Article Text

Download PDFPDF

How, where and with whom? Physical activity context preferences of three adult groups at risk of inactivity
  1. Nicola W Burton1,
  2. Asaduzzaman Khan2,
  3. Wendy J Brown1
  1. 1School of Human Movement Studies, The University of Queensland, St Lucia, Brisbane, Australia
  2. 2School of Health and Rehabilitation Sciences, The University of Queensland, St Lucia, Brisbane, Australia
  1. Correspondence to Nicola Winship Burton, Human Movement Studies, The University of Queensland, Blair Drive, Brisbane, Queensland 4072, Australia; nburton{at}hms.uq.edu.au

Abstract

Background Strategies to promote physical activity may be more successful if they reflect people's interests.

Purpose To explore physical activity contexts preferred by three adult groups at risk of inactivity: older adults, those with low income and those with high body mass index (BMI).

Methods Cross-sectional data were from a mail survey of 7873 adults aged 42–67 years. Respondents indicated the extent of disagreement or agreement with a preference for each of 14 contexts relating to format (eg, vigorous), location (eg, outdoors) and social setting (eg, done alone). Data were analysed using multilevel multinomial logistic regression. Adjusted OR and 95% CI are reported.

Results Those aged 60–67 (vs 42–49) years had significantly higher odds to prefer activities with people of same age, and significantly lower odds to prefer activities that are at a fixed time with scheduled sessions, competitive, team-based or vigorous. Adults with low (vs high) income had significantly higher odds to prefer activities that are low cost, not just about exercise, team-based, supervised, skill-based or that can be done alone, and significantly lower odds to prefer vigorous or outdoor activities. Adults with BMI 30+ (vs<25) had significantly higher odds to prefer activities that are supervised, with people the same sex, team-based, with people the same age, or at a fixed time with scheduled sessions.

Conclusions These three groups had distinct preferences for how, where and with whom the physical activity is done. This information could be used by those who promote, design, deliver and evaluate physical activity opportunities.

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Background

The benefits associated with regular physical activity participation are well documented, and include reduced risk of premature mortality, obesity, cardiovascular disease, some cancers, musculoskeletal problems, type II diabetes, psychological difficulties and healthcare usage.1,,4 Benefits can be experienced by people who become active even after an extended period of inactivity.5 In many Western countries, however, a large proportion of adults are insufficiently active, in particular those who are aged 60+ years, those who are socioeconomically disadvantaged and those with high body mass index (BMI).6,,10

Strategies to increase physical activity participation may be more effective if they reflect people's interests. Research on adults' activity preferences is, however, scarce, and often limited to activity type or intensity. Studies on activity context have tended to focus on activity done alone or in a group.11,,16 Other researchers have studied supervision and cost,11 structure,17 ease of learning18 and the mode of delivery.19 More comprehensive assessment has included multiple contextual variables of location (home, outdoors, work, centre-based), companions (alone or with a family member, coworker or friend), time of day (morning, afternoon, evening), variable scheduling (same or different time each day), structure (flexible or scheduled), supervision and other aspects such as music, fun, convenience and hard exercise.20,,22

Only a few studies of physical activity context have been conducted with adults in the general population instead of clinical or intervention participants. In a telephone survey with 3305 adults in the USA in 1986 to 1987, 69% of mid-aged (50–64 years) and 67% of older (64+ years) participants reported a preference to exercise alone with some instruction rather than in a group with an exercise leader.14 An interview-based study of 2298 inactive Australian adults in 1990 to 1991 found that 27% of those aged 40–59 years and fewer than 20% of those aged 60–78 years wanted to exercise with a group, and fewer than 5 per cent of adults aged 40+ years preferred team sports.13 A UK interview study of 947 adults aged 30–91 years demonstrated no age-related differences in preferring to exercise alone, but that if doing group-based exercise, people preferred same-aged participants.12 Data from this study also indicated a preference for gender-segregated exercise groups, particularly among those with a BMI ≥25.23 In a mail survey of 1332 Australian adults aged 18 to 60+ years in 1998, more people reported enjoying an unstructured than a structured activity (53% vs 31%).17 Most of these studies assessed a limited number of activity contexts and presented descriptive results. Only a few explored group differences. There is a need, therefore, for more population-based studies that assess a range of contextual variables, and with the systematic assessment of group differences.

The aim of this study was to assess the physical activity contexts preferred by three groups of adults in the general population at high risk of inactivity: older adults, those with low income and those with high BMI. We assessed 14 physical activity contexts related to format, location and social setting, and examined if there were preference differences between (A) adults aged 60–67 and 42–49 years, (B) low and high income groups and (C) those with BMI >30 and those with BMI <25.

Methods

Data were collected as part of the HABITAT (How Areas in Brisbane Influence Health and Activity) study, which was awarded the ethical clearance by the QUT Human Research Ethics Committee. Survey return was taken as informed consent.

Sample design and data collection

Data were from the 2009 HABITAT survey which was the first wave to include the physical activity context items. Details on the design, sampling and data collection for HABITAT have been published elsewhere.24 A multistage probability sampling design was used to select a stratified random sample of 200 Census Collectors' Districts (CCD) in Brisbane, the state capital of Queensland, Australia. Using data from the Australian Electoral Commission, a random sample of 85 people aged 40–65 years was selected from within each CCD. A mail survey was administered during May to July 2007 and again in 2009 using a method that included advance mail notice, personalised mail, a thank you/reminder notice, resending to non-respondents and a final letter to the non-respondents.25 From 17 000 eligible participants in 2007, 11 037 surveys were returned (68.5%). Between 2007 and 2009, 117 people were unable to be contacted, 54 withdrew, 11 died and 11 moved overseas, leaving 10 844 eligible participants in 2009.

Measures

Fourteen questionnaire items were used to assess the physical activity contexts related to format (how), location (where) and social setting (with whom). Items are shown in the Appendix. Respondents indicated the extent to which they agreed or disagreed with a preference for each of the contexts using a five-point Likert scale (strongly disagree, disagree, no preference, agree, strongly agree). Responses were collapsed into three categories of disagree, no preference or agree.

Age, income and BMI were also assessed by a questionnaire. Respondents provided date of birth, the birth year was subtracted from 2009 and age was categorised into one of three groups: 42–49, 50–59 or 60–67 years. Respondents indicated the gross (ie, before tax) annual household income from 11 response options, which were then collapsed into four categories (gross annual household income): <41 600; 41 600 to 71 799; 72 800–129 999 or 130 000+ per annum. The lowest income category is below the national average in 2006.26 Respondents reported height and weight which were used to derive BMI which was then categorised into one of three groups (kg/m2): <25, 25 to <30, or 30+.

The sociodemographic and health variables used in the analyses are shown in table 1. Questionnaire items were used to assess sex, household composition, education, employment status and overall health status. The K6 scale was used to assess psychological distress.27 Physical activity was assessed using items from the Active Australia Survey to indicate the time spent in the previous week doing walking, vigorous physical activity and moderate physical activity (excluding household chores and gardening).28 Responses were quantified in MET min/week as ((walking minutes×3 METS) + (moderate minutes×3 METS) + (vigorous minutes×7.5 METS)) and categorised into one of five levels: <90, >90–<450, >450–<900, >900–<1800, or >1800.

Table 1

Sociodemographic and health characteristics of the 2009 HABITAT respondent sample (N=7873)

Analyses

Analyses were conducted in 2011. To reflect the multistage sample selection, a multilevel approach was used. Multilevel multinomial logistic regression analyses with random effects were conducted using GLLAMM (Generalised Linear Latent And Mixed Models) commands of Stata version 10.1 (StataCorp, College Station, Texas, USA) to examine the possible relationships between preferred physical activity contexts and (A) those aged 60 to 67 years, (B) those with low income and (C) those with high BMI. Such relationships were adjusted for a variety of sociodemographic and health variables which were significantly associated with the outcomes of interests at p<0.005. A smaller p value was used in order to deal with the potential increase of the likelihood of type I error due to large sample size. A random intercept for CCD was used in the modelling to account for clustering within CCDs. The assumption of parallel regression, essential for ordinal logistic regression, was assessed using the Brant test. The estimation procedure used was numerical integration (10 integration points) with adaptive quadrature in order to obtain more reliable estimates of parameters.29 Disagreement with the activity preference was considered as the reference category for multinomial modelling and the models comparing the agreement and disagreement are presented. The adjusted OR and their 95% CI are reported where OR>1 is interpreted as a more likely preferred context, and OR<1 is interpreted as a less likely preferred context.

Results

Participants

Of the 10 844 surveys sent in 2009, 322 participants declined participation, 161 were ineligible (deceased, unable to respond), 2488 did not respond and 7873 returned surveys with data (74% of eligible cases). The sociodemographic and health profile of respondents are provided in table 1.

Activity context preferences

The proportion of people, by age, income and BMI group, indicating agreement or disagreement with each of the activity contexts is presented in table 2. In terms of format, over 75% of respondents in each of the three groups preferred activities at little or no cost. At least 60% indicated a preference for activities that are not just about exercise, and a preference against activities that are competitive. In terms of location, over 80% of respondents in each of the three groups preferred activities that can be done close to home, and at least 60% preferred outdoor activities. In terms of social setting, almost 75% of respondents in each of the three groups indicated a preference for activities that can be done alone. A summary of the multivariable results comparing (A) the adults aged 60–67 and 42–49 years, (B) low and high income groups and (c) those with BMI>30 and those with BMI<25, is provided in table 3.

Table 2

Proportion of respondents who indicated agreement/disagreement* with a preference for each activity context, by age, income and BMI

Table 3

Summary results of the multilevel multinomial logistic regression* to assess group differences in activity context preference (OR and 95% CI)

Adults aged 60–67 years had significantly higher odds than those aged 42–49 years to agree with a preference for activities with people the same age, and significantly lower odds to agree with a preference for activities at a fixed time with scheduled sessions, that involve competition, that are team-based and activities that are vigorous. There were no significant differences between the two age groups for activities that can be done alone, that are outdoors, with people of the same sex, with little or no cost, with supervision, in a set routine or a format, that are not just about exercise, that require skill and practice, or activities that are close to home.

Adults in the lowest income group had significantly higher odds than those in the highest income group to agree with a preference for activities that are low cost, are not just about exercise, are team-based, are supervised, require skill and practice, and activities that can be done alone. They had significantly lower odds to agree with a preference for activities that are done outdoors and vigorous activities. There were no significant differences between the two income groups for activities with people the same sex, at a fixed time with scheduled sessions, with people of the same age, that involve competition, with a set routine or a format, or activities near home.

People with BMI 30+ had significantly higher odds than those with BMI <25 to agree with a preference for activities that are supervised, with people the same sex, team-based, with people the same age and activities at a fixed time with scheduled sessions. There were no significant differences between the two BMI groups for activities that are outdoors, involve little or no cost, involve competition, are not just about exercise, are near home, can be done alone, have a set routine or a format, require skill and practice, or activities that are vigorous.

Discussion

This study examined the physical activity contexts preferred by three adult groups at high risk of inactivity: older adults, those with low income and those with high BMI. It is unique in that it used a large population-based sample; examined multiple contextual factors related to format, location and social setting; and explored group differences using multivariable analyses with adjustment for a range of sociodemographic and health-related variables including sex, living arrangement, physical and psychological health, employment status and current physical activity level. By asking participants to indicate a preference for or against each context separately, we avoided possible masking. Others who have asked a dichotomous question, for example, ‘exercising in a group with an exercise leader, or on your own with some instruction’, have combined the attributes of the social setting (group vs alone) and format (exercise leader vs some instruction), which excludes other options (eg, exercising with others with no leadership) and may lead the respondents to choose the least objectionable option. Beauchamp et al, for example, noted that a preference for solitary activity may be present if the respondents anticipate the group context to be (unappealingly) composed of people of a different age group.12 By assessing a range of preferences across domains of format, location and social setting, we were able to see how they are combined.30

General physical activity context preferences

In terms of format, the majority of respondents in each of the three groups (60–67 years, low income, BMI>30) indicated a preference for physical activities that involve little or no cost, or activities that are not just about exercise; and a preference against competitive activities. Costs have been identified previously as a potential barrier to activity for older adults,31 ,32 and those with low income,33,,35 or weight concerns,11 and can include membership or access fees, associated resources (eg, equipment, clothing) and indirect costs (eg, travel). A preference for activities that are not just about exercise may reflect a lack of enjoyment in exercise, which is a major barrier to activity participation.33 ,36 ,37 Another study indicated that aerobics and exercise classes were viewed with indifference by most blue-collar worker participants, and with antipathy by older women.38 Participants indicated that it was difficult to enjoy a structured physical activity if not an ‘ideal’ size and shape, which can be an issue for those with high BMI. The preference of these groups against competitive activities is consistent with other studies indicating that the interest in competition is predominant among young men,39 ,40 or those engaged in vigorous activity.41

For location, over 80% of respondents in each of the three groups preferred activities done close to home, and at least 60% preferred outdoor activities. Other studies have shown that adults are more likely to use facilities located close to home than elsewhere,42 and that the proximity or convenience of destinations and facilities is positively associated with activity participation,43 ,44 in particular walking for transport.45 Outdoor exercise has been associated with greater mean duration of exercise sessions than exercise done at home, work or at gym/health club,46 and additional benefits for health and well-being.47 ,48

In terms of the social setting, almost 75% of respondents in each of the three groups preferred physical activities that can be done alone. This is consistent with other research with mid-aged,13 and older adults.13 ,14 ,49 Activities that are done alone reduce reliance on others, and avoid difficulties related to scheduling and shared access. People may prefer to exercise alone because of the negative social comparisons associated with, for example, poor ability, unfavourable appearance. As previously noted, a preference for solitary activities may also present if respondents anticipate the group to be composed of people of a different age group.12 Similarly, the respondents may have other concerns with group members, such as their level of ability.

Activity context preferences of older adults

Adults aged 60–67 years were more likely than those aged 42–49 years to prefer activities done with people the same age, and less likely to prefer activities at a fixed time with scheduled sessions, activities that involve competition, activities that are vigorous, or activities that are team-based. This is consistent with other research indicating no age differences in the preference to exercise alone,12 and that older adults prefer to exercise with people of the same age,12 and not in team-based activities.13 As part of the transition away from the structured time demands associated with mid-adulthood (eg, work), older adults may want, and have the capacity for, a more flexible approach to scheduling activities. This flexibility may also help older adults to manage new time commitments, for example, caring for grandchildren. Time management problems and difficulty scheduling activity have previously been identified as a barrier to activity participation among older adults.50 Other commonly reported barriers for this age group include poor health and fear of injury,13 ,40 ,51,,54 which could contribute to the preference against activities that are vigorous or competitive. Also, some older adults consider vigorous exercise as not age appropriate.55 Accordingly, the preference for activities with others around the same age may be so as to be with people at a similar level of ability; younger adults may be at too high a level, and older adults too low.

Activity context preferences of adults with low income

Adults in the lowest income group were more likely than those in the high income group to prefer low cost activities, activities that require skill and practice, team-based activities, supervised activities, or activities that can be done alone. Cost is a commonly identified barrier to physical activity among those with socioeconomic disadvantage.33,,35 ,52 ,56 ,57 Other research has indicated that competence is an influence on activity participation among the individuals of low socioeconomic status,33 ,56 ,58 which is consistent with the preference for skill-based activities. The preference for team-based activities may reflect sporting interests, for example, football, netball. Social support and social participation have been reported as positively associated with activity participation among socioeconomically disadvantaged groups,59,,61 and team-based activities may, therefore, be an important source of social capital. Skill-based and team-based activities often involve a referee, coach, or trainer, which may contribute to the preference for supervised activities. It seems contradictory to have a preference for team-based activities as well as a preference for activities that can be done alone. This may be, however, because of the interest in a range of activities that involve these attributes, for example, team-based sport and swimming, or because of a disinterest in the group-based physical activity such as exercise classes.

The low income group was also more likely to prefer activities that are not just about exercise, and less likely to prefer vigorous or outdoor activities. Disadvantaged people are more likely to perceive the neighbourhood as unattractive and unsafe,33 ,62,,66 which could constrain outdoor activities, such as walking in the local area. The preference for activities that are not just about exercise and against vigorous activity could indicate a general disinterest in exercise, in particular the more traditional forms such as fitness classes. Other research has indicted that socioeconomically disadvantaged adults are less likely to value healthy activities such as exercise,67 and more likely to perceive health behaviour as compromising the quality of life.38 Individuals may also be disinterested in recreational activity because of physically demanding occupations.68

Activity context preferences of adults with high BMI

Adults with BMI 30+ were more likely than those with BMI<25 to prefer activities with people of the same sex, or activities with people around the same age. Another study also found a preference for same sex exercise groups among people with BMI ≥25.23 People with high BMI can feel embarrassed and intimated by exercise,69 ,70 and research on obese individuals' attitudes to physical activity has highlighted the barrier of stigmatisation, and the importance of a sense of solidarity, and feeling emotionally secure and socially accepted.71 ,72 Accordingly, activity companions of the same age and same sex may be seen as potentially more empathic and less critical than those of the opposite sex and a different age.

The high BMI group also preferred activities that are supervised, team-based activities, or activities at a fixed time with scheduled sessions. This may indicate a desire for high levels of extrinsic support and structure. Other researchers have reported that obese individuals can find it very difficult to exercise on their own, and that they want someone else to take responsibility for motivating them.71 Activity supervisors and team members can assist with motivation (eg, encouragement, feedback), provide individualised instruction on a suitable type and level of activities, assist with goal setting and problem-solving barriers, emphasise the importance of participation instead of mastery and be the focus of a sense of social commitment, all of which have been positively associated with the physical activity participation.73 Scheduled sessions can create a routine and social commitment that also helps to maintain an ongoing participation.

Methodological considerations

A comparison of the 2007 baseline sample with national census data indicated an under-representation of those with school-only education, not in the workforce, and living in disadvantaged areas,74 which may have influenced the results given that there were socioeconomic (income-based) differences in preferences. The multivariable analyses were, however, adjusted for education and employment status. We did not ask the respondents to rank their preferences, which can lead to seemingly conflicting results, such as a preference for activities done alone and a preference for activities with same sex companions. This may, however, reflect conditional interests, for example, a preference for solitary activities, and a desire for same sex groups if doing group-based activities. As preferences may be (positively or negatively) influenced by experience, physical activity level was included in the multivariable analyses. Contextual preferences may not always reflect the actual physical activity participation, which may be influenced by other factors such as pragmatics and opportunity. As this study was conducted with mid-aged adults, results may not be generalisable to young adults.

Conclusions

Information on physical activity contextual preferences can be used by health professionals, service providers and others who advocate for, promote, design, deliver and evaluate physical activity opportunities. The general preference for activities done close to home highlights the need for facilities (eg, recreational clubs) and infrastructure (eg, walking paths) across residential areas. In response to the preference for low cost activities, service providers could consider strategies such as loyalty programmes, group discounts, discounted bulk purchasing, financial incentives to promote and reward ongoing participation, and special offers for those with low income. At a population level, tax incentives can also reduce the cost burden.75 Given the general preference for activities that are not just about exercise, promotional efforts should highlight the additional attributes of activity participation. One study of the elements of a group exercise programme needed to encourage attendance found that 85 per cent of respondents wanted ‘good music’ and ‘fun exercise’.22 Another study that focused on engaging low income groups in activity programmes found that high levels of social interaction, interest and enjoyment were associated with improved levels of retention.33 The general preference for outdoor activities may also reduce the focus on exercise per se.

What this study adds

  • This text provides a novel information on the contexts that people prefer for physical activity, and demonstrates that mid-aged adults with high body mass index (BMI), those with low income and those over 60 years of age have distinct preferences for how, where and with whom they are physically active.

  • Those aged 60 to 67 years had a preference for activities with people the same age, and against activities that are vigorous, competitive, team-based and at a fixed time with scheduled sessions.

  • Adults with low income had a preference for activities that can be done alone, are supervised, are team-based, are skill-based, are low cost and that are not just about exercise; and against vigorous activities and outdoor activities.

  • Adults with BMI 30+ had a preference for activities with people the same age, with people the same sex, with supervision, in teams and activities at a fixed time with scheduled sessions.

Workplace team events, such as pedometer-based step challenges, may be more appealing to those with low income or high BMI, given the preference for team-based activities in these two groups. As these two groups were also more likely to prefer supervised activities, minimal contact and self-directed activity programmes may not be effective. Exercise-based group programmes for those with high BMI should involve participants of the same age and sex, and scheduled sessions. Group programmes for older adults should involve the participants of the same age, and include flexible timing. Community-based approaches to promote physical activity in low income neighbourhoods could include providing indoor opportunities and team sports, and should highlight the skill component.

Previous research indicates that physical activity which is consistent with individuals' preferences for mode is associated with less negative effect, fatigue and perceived exertion; and higher well-being and programme attendance.76,,78 The mode of activity is, however, only one aspect of participation. Most types of activity can be done in a range of formats, locations and social settings. For those with little intrinsic interest in physical activity, the context may be a particularly important influence on participation. Understanding how, where and with whom people prefer to do physical activity can be the additional useful information to tailor activity opportunities to be more appealing and thereby increase participation.

Acknowledgments

The HABITAT study was supported by project grants from the (Australian) National Health and Medical Research Council (NHMRC) (ID 339718, 497236), with some area-level data provided by the Brisbane City Council. NB was supported by a Heart Foundation fellowship (PH08B3905) and a NHMRC Program Grant (569940). The authors acknowledge Mr Martin O’Flaherty for his assistance with data coding and cleaning, Ms Robyn Baguley and Ms Sophie Miller for project management, and Ms Kylah McCarthy for her assistance with the literature.

References

Footnotes

  • Funding (Australian) National Health & Medical Research Council.

  • Competing interests None.

  • Ethics approval QUT Human Research Ethics Committee.

  • Provenance and peer review Not commissioned; externally peer reviewed.