The sample used for this study was derived from WIC clients across 21 US states, who completed a wichealth.org lesson during a 34-month period in the government’s 2014-2016 fiscal years. Participating subjects chose to complete a wichealth.org lesson as a means of meeting secondary contact requirements associated with the WIC program. Data-collection protocols using wichealth.org were approved for use by the Western Michigan University’s Human Subjects Institutional Review Board. Online informed consent was provided prior to completion of the online survey.

Data utilized in this study were collected from 4460 wichealth.org uses from the system database for WIC clients who started an English-based lesson in one of the first three stages of the Transtheoretical Model (ie, precontemplation, contemplation, and preparation), completed the lesson, and viewed at least one Health eKitchen video during their session. Only the individuals who began a lesson in the early stages of readiness to change, not those in the active change stages, could have measurable progression at the time of lesson completion.

All lessons were completed using the wichealth.org website application. Data consisted of six system-collected measures (links viewed, link view time, device type, beginning stage of change, ending stage of change, and lessons completed), four profile items (ethnicity, race, pregnancy status, and age), and time at which the Health eKitchen video was viewed. The staging algorithms used to identify the beginning and ending stages were based on criteria previously used to determine stages of change and progression, which have been described in detail elsewhere [3,4].

Binary progression—irrespective of whether a subject advanced in stage of change intent—was the dependent variable used. The primary independent variable of interest was subject exposure to a Health eKitchen video prior to completion of a wichealth.org lesson. Over the past decade, program evaluations conducted every 6 months demonstrated the importance of other variables currently collected from users of wichealth.org. These include user demographic characteristics of race, ethnicity, pregnancy status, and age as well as system access characteristics including device type (fixed, such as a desktop or laptop, vs mobile), lesson type (feeding behavior focus vs other), number of internet resource links accessed (depth), and duration of all resource links used. Each of these variables was considered a possible effect modifier or confounder of the relationship of interest. Depth and duration of link use are key performance indicators for wichealth.org previously demonstrated to be strongly associated with users’ stage of change progression [24]. Lesson type was also assigned based on whether the wichealth.org lesson content focused on feeding behaviors that would be relevant to Health eKitchen videos about food preparation and recipes. Lesson type was included in the model to determine whether the impact of a video view on a lesson is associated with whether the lesson addresses content related to food preparation and recipes.

In order to control for motivation bias potentially confounding the positive relationship between lesson engagement and viewing a Health eKitchen video prior to wichealth.org lesson completion, users who chose to view a Health eKitchen video after completing a wichealth.org lesson were used as the comparison group. Similar to users who viewed a Health eKitchen video prior to their lesson, these individuals were motivated to view a video, but they viewed the video after their lesson. As users access Health eKitchen by their own choice and motivation, it is likely they are motivated to be engaged to a greater extent and subsequently more likely to progress in the stage of readiness change compared to users that opt not to access Health eKitchen at all.

A two-sample test for equality of proportions was performed to test the difference in the likelihood of progression between comparison groups. Likewise, this difference was tested for each of the covariates: Welch two-sample t tests were performed to test whether link depth and duration were associated with user progression and each of the covariates. The Welch t test adjusts the number of degrees of freedom when the variances are not equal between groups. A logistic regression was conducted to evaluate the odds of progression for users who viewed a Health eKitchen video prior to subsequent completion of a wichealth.org lesson compared to those who did not. All data were analyzed using the R statistical package [25].

Video use has been well demonstrated as a valuable and impactful tool in the online educational setting. This study evaluated the effect of video viewing on intent to change parent-child feeding behavior when the video was viewed before beginning the educational lesson. Potential for confounding caused by motivation or volunteer bias associated with learners who chose to view a Health eKitchen video prior to their lesson was minimized, as the comparison group also chose to view Health eKitchen videos but did so after completion of their educational session.

The results of this study are important, because even though online nutrition education programs have been found to have an impact on various aspects associated with nutrition behaviors among WIC clients, including belief in ability and intent to positively impact parent-child feeding, knowledge and behaviors surrounding eating breakfast, and reduction in salt intake [3-5,26,27], the specific effect of video viewing on achieving online educational objectives when viewed by learners before engaging in a lesson of interest is not well reported in the literature, especially in the WIC setting. To our knowledge, only one published study deemed the use of a training video on how to access a WIC online nutrition education program as helpful in reducing challenges with online access [5]. However, examining the usefulness of instructional video on tasks associated with logging in and negotiating a website is very different from determining the relationship between video viewing and key performance indicator engagement, which was the basis of the current study.

Findings from this investigation suggest that early stage wichealth.org users were more successful in progressing in stage of behavior change intent when the session focusing on that behavior was preceded by a Health eKitchen video view as compared to when it was followed by video viewing while controlling for other key predictors of progression. It was expected that if an effect was identified, it would be limited to lessons focusing on meal-preparation behaviors, as Health eKitchen videos would mostly be relevant to such behaviors, rather than lessons focusing on, for example, behaviors such as physical activity or dental hygiene. However, users of non-meal preparation lessons benefited to a similar extent as users of meal preparation lessons when they used Health eKitchen before completing their lesson, suggesting that the content of the video may be less important than the arousal or level of engagement that may be incited in the learner following video viewing. A previous study on the use of video and its ability to enable deep learning that touched on the subject reported that the relevance of the video with a learning objective should be carefully considered and clear to the user [28]. In contrast, the current study indicated that the use of video may have a positive impact regardless of whether it is specifically related to a behavior change focus. Here, the degree of relevance may be important. In addition, the fact that the Health eKitchen videos are related to personal health choices may be relevant enough to stimulate the user’s motivation and interest in learning more about different, yet related, topics.

Subjects chose whether they wanted to use wichealth.org as the method for completing secondary contact requirements, which limits the ability to generalize the results of this study to the entire WIC population. In addition, Health eKitchen is a separate feature that does not qualify for secondary contact credit. As such, the two groups compared in this study may differ in how they approach educational learning. Nonetheless, the use of video prior to initiating an online learning session appears to be an important strategy to employ in e-learning environment designs in order to maximize user engagement and subsequent achievement of learning objectives.

We recommend that a more rigorous study be conducted to confirm our findings and allow generalization of the results to a broad audience of WIC participants. Additional studies could focus on introducing video learning in other online education and behavior change programs used by WIC to determine if our findings are applicable to any online learning approach with WIC clients or specific to the Health eKitchen feature available within the wichealth.org nutrition education platform. The use of video prior to the use of other traditional educational formats available within WIC clinics, such as information malls and one-on-one nutrition counseling, could be explored as a means for potentially impacting client engagement. Many clinics provide kiosks or looping videos in clinic waiting areas where nutrition-related videos could be shown. Although not the same as an online format, the implication that the use of video may, to some degree, influence client engagement warrants exploration of this method of learning.

This study is useful to nutritionists and WIC staff for developing or adopting online nutrition education and behavior change interventions, especially for WIC populations. This study showed that viewing a video prior to lesson completion was positively related to subject engagement in the online educational session.