Impact of COVID‐19 on a worksite weight loss program for employees with overweight and obesity

Abstract

Objective

The COVID‐19 pandemic has been shown to be negatively associated with physical activity engagement, adherence to healthy diet, and weight management among people with obesity. The current study examined COVID‐19‐related changes in weight, physical activity (PA), and diet among employees with obesity or overweight who participated in Vibrant Lives (VL), a worksite weight loss program.

Methods

School district employees participated in the 6‐month VL weight loss program and were categorized into non‐COVID‐era participants and COVID‐era participants. Participants completed questionnaires about PA and dietary intake at baseline and follow‐up. COVID‐era participants reported the effects of pandemic on their behaviors. Changes in weight, PA, and diet were compared between groups using multilevel linear mixed models and logistic regression models.

Results

A total of 266 participants (non‐COVID, n = 173; COVID, n = 93) were included. Significant weight loss (non‐COVID, −2.3 kg vs. COVID, −1.3 kg) and increases in moderate‐to‐vigorous PA minutes (non‐COVID, 48.7 min vs. COVID, 61.5 min) were observed associated with the program, but no significant differences in changes between the groups were found. Compared to non‐COVID participants, COVID participants decreased fast food consumption (p = 0.008) and increased sugar‐sweetened beverage intake (p = 0.016). Higher frequency of snacking and overeating were reported as barriers to a healthy diet.

Conclusion

The COVID‐19 pandemic was negatively associated with healthful dietary behaviors. The information obtained from participants regarding the reasons for their pandemic‐related changes in diet may help identify strategies to encourage healthier behaviors and weight management among people who have been negatively affected by the COVID‐19 pandemic.

1. INTRODUCTION

Worksite weight loss intervention programs have the potential to decrease the prevalence of obesity and obesity‐related comorbidities among employees. ¹ , ² , ³ , ⁴ Worksites are ideal settings in which to implement health‐promotion programs, in part because they commonly include robust communication and support systems for employees. ⁵ Worksite‐based educational programs that offer behavioral strategies for promoting physical activity (PA) and healthier diets have helped employees lose weight ⁶ , ⁷ and improve their diet and levels of PA. ⁸ , ⁹ Programs encouraging behavioral changes may help employees manage their weight and avoid regaining lost pounds.

After COVID‐19's emergence, many governments imposed stay‐at‐home orders (lockdowns) to limit the spread of infection. During this time, workplaces were closed and many people were restricted to engaging in only outdoor activities with social distancing or staying at home altogether. ¹⁰ Working from home may disrupt daily routines; for example, employees may decrease their occupational PA and increase their time spent in front of a computer or television (screen time). They may also lower adherence to healthy diet practices; some may increase their snacking and caloric intake owing to increased screen time ¹¹ and boredom from activity limitations. ¹² Indeed, people reported more snacking and overeating during the COVID‐19 lockdown. ¹³ Moreover, people gained weight at a rate of 1.5 kg per month during the stay‐at‐home‐order period. ¹⁴

Among individuals with obesity, COVID‐related barriers to engaging in PA and healthy eating behaviors may prevent weight loss. Previous studies highlighted the challenges adults with obesity have faced regarding weight gain and adherence to healthy behaviors for weight loss since the outbreak of the pandemic. ¹⁵ In particular, individuals with obesity have reported difficulty achieving their weight loss goals, ¹⁶ , ¹⁷ and weight loss program participants have shown a decrease in short‐term weight loss and food intake self‐monitoring ¹⁸ during the pandemic. Worksite weight loss programs for employees with overweight and obesity may have likewise been affected by COVID‐19; working from home, exercise facility closure, and difficulty in following a healthy diet may have decreased the efficacy of such programs.

To facilitate employers' providing effective weight loss programs for employees with obesity or overweight, there is a need to explore how COVID‐19 has affected healthy behaviors and to identify the barriers to weight loss, engagement in PA, following a healthy diet, and the implementation of worksite weight loss programs during the pandemic. The purpose of this study was to examine the effect of COVID‐19 and the pandemic shutdown on weight and behaviors of school district employees who participated in Vibrant Lives (VL), a multi‐year digital worksite weight loss program. Previous VL study using the objectively measured physical activity data (Fitbit) from this program found that study participant during the 2019–2020 school year had a greater increase in vigorous physical activity than those who participated in the previous year, despite a greater reduction in overall activity. ¹⁹ Here COVID‐19 related effects on weight, dietary behavior, and self‐reported physical activity were reported. It is hypothesized that, compared to people participating in VL before the pandemic, those participating during the pandemic would be less likely to engage in PA program dietary/eating behavior recommendations, and lose weight.

2. METHODS

This study was a secondary analysis of data from the VL program. The VL program is a digital weight loss program focused on behavioral changes, such as increasing PA and developing healthy eating habits using the adapted Diabetes Prevention Program. ²⁰ The VL program was offered during 3 school years, 2017–2018 (year 1), 2018–2019 (year 2), and 2019–2020 (year 3).

2.1. Participants and intervention

Participants in the VL program were employees of the Pasadena Independent School District in southeast Texas. Eligible participants had a body mass index (BMI) greater than 25 (in year 1) and greater than 27 (in years 2 and 3), signed a medical release form, and completed a baseline survey. The enrollment started in late September/early October of each school year with the intervention starting in November. Participants were divided into 2 groups based on the school year in which they participated: non‐COVID (2017–2019 school years) and COVID (2019–2020 school year). The study was approved by the Institutional Review Board of The University of Texas MD Anderson Cancer Center, and all participants gave their written informed consent before data collection.

All participants enrolled in the 6‐month VL program were educated about healthy eating behaviors and PA and were taught skills to support their efforts. The program included materials adapted from the Diabetes Prevention Program (DPP), which were weekly sent to participants through email about tips for being physically active, reducing calorie intake and healthy choice, and weight loss. The behavioral goals emphasized in the DPP included 150 min of moderate to vigorous intensity activity or 10,000 steps per day, a calorie goal of 1200–2000 calories per day and a fat gram goal of 33–55 g per day, depending on starting weight. Participants received text messages with tips for behavior changes, Fitbit activity monitors, and Wi‐fi‐connected scales (Fitbit Aria Air Smart Scale, San Francisco, CA). Participants were encouraged to use the Fitbit mobile app to monitor dietary intake and physical activity. They also participated in challenges each month from November–April in which they competed for prizes if they met the challenge objectives. Challenges included the “No gain challenge” (no weight gain from Thanksgiving to New Year's Day), the Heart Health Challenge to increase active minutes (as identified by the Fitbit) by 15%. In the 2018–2019 and 2019–2020 participants had the opportunity to engage with a private Facebook group in which weekly challenges were posted, for example, try a new vegetable, scavenger hunt, and participants were encouraged to post pictures related to the challenge.

2.2. Measures

Participants completed a survey that asked for basic demographic information such as age, race, sex, education level, and marital status. Weight was measured using Fitbit Aria Smart Scale at baseline and after the program.

Self‐reported PA and dietary intake were assessed using items from the Health Information National Trends Survey (HINTS, PA and fruit and vegetable intake) ²¹ and the Health of Houston Survey (other dietary intake questions). ²² The questions for PA and dietary intake were listed in the Table 1. For aerobic PA, a total minutes of moderate‐to‐vigorous physical activity (MVPA) during the week were calculated by multiplying the number of days per week participants did leisure‐time MVPA and the typical length of these sessions. A dichotomous variable signifying whether the participant met recommendations was calculated based on whether he/she did at least 150 min of MVPA and two sessions of muscle‐strengthening exercise per week was calculated. For dietary intake, responses were recorded according to whether participants (1) consumed red meat (≤2 times/day, score = 1; ≥3 times/day, score = 0); consumed fruit (≥1 cup/day, score = 1; <1 cup/day, score = 0); consumed vegetables (≥3 times/day, score = 1; <3 times/day, score = 0); consumed fast food (≤2 times/week, score = 1; ≥3 times/week, score = 0); ate breakfast (every day in the past week, score = 1; 0–6 days in the past week, score = 0); consumed sugar‐sweetened beverages (never in the past week, score = 1; at least once in the past week, score = 0); and read the nutritional information on a menu (often or always, score = 1; never or sometimes, score = 0).

TABLE 1.

Questions for behaviors using items from HINTS and Health of Houston Survey

	Questions
Physical activity	In a typical week, how many days do you do any physical activity or exercise of at least moderate intensity, such as brisk walking, bicycling at a regular pace, or swimming at a regular pace?
	On days that you do any physical activity or exercise of at least moderate intensity, how long do you typically do these activities? (minutes)
	In a typical week, outside of your job or work around the house, how many days do you do leisure‐time physical activities specifically designed to strengthen your muscles?
Dietary intake	During the past week, how often did you eat red meat such as beef, pork, ham, or sausage?
	How many cups of fruit (including 100% pure fruit juice) do you eat or drink each day?
	How many cups of vegetables (including 100% pure vegetable juice) do you eat or drink each day?
	How many times did you eat fast food?
	How many mornings did you have something for breakfast?
	How often did you drink sweetened drinks?
	When available, how often do you use menu information on calories in deciding what to order?

Note: The bold values emphasized what were measured in the survey questions.

Abbreviation: HINTS, Health Information National Trends Survey.

For COVID‐19 participants only, a COVID‐19 questionnaire was added at follow‐up to ask whether the COVID participants had changed their behaviors because of the pandemic. In this study, participants reported whether their food quality, snacking frequency, overeating, meal‐cooking frequency, meal‐cooking variety, and alcohol consumption had (1) decreased, (2) not changed, and (3) increased. If applicable, participants also selected the reason(s) that their dietary intake had been affected by COVID‐19.

2.3. Statistical analyses

Descriptive data at baseline were analyzed for participants who completed and dropped out as well as in the non‐COVID and COVID groups. One‐way analysis of variance tests were used for continuous variables (age and BMI), and Pearson's chi‐squared tests were used for categorical variables (i.e., race, sex, education). Multilevel repeated‐measure linear mixed models (for continuous variables) and logistic regression models (for dichotomous variables) were used to assess changes in weight, PA, and dietary intake between baseline and follow‐up (time) among the non‐COVID and COVID groups (group) and the interaction between time and group (time*group). Participants were nested within schools in the analyses. All statistical analyses were performed using STATA statistical software, version 15.1 (StataCorp LP, College Station, TX) and unstandardized beta values (β), odds ratios (ORs) with standard errors (SEs), and p‐values were reported. A p‐value of 0.05 was considered statistically significant.

3. RESULTS

Table 2 shows demographic data retrieved at baseline for the all VL participants (n = 380). Completers are based on participants who completed follow‐up assessments, while drop‐outs are based on those who did not complete follow‐up assessments. Drop‐out rates of non‐COVID and COVID groups were 27% (n = 64) and 35% (n = 50), respectively, which were not significantly different (χ = 2.692, p = 0.101). No statistical differences of baseline characteristics between completers and drop‐outs were observed. The analyses included 266 participants who completed baseline and follow‐up assessments. The participants' baseline characteristics by group are reported in Table 3. There were no statistical differences between the non‐COVID and COVID participants at baseline. Differences in outcomes between the 2017–2018 and 2018–2019 (non‐COVID) school years were tested; as there were no significant group by time interaction effects, the data from these two years were combined (data not shown).

TABLE 2.

Participants who completed and dropped out the VL program

	Completers (n = 266, 70%)	Drop‐outs (n = 114, 30%)	p‐value a
Mean age, years (SD)	43.4 (10.3)	41.9 (9.9)	0.200
Race, n (%)			0.650
Hispanic white	97 (36)	43 (38)
Non‐Hispanic white	122 (46)	57 (50)
Hispanic Black	1 (1)	1 (0)
Non‐Hispanic Black	30 (11)	8 (7)
Asian	7 (3)	1 (1)
Other	9 (3)	4 (4)
Female sex, n (%)	242 (91)	98 (86)	0.145
Education, n (%)			0.239
HS/GED or less	31 (12)	8 (7)
Tech/vocational degree	6 (2)	1 (1)
Some college	42 (16)	17 (15)
Bachelor's degree	84 (32)	45 (40)
Master's degree	94 (35)	37 (32)
Doctoral degree	9 (3)	6 (5)
Marital status, n (%)			0.914
Single	52 (20)	18 (16)
Married/cohabiting	179 (67)	79 (69)
Divorced/separated	35 (13)	17 (15)
Widowed	0 (0)	0 (0)
BMI category, n (%)			0.546
Overweight	66 (25)	25 (22)
Obese	200 (75)	89 (78)
Mean BMI, kg/m2 (SD)	35.4 (6.7)	36.3 (6.9)	0.630

Abbreviations: BMI, body mass index; GED, general educational development; HS, high school; SD, standard deviation.

^a One‐way analysis of variance tests were used for continuous variables (age and BMI) and Pearson's chi‐squared tests were used for categorical variables (race, sex, education, marital status, and BMI category).

TABLE 3.

Participants' baseline characteristics

	Total participants (n = 266) (2017–2020)	Non‐COVID group (n = 173) (2017–2019)	COVID group (n = 93) (2019–2020)	P‐value a
Mean age, years (SD)	43.4 (10.3)	43.1 (10.3)	43.9 (10.3)	0.544
Race, n (%)				0.064
Hispanic white	97 (36)	68 (39)	29 (31)
Non‐Hispanic white	122 (46)	82 (47)	40 (43)
Hispanic Black	1 (1)	1 (1)	0 (0)
Non‐Hispanic Black	30 (11)	14 (8)	16 (17)
Asian	7 (3)	5 (3)	2 (2)
Other	9 (3)	3 (2)	6 (7)
Female sex	242 (91)	158 (91)	84 (90)	0.785
Education, n (%)				0.584
HS/GED or less	31 (12)	23 (13)	8 (9)
Tech/vocational degree	6 (2)	5 (3)	1 (1)
Some college	42 (16)	28 (16)	14 (15)
Bachelor's degree	84 (32)	56 (32)	28 (30)
Master's degree	94 (35)	56 (32)	38 (41)
Doctoral degree	9 (3)	5 (3)	4 (4)
Marital status, n (%)				0.286
Single	52 (20)	35 (20)	17 (18)
Married/cohabiting	179 (67)	116 (67)	63 (68)
Divorced/separated	35 (13)	22 (13)	13 (14)
Widowed	0 (0)	0 (0)	0 (0)
BMI category, n (%)				0.346
Overweight	66 (25)	37 (21)	29 (31)
Obese	200 (75)	136 (79)	64 (69)
Mean BMI, kg/m2 (SD)	35.4 (6.7)	35.7 (6.8)	34.9 (6.6)	0.078

Abbreviations: BMI, body mass index; GED, general educational development; HS, high school; SD, standard deviation.

^a One‐way analysis of variance tests were used for continuous variables (age and BMI) and Pearson's chi‐squared tests were used for categorical variables (race, sex, education, marital status, and BMI category).

Table 4 shows weight, PA level, and dietary variables of the non‐COVID and COVID groups at baseline and follow‐up. Significant weight loss were observed (non‐COVID, −2.3 kg vs. COVID, −1.3 kg; β = −2.35; SE = 0.35; p < 0.001) and increases in MVPA minutes (non‐COVID, 48.7 min vs. COVID, 61.5 min; β = 49.10; SE = 11.63; p < 0.001) in each group, but no significant differences in the changes between the groups (weight, β = 1.08; SE = 0.59; p = 0.068; MVPA, β = 12.40; SE = 19.14; p = 0.517). Non‐COVID group showed more clinically significant weight loss (5%) compared to the COVID group (21.3% vs. 13.9%; Pearson's chi‐square test; p = 0.048). Participants in both groups reported significant reduction in red meat (≤2 times/week; OR = 1.49; SE = 0.25; p = 0.019) and fast food consumptions (≤2 times/week; OR = 2.08; SE = 0.35; p < 0.001), increases in consumptions of fruit (≥1 cup/day; OR = 2.14; SE = 0.38; p < 0.001) and vegetables (≥3 cups/day; OR = 1.71; SE = 0.38; p = 0.016), and use of menu calorie information to choose foods (always/often; OR = 3.08; SE = 0.56; p < 0.001) after the VL program. Compared to the non‐COVID group, the COVID group were more likely to limit their fast food consumption (≤2 times/week; OR = 2.19; SE = 0.65; p = 0.008) but less likely to limit their sugar‐sweetened beverage consumption (never in the past week; OR = 0.95; SE = 0.54; p = 0.016) after the program.

TABLE 4.

Weights, physical activity amounts, and dietary intakes of the non‐COVID and COVID groups at baseline and follow‐up

	Non‐COVID (2017–2019) n = 173		COVID (2019–2020) n = 93		Time	Group	Time*Group
	Baseline	Follow‐up	Baseline	Follow‐up	p‐value	p‐value	p‐value
Weight
Mean weight, kg (SD) a	96.2 (19.9)	93.9 (20.4)	95.9 (21.6)	94.6 (21.1)	<0.001	0.903	0.068
5% weight loss, n (%) b	37 (21)		13 (14)		‐	0.048	‐
Physical activity
Mean total minutes (SD) a	79.8 (83.1)	128.5 (114.0)	73.8 (98.0)	135.3 (185.5)	<0.001	0.494	0.517
MVPA ≥150 min/week, n (%) c	25 (17)	55 (32)	14 (15)	32 (34)	<0.001	0.756	0.573
Strengthening exercise ≥2 times/week, n (%) c	33 (19)	69 (40)	18 (19)	40 (43)	<0.001	0.956	0.747
Dietary intake, n (%)
Red meat (≤2 days/week) c	73 (42)	90 (52)	40 (43)	42 (45)	0.019	0.898	0.279
Fruit (≥1 cup/day) c	48 (28)	78 (45)	30 (32)	45 (48)	<0.001	0.441	0.783
Vegetables (≥3 cups/day) c	31 (18)	47 (27)	23 (25)	27 (29)	0.016	0.189	0.879
Fast food (≤2 days/week) c	95 (55)	124 (72)	40 (43)	72 (77)	<0.001	0.065	0.008
Menu calorie information used to choose foods (always/often) c	49 (28)	95 (55)	34 (37)	52 (56)	<0.001	0.168	0.263
Breakfast (every day) c	71 (41)	76 (44)	33 (35)	43 (46)	0.454	0.376	0.222
Sugar‐sweetened beverage consumption (never) c	69 (40)	63 (36)	44 (47)	13 (14)	0.378	0.243	0.016

Note: The bold values mean statistical significance of analyses.

Abbreviations: MVPA, moderate to vigorous physical activity; SD, standard deviation.

^a Repeated measure mixed models.

^b Pearson's chi‐squared test.

^c Repeated measure logistic regressions.

Participants in the COVID group also reported changes in eating behavior related to the COVID‐19 pandemic and the reasons for these changes. According to the responses, 66% of participants reported an increase in snacking frequency, 83% reported meal cooking more frequently, and 47% reported an increase in overeating, behaviors which can affect diet quality and weight loss during the pandemic (Figure 1).

FIGURE 1.

Pandemic‐related changes in dietary intake and participants' reasons for these changes.

4. DISCUSSION

The current study examined how the COVID‐19 pandemic affected changes in weight, PA, and dietary intake among school district employees with overweight and obesity who participated in a 6‐month digital weight loss program. Other studies have shown the negative effects of COVID‐19 on healthy behaviors and weight, including difficulty losing weight, decreased PA participation, and difficulty maintaining a healthy diet. ¹⁵ , ¹⁶ , ¹⁷ , ²³ In particular, adults with obesity gained weight during the pandemic, ¹⁵ , ¹⁶ , ²³ decreased PA, ¹⁶ and had unhealthy diet habits. ¹⁶ , ¹⁷ In terms of physical activity, most dietary behaviors and weight loss, our data showed improvement among school employees with overweight and obesity who participated in the VL weight loss program in both the COVID and the non‐COVID groups. The mean weight loss was lower for participants in the COVID‐19 years than for those in the non‐COVID‐19 years, but the difference between the two groups was not statistically significant. However, our data showed that clinically significant weight loss (5%) occurred more often in the non‐COVID group than in the COVID group. ²⁴ Likewise, in a recent study, significantly less weight loss was observed among the COVID‐19 cohort (2020) compared to the control cohort (2019). ¹⁸ Given that clinical weight loss may be particularly important for people with overweight and obesity for preventing comorbidities, the fact that the COVID‐19 pandemic has been shown to be negatively associated with desirable weight loss and health outcomes is especially concerning from a public health perspective. ²³

Interestingly, this study found that the COVID group showed a greater increase in the mean number of weekly self‐reported MVPA minutes and in the percentage of participants meeting MVPA and strengthening exercise recommendations, although these group differences were not statistically significant. This is consistent with our previous paper showing higher “very active minutes” as recorded by participants' Fitbits. ¹⁹ Previous studies yielded similar results, including an increase in self‐reported exercise on weekdays ²⁵ and an increase in MVPA minutes ¹⁹ during the COVID‐19 lockdown. However, most previous studies have shown decreases in MVPA related to the COVID‐19 lockdown. ²⁶ , ²⁷ , ²⁸ , ²⁹ More than half of the COVID group participants in our study reported that their aerobic MVPA and strengthening exercises either had not changed, remained the same amount of PA as before the pandemic but that they had substituted different activities, or increased overall. Participants reported that the pandemic had led them to try to get fitter to protect their health, find new resources for engaging in PA at home, and have more time and motivation to exercise. ¹⁹ A recent study found that patients who received telemedicine obesity care were about 2.5 times more likely to lose weight and increase exercise participation compared to patients who did not receive telemedicine care during the COVID‐19‐related lockdown. ³⁰ Internet‐based, digital, or telemedicine interventions may have the advantage of not being affected by the pandemic lockdown. It is possible that involvement in the Vibrant Lives weight loss program, in which evidence‐based content was delivered virtually and thus not interrupted by the COVID shutdown, helped to buffer participants against COVID‐related reductions in physical activity less healthful eating patterns.

Nevertheless, many Vibrant Lives participants in the COVID group reported struggles in following a healthy diet during lockdown. Consistent with the results of previous studies, ²⁶ , ²⁸ , ³⁰ , ³¹ , ³² , ³³ these results showed a pandemic‐related increase in snacking frequency and overeating among participants. In particular, those in the COVID group significantly increased their consumption of sugar‐sweetened beverages after the intervention. Participants may have increased their snacking and overeating during the pandemic for the same reasons that they reported changing their overall eating behaviors such as that they had more time for cooking and baking sweets, more frequent stress‐induced eating, more difficulty resisting snacking at home, and a higher likelihood of eating because they were bored. However, participants in the COVID‐19 group showed significantly less fast food consumption than those in the non‐COVID group. A similar result was found in a previous study, ²⁸ indicating that people were less likely to dine at or pick up fast food restaurants during lockdown. Although our results regarding PA related to the pandemic were mixed, most relationships of COVID‐19 with diet and eating behavior were negative, indicating that the pandemic put up more barriers to adherence to a healthy diet. These barriers may be affected by other factors, such as stress and mood. ³³ , ³⁴ More research is required to identify ways to overcome the pandemic‐related barriers to following a healthy diet and managing stress.

Our study had some limitations that could affect interpretations of its results. The changes of PA and dietary intake were subjectively assessed using surveys, which have well‐documented limitations. On the other hand, a strength of our study was its use of a participant survey, the results of which could be used to determine whether changes in weight, PA, and dietary intake are correlated with COVID‐19‐related perceived barriers and facilitators. The information obtained from participants regarding the reasons for their pandemic‐related changes in PA and diet may suggest strategies to encourage healthier behaviors and weight management among people who have been negatively affected by COVID‐19 pandemic. In addition, the VL project was implemented by digital program of multiple years, which allowed for minimal changes to programming during the lockdown, and thus this allowed to examine the impact of the COVID‐19 independent on behavioral changes and weight loss.

5. CONCLUSION

The COVID‐19 pandemic was negatively associated with health‐compromizing eating behavior among the VL program participants, and these participants showed efforts to keep healthy behaviors during lockdown. Future studies are needed to investigate and provide solutions for the physical, behavioral, and mental health problems experienced by people who have been affected by the long‐term COVID‐19 pandemic.

AUTHOR CONTRIBUTIONS

Che Young Lee and Karen M. Basen‐Engquist conceived study design. Che Young Lee conducted the data analysis and wrote the manuscript with Michael C. Robertson, Kendahl Servino, Thuan Le, Margaret Raber, Katherine Oestman, and Karen M. Basen‐Engquist. Michael C. Robertson and Thuan Le participated in data collection and management. Kendahl Servino implemented literature search. All authors were involved in writing and revising the manuscript and approved the submitted and published versions.

CONFLICT OF INTEREST

The authors declared no conflict of interest.

Lee CY, Robertson MC, Servino K, et al. Impact of COVID‐19 on a worksite weight loss program for employees with overweight and obesity. Obes Sci Pract. 2023;9(4):395‐403. 10.1002/osp4.653