T H E

S

O F

i E N

L I F E S T Y L E

C H A N G E

Parks and Recreation

The Potential for Pocket Parks to Increase Physical Activity Deborah A. Cohen, MD, MPH; Terry Marsh, MPH; Stephanie Williamson, BA; Bing Han, PhD; Kathryn Pitkin Derose, PhD, MPH; Daniella Golinelli, PhD; Thomas L. McKenzie, PhD PURPOSE

Abstract Purpose. To assess the use of new pocket parks in low-income neighborhoods. Design. The design of the study was a quasi-experimental post-test only comparison. Setting. Los Angeles, California, was the setting for the study. Subjects. Subjects were park users and residents living within .5 mile of three pocket parks and 15 neighborhood parks. Intervention. The creation of pocket parks. Measures. We used the System of Observing Play and Recreation in Communities (SOPARC) tool to measure park use and park-based physical activity, and then surveyed park users and residents about their park use. Analysis. We surveyed 392 and 432 household members rvithin .5 mile of the three pocket parks before and afler park con.üruction, respectively, as well as 71 pocket park users, and compared them to 992 neighborhood park users and 342 residents living within .5 mile of other neighborhood parks. We compared pocket park use to playground area use in the larger neighborhood parks. We used descriptive statistics and generalized estimating equations for the analysis. Results. Overall, pocket park use compared favorably in promoting moderate-to-vigorous physical activity with that of existing playground space in nearby parks, and they were costeffective at $0.73/MET hour (metabolic equivalent hour) gained. Pocket park visitors xualked an average of .25 miles to get to a park. Conclusions. Pocket parks, ruhen perceived as attractive and safe destinations, may increase physical activity by encouraging families with children to walk there. Additional stratèges and programs may be needed to encourage more residents to use these parks. (AmJ Health Promot 2014;28[3s]:Sl 9-S26.)

Key Words: Physical Activity, Parks, Pocket Parks, Accessibility, Prevention Research. Manuscript format: research; Research purpose: intervention testing/ program evaluation; Study design: quasi-experimental; Outcome measure: behavioral, financial/economic; Setting: local community; Health focus: physical activity; Strategy: built environment; Target population age: yotith, adults, seniors; Target population circumstances: income level, geographic location

Deborah A. Cohen, MD, MPH; Terry Marsh, MPH; Stephanie Williamson; Bing Han, PhD; Kathryn Pitkin Derose, PhD, MPH; and Daniella Golinelli, PhD, are with the RAND Corporation, Santa Monica, California. Thomas L. McKenzie, PhD, is with San Diego State University, San Diego, California. Send reprint requests to Deborah A. Cohen, f776 Main Street, Santa Monica, CA 90405; [email protected]. TJm manu.icript was submitUd April 30, 2013; rexiisions xoere reqiiesled Juty 30, 2013; the manuscript was accepted for ¡mbtkaiion August 13, 2013. ('•itftyrigtU © 2014 by American Journal of Health Promotion, Inc. 0H9()-ini/14/$5.00 + 0 DOI: I0.427H/ajhpl30430-QUAN-213

American Journal of Healfh Promofion

Physical inactivity is an important and malleable risk factor accounting for 10.8% of all-cause mortality, as well as 6.7% to 12.4% of mortality from heart disease, diabetes, colon, and breast cancer.' Inactivity is associated with multiple mtisculoskeletal problems and a host of other chronic diseases. Fewer than 5 percent of adults and half of children in the United States meet the national guidelines of 150 minutes per week and 60 minutes per day of moderate-to-vigorotis physical activity (MVPA), respectively. Yet, adults need routine physical activity to limit the risk of chronic diseases and children need to engage in moderate and, in partictilar, vigorous physical activity for healthy bone and muscle development." Pre\ious research indicates that children are most likely to be active when they are outdoors.^ Indoor environments typically support sedentary behaviors and have substantial constraints to activities like running, climbing, jumping—movements that engage large muscle groups and strengthen bones. Inner city neighborhoods, in particular, have limited access to green space where children can be active. Lack of safety is also a widespread fear in lowincome, high-crime neighborhoods.^'*' To address this problem and facilitate more otttdoor play and physical activity (PA), philanthropic organizations have made efforts to improve the conditions of parks, as well as to convert vacant lots and ttndesirable urban parcels into pocket parks, also called mini-parks, vest-pocket parks, or parkettes. Pocket parks are often quite small (less than 1

January/February 2014, Vol. 28, No. 3 Supplemenf

SI9

Table 1 Neighborhood and Park Characteristics Households in Poverty, %

Park Name Marson Comparison parks for Marson (n = 6)—Average Alpine Baldwin Hills Denker Harvard Rancho Ciénega Ramon Garcia Beverly Comparison parks for Beverly (n = 6)—Average Algin Sutton Lincoln Heights Mt. Carmel Ross Snyder South Trinity Broadway Comparison parks for Broadway (n = 3)—Average Hoover Toberman 109th Street

Hispanic (Ethnicity), %

African-American (Based on Race), %

Population Density (0.5 Mile)

Size (Acres)

30.3 30.3

70.2 51.1

5.7 31.3

10,726 16,088

0.2 10.3

30.1 29.6 32.1 30.0 28.3 31.7

29.8 44.9 58.7 45.3 35.4 92.3

3.9 45.2 35.8 48.7 53.7 0.5

11,270 13,785 13,684 14,443 22,924 20,423

1.9 10.0 2.9 12.0 28.0 7.2

39.9 39.1

82.1 71.2

31,320 19,398

0.3 8.7

39.1 37.6 39.3 40.2 40.7 37.5

54.0 67.5 59.0 80.8 79.6 86.3

3.1 20.4 44.2 1.3 35.4 14.4 19.5 7.8

16,351 15,239 19,967 17,670 28,520 18,639

16.0 2.0 3.4 11.0 18.0 2.0

40.2% 42.0

75.3% 67.9

23.5 17.0%

18,644 18,695

41.1 42.4 42.5

65.7 77.9 60.2

8.2 4.3 38.4

23,035 20,465 12,586

0.2 2.9 2.9 2.7 3.1

Pocket park neighborhoods defined as within 0.5-mile radius. Comparison park neighborhoods withir11 -mile radius, except for population density. Both based on 2000 Census Block Groups.

acre) compared to neighborhood or community parks, and they generally serve the immediate population living within .25 to .5 miles of the park. Pocket parks also usually have limited facilities, offer few or no programs, lack indoor facilities, and are not staffed. To increase .safet)' and reduce crime, the entire area is typically fenced and can be locked outside the hours of operation. Although there is a growing literature on physical activity in parks,^"''^ to our knowledge this body of work has not explicitly addressed the role that pocket parks might play in physical activity promotion. The construction of three pocket parks in Los Angeles, Galifornia, represented a change in the built environment in their surrounding communities and created a natural experiment to study their effects on park use and physical activity among neighborhood residents. Building new parks can be expensive, given numerous rigorous codes that dictate the quality of play surfaces and equipment safety features. The three pocket parks

S20

American Journal of Health Promotion

in this study were built for a total of $3 million, or about $1 million each. This article reports our findings from before and after assessments of physical activity in the new pocket park developments and compares their use to that of playground areas in nearby, larger neighborhood parks. METHODS Design Because the pocket park spaces were not used prior to construction, we compared their use after construction to existing neighborhood parks that served similar sociodemographic populations. Sample The three pocket parks in this study were developed by a local nonprofit group and its funders. The census tracts around the sites had high rates of household poverty (range 30%41%) and substantial minority populations (Latino range: 70%-80%; African-American range: 3%-17%; Asian:

0%-16%). The number of individuals living within .5 mile of these parks varied substantially, with Marson Park at 10,726 people; Broadway Park at 18,644; and Beverly Park at 31,320. Table 1 describes the characteristics of the pocket parks and the local parks to which they were compared. To evaluate park use, we assessed the number of people visiting the pocket park locations before and after construction. We compared the numbers of people using those parks postconstruction to the numbers using a sample of playgrounds in larger neighborhood parks that were matched to each of the pocket parks by the percentage of households in poverty (2000 U.S. Gensus). Marson and Beverly each had six matched parks and Broadway had three. Observation of Park Users We assessed park use with the System for Observing Play and Recreation in Gommunities (SOPARG) tool, which was designed to measure park use and physical activity^^ and has high inter-

January/February 2014, Vol. 28, No. 3 Supplement

Table 2 MVPA and METs and Total People/Number Observed in Pocket Parks and in Play Areas of Comparison Parks Park Name

Estimated METS During Observation Times

Number Observed Engaging in MVPA

Total Number Observed

Pocket parks Marson Beverly Broadway Average

m

327 554 90 324

148

M-

260

14 36

t47

Marson matched cluster Marson Alpine Baldwin Hills Denker Harvard Rancho Ciénega Ramon Garcia Average

327 189 150 3 306 483 549 280

m

148 68 51 2

' .

m n

.103 152 204 97

il

260 295.

m ü

Beverly matched cluster Beverly Aigin Sutton Lincoin Heights Mt. Carmel Ross Snyder South Trinity Average Broadway matched cluster Broadway Hoover Toberman 109th Street Average Overall average

554 944 222 195 374 1017 404 526 90 612 107 66 262 374

142 24

98 70

M

m

141 46 m

147 373 173 193

. , -, •

14

'



i m.

m.

•"'

106 12 •

m -



'

"



200 53 23 92 134

METS indicates metaboiic equivalents; MVPA, moderate-to-vigorous physical exercise.

observer reliability, ranging from .80 to .9^}'^ Trained promotoras (i.e., community health workers) observed all areas of the pocket parks and comparison playground areas four times a day during each of the seven days of the week at baseline and at follow-up. The four daily observation start times were divided into early morning, (7:30, 8:30, or 9:30), late morning to noon (10:30, 11:30, or 12:30), afternoon (1:30, 2:30, or 3:30), and evening (4:30, 5:30, and 6:30), with different hours observed in each of the four time periods on different days to cover all the hours. We coded the apparent gender, age group (child, teen, adult, senior), race/ethnicity (Latino, black, white, other), and activity level (sedentary.

American Journal of Healfh Promotion

walking, vigorovis) of each observed park user. For each park area we also recorded whether or not it was accessible, usable, equipped, supervised, or dark, and if an organized activity was taking place. Observations were not conducted on holidays, and any observations cancelled because of inclement weather were made up at the same time of day on the same day of the next week that had clement weather. Baseline pocket park observations were conducted between midjuly and mid-August 2006, and follow-up assessments were completed during the same season in 2008. Assessments of comparison parks were done in 20082009. SOPARC observations were

conducted at similar times of day at the pocket parks and their comparison parks, and each park was assessed on 28 occasions during a week, making the findings comparable. Surveys: Park Users and Residents In addition to conducting SOPARC observations, the promotoras, who were bilingual in English and Spanish, interviewed pocket park users and residents living within a half mile. Only residents were interviewed at baseline because the parks had not yet been constructed. The interviews were conducted in Spanish or English as preferred by the respondent. Respondents in pocket parks and the larger neighborhood parks were systematically recruited from the most and least busy areas throughout the park by gender and activity level. A random sample of household addresses within .25 mile of the pocket park and another between .25 and .5 mile of the park was selected, and field staff went door-to-door to conduct the surveys. If the first survey attempt was unsuccessful, data collectors returned to a house up to four additional times (three additional times at comparison parks), attempting to survey the occupant 18 years or older at the next birthday before replacing the address with an alternate address in the same strata. We attempted to administer surveys at the same addresses at baseline and follow-up. Many houses around the parks with the highest poverty rate were not accessible (i.e., gated or fenced), so in-home resident surveys were sometimes not possible. In these cases we replaced the in-home resident surveys with intercept surveys conducted at high pedestrian traffic areas (e.g., bus stops, store fronts) within a half mile of the park. Eligible survey respondents were 18 years of age or older and resided within a half mile of the pocket park (or 1 mile of the comparison parks). If the respondent had a child under the age of 18 years, we also asked questions about the child's park use. The suirey content and administration were similar for pocket and comparison parks. We weighted the survey sample so the distribution of respondents was simi-

January/February 2014, Vol. 28, No. 3 Supplement

S21

Table 3 Self-Reported Park Use From Household Surveys Conducted Before and After Pocket Park Construction Pocket Parks Residents Before (n = 392), %

Residents After (n = 432), %

11.1 25.8 9.6 71.7 10.8

33.9 35.7 14.4 71.1 21.8

Adult visits any park > once per week Engage in leisure time exercise Exercise in park Half or more of leisure time exercise is vigorous Use of other parks > once per week

lar to the underlying local population as measured in the 2000 U.S. Census. To estimate distance traveled to the park, we asked survey respondents to provide the nearest street intersection to their house. These intersections and park addresses were geocoded using AicGIS, and Cartesian distances between the two were estimated. Pocket Park Descriptions

Three pocket parks were developed, two in previously vacant lots and the third in a former community garden site. A local charitable organization funded the construction and a nonprofit community-based organization managed the projects. All three pocket parks had playground equipment and benches installed, and a walking path was developed around the perimeter of the Beverly, the largest park. All were fenced and enclosed by gates that could be locked. Other than for opening day ceremonies, substantial outdoor marketing efforts (e.g., banners, flyers, notices) were not visible in or near the pocket parks. Analysis

We tested the mean difference between the number of park users of pocket parks and the users of comparison park playgrounds by a repeatedmeasure Poisson regression and using the generalized estimating equation (GEE) method. This model controlled for weekend dates, proportion of households in poverty, and total population density within 1 mile of the park. We also included indicators for each pocket park and its matched comparison parks to account for the potential effects due to artificial matching.

S22

American Journal of Health Promotion

Estimating Cost-Effectiveness We estimated cost-effectiveness by amortizing the cost of building each park over 30 years. Although playground equipment is usually warrantied for 15 years, the construction involved considerable costs to prepare the site and mitigate existing conditions (e.g., asphalt and concrete surfaces); these composed a significant part of the construction expense. We did not include the costs of park maintenance. We also assumed that the metabolic equivalents (METs) expended dtiring the week of measurement were similar to the 329 days (47 weeks) of the year when there is no precipitation in Los Angeles. We calculated the dollars spent per MET-hour expended in the parks per year. The method interprets cost-effectiveness based upon achieving the nationally recommended guidelines of 150 minutes of MVPA per week or 2.5 hours at 4.5 MET-hours (11.25 METs) in light of the cost of per capita health care and the contribution of physical inactivity to health care costs (about 2.5%-5%). Given the current cost of health care, it is theoretically cost-effective to spend between $0.50 and $1.00 per METhour.'••"' RESULTS Park Characteristics

Table 1 describes the physical characteristics of the pocket parks and their comparison parks, as well as the sociodemographic characteristics of the poptilation they serve in a half-mile radius for pocket parks and a mile radius for neighborhood parks. Neighborhood parks are about 15 to 50 times larger on average than pocket

The potential for pocket parks to increase physical activity.

To assess the use of new pocket parks in low-income neighborhoods...
8MB Sizes 0 Downloads 0 Views