Data missingness and representativeness

The survey was completed by 1429 respondents residing in Scotland. Missingness across the dataset was low at 3.9% of the total observations spread across approximately 53% of variables; see [25] for more details. Participants were slightly imbalanced with respect to deprivation using the SIMD quintiles; the most deprived group is proportionally under-represented with respect to the reference population (persons 60 years and over in Scotland), and the least deprived quintile appears to be over-represented, whereas we observe an over-representation of the least deprived quintile. The second to fourth quintiles appear to be fairly represented. This imbalance is likely due to both the surveying method (using predominantly online recruitment methods) and the differing non-response rates by socio-economic status as is commonly reported [41]. Therefore, the imbalance needs to be considered when interpreting the results.

For those under the age of 60 years, the reference population (defined as individuals self-identifying as ‘at-risk’ or vulnerable to Covid-19) is such that very little, if any, information exists about the key variables, thus making direct comparisons to assess sample balance across them unreliable or impossible. Therefore, although data were collected on participants representing a wider age range (22–98 years), only data on participants aged 60+ are presented in this paper; these were 84% (n = 1198) of the original 1429 participants that completed the survey.

Descriptive statisticsSocio-demographics and health behaviours

Participant characteristics and socio-demographics are shown in Table 1. The mean (SD) age was 67.3 (5.4) years; participants were aged between 60 and 98 years although the majority were between 60 and 70 years; median and IQR 66 (63–71). The majority reported their ethnicity as ‘White British’ (n = 1185, 99%) and reflects the Scottish population [42]. Most participants also fitted into the ‘retired’ (n = 811, 68%) or ‘Employed/self-employed’ (n = 214, 18%) categories, and two income brackets earning <£10,000 n = 106, 11%; earning ≥ £30,000 n = 376, 38%; ‘Prefer not to say’ n = 194). Of the 52% with a diagnosis of a health condition, 94% were taking prescribed medication for it. Most participants reported being non-smokers (95%).

Table 1 Sociodemographic characteristics of the sample aged > 60 years (n = 1198)Physical activity, sedentary behaviour, screen time and sleep

Physical activity, sedentary behaviour, screen time and sleep data are displayed in Table 2. In accordance with the IPAQ manual, total physical activity derived from the IPAQ variables (including vigorous, moderate, and walking) was used to determine physical activity categories for each participant shown in Table 3. As such, three-quarters of the participants reported physical activity volumes that categorized them as meeting physical activity guidelines. With regards to strength training, the overwhelming majority (87%) reported engaging in strength training less than two days weekly and therefore did not meet the strength training component of the physical activity guidelines. The participants’ reported time in bed was used to determine sleep categories for each participant based on the National Sleep Foundation recommendations. A minority of the participants were classified as ‘long sleepers’, with a relatively even split of participants across the sample being categorised as either ‘ideal sleepers’ or ‘short sleepers’.

Table 2 Physical activity, sedentary behaviour, screen time and sleep descriptive results, overall and stratified by genderTable 3 Proportion of participants meeting physical activity and sleep guidelines, stratified by genderHealth-related quality of life, loneliness, and social support/contact

On the EQ. 5D-3L, participants reported moderate health-related quality of life, and self-reported health rating; the total scores for this as well as mean loneliness score, social network size, perceived social support, and social contact time are shown in Table 4. The present sample appeared to report better mobility, self-care, and ability to perform usual activities than normative values for age group, but more problems with pain and anxiety/depression [43]. Table 4 also reports on the variables included in the models. These variables (loneliness, social support and social contact) have been discussed in greater detail elsewhere [25].

Table 4 Scores for psychosocial variables included in the modelChange in physical activity, sedentary behaviour, screen time, and sleep

Perceived changes from before social distancing began are shown for all variables where this question was asked in the 60+ years group in Table 5. Reported changes in vigorous- and moderate-intensity physical activity were similar with approximately half of the participants reporting having similar pre-social distancing levels. Just over a quarter of the participants reported walking more since social distancing. For strength training, 22% of participants reported doing less compared to before social distancing, while the majority of the participants reported that their strength training levels were the same, which for the majority was less than the two days per week guidelines already. For both screen time and sitting time, most participants reported doing more compared to before social distancing.

Table 5 Perceived changes since pre-social distancing for participants (n = 1198)Association of physical activity with loneliness and health-related quality of lifeCategorical predictor variables

In this manuscript, associations with the physical activity variables (physical activity category, reported walking time) is the focus. Figure 2 presents estimate from the joint model looking at associations between categorical predictor variables with loneliness and health-related quality of life scores, on the outcome scale.

Fig. 2

Parameter estimates on the outcome scale for the effect of categorical variables included in the model, by outcome (UCLA Loneliness and EQ. 5D health-related quality of life). The dashed line indicates the model intercept (the model average when everything is kept at the reference category)

Figure 2 shows that being female related to higher loneliness and worse health-related quality of life. Further, being in the high or medium active physical activity category was associated with better health-related quality of life (lower EQ. 5D score). Physical activity level was not associated with loneliness. Finally, living in a remote rural location was associated with better health-related quality of life, but less than ideal sleep was related to worse health-related quality of life.

Scale predictor variables

Figure 3 presents plots of the estimated shape of the relationship (spline plot) between walking (min/wk) with the two outcomes. The number and location of observations across the range of the independent variable is also shown as a density rug; the grey shaded bands indicate uncertainty around the estimates. Supplementary File 1 contains the complete output of the statistical model, while Supplementary File 2 contains all the spline plots based on it.

Fig. 3

Spline plots showing the estimated relationship between health-related quality of life (EQ. 5D, top) and loneliness (UCLA, bottom) with walking; the dashed line indicates the model average when all other variables are kept at the average or reference level

With reference to Fig. 3, reported walking of less than ±150 minutes per week had no association with loneliness scores. Between roughly 200 and 300 minutes of walking per week appeared to be associated with higher loneliness scores. Walking more than 600 minutes per week was associated with a lower UCLA score, although this is based on a limited number of data points, most of the sample reported walking less than 200 minutes per week. Greater volumes of reported walking time appeared to be associated with a lower EQ. 5D score (indicating better health-related quality of life), although very high volumes of walking (in excess of ±500 minutes per week) did not appear to yield better health-related quality of life scores in comparison to walking ±300 minutes per week.

Qualitative resultsChanges in physical activity and strategies to engage in physical activity (qualitative)

Overall, the responses to the open-ended questions in the survey indicated that there was substantial variation in not only the experience of social distancing (reported previously [25]), but also in the changes to physical activity and strategies used to engage in physical activity. These variations appeared to differ on an individual basis and the qualitative data presented indicates the complexity of the effect of social distancing guidelines (in addition to stay-at-home orders and closure of leisure facilities) on adults aged over 60 years.

The most common change in physical activity and/or most commonly reported ‘new’ physical activity was active travel. Many participants reported changing or replacing their motorised transport (e.g., bus and car) with walking or cycling: For example, one participant wrote, “Walking to local shops and parks when before I would have used the bus at least one of the ways.” (71y Male, Rest of Scotland). It was also positive to note that some participants indicated that they had intended to sustain this particular change: “I haven’t used public transport since lockdown. I have walked everywhere and I [intend] to continue to do so.” (68y Female, Rest of Scotland).

Many participants also noted that they optimised or maximised their walking opportunities. This took many shapes; some participants reported replacing what would have been their work commute with walking, some reported deliberately taking longer routes to the shops or other destinations, and some reported how their walking pattern(s) had changed: “do a nightly walk roughly equivalent in distance to my usual commute to work” (60y Male, Rest of Scotland) and “Each morning I walk to buy a newspaper during lockdown I have extended this walk. Instead of going directly to the shop my walk now takes about an hour each day instead of about 10 minutes.” (62y Male, Rest of Scotland).

Participants reported that access to facilities (or the lack thereof) was a major reason for the change in their physical activity. It was also a reason for the introduction of new physical activities. Several participants did mention that they missed the gym along with the social interaction that often accompanied their exercise. However, it was positive to note that some participants were able to continue with their usual activities such as Pilates, dance, and aerobics classes which had moved to an online format. Additionally, some participants mentioned that where there was not an online alternative, walking or cycling served as a complementary exercise: “I started doing daily exercise classes online instead of going to the gym. I’m doing a mix of Yoga, Pilates and Taichi. I’m also started to go for a walk as a routine.” (64y Female, Rest of Scotland). However, there were also participants who reported that none of their usual activities were options, even online, so their modality of physical activity changed. For instance, “Cycling or walking or both undertaken every day during lockdown with only two exceptions. Because other activity options were unavailable exercise opportunities became very important.” (63y Female, no postcode provided), and “I usually swim 5 or more days per week for at least an hour but with pool being closed I now walk instead” (67y Female, accessible rural). Three other topics that came up far less frequently but are worth mentioning with reference to online physical activity specifically, included 1) internet connectivity issues; 2) lack of enthusiasm to start or attempt a new physical activity; and 3) lack of motivation to persist with a new physical activity.

Many participants reported health and injury as a reason for their change in physical activity. It is noteworthy that only two participants explicitly stated Covid19 when answering the questions about physical activity. These two participants reported that they had contracted Covid19 and had restricted their physical activity due to their recovery. One of the two participants mentioned that they had been hospitalized. Commonly mentioned injuries and health issues that affected physical activity included cardiac issues and musculoskeletal injuries, some of which occurred during the period of social distancing and were further affected by the restrictions. Examples include, “I started walking more, but I have hurt my knee, so walking has been curtailed” (67y Female, Rest of Scotland) and:

“A few weeks before social distancing I had a heart attack and 2 stents fitted. Lock down precluded cardiac rehab classes. Cardiac physio sent exercise DVDs from cardiac rehab and [month] ago I was referred to local council. … I currently use the online videos” (66y Female, Rest of Scotland).

It is also worth noting that several participants reported that social distancing had assisted with their injury management, and so although they were injured and/or recovering, the regulations on social distancing and lockdown afforded them time to recover and improve their physical activity: “Lots more walking. Up to 6 miles a day. Weather dependent though. Had hip replacement in December and this had been FANTASTIC for it. And for me in general. Loving the activity.” (66y Female, accessible rural).

The influence of shielding (or stay-at-home orders) on physical activity changes was mentioned by both participants who were shielding (due to their own health) as well as those who were residing with someone who was shielding (e.g., a child or partner); the role of caring is detailed later. It is important to note that responses regarding shielding did appear to be different depending on when the survey was completed (see Fig. 1 for shielding guidance and the restrictions physical activity in relation to timing of data collection). An additional point that was often mentioned alongside shielding was the type of housing or general space available for physical activity, which unsurprisingly was an important factor for individuals who were advised to not leave home for their safety. Participants wrote: “Since I am shielded my only exercise is housework” (88y Male, rest of Scotland), “Shielding required a complete change in exercise behaviour. 12 weeks confined to house & garden so indoor cycling and brisk walks round garden. Now allowed out” (67y Female, accessible rural), and “Whilst shielding I started walking round garden as I couldn’t go out” (64y Female, no postcode provided).

Several participants mentioned that their pets played a role in their change of or new physical activities during lockdown. One participant mentioned engaging in daily horse-riding exercise although dogs were the most commonly reported pet influence. For most participants that mentioned their dogs, a common theme was that they were either walking for longer and/or walking more frequently than before social distancing: “I now walk my dog twice a day 7 days a week, instead of only 2 walks a day at weekends. I now walk approx 21 hours per week compared to 4 hours prior to lockdown” (60y Female, accessible rural).

The first lockdown in Scotland occurred during the spring/summer months. Based on the responses of several participants, fair weather made outdoor physical activities such as walking and gardening possible. Conversely, ‘bad’ weather appeared to be a deterrent to being outdoors – particularly as the end of summer approached:

“I haven't started any new physical activities. I walk almost every day round a park, to shops etc. I have been cycling a bit more recently. Part of the reason for that is the nice weather we have experienced in recent weeks. … My routine has not changed a lot because I am retired, and I walk as much as I can, anyway.” (61y Female, accessible rural).

It was interesting to note how few responses included the element of time (or lack thereof) with regard to changes in physical activity relative to other responses detailed above. Based on the volume of responses, it appeared that in the context of all pandemic-related challenges, time was not a major barrier for adults aged 60 years and over. However, when time was mentioned, it was both a facilitator (where someone reported having more time) and a barrier (where someone reported having less or no time), albeit less frequently. In several instances, time was linked to employment (or lack thereof due to being furloughed) or other responsibilities: “Knowing that I was going to be home all day for a rather long time, I decided to incorporate a workout into my daily schedule.” (65y Female, Rest of Scotland).

Beyond the reported changes in physical activity due to social distancing, we were interested in responses that described successful strategies to maintain physical activity as well as responses that reflected positive changes related to physical activity that people made during social distancing (and/or because of stay-at-home guidance and closure of leisure facilities). These anecdotes ranged from participants’ positive experiences in discovering something new, to enjoying domestic but physical tasks, to seeing measurable improvement in their physical ability, and in several instances, participants reported setting goals or targets and reported how this motivated them: For example, “I have decided to walk at least 10k steps daily over lockdown and I have done so on the vast majority of days” (72y Female, rest of Scotland), and:

“I started exercising accompanied by better diet and using a treadmill. Going for more walks has increased my fitness, lowered my weight and increased my mobility. The increased mobility has meant that I can now manage steeper inclines, walk faster and for longer periods. The change is that I'm actually exercising, which I found almost impossible before. I didn't do anything physical because I was overweight and in discomfort. I've lost almost 1st 11 lbs and at almost 67 yrs old, I feel so much better. I probably wouldn't have done it if lockdown hadn't happened.” (66y Female, accessible rural)

and

“I have a gym membership but [don’t] like classes because they are too busy so online zoom has been brilliant for me. I can see a change in my shape for the better. I am going to continue running several times a week. I didn't run at all before lockdown” (66y Female, rest of Scotland)

Regarding the location of physical activity, the most frequently identified location of physical activity for participants was the local area (n=646, 54%), followed by in the house (n=559, 47%), in the garden (n=512, 43%), and online (n=195, 16%). One hundred and sixty-nine participants (14%) selected the ‘other’ option. In these responses, the most frequently reported ‘other’ options included golf courses, blue space (including the beach and lochs), in the woods or hills, and places of work. Several participants used this space in the survey to elaborate on their physical activity location and the reasoning: “I live in the country and have horses and sheep and dogs do plenty physical exercise!” (73y Female, accessible rural), and “At work. I'm employed in grounds maintenance. It can be very physical.” (64y Male, rest of Scotland).In response to the final question of the survey, for which participants were prompted to add anything that they felt had not been covered in the survey, the overwhelming majority of responses included stories of frustration, grief and longing to see grandchildren (that have been described in detail elsewhere [25]). However, several participants did elaborate on their physical activity, often linking it to their social activity and other themes identified in the physical activity-specific questions, such as health and injury: “I feel much less fit and strong, and am of an age where regaining those may prove difficult...” – 81y Female, rest of Scotland) and:

“I had covid and now have long tail covid. My family are wary that I may still be infectious. My physical activity is very curtailed. I used to row competitively, cycle, yoga, walk everywhere. I can only manage slow, short walks. I am signed off sick from work.” (62y Female, rest of Scotland).

Several participants also elaborated further on their lived experience and how the time of social distancing paved the way for a positive future, sometimes through their strategies to remain physically active:

“I had time on my hands during lockdown and was worried about my weight and my diabetes putting me at greater risk of a bad outcome if I contracted covid-19. I decided to join an online health and fitness plan, which has been a great success. In 2 months I have lost over 2 [stone]. I now walk every day routinely more than 10,000 steps a day. I have also joined the One Million Step Challenge. In addition to this, I have discovered another strand of support in my online group and online coach. As a result of joining this programme, I am now [eating] only wholefoods and following a low-carb diet. I believe my blood sugars are back under control, I no longer have sugar cravings, I am sleeping much better and I am now not nearly as anxious as I was previously.” (61y Female, rest of Scotland),