Evaluation of age-friendly guidelines for public buses

Transportation Research Part A 53 (2013) 68–80

Contents lists available at SciVerse ScienceDirect

Transportation Research Part A journal homepage: www.elsevier.com/locate/tra

Evaluation of age-friendly guidelines for public buses Kieran Broome a,b,⇑, Linda Worrall b, Jennifer Fleming b,c, Duncan Boldy d a

School of Health & Sport Sciences, University of the Sunshine Coast, Maroochydore DC, Queensland 4558, Australia School of Health & Rehabilitation Sciences, The University of Queensland, Brisbane, Queensland 4072, Australia Princess Alexandra Hospital, Brisbane, Queensland 4072, Australia d Centre for Research on Ageing, Curtin Health Innovation Research Institute, Curtin University of Technology, GPO Box U1987, Perth, Western Australia 6845, Australia b c

a r t i c l e

i n f o

Article history: Received 16 July 2012 Received in revised form 26 April 2013 Accepted 30 May 2013

Keywords: Buses Age-friendly Elderly Evaluation Useability

a b s t r a c t Background: Older people commonly experience difficulty using public transport, and identify a range of barriers and facilitators to bus use. Purpose: This study aimed to evaluate the impact of implementing age-friendly guidelines for public buses on bus use, useability and social participation for older people. Methodology: Age-friendly changes were implemented in Hervey Bay, Queensland, over a 3 year period. Changes included more lower floor buses, age-awareness training for bus drivers, more frequent scheduling, improved pedestrian infrastructure, a bus buddy program and trial of a flexible route bus. Brisbane, Queensland, served as a control site. This study adopted two complementary methodologies for evaluating the impact of agefriendly changes. The first study incorporated on-board satisfaction surveys with 335 bus users. The second involved a repeated measures cohort study (n = 100) with bus users and non-users. Results: Implementation of age-friendly changes in Hervey Bay, Queensland, resulted in improved satisfaction with, and perceived useability of, the bus system. Participants in Hervey Bay maintained their bus use and social activity participation, while participants in Brisbane experienced a decline in ease of bus use and social activity participation over the intervention period. Conclusion: The implementation of age-friendly guidelines for public buses may improve the useability of bus systems for older people and result in maintenance of participation in social activities. Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction The World Health Organization encourages the provision of age-friendly transport as a core component of age-friendly cities (World Health Organization, 2007a). The useability of bus systems is associated with frequency of bus use, participation in community activities and quality of life for older people (Broome et al., 2009; Dent et al., 1999; Gabriel and Bowling, 2004; Stahl, 1987). Older people typically encounter a range of barriers when using public bus systems (Broome et al., 2010; World Health Organization, 2007b). By overcoming these barriers, older people may experience greater opportunity to access the community, remain independent and retain quality of life. Age-friendly guidelines provide a set of principles and focus areas for improving community accessibility for older people. The development of age-friendly guidelines requires a ⇑ Corresponding author. Address: School of Health & Sport Sciences, Faculty of Science, Health and Education, University of the Sunshine Coast, Maroochydore DC, Queensland 4558, Australia. Tel.: +61 7 5456 5163; fax: +61 7 5456 5010. E-mail addresses: [email protected] (K. Broome), [email protected] (L. Worrall), j.fl[email protected] (J. Fleming), [email protected] (D. Boldy). 0965-8564/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.tra.2013.05.003

K. Broome et al. / Transportation Research Part A 53 (2013) 68–80

69

bottom-up participatory approach, where older people are consulted on their perspectives, experiences and needs (World Health Organization, 2007b). A number of core priorities have been suggested for the provision of age-friendly bus systems. Seven key priorities include physically accessible vehicles, friendly and helpful bus drivers, appropriate timetabling and scheduling, bus stops located close to homes and destinations, accessible pedestrian infrastructure, adequate training and information on the use of buses, and routes and destinations matched to the needs of older people (Broome et al., 2010). A number of interventions for each priority have been suggested across the literature that could potentially improve the useability of public bus systems for older people. In order to achieve these priorities, transport providers and policy makers should be guided by evidence that supports each of these interventions. Some interventions, such as the introduction of lower floor buses, age-awareness training for bus drivers, increased frequency of buses, flexible route bus systems, video training for older people and concession fares have initial evidence supporting their use in age-friendly bus systems (Broome et al., 2011, 2012; Balcombe et al., 1998; Shaheen and Rodier, 2006; Suen and Sen, 1999; White et al., 1992). Further research is required to explore and compare the efficacy of other interventions, including providing age-friendly information and including the needs and preference of older people in route design. While some interventions have been examined individually, no research has examined the overall impact of implementing age-friendly guidelines for public buses. Older people may experience multiple barriers to catching the bus, and therefore by implementing multiple changes simultaneously they may be able to access public bus systems with greater ease. The success of replacing a fixed route service with a flexible route service (an increase in patronage by older people of 89%) is a case in point (Broome et al., 2012). The introduction of the flexible route service addressed several of the priorities at once. While the sample site used in the study had relatively high satisfaction scores for bus driver friendliness and helpfulness and concession fares prior to the study, the introduction of a flexible route service additionally met the needs for close bus stops and accessible pedestrian infrastructure (as passengers no longer had to walk to the bus stop), as well as providing greater variability in destinations. The increase in patronage attributed to the flexible route introduction was accompanied by very high scores in satisfaction across most aspects of the bus system. Single interventions, such as the introduction of low floor buses (an increase in general patronage by 10%) (Suen and Sen, 1999) typically lead to smaller improvements. This study aimed to evaluate the effectiveness of implementing age-friendly guidelines for public buses. Evaluating the effectiveness of age-friendly interventions is a multifaceted phenomenon. Ease of use, satisfaction with buses and frequency of bus use are potential features of age-friendliness as applied to bus use and have been used as primary outcome measures (Lui et al., 2009). Each of these variables can be measured in multiple ways, using varying methodologies. Additionally, the aim of establishing age-friendly bus systems is not to increase bus use for the sake of bus use, but to allow older people greater opportunity to access the community. Therefore, the frequency of participation in meaningful community activities should also be considered a measure of the success of age-friendly interventions (Lui et al., 2009). The aims of this study were to test the hypotheses that the implementation of age-friendly guidelines would result in (1) improved satisfaction of older bus users with the buses; (2) an increase in the subjective ease of bus use for both older bus users and older non-users; (3) an increase in the frequency of bus use for older people; (4) increased frequency of community participation for older people; and, (5) a change in the barriers and facilitators to bus use reported by older people. 2. Methodology This study aims to evaluate the effectiveness of implementing age-friendly guidelines for public buses using the case study of Hervey Bay, Queensland, Australia. Whilst a stringent control site was not feasible, due to the multiple variables in the study, the experience of north Brisbane, Queensland, Australia, where age-friendly guidelines were not explicitly implemented, was used as a ‘‘usual practice’’ control. This pre-post intervention study with a usual practice control utilised two methodologies to establish the efficacy of agefriendly guidelines. The first methodology involved satisfaction surveys conducted on the bus and at bus interchanges with bus users immediately before and after the intervention period. Participants were not matched comparisons, but were all bus users. The second methodology followed the same cohort of older people before and after the intervention period, and examined changes in frequency and perceived ease of bus use, as well as reported participation in social activities. These two methodologies will subsequently be referred to as study 1 and study 2 respectively. A third planned methodology (monitoring senior ticket sale data) could not be employed, due to a change in ticketing system at one site during the intervention period. The study received ethical approval from The University of Queensland Behavioural and Social Sciences Ethical Review Committee. 2.1. The intervention The intervention period took place over 2 years, during which the researchers collaborated with the transport provider (Wide Bay Transit), policy maker (Queensland Transport) and local government council in Hervey Bay to assist and monitor the implementation of age-friendly guidelines. In Brisbane, whilst age-friendly guidelines were not explicitly implemented, some changes that could be considered age-friendly took place as part of the usual quality improvement practice of transport

70

K. Broome et al. / Transportation Research Part A 53 (2013) 68–80

Table 1 Age-friendly changes in bus systems during the period of study. Priority

Changes in Hervey Bay

Changes in Brisbane

1. Low floor buses

Increase in percentage of lower floor buses

2. Friendly and helpful bus drivers 3. Timetables and scheduling

Age-awareness training for all bus drivers Increased number of buses on all routes

4. Bus stops and shelters available and close to home and destinations 5. Accessible pedestrian infrastructure

Replacement of one fixed route with a flexible route service No improvement in the quality of bus stops and shelters themselves Council plan in place for pedestrian infrastructure upgrade in line with agefriendly principles Bus buddy program introduced (but minimal uptake) Replacement of one fixed route with a flexible route service

Increase in percentage of lower floor buses Nil Increased number of buses on some routes Nil

6. Training and information 7. Routes and destinations

Nil Nil Nil

providers. Table 1 illustrates changes that took place during the intervention period at both sites. Improvements in the physical accessibility of buses at both sites were associated with the requirements of the Disability Standards for Accessible Public Transport 2002 (Williams, 2002). 2.2. Study 1: Pre- and post-intervention on-board satisfaction survey 2.2.1. Sampling Volunteer sampling was used to select participants. The Hervey Bay sample area included bus services in the Hervey Bay region provided by the local bus company, Wide Bay Transit. Hervey Bay is a regional coastal town approximately 300 km north of Brisbane. The total population was 48,157 (Australian Bureau of Statistics, 2007a). Twenty-one percent were aged 65 and over. The population density was 121.9 persons/km2. Other public transport services included taxis. The Brisbane sample area included Brisbane Transport bus services in the north Brisbane region that serviced the Chermside and Toombul interchanges or serviced the region between Gympie Road and Sandgate Road, bounded by the Gateway Motorway in the north. The total population of Brisbane (a large city) was 1,763,131 (11.2% aged 65 and over), with a population density of 296.3 persons/km2 (Australian Bureau of Statistics, 2007b). Other public transport services included taxis and trains. In both Hervey Bay and Brisbane, the majority of the population were Australian-born and English-speaking, and overall transport oversight was governed by the state department of transportation. Although all bus users encountered were invited to participate in the survey, only participants who reported their age as 60 years and over were included in this analysis. Approximately half of all people surveyed were aged 60 years and over. Given the volunteer sampling used at both time periods, basic demographic information was obtained to ascertain whether the two samples were similar. 2.2.2. Outcome measures The satisfaction survey was a brief 5-min survey collecting data on basic demographics, driving, bus use and satisfaction with various aspects of the bus system. Demographics included age (60 and over or 60 and under), gender, sample site (Hervey Bay or Brisbane) and presence of a disability or medical condition that makes it difficult to catch a bus. Participants were asked how frequently they used the bus (3+ days per week, 1–2 days per week, fortnightly, monthly or less than monthly) as well as whether they drove a car or lived with someone who drove. Participants were asked to rate their satisfaction with the bus system using a 5-pt Likert scale from ‘‘very satisfied’’ to ‘‘very dissatisfied’’. The 5-pt Likert scale was preferred, as it was sensitive to change without being onerous on participants. They were asked to rate their overall satisfaction, as well as satisfaction with eleven aspects of the bus system including finding information about buses, understanding information, when and where buses go, ticket prices and types, signage of the route on buses, moving on, off and around the bus, driver friendliness, driver helpfulness, distance to and from the bus stop, the bus stops and shelters and their own level of knowledge about using buses. The specific criteria were drawn from previous research on age-friendly bus system priorities (Broome et al., 2010) to promote validity. Respondents’ comments regarding satisfaction with the bus system were also written on the form by the researcher. After the intervention, participants were also asked to rate whether buses were the same, better or worse compared with 2 years ago. 2.2.3. Procedure Researchers conducted surveys for approximately 4 days at both sites before and after the intervention period. A standardised script, approved by the ethics committee, provided information on the study and was read to potential participants on first contact. Verbal consent was sought prior to collecting data. 2.2.4. Data analysis Exploratory statistical analysis was conducted using the Statistical Package for the Social Sciences (SPSS) Version 11. In the initial exploratory analysis, demographic data of the two samples (before and after) were compared using descriptive statistics and Chi-square tests. To address hypotheses one and three, changes in satisfaction and transport use were analysed

K. Broome et al. / Transportation Research Part A 53 (2013) 68–80

71

using non-parametric inferential statistics (Chi-square and Mann–Whitney U tests). To address hypothesis five, an analysis using the methodology of Koffman and Salstrom (2001), where the satisfaction with each criteria is subtracted from overall satisfaction to elucidate their relative impact, was conducted to establish which aspects of the bus system at each time point had the greatest impact on overall satisfaction scores. In addition to the exploratory analysis and the Koffman and Salstrom methodology, a secondary analysis of the data, using Stata Version 10, was conducted using multinomial logistic regression to devise a model to explain the data. The multinomial logistic regression also allowed the effect of confounding variables within the data to be taken into account. The outcome variable was location and time period (Hervey Bay pre-intervention, Hervey Bay post-intervention, Brisbane preintervention, Brisbane post-intervention). Explanatory variables included gender, driving status, disability status, bus satisfaction and frequency of use. The model addressed hypotheses one and three. For the purpose of the multinomial logistic regression, demographic categories were converted to binomial categories (e.g., current driver, retired driver and never driven were collapsed to driver and non-driver), overall satisfaction was collapsed to three categories (aggregating neutral to very dissatisfied due to a low number of responses in these categories) and frequency of bus use was collapsed to three categories (aggregating fortnightly to less than monthly due to a low number of responses in these categories). The multinomial logistic regression addressed hypotheses one and three. 2.3. Study 2: Repeated measures cohort study 2.3.1. Sampling The sample was a volunteer sample of older people (aged 60 and over). Eligibility criteria included living in the community (i.e., not in hospital or residential facilities) and sufficient cognitive and language ability to complete the questionnaire. The two sample sites were Hervey Bay and north Brisbane. Further details regarding the sampling procedures can be found elsewhere (Broome et al., 2010). 2.3.2. Outcome measures Outcome measures included a demographic questionnaire, questions relating to transport use and completion of the Social Activities Checklist (SOCACT) (Cruice et al., 2005). Demographic information collected prior to the intervention period included age, gender, sample site (Hervey Bay or Brisbane) and length of residence in the area (5 or more years, 1–5 year, or less then 1 year). Following the intervention period, participants were asked whether they were living in the same suburb. Transport variables included frequency of bus use (frequently, occasional or rarely/never), self-rated ease of bus use (measured on a 10 cm analogue scale with higher scores indicating greater ease), driving status (current driver, retired driver or never driven) and reliance on another driver. After the intervention, participants were also asked to rate whether buses were the same, better or worse, and whether they were using the buses more, the same or less frequently, compared with 3 years ago. Participants were also asked to list the three greatest barriers to bus use after the intervention period, to compare with nominal group technique data collected from a previous study (Broome et al., 2010). Social activity participation was measured using the SOCACT. The SOCACT is a measure of frequency and satisfaction with social activity participation based on a range of gerontology, stroke and mental health questionnaires (Cruice et al., 2005; Worrall and Hickson, 2003). Items include frequency of participation in twenty social and societal activities including leisure, informal and formal activities. A final question explores whether the person is satisfied with their activity level. 2.3.3. Procedure Approximately 6 months prior to the intervention period, the volunteer cohort of older people from Hervey Bay and north Brisbane were involved in a research project investigating the barriers and facilitators to bus use (Broome et al., 2010). At that time, the participants received information on the study and completed participant consent forms, demographic, transport use and social activity questionnaires, and the majority (60.4%) agreed to take part in a follow-up questionnaire after the intervention period. These questionnaires were completed individually, prior to being involved in a nominal group technique discussion. Following the intervention period, the participants were contacted via mail and invited to complete the followup questionnaire to explore changes in bus use and social activity participation across the 3 years (including the 2 year intervention period). The majority returned the questionnaires by post, with a small number electing to complete a telephone survey. 2.3.4. Data analysis In the initial analysis, descriptive and non-parametric inferential statistics (Chi-square, Mann–Whitney U tests, Kruskal Wallis tests, Spearman rank correlations) were used to describe the sample and associations between variables before and after the intervention period. Descriptive statistics were used to address hypotheses one and three. To address hypotheses two, three and four, changes in transport use and social activity participation over the intervention period were analysed using non-parametric inferential statistics (Chi-square and Wilcoxin signed-ranks tests). The principles of the nominal group technique (Broome et al., 2010) were used to analyse the three greatest barriers to using the bus after the intervention in order to address hypothesis five. Exploratory statistical analysis was conducted using the Statistical Package for the Social Sciences (SPSS) Version 11. A secondary analysis, using Stata Version 10, was conducted using multinomial logistic regression with clustered effects to devise a model to explain the data, and to help to control for the effects of

72

K. Broome et al. / Transportation Research Part A 53 (2013) 68–80

confounding variables within the data. The outcome variable was location and time period (Hervey Bay pre-intervention, Hervey Bay post-intervention, Brisbane pre-intervention, Brisbane post-intervention). Explanatory variables included age, gender, length of time living in suburb, driving status, bus ease and frequency of use, and satisfaction with overall social activity levels. The model addressed hypotheses two through four. For the purpose of the multinomial logistic regression, demographic categories were converted to binomial categories. Driving status was recoded from current driver, retired driver and never driven to driver and non-driver. Length of residence was collapsed to under 5 years and 5 years and over. Age was collapsed to three categories (60–69 years, 70–79 years, and 80 years and older). Ease of bus use was collapsed to three categories (mild to no difficulty catching the bus, moderate difficulty catching the bus, and severe difficulty catching the bus). 3. Results 3.1. Demographics 3.1.1. Study 1: Satisfaction survey Demographic characteristics of the four groups (Hervey Bay pre-intervention, Hervey Bay post-intervention, Brisbane pre-intervention and Brisbane post-intervention) were compared. Prior to the intervention period, the demographic profiles of the two samples were similar, with no significant difference for any of the variables between Hervey Bay and Brisbane. The demographic characteristics of the samples are shown below in Table 2. After the intervention, the samples in Hervey Bay and Brisbane differed significantly in the proportion of participants who currently drove (X2 = 5.372, p = 0.020) with a smaller proportion of current drivers in the city of Brisbane. After intervention, the proportion of participants who experienced a disability that made it difficult to catch a bus was significantly lower in Hervey Bay than Brisbane post-intervention (X2 = 8.354, p = 0.004) or Hervey Bay pre-intervention (X2 = 5.488, p = 0.019). The proportion of survey participants who were male or lived with someone who drove was similar for all samples. 3.1.2. Study 2: Cohort study Of the original 227 participants in the nominal group technique, 100 completed the follow-up survey. These 100 participants are the subject of the subsequent analysis. Those who took part in the follow-up survey were of similar age, gender, length of residence in the area, driving status, reliance on other drivers, frequency of bus use and social activity participation compared with those who did not take part. There were 48 participants from Hervey Bay and 52 participants from Brisbane. There was a mean age of 72.4 years (SD = 7.4 years) at baseline with no significant difference between the two sites. Participants from Hervey Bay and Brisbane had similar gender (22.9% male and 19.2% male respectively), length of residence (72.3% over 5 years and 76.9% respectively), driving status (60.4% current drivers, 20.8% retired drivers and 18.8% never driven, and 55.8% current drivers, 21.2% retired drivers and 23.1% never driven respectively) and reliance on other drivers (45.7% relying on another driver in Hervey Bay and 42% in Brisbane). Over the intervention period 11 participants who were current drivers retired from driving. Of these, 6 were from Hervey Bay and 5 were from Brisbane. Five participants lived in a different suburb at follow-up (2 from Hervey Bay and 3 from Brisbane). 3.2. Hypothesis one; Hervey Bay post-intervention would experience improved satisfaction of older bus users with the buses This hypothesis was addressed only by the satisfaction survey. Prior to the intervention, there was a non-significant trend towards a difference in overall satisfaction with the bus system between Hervey Bay and Brisbane (Z = 1.748, p = 0.080), with Hervey Bay participants more satisfied. This trend was statistically significant for all individual aspects of the bus system (excluding when and where buses go, and ticket prices and types) with Hervey Bay participants more satisfied in all instances (2.082 > Z < 6.716, 0.037 > p). The satisfaction scores from the samples are shown below in Table 3. Following the intervention there was a significant improvement in overall satisfaction with the bus system in Hervey Bay (Z = 2.830, p = 0.005). This addressed hypothesis one. The only change in satisfaction with specific aspects of the system was that satisfaction with signage of the route on buses decreased at follow-up (Z = 2.657, p = 0.008).

Table 2 Demographic characteristics of satisfaction survey participants. Variable

Total participant numbers Gender (male) Current driver Lives with someone who drives Presence of a disability that makes it difficult to catch a bus

Before intervention N (%)

After intervention N (%)

Hervey Bay

Hervey Bay

Brisbane

97 29 (29.6) 28 (28.9) 21 (21.9) 6 (6.1)

24 12 20 17

26 22 17 16

95 (27.4) (23.2) (17.9) (16.8)

Brisbane 16 17 12 12

60 (26.7) (28.3) (20.0) (20.0)

83 (28.9) (14.5) (24.1) (12.7)

73

K. Broome et al. / Transportation Research Part A 53 (2013) 68–80 Table 3 Satisfaction with the bus system. Satisfaction

Overall Finding information about buses Understanding information When and where buses go Ticket prices and types Signage of the routes on buses Moving on, off and around the bus Driver friendliness Driver helpfulness Distance to and from the bus stop The bus stops and shelters Own level of knowledge about using buses

Before intervention mean (SD)

After intervention mean (SD)

Hervey Bay

Brisbane

Hervey Bay

Brisbane

1.41 1.27 1.34 1.83 1.21 1.30 1.37 1.03 1.03 1.29 1.94 1.30

1.52 1.68 1.67 1.98 1.32 1.83 1.72 1.52 1.48 1.75 2.30 1.43

1.17 1.21 1.27 1.60 1.25 1.63 1.36 1.07 1.07 1.30 1.84 1.40

1.86 1.69 1.32 1.85 1.22 1.80 1.43 1.58 1.58 1.31 2.38 1.42

(0.679) (0.721) (0.696) (1.170) (0.670) (0.774) (0.790) (0.177) (0.177) (0.785) (1.480) (0.669)

(0.566) (0.892) (0.877) (1.112) (0.567) (0.942) (1.059) (0.567) (0.567) (0.950) (1.124) (0.594)

(0.427) (0.614) (0.688) (0.978) (0.711) (1.018) (0.835) (0.261) (0.261) (0.742) (1.199) (0.761)

(0.891) (1.147) (0.788) (1.119) (0.592) (1.289) (0.805) (0.849) (0.849) (0.752) (1.374) (0.835)

Conversely, in Brisbane following the intervention period there was a significant decline in overall satisfaction with the bus system (Z = 2.435, p = 0.015). With regards to specific aspects of the bus system, there was an improvement in satisfaction with understanding information (Z = 2.948, p = 0.003) and distance to the bus stop (Z = 3.815, p < 0.001), and no change with other aspects of the bus system. After the intervention, participants were asked to rate whether buses were the same, better or worse compared with 2 years ago. This provided additional information to address hypothesis one. There was a significant difference in response between Hervey Bay and Brisbane participants (X2 = 39.236, p < 0.001). In Hervey Bay 67.4% of participants reported that the buses were better, 13.7% the same and no participants rated the buses as worse. Almost a fifth of participants (18.9%) were recorded as ‘‘not applicable’’. Nearly all of the ‘‘not applicable’’ respondents were new users who were not catching the bus 2 years ago. In Brisbane, 38.6% of participants reported that buses were better, 50.6% the same and 6.0% worse. Again, 4.8% were recorded as not applicable as they were new users of the bus system since the beginning of the intervention period. The direction of the change for overall satisfaction and satisfaction with specific aspects at each site were different, and appears contradictory. Overall satisfaction is likely to be a more valid measure of satisfaction with the bus system, as changes in overall satisfaction measured at the two-time points is consistent with self-ratings of improvement in the bus system over the past 3 years. The methods of Koffman and Salstrom (2001) were used to establish which aspects of the bus system had the greatest impact on the bus system, both before and after the intervention and in both Hervey Bay and Brisbane. This addressed hypothesis five. The relative impact is shown in Table 4, where positive scores indicate a positive impact and negative scores a negative impact. Both prior to and following the intervention, driver friendliness and driver helpfulness had the greatest positive impact on the overall satisfaction with the bus system in Hervey Bay. The third greatest positive impact changed from ticket prices and types before the intervention, to finding information after the intervention. In Brisbane, the positive impact on overall satisfaction was mostly attributable to ticket prices and types, own knowledge about using buses and driver helpfulness prior to the intervention, and ticket prices and types, distance to the bus stop and understanding information following intervention. The greatest negative impact on the bus system both before and after intervention, and in both Hervey Bay and Brisbane, was attributable to when and where the buses go, and bus stops and shelters. Before the intervention the third most negative aspect in Hervey Bay was moving on, off and around the bus and in Brisbane was signage of the routes on buses. After inter-

Table 4 Relative impact of each aspect of the bus system on overall satisfaction.

Finding information about buses Understanding information When and where buses go Ticket prices and types Signage of the routes on buses Moving on, off and around the bus Driver friendliness Driver helpfulness Distance to and from the bus stop The bus stops and shelters Own level of knowledge about using buses

Before intervention mean (SD)

After intervention mean (SD)

Hervey Bay

Brisbane

Hervey Bay

0.149 0.074 0.415 0.202 0.117 0.043 0.383 0.383 0.128 0.521 0.117

0.150 (0.709) 0.133 (0.724) 0.450 (0.872) 0.217 (0.585) 0.300 (0.979) 0.183 (1.172) 0.017 (0.725) 0.050 (0.746) 0.217 (0.904) 0.767 (1.170) 0.1 (0.543)

0.041 0.102 0.449 0.041 0.408 0.153 0.092 0.092 0.135 0.663 0.224

(0.816) (0.779) (0.955) (0.798) (0.828) (0.938) (0.674) (0.674) (0.833) (1.381) (0.866)

(0.641) (0.681) (0.801) (0.785) (1.092) (0.817) (0.478) (0.433) (0.609) (1.175) (0.711)

Brisbane 0.241 0.530 0.012 0.699 0.036 0.422 0.349 0.349 0.556 0.482 0.482

(1.551) (1.086) (1.215) (1.101) (1.357) (1.083) (1.098) (1.098) (1.173) (1.525) (1.203)

74

K. Broome et al. / Transportation Research Part A 53 (2013) 68–80

vention the third negative aspect in Hervey Bay was signage of the routes on buses and in Brisbane was again signage of the routes on buses. The relative impact of various aspects of the bus system on overall satisfaction highlighted that when and where the buses go, as well as the bus stop and shelters, were the greatest negative influence both before and after intervention. Neither site included modifications to bus stops or shelters as part of the intervention, therefore this stability is expected. It is interesting to note that bus stops and shelters themselves did not attract as high a priority in either nominal group technique or focussed ethnography research (Broome et al., 2010). Therefore, in setting age-friendly priorities, satisfaction surveys may be a valuable supplement to these other techniques. With regards to when and where buses go, both sites experienced increases in the frequency of buses during the intervention period. In Hervey Bay especially, the changes that took place fit well with the findings of the original nominal group technique research, namely requiring more frequent buses during the day as well as during the evening and weekends. Pilot research suggests that, to meet the preferences of 90% of the population, buses should be scheduled at least every 15 min at all times. Following the intervention period very few services in Brisbane and no services in Hervey Bay met that requirement. This could explain why, despite increases in frequency, the frequency of buses does not yet meet the preferences of older people. It was also noted that the signage of routes on buses remained as one of the most negative factors at follow-up for Brisbane, and was rated by Hervey Bay participants even more poorly at follow-up than prior to intervention. It should be noted that, from comments collected, the newer lower floor buses that have been introduced for accessibility do not have appropriate signage. Participants noted that signage should be low glare (especially in direct sunlight), should have white text on a black background (rather than yellow on black) and should have signage on the front, back and sides of the bus. Lower floor bus designs in Brisbane and Hervey Bay do not routinely meet these requirements, and thus the true ‘‘accessibility’’ of these new buses is limited. It should by noted, however, the comments from the majority of participants were positive regarding the new lower floor buses, resulting in increased ease getting on and off the bus. In order to control for confounding variables, multinomial logistic regression was used to model explanatory factors for the differences between the four samples; Hervey Bay pre-intervention, Hervey Bay post-intervention, Brisbane pre-intervention and Brisbane post-intervention. The model is shown below in Table 5. Hervey Bay pre-intervention was used as the reference category.

Table 5 Multinomial logistic regression of the satisfaction survey. B

SE

95% Confidence interval for Exp(B) Lower bound

Odds ratio

Upper bound

0.014a 0.517 0.706 0.707 0.053 0.013a 0.003b 0.555 0.239

0.639 0.401 0.538 0.131 0.094 0.066 0.351 0.077

1.248 0.863 1.158 0.364 0.268 0.195 1.570 0.382

2.439 1.856 2.493 1.013 0.761 0.576 7.009 1.896

0.922 0.370 0.403 0.413 0.406 0.575 0.575 0.817 0.823

0.583 0.591 0.863 0.95 0.383 0.076 0.409 0.818 0.483

0.397 0.424 0.457 0.643 0.117 0.201 0.167 0.355

0.820 0.933 1.026 1.425 0.361 0.622 0.828 1.781

1.693 2.054 2.304 3.158 1.113 1.921 4.111 8.938

0.807 0.357 0.445 0.398 0.412 0.609 0.627 0.644 0.640

0.067 0.913 0.063 0.175 0.929 0.576 0.815 0.002b 0.177

0.517 0.183 0.787 0.430 0.216 0.252 0.038 0.120

1.039 0.437 1.716 0.964 0.711 0.864 0.134 0.421

2.091 1.045 3.743 2.162 2.346 2.954 0.473 1.477

Hervey Bay post-intervention vs. Hervey Bay pre-intervention Intercept Male Current driver Reliance on another driver Presence of a disability that makes it difficult to catch a bus Bus frequency 3+ times per week Bus frequency 1–2 times per week Very satisfied overall Fairly satisfied overall

1.148 0.222 0.147 0.147 1.009 1.318 1.633 0.451 0.963

0.869 0.342 0.390 0.391 0.522 0.533 0.552 0.763 0.818

Brisbane pre-intervention vs. Hervey Bay pre-intervention Intercept Gender Driving status Reliance on another driver Presence of a disability that makes it difficult to catch a bus Bus frequency 3+ times per week Bus frequency 1–2 times per week Very satisfied overall Fairly satisfied overall

0.317 0.199 0.070 0.026 0.354 1.020 0.475 0.188 0.577

Brisane post-intervention vs. Hervey Bay pre-intervention Intercept Gender Driving status Reliance on another driver Presence of a disability that makes it difficult to catch a bus Bus frequency 3+ times per week Bus frequency 1–2 times per week Very satisfied overall Fairly satisfied overall

1.503 0.039 0.828 0.540 0.037 0.341 0.147 2.012 0.865

Note: Pseudo-R2 = 0.0977. Model X 2ð24Þ ¼ 93:565, p < 0.001. a p < 0.05. b p < 0.01.

Sig.

K. Broome et al. / Transportation Research Part A 53 (2013) 68–80

75

The overall model was significant. Demographic variables, including gender, driving status, reliance on other drivers, and disability that made it difficult to catch a bus, were not significant variables in the model. Overall satisfaction with the buses was one of the significant contributors to the model. There were no significant differences between Hervey Bay and Brisbane pre-intervention. Compared with Hervey Bay pre-intervention, participants were 7.479 times more likely to be in the Brisbane post-intervention group if they were not very satisfied with the bus system overall suggesting a decrease in satisfaction in the Brisbane sample at follow-up. The model was consistent with the findings from the initial exploratory analysis. Overall, hypothesis one is supported, and it appears that the implementation of age-friendly guidelines improves satisfaction with the bus system. 3.3. Hypothesis two; Hervey Bay post-intervention would experience an increase in the subjective ease of bus use for both older bus users and older non-users This hypothesis was addressed by the cohort study only. There was a minor sampling bias from those who volunteered to complete the follow-up survey. Those who took part in the follow-up survey had significantly higher (Z = 2.397, p = 0.017) self-rated of ease of bus use than those who did not take part (mean of 6.46 versus 5.39). With regards to those who did take part, prior to the intervention there was also a significant difference in self-rated ease of bus use between sites (Z = 2.111, p = 0.035), with Brisbane participants reporting greater ease of bus use (mean = 7.19, SD = 2.701) than Hervey Bay participants (mean = 5.58, SD = 3.520). Following the intervention period, there was no change in self-rated ease of bus use for Hervey Bay participants (Z = 0.452, p = 0.651). Despite the lack of significant change in self-rated ease and frequency of bus use, 61.0% of Hervey Bay participants reported that buses had improved over the past 3 years, 34.1% said they remained the same and 4.9% stated they found buses more difficult. Following the intervention period, there was a significant decrease in self-rated ease of bus use (Z = 2.341, p = 0.019) for Brisbane participants from to a mean of 7.19 (SD = 2.701) to a mean of 6.21 (SD = 3.064). At the follow-up, 29.4% of Brisbane participants reported that buses had improved over the past 3 years, 62.7% said they remained the same and 7.8% stated they found buses more difficult. The proportion of participants who felt that the buses had improved was significantly higher in Hervey Bay, while the proportion of participants who felt that buses had become more difficult was higher in Brisbane (X2 = 8.941, p = 0.011). In order to control for confounding variables, multinomial logistic regression with clustered effects was used to model explanatory factors for the differences between the four samples in the cohort study; Hervey Bay pre-intervention, Hervey Bay post-intervention, Brisbane pre-intervention and Brisbane post-intervention. Clusters were defined as the matched pre and post intervention participant numbers. The model is shown below in Table 6. Hervey Bay pre-intervention was used initially as the reference category. The overall model was significant. Ease of bus use was not a significant variable when Hervey Bay pre-intervention was used as the reference category, but as the exploratory analysis suggested a change in bus ease of use for Brisbane over the intervention period, the multinomial logistic regression with clustered effects was repeated using Brisbane pre-intervention as the reference category. A comparison between Brisbane pre-intervention and post-intervention suggests that there was a significant change in bus ease of use (p = 0.019) over the intervention period with participants being less likely to have mild to no difficulty catching the bus at follow-up (odds ratio: 0.252, 95% confidence interval: 0.079, 0.799). The model was consistent with the findings from the initial exploratory analysis. Hypothesis two was not supported. Despite the fact that the majority of participants in Hervey Bay felt that the buses had improved across the intervention period, there was minimal change in self-rated ease of bus use. On the other hand, in Brisbane it was noted that, despite minimal changes during the intervention period, participants reported decreased ease of bus use. While this may seem unusual, it is important to consider that the time between the initial measures and follow-up was approximately 3 years. During that time, participants aged 3 years, and it was likely that some experienced changes associated with ageing that restricted their ease of bus use. An alternative hypothesis, supported by the data might be that the implementation of age-friendly guidelines to buses in Hervey Bay assisted participants to maintain their bus use, while participants in Brisbane experienced a decline in bus useability as a result of ageing. 3.4. Hypothesis three; Hervey Bay post-intervention would experience increase in the frequency of bus use for older people This hypothesis was addressed by both studies. For the satisfaction survey, bus use characteristics of the four groups (Hervey Bay pre-intervention, Hervey Bay post-intervention, Brisbane pre-intervention and Brisbane post-intervention) were compared. Prior to the intervention period, the majority of participants in both Hervey Bay and Brisbane caught the bus at least weekly or more frequently (92.6% and 81.7% respectively). There was a significant difference in the frequency of bus use in Hervey Bay and Brisbane (X2 = 12.739, p = 0.013). Hervey Bay participants were more likely to use the bus every day and less likely to use the bus either once a fortnight or once a month. Following the intervention, there was a decrease in the proportion of Hervey Bay participants sampled who caught the bus on a weekly basis, with an increase in the proportion who used the bus at least once a fortnight or once a month (X2 = 13.273, p = 0.010). Following the intervention, there was an increase in the proportion of Brisbane participants who

76

K. Broome et al. / Transportation Research Part A 53 (2013) 68–80

Table 6 Multinomial logistic regression with clustered effects of the cohort study. B

(SE)

Sig.

95% Confidence interval for Exp(B) Lower bound

Odds ratio

Upper bound

Hervey Bay post-intervention vs. Hervey Bay pre-intervention Intercept 2.039 Moderate difficulty catching the bus 1.127 Mild to no difficult catching the bus 1.106 70–79 years of age 0.512 80 years and older 0.390 I would like to be doing more activities 0.662 Occasional bus user 0.058 Rarely or never uses the bus 0.517 Gender male 0.163 Lived in same suburb for 5 years or more 0.107 Lives in the same suburb at follow-up 0.182 Current driver 1.082 Reliance on another driver 0.047

(0.786) (0.674) (0.608) (0.363) (0.627) (0.479) (0.568) (0.717) (0.267) (0.284) (0.403) (0.521) (0.278)

0.094 0.069 0.158 0.534 0.167 0.919 0.471 0.541 0.708 0.651 0.038a 0.867

0.086 0.100 0.294 0.198 0.202 0.310 0.146 0.698 0.515 0.545 0.122 0.553

0.324 0.331 0.599 0.677 0.516 0.944 0.596 1.177 0.899 1.200 0.339 0.954

1.214 1.091 1.220 2.313 1.320 2.873 2.432 1.986 1.569 2.643 0.941 1.647

Brisbane pre-intervention vs. Hervey Bay pre-intervention Intercept 0.157 Moderate difficulty catching the bus 0.253 Mild to no difficult catching the bus 0.675 70–79 years of age 0.710 80 years and older 0.505 I would like to be doing more activities 0.468 Occasional bus user 1.397 Rarely or never uses the bus 3.251 Gender male 0.462 Lived in same suburb for 5 years or more 0.015 Lives in the same suburb at follow-up 1.106 Current driver 0.277 Reliance on another driver 0.093

(1.702) (0.857) (0.766) (0.666) (1.150) (0.587) (0.656) (1.129) (0.589) (0.649) (1.361) (0.779) (0.632)

0.768 0.378 0.286 0.660 0.425 0.033a 0.004b 0.432 0.981 0.417 0.722 0.884

0.240 0.438 0.133 0.174 0.198 0.068 0.004 0.199 0.276 0.230 0.286 0.264

1.288 1.965 0.492 1.658 0.627 0.247 0.039 0.630 0.985 3.021 1.319 0.912

6.912 8.809 1.813 15.780 1.979 0.895 0.354 1.995 3.517 43.529 6.079 3.148

Brisane post-intervention vs. Hervey Bay pre-intervention Intercept 2.003 Moderate difficulty catching the bus 0.492 Mild to no difficult catching the bus 0.704 70–79 years of age 0.820 80 years and older 0.193 I would like to be doing more activities 0.349 Occasional bus user 0.948 Rarely or never uses the bus 37.491 Gender male 0.322 Lived in same suburb for 5 years or more 0.112 Lives in the same suburb at follow-up 0.633 Current driver 0.411 Reliance on another driver 0.271

(1.179) (0.722) (0.639) (0.632) (1.063) (0.569) (0.650) (0.800) (0.635) (0.636) (1.012) (0.747) (0.602)

0.496 0.271 0.195 0.856 0.539 0.145 0.000c 0.612 0.861 0.532 0.582 0.652

0.148 0.141 0.128 0.151 0.231 0.108 0.000 0.209 0.257 0.259 0.153 0.234

0.612 0.495 0.441 1.213 0.705 0.388 0.000 0.724 0.894 1.884 0.663 0.762

2.519 1.730 1.520 9.742 2.151 1.385 0.000 2.515 3.109 13.700 2.866 2.480

Note: Pseudo-R2 = 0.150. Model X 2ð36Þ ¼ 15365:19, p < 0.001. a p < 0.05. b p < 0.01. c p < 0.001.

caught the bus every day (X2 = 10.438, p = 0.034). As a result of the changes, there was no significant difference between the frequency of bus use in Hervey Bay and Brisbane at follow-up. These findings were confirmed in the satisfaction survey multinomial logistic regression. As demonstrated in Table 5, frequency of bus use was a significant variable in the model. There were no significant differences between Hervey Bay and Brisbane pre-intervention. Compared with Hervey Bay pre-intervention, participants were more likely to be in the Hervey Bay post-intervention group if they were not using buses daily or weekly, and thus were more likely to be using the buses less than weekly in the post-intervention sample. It is seemingly contradictory that, despite the fact the participants in Hervey Bay were more satisfied with the bus service at follow-up the participants were catching the bus less frequently than the sample prior to the intervention. A limitation of the on-board satisfaction survey study design made it impossible to follow participants over time. It may be that there are new people who are now using the bus as a result of the intervention and who are only using it infrequently (i.e., not every day). Therefore, a smaller proportion of those sampled would be frequent users, bringing down the mean frequency of use. This is supported by the fact that there were more people in Hervey Bay than Brisbane who could not answer the self-rating of change in the bus system over the past 2 years as they were not using the bus 2 years ago. This indicates that Hervey Bay may have attracted more new users. This hypothesis also fits with the change in demographics of the sample. There was a

77

K. Broome et al. / Transportation Research Part A 53 (2013) 68–80

reduction in the proportion of people in Brisbane at follow-up who were current drivers. Current drivers are more likely to drive if alternate transport is not of good quality (Broome et al., 2009). Therefore, this reduction in the proportion of current drivers sampled on buses in Brisbane may reflect the increased dissatisfaction with the bus system, causing an increased reliance on cars. Another ambiguous finding is that the proportion of participants with ‘‘a disability that makes it difficult to catch a bus’’ in Hervey Bay decreased at follow-up. This may be because following the intervention people with disabilities are not using the buses, or alternatively, improvements in the bus system led to a decrease in people with a disability who found ‘‘it difficult to catch a bus’’. It is important to supplement these findings with alternative methodologies, such as the cohort study, to provide a clearer picture. The cohort study showed somewhat different results. Prior to the intervention there were significantly more (X2 = 14.582, p = 0.001) frequent bus users in Brisbane (57.7%) than Hervey Bay (27.7%). There were similar proportions of occasional users (approximately 35%) in both Hervey Bay and Brisbane prior to the intervention. Prior to the intervention period, there was a significant association between ease of bus use and frequency of bus use (X2 = 10.732, p = 0.005) with those who never used buses experiencing more difficulty than those who used buses occasionally or frequently. There was also a significant association between driving status and frequency of bus use (X2 = 16.562, p = 0.002) with those who had never driven using the bus more frequently than retired drivers, who used the bus more frequently than current drivers. All participants who had never driven used the bus at least occasionally, while 14.3% of retired drivers never used the bus. Following the intervention period 12.2% of participants in Hervey Bay had increased their frequency of use compared to their pre-intervention responses, while 13.1% had decreased their frequency of bus use. When participants reported their self-rated change in bus use over the past 3 years, 25.0% reported they used the buses more frequently, 61.4% the same and 13.6% less frequently. Following the intervention period 9.6% of participants in Brisbane had increased their frequency of use compared to their pre-intervention responses, while 13.4% had decreased their frequency of bus use. When participants reported their selfrated change in bus use over the past 3 years, 30.8% reported they used the buses more frequently, 51.9% the same and 17.3% less frequently. Due to the difference in pre-intervention frequency of use, the cohort study multinomial logistic regression with clustered effects was used to control for confounding variables. Frequency of bus use and current driving status were significant contributors to the model. There were significant differences between Hervey Bay and Brisbane pre-intervention, with Brisbane participants using the bus more frequently. This difference was also found between the Hervey Bay pre-intervention and Brisbane post-intervention groups. Data for rarely or never uses the bus system should be interpreted with caution due to the small number of participants in each of these categories for Brisbane (4 pre-intervention and 2 post-intervention). Hypothesis number three was not supported. Again, this is incongruous with the improved satisfaction that Hervey Bay participants experiences and decreased ease of use that Brisbane participants experienced. The measure of frequency of bus use may have been limited by the sensitivity of the instrument, which measured frequency as only ‘‘frequent’’, ‘‘occasional’’ or ‘‘rarely/never’’. Alternatively, it may be hypothesised that participants in Brisbane continued to use the bus, despite difficulty. If this was the case, then it is likely that participants would prioritise non-discretionary activity (e.g., shopping), and a decrease in discretionary social activities (e.g., going to museums or libraries) would be seen in Brisbane.

3.5. Hypothesis four; Hervey Bay post-intervention would experience increased frequency of community participation for older people This hypothesis was addressed by the cohort study only. A comparison of the frequency of the social activity participation before and after intervention was conducted for both Hervey Bay and Brisbane. In Hervey Bay, there was no change in frequency of participation in all twenty activities measured with the SOCACT. Similarly, there was no significant change with the level of satisfaction with activities. In contrast, Brisbane participants experienced a significant decrease in frequency of participation for seven of the activities, including visiting exhibitions, museums and libraries, going to restaurants, exercise or playing sports, travelling or going on tours, playing with or helping children/grandchildren, participant in community or voluntary organisations, going to professional events or union meetings, and going to classes or lectures. These significant changes in Brisbane are shown below in Table 7. Between a quarter and a half of Brisbane participants experienced a reduc-

Table 7 Significant changes in participation in social activities in Brisbane before and after intervention.

Visit exhibitions, museums, libraries Go to restaurants Exercise or play sports Travel or go on tours Play with or help children/grandchildren Community voluntary organisations Go to professional events or union meetings Go to classes or lectures

Before intervention mean (SD)

After intervention mean (SD)

Mean difference

Z

p

1.84 1.96 3.25 1.75 2.16 2.24 0.65 1.46

1.51 1.61 2.70 1.39 1.67 1.67 0.41 1.02

0.33 0.35 0.55 0.36 0.48 0.57 0.24 0.44

2.100 2.442 1.954 2.107 2.165 2.779 1.887 1.985

0.036 0.015 0.051 0.035 0.030 0.005 0.059 0.047

(1.302) (1.183) (1.354) (1.007) (1.599) (1.380) (0.881) (1.501)

(1.160) (1.239) (1.768) (0.953) (1.569) (1.477) (0.670) (1.234)

78

K. Broome et al. / Transportation Research Part A 53 (2013) 68–80

tion in their frequency of use in the aforementioned activities, depending on the activity. Again, there was no significant change in satisfaction with activity levels. There were some associations found between social activity and demographic and bus use variables. Prior to the intervention there was a significant association between frequency of bus use and playing cards or other indoor games (X2 = 6.571, p = 0.037). Those who used buses occasionally engaged in cards and indoor games more frequently. After the intervention, there were no significant associations between frequency of bus use and frequency of activity participation. Prior to intervention, ease of bus use was significantly and positively correlated with frequency of visiting or helping friends and relatives (rs = 0.244, p = 0.020). Following the interventions, ease of bus use was significantly and positively correlated with exercise or playing sports (rs = 0.297, p = 0.006), travelling or going on tour (rs = 0.274, p = 0.012), playing cards of other indoor games (rs = 0.321, p = 0.005) and playing with or helping children/grandchildren (rs = 0.291, p = 0.013). Driving status prior to intervention was significantly associated with frequency of visiting exhibitions, museums and libraries (X2 = 6.924, p = 0.031), going to restaurants (X2 = 6.683, p = 0.035), visiting or helping friends and relatives (X2 = 6.560, p = 0.038) and going to classes or lectures (X2 = 6.067, p = 0.048). Following intervention, associations with driving status were significant for participation in community voluntary organisations (X2 = 8.515, p = 0.014) and going to classes or lectures (X2 = 8.440, p = 0.015). Broadly, hypothesis four was not supported by the data. An alternative hypothesis would be that age-friendly changes assist in maintaining social activity participation despite ageing. Supporting evidence is that the social activities that reduced in frequency in Brisbane following the intervention were all activities which were associated with either ease of bus use or driving status. This would suggest that an age-friendly approach to transport provision does benefit participation in a range of social activities, including visiting exhibitions, museums and libraries, going to restaurants, exercise or playing sports, travelling or going on tours, playing with or helping children/grandchildren, participation in community voluntary organisations and going to classes or lectures. 3.6. Hypothesis five; Hervey Bay post-intervention would experience a change in the barriers and facilitators to bus use reported by older people This hypothesis was addressed by the cohort study only. Participants at each site were asked to list the three most important barriers to using the bus at follow-up, with the purpose of comparing these to nominal group technique (NGT) data from a previous study (Broome et al., 2010). Thirty-three (33) participants in Hervey Bay responded to this question. The three most important barriers to using the buses in Hervey Bay after the intervention were, in order, poor vehicle accessibility (primarily difficulty getting on and off the bus), distance to and from the bus stop, and unsatisfactory timetables and scheduling (commonly characterised by infrequent buses). Similarly, in Brisbane, 42 participants answered the questions with the most important barriers being, in order, inappropriate timetables and scheduling, distance to and from the bus stop, and poor vehicle accessibility. This is closely aligned to the top three barriers from the original NGT data (namely limited scheduling of buses, long distance to the bus stop and poorly accessible bus entry and exit) (Broome et al., 2010). This indicates minimal change to primary barriers at both sites, and the hypothesis is not supported. Replacement of some older buses with new lower floor buses, as well as an increase in frequency of buses, occurred at both sites to different degrees. The increase in the proportion of lower floor buses was in line with the requirements of the Disability Standards for Accessible Public Transport 2002 (Williams, 2002). This highlights that while increases in proportion of lower floor buses and frequency of buses may take place, the full impact of implementing age-friendly guidelines may not be experienced until 100% compliance is reached. Therefore, a graded approach to improving accessibility may have limited effectiveness until the end goal is reached. 4. Discussion Synthesising the findings, it could be suggested that implementation of age-friendly guidelines described for Hervey Bay may improve satisfaction with bus use and maintain ease of bus use for older people. 4.1. Strengths and limitations of the research Both methodologies provide different, but complementary, views of the age-friendliness of bus systems. This was a strength of the study as it provided triangulation of results, and opportunity to explore and explain discrepancies. Both methods have strengths and weaknesses. The on-board satisfaction survey was able to provide highly valid perspectives on the bus system from older people who had recently used and experienced the bus system. In contrast, it was limited in its ability to monitor changes in frequency and ease of use, as individuals could not be followed up over time. Satisfaction survey participants were easier to recruit than the cohort study (resulting in larger feasible sample sizes), and participation in the study was less onerous. The cohort study, with repeated measures, was more reliable in monitoring changes in ease and frequency of bus use and social participation. The cohort study also captured the perspectives of bus non-users, who are an important target group in accessibility studies. Limitations of the cohort method were that volunteer sampling bias is likely to increase due to the increased commitment required and that appraisal of the bus system may have occurred outside

K. Broome et al. / Transportation Research Part A 53 (2013) 68–80

79

the context of recent experience of the bus system. It is recommended that future research considers the need for multiple methods to address the limitations of single methodologies. Future studies should also monitor patronage data if possible to provide further triangulation. The study was also limited by the scope of age-friendly changes. While there were many age-friendly changes in Hervey Bay and limited age-friendly changes in Brisbane during the intervention period, neither site fully met the most important requirements of an age-friendly bus system. Therefore, further age-friendly changes may result in greater improvements in satisfaction, bus use and social activity. Subsequent research may aim to develop further age-friendly technologies and innovations, and evaluate the impact of these on age-friendly outcomes such as satisfaction, frequency and ease of use, and social activity participation. While the outcome measures used in the study provided a broad range of information, and were useful in analysis, future researchers may consider improvements in the development of outcome measures. The 5-pt Likert scale and associated wording used in the satisfaction resulted in a ceiling effect, and limited the tools sensitivity to measure change. The small number of categories in frequency of use also limited the sensitivity of the tool. Future researchers may consider alternate measures such as the number of times using the bus in the past month.

4.2. Implications for transport providers and policy makers The findings from this study suggest that an age-friendly approach to bus systems has the potential to assist older people to maintain bus use and social participation as they age. By addressing the priorities set by older people, older people will become more satisfied with the bus system and are more likely to report an improvement in the bus system. This can occur even when not all priorities are completely met, as was the case in this study. The minimal changes in barriers post-intervention (hypothesis five) suggests that further improvements can be made. It is possible that more extensive changes may result in improvements, not just maintenance, of ease and frequency of bus use, and social participation. The changes achieved in this study were a result of a concerted, albeit limited, effort to implement the seven priorities for an age-friendly bus system. Table 1 captures the extent to which each priority was achieved. If transport providers and policy makers seek to replicate or improve upon the results of this study, they should be cognisant of the seven priorities, approaches to achieving these priorities and relevant barriers and facilitators to change. The first priority for an age-friendly bus system is improving the accessibility of buses by providing lower floor buses that make it easier to get on and off. In this study, the increasing proportion of lower floor buses was legislatively mandated at a minimum of 25% of all bus services, although full compliance is not expected until 2022 according to the compliance schedule (Williams, 2002). Since getting on and off buses remained a critical barrier after the intervention period suggests that satisfactory progress in this area will not have been made until a far higher percentage of buses are lower floor buses. Cost is frequently implicated as the main barrier to services using lower floor buses. Policy makers will therefore need to consider cost implications when considering planning for an age-friendly bus system. The second priority for an age-friendly bus system involves facilitating the friendliness and helpfulness of bus drivers. A pilot study (Broome, 2010) suggests that age-awareness training for bus drivers has the potential to increase the frequency of age-friendly bus driver behaviours. The training in this study was a one-off session of 1.5 h duration, requiring minimal investment by transport providers. The third priority for an age-friendly bus system is the provision of frequent buses, including during the evenings and on weekends. Similar to lower floor buses, it appears that high satisfaction is unlikely to be achieved until all services meet standards higher than those used in this study (e.g., a maximum of 15 min intervals). Again, policy makers should consider cost implications when improving the headway of buses through more frequent services. The fourth priority is providing bus stops that are close to homes and destinations. Flexible route transport used on one service in Hervey Bay was a powerful example of a transport innovation that could meet multiple needs of older people. Where contexts are suitable, such as in medium to low population density areas, transport providers are encouraged to trial this system, as it has significant benefits for useability, patronage and satisfaction (Broome et al., 2012). Alternatively bus routes can be designed to provide maximum coverage of all areas within a system, aiming to be within 200 m of all residences and destinations. The satisfaction survey findings from this study also suggest that attention should be given to the accessibility of bus stops and shelters themselves. Shelters should provide accessible seating, provide shelter from the weather at all times and integrate well with vehicle boarding. The fifth priority involves providing accessible pedestrian infrastructure, including footpaths and pedestrian crossings. The outcomes from this study were associated with a comprehensive plan by the local council in Hervey Bay to provide walkable neighbourhoods, with footpaths connecting all areas. An audit of pedestrian crossing times (Amosun et al., 2007) was not conducted, which may also be beneficial if concerns exist. The sixth priority was providing appropriate information and training for older people on how to use buses. In this study, prior to intervention, bus users rated their satisfaction with their own knowledge about using buses positively compared to other aspects of the bus system. The bus buddy program introduced in Hervey Bay had minimal uptake. Satisfaction with older bus users’ own knowledge declined during the research period. This suggests that alternative approaches need to be considered, for example the provision of age-friendly information or strategies to target specific populations such as new residents or older people retiring from driving (Liddle et al., 2006).

80

K. Broome et al. / Transportation Research Part A 53 (2013) 68–80

The seventh priority involves ensuring that bus systems provide access to destinations of interest to older people. One approach is to involve the needs of older people in route design. The destinations preferred by older people (Broome, 2010) include cross-suburban travel, leisure destinations, shops, CBD, medical services and other towns. Alternatively, flexible route bus systems also meet this need by provide flexibility in the destinations provided. 5. Conclusion Improving the useability of buses is an important endeavour in facilitating age-friendly communities. The application of age-friendly guidelines to public bus systems may facilitate maintenance of bus use, and participation in social activity, with ageing. This is one of the first studies to measure the effect of age-friendly changes on social participation. The implementation of age-friendly guidelines in a systematic way results in improved satisfaction with the bus system and a subjective perception of improvement in the bus system. A combined approach, addressing a number of priorities, is likely to have a greater impact than individual interventions alone. Transport providers and policy makers should be guided by research evidence, selecting effective technologies and innovations that meet the priorities of older people and have a documented impact on satisfaction, bus use, ease of use and ultimately participation of older people in communities and society. Acknowledgements This study forms part of a project supported by the Australian Research Council, Queensland Transport and the Queensland Department of Communities which aims to establish processes for creating, implementing and evaluating age-friendly guidelines using the case study of public buses. Particular acknowledgement is given to Wide Bay Transit for providing data and facilitating access to the buses. References Amosun, S.L., Burgess, T., Groeneveldt, L., Hodgson, T., 2007. Are elderly pedestrians allowed enough time at pedestrian crossings in Cape Town, South Africa? Physiotherapy Theory and Practice 23, 325–332. Australian Bureau of Statistics, 2007a. 2006 Census QuickStats: Hervey Bay City Part A. . Australian Bureau of Statistics, 2007b. 2006 Census QuickStats: Brisbane (Statistical Division). . Balcombe, R.J., Astrop, A.J., Hill, E., 1998. Concessionary Fares: Trip Generation among Elderly Passengers. Transport Research Laboratory, Crowthorne. . Broome, K., 2010. Age-Friendly Buses: Development, Evaluation and Occupational Therapy Theory. PhD Thesis. The University of Queensland, Brisbane. Broome, K., McKenna, K., Fleming, J., Worrall, L., 2009. Bus use & older people: a literature review applying the person-environment-occupation model in macro practice. Scandinavian Journal of Occupational Therapy 16, 3–12. Broome, K., Worrall, L., McKenna, K., Boldy, D., 2010. Priorities for an age-friendly bus system. Canadian Journal on Aging 29, 435–444. Broome, K., Worrall, L., Fleming, J., Boldy, D., 2011. Identifying age-friendly behaviours for bus driver age-awareness training. Canadian Journal of Occupational Therapy 78, 118–126. Broome, K., Worrall, L., Fleming, J., Boldy, D., 2012. Evaluation of flexible route bus transport for older people. Transport Policy 21, 85–91. Cruice, M., Worrall, L., Hickson, L., 2005. Personal factors, communication and vision predict social participation in older adults. Advances in Speech Language Pathology 7, 220–232. Dent, O.F., Broe, G.A., Creasey, H., Waite, L.M., Cullen, J.S., Grayson, G.A., 1999. Transportation needs of community-living older people in Sydney. Australasian Journal on Ageing 18, 186–190. Gabriel, Z., Bowling, A., 2004. Quality of life from the perspectives of older people. Ageing & Society 24, 675–691. Koffman, D., Salstrom, R., 2001. How to Best Serve Seniors on Existing Transit Services. Mineta Transportation Institute, San Jose. Liddle, J., McKenna, K., Bartlett, H., 2006. Improving outcomes for older retired drivers: the UQDRIVE program. Australian Occupational Therapy Journal 53, 303–306. Lui, C.W., Everingham, J.A., Warburton, J., Cuthill, M., Bartlett, H., 2009. What makes a community age-friendly: a review of international literature. Australasian Journal on Ageing 28, 116–121. Shaheen, S.A. and Rodier, C.J. (2006) Video Transit Training for Older Travelers: A Case Study of the Rossmoor Senior Adult Community, California. . Stahl, A., 1987. Changing mobility patterns and the aging population in Sweden. Transportation Research Record 1135, 37–41. Suen, S.L., Sen, L., 1999. Mobility options for seniors. Transportation Research Board, Bethesda, Maryland. White, P.R., Turner, R.P., Mbara, T.C., 1992. Cost benefit analysis of urban minibus operations. Transportation 19, 59–74. Williams, D., 2002. Disability Standards for Accessible Public Transport 2002. Australian Government, Canberra. World Health Organization, 2007a. Global Age-friendly Cities Project. . World Health Organization, 2007b. Global Age-friendly Cities: A Guide. World Health Organization, Geneva. . Worrall, L., Hickson, L., 2003. Communication Disability in Aging: From Prevention to Intervention. Thomson Delmar Learning, Clifton Park, NY.