Introduction

The built environment has been increasingly recognized as an important dimension in the framework of social determinants of health [1, 2]. There is a growing body of literature linking the built environment to physical and mental health [3,4,5,6,7,8,9]. For example, the design of urban environments, such as neighborhood walkability, transportation, food density, parks and recreational facilities, and aesthetics, has been associated with health behaviors such as diet [10] and physical activity [11]. These health behaviors, in turn, contribute to the development of diseases such as obesity, type 2 diabetes, cardiovascular disease, and cancer [12,13,14,15].

While objective physical attributes of the built environment are important (e.g., concrete amount of green space), perceptions of the environment, as proxies of how people experience their environment, may be just as critical. For example, built environment attributes have been found to have an indirect effect through the perception of the built environment on moderate-to-vigorous physical activity [16]. Indeed, perceptions of the built environment, such as the sense of satisfaction with one’s neighborhood, has been shown to mediate the relationship between objective measures of the environment and health [17, 18]. The perceived physical environment has also been found to be positively associated with physical health outcomes such as obesity [19] and mental health outcomes such as depression [20,21,22].

While the conventional biomedical model emphasizes disease outcomes, quality of life, as measured by various validated survey instruments, is increasingly seen as an important health outcome in its own right [23]. Research on the relationship between the built environment and measures of health-related quality of life (HR-QoL) is limited but emerging [3, 8, 24,25,26,27,28]. In one study, land-use heterogeneity and housing density were found to be associated with HR-QoL [28]. Perceptions of the built environment, such as perceived street noise and traffic safety, have also been related to mental and physical health components of quality of life [3, 27]. Furthermore, research has shown a positive correlation between neighborhood walkability and HR-QoL in Australia and China [26, 27]. Increased perceived diversity, safety and esthetics were found to be associated with higher physical and mental HR-QoL [27]. However, to our knowledge, little is known about these relationships in minoritized populations in the United States.

To bridge this gap, we first tested the psychometric properties of the Neighborhood Environment and Walkability Survey (NEWS) in New York City (NYC), recognizing its unique built environment compared to other major American cities. Subsequently, we examined the association between perceptions of the neighborhood environment and HR-QoL in a sample of mostly Black and Latino residents in NYC. We hypothesized that positive perceptions of the neighborhood environment would be associated with higher physical and mental components of HR-QoL.

Method

Study design

This study utilized the baseline adult survey data from the Physical Activity and Redesigned Community Spaces (PARCS) Study, a natural experiment evaluation on the impact of citywide park redesign and renovation on health and wellbeing [29]. The baseline survey data were collected from Summer 2016 – Spring 2018 (except during winter months) in 54 park neighborhoods throughout the five boroughs of NYC. A convenience sample of adult residents was invited to participate in the study survey.

Park neighborhood selection

To be included in the study, NYC Parks and Recreation identified parks in neighborhoods which met two of three selection criteria: high poverty (≥ 20% population below poverty line), high population growth (≥25% growth during 2000–2010) or high population density (≥ 110 people/acre). For the purpose of the PARCS evaluation study, study park buffer zones were defined as the area within a 0.30-mile radius from the perimeter of each park.

Sample selection

To meet the PARCS Study eligibility criteria, participants needed to live within the designated 0.3-mile buffer zone. Eligible participants were adults ≥18 years of age who owned a smart phone, spoke English, Spanish or Chinese (Mandarin or Cantonese), did not have any mobility issues and were intending to stay in the neighborhood for at least 4 years (due to the longitudinal nature of the parent study). This paper included observations from 1252 participants who provided any data on the relevant survey items described below.

Measures

Health-related quality of life (HR-QoL)

HR-QoL is a multidimensional concept that includes individuals’ perception of their physical and mental functioning, limitations due to physical or mental health problems, bodily pain, vitality, general health, and social functioning [30]. We used the well-validated Short-Form 12 (SF-12) survey [31] in this study. Eight domains were covered by the SF-12 survey questions including 1) limitations in physical activities because of health problems 2) limitations in social activities because of physical or emotional problems 3) limitations in usual role activities because of physical health problems 4) bodily pain 5) general mental health 6) limitations in usual role activities because of emotional problems 7) vitality 8) general health perceptions. For the purpose of our analysis, we used the composite scores on mental health and physical health as our two primary outcome variables, using the standard scoring protocol for SF-12 [32]. The higher the score, the better the self-reported HR-QoL in mental health and physical health (range 0–100).

Neighborhood Environment and Walkability Survey (NEWS)

NEWS was developed in 2002 to measure resident perceptions of neighborhood characteristics [33], including residential density, land use mix, street connectivity, infrastructure for walking/cycling, neighborhood aesthetics, traffic and crime safety, and neighborhood satisfaction [33]. This survey we used included: places for walking and cycling (5 items), neighborhood surroundings (6 items), safety from traffic (8 items), safety from crime (6 items), and neighborhood satisfaction (18 items) [34]. Items for the first four original subscales all had four response options (1 = strongly disagree, 2 = somewhat disagree, 3 = somewhat agree, 4 = strongly agree) while items in the neighborhood satisfaction subscale had five response options (1 = strongly dissatisfied, 2 = somewhat dissatisfied, 3 = neither satisfied nor dissatisfied, 4 = somewhat satisfied, 5 = strongly satisfied). Items were coded such that the higher the subscale score, the more positive the attitude regarding the construct.

Covariates

Demographic variables included age, gender (female vs. male as reference group), ethnicity (Black, Other vs. Latino as reference group), BMI (overweight 25–29 kg/m2, obese > = 30 kg/m2 vs. under and normal < 25 kg/m2 as reference group,), annual household income (< $20,000 vs. ≥ $20,000 as reference group), education (high school graduate, some college or higher vs. less than high school as reference group), employment (self-employed, homemaker, student, retired, unable to work, unemployed for 1 year or more, unemployed for less than 1 year, don’t know/not sure vs. employed as reference group), and marital status (divorced, widowed, separated, never married, a member of an unmarried couple living together vs. married as reference group).

Statistical analyses

NEWS was originally developed in Seattle and cross-validated in Baltimore [34]. Factor analysis has been used to adapt to countries other than the United States, such as Australia and Western Europe [35]. Due to the unique neighborhood characteristics of NYC, we first tested the psychometric properties of NEWS using exploratory factor analysis (EFA) with oblimin rotation [36]. Items with factor loading < 0.40 were dropped in subsequent analysis [37]. We used Cronbach’s alpha [38] was used to assess the internal consistency of the newly constituted subscales based on the factor analysis, considering Cronbach’s alpha > 0.7 as good and acceptable. Spearman correlation tests were performed between NEWS subscales and HR-QoL components.

To deal with missing data, we used sequential regression modelling to impute the missing values in the NEWS scales, QoL scores and covariates [39]. Twenty imputations were generated, taking into account 14% of missing cells in the original dataset. Descriptive statistics were estimated for all the NEWS subscales, HR-QoL components, and covariates with and without imputation. Then, we regressed mental and physical HR-QoL scores (in separate models) on the newly formed NEWS subscales, adjusting for covariates, with multiple imputed dataset. Generalized Estimating Equation (GEE) models [40] were used to account for the clustering effect of park neighborhoods. Sensitivity analyses were conducted using the complete cases for GEE modeling.

To test the multicollinearity between NEWS subscales, variance inflation factors (VIF) were calculated in the GEE models. The VIF for the NEWS subscales ranged from 1.1 to 1.6, showing no multicollinearity among NEWS subscales [41].

All analyses were performed using R version 4.0.4 [42]. The package of “psych” in R was utilized for the EFA [43]. We used the “MICE” package in R for multiple imputations [44]. The “geepack” package in R was utilized to test GEE models [45]. Alpha was set at p < 0.05.

Results

Factor analysis of NEWS subscales

Using the raw data, Cronbach’s alpha showed poor internal consistency for three of the original survey subscales: places for walking and cycling (0.57, 95% Confidence Interval (CI) = 0.54–0.61), safety from traffic (0.64, 95% CI = 0.61–0.67), safety from crime (0.66, 95% CI = 0.63–0.69). Neighborhood surroundings had good internal consistency with a Cronbach’s alpha of 0.79 (95% CI = 0.77–0.80) and neighborhood satisfaction had high internal consistency with a Cronbach’s alpha of 0.91 (95% CI = 0.90–0.92).

To adapt NEWS for NYC, EFA was performed for the above scales besides neighborhood satisfaction. EFA yielded two subscales: enablers (e.g., trails, sidewalks, esthetics) vs. barriers (e.g., high crime and traffic) of walkability. EFA results for the first two subscales with items demonstrating a factor loading ≥0.40 are shown in Table 1. Cronbach’s alpha was 0.84, 0.78 and 0.91 for walkability enablers, walkability barriers, and neighborhood satisfaction, indicating good internal consistency. The higher the NEWS subscale value, the stronger agreement on the statements for enablers and barriers.

Table 1 Exploratory factor analysis of the Neighborhood Environment Walkability Scale (NEWS)
Full size table

Participant characteristics and bivariate associations of NEWS and HR-QoL

Table 2 provides the descriptive statistics of the sample with and without imputation. The imputed results closely matched those of the non-imputed data. The sample had a mean (±standard error) age of 38.8 ± 12.5 years and was 80.1% female. The vast majority of the participants were Latino (54.1%) or Black (34.6%). Just over half of the participants were in a household with an annual income lower than $20,000 while almost half had some college or a higher degree. In terms of employment, 37.5% of the sample was employed and 11.4% was self-employed). Just over one-third of the participants were married or living with a partner. Table 3 shows significant crude associations between NEWS scales and HR-QoL. The correlation coefficients of NEWS walkability enablers, NEWS walkability barriers, and NEWS neighborhood satisfaction with physical HR-QoL are 0.086, − 0.151 and 0.098 (p < 0.05). The correlation coefficients of NEWS walkability enablers, NEWS walkability barriers, and NEWS neighborhood satisfaction with mental HR-QoL are 0.180, − 0.143 and 0.218 (p < 0.05).

Table 2 Descriptive statistics
Full size table
Table 3 Spearman correlation of NEWS subscale and HR-QoL
Full size table

Regression analysis

Using separate GEE models with multiple imputations, mental and physical HR- QoL scores were regressed on NEWS subscales adjusting for covariates (Table 4). Results (with multiple imputations) showed mental HR-QoL was negatively associated with barriers of walkability (β ± SE = − 1.63 ± 0.55, p = 0.003) but positively associated with neighborhood satisfaction (β ± SE = 1.55 ± 0.66, p = 0.02), after adjusting for covariates. Physical HR-QoL was associated with only barriers of walkability (β ± SE = − 1.13 ± 0.57, p < 0.05). Sensitivity analyses showed the regression results were similar in imputed vs. non-imputed data (data not shown).

Table 4 GEE model results for mental and physical health-related quality of life (HR-QoL)
Full size table

Discussion

This is one of the first studies to examine HR-QoL in relation to perceived neighborhood environment. In particular, our study adds to the emerging literature with a specific focus on Latino and Black residents in lower income neighborhoods. We found that high perceived barriers of walkability was associated with both lower physical and mental HR-QoL. In addition, a global scale of neighborhood satisfaction was positively related to mental, but not physical, HR-QoL.

Despite the ubiquity of NEWS and SF-12 (or SF-36) in the literature, there have been surprisingly few studies investigating the relationship between these two sets of measures. In prior research, an Australian study on people aged 75 or greater showed a positive correlation between the original NEWS subscales and physical and mental HR-QoL using SF-36, but the study did not consider covariates [26]. Another study conducted in adults of different ages in 6 urban centers of China used multivariable models to examine each original NEWS subscale in relation to mental and physical HR-QoL (measured by SF-12) and found that higher perceived land use diversity, safety and esthetics were associated with higher physical and mental well-being [27]. Our study is the first to examine the association of the NEWS in relation to HR-QoL in the United States, especially among minority populations.

HR-QoL is an important public health outcome given the growing body of literature showing it to be an independent predictor of diverse clinical outcomes [46,47,48,49,50,51,52,53,54]. Physical HR-QoL as measured by SF-12 or SF-36 has been associated with the mortality of patients with hemodialysis [48], after coronary artery bypass graft surgery [54], and within 48 hours of admission to the ICU [52], as well as the development of obesity [55], diabetes [56], cardiovascular disease [57], and several cancers (e.g., oral [53] and advanced breast cancer [51]). In addition, patients with better mental HR-QoL scores were shown to be more likely to improve after lumbar fusion [49]. Mental HR-QoL has also been associated with mental health outcomes, such as anxiety [58], depression [59], and relapse of schizophrenia at 24-month follow up [47]. Therefore, HR-QoL can be considered a proxy of global well-being but more research is warranted on how to intervene on HR-QoL, including potentially via environmental strategies such as improving neighborhood walkability.

The items for the barriers for walkability subscale were a function of perceived traffic and crime. Our results are corroborated by prior research linking these environmental factors to health. For example, a previous study on children found that children who were exposed to high traffic volumes had significantly higher odds of asthma [60]. In addition, exposure to traffic congestion has been associated with on-the-job elevations of urinary catecholamines (a marker of stress) among bus drivers [61]. Elsewhere, the safer or less crime individuals feel in their neighborhood, the better mental health outcomes they have such as lower distress [62, 63] and better physical health [64]. Collectively, these and our findings point to the potential importance of transportation and neighborhood design in urban areas where the alleviation of traffic [65] and improvement in perceived safety [62, 66,67,68] may contribute to population well-being, which in turn may have a downstream impact on reducing health disparities.

It was notable that neighborhood satisfaction was significantly associated with mental but not physical HR-QoL. Neighborhood satisfaction has previously been found to be a predictor of mental health outcomes [69]. Neighborhood satisfaction has also been studied as a significant mediator between the quality of green space in a neighborhood and general health outcomes [18]. It is possible that mental HR-QoL serves as a mediator between neighborhood satisfaction and physical HR-QoL over time; future longitudinal studies are needed to further examine this hypothesis.

This study highlights a caveat for the generalizability of well published psychosocial scales such as NEWS without further psychometric testing. NEWS has been widely used in research globally [35, 70,71,72]. However, items in NEWS may be context-specific and may need to be customized to specific study populations [26, 70]. An important contribution of this paper is the application of EFA to reconstruct the factors, an approach that could be considered for future studies using surveys of perceptions of the environment.

This study and its findings add to the literature on urban livability. Urban livability is a multifaceted concept that incorporates diverse aspects of the neighborhood environment, including physical, biological and socioeconomic characteristics and their interactions [73, 74]. HR-QoL in this literature may be conceptualized, for instance, in terms of the number or density of health-related facilities and services [73]. Our study shows that the lived experience of community residents is an additional important dimension to consider. As such, this study has important research and policy implications that require the convergence of public health, urban planning and design, and other fields in a more holistic approach to urban livability.

Several limitations are inherent in this study. The study was cross-sectional; thus, no causality could be inferred. The target population included mostly residents of low-income, minority neighborhoods, limiting the generalizability of our findings to all of NYC or elsewhere. In addition, we cannot rule out the possibility of selection bias among those who chose to participate in the survey study, thus study findings may not be representative of the entirety of the underlying communities. The population focus, however, was also a strength of the study given the heightened health disparities experienced by Latino and Black communities in the United States.

In conclusion, perceptions of the built environment appear to be important factors in the HR-QoL of low-income residents in NYC. Further research is warranted to investigate the pathways by which such perceptions influence HR-QoL, including potential stress mechanisms. The current paper adds to the literature on urban health and urban planning and shows the potential value of incorporating community members’ experiences of the built environment and robust HR-QoL measures in studies of population well-being and environmental livability.