Demographics
Of 1723 participants, mode age was 24.9 years (age range collected), range 16-80 years, with 83.1% (n=1431) under 30 years. The majority (97.2%) were assigned the female gender at birth (n=1675; male n=46, prefer not to say n=2), with 91.6% (n=1578) identifying as female. Forty-five participants (2.6%) identified as Aboriginal or Torres Strait Islander. Most participants lived in urban areas (81.5%), over half with their families (55.7%) or partner (16.4%), and 9.3% were parents. The cohort was well educated, with 40.9% having graduated high school and an additional 51.9% higher qualification or degree, and 68.3% were in paid employment prior to the pandemic. Full demographics in Table 1.
Pre-existing eating disorder illness factors
Thirty-nine percent (n=679) self-reported a current DSM-5 (APA, 2013) eating disorder (ED) diagnosis, 20.1% (n=346) a previous diagnosis, with 40.5% (n=698) not having ever received formal diagnosis but self-identified ED symptoms. Table 1 shows the proportion of the sample by diagnostic group, the largest proportion reporting a lifetime experience of Anorexia Nervosa (AN) (42.3%). To note, participants could identify more than one diagnosis, with 33.1% identifying multiple.
This cohort reported significant current eating disorder symptomatology. Of those who completed the EDE (n=1267), the mean EDE Global Score was 4.08 (SD=1.18, range 0.06 to 6.00), comparable to expected norms for a clinical population (x=4.02, SD=1.28) and distinct from a non-clinical population (x=0.93, SD=0.86), (Aardoom, Dingemans, Op't Landt, & Van Furth, 2012). Regarding identified ED symptoms prior to the pandemic (i.e., prior to March 1, 2020): 90.8% (n=1565) reported body image concern, 75.5% (n=1300) food restriction/dieting, 55.6% (n=958) over/binge eating, 25.0% (n=430) self-induced vomiting, 37.1% (n=639) driven over-exercise, and 14.9% (n=257) laxative or water pill misuse (multiple responses allowed). Only 100 participants reported diet pill misuse (Table 2) but as reported increase in symptom was low for this group (n=4), no further analysis was interpreted (Table 3).
Participants reported a history of ED symptomatology for a mean (x) 9.28 years (SD=8.45, range 1-53 years, Median 7 and Interquartile range 3-12 years). Prior to the pandemic, the vast majority (95.9%) reported experiencing ED symptoms at least some of the time: 20.5% all the time, 45.0% most of the time, 16.4% half of the time, 14.0% some of the time, with 4.1% none of the time (but previously). Nearly half (47.9%) were receiving some form of ED treatment during the pandemic, 25.7% had received treatment in the past, and 26.5% had never received any treatment. To note, there was a significant difference between EDE Global Score for those with a treatment history (x=4.19, SD=1.16) compared to those with none (x=3.78, SD=1.19), t(1265)=5.58, p<.001. One hundred and twenty-seven participants had an inpatient admission within the past twelve months, and an additional 168 reported an admission prior to that, representing 17.1% of the sample reporting receipt of inpatient care.
Pre-existing co-occurring mental health factors
Seventy-nine percent reported a current or lifetime co-occurring mental health condition: 71.1% obsessive-compulsive disorder, 70.9% anxiety, 55.2% depression, and 65.2% a history of self-harm, suicidal ideation or at least one suicide attempt (24.7% currently, 40.5% previously), detailed in Table 1 (multiple responses allowed). Considering mental state within the week of baseline reporting, participants reported on average mild depression (x=12.36, SD=5.98), mild anxiety (x=8.86, SD=5.17) and normal stress (x=11.99, SD=5.03) on the Depression Anxiety Stress Scale (DASS21) and elevated feelings of loneliness (UCLA Loneliness Scale: x=54.54, SD=12.10, range 2-80 with a maximum of 80). Participants rated their overall quality of life as an average of 44.96 on a visual analogue scale [range 0 (very poor) - 100 (very high), SD=23.46].
Impact of COVID-19
Participants were asked about the direct impact of COVID-19 on themselves or an immediate family member: 629 self-quarantined after being a close contact with an identified case, with 108 of that group falling physically ill with the virus. An additional 65 people were hospitalised, with 20 of those passing away. To note, 107 participants declined to answer. Participants rated feelings related to their experience of the pandemic overall on a visual analogue scale: Worry x=38.93 [0 (very worried) - 100 (not worried), SD=24.94]; Fear x=44.41 [0 (very fearful) - 100 (not fearful), SD=25.23]; Confidence x=41.91 [0 (not at all confident) – 100 (very confident), SD=22.15]; and Hopefulness x=43.25 [0 (not at all hopeful) – 100 (very hopeful), SD=22.98], although wide variance is noted. Participants with higher scores on the EDE Global Score, DASS and UCLA Loneliness, were significantly more likely to report more worry and fear, and less confidence and hopefulness, relating to their pandemic experience (Table 3).
Impact of COVID-19 on eating disorder symptoms
Participants were asked to rate degree of change of ED symptoms at assessment (within pandemic) compared to before the pandemic (pre-pandemic). Presented in Table 2 are responses by the total cohort and by subgroups dependent on whether the ED symptom was present pre-pandemic or not. As can be seen overall, there is a reported increase across most ED symptoms within pandemic: 88.0% of the total sample reported an increase in body image concern (67.2% a lot, 20.8% somewhat), 74.1% in food restriction/dieting, 66.2% in binge/over-eating, 48.6% in driven/over-exercise and 25.4% in self-induced vomiting. Of the 1565 who reported body image concern as a current symptom prior to the pandemic, 90.5% reported an increase (70.3% a lot, 20.2% somewhat), of the 1300 who reported food restriction/dieting as a current ED symptom prior to the pandemic, 82.1% reported an increase, and of the 958 who reported binge/over-eating as a current ED symptom prior to the pandemic, 89.2% reported an increase.
Although applying to much smaller proportions of the cohort, reporting re-emergence and first-time emergence of ED symptoms during the pandemic was not uncommon. As seen in Table 2, 59.3% of participants reporting previous (i.e., a lifetime ED symptom that was not active prior to the pandemic) body image concerns reported symptom return during pandemic. Similarly, 47.7% who reported lifetime food restriction/dieting, 47.2% binge/over-eating and 40.5% driven/over-exercise reported re-emergence of those symptoms. For some participants who reported they had never experienced (or the symptom was not applicable) a particular ED symptom prior to the pandemic, they experienced it for the first time during the pandemic: 65.3% reporting increased body image concerns; 37.5% food restriction/dieting, and 26.8% binge/over-eating.
Independent samples t-tests were used to explore the relationship (Mean Difference; MD) between change in ED symptoms reported during the pandemic with ED severity. For all ED symptoms, participants who reported an increase in ED symptoms during the pandemic (within pandemic) were significantly more likely to have a higher eating disorder severity score as indexed by the EDE Global Score, for example, body image concern (MD=0.98; t(1325)=10.05, p<.001) and food restriction/dieting (MD=0.62; t(1325)=8.62, p<.001) (Table 3).
Pearson’s chi-square test of independence with corresponding relative risk (RR) and confidence intervals (CI) were performed to explore relationships between change in ED symptoms within pandemic with ED diagnostic status and ED treatment status (Table 4). The relationships between ED symptom and presence of ED diagnosis were mixed, for example, participants with current ED diagnosis (39.4%) were 1.6 times more likely to report an increase in self-induced vomiting during the pandemic as those without diagnosis (RR 1.62; 95% CI, 1.442 to 1.811), but 25% less likely to report an increase in binge/over-eating (RR 0.75; 95%CI, 0.667 to 0.842). The relationship between ED treatment status (i.e., whether participants were engaged in active treatment or not) and ED symptom change were also mixed. The 47.9% of the sample engaged in ED treatment were more likely to report an increase in self-induced vomiting (1.3 times), laxative and/or pill misuse (1.4 times) and driven/over-exercise (1.2 times), but 22% less likely to report an increase in binge/over-eating.
Relationship of mental state to eating disorder symptom change
Independent samples t-tests were used to explore the relationship between ED symptom change during the pandemic with mental state (DASS-21) and loneliness (UCLA Loneliness Scale) (Table 3). Higher DASS depression scores were associated with an increase in all ED symptoms within pandemic, such as for body image concern (MD=1.90; t(1216)=3.54, p<.001), food restriction/dieting (MD=1.50; t(1216)=3.84, p<.001), binge/over-eating (MD=0.94; t(1216)=2.63, p=.009), and self-induced vomiting (MD=2.29; t(1216)=5.84, p<.001) with a moderate to high effect size. Higher DASS anxiety and stress scores were significantly associated with increased body image concern, food restriction/dieting, self-induced vomiting, driven/over-exercise and laxative and/or pill misuse. Finally, higher loneliness scores were associated with an increase in all ED symptoms except for driven/over exercise.
Pearson’s chi-square test of independence with corresponding relative risks and confidence intervals were performed to examine the relation between the presence of at least one co-occurring mental health condition (79.0%) and change in ED symptom during the pandemic (Table 4). Participants who self-reported a current co-occurring mental health condition were 1.1 times more likely to experience an increase in body image concern (RR 1.11; 95%CI, 1.018 to 1.219) and food restrictions/dieting (RR 1.07; 95%CI, 1.006 to 1.135) within pandemic, noting modest effect sizes. Regarding the impact of the pandemic on other aspects of mental health: 87.5% of participants reported guilt over buying food, 69.3% a poorer quality of sleep, 50.3% increased alcohol use, 34.5% increased use of prescription medication, 32.3% increased smoking and 31.7% increased recreational drug use (Table 2).
Impact of public health measures
Presented in Table 5, most participants were negatively impacted by a variety of public health measures and pandemic experiences, for example, 83.0% were negatively impacted from change in daily routine, 81.0% from restricted access to family and friends, 68.0% from social media reaction to the pandemic and 62.0% from news coverage. Independent samples t-tests were used to explore the relationship between eating disorder severity as indexed by the EDE Global Score and whether these experiences caused a negative impact, as opposed to a neutral/positive impact. There was a significant relationship between higher EDE Global scores in 12 out of 18 experiences, such as promotion of exercise as an essential activity (MD=0.55; t(1243)=8.49, p<.001), change in daily routine (MD=0.54; t(1254)=6.08, p<.001), availability of safe foods on meal plan, increased focus on cleaning/hygiene, and change in access to health professionals/treatment, to name a few.
Residents of the state of Victoria (45.6%) experienced stricter public health measures for most of the study duration so a comparison of impact was made to the rest of Australia. Current ED illness level as indexed by EDE Global score (MD=0.01; t(1221)=0.15, p=0.88) was not significantly different for Victorian residents (Table 3), and the only significant ED symptom increase was for food restrictions/dieting behaviour (RR 1.2; 95%CI, 1.07 to 1.38) (Table 4). While there was no significant difference in stress, anxiety or loneliness, Victorian participants reported significant difference from the rest of Australia for depression (MD=1.93; t(1194)=2.72 p=0.007) to a large effect size (Table 3).