In this study, multidomain lifestyle intervention resulted in favourable changes in participants’ dietary habits and physical activity, and part of the improvements remained also during the long-term post-intervention follow-up. Although no significant differences between the intervention and control groups were evident in other modifiable factors, we showed that, in the whole cohort, social and cognitive activities increased, and self-evaluated health and memory improved during the active study period but decreased thereafter. Alcohol use tended to decrease over time. Smoking decreased more in the intervention group during the intervention period, but this difference was not maintained in longer term. During the COVID-19 pandemic, frequency of participants’ physical and cognitive activities increased and so did binge drinking.
In line with previous studies [7,8,9], physical activity increased, and diet improved during the intervention in the intervention group. The effect of the intervention on physical activity was no longer observed during the long-term post-intervention follow-up, but improvements in diet achieved during the intervention were mainly sustained. These results are fairly similar to those from the Finnish Diabetes Prevention Study targeting lifestyle intervention to overweight, middle-aged persons with impaired glucose tolerance; favourable effects of intervention on dietary habits and weight were sustained, but not the effects on physical activity in the long-term post-interventional follow-up period [12]. Not many earlier studies have reported post-intervention sustenance results, but in our study the main reason for this difference is most likely the set-up of the original intervention. During the original intervention period free access and instruction to gym was provided, but afterwards participants were required to conduct and also pay their training independently. Original intervention included support to plan how to continue training individually, but better integration of the training option provided by community and other stakeholders could have improved the long-term results in physical activity. In dietary intervention, on the other hand, the main content during the original intervention was learning to adopt healthy diet independently at home, which is likely to be better maintained after the supported period is over. Also, as our maintenance intervention consisted of text messages to provide small tips and reminders, and may thus have been more suited to support smaller changes adopted at home.
Frequency of cognitive and social activities increased, and self-evaluated health and memory improved during the active study period, but the changes were largely similar in both groups. These changes may be partly associated with the participation in the study, i.e. Hawthorne effect. While people in the intervention group had many individual and group intervention sessions, the control group similarly visited the study nurse and physician several times during the first two years. During the study visits, the study nurse gave all participants oral and written information and advice on healthy lifestyle (diet, physical, cognitive, and social activities) for supporting healthy aging. It is a typical finding in lifestyle interventions that the active study period has effects on health behaviour on both intervention and control groups [11, 12, 14].
A decrease in social activity has been shown to be associated with cognitive decline [25]. In our study, social activity decreased in post-intervention follow-up, and the pandemic did not notably influence such trend in older people at risk of developing dementia. A Finnish study [26] investigating the effect of the pandemic on activity in older people reported that many kinds of activities (including social, cognitive, and physical) decreased during the pandemic. In a study from Australia, participants reported decreasing trend in social activities but an increase in cognitive activity during the COVID-19 pandemic [22]. In that study, the increase in cognitive activity was, however, at least partly associated with engagement in a public health program targeting dementia risk reduction rather than with the pandemic. In our study, cognitive activity increased during the pandemic as opposed to previous trend. It is possible that decrease in other activities outside home during the restrictions were partly compensated by cognitive activities, which can take place at home, such as reading, writing, or crosswords. This could indicate that people were able to adjust their activities in the restrictions and such flexibility may be important for coping with the situation.
In line with previous studies [19, 21,22,23], smoking was rare among older adults and reduced during the pandemic. Regarding alcohol use, consumption frequency decreased but frequency of binge drinking increased which has been found also in other high-income countries [27] and could be related to the polarization of drinking habits that has become more obvious during the pandemic [28]. Decline in social activities and limited access to places outside home can also possibly explain overall lower frequency of drinking.
The biggest contradicting finding, as opposed to other studies, was the change in physical activity during the COVID-19 pandemic period. In our study, frequency of physical activity increased significantly whereas in most of previous studies, physical activity has decreased [17,18,19,20,21, 29]. In many of those studies, participants have self-evaluated the change in physical activity (e.g. decreased, increased, remained the same) which may have led to the differences in our results where change in physical activity was measured as a change in physical activity level between different study points. Also in our COVID-19 survey participants evaluated how they felt their physical activity was changed (more, less, the same than before the pandemic), and we have previously reported that participants considered that their physical activity decreased [24], whereas we see an increase when analysing level reported before the pandemic and during the pandemic in this study. One reason for the discrepancy between subjective feeling of change and longitudinally measured change could be availability of other activities; as most gatherings and outside home activities were closed down during the pandemic (reflected also in social activities in this study), people had more free time at home. Thus, their experience of time spent in physical activities could have been lower simply due to more overall free time. In addition, the type of activity may have changed when gyms and most instructed exercise sessions were closed, but there was still possibility for going outdoors, and thus increased frequency may come with change in the type of sports and/or decreased intensity. Present study only measured moderate-intensity activity of at least 20 minutes at the time which is not likely to represent the total physical activity of the study participants. However, the used assessment is likely to capture various types of exercise conducted outdoors (e.g. walking and jogging) which were more accessible also during the pandemic and related restrictions.
Based on these results, the effect of the pandemic period on lifestyle was overall more positive than expected as we did not find any statistically significant negative changes between the 7-year time point and the pandemic; and we did observe increased frequency of physical activity. Older people are used to accommodate their activity because of declining health and function along with ageing [30] which may partly explain the results. Also, pandemic related changes in everyday life in older adults in our study group, who were mostly retired and still living independently at home, could have been smaller than for younger people who were used to going to school or work and meeting more people outside home; or for those older adults who were living in institutions outside home. It’s worth noting that this questionnaire was completed after the first wave of the pandemic with a relatively strict restrictions ongoing but still after the lockdown related measures were already lifted. People may have felt more positive after the most difficult period was over. Also, these changes represent short-term effects of the pandemic during its first 6 months and effects in longer term may be different. It would also be important to identify groups that were maybe more vulnerable to pandemic. In our earlier analysis with self-evaluated changes those living alone reported more negative changes, e.g. in physical activity and overall health, but these findings were not evident here with longitudinal data.
Strengths and limitations
In the present study, we reported results from long-term follow-up (ten years) of the multidomain intervention trial. This enabled us to assess the effect of the intervention on lifestyle during the active trial and also in long-term. We had longitudinal data on different outcomes from many years before the COVID-19 pandemic, and thus, we were able to compare the pandemic-associated changes to the preceding trends, not only to the data based on participants’ perceived changes that has been done in the most studies reporting pandemic-associated health changes and that is more prone to recall bias.
Methodological limitations include reliability of the self-report in outcomes, without objective measures. However, measures were mainly the same at every follow-up point which make them comparable to each other. The purpose of this study was to describe general trends in older people’s lifestyle with and without intervention and before and during the pandemic. Therefore, we reported changes in lifestyle using crude measures of different lifestyle factors which were also available from postal survey during the pandemic. More detailed analyses on changes in individual lifestyle factors with more specific measures taken during study visits and more individualised approach are forthcoming after the post-pandemic follow-up visits are completed. Future studies should involve more objective indicators, such as wearables for physical activity or biomarkers for dietary intake.
Another main limitation is the attrition. The study population participating at the 10-year follow-up was still relatively healthy and mostly independently living, because people engaging in lifestyle intervention and still participating after ten years are likely to be selected and may thus not represent all older adults in Finland, although the cohort is recruited from population-based sample. Participation rate among eligible participants was high throughout the study, but due to deaths, illnesses, and withdrawals only 60% were still included at the end of this follow-up. The original selection of at-risk group attenuates the typical selection of healthy and health-conscious people in studies, but cohort attrition is likely to influence the results in the latest visits. There were no differences between the original intervention and control groups in terms of attrition, making the group comparisons less likely biased, but likely that true changes over time would be more negative than in this selected population.
A pandemic-associated limitation is that the pandemic follow-up point was during the very first months after the pandemic started, right after the first wave although after most restrictions were already lifted. Based on these data, long-term consequences of the pandemic on lifestyle factors are not known. On one hand, it is also possible that lifestyle was more affected during this period as compared to later waves after vaccinations were introduced and restrictions were less strict. On the other hand, it’s also possible that people felt more optimistic after the most restricted period was over. Also, the questionnaire was sent during summer, which made it easier to move outdoors and meet people, when there were still restrictions on gatherings indoors. Originally participants came to the study visits different times of year.