Does implementation of a low emission zone change perceptions of air quality, acceptability and travel behaviour? a longitudinal and repeated cross-sectional survey study in Bradford, UK

Patient and Public Involvement

A pollution research advisory group was formed including 12 members of the general public living within the CAZ boundary in Bradford. They met regularly throughout the duration of the study to provided advice on all aspects of the wider study (McEachan et al., 2022). For the current study, they gave specific feedback on the wording of participant information sheets and content of questionnaires, and have helped to interpret key findings.

Study design

This study was part of a wider evaluation of the heath and economic impacts of a Clean Air Zone (CAZ) which was implemented in the city of Bradford, UK, in September 2022 (McEachan et al., 2022). We conducted a population-based survey prior to, and 12 months after the launch of the CAZ. The first survey was conducted between April 2021 and December 2021 (previously reported in Mebrahtu et al., 2023). The second survey was conducted one year after the CAZ implementation (between July – November 2023) and included both participants who responded to the first survey, and additional participants responding to the second survey only. The study thus includes a longitudinal and repeated cross-sectional element.

Setting

Bradford is the fifth largest metropolitan district in England, with a population of over 546,000 (Office of National Statistics, 2022). Residents of the city are predominantly of white British (57%) and Pakistani origin (25%) (Bradford Metropolitan District Council (BMDC), 2022). A third of the population live in the most deprived Index of Multiple Deprivation (IMD) decile according to England averages (Bradford Metropolitan District Council (BMDC), 2019). The city has higher than average levels of respiratory morbidity (Mebrahtu, 2015; Mebrahtu et al., 2016; Mebrahtu et al., 2015). It is home to the representative longitudinal Born in Bradford (BiB) birth cohort, which recruited 12,453 mothers, with 13,776 pregnancies and 3,448 partners between 2007–2011; 42% of the BiB cohort are of Pakistani origin and 37% are of White British origin (McEachan et al., 2024).

In 2018, the UK government identified Bradford as exceeding legal limits of pollutants at several parts of the city and directed the local authority to develop and implement a charging Clean Air Zone as part of a wider plan to tackle pollution (McEachan et al., 2022). The resulting Bradford Clean Air Plan included a Class C CAZ, where non-compliant buses, coaches, heavy goods vehicles, vans, minibuses, taxis and private hire vehicles are charged a daily fee to enter the zone. Private vehicles are not charged (McEachan et al., 2022).

Procedure and participants

Details of baseline survey recruitment are reported in detail elsewhere (Mebrahtu et al., 2023). Eligible participants were either i) participants of the BiB cohort or ii) members of the general public. Our baseline sample included 1137 BiB participants and 812 members of the general public who completed the survey online, via telephone or via post. For the follow-up survey we recontacted all baseline BiB participants, and the general public participants who had provided contact details (N=340) via email (asking participants to complete an online survey) or telephone interview. These participants’ questionnaires were given a unique ID code which allowed baseline and follow-up surveys to be matched. To boost the sample at follow-up, we recruited additional members of the general public online via a survey link which was distributed widely via online mailing lists. We also publicised the link in person via pop-up stands in three local hospital sites, 9 general practice sites, and at 12 community events (e.g. park fun days). See Figure 1.

Figure 1. Survey recruitment flow chart.

Survey variables

Three categories of variables were collected: 1) attitudes towards the CAZ, 2) air quality perceptions, 3) travel behaviours, see Table 1. Demographic data were already available for the BiB families. We asked members of the general public to complete demographic questions; however, in order to maximise completion rates, these questions were not mandatory and were located at the end of the questionnaire.

Statistical analysis

First, sociodemographic characteristics of participants were tabulated and descriptively summarised. Individual items for perceptions of air quality, perceptions of the CAZ and travel mode behaviour were also summarised descriptively. Comparisons between the cross-sectional baseline and follow-up responses were made using chi-square tests and tests of proportions. We identified differences between patterns of responses for the two samples (BiB vs general public) in the baseline sample, and thus, for the repeated cross-sectional analysis, we report BiB participants and members of the general public separately.

To explore changes in attitudes towards the CAZ and perceptions of air quality (i.e., longitudinal analysis), response options were re-categorised to simplify interpretation of results (Table 1, final column). Multi-collinearity was first assessed by running a series of ordinary logistic least square regressions of each independent variable against all other variables. The mean variance inflation factor for all regressions was <2, indicating limited risk of multi-collinearity between the independent variables. Multinomial regressions, which are appropriate for multi-category outcomes, were used to explore whether changes in attitudes (reference: no change) were associated with demographic factors including ethnicity (white British[reference], Pakistani origin, other), sex (male[reference], female), Age (18–34[reference], 35–44, 45+), and Bradford IMD quintile (reference IMD 1 – most deprived) and whether respondents lived inside (reference) or outside the CAZ zone. Analyses were conducting using the mlogit command in Stata (v17 Statacorp 2023). The analyses were based on those who provided complete data (i.e., complete case analysis).

Survey population characteristics

A total of 1,949 (BiB families=1,137; general public=812) and 2,085 (BiB families=640; general public=1,445) individuals participated in the baseline and follow up surveys, respectively. Demographic data were available for most of the BiB families but were missing for a substantial proportion of the general public sample (Table 2). BiB participants were predominately female (90% baseline, 92% follow-up), aged 35–54 (88% baseline, 85% follow-up). At baseline, 44% were of white British origin and 38% were of Pakistani origin. At follow up, these figures were 49% and 32% respectively. Around a quarter lived within the CAZ boundary at baseline and follow-up. At baseline, 42% lived in the most deprived quintiles in Bradford, compared with 35% at follow-up.

Where demographic data were reported for the general public sample, around half were female (52% baseline, 41% follow-up), two-fifths were aged 18–44 (41% baseline and 35% follow-up) and just over one-fifth reported living in the CAZ boundary (22% at baseline and 23% follow-up). At follow-up there was an increase in the proportion of Pakistani (19%) and other ethnic origin participants (14%) and more reported being living in the most deprived quintiles of Bradford (31% vs. 23% at baseline). See Table 2.

Acceptability of the Clean Air Zone

Cross-sectional analysis: In our repeated cross-sectional data, acceptability of the CAZ at follow-up was lower for both BiB and general public samples than at baseline, see Figure 2. At follow-up, 64.2% of BiB participants, and 59.4% of general public participants said that they felt the CAZ was a good idea; (a drop of 4.9% (95% CI: -9.4 to -0.3%, and 10.8% (95% CI: -14.9 to -6.8%), respectively). Fewer participants reported that they felt the CAZ covered the right geographical areas (i.e., responded agree or strongly agree) with 28.6% of BiB participants (a reduction of 8.7%, 95% CI: -13.0 to -4.2%) vs 31.6% of general public participants agreeing or strongly agreeing (reduction of 8.0%, 95%CI: -12.1 to -3.8%).

Figure 2. Percentage of respondents reporting positive CAZ attitudes at baseline, and follow-up.

The numbers of respondents who agreed or strongly agreed that the CAZ would improve the health of their family or community also reduced. Agreement dropped by 13.4% for BiB participants (to 51.1%) and by 13% for general public participants (to 48%). See Table 3 and Figure 2.

Longitudinal analysis. We include the full descriptives for CAZ and air quality perceptions for the N=814 participants completed both surveys in extended data (see Table S1), These are not discussed further. Table 4 and Figure 3 show the percentage of people whose attitudes remained stable, or changed. The majority of participants attitudes towards the CAZ remained stable pre and post CAZ launch. Just over half (55.9%) of the longitudinal sample felt that the CAZ was a good idea, and two fifths (41%) agreed it would improve the health of their community at both baseline and follow-up (i.e. maintaining consistent attitudes).

Figure 3. Changes in attitudes towards the clean air zone from baseline to follow-up.

Multinomial regressions comparing those whose attitudes had changed from baseline found few consistent relationships with socio-demographic variables, full results can be found in the extended data (see Table S2). Significant relationships are highlighted in Figure 4.

Figure 4. Summary of significant associations from multinomial regressions exploring changes in attitudes.

Positive change in support for the CAZ: Those living outside the CAZ boundary were more likely to have a positive change in their support of the CAZ than those who lived inside the boundary (coefficient=0.71; 95% CI: 0.10 to 1.32). Participants from less deprived areas (quintile 3: coefficient=-0.82; 95% CI:-1.56 to -0.07; quintile 4 (coefficient=-0.87; 95% CI: -1.67 to -0.08) were less likely to be in this group compared with the most deprived participants.

Negative change in support for the CAZ: Compared to white British participants, the other ethnic minority participants were less likely to be in this group (coefficient=-0.79; 95% CI: -1.53 to -0.04). Those living in the in the least deprived areas were less likely to become negative about the CAZ than those in the most deprived category (coefficient=-1.25; 95% CI: -2.51 to -0.01).

Decrease in agreement, or becoming unsure that CAZ will improve health: White British participants were more likely to be in the group that reported a decrease in agreement that the CAZ would improve health compared with Pakistani participants (coefficient: -0 .77, 95%CI: -1.51 to -0.04), or be in the group that reported becoming neutral compared with other ethnic minority participants (coefficient: -0.76, 95%CI: (-1.27 to -0.25).

Perceptions of air quality

Cross-sectional analysis: Perceptions of air quality appeared relatively similar between baseline and follow-up for BiB participants but varied more for members of the general public (Table 3). For BiB participants, at follow-up, just over a third of the sample (37%) felt air quality was ‘poor’ or ‘very poor’, and 38% felt it was the same as other areas in the UK. Thirty-four percent were moderately to extremely concerned about air quality, and 18% thought other people in Bradford were moderately to extremely concerned. There was a drop of around 10% (95% CI: -15 to -0.05%) in the numbers reporting that it was ‘extremely important’ to improve air quality.

For the general public sample, there were significant differences in the distribution of responses to most air quality items. At follow-up, just over a third (36.5%) reported air quality to be ‘poor’ or ‘very poor’; a drop of 24.7% (95% CI: -28.8 to -20.5 %). The proportions of people that reported air quality in Bradford was worse than other areas (46.9%), those that reported fewer people were ‘extremely concerned’ (32%) and those that felt it was ‘extremely’ important to improve air quality (70%) also fell by 20.6% (95% CI:16.5 to 24.7%), 14.3% (95% CI: -18.0 to -10.5 %) and 13.8% (95% CI: -17.8 to -9.7 %)) respectively (see Table 3). Perceptions of ‘others’ concern was more similar, with around 20% reporting that others were moderately or extremely concerned, with no significant change from baseline.

Longitudinal analysis: Attitudes towards air quality remained mostly stable pre and post CAZ launch. (Figure 5, Table 4). Where changes did occur, roughly equal numbers changed to a more positive standpoint (e.g. thinking air quality is better: 14%, compares better to other areas of the UK: 13.9%, and less concerned (e.g. reporting less concern: 14.0%, reporting others concern reduced 19.2%) compared with a more negative one (thinking air quality is worse: 13.6% and compares worse with other areas: 16.2%) or more concerned standpoint (14% and others more concerned: 12.5%).

Figure 5. Changes in attitudes towards air quality from baseline to follow-up.

Multinomial regressions identified no relationships between any demographic factors and those whose attitudes or concern towards air quality had worsened (e.g. more likely to rate it worse, and compare it negatively to other areas of the UK) and those whose concern had either increased or reduced (see Extended data). However, there were some significant associations as follows, as outlined in Figure 4.

Positive change in attitudes towards air quality: Those in the ‘other’ ethnic origin category were more likely to have positive change of attitude about Bradford air quality than white British ethnic origin participants (coefficient=0.78; 95% CI:0.26 to 1.31).

Positive change in how Bradford compares to the rest of the UK: Compared to the most deprived group, those in IMD quintile 2 were more likely to report change comparing Bradford favourably to the rest of the UK compared to those in the most deprived quintile (co-efficient=0.83, 95%CI - 0.03 to 1.62).

Increase in perceptions of others concern about air quality: Those in the oldest age group (45+) were more likely to be in the group whose perceptions of others concern about air quality increased compared with those in the younger age group (18–34, coefficient=1.32, 95%CI 0.22-2.22).

Importance of improving air quality increased: There were some differences by IMD quintile. Compared with the most deprived areas, those in more affluent areas were less likely to report increased importance of tackling pollution (compared with quintile 1: quintile 3 coefficient= -0.87, 95%CI -1.64 to -0.10; quintile 5 coefficient=-1.11; 95% CI: -2.06 to -0.17).

Importance of improving air quality decreased. Those 45+ years old were less likely to say that importance of improving air quality decreased (coefficient=-1.13; 95% CI: -2.00 to -0.26) compared with those 18–34 age group.

Travel behaviours

Private transport was the dominant mode of travel for getting to work and travelling for leisure for both samples. There were slight increases in the numbers reporting private transport for work in our repeated cross-sectional sample (64.6% baseline to 70.5% at follow up for BiB participants, and from 47.5% baseline to 57.5% follow up for general public participants). There was a slight decrease in the numbers of BiB participants reporting private transport for leisure (88.1% at baseline and 79.1% follow up) and a minor increase for the general public (64.4% baseline and 66.2% follow up) participants.

More baseline BiB families than the follow up BiB families participants used active travel (walked or cycled) to take children to school (55.6% baseline vs 49.4% follow up) whilst only a quarter of the general public used active travel during the same period (20.4% baseline vs 22.1% follow up).

Private transport was the most used mode of transport for leisure by the BiB families The proportion of using a car for a 10 minutes’ walk by the BiB families (7% baseline vs 6% follow up) and general public (4% baseline vs 8% follow up) remained small in both surveys (Table 3).