Self-monitoring, weight, dietary intake, and physical activity before and after hurricane Helene among weight-loss seeking adults in South Carolina

Abstract

Objective

This study examined differences in adherence to self-monitoring and weight, dietary intake, and physical activity before and after Hurricane Helene among adults with knee replacement participating in a 12-month multicomponent weight-loss program in South Carolina, USA.

Methods

Self-monitoring included weight (smart scale), dietary intake (smartphone app), and steps and active minutes (Fitbit). Accounting for covariates, multilevel logistic regression was used to predict the likelihood of adhering to the self-monitoring behaviors in the 30 days before compared to the 30 days after the hurricane in September 2024. Multilevel linear regression with piece-wise analysis predicted differences in daily energy intake, physical activity, and body weight before and after the hurricane.

Results

Participants (n = 26) were 73.1 % female, 66.3 ± 7.1 years, body mass index (BMI) 32.4 ± 4.2 kg/m² and had spent 134 ± 108 days in the weight-loss intervention. After the hurricane, participants were significantly less likely to self-monitor across all behaviors and the longer time spent in the study further decreased the likelihood of adhering to self-weighing and physical activity monitoring post-hurricane. Steps/day increased during the post-hurricane period (b = 26.22 steps/day, 95 % CI: 6.14, 48.62) but no other significant changes in behaviors were found post-hurricane.

Conclusions

Following Hurricane Helene and widespread power outages, adults with knee replacement participating in a weight-loss intervention were less likely to self-monitor but increased steps per day. There were no significant changes in energy intake or weight post-hurricane. Participation in a multicomponent weight-loss intervention may offer some protection against detrimental effects of hurricane exposure on health behaviors.

1. Introduction

In September 2024, Hurricane Helene, a category four hurricane, swept through the southeastern United States. In South Carolina, millions of people were affected by the disaster which brought major disruptions including severe weather, property damage and power outages (South Carolina State Climatology Office, 2024). Approximately 1.2 million homes in South Carolina were without power on the day of the hurricane (Hurricane Helene, 2024), and many experienced outages of a week or more, particularly in the upstate of South Carolina where thousands of homes were damaged or destroyed (Hagen et al., 2025).

The impact of disasters may extend beyond initial physical or environmental harms to significant effects on health-related behaviors such as increased alcohol and cigarette use (Flory et al., 2009; North et al., 2011), consumption of energy dense foods (Turner-McGrievy et al., 2020), reduced physical activity and weight gain (Bell et al., 2019). The potential consequences of these behaviors on health is particularly concerning for areas like South Carolina where obesity, physical inactivity, diabetes and hypertension are already prevalent (Trust for America's health. State of obesity Report, 2023). For individuals actively engaged in health promoting behaviors (e.g., self-monitoring, physical activity or dietary restriction), the impact of a disaster may adversely affect their efforts. For instance, power outages or travel restrictions could limit access to fresh foods and the stress or demands of clean-up efforts may reduce motivation for physical activity (Pendergrast et al., 2021) or self-monitoring. However, no studies to our knowledge have explored the effects of a hurricane on the health-promoting behaviors of people enrolled in a weight-loss intervention.

Therefore, the purpose of this study was to examine differences in the adherence to self-monitoring of weight, dietary intake, and physical activity before and after Hurricane Helene among adults with knee replacement in South Carolina who were enrolled in a multicomponent weight-loss intervention. We anticipated that following Hurricane Helene, participants would have reduced self-monitoring behaviors, consumed more calories, engaged in less physical activity and increased body mass compared to before the disaster. This analysis may provide valuable information about the effects of these events on motivated individuals and inform efforts to support health-promoting behaviors in areas at risk of disasters.

2. Methods

2.1. Overview and weight-loss intervention

This retrospective secondary data analysis used data from an ongoing randomized controlled trial investigating the effectiveness of a multicomponent (diet and physical activity) weight-loss intervention in South Carolina adults who have received a knee replacement. All procedures were approved by the University of South Carolina's Institutional Review Board (2021, Pro00109459) and followed institutional guidelines for the protection of human subjects. Participants provided written informed consent before participation.

The parent study is a two-arm randomized controlled trial in which participants are randomized to a 12-month weight loss intervention or control group (NCT05190666). Participants randomized to the weight-loss intervention receive tailored goals for weight loss (7 % of starting weight), calorie intake (500-1000 kcal deficit/day) and physical activity (bi-weekly progression). Additionally, participants receive behavioral lessons and phone calls from a health coach. Coaching calls occur weekly during the first four months, bi-weekly in months five and six, and monthly for the remaining six months. Throughout the intervention, participants are encouraged to self-monitor their weight, dietary intake, and physical activity. To facilitate this, they are provided with self-monitoring tools including a smart body weight scale, access to a free online dietary tracking tool and an activity tracker. Data from the self-monitoring tools are remotely transmitted and is accessible to the study team close to real time.

2.2. Participants

Inclusion criteria for the parent study were a body mass index (BMI) between 25 and 45 kg/m², a knee replacement less than six months prior to baseline assessment, a phone or tablet with internet access; and ability to read and write in English. Exclusion criteria were any contraindications to diet or weight loss, had bilateral knee replacements or were expecting a further knee replacement within the next 18 months, had a mobility limiting comorbidity (other than knee replacement), were enrolled in a formal weight loss program, using or planning to use medical weight-loss methods (medications, bariatric surgery) or were planning to relocate out of the study locations (Midlands or Upstate South Carolina).

In this secondary analysis, participants were actively enrolled in the parent study, randomized to receive the weight-loss intervention and had self-monitored at least one of the self-monitoring behaviors (weight, dietary intake or physical activity) in the 30 days before Hurricane Helene made landfall in South Carolina (09/27/2024).

2.3. Outcomes

Height (m) and weight (kg) were measured at baseline by trained assessors using a stadiometer (Seca 213) and digital scale (Seca 869), respectively. BMI was then calculated as kg/m². For all other outcomes, only data collected in the 30 days before and after Hurricane Helene made landfall in South Carolina (8/28/24–10/26/2024) were used in this analysis.

Daily body weight, energy intake and physical activity before and after the hurricane were obtained via participant self-monitoring. Body weight (kg) was collected using a smart bodyweight scale (FitBit Aria Air), which remotely transferred data to the study's coaching interface. Energy intake (kcal/day) was reported by participants using a smartphone app (FatSecret) and accessed by the study team through FatSecret Professional. Adherence to dietary self-monitoring was defined as recording ≥1000 kcal/day. Physical activity (steps/day and active minutes i.e., moderate-vigorous physical activity/day) was collected using a wrist-worn activity tracker (FitBit Inspire 2) and accessed via the study's coaching interface. Adherence to physical activity self-monitoring was defined as having ≥1000 steps/day.

2.4. Statistical analysis

Descriptive statistics were used to describe baseline characteristics and to summarize adherence to self-monitoring and health behaviors (e.g., steps/day) using participant-level means for the 30 days before and after the hurricane.

Multilevel logistic regression models were used to predict the likelihood of adhering to the self-monitoring behaviors in the 30 days before the hurricane (reference period) compared to the 30 days after. Multilevel linear regression models with piece-wise analysis evaluated changes in daily energy intake, physical activity (active minutes and steps) and body mass during the two time periods. Active minutes, steps and energy intake were transformed (square root) to test whether the normality of residuals improved. However, despite improved normality, this did not change the model results. Therefore, to aid interpretation of results, original data was used in all models. Age, sex and BMI were included as covariates in both models. Additionally, as the level of behavioral support decreases during participation in the intervention (coaching call frequency reduces from weekly to bi-weekly to monthly), time spent in the intervention was included as an interaction term in both models to examine whether this variable moderated changes in outcomes. Due to variables being on different numeric scales, age, BMI and time in the intervention were transformed using z-score. Geographic location (Midlands or Upstate South Carolina) was considered as a covariate; however, location did not improve model fit or affect the outcomes. To account for the moderate sample size, parametric bootstrapping (1000 simulations) was applied to estimate confidence intervals for all outcomes. All analyses were conducted using R (version 4.4.3).

3. Results

Twenty-six participants residing in the Midlands (n = 17) and Upstate (n = 9) South Carolina were included (73.1 % female, aged 66.3 + 7.1 years, BMI 32.4 ± 4.2 kg/m²) and had spent an average of 134 ± 108 days in the weight-loss intervention.

3.1. Adherence to self-monitoring

The participant-level mean values for adherence to self-monitoring before and after the hurricane and the results of the regression analyses are displayed in Table 1. In the 30 days before the hurricane, participants self-monitored body weight, energy intake, and physical activity on 63.3 %, 56.6 % and 90.0 % of days, respectively. In the 30 days after the hurricane, weight was self-monitored on 53.3 % of days, energy intake on 50.0 % of days and physical activity on 83.3 % of days (Fig. 1). Participants were less likely to self-monitor weight, dietary intake and physical activity after the hurricane compared to before (shown with 95 % confidence intervals [CI], Table 1). Additionally, longer time spent in the study was associated with decreased likelihood of monitoring body weight and physical activity post-hurricane.

Table 1.

Participant-level mean values and results of multilevel linear regression models for health behaviors before and after Hurricane Helene among adults in South Carolina, 2024.

Adherence to self-monitoring - Multilevel logistic regression models
Self-monitoring tool	Time	Days adhered (max 30), mean (SD)	Fixed Effects	b	SE	95 % CI (Bootstrap)
Body Weight (Smart scale)	Before	19 (11)	After vs before	−0.77	0.20	−1.05, −0.37
	After	16 (12)	Time in study	−0.67	0.47	−1.38, −0.41
			After vs before x Time in study	−0.33	0.16	−0.67, −0.04
Dietary Intake (FatSecret)	Before	17 (11)	After vs before	−0.51	0.21	−0.62, −0.03
	After	15 (11)	Time in study	0.11	0.54	−0.91, 0.05
			After vs before x Time in study	−0.02	0.14	−0.39, 0.32
Physical Activity (FitBit)	Before	27 (6)	After vs before	−1.63	0.32	−2.12, −1.03
	After	25 (10)	Time in study	−0.47	0.62	−1.49, 0.07
			After vs before x Time in study	−0.51	0.22	−1.03, −0.11
Changes in health-related behaviors - Multilevel linear regression models with piece-wise analysis
Health-related behavior	Time	Unadjusted Means (SD)	Fixed Effects	b	SE	95 % CI (Bootstrap)
Body weight (kg)	Before	87.1 (14.5)	Before	0.00	0.02	−0.006, 0.007
	After	84.7 (13.0)	After	0.01	0.02	−0.001, 0.01
			Time in study	−1.98	1.81	−2.10, −1.85
			Before x Time in study	0.01	0.00	0.006, 0.02
			After x Time in study	0.03	0.00	0.02, 0.03
Energy intake/day (kcal)	Before	1390 (205)	Before	−4.21	1.10	−6.23, −2.03
	After	1363 (232)	After	1.17	1.10	−1.03, 3.34
			Time in study	37.45	33.88	−3.55, 71.32
			Before x Time in study	3.38	1.11	1.05, 5.51
			After x Time in study	0.26	1.15	−2.02, 2.41
Steps/day	Before	7511 (3513)	Before	−16.02	11.82	−36.45, 5.26
	After	7790 (3614)	After	26.22	12.27	6.14, 48.62
			Time in study	493.64	628.19	76.20, 899.46
			Before x Time in study	−10.51	11.20	−33.11, 11.83
			After x Time in study	−18.53	12.19	−41.65, 5.88
Active minutes/day	Before	25.1 (22.7)	Before	−0.05	0.12	−0.29, 0.19
	After	31.9 (27.4)	After	0.30	0.13	−0.08, 0.56
			Time in study	1.35	4.22	−3.39, 5.63
			Before x Time in study	−0.06	0.13	−0.32, 0.20
			After x Time in study	−0.17	0.14	−0.44, 0.11

SD = standard deviation; b = regression estimate; SE = standard error; CI = confidence interval.

Fig. 1.

Individual adherence to self-monitoring before and after Hurricane Helene among adults in South Carolina, 2024.

Foot note: Each point represents a participant's adherence to self-monitoring behaviors before (blue) and after (red) the hurricane. Black circles represent mean adherence. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

3.2. Energy intake, physical activity and body weight

The unadjusted mean values for each health-related behavior before and after the hurricane are displayed in Table 1. No significant changes in body weight occurred in the 30 days before or after the hurricane, however longer time spent in the intervention was associated with less weight loss during both time periods. Energy intake/day decreased during the 30 days before the hurricane (regression estimate [b] = −4.21, 95 % CI: −6.23, −2.03); however, longer time spent in the study was associated with less of a decrease in energy intake (b = 3.38, 95 % CI: 1.05, 5.51). No changes in energy intake/day were observed after the hurricane. Steps/day increased in the 30 days after the hurricane (b = 26.22, 95 % CI: 6.14, 48.62) which was not affected by time spent in the study.

4. Discussion

This study found that adults in South Carolina who were engaged in a weight-loss intervention were less likely to self-monitor health-related behaviors in the 30 days following Hurricane Helene as compared to before. However, participants engaged in more steps during the post-hurricane period, and weight and energy intake were maintained. Additionally, several behaviors, including the likelihood to adhere to self-monitoring were influenced by the time spent in the weight-loss intervention, with longer time in the study being associated with lower likelihood of adhering to self-monitoring after the hurricane compared to before.

Compared to before Hurricane Helene, participants were less likely to self-monitor following the hurricane, which was amplified with longer participation in the intervention. Self-monitoring may have been influenced by power outages making charging devices difficult, however the decline in adherence is consistent with literature suggesting other events that disrupt daily routines, such as holidays, have negative impacts on adherence to self-monitoring (Robertson et al., 2021; Fahey et al., 2019). Further, self-monitoring adherence tends to decline over the course of a behavioral intervention (Robertson et al., 2021; Turner-McGrievy et al., 2019), due to various inter-individual reasons including reduced motivation (Webber et al., 2010). Therefore, while self-monitoring may be susceptible to interruptions in everyday life, being in the early stage of an intervention may mitigate some of the impact.

Despite reduced adherence to self-monitoring, health related behaviors were not considerably affected as participants' weight and energy intake did not change during the post-hurricane period. Moreover, daily steps increased after the hurricane. Prior evidence on physical activity following hurricanes contrasts with these findings, including reductions in the use of bicycle/pedestrian trails (Doubleday et al., 2019) and reports of physical activity being hindered by stress, reduced motivation, and exhaustion from clean-up efforts (Pendergrast et al., 2021). Additionally, exposure to disasters has been associated with weight gain (Bell et al., 2019) and increased stress with consumption of unhealthy foods (Hill et al., 2022). It is possible that the current participants, who had spent approximately four months in the intervention, were able to maintain weight and energy intake and increase physical activity despite less self-monitoring due to habit formation and/or the support provided through the ongoing intervention. The current intervention, which combines self-monitoring with other effective behavior change techniques (i.e., goal setting, feedback, reinforcement) targets multiple constructs from social cognitive and self-determination theory (e.g., autonomy, self-efficacy, social support) (Bandura, 2004; Deci and Ryan, 2000). Hence, it may be that theory-based multicomponent interventions can offer protective effects against changes in health behaviors in the face of disasters like Helene.

This study has several limitations including a small sample size of adults with recent knee replacement, thus the findings may not be generalizable. Additionally, although bootstrapping was used to provide more robust estimates, we acknowledge that small samples can lead to spurious findings. Further, we were unable to quantify participants' exposure to the hurricane, or the level of disruption experienced. Despite these limitations, this was one of the first studies to offer novel insight into the impact of disasters on the health-related behaviors of individuals enrolled in a weight loss intervention.

Following Hurricane Helene, adults with knee replacement participating in a weight-loss intervention were less likely to self-monitor weight, energy intake and physical activity compared to pre-hurricane but increased steps/day and maintained both energy intake and weight. In areas like South Carolina where the frequency and severity of health-related diseases and disasters pose an increasing threat, efforts to minimize any detrimental effects on health-related behaviors should be encouraged. These findings suggest that participation in theory-based multicomponent weight-loss interventions may offer some protective effects on these behaviors in the aftermath of a hurricane.

CRediT authorship contribution statement

Clare Kennerley: Writing – review & editing, Writing – original draft, Visualization, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Chih-Hsiang Yang: Writing – review & editing, Visualization, Methodology, Formal analysis, Conceptualization. Jongwon Lee: Writing – review & editing, Visualization, Methodology, Formal analysis. Kailyn Horn: Writing – review & editing, Writing – original draft, Investigation. Scott Jamieson: Writing – review & editing, Writing – original draft, Investigation. Gabrielle Turner-McGrievy: Writing – review & editing, Methodology, Conceptualization. Elizabeth Regan: Writing – review & editing, Methodology, Conceptualization. Christine Pellegrini: Writing – review & editing, Writing – original draft, Supervision, Project administration, Methodology, Investigation, Funding acquisition, Conceptualization.

Funding

Research reported in this publication was supported by the National Institute On Aging of the National Institutes of Health under Award Number R01AG070004. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.