college degree and above education (p<0.019); and expressing likelihood to use a self-testing (p<0.045) (table 3). The multivariate analyses confirmed that females, in comparison with males, would be more likely (adjusted PR 1.05, 95% CI 1.00 to 1.09, p<0.018) to pay for a self-test; that, compared with people with primary education only, people with completed secondary education (adjusted PR 1.18, 95% CI 1.02 to 1.37, p<0.024) and college degree of higher education (adjusted PR 1.18, 95% CI 1.02 to 1.37, p<0.020), would be more likely to pay for a self-test; and that, compared with respondents expressing unlikelihood to use self-testing, those expressing likelihood to use self-testing (adjusted PR 1.08, 95% CI 1.01 to 1.16, p<0.0.24), would be more willing to pay for a self-test device (table 3).

Actions on self-testing for COVID-19

The respondents were asked what they would do if they self-tested negative, but they knew that they are a close contact with a COVID-19 patient and they feel they have COVID-19 compatible symptoms (table 4). In this scenario, 58.49% (95% CI 53.84% to 63.12%) would stop self-isolating. Just 6.19% (95% CI 3.92% to 8.46%) would stop wearing masks, and 26.61% (95% CI 22.44% to 30.76%) would stop social distancing. Univariate analyses suggested a significant association between agreement with the idea of self-testing and expressing that they would stop wearing masks following a negative result, even if symptomatic (p<0.04).

Table 4

Likely actions on self-testing for COVID-19

In the event that the respondents self-tested and its result were positive, the majority would report it (86.93%, 95% CI 83.74% to 90.1%); request post-test counselling (73.62%, 95% CI 69.47% to 77.77%); self-isolate (93.12%, 95% CI 90.73% to 95.50%); and, warn their contacts (85.32%, 95% CI 81.98% to 88.65%) (table 4). In Valle del Mantaro, manifestations of willingness to warn close contacts were less frequent both among females (76.07%, 95% CI 68.22% to 83.91%) and males (83.96%, 95% CI 76.86% to 91.06%) in comparison with urban areas (91.76% females, 95% CI 85.80% to 97.72%; 91.27% males, 95% CI 86.27% to 96.26%).

Univariate analyses suggested that, following a positive self-test result, willingness to report it was associated with willingness to pay (p<0.043); willingness to self-isolate was associated with an agreement with the idea of COVID-19 self-testing (p<0.004), and the likelihood to use self-testing (p<0.001); while willingness to warn contacts was associated with an agreement with the idea of self-testing (p<0.001), willingness to pay (p<0.02) and the likelihood to use self-testing (p<0.001). Bivariate analyses suggested that respondents with any education other than formal education (coefficient −1.99, 95% CI −3.45 to −0.54, p<0.007) could have less odds to comply with recommended actions following a positive self-test; and that respondents who expressed likelihood to use self-testing (coefficient 0.48, 95% CI 0.21 to 0.75, p<0.001) could comply better with recommended actions following a positive self-test in comparison with respondents who did not express likelihood to use self-testing (table 3). The OLS regression model did not confirm any of these associations.


We carried out a survey to understand what could potentially be the general public’s values and attitudes towards SARS-CoV-2 self-testing in Peru. This survey was conducted in a rural and in an urban area in Peru four months before self-testing was approved by the country’s Ministry of Health.21 22 Despite the unavailability of self-testing for the public at the time, most respondents in our survey manifested agreement with the idea of self-testing (86.96%), likelihood of using self-testing if needed (78.95%) and willingness to pay for self-test devices if these were not provided for free by health authorities (94.75%).

The average maximum payment for a device that our respondents in Lima would consider acceptable (US$14.7 females; US$14.8 males) was twofold the average payment that our respondents in Valle del Mantaro would find acceptable (US$5.9 females; US$7.1 males). This finding is aligned with other studies on acceptability of self-testing which have reported that inhabitants in rural areas might be less willing to pay—or, have less means to pay—for a self-testing than inhabitants in urban areas.24 25 This finding suggests the need to assess the feasibility of subsidising the price of self-testing for end-users or to set a ceiling retail price on self-test kits, as the Government of Malaysia did in 2022,26 to make access to self-testing more affordable for the less economically advantaged households.

In our survey, although willingness to comply with recommended actions following a positive self-testing were optimal in both study geographies, there were less respondents expressing likelihood to warn their close contacts in Valle del Mantaro (76.07% females; 83.96% males) than in Lima (91.76% females; 91.27% males). In this regard, since self-testing kits have finally entered the Peruvian market, delivery of education using a wide array of communication formats on the actions the general public should take on if they receive a reactive self-test result might be needed in rural and lower-income settings to ensure that this approach will be used in compliance with public health authorities’ recommendations in the country.

Some strengths and weaknesses must be noted. Although meaningful, our findings need to be cautiously interpreted. A strength of this study is its large sample, which is characterised by diversity regarding age, sex, education or geography. Since this survey in two locations in Peru is a component of a multicountry study,23 their comparison with other countries’ findings is possible. Nevertheless, a limitation we must acknowledge is that, since this is cross-sectional exploratory research, it is not possible to fully understand what the causal drivers of self-testing acceptability are. Social desirability might have influenced how our respondents expressed how they would react if they received a positive self-test result.

An additional consideration is that our fi ndings might have been different if our methodology had combined street and household-based recruitment strategies. Had we opted for a household-based strategy, as a survey carried out in mid-2021 did in Kenya,24 we could perhaps have approached more individuals who seldom leave their homes (e.g., due to fear of COVID-19). Also, we consider that the respondents’ perspectives on their likelihood of using self-testing could have been different if our data collection had been conducted during a peak of COVID-19 incidence—when, supposedly, individuals’ perceptions of their risk of COVID-19 acquisition and concerns about the availability of diagnostic services might be higher.

We consider that the survey findings can be considered—as a proxy for what the Peruvian public perspectives on self-testing might be—as guidance for programming of self-testing delivery. However, we cannot affirm that the survey findings are representative of the whole country. In Peru, local governments were actively engaged in the public health response against the pandemic.27 Restrictions to travelling, working and social activities were not homogeneously applied across Peru. At the time of survey conduct in Lima and Valle del Mantaro, there were geographies other than our study settings which were under lockdown. Hence, attitudes towards self-testing might have been different in municipalities which were under worse epidemiological scenarios.

Despite these limitations, the added value of our survey is that its findings are comparable to those of other acceptability studies. In surveys carried out with the same methodology in South Africa,28 Kenya24 or Indonesia,29 acceptability of self-testing was also high as 90.39% (n=531), 81.4% (n=419) and 60.79% (n=630) of respondents in these countries, respectively, expressed likelihood of using a self-test if they needed it. Other quantitative studies carried out in high-income countries such as Cyprus and Greece30 or Germany31 have also reported high acceptability. In Cyprus and Greece, 79% of respondents (n=248) expressed willingness to use self-testing. In Germany, 89.3% of teachers (n=711) invited to participate in a school-based pilot project accepted to use self-testing and submit their results for evaluation. Qualitative research has also provided evidence of acceptability of self-testing among minoritised groups, such as black/African Americans in the USA,32 or among inhabitants of remote rural areas in South Africa.25

In essence, this study adds to the literature of acceptability studies in low-income and middle-income countries (LMICs). With the exception of HIV self-testing studies,18 19 33 literature on devices for self-diagnosis of infectious diseases are still scarce in resource-constrained territories even if, due to lack of health personnel and decentralised laboratory capacities, self-diagnostics could impact their populations’ health and well-being. Our survey proposes a methodology which, prior to its application in this study, has long been used by FIND, the Global Alliance for Diagnostics in settings such as Kyrgyzstan34 or Rwanda35 in the frame of rapid hepatitis C virus antigen-detection self-testing studies. There is a need to advocate for the conduct of acceptability studies in LMIC prior to the introduction of self-tests that have been only tested in high-income, industrialised countries. Population-based surveys, conducted alongside qualitative studies and self-tests performance and usability studies, can help emergency preparedness programmers design their introduction and scale up of self-testing kits for emerging and re-emerging infectious diseases in full consideration to people’s expressed rumours, fears, preferences, and to their own cosmovision of diseases affecting them.

Self-diagnostic acceptability studies can have special value for authorities to plan strategies to enhance users’ linkage to care after receiving a positive self-test result. We recommend that the predictors that our multivariate analyses identified with regard of self-testers’ potential non-compliance with recommended actions that would enable them to effectively link to care (i.e, report result, request counselling) be carefully considered by self-testing programmes planners in Peru and other resource-constrained countries in the region. There might be other predictors of non-compliance which this survey has failed to identify. Barriers to linkage to care identified by other qualitative studies on SARS-CoV-2 self-testing include fear of stigma, healthcare workers’ negative attitudes, cost of healthcare or denialism of the disease.25 32 33 In our study context, future research will be needed to understand how to ensure barriers to access healthcare do not deter Peruvians to report a positive self-test and request confirmatory testing, as well as to measure the impact of SARS-CoV-2 self-testing in case detection and pandemic control.

In conclusion, this survey shows acceptability of SARS-CoV-2 self-testing for the general public of both sexes who participated as respondents in a rural and in an urban area in Peru. This survey suggests that individuals living in large cities in Peru might be more compliant than rural inhabitants with recommended actions following a negative or a positive result. Interventions to educate the public on the purpose and utility of self-testing, and on processes to link to COVID-19 care must be tailored to the public paying careful consideration to geography-based and gender-based constraints to access healthcare. Health authorities in Peru should consider measures to subsidise access for low-income households to self-testing as an evidence-based screening approach to complement healthcare facilities-led COVID-19 case detection efforts.

Data availability statement

Data are available on reasonable request. Data will be available on reasonable request by contacting the corresponding author in this article.

Ethics statements

Patient consent for publication

Ethics approval

This study involves human participants and the study received ethical clearance by the Universidad Peruana Cayetano Heredia (Ref.: CIEI Code 205954). Participants gave informed consent to participate in the study before taking part.


The authors of this article are indebted to all survey respondents and surveyors in Lima and in Valle del Mantaro, who made this research possible.



  • Contributors SS, EIR and GZM-P developed the initial research project. CC and PSMS adapted research protocol and led study implementation in Peru. GZM-P, SS and CC wrote the manuscript. PSMS, PPO and PAT-S coordinated data collection. GZM-P, AM, DB and AB performed data analyses. All authors have reviewed the manuscript and contributed to its conduct. SS and EIR act as guarantor.

  • Funding This work was supported with funding provided by the German Ministry for Education and Research (BMBF) through KfW (grant number n/a).

  • Disclaimer The funders played no role in the study design; in the collection, management, analysis, or interpretation of the data; in writing the report; or in the decision to submit the report for publication.

  • Competing interests None declared.

  • Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

  • Provenance and peer review Not commissioned; externally peer reviewed.