Participants in an online weight loss program can improve diet quality during weight loss: a randomized controlled trial

Obesity is a major public health problem that has reached epidemic proportions worldwide [1]. In Australia, the prevalence of overweight and obesity has dramatically increased over the past two decades [2]. In the 2011–12 National Health Survey, 70% of adult men and 56% of adult women were assessed as being overweight or obese [3]. Excess body weight increases chronic disease risk, including type 2 diabetes, gall bladder disease, heart disease and some cancers, such as cancer of the kidney, colon, and breast [4]. As a result, the prevention of excess weight gain has become a national health priority in Australia [2]. Achieving a modest weight loss of 5% of initial body weight is associated with important health benefits [5], including prevention or improved control of type 2 diabetes [6–9], reduced cardiovascular disease risk (CVD) [6, 10] and improvements in other existing health conditions [5], including kidney disease [11] and sleep apnea [12].

Diet quality may influence the development of overweight and obesity. Diet quality refers to the nutritional adequacy of a diet [13], measured by evaluating how closely food patterns align with national dietary guidelines and how varied healthy food choices are within core nutrient-dense food groups [14]. A recent systematic review of prospective cohort studies found a strong association between poor diet quality and greater weight gain [15] in both men and women. For example, in a cohort of 2,245 adult men and women, those who achieved a lower Diet Quality Index gained more weight over eight years compared to those with higher scores [16]. In addition, several studies have found that higher diet quality is inversely related to chronic disease risk, all-cause mortality and cause-specific mortality [17, 18], including CVD and cancer risk.

The role that modifying diet quality plays in assisting overweight and obese individuals to lose weight and maintain weight loss is less clear. To our knowledge, only one previous randomized controlled trial (RCT) has evaluated diet quality changes during a weight loss intervention [19]. The study used the Healthy Eating Index-2005 (HEI-2005) and found the diet quality of women (n = 66) significantly improved (HEI-2005 score 53.9 vs. 57.4, P = 0.002) from baseline to post-intervention after a 16-week behavioral weight loss program. Furthermore, participants with a weight loss of ≥5% had a significantly greater improvement in their diet quality score compared to those with <5% weight loss [19]. This study was a secondary analysis of an RCT with participants randomized to one of two behavioral weight loss programs. The authors did not compare diet quality between the two groups. Data from both groups were combined for the analysis as no significant difference in weight loss was demonstrated between the two treatment groups. Consequently, it is not known whether a weight loss intervention improves the quality of dietary intake relative to controls, nor if different approaches to delivery of weight loss interventions influence diet quality.

The primary aim of this study was to compare changes in diet quality, assessed using the Australian Recommended Food Score (ARFS) [15], in overweight and obese adults randomized to a basic or enhanced version of a commercial web-based weight loss program or a wait-list control group for 12 weeks. The secondary aim was to determine whether there was an association between the extent of weight loss and diet quality at 12 weeks. It was hypothesized that diet quality would improve during the weight loss program in intervention participants and that a higher diet quality score at 12 weeks would be associated with a greater percentage weight loss. We also hypothesized that the enhanced intervention group receiving specific feedback on dietary intake would have greater improvements in diet quality compared to the basic intervention and control groups.

The primary aim of this study was to compare changes in diet quality in an online RCT with two online treatment arms of varying intensity during a 12-week web-based weight loss program. A secondary aim was to determine whether there was an association between diet quality score at 12 weeks and percentage weight loss. The change in overall diet quality in the enhanced group was significantly greater than the control group. However, there were no significant differences between the enhanced and basic, or basic and control groups. Importantly, improvements in diet quality scores were associated with a greater percentage weight loss, regardless of treatment group.

No significant differences were found in change in ARFS subscales from baseline to 12 weeks between study groups, however two ARFS subscales improved significantly within all study groups over this time. Improvements in the fruit, meat, wholegrain and water subscales were seen in the enhanced group, while the meat and water subscales improved in both the basic and control groups. There was no improvement in the vegetable subscale. It was hypothesized that improvements in diet quality in the enhanced group would be due to the increased variety in fruit and vegetables, as a result of the additional feedback the group received regarding consumption of these foods. Similar results were found in Webber & Lee’s (2011) study, where the increase in fruit score was significant (P < 0.05) and the increase in vegetable score was not [23]. A study by Booth et al. (2008) suggested that different food types might have different adjustment periods. For example, improvement in fruit intake can be achieved fairly quickly, compared to an increase in vegetable intake which may require a longer time period to achieve [24]. This suggests that vegetable intake may be more difficult to improve in a 12-week study. Additional strategies other than feedback may be required to improve the variety of vegetable intake.

The present study found that higher diet quality scores at 12 weeks were associated with greater percentage weight loss (as diet quality scores increased, percentage weight loss increased). This is important, as previous studies examining the relationship between diet quality and weight change have primarily focused on weight gain [25–27]. Also, a focus on improvement in diet quality can be incorporated within dietary intervention goals. Our findings are consistent with the study by Webber & Lee (2011), which found that participants with a weight loss of ≥5% had a significantly greater improvement in diet quality score compared to those with <5% weight loss [23].

From the current study findings, weight loss increased by an average of 1% of body weight when the ARFS increased by ten points (or 0.1% for each one point increase). In simplified terms, an increase of one ARFS point equates to consuming one new/different food at least once per week. Also we demonstrated for some ARFS sub-scales (e.g. meat, wholegrain and water) that a one-point increase in score increased weight loss (0.7% meat, 0.4% wholegrain and 1.7% water). At a population level, a 0.1% decrease in weight may be important. We have demonstrated that an enhanced web-based weight loss intervention can significantly improve diet quality (i.e. mean increase in ARFS of 2.2 in 12 weeks). Therefore, further investigation of dietary and/or weight loss interventions that can be implemented at the population level to improve diet quality are warranted.

Analysis of the association between ARFS subscales at 12-weeks and percentage weight loss found several significant positive associations, including fruit, meat, wholegrain, dairy and water subscales. This suggests that as participants consumed a greater variety of foods from within the fruit, meat, wholegrain, dairy or water sub-scales on a regular basis (i.e. ≥ once per week) the amount of weight loss increases, independent to their total energy intake. For example, they may have increased their intake of breakfast cereals, oats, rice, wholegrain bread, noodles or pasta to ≥ once per week each to increase their score in the wholegrain subscale. Several studies have found higher intakes of these foods are associated with weight loss [28–39]. However, these studies usually measure quantity rather than variety of consumption. Studies found that higher intakes of fruit, protein and high-fiber breads and cereals were associated with greater weight loss [28–30, 39]. Higher intakes of fruit and wholegrain were also shown to be a significant predictor of weight loss [31, 32]. Several studies suggest there is an association between higher dairy intake and greater weight loss [33–35, 38], however some studies are inconclusive [35, 40] and others have found that high intakes of dairy to lead to weight gain [41, 42]. Studies have also shown that increasing water consumption during weight loss interventions can lead to greater weight loss [43, 44]. Therefore, it may be effective to focus on these particular food groups during weight loss interventions to facilitate greater weight loss. Further studies examining the association between these food groups, in the context of diet quality and weight loss during weight loss interventions are needed to support these findings. Consumption of a wide variety of nutrient-dense foods, thereby improving diet quality, may lead to greater weight loss. Using a brief tool such as the ARFS to measure diet quality would carry a lower respondent burden than completing an FFQ. Diet quality tools only include foods that are consistent with national dietary guidelines, so have a smaller number of items than an FFQ, and fewer frequency categories [15].

A limitation of this study is the use of a self-reported tool to assess dietary intake. There is also potential bias from the training effect of the FFQ and under-reporting, common among overweight and obese individuals [45, 46]. In addition, participation in the intervention may have influenced dietary recall, making participants more aware of their eating patterns. Another limitation is the potential low sensitivity for scoring given that having a recommended food once per week adds one point to the total score in the same way as having the same food three or more times per week [15]. Moreover, the ARFS does not reflect energy-dense nutrient-poor foods. However, the FFQ and the ARFS are both validated reliable tools to assess dietary intakes in Australian adults [19].

Another limitation is in respect to generalizability. The study sample included a higher proportion of high-income earners (those with a weekly household income > AUS$1500 were regarded as having ‘high incomes’ at baseline [47]), university educated individuals, and those born in Australia, compared to the Australian population. Therefore, the results may not be applicable to other ethnic groups and those of lower-socioeconomic status. However, the study recruited almost equal proportions of male and females, which is unique, as the majority of weight loss trials recruit predominantly females [48]. Also, this was a secondary analysis of an RCT powered to detect differences in BMI change between groups. Therefore, the magnitude and significance of the difference between the groups of the measures in this analysis may have been limited by the sample size. Consequently, additional investigation is required with a larger sample size, powered to detect a significant difference in diet quality. The main strength of the study is that this is only the second study to explore changes in diet quality as a result of participating in a weight loss intervention. Additional strengths include the RCT design and use of ITT analysis using BOCF.

Significant improvements in diet quality were detected in the enhanced intervention group, from baseline to 12 weeks, compared to control. In addition, higher diet quality scores at 12 weeks were associated with greater percentage weight loss. However, despite improvements, diet quality scores remained low in the enhanced group. This study highlights the importance of addressing diet quality in weight loss interventions and that a brief diet tool can be applied to measure diet quality as part of a weight loss intervention. The scoring method of a diet quality index suggests it is a simple tool to evaluate diet in the context of weight loss and identify areas for improvement.