Survey design

The COVID-19 first emerged in China in December 2019. Nationwide stay-at-home orders were in effect by late January and early February, 2020 [20]. In April, an online survey was conducted in China as well as in the United States of America (USA) (stay-at-home directives were issued during mid-March in the USA). The survey requested no identifier items, such as name, address, etc., and involved no medical records or archived samples at all. The study was deemed exempt from requiring human subjects approval by the Institutional Review Board of the University of Pennsylvania. Moreover, potential participants must be 18 years of age or older (self-reported). Once started, they were given the flexibility of skipping any questions or discontinuing at any time.

The survey’s main purpose was to investigate how the COVID-19 pandemic and the stay-at-home directives affected matters relating to food at the household level. The survey included 53 questions (single or multiple responses, numerical or descriptive choices) designed to obtain data on demographics of the participants (e.g., age, number of weeks quarantined at home); food sourcing and vigilance at the market-consumer interface during the quarantine, and parameters reflecting food wastage (food purchasing, at-home food related activities, and food-use behavior), and other relevant household food information such as body weight change (S1 Table).

The questionnaire, initially developed in English, was translated into Chinese. The latter was nearly identical in content to the English version except for a few items that were adapted to be applicable to a Chinese situation, for example, race/ethnicity choices, household income range (in Chinese Yuan instead of USD$), height (in meters instead of feet), and weight (in kilograms instead of pounds). Additionally, the Chinese version contained an extra item in the question of food sourcing once it became evident that the use of the CoGGO food-acquiring mechanism was rapidly spreading throughout China.

Survey distribution

After pilot testing, both the English and Chinese versions of the survey were placed on the same platform (Qualtrics, a web-based survey platform, www.qualtrics.com, United States of America).The distribution of the survey to the public was achieved through individual as well as institutional networks and snowballing via social media. In China, dissemination of the survey was conducted through social media such as QQ and WeChat. In particular, a number of university faculty members participated in the distribution of the survey through a network of former and current students, who scattered around the country while staying-at-home due to the pandemic.

The collection of data online took place from 17 to 27 in April, 2020 (U.S.) and April 22 to May 5, 2020 (China). A comprehensive analysis of the survey data with cross-national comparisons has been published [2]. The present study focuses on two aspects of survey findings that are unique with the Chinese participants that had not been reported elsewhere. First, we identify the community-based online grocery group-ordering as one-of-its-kind food sourcing mechanism that mushroomed in China during the pandemic. Second, we examine how different age groups behaved two-month into the quarantine regarding food attitude and behavior. We discuss broader implications of these findings. It is also necessary to note that the current study is based on a larger pool of Chinese participants (totaling 2126) than that used in the previous report [2] totaling 1732, because of extended online data collection (April 22 to 27 in previous report, April 22 to May 5 in current study).

Age grouping

Unlike the USA cohort that had a more-or-less even distribution in age groups [2], the Chinese participants were skewed towards a younger demographic. For example, only 2.1% of the 2,126 participants were 60 years of age or older. A recent survey by Zhao et al. investigating dietary diversity in China during the COVID-19 pandemic [21] found a similar phenomenon in which only 16.4% of the participants (n = 1,938) were older than 45 years of age. Despite potential limitations associated with the method of convenience sampling via online survey, meaningful observations can be derived from the given cohort.

The 2,126 participants were divided into three age groups described below. Considerations for the grouping included factors such as psychological maturity, socioeconomic status, and typical food roles based on Chinese culture and family dynamics. The grouping chosen was similar to Bilska et al. [13] whose pre-COVID-19 study investigated the influences of certain demographic factors (age, gender, education etc.) on food waste and food management behaviors among Polish consumers. One exception is that the current study did not separately analyze the responses from 45–59 or 60 and older groups as Bilska et al. Our considerations are discussed below (in Group III).

Group Ⅰ (18–24 years of age, n = 796). Survey participants in this group were most likely to be single, in most cases still attending college or graduate school, and not yet financially independent. The food behavior of this age group was generally affected by greater spontaneity, an alignment towards convenience, and limited experience in food handling or dealing with trade-offs between food consumption priorities and food management [22]. In particular, survey respondents of this age group were most likely staying at home with their parents because the outbreak of the COVID-19 pandemic coincided with the beginning of Chinese New Year, which families traditionally celebrate together. However, their food attitudes and behaviors have a great impact on the family food consumption [23]. For example, the food preferences of these young people while staying at home with families affect the purchasing activities of family food buyers significantly [24], including food category, quantity etc. And these people was easily to waste food more frequently and in larger quantities [25], which has led to a serious food waste in the household level.

Group Ⅱ (25–34 years of age, n = 789). This was a group containing individuals transitioning into the workforce and general society, gradually gaining financial independence, with a growing sense of responsibility.

Group Ⅲ (≥35 years of age, n = 541). In Chinese society, people in their 30s are typically married, with many already young parents. People in this age group were generally financially self-sufficient with a steady income. In particular, these participants would have had an expanding sense of responsibility for their young families, and to a certain extent, their parents. In essence, these people tended to be mature and responsible. Unlike Bilska et al. who had two older groups (45–59 years of age; 60 and older) [13], we did not separate them but included them together in Group III. This was based on two considerations. First, many Chinese families have the tradition of 3-in-1 unity when it comes to daily food matters, with grandparents taking care of food chores (shopping, preparation, cooking) for their busy-working sons or daughters (typically in their 30’s) and grandchildren (typically preschoolers). Our data showed that even the people over 60 are grouped independently, the results were not much difference with the current three age groups, and does not affect the main conclusions of this article (S1 Fig). Second, number of survey respondents older than 45 are small (total 255, or 12% of the cohort). Separating them from Goup III would skew the size of groups, which would undermine the precision and significance of statistical analysis.

Data analysis

Raw data in Qualtrics were exported to SPSS (IBM SPSS Statistics Version 22) for analyses. Descriptive analyses included calculation of means [with 95% confidence interval (95% CI)], standard deviation, median, interquartile range (IQR) of continuous variables and list of categorical variables. Frequency counts and percentages were used for statistical analyses of categorical variables. Chi-square tests were used to compare categorical variables. Analysis of Variance (ANOVA) was used to test continuous variables and for statistical analysis of significance among variables. P<0.05 was set as the standard for statistical significance.

General description

Of the 2,126 respondents, age ranged from 18 to 82 years with a median of 27. A total of 94.4% respondents reported following the stay-at-home orders and having stayed at home for 7–12 weeks (median 10 weeks) at the time of the survey. The remaining 5.6% were presumably in the cadre of emergency or essential personnel with various service or administrative duties. The number of people per household had a median of 4 and a 25-75th percentile of 3 to 5. Table 1 shows a statistical breakdown of the gender, ethnicity, and other aspects of the survey participants.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Demographic information of survey participants.

https://doi.org/10.1371/journal.pone.0260244.t001

It is worth noting that the number of responses to individual questions varied, in some cases considerably lower than the grand total (2126) of survey respondents. This is mainly because the survey allowed participants to skip questions or to exit at any time. In particularly, two questions (gender and race) had the lowest responses (26–27% lower than the grand total). This is possibly related to their ‘positioning’, since these two questions were placed in the last part of the questionnaire, making it easy to miss or skip.

Food sourcing

Under the influence of the COVID-19 pandemic, the sources of household food purchases have undergone dramatic changes. There was a sharp increase in online food purchases (Fig 1A). More than half of all respondents reported online grocery shopping for delivery, either for the first time or with a higher frequency (Fig 1A). This shift is not surprising given the unprecedented extent of the health threat and uncertainty posed by the COVID-19 pandemic. Increased online grocery shopping during the pandemic has been reported elsewhere (e.g. [1–5, 26]).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Household food sourcing change during the pandemic compared with before of the whole respondents (a) and different age group (b, p<0.001).

https://doi.org/10.1371/journal.pone.0260244.g001

Notably, there were differences in specific online shopping preferences among people of different ages (p<0.001, Fig 1B). Of the respondents in Group III, 66.3% (N = 404) reported purchasing food online for delivery, either for the first time or with increased frequency. In comparison, this food sourcing mode appeared to be proportionally less popular with Group I and II (48.8%, N = 617, and 57.5%, N = 640, respectively). This may be explained that the younger the group, the less responsibility of food shopping in the family: 24.8%, 42.0%, and 56.0% for Groups I, II, and III, respectively (S2 Fig). This also indicated that Group III quickly adopted online food shopping since this group shouldered more responsibility for providing food in households (S2 Fig). People in this group are likely to retain online food purchasing practices in the future, i.e., post-pandemic. Online food purchase here refers to the purchase of food for family consumption through Internet-based e-commerce platforms [27] such as Walmart App, including fresh food, cooked food, etc. A special form of Internet-supported food sourcing during the pandemic, CoGGO, is described later.

On the other hand, the younger age groups showed a stronger preference for community-based online group grocery-ordering (CoGGO, see more details below), with 50.9% (N = 617) of Group I reporting purchasing food especially fresh food via CoGGO compared to 32.4% (N = 404) in Group III. CoGGO is a specific food sourcing format that enables community-based group ordering for all type of foods (fresh or dry, processed or un-processed, plant- or animal-source food) via internet based platform such as WeChat or other Apps. The younger generation has a greater acceptance for novel ideas [6]. They are more adept and adaptable to this new household food sourcing method. It would be of interest to determine whether CoGGO attracted yet greater participation in the later phase of the pandemic and the relevant demographic differentiations.

CoGGO as a special form of online-based food acquisition mechanism was reported by 42% (N = 1661) of survey respondents (Fig 1A). To our knowledge, this mechanism of community-based food sourcing for households has not been reported elsewhere. Fig 2 shows the framework and basic function of CoGGO. Under CoGGO, residents in a community or neighborhood form an online group, with a group leader (usually the individual who initiated the community group) acting as the liaison between the group and grocery supplier(s). The digital platform is provided via WeChat, a Chinese mobile phone app that supports the logistics of individualized browsing and selection with a built-in digital wallet to facilitate monetary transaction. For each ‘sale’ event (once or multiple times a week, depending on group size, demands, and food flow), the group leader would share the food inventory with group members, gather orders within a set time frame, and send the order to the supplier for fulfillment, all via the WeChat platform. A group order was usually fulfilled and delivered within a day or two. Delivery was typically to the entrance of the gated community (which is common in Chinese cities) for retrieval by the individuals.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. A schematic illustration of CoGGO (community-based online grocery group-ordering) that mushroomed in China during COVID-19 pandemic.

https://doi.org/10.1371/journal.pone.0260244.g002

Internet-based group ordering as a form of e-commerce emerged in China prior to the pandemic [27, 28], although pre-pandemic orders tended to mainly be for non-food items or non-perishable foods. The onset of the pandemic quarantine spurred the rapid adoption of CoGGO by households as a new and safer method to obtain food. The adoption of CoGGO apparently spread throughout China as survey participants represented 30 out of 34 provincial-level administrative regions. Interestingly, a recent report from an industry-sponsored survey provided additional evidence of the rapid growth of CoGGO in China, with fruits, vegetables, meats, and eggs topping online food sales (https://www.iimedia.cn/c400/74181.html).

The potential implications of CoGGO as an alternative mechanism for household food sourcing during the pandemic are multifaceted. Grocery stores in Chinese cities are oftentimes tightly spaced and overly crowded. CoGGO provided a safer alternative approach for households to meet their food requirements while cutting down potential viral exposure, thereby helping to keep viral infections down. Furthermore, CoGGO events helped to cultivate a sense of community connection among participating members (oftentimes neighbors) at a time of physical isolation and psychological stress. The “buying group” model described by Si et al. as organized by residents in Wuhan had a similar effect [6]. Further research is needed to address several issues, such as how CoGGO has evolved post-pandemic, how it may affect food distribution efficiencies as well as potential carbon footprints, what long-term impacts it may have pertaining to food system resilience, food access and security at the household level, as well as food safety and food justice.

Food vigilance

36.5% of the respondents reported that the frequency of food shopping trips prior to the pandemic was once a week on average, 34.4% of them was 2–3 times a week, 14.4% of them was 4 or more times a week, and 15.0% reporting a frequency of less than weekly (Fig 3A, N = 1749). However, survey participants reported far fewer food shopping trips to marketplaces during the pandemic (Fig 3B). Participants compensated for the reduced number of physical trips to the grocery store by increasing online-based food acquisition (Fig 1B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3.

Frequency of grocery shopping trips before the pandemic (a), changes during the home quarantine (b), and age groups differed in cutting down grocery shopping trips (c, p<0.001) during the COVID-19 pandemic.

https://doi.org/10.1371/journal.pone.0260244.g003

In terms of age difference, there was no difference in the frequency of food shopping prior to the pandemic (S3 Fig). In contrast, during the home quarantine, a higher proportion (59.6%, N = 431) of respondents in Group III reported “much fewer” grocery shopping trips compared to Groups I and II of 41.3% (N = 651) and 47.5% (N = 668) lower, respectively (Fig 3C). Those reporting “much fewer” plus those reporting “somewhat fewer” grocery trips accounted for 81.7% (N = 431) in Group III, compared to 71%–72% in the younger groups. This result can be attributed to the older age group exhibiting a higher propensity to protect themselves and families from potential viral-exposure risks.

Participants also compensated for reducing physical trips to grocery stores by purchasing extra food compared to before the pandemic (Table 2). The five food categories in which this pattern was most prevalent were vegetables, meat/fish/eggs, rice/flour/dried beans, fruits, and milk and dairy, and this pattern was consistent across all age groups. Not surprisingly, 46.6%–56.4% of respondents in Group III reported purchasing extra food within these five food categories, whereas substantially lower percentages were reported in Group I (29.9%–44.0%), and that in Group II falling between the two extremes. This result is an additional indication that Group III demonstrated a heightened sense of responsibility for household food security, consistent with Chinese culture and family structure. It should be noted that the particular question in the questionnaire relating to extra food purchases was descriptive/qualitative rather than quantitative. Therefore, it could not be determined whether any of the reported extra purchases may have constituted stockpiling or hoarding. It is assumed that this was not the case since the Chinese respondents gave a higher rating to food supply availability and diversity at the market-consumer interface compared to the respondents in the USA survey [2].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Percentage of survey participants reporting extra food purchases during the COVID-19 home quarantine compared to before the pandemic^a.

https://doi.org/10.1371/journal.pone.0260244.t002

Food waste reduction

Roughly 40.1% (N = 1612) of all respondents reported less food wastage during the home quarantine compared to prior to the pandemic (Fig 4). A greater proportion (47.2%, N = 398) of respondents in Group III reported less food wastage than the other two groups (37.6%, N = 598 and37.8%, N = 616) (Fig 4). This result was consistent with that of several pre-pandemic studies [9–13]. The three reasons given most often in the current study among all age groups for wasting less food were: (1) a greater appreciation of food and awareness of a potential shortage; (2) more time for meal planning and cooking; (3) avoidance of needing to visit the grocery store (Table 3). Respondents in Group I (53.5%, N = 591) indicated that they had more time for meal planning and cooking, thereby leading to less food waste. Within Group III, respondents gave the avoidance of needing to visit the store as the most important reason for reducing food wastage (Table 3). There were significant differences in attitudes to food between the different age groups (p = 0.037). Although motivations for reducing food waste reduction may have diminished as the home quarantine was lifted in China, it remains unclear whether this resulted in a return to pre-pandemic food waste levels. Nevertheless, most of GroupⅠrespondents (51.2%, N = 598) indicated an awareness of food wastage issues, and nearly one-third (32.8%, N = 591) reported being more creative in using food items during the lockdown. It is possible that the positive impact of the pandemic on food behavior (e.g., wasting less food) may persist beyond the duration of the lockdown. It is important to develop effective interventions to extend the positive changes that took place during the pandemic and to engage the public to encourage sustainable food behaviors.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Wasted food amount change from all respondents (Solid triangle) and different age groups (bar) during the pandemic (p = 0.006).

https://doi.org/10.1371/journal.pone.0260244.g004

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Reasons, categories and behaviors for discarding food during COVID-19 quarantine, as reported by different age groups.

https://doi.org/10.1371/journal.pone.0260244.t003

The responses of participants to a question relating to the types of food thrown away prior to the survey (i.e., as remembered most accurately) were most often spoiled fruits and vegetables, followed by leftovers of home cooking, and finally food past the expiration date (Table 3). Interestingly, most respondents (59.1%–75.2%) reported checking the expiration date of food items before deciding on whether to discard them, suggesting the critical importance of date-labeling in influencing consumer food waste behavior, even during times of crisis. Respondents also reported using sensory-based judgments (e.g., smelling if the food is still okay) to decide whether to dispose of food items. Furthermore, respondents commonly reported cutting off spoiled parts of fresh fruit and vegetables or creating a new dish from leftover food (Table 3).

Positive behavioral changes including reductions in food wastage during the COVID-19 pandemic have been reported more broadly. For instance, Jribi et al. found that most of the respondents in a small-scale survey (n = 284) in Tunisia reported having food conservation strategies in place, with nothing of what they bought being discarded [18]. In an Italian survey [29] found that most of the surveyed 1,078 households reported a reduction in food waste during the pandemic lockdown. Positive food attitudes with more prudent use of food and less wastage were attributed to heightened awareness and uncertainties people were facing regarding the trajectory of COVID-19 [2]. Regardless of the pandemic, food waste is a serious challenge undermining global food security and sustainability [30], and therefore reducing food loss and waste throughout the food chain, particularly at the consumer level, is paramount [31]. In addition, finding ways to sustain the positive changes and to broaden relevant reach is vitally important within the pursuit of a resilient and sustainable food future.

Weight change

After being housebound for about two months with limited physical space and restricted daily activities, self-reported weight-change differed between age groups (Fig 5). Notably, those reporting weight gain of >4.5 kg (clinically significant weight gain) accounted for 12.4% (N = 581) in Group I, compared to 9.3% (N = 601) and 8.0 (N = 376) in Groups II and III, respectively. Respondents reporting >2.3 kg weight gain (i.e. the two highest categories added together) accounted for 27.9% in Group Ⅰ, 23.6% in Group Ⅱ, and 19.1% in Group Ⅲ. The greatest proportion of all age groups reported no weight change (42.0%, 48.4%, and 49.7% for Group I, II, and III, respectively). Only a small percentage in each group indicating weight loss, mostly in the <2.3 kg category (Fig 5).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Body weight change during the COVID-19 pandemic in different age groups (p = 0.018).

https://doi.org/10.1371/journal.pone.0260244.g005

There have also been reports of pandemic-induced changes to weight more broadly. For example, Ammar et al. in a survey of 1,047 participants in Africa, Asia, and Europe identified that home quarantine negatively affected food consumption patterns, with overeating was reported [19]. In a survey conducted in Italy, 19.5% of participants reported gaining weight and 42% indicated greater intake of comfort food (chocolate, dessert, ice cream, etc.), which was attributed to higher anxiety [17]. In our previous report comparing U.S. vs. Chinese cohorts [2], the Chinese cohort (n = 1,732) had 11% self-reporting >4.5 kg weight gain, compared to just ~2% of the USA cohort (n = 1,547). Weight change may be related to alteration in food purchasing and eating patterns etc. [2, 17], which could reflect the impact caused by the household food consumption behavior changes during the pandemic to some extent. Apparently, the younger Chinese participants were more prone to weight gain under the same conditions of home quarantine as their older counterparts. This may be explained as the young people might have enjoyed delicious home-made meals while staying with parents during the quarantine. They may also be more resilient to stress or life-changing events under the given circumstances in the extend-family structure in China. This is mainly because they are not financially independent, therefore their senses of financial responsibilities and worries are yet to fully develop. The implications of such resilience as well as the susceptibility to weight-gain and long-term health implications, if any, would be of interest for future research.

Other food consumption behaviors

Additionally, the age groups differed significantly on the following issues: younger people were more optimistic pertaining to perceived food security and food supply stability locally or nationally; the older ones indicated more frequent home-cooked meals as well as household members cooking/eating together more often; the older group were also more in charge for household food matters such as food handling and cooking (S2, S4 and S5 Figs). On the other hand, there were no differences between age groups regarding: (1) changes (increase or decrease) in types of food consumed during vs. before the pandemic, which were consistent across the three groups; (2) most respondents regardless of age group expressed that the COVID-19 pandemic would likely influence their food behavior in the future concerning food purchase or food handling; (3) food waste disposal was primarily via garbage services, alternatives such as composting was minor.

Considered together, the results of this survey indicated that people adjusted and changed their food consumption behaviors while responding to the impact of the pandemic, with these changes differing among different age groups. Understanding these changes and following up on the impact of the pandemic is important to enhance the resilience of the food system to shock events. Key points of knowledge and implications can be derived from the survey results, including: (1) The pandemic-related lockdown accelerated the development of online food shopping, and particularly in the use of CoGGO as a new household food-sourcing for households. Going forward, the regulation of the new online food supply methods and the balance between online and traditional food supply modes will present a challenge to policy makers and business leaders. (2) In this study, under the dual pressure of an increased demand for food at home and a substantial reduction in the availability of in-store food during the pandemic, Group III respondents rapidly adopted online-based food sourcing methods, including CoGGO. Future development of food supply channels/mechanisms to target the group of consumers shouldering most of household food-purchasing responsibilities, i.e., the older group, will be important to ensure household food security under a pandemic crisis. (3) retaining changes in food disposal behavior among the different age groups that occurred due to the influence of the pandemic through to the post-pandemic period can help to reduce food wastage in the future. In particular, it is important to retain changes seen in younger people of becoming more aware of potential food shortages and the uptake of improved food processing methods. On the other hand, further study from a public health perspective on gains in weight among younger people under the extended family structure in China may be warranted.