Harnessing Technology and Citizen Science to Support Neighborhoods that Promote Active Living in Mexico

Abstract

Middle- and low-income countries bear 80 % of the global chronic disease burden. Population-level, multi-sectoral approaches to promoting healthful lifestyles that take into local physical, socioeconomic, and sociocultural characteristics of both the environment and the population are needed. The “Nuestra Voz (Our Voice)” is one such approach that involves neighborhood residents acting as “citizen scientists” to systematically gather information on the barriers and facilitators of physical activity in their neighborhoods and then use their data to collectively advocate for local environmental- and policy-level changes to support active living. We pilot tested this approach in Cuernavaca, Mexico with adults and adolescents. This community-engaged and participatory approach is driven by residents, who utilize a GPS-enabled electronic tablet-based application with simple audio-based instructions to take photographs and record audio narratives of facets of their neighborhood that promote or hinder active living. After collecting these data, the citizen scientists come together in a community meeting and use their data to prioritize realistic, multi-level changes for promoting active living in their neighborhoods. A survey assessed participants’ acceptability of the approach. Participating citizen scientists included 32 adults and 9 adolescents. The citizen scientists rated the acceptability of five of the nine acceptability survey items with an average of 4.0 or higher out of 5.0, indicating they thought it was “fun,” were comfortable carrying the tablet, were likely to use it again, and would recommend it to friends and family. Items with average scores of less than 4 were all related to safety concerns. The most common barriers reported by citizen scientists using the tablet were poor sidewalk quality, presence of trash, negative characteristics of the streets, unpleasant aesthetics (e.g., graffiti), and presence of parks and recreational facilities. The Our Voice citizen scientist approach using the Discovery Tool has high potential for assisting communities in diverse settings to begin to identify both local barriers to active living as well as potentially useful strategies for promoting physical activity in culturally congruent ways that are appropriate and feasible in the local context.

Background

Chronic diseases such as diabetes and cardiovascular disease cause more deaths worldwide than all other causes combined.¹ Increasingly, attention is being given to middle- and low-income countries that bear 80 % of the global chronic disease burden.¹ In a middle-income country such as Mexico, age-adjusted deaths due to chronic disease are estimated at 542.6 per 100,000 men and 411.7 per 100,000 women, which is higher than in the USA (458.2 per 100,000 men and 325.7 per 100,000 women).¹ Additionally, the prevalence of obesity, a major risk factor for multiple chronic diseases, has increased in middle-income countries such as Mexico to levels comparable to the USA.^{1 – 4} For example, 38 % of women in Mexico and the USA are classified as obese according to the most recent national surveys in each country.^{3 , 4}

Fortunately, chronic diseases are largely preventable through reduction of behavioral risk factors such as physical inactivity and unhealthful diets.⁵ Regular physical activity is beneficial for weight management and reduces the risk of a range of chronic diseases, particularly high blood pressure, diabetes, depression, and some cancers.⁶ The prevalence of physical inactivity in middle-income countries is also approaching levels of high-income countries.¹ An estimated 38 % of Mexicans are physically inactive, similar to the 41 % classified as inactive in the USA.⁷ Innovative, scalable, and effective models for primary and secondary prevention of chronic disease that promote physical activity are needed to address the growing burden of chronic disease in middle-income countries such as Mexico.⁸

Population-level, multi-sectoral approaches to promoting healthful lifestyles can be successful and potentially cost-effective.^{1 , 8} Modifying environments in which individuals live, work, and play to support and encourage healthy lifestyle choices on a population level is one potential approach.^{6 , 9 , 10} Support for this approach is drawn largely from observational studies linking physical activity with specific features of the built environment, including residential density, street connectivity, land use mix, pedestrian infrastructure, aesthetics, and access to recreational facilities and public transit.^{9 , 11 – 19} Given this increasing evidence, there is an urgent need for innovative, effective, and scalable approaches that can overcome the challenges of making such infrastructural modifications to the built environment to support active lifestyles at the population level.

“Bottom-up” approaches that harness the expertise and influence of local neighborhood residents may hold particular promise for supporting infrastructural changes that are relevant for a given neighborhood, feasible given existing resources and constraints, and culturally appropriate for specific populations. One such approach involves neighborhood residents acting as “citizen scientists” to systematically gather information on the barriers and facilitators of physical activity in their neighborhoods and then use their data to collectively advocate for local environmental- and policy-level changes to support active living. This approach, referred to as “Our Voice,” has been shown to be acceptable and feasible in several locales in the USA,^{20 – 24} including in a low-income largely Latino neighborhood in northern California.²⁰ The benefit of “Our Voice” is the ability to identify strategies that are successful for promoting physical activity in specific environments that take into account existing resources and constraints, and local physical, socioeconomic, and sociocultural characteristics of both the environment and the population.

A citizen-engaged approach such as “Our Voice” may hold particular promise for translation to low- to middle-income countries given that strategies are developed by residents themselves and therefore are relevant to the local context. However, the acceptability and feasibility of implementing and disseminating a citizen scientist approach for promoting active living in low- to middle-income countries like Mexico has not yet been explored. The Nuestra Voz (Our Voice) Mexico study builds on previous pilot studies using a citizen scientist engagement approach and a simple e-tablet application (the Healthy Neighborhood Discovery Tool) to crowd source information on barriers to and facilitators of physical activity that are potentially modifiable and realistic given local resources, priorities, and values.^{20 – 24}

Methods

The purpose of the Nuestra Voz-Mexico pilot study was to examine the acceptability and feasibility of using a citizen scientist approach with the Discovery Tool in Mexico in neighborhoods with varying levels of objectively defined socioeconomic status and walkability.

The study was approved by the Institutional Review Boards of Stanford University and the National Institute of Public Health of Mexico. All adult participants signed informed consent forms, and adolescent participants verbally assented to participate in the study while their parents or guardians provided written informed consent.

Study Site

The Nuestra Voz-Mexico pilot study was conducted in Cuernavaca, Mexico. Cuernavaca was an ideal setting for pilot testing based on demographics and existing research infrastructure. Cuernavaca, an urban setting with a population of 365,168 inhabitants located 76 km south of Mexico City, has a similar prevalence of overweight and obesity compared to the whole country,³ and can be considered a standard Mexican city in size and structure.^{25 , 26} Additionally, Cuernavaca is home to the headquarters of the Mexican National Public Health Institute (Instituto Nacional de Salud Pública, INSP). Moreover, Cuernavaca served as the Mexican site for the International Physical Activity Environment Network (IPEN) Adult Study. Briefly, IPEN was a multinational study involving 12 countries, with the objective of accurately assessing the strengths of association of built environment features with physical activity levels among adults, using a robust pooled sample from around the world.²⁷ We took advantage of existing ties to local neighborhoods, the recruitment experience, and the availability of neighborhood-level walkability and socioeconomic status data from IPEN-Mexico to conduct the Nuestra Voz-Mexico pilot study in Cuernavaca.

Sampling Strategy, Eligibility Criteria, and Recruitment

We selected four neighborhoods out of the 32 IPEN-Mexico census tracts for the Nuestra Voz-Mexico pilot study with the purpose of examining the acceptability and feasibility of the resident engagement process in diverse neighborhood contexts in Mexico. For IPEN-Mexico, Cuernavaca census tracts were stratified according to socioeconomic status (SES) (median split: high/low) and walkability (median split: high/low) and eight census tracts were randomly selected per SES/walkability stratum.^{25 , 26} The walkability index was calculated using z-scores of intersection density (number of 4-way intersections over total area per census tract), land use mix (diversity of land use types per census tract, using a normalized entropy score ranging from 0 to 1),²⁸ proportion of commercial land use (over total census tract area), and net residential density (total residences over area destined for residential use per census tract).²⁷ We selected one neighborhood from each stratum (low SES/high walkability; low SES/low walkability; high SES/high walkability; high SES/low walkability) to examine the acceptability and feasibility of the Nuestra Voz approach to measuring and potentially modifying local neighborhood environments in distinct types of Mexican neighborhoods. The selection of each neighborhood per stratum was based on safety considerations for data collectors (known high-crime areas were avoided), and neighborhoods with higher proportions of households with mid-life to older adults (based on census data) were favored for inclusion. The target sample size was 30 adults aged 45 years and older (8 to 11 per census tract) to achieve saturation of the neighborhood information provided, as seen in US-based studies.^{20 – 24} In these latter studies, it was found that “saturation” in relation to basic neighborhood environment issues being identified could be achieved with approximately 8–10 residents from a neighborhood. We also aimed to include a smaller total sample of 10 adolescents ages 11 to 14 years (4 to 5 adolescents per census tract) to complement the information obtained from the adults.

Eligible participants were permanent residents of the selected neighborhoods and were able to move around their neighborhood with or without an assistive device. Recruitment occurred using a systematic community-based door-to-door approach, proven to be successful for IPEN-Mexico and other studies in Mexico.^{3 , 26}

Procedures

The Our Voice citizen science approach has been described previously.^{20 – 24 , 29} Briefly, this community-engaged and participatory approach is driven by residents, who utilize a GPS-enabled electronic tablet-based application (the Discovery Tool) with simple audio-based instructions to take photographs and record audio narratives of facets of their neighborhood that promote or hinder active living. The tool also uses GPS to map the walking routes taken by the citizen scientists and geocode their pictures and audio recordings. After collecting these data, the citizen scientists come together in a community meeting and use their data to prioritize realistic, multi-level changes for promoting active living in their neighborhoods. In addition, they learn how to advocate for these changes with local policy makers or use other types of community engagement and social mobilization strategies (e.g., working together with fellow residents) that they identify.

To facilitate pilot testing in Mexico, the Spanish language version of the Discovery Tool that was developed for Latino residents in the USA was adapted to reflect linguistic differences in Spanish spoken in Mexico. The audio instructions included in the app were re-recorded to reflect these linguistic adaptations. Nuestra Voz-Mexico participants (citizen scientists) were asked to take a short walk (approximately 15–30 min) around their neighborhoods, and to use the Discovery Tool to identify barriers to or facilitators of physical activity. The application is designed to be easy to use for individuals with limited literacy and technology capabilities, and typically requires only brief training (about 5 min) to be able to use it effectively. For this pilot study, a research assistant accompanied the participants on their walk.

Although full deployment of the community advocacy training portion of the citizen science community engagement model was beyond the scope of this pilot study,³⁰ participants from one neighborhood were invited to a community meeting to identify issues relating to addressing barriers to active living in their neighborhoods and to discuss how relevant changes could be made in Mexico. The research assistant invited all participants from the low SES/low walkability neighborhood to participate in the community meeting, which took place at the main offices of the National Public Health Institute in Cuernavaca. This neighborhood was selected because of the residents’ enthusiasm for the project, as evidenced by a particularly high response rate across the four neighborhoods (i.e., 58 % relative to 44 to 14 % across the other three neighborhoods). As has been done in prior implementations of the Our Voice approach,²⁰ the research assistant provided a 1-h training to residents in community advocacy related to the built environment and physical activity using previously developed Spanish language materials. The training focused on balancing feasibility of needed changes with the potential impact that the changes could have on neighborhood physical activity levels. The research assistant then presented the group with the results pertaining to the topics they had identified through their photos and audio files, and facilitated a discussion of the findings. The meeting attendees then voted on their top priorities for change and discussed strategies for advocating for these changes.

Measures

Participants responded to an interviewer-administered survey that assessed basic sociodemographic information, initial self-efficacy for using the Discovery Tool, perceptions of the neighborhood built environment (adapted version of the abbreviated Neighborhood Environment Walkability Scale (NEWS-A)^{31 , 32}), and the acceptability of the citizen scientist approach using the Discovery Tool. The instrument to assess self-efficacy for using the Discovery Tool was adapted from an existing measure (computer self-efficacy measure³³) and used a 10-point scale (1 = not at all sure; 10 = very sure) to assess participants’ confidence in using the Discovery Tool with varying levels of assistance. The NEWS-A is a shorter adaptation of the original NEWS, has good test-retest reliability, and has been previously used in Mexico.^{18 , 31 , 32} We assessed the following NEWS domains: (1) land use mix-diversity (23 items, using a 5-point scale with an additional option for “don’t know”): proximity of destinations from home ranging from less than a 5-min walk to more than a 30-min walk from home; (2) land-use mix access (3 items, using a 4-point scale): access to services in the neighborhood; (3) street connectivity (2 items, using a 4-point scale): number of cul-da-sacs and intersections; (4) Walking/cycling (6 items, using a 4-point scale): presence of sidewalks, pedestrian and cycle trails and physical barriers between walk/cycle paths and roads; (5) aesthetics/neighborhood surroundings (4 items, using a 4-point scale): presence of trees, attractive sights, and trash; (6) safety from traffic (3 items, using a 4-point scale): volume and speed of traffic, availability of pedestrian crossings; and (7) safety from crime (3 items using a 4-point scale): crime rates.

We also used a nine-item acceptability survey that uses a 6-level Likert agreement scale (1 = not at all, 2 = not very, 3 = a little, 4 = more or less, 5 = somewhat, 6 = very). The acceptability survey included items to determine interest, enjoyment, ease of use, comfort, safety perceptions, likelihood to use again or recommend to others, and overall participants’ acceptability or concerns with regards to using the Discovery Tool for assessing their neighborhood. The acceptability survey was used in a prior pilot test of the citizen scientist approach and allows for comparability of results.²⁰ The survey included a final open-ended question to solicit feedback on participation in the study.

Analysis of Survey Data

Descriptive statistics (means, proportions) for sociodemographic information, cellular phone ownership (used as a marker of familiarity/access to technology), and self-rated self-efficacy for using the Discovery Tool were calculated for adults and adolescents. We used a Wilcoxon rank sum test to compare adults’ and adolescents’ self-efficacy for using the Discovery Tool. To describe the citizen scientists’ perceptions of the neighborhoods in which the study was conducted, we calculated scores for key built environment constructs (residential density, connectivity, land use mix, availability and accessibility to facilities, safety from crime, safety from traffic, and aesthetics) from the NEWS-A survey using standardized scoring procedures.³² Higher scores reflect more favorable environments. The non-parametric Kruskal-Wallis test was used to test for significant differences in perceptions across the four different neighborhoods.

To assess acceptability and feasibility of the citizen scientist approach using the Discovery Tool, we tracked the number of photo and audio recordings for each participant and summarized the information derived from the 9-item acceptability survey. The responses for questions inquiring about a negative concept (e.g., how nervous did you feel wearing or carrying the device?) were reverse coded so that a higher response always reflected higher acceptability. Acceptability was defined, per item, as having a score of 4 or above (out of 5). In addition, we reviewed responses to open-ended questions completed at the conclusion of the survey to identify common themes. The Kruskal-Wallis test was used to compare acceptability and feasibility across the four different neighborhoods.

Analysis of Photos and Audio Recordings

To further assess the acceptability and feasibility of this approach, we examined the identified barriers, facilitators, and potential solutions to issues identified by the citizen scientists to identify common themes. A previously developed coding schema was used and augmented by additional categories of relevance for this particular setting, for example, dog-related issues, graffiti, vacant lots, and pedestrian bridges.^{20 , 24} Additionally, photos and audio recordings were coded as a barrier or facilitator according to the participants’ descriptions. The coding of all the photographs and audio narratives was conducted by a trained research assistant. We also tracked the number of attendees at the community meeting and the proposed changes to local built and social environments that they prioritized.

Results

We recruited 32 adults and 9 adolescents to participate as citizen scientists in the pilot study. The response rate from the door-to-door recruitment procedure was 38 % for adults and 43 % for adolescents. The response rate was highest in the low SES/low walkability neighborhood (58 %) and lowest in the low SES/high walkability neighborhood (14 %). Both high SES neighborhoods had similar response rates, with 44 % for the low walkability neighborhood and 42 % for the high walkability neighborhood.

Participants

Participating adult citizen scientists were primarily women (75 %) and had a mean age of 57 years (SD 8.7) (Table 1). The sample represented adults from various education levels, including those with less than an elementary school education (19 %) and those with some college education (25 %). Eighteen (56 %) of the 32 adults owned a cell phone, two of which owned a Smartphone. Youth citizen scientists included five girls and four boys with a mean age of 13 years (SD 1.6). Seven of the nine youth reported owning a cell phone, two of which owned a smartphone. The citizen scientists’ initial overall self-efficacy for using the Discovery Tool across varying levels of assistance was higher among adolescents compared to adults (9.3 vs. 7.7, p < 0.01). Adolescents were “sure” or “very sure” that they could use the Discovery Tool with no help or instructions from others (mean score of 9.4 [SD 0.5] on a scale of from one [not at all sure] to 10 [very sure]). Adults reported an average rating of 7.1 (SD 2.9) for using the Discovery Tool with no help or instructions, 7.7 (SD 2.5) if the device was similar to one used before, and 7.5 (SD 2.7) if they could see someone use the tablet first.

TABLE 1.

Characteristics of adult and adolescent participants in the Nuestra Voz-Mexico study, May–June, 2014. (n = 41)

Characteristic	Total		Adolescents		Adults
	(n = 41)		(n = 9)		(n = 32)
Mean age in years, n (SD)			13.3	(1.6)	57.3	(8.7)
Gender, n (%)
Female	28	(68 %)	4	(44 %)	24	(75 %)
Male	13	(32 %)	5	(56 %)	8	(25 %)
Educational attainment, n (%)
Elementary or less	9	(22 %)	3	(33 %)	6	(19 %)
Middle school	13	(32 %)	4	(44 %)	9	(28 %)
Some high school	2	(5 %)	1	(11 %)	1	(3 %)
High school or the equivalent	1	(2 %)	0	(0 %)	1	(3 %)
Some college	9	(22 %)	1	(11 %)	8	(25 %)
Finalized college	6	(15 %)	0	(0 %)	6	(19 %)
Post-graduate studies	1	(2 %)	0	(0 %)	1	(3 %)
Phone ownership, n (%)
Cell phone ownership	23	(56 %)	5	(56 %)	18	(56 %)
Smartphone ownership	4	(18 %)	2	(40 %)	2	(12 %)
Initial self-efficacy for using Discovery Toola, mean (SD)
Overallb	8.1		9.3		7.7
Use with no help or instructions	7.7	(2.8)	9.4	(0.5)	7.1	(2.9)
Use if the device was similar to one you had used before	8.1	(2.3)	9.4	(0.7)	7.7	(2.5)
Use if I could see someone use the tablet before me	8.0	(2.5)	9.4	(0.8)	7.5	(2.7)
Use if there were written instructions on how to use it	8.3	(1.8)	9.3	(0.9)	7.9	(1.9)
Use if somebody taught me how to use it before using it	8.2	(1.9)	8.9	(1.6)	8.0	(1.9)
Use if somebody could help me while I’m using it	8.0	(1.6)	9.2	(0.9)	7.7	(1.6)

^aParticipants rated their confidence using the tablet to take photos and record their thoughts when they walked around their neighborhood in each scenario on a scale from 1 to 10 (1 = not at all sure, 10 = very sure)

^bMean combining the responses to each scenario, P < 0.001

Neighborhood Environment Perceptions

Citizen scientist perceptions of neighborhood walkability using the NEWS-A varied by neighborhood across built environment domains, with statistically significant differences by neighborhood in the domains of land use mix-diversity (p = 0.05) and crime safety (p = 0.02) (Table 2). Participants in the high SES/high walkability rated the land use mix-diversity as highest (p value = 0.05), while participants from the high SES/low walkability neighborhood rated crime safety as highest (p value = 0.02). Overall across all neighborhoods, the highest scores were observed in the domains of land use mix-access, street connectivity, and walking/cycling facilities. The lowest scores were observed in the domains of safety from traffic and crime.

TABLE 2.

Mean scores on the neighborhood environment walkability scale by neighborhood (n = 41), Cuernavaca, Mexico May–June, 2014

	Total	Neighborhood 1 Low	Neighborhood 2	Neighborhood 3	Neighborhood 4
Socioeconomic status (low or high)		Low	Low	High	High
Walkability (low or high)		High	Low	High	Low	p value
	n = 41	n = 4	n = 18	n = 10	n = 9
	Mean (SD)	Mean (SD)	Mean (SD)	Mean (SD)	Mean (SD)
Land use mix-diversity	3.6 (0.29)	3.3 (0.11)	3.6 (0.33)	3.8 (0.27)	3.7 (0.17)	0.05
Land use mix-access	3.0 (0.47)	3.0 (0.50)	2.9 (0.45)	3.1 (0.52)	3.1 (0.49)	0.55
Street connectivity	2.7 (0.46)	3.1 (0.17)	2.6 (0.45)	2.7 (0.40)	2.6 (0.60)	0.23
Walking/cycling facilities	2.7 (0.63)	2.8 (0.88)	2.5 (0.53)	2.6 (0.51)	3.3 (0.62)	0.07
Aesthetics	2.2 (1.1)	2.4 (1.3)	2.2 (1.2)	2.4 (1.2)	2.1 (0.99)	0.85
Pedestrian/traffic safety	2.2 (0.57)	2.2 (0.88)	2.4 (0.54)	2.1 (0.55)	2.1 (0.49)	0.19
Crime safety	1.7 (0.94)	1.3 (0.50)	1.3 (0.67)	2.1 (1.1)	2.3 (1.0)	0.02

Acceptability of the Citizen Scientist Approach Using the Discovery Tool

The average duration of the Discovery Tool walks was 14.4 min (SD 15.0). During the walks around the neighborhood, adult and adolescent citizen scientists took a total of 145 photos of neighborhood characteristics that they believed promoted or hindered physical activity, or an average of 3.6 (SD 4.7) photos per adult and 3.4 (SD 2.7) photos per adolescent (Table 3). The number of photos per participant ranged from 0 photos to 26 photos. The citizen scientists recorded a total of 89 audio narratives or an average of 1.9 (SD 2.1) per adult and 3.0 (SD 2.5) per adolescent. The number of audio narratives per participant ranged from 0 to 10. Adult citizen scientists in the high SES/low walkability neighborhood took the highest number of photos (mean = 7.1, SD 8.1) and recorded the highest number of audio narratives (mean = 2.5, SD 3.3), although the differences were not statistically significant across neighborhoods (range across the different neighborhoods for photos = 7.1, 1.3; for audio narratives = 2.5, 1.3).

TABLE 3.

Data collected using the discovery tool according to neighborhood type, Cuernavaca, Mexico, May–June, 2014. (n = 41)

Sorted by Adults n = 32	Total	Neighborhood 1	Neighborhood 2	Neighborhood 3	Neighborhood 4
Socioeconomic status (low or high)		Low	Low	High	High	p value
Walkability (low or high)		High	Low	High	Low
	n = 32	n = 4	n = 12	n = 8	n = 8
Photos
Total	116	5	31	23	57	0.13
Mean (SD)	3.6 (4.7)	1.3 (1.5)	2.6 (1.2)	2.9 (3.1)	7.1 (8.1)
Range (min, max)	(0, 26)	0, 3	(1, 5)	(0, 10)	(1, 26)
Audio
Total	60	5	23	12	20	0.97
Mean (SD)	1.9 (2.1)	1.3 (0.5)	2.2 (2.3)	1.5 (1.2)	2.5 (3.3)
Range (min, max)	(0, 10)	(1, 2)	(0, 8)	(0, 4)	(0, 10)
Sorted by adolescents n = 9	Total	Neighborhood 1	Neighborhood 2	Neighborhood 3	Neighborhood 4
Socioeconomic status (low or high)		Low	Low	High	High	p value
Walkability (low or high)		High	Low	High	Low
	n = 9	n = 0	n = 6	n = 2	n = 1
Photos
Total	29	0	22	1	6	0.14
Mean (SD)	3.4 (2.7)	0	3.7 (3.1)	0.5 (0.7)	6
Range (min, max)	(0, 9)	0	(1, 9)	(0, 1)	6
Audio						0.23
Total	29	0	18	3	8
Mean (SD)	3 (2.5)	0	3 (2.4)	1.5 (2.08)	8
Range (min, max)	(0, 8)	0	(0, 6)	(1, 2)	8

The citizen scientists rated the acceptability of five of the nine acceptability survey items with an average of 4.0 or higher out of 5.0; they indicated that they thought it was “fun” to use the Discovery Tool, they were comfortable carrying the tablet, were likely to use it again, and would recommend it to friends and family (Table 4). The highest average acceptability scores on six of the nine items were reported in the neighborhood characterized by low SES and low walkability, although only the survey items related to comfort using the tablet and likelihood of using the tablet again reached statistical significance in neighborhood comparisons (p < 0.05). Items with average scores of less than 4 were all related to safety, including comfort level carrying the tablet, nervousness bringing the tablet on the walk, and level of danger in carrying the tablet. Acceptability scores related to safety were all higher in the high SES neighborhoods compared to the low SES neighborhoods. Citizen scientists from the high SES/low walkability neighborhood rated their level of nervousness carrying the tablet higher than all other neighborhoods (p = 0.05). In addition to acceptability, the majority of participants (64 %) reported at the end of the study that they were more conscious of the characteristics of their neighborhood and their potential influence on their own activity levels as a result of participating in the Nuestra Voz-Mexico pilot study in open-ended feedback at the end of the survey.

TABLE 4.

Discovery tool feedback (n = 41) Cuernavaca, Mexico May–June, 2014

	Total		Neighborhood 1		Neighborhood 2		Neighborhood 3		Neighborhood 4
Socioeconomic status (low or high)			Low		Low		High		High
Walkability (low or high)			High		Low		High		Low
	n = 41		n = 4		n = 18		n = 10		n = 9
Survey questiona	mean	(SD)	mean	(SD)	mean	(SD)	mean	(SD)	mean	(SD)	p value
How much fun was it to use the tool (the electronic tablet) today?	4.2	(1.0)	3.8	(0.96)	4.6	(0.98)	4.0	(0.67)	3.8	(1.3)	0.13
How comfortable did you feel carrying the electronic tablet?	4.0	(1.3)	3.3	(0.50)	4.5	(1.2)	3.2	(1.5)	4.2	(1.2)	0.04
How nervous were you bringing the electronic tablet with you?	3.2	(1.5)	1.8	(0.96)	3.0	(1.5)	3.3	(1.6)	4.2	(1.1)	0.05
How much attention do you feel you attracted by carrying the electronic tablet with you?	2.8	(1.8)	2.5	(1.9)	2.1	(1.5)	3.1	(1.7)	4.1	(1.8)	0.07
How dangerous did it feel carrying the electronic tablet with you?	2.9	(1.9)	2.8	(1.7)	2.4	(1.7)	3.2	(2.1)	4.0	(1.8)	0.25
How much fun was it using the electronic tablet to document aspects of your neighborhood?	4.2	(1.0)	3.8	(1.3)	4.4	(0.92)	4.0	(1.1)	4.1	(1.1)	0.54
How likely are you to recommend this device to your friends and family?	4.7	(1.2)	3.7	(1.2)	5.1	(0.54)	4.3	(1.5)	4.6	(1.6)	0.27
If you could, how likely are you to use this device again?	4.2	(1.5)	3.8	(0.96)	5.0	(0.91)	3.7	(1.4)	3.6	(1.9)	0.04
If you could, how likely are you to use this device for a longer period of time?	3.8	(1.4)	4.0	(1.4)	4.5	(0.92)	3.5	(1.6)	2.8	(1.4)	0.03

^aScale of 1 to 5, 1 = lowest acceptability score, 5 = highest acceptability score

Barriers to and Facilitators of Physical Activity

The most common barriers reported by citizen scientists across the different types of neighborhoods using the Discovery Tool were poor sidewalk quality, presence of trash, negative characteristics of the streets, and unpleasant aesthetics (e.g., graffiti), and presence of parks and recreational facilities (Table 5). Citizen scientists from both high SES neighborhoods (with high and low walkability) and the low SES/high walkability neighborhood reported the same most common barriers of poor sidewalk quality, negative street characteristics, and presence of trash (see Fig. 1). Like the other neighborhoods, citizen scientists from the low SES/low walkability neighborhood commonly reported poor sidewalk quality as a barrier. Yet, in contrast to the other neighborhoods, participants from the low SES/low walkability neighborhood noted the presence of parks or recreational facilities and unpleasant aesthetics as common barriers. The presence of parks or recreational facilities was also cited as a facilitator of physical activity in the low SES/low walkability neighborhood. The Mexican citizen scientists generally noted few additional facilitators of physical activity across the different neighborhoods.

TABLE 5.

Barriers identified using the discovery tool, Cuernavaca, Mexico, May–June, 2014. (n = 41)

	Neighborhood 1			Neighborhood 2			Neighborhood 3			Neighborhood 4
Socioeconomic status (low or high)	Low			Low			High			High
Walkability (low or high)	High			Low			High			Low
Total photos/audio files	N = 10			N = 70			N = 47			N = 77
	Theme	n	(%)	Theme	n	(%)	Theme	n	(%)	Theme	n	(%)
Barrier 1	Poor sidewalk quality	6	(60 %)	Poor sidewalk quality	15	(21 %)	Street characteristics	13	(28 %)	Poor sidewalk quality	48	(62 %)
Barrier 2	Trash	3	(30 %)	Parks	19	(14 %)	Poor sidewalk quality	11	(23 %)	Street characteristics	18	(23 %)
Barrier 3	Street characteristics	1	(10 %)	Unpleasant aesthetics/beauty	5	(7 %)	Trash	5	(11 %)	Trash	7	(9 %)
Facilitator 1				Parks	10	(14 %)	Good sidewalk quality	3	(6 %)
Facilitator 2				Good sidewalk quality	2	(3 %)
Facilitator 3				Pleasant aesthetics/beauty	2	(3 %)

FIG. 1.

Example photos from the Discovery Tool; Nuestra Voz, Cuernavaca, Mexico, 2013.

Community Meeting

A total of 18 residents from the low SES/low walkability neighborhood participated in the community meeting, including 9 adult citizen scientists, 3 adolescent citizen scientists, and 6 family members of the adolescents. The attendees discussed 10 themes that were identified in their photos and audio narratives, including the following: sidewalk quality, street characteristics, parks, trash, aesthetics, safety, dogs, destinations, mobility for handicapped, and pedestrian crossings. After the discussion, attendees identified their top three priorities through voting. The top priorities voted by attendees were safety, dogs, and unpleasant street characteristics, such as graffiti. The discussion on how to advocate for changes to impact these issues focused primarily on what the neighborhood residents could organize to do themselves, such as creation of a neighborhood committee to help neighbors leash and clean up after their dogs, neighborhood watch programs to combat crime, and strategies to promote increased social cohesion in the neighborhood. The neighborhood residents did not identify communication with local policy makers or key stakeholders on their own as promising strategies to impact change.

Discussion

Results from the Nuestra Voz-Mexico pilot study supported the acceptability and feasibility of a citizen scientist approach using the Healthy Neighborhood Discovery Tool in Mexico, a middle-income Latin American country. This approach previously has been successfully pilot tested in diverse locations in the USA, including in low-income Latino neighborhoods.^{20 – 24} However, it has never been pilot tested in a low- or middle-income country such as Mexico. Low- and middle-income countries are high priorities for physical activity promotion because of increasing levels of physical inactivity, obesity, and the high burden of chronic diseases.^{7 , 34 – 36} However, a “one-size-fits-all” approach is unlikely to be successful in the diverse settings that they represent. The Our Voice citizen scientist approach using the Discovery Tool has high potential for assisting communities in diverse settings to begin to identify both local barriers to active living as well as potentially useful strategies for promoting physical activity in culturally congruent ways that are appropriate and feasible in the local context.

We examined the acceptability and feasibility of this citizen scientist approach in four types of neighborhoods in Mexico defined by neighborhood SES and walkability (low SES/high walkability; low SES/low walkability; high SES/high walkability; high SES/low walkability). Engagement and acceptability of this approach was highest in the neighborhood characterized by low SES and low walkability. This neighborhood had the highest response rate (58 versus 40 % overall) and the highest scores on the acceptability survey. Additionally, the average number of photos taken and audio recordings per adult were second highest in this neighborhood, just behind the high SES/low walkability neighborhood (the averages for adolescents are difficult to compare due to small sample size). High levels of engagement and acceptability in this neighborhood could be due to several factors. First, residents of this neighborhood rated their perceptions of crime as lowest of all the neighborhoods on the Neighborhood Environment Walkability Scale, which could have facilitated their willingness to participate in a study that involved walking around the neighborhood with a tablet computer. Second, the residents of this neighborhood may have been motivated to change their neighborhood due to low levels of objectively measured walkability as well as the lowest scores of all neighborhoods on two NEWS domains focused on access to services in the neighborhood and availability of features that promote walking and biking. A third factor contributing to high engagement and acceptability could have been the attraction and novelty of using a tablet-based application. The citizen scientists from this neighborhood reported the highest scores on the acceptability survey for items related to how “fun” it was to use the device.

In contrast, the neighborhood characterized by low SES and high walkability had the lowest levels of engagement and acceptability. The response rate was 14 versus 40 % overall. Compared to the overall sample, the adults in this neighborhood averaged fewer average photos taken by adults (1.3 vs. 3.6 photos per adult) and fewer audio recordings (1.3 vs. 1.9 audio recordings per adult). It appeared that the low levels of engagement and acceptability in this neighborhood were due to factors related to safety based on findings from the acceptability survey and the NEWS. The citizen scientists from this neighborhood reported the lowest acceptability scores on questions related to safety, including comfort carrying the tablet, nervousness bringing the tablet on the walk, and how dangerous it felt carrying the tablet. Additionally, the residents in this neighborhood reported the highest scores for crime on the NEWS. The safety issues in this neighborhood may be related to specific characteristics of high walkability neighborhoods in Mexico. Characteristics of neighborhoods classified as highly walkable include high intersection density, highly connected streets, and high land use mix. It is possible that these characteristics are associated with high levels of crime in Mexico, thus presenting barriers to using the tablet-based application. Potentially, the citizen scientist approach could be used with mobile phones, which are smaller and more discrete than tablets, to reduce concerns about safety.

Results from the IPEN-Mexico study provide important context in which to interpret the findings from this pilot study and provide a rationale for a citizen scientist approach for promoting physical activity. The IPEN-Mexico study revealed an unexpected inverse association between the walkability metrics that were applied, that have been developed and used by urban planning and transportation experts in a number of other countries, and objectively measured physical activity among adults in Cuernavaca, including in the neighborhoods that were included in this study.³⁷ This contrasts with findings from high-income countries (e.g., USA, Australia) where walkability and objectively measured physical activity have been positively correlated across age groups.^{11 , 38} Access to public transit provides a specific illustration of this contrast in high-income and middle-income countries. In the USA and Canada, access to public transit has been shown to be positively associated with physical activity.^{39 – 41} This is likely due to the fact that people who live close to public transit will walk more to access that transit and people who do not live close to public transit are more likely to use their cars. In contrast, findings from IPEN-Mexico document a negative association between proximity to public transit and physical activity.³⁷ This may be due to the proportion of the population that owns personal vehicles. Compared to the USA, fewer Mexicans own personal cars (797 cars per 1000 people in the USA and 275 cars per 1000 people in Mexico⁴²), and as a result do not typically have choices between driving or using public transit. Thus, the closer an individual lives to public transit the less walking he or she may have to do more generally, given that car use is often not an option. The negative association between proximity to public transit and physical activity may also be due to the structure of the transit system in Mexico. In Mexico, buses stop anywhere along the route if signaled by a passenger, thereby enabling the passengers to board and disembark closer to their points of origin and desired destinations and thus minimize walking.³⁷ Residents who live in high walkability neighborhoods in Mexico are likely to live close to the bus routes and thereby minimize the distance they walk to get on the bus, whereas residents of low walkability neighborhoods are likely to live farther from the bus route, requiring more walking. This may partially explain why the residents in the low walkability neighborhood were highly accepting of the citizen scientist approach because they walk more in their neighborhood and thus may be motivated to improve it. The results from IPEN-Mexico provide an excellent example of the impact of local environmental nuances on physical activity and underscore the importance of “bottom-up” approaches driven by local residents.

Findings from IPEN-Mexico on individual-level determinants of physical activity may also be informative for identifying potential strategies to increase engagement in this citizen scientist approach. The IPEN-Mexico study found that certain individual-level sociodemographic characteristics such as being male and not owning a car were associated with higher levels of physical activity.²⁵ In addition, Mexican adults were more active during the week as opposed to the weekend, potentially signifying that transport and occupational physical activity may be larger contributors to physical activity than leisure-time physical activity.²⁵ Future studies could aim to engage citizen scientists from the most “at risk for inactivity” population subgroups (women, car owners) to identify their perceptions, priorities, and potential solutions for improving their neighborhoods in ways that would make them more likely to use their neighborhoods to be physically active. Encouragingly, the majority of the adult residents who volunteered to learn how to become citizen sciences in the Nuestra Voz pilot were in fact women.

We have also pilot tested the Our Voice citizen scientist approach using the Discovery Tool with 10 adolescent and 10 adult participants from Fair Oaks, California, a low-income Latino immigrant neighborhood.²⁰ This pilot study, similar to the Nuestra Voz pilot, was successful in bringing older adults and adolescents together to identify characteristics of their neighborhood that hinder and promote physical activity. Older adults and adolescents are important demographic groups for physical activity promotion as they have the lowest levels of activity within their respective age groups (adults and children).⁴³ Additionally, the citizen scientists in the Nuestra Voz-Fair Oaks and Nuestra Voz-Mexico projects revealed similar themes concerning the built environment and its association with physical activity. The most commonly identified themes in Nuestra Voz-Fair Oaks were sidewalk characteristics and trash. These were also commonly identified themes in Nuestra Voz-Mexico. However, the severity of the negative aspects was noticeably higher in Mexico than in the USA (e.g., a smaller crack on the sidewalk may be perceived as “low sidewalk quality” in Fair Oaks, while a deep, dangerous ditch in a sidewalk was provided as an example of “low sidewalk quality” in Cuernavaca). Differences in the two pilot projects included the number of photos and audio recordings, the duration of the walks, and the focus of the community meetings. Adult citizen scientists in Fair Oaks took an average of 11.3 photos with accompanying audio files compared to 3.6 photos per adult in Cuernavaca. Additionally, citizen scientists in Fair Oaks walked an average of 30 min, compared to 14 min in Cuernavaca. Finally, the community meetings had different outcomes in Fair Oaks and Cuernavaca (although, due to funding limitations for the Mexico pilot, community meetings were able to be conducted in only one neighborhood there). In Fair Oaks, the citizen scientists prepared, with guidance from local community facilitators and researchers, a presentation of their results to local policy makers and were successful in communicating their priorities and gaining experience in how to advocate for policy- and environmental-level change at the local level. In Cuernavaca, the citizen scientists focused to a greater extent on what they could do themselves to change the built environment for promotion of physical activity as opposed to influencing local policy makers. For example, to control dog waste, Mexican citizen scientists thought of tying plastic bags to lamp posts to make it convenient to clean up after their dogs. Also in Mexico, adults and adolescents engaged in discussions about how to provide opportunities for younger residents, in particular, to express themselves through appropriate types of neighborhood art because at least some adults viewed graffiti as a barrier to physical activity. Given the relatively similar demographic characteristics of the groups in Fair Oaks and Mexico (adolescents/adults; Mexican descent), the difference in interaction with policy makers may be a reflection of the local policymaking context as opposed to cultural differences in willingness to interact with policy makers. As such, the citizen scientist approach may be a vehicle for increased participation in the policymaking process in settings such as Mexico where this is traditionally not the case. This issue deserves further systematic study in Mexico and other low- and middle-income countries, where other “paths to policy” may merit additional exploration. As part of the Our Voice Mexico pilot study, as well as another pilot study being conducted in Bogotá, Colombia (King et al., TJACSM, 2016), initial observations have been made about the potential utility of harnessing the influence of researchers themselves in such regions as part of the communication pathway between local residents and decision-makers.

Nuestra Voz-Mexico was our first attempt to pilot test the citizen scientist approach in a middle-income setting. Among the limitations of this first-generation pilot study were the sample sizes, which, while small, were sufficient to allow some initial detection of differences across neighborhoods in several areas of interest, including perceived neighborhood characteristics as captured by the adapted NEWS-A, and participant perceptions and feedback in using the Discovery Tool. In addition, funding constraints limited the full application of the Our Voice citizen science approach, which typically can include additional community meetings and forums by citizens and relevant local organizations and stakeholders to more fully explore and problem-solve around barriers to active living and other relevant behaviors in the local community.²⁹ Importantly, numerous challenges also remain to fully realize the potential for scalability and widespread dissemination of this approach. First, as identified by the Mexican citizen scientists, the safety factor will be important to address in contexts where tablet computers may be less common. Second, while we conducted the pilot test in four types of neighborhoods in one urban setting, additional work is needed to understand how this approach may be adopted in rural settings and megalopolises like Mexico City. Third, strategies that promote scalability are needed to put the application in the hands of large numbers of potential citizen scientists to maximize the potential of this approach to shape the environments in which we live, work, and play to support and promote active lifestyles. Finally, longer-term studies are merited to better ascertain how the types of local built and social environmental changes facilitated by the Our Voice citizen science approach impact the physical activity behaviors and related outcomes that this approach is aiming to promote.

Conclusions

The Nuestra Voz pilot study provides information on the acceptability and feasibility of using a citizen scientist approach involving the Stanford Healthy Neighborhood Discovery Tool in a middle-income country setting. Catalyzing communities to develop local, impactful, and feasible strategies to promote active lifestyles across the life course is critical for curbing the worldwide chronic disease epidemic. A citizen scientist approach using the Discovery Tool provides a useful framework, process, and set of tools to support this goal. As shown by this pilot study, a key strength relating to potential scalability includes the ability of this approach to be adapted to local contexts so that resulting strategies can be better tailored to the local environment, culture, and political structure. Only through more fully leveraging the perspectives of residents themselves can relevant and truly translatable population health solutions in this and other health areas emerge.

IPEN, International Physical Activity Environment Network; NEWS-A, Neighborhood Environment Walkability Scale; SD, standard deviation; SES, socioeconomic status; USA, United States of America

Compliance with Ethical Standards

The study was approved by the Institutional Review Boards of Stanford University and the National Institute of Public Health of Mexico. All adult participants signed informed consent forms, and adolescent participants verbally assented to participate in the study while their parents or guardians provided written informed consent.

Funding

This study was funded by a grant from the Stanford Center for Innovation in Global Health (title: Nuestra Voz-Mexico). At the time of manuscript preparation DS was funded as a Postdoctoral Research Fellow by the Michael & Susan Dell Foundation. The funders did not play a role in the design of the study or collection, analysis, or interpretation of data.