Engaging citizen scientists to build healthy park environments in Colombia

Abstract

Colombia’s Recreovía program offers community-based free physical activity (PA) classes in parks. We evaluated built and social environmental factors influencing Recreovía local park environments, and facilitated a consensus-building and advocacy process among community members, policymakers and academic researchers aimed at improving uptake and impact of the Recreovía program. We used a mixed-methods approach, with individual and contextual PA measurements and a resident-enabled participatory approach (the Our Voice citizen science engagement model). Recreovía participants were likely to be women meeting PA recommendations, and highly satisfied with the Recreovía classes. Reported facilitators of the Recreovía included its role in enhancing social and individual well-being through PA classes. Reported barriers to usage were related to park maintenance, cleanliness and safety. The Our Voice process elicited community reflection, empowerment, advocacy and action. Our Voice facilitated the interplay among stakeholders and community members to optimize the Recreovía program as a facilitator of active living, and to make park environments more welcoming.

INTRODUCTION

It is vital to evaluate and scale-up interventions that have the potential to promote active and healthy living (Sallis et al., 2016). Physical inactivity increases the global prevalence of non-communicable diseases (NCDs) such as heart disease, stroke, diabetes, depression and eight types of cancers (2018 Physical Activity Guidelines Advisory Committee, 2018). Nearly a quarter of the world’s adult population is physically inactive, with even higher rates found among women (WHO, 2018). Latin America and the Caribbean have the highest levels of insufficient activity among women (>40%) (Guthold et al., 2018) highlighting a trend related to economic development (WHO, 2018). In fact, the greatest burden of NCDs occurs among women from low- and middle-income countries (Guthold et al., 2018).

Utilizing public spaces—particularly parks—has emerged as a potentially low-cost and widely accessible approach to promoting physical activity (PA) and healthy living (Arena et al., 2017; Díaz del Castillo et al., 2017). Latin America has numerous park-based health promotion programs (Hoehner et al., 2013; Díaz del Castillo et al., 2017). These feature regular free programming, including music-accompanied gymnastics, tropical rumba and similar classes in park and plaza public spaces (Cohen et al., 2017; Rios et al., 2017; Gaffney et al., 2019). Previous studies in Brazil and Colombia reported that women are the main users of park-based PA classes, and that people visiting parks with such programs are more likely to engage in moderate-to-vigorous physical activity (MVPA) compared to people visiting parks without formal PA programming (Torres et al., 2016; Sarmiento et al., 2017).

Beyond their potential use for PA promotion and programming, the neighborhood-level availability of public parks has been associated with higher leisure time PA (Zhang and Li, 2017), lower incidence of chronic diseases (Besenyi et al., 2014) and higher levels of mental health (Wood et al., 2017). In addition, benefits of activity-friendly environments can extend beyond individual-level impacts, increasing community-wide well-being by positively impacting neighborhood social cohesion and social capital (Torres et al., 2013; Díaz del Castillo et al., 2017; Atkins et al., 2018). Understood in this light, these types of community-based programs can be seen as opportunities for social integration, especially for marginalized populations (Plane and Klodawsky, 2013), and also contribute to decreasing gender disparities in PA (Torres et al., 2016).

Previous research has shown that deterrents to park use can include lack of safety and inadequate facilities (Rachele et al., 2016; Douglas et al., 2018). These effects are not distributed equally across socio-economic groups, as low-income populations generally have poorer access to safe, high-quality parks (,Rigolon, 2017; Han et al., 2018; Douglas et al., 2018). In addition, the determinants of PA in parks typically include the interactive effects of individual characteristics, social and physical environments and policies (Sallis et al., 2006; Sawyer et al., 2017).

The municipal government of Bogotá, Colombia offers free PA classes in urban public parks through a broader set of initiatives called ‘Recreovía’ (Díaz del Castillo et al., 2017). Designed in 1995 to promote PA at a community level, the Recreovía program has undergone rapid scale-up in Bogotá (Sarmiento et al., 2017; Abolghasem et al., 2018). By 2018, >1.2 million people had participated in 10 894 activities across both weekends and weekdays (IDRD, 2019). Recreovía goals include increasing park use and promoting healthy lifestyle habits (Rios et al., 2017). In Colombia’s current post-conflict era, the Recreovía program also seeks to build social capital and decrease inequalities in the availability of recreational spaces (Rios et al., 2017).

Despite broad uptake, there is notable variability in Recreovía participation (Torres et al., 2016). To better understand characteristics and interactions between built and social environmental factors that support communities’ regular PA, and ultimately inform effective PA policies, we applied a mixed-methods design through a sequential explanatory approach (Pluye and Hong, 2014). As the framework for our study we followed the Our Voice ‘by the people’ citizen science model. Our Voice uses accessible technologies to engage community members in data collection and analysis, offering a channel for providing feedback on existing programs and empowering residents to promote data-driven changes in the local built and social environments (Buman et al., 2013; King et al., 2016). While the model has been applied in a variety of contexts, previous to this study it had not been employed for systematic evaluation of parks and programs like Recreovía (Rosas et al., 2016; Sheats et al., 2017; Chrisinger et al., 2018; King et al., 2019 ).

In a context of growing interest in socio-ecological models to better target interventions aimed at health equity and behavior change (Blacksher and Lovasi, 2012), participatory strategies have become a basis for promoting a culture of active living, prompting citizens to serve as powerful agents of change (Sallis et al., 2006; Bauman et al., 2012; Arena et al., 2017; Koorts et al., 2018;). Our Voice applies the socio-ecological framework and draws from empowerment theory and social cognitive theory to engage residents as behavioral scientists in understanding and affecting the factors that influence their behaviors. Our Voice targets individuals, social environments, physical environments and policies to achieve behavioral changes for health promotion. It encourages people to gain confidence and skills to improve their quality of life through strengthening their self-efficacy to change relevant health-promoting behaviors and environments (Glanz and Rimer, 2005; Hinckson et al., 2017).

The specific aims of this study were to (i) evaluate built and social environmental factors influencing Recreovía local park environments; (ii) identify perceived factors that influence participants’ access to and engagement in Recreovía; and (iii) facilitate a consensus-building and advocacy process among community members, policymakers and academic researchers aimed at improving uptake and impact of the Recreovía program.

MATERIALS

Study setting

This study was conducted in Bogotá, Colombia, a city with 7.2 million inhabitants (Departamento Administrativo Nacional de Estadística, 2019). Currently, Bogotá has 5112 parks (IDRD, 2018). Recreovía primarily takes place in low- to middle-income neighborhoods (Díaz del Castillo et al., 2017; IDRD, 2019). It offers free classes daily, for 2–3 h on weekday mornings and/or evenings, and 3–6 h on Sundays and holidays. In 2018, the Bogotá Recreovía program had 70 PA instructors with an annual budget of US$1.3 million (personal communication with Recreovía Program Coordinator, September 2018).

Study design based on the Our Voice citizen science model

Our Voice in the Recreovía employed an explanatory sequential design (Pluye and Hong, 2014) to account for multiple-level device based and reported data through observation and participatory approaches integrating quantitative and qualitative methods. First, we performed direct observations to characterize the parks in relation to park occupancy, PA levels and contextual features. Then, we implemented the Our Voice model to engage community members in data collection and analysis, and advance data-driven changes to promote environmental changes with policymakers. Our Voice in the Recreovía is part of the global Our Voice Citizen Science for Health Equity Research Initiative, and is a socio-ecological model that uses accessible technology and participatory qualitative and quantitative methodologies (King et al., 2016).The entrée into this process is the Stanford Discovery Tool, a mobile data-gathering application (app) that allows residents to document neighborhood features through geocoded photographs, audio narratives and GPS-tracked walking routes (Buman et al., 2013). The tool has been culturally adapted and translated for use in Colombia (Zieff et al., 2018). Through a facilitated process, the citizen scientists then use their data to promote changes in local built and social environments that can generate positive health-related impacts (King et al., 2019). The study included four phases: (i) planning and recruitment; (ii) data collection; (iii) data analysis and consensus-building; and (iv) data harmonization and ripple effects capture.

Planning and recruitment

We selected two public parks in urban areas based on official Recreovía attendance reports. Among parks with Recreovía programs, Santa Isabel Park represented the lowest 10% in attendance rates, while San Andrés represented the upper 10% (personal communication with Recreovía Program Coordinator, September 2018).

In October 2017, participants in Recreovía PA classes at both parks were invited to engage in the Our Voice process as citizen scientists. The study was advertised via verbal announcements during and after the classes with the support of the PA instructors. Individuals 18 years of age and older were eligible to participate. Subsequently, among eligible participants we selected self-identified 25% ‘community leaders’ and 75% ‘non-community leaders’ for the final sample. We approached community leaders from pre-existing elderly groups, Community Action Boards and other neighborhood groups. Recruitment was stated as completed upon reaching saturation in the discussions with citizen scientists about positive and negative environmental features from each park. All participants read and signed a university-approved informed consent form.

Data collection and analysis

Community walks and survey

Between October and December 2017, the research team trained citizen scientists to use the Discovery Tool on mobile phones supplied by the project staff. Per the standard Our Voice Discovery Tool protocol (Buman et al., 2013), citizen scientists took photographs and recorded audio narratives describing environmental features that negatively and positively impacted their experience related to engaging in PA in the park. Anonymous, de-identified Discovery Tool data were then uploaded to a secure server at Stanford University.

Participants then completed a brief two-part survey. Socio-demographic variables assessed included participant age, gender, education and household socio-economic status. The survey also addressed participation in and satisfaction with the Recreovía program, general perceptions of the park and perceived levels of community cohesion and empowerment among neighborhood residents. Community perceptions were evaluated using five-point Likert scales from the standard Our Voice protocol (King et al., 2016).

PA levels: individual and park level

Individual-level PA and anthropometric measurement

To assess objectively the PA levels of participants, we used the ActiGraph GT3X and GT3X+ Accelerometer (ActiGraph LLC, Pensacola, FL, USA), initialized to collect data at 60-s epochs. Participants were asked to wear the accelerometer 24 h per day for 7 days using an elasticized belt around the waist at the right mid-axillary line. For wear-time validation, a minimum of 5 days with at least 10 h of wear during the awake time was required (Salvo et al., 2014). Accelerometry data were scored using Freedson cut-points for adults (Freedson et al., 1998). The cut-off point used with the accelerometer data was an average MVPA per day of >22 min, equivalent to meeting the weekly PA recommendation of ≥150 min/week. These analyses were conducted with R 3.3.2 (R Core Team, 2013).

Anthropometric data (i.e. body weight and height) were measured by trained researchers using a portable Tanita SC-240 Body Composition Analyzer (Tanita, Arlington Heights, IL, USA), and a portable Seca 213 stadiometer (Seca, Hamburg, Germany).

Parks characteristics and PA levels

To characterize the park in relation to park occupancy and MVPA, we used the System for Observing Play and Recreation in Communities (SOPARC). SOPARC is an observational method developed to assess the number of park users by gender and age group, and their PA levels (McKenzie et al., 2006). It consists of systematic micro-scans of selected target areas over a specified period of time. The method has been adapted previously for application in Colombia (reliability >80%) (Santos et al., 2016). Data were collected by five trained observers under the supervision of two field coordinators.

Commensurate with SOPARC methodology, parks were divided into target areas according to characteristics that facilitate the practice of PA (e.g. courts, parking lots, fields, walking/running tacks, exercise areas and open areas). Santa Isabel included 42 target areas and San Andrés included 25 target areas. Observations were made according to Recreovía schedules of each park. In Santa Isabel, park observations were made during three Sundays between March and May 2017. In San Andrés, park observations were made during three weekdays between May and July 2017. On each observation day in the Recreovía target areas, three 20-min observation periods were conducted. In Santa Isabel, observations were made at 9:00 am, 10:00 am and 11:00 am. In San Andrés, observations were made at 7:00 am, 8:00 am and 9:00 am. Total park users and their mean characteristics were calculated for each park on a daily basis. Observed user characteristics included gender, age group and current PA level.

To systematically assess the contextual characteristics of both parks we used the Physical Activity Resource Assessment (PARA) instrument (Lee et al., 2005). Through PARA we evaluated the quality of parks by qualifying their conditions into five domains (i.e. features for PA practice, amenities, incivilities, services and accessibility) with scores ranging from 0 to 4 (Sarmiento et al., 2017). The quality score was computed as the sum of the items.

We used descriptive statistics to analyze park and users’ characteristics. Statistical significance was tested among categorical variables using a Chi-square test, and among continuous variables using a t test. All analyses were performed using SAS 9.4 (SAS, Cary, NC, 2007).

Data analysis by the citizen scientists

Discovery tool data preparation

After the Discovery Tool data were collected by the citizen scientists, the research team prepared documents with verbatim transcriptions of each participant’s audio narratives paired with their respective photographs. Each participant received printed copies of their own data.

Community meetings to discuss the environmental data

All citizen scientists who participated in data collection were invited and encouraged to participate in community meetings organized by the research team. During March 2018, an anthropologist facilitated six 1-h community meetings (two in Santa Isabel Park and four in San Andrés Park), which were attended by a total of 24 citizen scientists (50%). Participants collectively reviewed their photos and audio transcripts and identified relevant issues to address in each park. Lawyers from the research team provided advocacy training on how to use participatory mechanisms described in the Colombia Constitution. All participants received a snack and sweatband in gratitude for their participation, as well as their anthropometric and accelerometer results with recommendations to improve health.

Following each community meeting, the research team organized the categories identified into clusters based on the themes that emerged during the citizen scientist discussions. Saturation was reached when the themes reflected similar information for each park regarding perceived enablers and barriers to practicing PA.

Community meeting with local policymakers for empowerment and change

In April 2018, four citizen scientists [Santa Isabel = 1, San Andrés = 3, (8.3%)] and three local policymakers attended a community meeting at the Institute of Sports and Recreation. The citizen scientists presented their findings and recommendations across the two sites using the visual, narrative and map data drawn from the Discovery Tool.

Members of the research team entered field notes from all community meetings into Excel (Microsoft Corporation, 2018) and thematically analyzed them by constructing narrative matrices. Themes included notions about healthy habits and healthy environments; personal experiences practicing PA; perceptions about the environment where they practice PA and ideas about how to promote PA. Participant quotes were translated from Spanish to English by members of the research team for dissemination purposes.

Data harmonization

To inform policies aimed at increasing park usage for the promotion of PA, and identify facilitators of and barriers to Recreovía participation, the research team harmonized the following data: surveys about park features, MVPA levels measured by accelerometry, SOPARC measurements and the themes that emerged from the community meetings.

RESULTS

Characteristics of parks and target areas

Both public parks were located in middle-income localities. However, Santa Isabel had higher monetary poverty and homicide rates than San Andrés (Bogotá Cómo Vamos, 2017). Santa Isabel Park is smaller (22.521 m²) than San Andrés Park (68.695 m²), but the Recreovía area of Santa Isabel was larger than the Recreovía area of San Andrés. Although both parks are public, Santa Isabel Park is crossed by streets and San Andrés Park is enclosed with a fence and has specific access hours. San Andrés Park had higher quality scores in features, amenities, safety and services compared to Santa Isabel Park (Table 1).

Table 1:

Characteristics of parks, features of the parks and socio-demographic characteristics of citizen scientists

Variable	Santa Isabel	San Andrés	p Value
Total observed park users	856	2.913
Park occupation by sex (%)			0.2487**
Female	58.0	66.9
Male	42.0	33.1
Park occupation by age (%)			<0.001**
Children	17.3	5.4
Adolescents	4.6	14.5
Adults	67.7	29.0
Older adults	10.4	51.2
Physical intensity level by sex (%)			0.290**
Female
Sedentary	14.1	13.0
Moderate	57.7	52.4
Vigorous	28.2	34.6
Male
Sedentary	16.8	15.4
Moderate	57.4	52.1
Vigorous	25.9	32.5
Characteristics of target areas
Number of areas	25	43	–
Size of areas (m2)
Average ± SD	181.65 ± 124.25	410.97 ± 583.8	0.017***
Minimum–maximum	1.6–452.4	1.08–2326.81
Median	170.6	230.0
Interquartile range	109.2–222.7	97.38–342
Type of area (%)
Recreovía área	8.00	4.65	<0.001**
Sports area	36.00	51.17	<0.001**
Playground area	20.00	6.97	<0.001**
Open area	8.00	16.28	<0.001**
Skating/roller track	4.00	13.95	<0.001**
Walking/running track	12.00	4.66	<0.001**
Strength/stretching exercise area	12.00	2.32	<0.001**
Conditions (%)
Usable	99.52	100.00	0.163*
Accessible	100.00	96.77	0.000*
Equipped	71.98	51.24	0.000*
Empty	29.95	31.59	0.679*
Organized	39.13	35.82	0.473*
Supervised	19.32	16.92	0.423*
Quality of park scores
Features	38	73	<0.001*
Amenities	8	18	<0.001*
Incivilities (e.g. cleaning, esthetics and safety)	15.00	26.00	0.009*
Services	1.00	6.00	0.003*
Accessibility	2.00	2.00	0.999*
Socio-demographic characteristics of citizen scientists
Total citizen scientists	17	31
Age (mean ± SD)	52.29 ± 16.17	48.90 ± 14.56	0.462*
Gender (%)
Female	88.24	87.10	0.909**
Male	11.76	12.90
Education (%)
None—primary school or less	5.88	0.00	0.470**
Non-completed secondary school—graduated from secondary school	35.29	29.03
Technical degree/1–3 years of university studies—undergraduate degree	52.94	58.06
Postgraduate degree	5.88	12.90
Household socio-economic status (%)
Low (1–2)	11.76	9.68	0.071**
Middle (3–4)	88.24	90.32
High (5–6)	0.00	0.00
Meeting with PA recommendations (by accelerometry) (%)	85.71	100.00	–
Mean daily MVPA by accelerometry (min)	54.6 ± 22.0	73.0 ± 32.6	0.066*
Anthropometric characteristics (%)
Normal	51.60	41.20
Overweight/obese	48.40	58.80
Self-perceived health status (%)
Excellent	29.41	19.35	0.428**
Very good	11.76	25.81
Good	41.18	48.39
Regular	11.76	6.45
Bad	5.88	0.00
Time spent in the Recreovía (min)	152.6 ± 63.24	125.0 ± 45.55	0.087*
Reasons to attend to the Recreovía (%)
Health	88.24	96.77	0.243**
Perceptions toward the park improvement during Recreovía days (%)
Parks quality	41.18	64.52	0.036**
Friendliness of the park	88.24	74.19	0.417**
Amount of garbage and debris	5.88	48.39	0.009**
High/very high satisfaction with Recreovía program (%)
Organization	100.00	96.77	0.454**
Accessibility to activities	94.12	96.77	0.66**
Duration of classes	94.12	93.55	0.232**
PA classes	100.00	96.77	0.454**
Professionalism of instructors	100.00	100.00	0.999**
Available space	94.12	93.55	0.938**
Somewhat agree/totally agree on questions of community empowerment (%)
This is a community where people support each other	64.71	74.19	0.222**
I can influence decisions that affect my community	41.18	54.84	0.664**
Working together with others in my community, we can influence the decisions that affect this community	70.59	83.87	0.224**
People in my community know who to talk to make changes in our community	47.06	54.84	0.055**

^* t Test.

^**Chi-square test.

^***Mann–Whitney–Wilcoxon test.

Park users and activities observed

We observed 3769 park users, 22.7% of whom were observed in Santa Isabel and 77.3% in San Andrés. Across both sites, most observed park users were women (64.9%) and adults (79.7%). However, in San Andrés, the percentage of older adults (51.2%) was higher than in Santa Isabel (10.4%). The percentage of users engaging in MVPA across both sites was virtually equal for men (84.8%) and women (86.3%).

Characteristics of citizen scientists

Table 1 shows the characteristics of citizen scientists who participated in the Our Voice process. A total of 48 citizen scientists were enrolled in the study (Santa Isabel = 17, San Andrés = 31), but for one participant, Discovery Tool data were not properly recorded. Citizen scientists were mostly women, and on average were 50 (SD = 15) years old, with more than a high school education, and from middle-income households. Citizen scientists in general had perceptions of good-to-excellent health status, reported spending an average of 135 min/per day (SD = 53.5) in the Recreovía, and described their main reason for participating as ‘health’ (94%). Reflecting the high PA levels of users of the Recreovía, most citizen scientists from both parks met PA recommendations and in general had perceptions of good-to-excellent health status, although 50% of the citizen scientists were overweight or obese. Overall, citizen scientists were highly satisfied with the Recreovía classes and reported that these classes increased the friendliness of the parks. However, citizen scientists from Santa Isabel were less satisfied with the cleanliness of the exercise class areas compared to San Andrés.

Citizen scientists from both parks reported perceiving that in their communities, people support each other and that working together could strengthen their possible influence on the decision-making process.

Discovery tool data and community meetings

Perceived enablers and barriers to practicing PA in parks

Overall, 299 photos and 297 audio narratives were recorded by 47 citizen scientists. Of these, 24 participated in the community meetings to review, analyze and discuss the Discovery Tool data. As assets in both parks, citizen scientists highlighted the encouragement of PA through the Recreovía program; availability of outdoor fitness equipment; maintenance of green areas and the diversity and maintenance of courts. They reported that their neighborhood parks are inclusive spaces offering a range of facilities adequate for diverse age and gender groups.

Regarding barriers to park usage, Santa Isabel’s residents were particularly attentive to safety and cleanliness concerns. They identified as top concerns poor management of garbage (including animal waste), lack of safety and limited interactions among neighbors. They suggested that this was due to the presence of a homeless population, poor community recycling practices and the use of the park for the sale and consumption of illegal drugs. Conversely, San Andrés residents expressed that safety and cleanliness were not challenges as their park is enclosed, has private security, prohibits pets and is administered by the mayor’s office. The only expressed sporadic concerns about the San Andrés Park were over the lack of maintenance of courts, trails and green areas, and limited night-time lighting. Other concerns from participants of both parks included occasional problems with the sound quality of the music accompanying the PA classes and the irregularities of the court surfaces for the PA classes.

Citizen scientists’ experiences accessing and engaging with PA classes in parks

During the community meetings, citizen scientists also noted the co-benefits of community-based PA programs, which went beyond PA promotion to include increasing self-confidence, social capital, mental health and civic engagement. Citizen scientists revealed that experiencing such outcomes supports their regular PA practice.

One older adult said that PA helps them to maintain their vitality and self-confidence.

I thank the doctor who told me to come here to walk, because otherwise I would be in a wheelchair (…) from the age of 40 I have exercised; I do not like pills. When I feel pain in my joints I go out and I heal myself with exercise.

(Santa Isabel)

Another participant mentioned that Recreovía park use amounts to taking care of mental health and building bonds with neighbors.

I come to the park to walk or do physical activity and I disconnect; I have fun, and I’m happy. And I’m with friends, sharing, having a good time.

(Santa Isabel)

A third citizen scientist noted the importance of supporting the District’s efforts to create programs that benefit public health.

These spaces allow people to exercise without costs (…). I think it’s very good that the government cares about this. It’s a good policy; so, we should come and take advantage of it.

(San Andrés)

Building consensus to advocate for healthier environments

During the community meeting with local policymakers, citizen scientists praised the high quality of the Recreovía program and the expertise of the PA instructors as critical in meeting their communities’ needs for healthy and active living. The dialog among users of both parks also elicited a proposal to create specific Recreovía hours for the elderly.

Policymakers, in turn, stressed the value of receiving data-driven feedback from the community, as this afforded them the opportunity to clarify information about district-level accountability for the Recreovía program and general park maintenance. At the conclusion of the meetings, policymakers committed to bringing the community’s concerns to the proper administrative offices.

Short-term outcomes and longer-term ripple effects of Our Voice in the Recreovía

Several short-term effects were documented 6 months following the conclusion of the Our Voice process. First, Our Voice contributed to strengthening and sustaining the Recreovía program in Santa Isabel, where low attendance had previously led to a plan to discontinue it. Now, in response to community interest and needs highlighted through the Our Voice process, the site continues operating on weekends, and an additional site has been developed nearby. Second, citizen scientists were empowered as advocates for the Recreovía program. During a presentation at the National Seminar on Policies and Programs to Promote Healthy Habits and Lifestyles held by the Ministry of Sports (Bogotá, October, 2018), a citizen scientist from the Santa Isabel assessment presented on Recreovía’s community benefits, highlighting that the Our Voice approach helped to foster dialog and shared solution-building within communities and with policymakers. Third, the relationships among community members, policymakers and academic researchers aimed at advancing activity-promoting changes in the Recreovía program were strengthened. Specifically, the coordinator of the Recreovía program became involved in an ongoing Recreovía evaluation project, which has resulted in the scheduling of monthly meetings and further steps to clarify the impacts of the program. In addition, before implementing new Recreovía hubs in underserved areas, the research team was called in to conduct a formal impact evaluation, and the coordinator presented results from research papers and policy briefs about Recreovía to the director of the Institute and at an international scientific congress. The data-driven decision-making process to open new hubs reflects policymakers’ engagement with the research activities and the Our Voice process (personal communication with coordinator, January 2019).

Six months after implementation, we observed project ripple effects related to local capacity building. The director of the Ministry of Sports’ Healthy Habits and Lifestyles Program invited the research team to present the results of Our Voice in the Recreovía and train the program’s municipal-level instructors on Our Voice methodology during the National Seminar (Bogotá, October, 2018). In total, 250 instructors participated in the training, which included virtual and in-person sessions, and featured a citizen scientist who shared her experience about Our Voice in the Recreovía. During the same month, the researchers were also invited to share the Our Voice research experience and results at the International Physical Activity Seminar organized by Bogotá’s Institute of Sports and Recreation.

DISCUSSION

The Our Voice in the Recreovía study, the first of its kind in a middle-income Latin American city, discovered relevant factors supporting regular community-based PA practice related to the interaction between individual behaviors and healthy environments. Quantitative measurements regarding characteristics of parks showed that San Andres Park had better quality than Santa Isabel Park. The Our Voice approach helped researchers and decision makers to understand that beyond characteristics of the parks, Recreovía users highlighted broad factors supporting their PA engagement. Over the course of the project, citizen scientists expressed high satisfaction with the Recreovía program, and reported that it improved the friendliness of the environment of both parks. They elucidated program co-benefits that extended to fostering self-confidence, social capital, mental health and civic engagement. At the same time, they reported that the main barriers to use were related to park maintenance, cleanliness and safety. The opportunity for residents of two different communities to listen to each other and engage in dialog with local policymakers allowed all actors to understand the complex array of individual, social and political factors that support regular PA. Documentation of short-term outcomes and longer-term ripple effects showed that both citizen scientists and policymakers can effectively engage with research activities to improve their environments, and enhance existing programs. In study follow-up, local and national policymakers have acknowledged the potential of the Our Voice process to strengthen community-based PA programs and to build local capacity for health-promoting change.

The results of Our Voice in the Recreovía are consistent with previous studies using the Our Voice approach, in which citizen scientists identified and advocated for improving aspects of their neighborhoods that can support healthy living. Building on this foundational research, Our Voice in the Recreovía contributes to understanding the impacts and promise of park-based programs by assessing a program with 1.2 million participants, mainly benefiting women, embedded in social policy for NCD prevention and PA promotion. Our Voice in the Recreovía may also serve as a model for direct resident engagement in building and strengthening a range of park-based community programs to promote healthy living. In fact, worldwide, similar Recreovía-type programs seeking to increase levels of PA and decrease health inequalities by reaching low-income women exist in countries like Argentina, Brazil, Chile, Cuba, Ecuador, Guatemala, Mexico, the USA and Australia (Díaz del Castillo et al., 2017).

While citizen scientists described adequate facilities in both parks (i.e. recreation courts, activities, outdoor fitness equipment), there was a big usage gap between the two that can be explained at least in part by environmental factors. Public parks were positively described as inclusive spaces with the potential to enhance social capital, self-care and civic engagement. However, concerns about safety hindered the use of the parks for PA. This is consistent with previous studies about lack of safety as an environmental deterrent for PA practice in public spaces (Francis et al., 2012; Rachele et al., 2016; Sawyer et al., 2017). However, Our Voice in the Recreovía provided more nuanced insights into how engaging residents and policymakers can create a synergy of bottom-up and top-down efforts to address safety concerns and encourage active living.

This study’s findings helped to clarify the varying levels of success of Recreovía programming across sites. Our Voice in the Recreovía proved to be a multi-level, multi-sectoral community-based participatory experience that drew on local knowledge to leverage PA champions and advocate for a sustained park-based PA program.

Strengths and limitations

Our Voice in Recreovía had three main strengths. First, at the methodological level, we combined quantitative, qualitative and technology-based methods. Second, at the policy level, we were able to stimulate productive data-driven dialog among community members and policy makers in a variety of contexts (e.g. meetings, policy seminars). Third, through ongoing contact with program coordinators, we have been able to track longer-term ripple effects, such as using citizen scientist data to leverage the maintenance and opening of new Recreovía sites.

This study has limitations that should be considered. First, citizen scientists were recruited during their participation in the Recreovía; therefore, we did not recruit general park users who were not also Recreovía participants. This might have created a complacency bias among citizen scientists, regarding acceptance of the program. In addition, as the Our Voice process includes several phases, the number of citizen scientists participating in each phase of the process diminished over time. Despite this attrition, participants were able to reach consensus around the central ideas and activities, underscoring that a relatively small number of engaged residents can enact positive change. Second, while the project enjoyed the engagement of key local stakeholders (e.g. the coordinators of the Recreovía program and the National Healthy Habits and Lifestyles program), the community meetings had relatively low policymaker participation. Finally, project ripple effects were evaluated over a relatively limited time frame (i.e. 11 months). Fortunately, the commitment to nurture the relationships formed during this initial period will allow further documentation of potential ripple effects.

CONCLUSIONS

This study underscores the importance of mixed-methods approach to understand factors supporting community-based PA behaviors. As such, relevant identified factors influencing the use of the Recreovía program include its role in enhancing social and individual well-being. Recreovía improves the friendliness of the park environment; additionally, it fosters self-confidence, social capital, mental health and civic engagement. The Our Voice process elicited community reflection, dialog, empowerment, advocacy and action. The process allowed residents to reflect on outcomes of their PA practice while systematically collecting and discussing data of direct relevance to their local community. This gave them an evidence-based ‘voice’ to advance pragmatic solutions that promote active living in their local community. Our Voice facilitated the interplay among researchers, citizen scientists and policymakers to optimize community programs. Combining quantitative and qualitative methods, this study provides an adaptable framework for using technology-driven participatory design to evaluate park based as well as other types of interventions.