Mediated intervention approaches have been put forth as promising public health approaches for physical activity promotion due to their potential for wide reach, accessibility, effectiveness and cost-effectiveness [28–30]. However, results from the current study suggest that the desirability of these interventions among the population, especially for telephone-based interventions is low and rather face-to-face physical activity programs (group-based or one-on-one with an instructor) that are more resource intensive to deliver are preferred. Given the need for more sustainable behaviour change approaches this is of concern and may represent a significant public health challenge. Whilst several reviews have shown that mediated interventions can be effective [28–30], information relating to participant representativeness and the reach, adoption and maintenance of the implemented interventions is scarce . As such, it is difficult to determine if these interventions will be effective in a real world-setting, where preference for them is low. In order to boost the public health impact of mediated interventions it may be necessary to combine them with other more preferred approaches, (which has been shown to enhance efficacy of telephone-based interventions ) and/or develop strategies to enhance their appeal. The results of this study provide some insight into what target groups may be likely to adhere to mediated intervention and what target groups may require additional encouragement or assistance to adhere to these approaches.
Comparison to previous research
Our results support the previous research findings that men are more likely to prefer mediated (print and online) interventions than women  and that people with lower physical activity levels are more likely to prefer mediated programs than people with higher levels of physical activity . The preference for mediated intervention among men may be due to men’s desire to self-monitor their health, maintain their regular activities, and to obtain health information independently and judge illness severity before seeking help . There may also be social factors contributing to this preference, such as men’s traditional social roles (e.g., difficulty relinquishing control; immunity and immortality; perception that men are not interested in prevention; and lack of male care providers) . Among those with lower physical activity levels, mediated interventions may be perceived as less confronting, and hence preferred to face-to-face interventions. This seems plausible, since low activity levels are associated with lower self-efficacy, knowledge and skills .
The finding that having a chronic disease is associated with an increased preference for print-based interventions (in the health model) in the current study is in contrast to studies exploring intervention preferences among individuals with diabetes  and cancer . In both of these studies, a strong preference for face-to-face interventions over mediated interventions (telephone, video-tape, pamphlet, and internet) was reported. One explanation for the different findings observed in the current study is that age may have had a confounding effect, since older age is associated with both the preference for print-based interventions and the likelihood of having a chronic disease [35, 36]. Alternatively, it may be that the distinct etymology associated with different chronic diseases impacts on an individual’s physical activity support needs. In the case of elderly people, a wide range of barriers associated with older age, such as lack of transportation facilities, financial considerations, lack of affiliation to the fitness center culture, or social embarrassment, may contribute to a preference for mediated interventions that are home-based .
To our knowledge, previous research has not examined the influence of geographical location, weight status or internet use on preference for intervention delivery mode. In doing so, we found some novel findings. Namely, living in a regional town was associated with preferring group-based interventions and living rurally was associated with preferring online interventions. This may be due to the unique social structures and availability of health services present in each of these settings [38, 39]. For example, individuals that live in rural Australia face transport and accessibility problems that are not experienced by those living in regional and urban areas , potentially making the practicality of online interventions more attractive. The interactivity of these interventions may also enhance their appeal, since social isolation is also often experienced in rural settings [38, 40]. Another novel finding was that obesity was strongly associated with preferring face-to-face interventions with an instructor over any other delivery mode. Possible reasons for this include being too shy or embarrassed to exercise in front of a group and/or a perceived need for greater support than what they believe could be provided via a mediated intervention . This could have public health implications, given the health and economic burden associated with obesity [42, 43] and the need for sustainable approaches. Finally, this study also found that being in the 35–44 age group, compared to being in the 18–34 age group, was associated with a greater preference for web-based interventions over face-to-face interventions. This may reflect the increasing use of the internet among this age group at work (51%), and at home (52% access internet at home daily) , and/or the difficulty or perceived difficulty of attending face-to-face sessions at this stage of life due to a busy social or professional life .
Strengths and limitations
The major strength of this study is that it is the first to examine the relationship between physical activity intervention delivery mode preference and individual characteristics among adults. A second strength is that the research was conducted in a large sample recruited from the general population. Research in this field is scare and is predominantly limited to research describing intervention preferences in special sub-population groups. Hence, this study has contributed to the literature by expanding current knowledge about adults’ physical activity intervention delivery mode preferences and by highlighting what factors may influence these choices.
This study also has some limitations that should be considered when interpreting results. First, there was over-sampling of older adults and under sampling of younger adults. This is potentially due to the use of landline only sampling. Whilst recent studies have shown that the exclusion of mobile only households does not (yet) significantly influence survey results , it is possible that this sampling method led to response bias, since younger people increasingly live in a mobile only household . Further, due to the under-sampling of adults aged 18–24 and 25–34 years it was necessary to combine these age groups in order to perform the analyses. Given age-related differences in several factors that may influence intervention preferences (e.g., stage of life and lifestyle) it is possible that examining these two age categories separately would yield different results. This will require further analysis. Second, the data for this study was collected in 2010. More recent data may yield different outcomes, given the rapidly changing telecommunications environment. However, our review of previous research suggests that our data reflects current intervention preference trends [10, 11]. Third, the individual characteristics explored in this study were not exhaustive. Although our study investigated a range of potential demographic, health and behavioural correlates of preference for intervention delivery mode, other potential correlates such as psychosocial factors were not explored. This may account for the small proportion (6%) of variance explained by the factors included in the combined model, since psychosocial factors typically explain 30-40% of the variance in physical activity behaviour [48–50]. Psychological factors that may be relevant to assess in future studies could include perceived physical activity barriers, preference for physical activity in groups or alone, social support and the importance placed on social interaction, as well as how each of these interacts with key socio-demographic, environmental and health factors . Finally, due to the low proportion (<1%) of people who reported a preference for telephone based counselling these were excluded from the analysis and thus we were unable to explore preference characteristics related to this delivery mode which is still currently used in research settings and by Government agencies.
Implications for practice
Despite accounting for only a small proportion of variance, the results may have implications for practice as they suggest that certain population groups, such as males and females, the young and the old, those living in regional and rural areas and those with a high body mass index may be responsive to different interventions types. However, the results also demonstrate that the preference for mediated interventions is low at a population level, especially among women, the obese, and those living in urban areas. If we are to successfully reduce the public health burden associated with physical inactivity, an increased effort to promote the desirability and accessibility of mediated approaches among these groups may be needed. Furthermore, attempts to adapt intervention strategies to the local context and target them to well-defined groups or individual characteristics may also further enhance efficacy . This study helps delineate these groups and provides some insight into what approach may be most appealing to whom.