The results clearly indicate that a video-tailored physical activity intervention is acceptable for potential users, and that developing such intervention is feasible. Furthermore, a lot of useful information was obtained to inform the future development of this type of interventions. It has often been suggested that internet-delivered interventions that are highly interactive and appealing in use will result in higher participant engagement and retention, and might also result in higher and longer term effectiveness [7, 31, 32]. Whilst to date there is no direct evidence to support that computer-tailored interventions are capable of increasing engagement and retention , they have previously shown to be successful in changing behaviour [11, 12] and incorporating this 'next generation' video enhancement (which increases their appeal), is anticipated to further increase their effectiveness. The detailed information presented in this study is entirely focussed on providing assistance with the development of these innovative interventions.
Nearly all the focus group participants supported the concept of a video-tailored intervention, whereas a much lower proportion of survey participants indicated using video would be their preferred method of delivery. This is likely due to the extensive introduction and illustration on this topic received by focus group participants, resulting in higher understanding and appreciation of the concept of video-tailoring when compared to survey participants who only received a brief explanation over the phone. Social desirability might also have contributed to focus group participants showing greater support for video-tailored interventions. However, when compared to an online survey, conducted by Marshall et al., the current support for video delivery of physical activity advice has risen to a level three times higher than what it was in 2003 . The survey, by Marshall et al., identified the preferred sources of advice on how to become more physically active and only 12% of internet users preferred receiving advice using video. As such, only 7 years later, the continued 'internet revolution' has rapidly changed user preferences, with now more than 35% of participants indicating that they would prefer video-based advice. The preference of text over video delivery by survey participants is likely a case of people being unfamiliar with a new and unknown type of intervention (video), and preferring what they know best (text). As internet speed, number of broadband connections and new websites rapidly keep evolving in a direction that is more suitable for video-delivery of content, the number of people that prefer video over text delivery will undoubtedly continue to grow. This is also reflected by having a higher proportion of participants less than 40 years of age who preferred to receive video-tailored advice (43.5%), as younger people are often setting new internet trends which are later adopted by the broader population.
A recent study indicated that those most likely to watch videos on the internet are young, male and have a broadband connection . This is in line with the current study with regards to those who are younger, however in this study more females (37.8%) indicated to prefer video-tailored messages when compared to men (33.8%). This might be due to the health related nature of the video messages, as males typically show less interest in their health when compared to females [45, 46]. In this study it was also found that participants with a slow internet connection more often preferred to receive text-based advice (54.7%) compared to participants with a fast internet connection (49.6%). Similarly, those with a fast internet connection more often preferred to receive a video-based advice (40.3%) compared to those with a slow internet connection (31.9%) (Chi2 = 632; P < 0.001; not reported in results section). In relation to this, a study conducted in 2005 showed that bandwidth constrains of 'video-rich' health behaviour change websites were too large to allow satisfactory use with dial-up modems . It is therefore encouraging that broadband connections are overtaking the marked at an astonishing rate, with currently only 10% of Australians continuing to use a dial-up connection . The survey results of this study were in line with this, indicating that for less than 20% of the sample downloading videos would be 'slow' or 'very slow'.
From these results it thus seems that many internet users are ready to receive a video-tailored physical activity intervention, and that current internet infrastructure is able to support it. However, the results from this study also indicate that it is too soon to go yet another step further and use mobile phone technology to implement this type of intervention. Even though 'smart phone' use is on the rise (according to market research, conducted in June 2010, 36% of Australians owns a smart phone and this is likely to increase to 50% within one year), the majority of both focus group and survey participants did not like the idea of using their mobile phones to receive this kind of intervention. Only 10% of survey respondents were in favour of this idea; although this number was nearly double for those under 40 years of age. The reason for the low support might result from the fact that the average age of participants in both the focus groups and the survey was relatively high (47.4 and 52.8 respectively). An American study, conducted in 2010, showed that the number of people that have watched a video on their mobile phone sharply drops with increasing age: 40% have done so in those aged 20 to 30 years, 20% in those aged 30 to 50 years, and only 6% in those aged 50 to 65 years . Thus, although it is too early to implement a video-tailored intervention though mobile phones to date, it is likely that this will change in the future.
The results from both focus groups and survey highlight a number of challenges that health promotion professionals will face when designing video-tailored interventions. Content presented by means of video is very information rich, and as such there are a lot more 'variables' that one can tailor to in order the make the advice more interesting or appealing, when compared to information that is presented in plain text format. On many occasions during the focus groups participants expressed very diverse preferences with regards to, for example, who presents the information in the videos, what approach the videos should have, at what interval new video-messages should be delivered and more. This is in line with the results from the survey which did not identify a clear preference as to whom should present the video-messages. Several focus group participants suggested that one should be offered the choice ('pick your presenter'). Whilst technically feasible, accommodating such preferences would be a huge logistical challenge, as each time a participant is offered such a choice it would double, triple or quadruple (depending on how many choices are offered) the entire database of video-messages that support the program. Unlike writing text, producing videos is difficult (focus group participants indicated that a professional production is required to be engaging and credible), time consuming and expensive. Due to a lack of evidence it is unclear whether accommodating such 'personalisation' of preferences would result in higher intervention effectiveness . However, two studies did examine this and their outcomes suggest that efforts should be made to personalise and tailor feedback on as much variables as feasibly possible. A HIV prevention study provided participants with the option to choose one of four virtual characters to guide them through the intervention, consistent with the focus group outcomes of the present study, preferences for the virtual presenters were very diverse (they were all selected by large proportions of participants) and participants responded very positive towards the intervention . Further, a study by Dijkstra et al. evaluated the effects of 'feedback, 'personalisation' and 'adaptation' in an attempt to uncover the working mechanisms of computer-tailoring, and concluded that both 'feedback' and 'personalisation' (but not 'adaptation') were effective to increase intervention effectiveness .
Another challenge is that video-messages need to be short above anything else. Over half of survey participants want the messages to be shorter than 5 minutes; 15% of men want them to be even shorter than one minute. The focus groups yielded similar outcomes with a majority of participants indicating that 5 minutes is the maximum length and that they would not watch messages that are longer. This requirement makes it hard for health professionals. Five minutes might be sufficient to communicate a message, but it is doubtful that this will be enough to change behaviour. Although not much is known about the actual exposure participants have to intervention materials, it is obvious from previous website-delivered physical activity interventions that higher exposure (often measured through number of logins into intervention website) leads to higher intervention effectiveness [7, 50]. A potential way of dealing with this challenge might be to develop multiple video messages that are reasonably short. As such, participants would be able to view the different videos available on a website at their own pace, or alternatively they could be provided with new videos at intervals set by health professionals. The focus groups revealed that participants are open to the idea receiving multiple messages at a set interval, which would allow them to receive more advice without causing an information overload.
A major strength of this study is that it combines qualitative focus group data with quantitative survey data, and as such overcomes some of the limitations that are prone to each of these research methods when used separately. For example, focus group participants were a convenience sample and selection bias might have occurred as those approached participated on a voluntary basis. Furthermore, social desirability bias may have emerged by discussing issues in a group, even though participants were encouraged to express their opinion even if it was different from others. However, representativeness issues were overcome by adding and comparing data on the same topic from a large state-wide survey which aimed to reflect the characteristics of the population taken from the most recent Australian Census. In turn, detailed and in depth information about a specific topic cannot be gathered using a large scale survey methods, this is where the focus groups show their added value, with their ability to discover hidden opinions and attitudes through discussions with peers, and their ability to gather extensive information in a relatively short period of time. Finally, one specific limitation with regards to the survey outcomes should be mentioned: a large proportion (up to 20%) of older survey respondents (over 60 years of age) did not answer the questions in relation to personalised video-tailored physical activity messages (not reported in Table 1), whereas this was minimal (generally less than 2%) in younger participants (less than 40 years of age). This is likely an indication that the oldest group of survey participants has not yet fully caught up with new aspects of internet technology, or have more difficulties understanding the concepts in the questions asked. As such, survey outcomes of the oldest group of respondents should be interpreted with caution.