Individuals receiving self-monitoring feedback in the behavioral weight loss intervention were active duty military personnel stationed at Lackland Air Force Base in San Antonio, Texas, enrolled in the Fit Blue weight loss study (2014-2017) [27,28]. Recruitment used posters and bulletins, on-base presentations, advertisements, and word of mouth. Those interested were phone screened by study staff to assess eligibility (ie, >1 year left on base, >18 years of age, BMI >25.0 kg/m², health care provider clearance, computer/email access). At baseline, 248 participants were randomized to either a counselor-initiated or self-paced 12-month intervention condition. Conditions varied in intensity and self-initiation required but were similar in intervention goals.

All feedback duration times were included in analyses regardless of condition. Descriptives (ie, median, mean, SD, first quartile, third quartile, quartile range) were observed for feedback duration across counselors by 12 annual quarters (ie, 3-month periods). For counselors who joined the team later (N=2), the first quarter they provided feedback was compiled with the first quarter of feedback from the original counselors. Differences in overall duration times by each consecutive annual quarter were analyzed using the Wilcoxon-Mann-Whitney test.

Counselors needed increasingly less time to review online self-monitoring data and compose personalized feedback to participants over the first 9 months the behavioral weight loss intervention was implemented, with a nonsignificant trend suggesting increased efficiency for the next 3 months as well. Thus, after 12 months, the mean amount of time spent in reviewing self-monitoring data and composing feedback decreased from 53 to 26 minutes. When examining median times, there was less variation across all quarters (ie, 25-30 minutes); however, a decrease in IQR in the first three quarters was notable. This narrowing time range indicated that counselors composed these feedback messages more consistently near the median (ie, 30 minutes) over time.

Although the standard deviation decreased, it remained high, likely due to variability across individual counselors and participant characteristics. Some participants might have logged similar data to previous weeks, or only logged one day, requiring shorter feedback messages. In response to these self-monitoring situations, counselors might write, “You continued to meet your calorie and fat goals and continue to make regular choices of fruits, vegetables, whole grains, and low- or no-calorie beverages” or “You met your calorie and fat goals on the one day that you were able to log. What were the barriers for logging on the other days?” Additionally, perhaps some counselors were quicker at crafting feedback than others.

Mean time was 32 minutes per feedback message, which is higher than the 10-15 minutes found by Hunter et al [26], during which counselors similarly provided feedback on food intake, exercise, and weight. However, only time spent providing feedback was reported, which—unlike the current study—did not include time reviewing self-monitoring data [26]. Further, it is unclear if the counselors in this previous research logged feedback duration contemporaneously or retrospectively [26].

The decrease in time for feedback messages might be influenced by multiple factors. Perhaps counselors became more efficient at reviewing self-monitoring data and constructing feedback with experience and additional training. Another possibility is that counselors provided less detailed feedback and became “sloppier” over time. However, this is less likely given that the overall intervention results indicated participants experienced significant weight loss at 4-month and 12-month outcomes [28], and periodic retraining on self-monitoring feedback was conducted to increase the likelihood of maintaining good quality feedback. Further, participants might have needed increasingly less feedback about their behaviors over the 12 months that they participated in the intervention. However, our study compared time periods specific to counselor experience, which included the three years that the 12-month intervention was implemented. Thus, individual participants cycled throughout the three years that our results were analyzed and overlapped with participants at other stages of intervention. Nonetheless, future research should rate the quality of feedback over time alongside the time required to compose it.

Regardless of increasing counselor efficiency, 25-30 minutes per feedback message is a high cost for interventions, even cost-effective interventions such as this one [24]. In order to improve dissemination of behavioral weight management programs to all who are eligible and interested, particularly in settings (eg, primary care) without individuals who may have the time, training, and supervising experience to provide self-monitoring feedback, it may be beneficial to develop strategies for decreasing the amount of time required for crafting each message. Some possibilities for reducing the costs of individualized counselor feedback would be generating computer-automated feedback or facilitating peer-group interaction to help promote self-monitoring behaviors [30]. Although computer-automated feedback was comparable to counselor feedback in the short term in previous research, automated feedback was less effective for long-term weight loss [12]. Given limited research on computer-generated feedback [12,31,32] and the common use of counselor-generated feedback [9-12], it is clearly still important to understand in detail the time costs associated with counselor feedback. These details can inform future program budgets, especially since counselor compensation is one of the largest costs [21,24,25]. However, since computer-tailored feedback might be more regularly incorporated into future interventions, a better understanding of the efficiency of counselors in crafting feedback may be beneficial to compare these different modalities.

There are several limitations to consider. Counselors self-reported time spent reviewing self-monitoring data and constructing feedback, which may be less accurate compared to objective measurements. However, this level of detail is much greater than previous studies. In addition, findings are based on lay counselors and may differ from other weight loss professionals (eg, dietitians). Finally, future studies might examine a consortium of behavioral weight management studies in order to have a larger sample size of interventionists in analyses that examine these questions.

Current findings suggest that counselors, across the initial 9 months of a behavioral weight loss intervention, become quicker at reviewing participant self-monitoring data and composing individualized feedback. Although there was individual variability, findings indicate that after 9-12 months of experience, counselors composed self-monitoring feedback more consistently in about 30 minutes. Despite indications of increased counselor efficiency, time per feedback message was only reduced to 25-30 minutes. Weight loss programs might consider testing computer-automated feedback with human tailoring to reduce counselor time.