Chronic diseases are a large and growing public health problem. Seven of the top ten leading causes of death in the United States are chronic diseases (Murphy, Kochanek, Xu, & Arias, 2015). About half of all adults in the United States has at least one chronic health condition, and many adults (especially middle to older) have multiple chronic diseases (Ward & Schiller, 2013). Approximately one-third of those 45–64 years of age and over 60% of those 65+ years of age have multiple chronic health conditions (Ward & Schiller, 2013). Obesity was not included as a chronic disease in these estimates; therefore, the prevalence of multiple chronic diseases is likely to be even higher.

Contributors to these high rates of chronic disease include unhealthy diet, physical inactivity, and tobacco use (Dietz, Douglas, & Brownson, 2016). These factors are modifiable and are commonly referred to as the leading causes of actual death (Mokdad, Marks, Stroup, & Gerberding, 2004). A number of specific dietary factors can help to prevent the development of comorbid chronic health conditions. There is evidence suggesting that increased fruit and vegetable consumption reduces the risk of diseases such as coronary heart disease, hypertension, stroke, cancer, and overweight, among others (Boeing et al., 2012). The evidence regarding lowering saturated fat intake is mixed, although it has been associated with reduced risk of coronary heart disease (Astrup et al., 2011). Additionally, increased intake of dietary fiber has been shown to reduce the risk of cardiovascular disease, diabetes, obesity, breast cancer, and colon cancer (Timm & Slavin, 2007). There is also evidence suggesting that dietary factors/behaviors can help to manage many chronic health conditions including diabetes (Pastors, Warshaw, Daly, Franz, & Kulkarni, 2002), hypertension (Appel et al., 2006), and obesity (Makris & Foster, 2011).

Unfortunately, a majority of American adults do not meet dietary guidelines (Krebs-Smith, Guenther, Subar, Kirkpatrick, & Dodd, 2010). Lifestyle interventions focusing on healthy eating may be one means for preventing chronic diseases and/or improving the health of those with chronic diseases. A number of dietary interventions focused on preventing or managing chronic diseases have been conducted, however the effectiveness of these programs on dietary adherence outcomes are mixed (Desroches et al., 2013). A majority of these interventions to date are group-based or individualized with a health professional (e.g. nurse, dietician) (Desroches et al., 2013). In an effort to combat this significant public health problem, there is a need for low-cost programs that are easily accessible and can reach a large number of people. One such intervention approach may be computer-tailored interventions, which have shown promise for improving dietary outcomes (Broekhuizen, Kroeze, van Poppel, Oenema, & Brug, 2012). Self-directed print dietary programs may also be appealing, as they require very few resources in terms of staff time and equipment, making large scale implementation feasible and potentially impactful. Unfortunately, self-directed programs that can successfully improve dietary behaviors that do not do not require internet or phone contact (Fries et al., 2005; Kristal, Curry, Shattuck, Feng, & Li, 2000) are rare..

As part of a larger trial designed to evaluate the effectiveness of a self-directed exercise program (First Step to Active Health^®) (Wilcox et al., 2015), a self-directed nutrition program (i.e., Steps to Healthy Eating) was developed to serve as the attention control condition. Results from the main trial have been published elsewhere (Wilcox et al., 2015), Briefly, participants in the exercise condition showed greater increases in physical activity than those in the nutrition group, and weight significantly decreased in the nutrition group at 9 months (~2 lbs), whereas there was no change in the exercise group. Although evaluating the effects of this program was not a main aim of the trial (Wilcox et al., 2015), the purpose of the secondary analyses conducted in this paper were to evaluate the effectiveness of the self-directed nutrition program on fruit and vegetable consumption and fat- and fiber-related behaviors among adults with chronic health conditions.

Of the 401 participants randomized, 321 had 12-week and/or 9-month data and were included in this study (80%). More women than men were retained at either time point (p=.04), and those retained also had higher fruit and vegetable consumption at baseline (p=.02). On average, participants were 56.5±10.5 years of age, had a BMI of 32.9±8.3 kg/m², and had 2.0±1.0 chronic health conditions. A majority were women (88%), White (65%), married (62%), and had at least some college education (88%). Chronic diseases were prevalent; 85% of participants were overweight or obese, and 100% reported arthritis, 49% high blood pressure, 42% high cholesterol, 13% osteoporosis, and 11% cancer; 94% had at least 2 chronic health conditions and 67% had at least 3. Table 2 shows the health-related characteristics of the sample by intervention group assignment. There were no significant baseline differences between intervention groups on any of these factors.

Participants in the nutrition group returned, on average, 11.0±2.6 (out of 12) logs. Among those returning any logs (n=162), 98.2% made it to Step 1 fruit, 92.6% made it to Step 2 vegetables, 81.5% made it to step 3 grains, and 64.1% made it to step 4 meat and beans. Among those returning a log in week 12 (n=143), 2.8% were on Step 1, 5.6% were on Step 2, 12.6% were on Step 3, and 79.0% were on Step 4.

Table 3 shows the adjusted baseline, 12-week, and 9-month means for dietary behaviors, and the within and between group effect sizes. There was a significant Group X Time interaction for fruit and vegetable consumption. There was a significant increase from baseline to 12 weeks but not baseline to 9 months for the nutrition group. There was no change at either time point for the exercise group. The nutrition group had significantly higher fruit and vegetable consumption than the exercise group at 12 weeks, but there was no difference between groups at 9 months. The effect sizes show that the magnitude of reductions were greater for the nutrition than the exercise group.

There was a significant Group x Time interaction for fat-related behaviors. There was a significant decrease from baseline to 12 weeks and baseline to 9 months for both the nutrition and exercise groups. The nutrition group had significantly lower scores than the exercise group at 12 weeks, but there was no difference between groups at 9 months. The effect sizes show that the magnitude of reductions were greater for the nutrition than the exercise group.

There was a significant Group x Time interaction for fiber-related behaviors. There was a significant decrease from baseline to 12 weeks and baseline to 9 months for both the nutrition and exercise groups. The nutrition group had significantly lower scores the exercise group at 12 weeks, but there was no difference between groups at 9 months. The effect sizes show that the magnitude of reductions were greater for the nutrition than the exercise group.

The public health burden of chronic diseases both in terms of prevalence and cost is significant (Gerteis et al., 2014). Although dietary interventions may be one means for reducing the burden of chronic diseases, self-directed dietary programs that can be widely disseminated at a low-cost, and do not do not require internet or phone contact are rare. Findings from this study show promise in that a low-cost, low-resource (i.e., staff and materials), self-directed dietary intervention that incorporates evidence-based behavioral strategies (Artinian et al., 2010) can improve dietary behaviors in adults with chronic diseases in the short-term. Although between group effect sizes were small, small changes across a large number of people have a greater public health impact than large changes across a small number of people (Dzewaltowski, Estabrooks, & Glasgow, 2004; Estabrooks & Gyurcsik, 2003).

While the short-term effects of this intervention are somewhat promising, the long-term effects are less clear. In the nutrition group specifically, fat-related behaviors were maintained at 9 months, and although fiber-related behaviors were not fully maintained, there was still a small effect at 9 months. However, increases in fruit and vegetable consumption were not maintained at 9 months. Furthermore, differences in dietary behaviors between groups no longer existed at 9 months. Although the intervention was low in intensity, participants were asked to complete and return (weekly) daily logs, and compliance was high. Logs were not required during the maintenance period of the study (i.e. after 12 weeks), and interestingly, the corresponding dietary outcomes during this time period were less impressive. The literature shows that self-monitoring is key for behavior change. A meta-regression analysis found that across 122 evaluations of physical activity and dietary change interventions, those that combined self-monitoring with at least one other technique related to self-regulation (e.g., planning, goal setting) were significantly more successful than interventions not including these techniques (Michie et al., 2009). Efforts to continue self-monitoring after the active intervention is complete (i.e. during the follow-up period) should be made, as it may improve maintenance of dietary changes.

Maintenance of behavior change, including dietary behaviors, is challenging and it is unclear how to best achieve it (Greaves et al., 2011). Many of the dietary interventions conducted to date have limited follow-up evaluation (Ammerman, Lindquist, Lohr, & Hersey, 2002; Desroches et al., 2013). Therefore, little is known about the long-term effectiveness of dietary interventions or how to prevent relapse (Ammerman et al., 2002), particularly among those with chronic health conditions. Follow-up measures, and extending the follow-up period even further (i.e., beyond 1 year) may be necessary to better understand the long-term effectiveness of these types of interventions (Ammerman et al., 2002).

This self-directed nutrition program is appealing, as it did not require following a strict diet (e.g., vegetarian/vegan, Mediterranean), fasting, or excluding certain foods. Such diets may be unrealistic from a public health perspective, as it may be difficult for the average person to understand and follow these diets long-term. This intervention was based on the U.S. dietary guidelines, incorporated behavior change strategies consistent with Social Cognitive Theory (Bandura, 1986) (i.e., self-monitoring, goal setting, enhancing self-efficacy), and was designed to be low cost and easily adapted to meet individual needs, making it more desirable and feasible for use at a population level.

Given the importance of physical activity in the prevention and management of chronic diseases (Durstine, Gordon, Wang, & Luo, 2013), public health interventions may want to consider including a dietary and a physical activity component. A self-directed exercise intervention delivered in the same manner as this intervention was shown to be effective in increasing physical activity (Wilcox et al., 2015). Interestingly, and unexpectedly, participants in the exercise group also showed improvements in fat- and fiber-related behaviors, albeit very small in magnitude, suggesting that an exercise intervention may also elicit some changes in dietary behaviors. Previous research has suggested that physical activity may be a ‘gateway behavior’ to engaging in other healthy behaviors (e.g. healthy diet) (Blakely, Dunnagan, Haynes, Moore, & Pelican, 2004; Jayawardene, Torabi, & Lohrmann, 2016). Regardless, targeting and changing both behaviors (physical activity and diet) simultaneously may result in more powerful and meaningful effects (King et al., 2013). Although the availability of such programs is the first step, efforts aimed at promoting use of these programs is essential, as evidence-based interventions focused on individuals with chronic health conditions (e.g., arthritis) known to be effective are underutilized (Hootman et al., 2012).

The findings of this study should be interpreted in the context of several limitations. Limitations include the self-reported diagnosis chronic health conditions, the use of self-reported dietary measures, which can lead to social desirability bias, and the measurement of dietary behaviors and not consumption (for fat and fiber). Our sample was mostly well-educated women who were likely more motivated to change health behaviors (i.e., they self-selected into the study). Finally, 20% of people did not have follow-up data at either time point, and no imputation methods were used.