This prospective study of a large and nationally representative cohort of older adults in England sought to describe changes in participants’ TV viewing time over a 2-year follow-up period, and model socio-demographic, behavioural and health factors associated with increases in TV viewing. Results showed that participants watched an average of over 5 hours of TV at baseline, and mean viewing time across the whole cohort increased slightly but significantly over time. While 41% of participants decreased their viewing over the two waves, half of participants increased their viewing, with a quarter of the sample watching at least one more hour of TV daily at follow-up. Increases in TV viewing were associated with lower socioeconomic status, depressive symptoms, higher BMI, lower levels of physical activity, and being a smoker.
To our knowledge, our study is one of the first to use prospective data to describe variation in TV viewing over time. In addition to observing mean increases in viewing time, we also investigated patterns of change according to baseline viewing duration. These showed a general tendency towards increased viewing among participants viewing fewer than 6 hours of TV per day at baseline. While the majority of those watching 6 or more hours per day at baseline reported decreased viewing time at follow-up, 29% of this group increased their viewing time by at least one hour. Previous research has suggested that the transition to retirement is associated with increases in TV viewing , but we found that retirement status did not predict change in viewing duration. Additionally, similar patterns of results were observed among a subsample of participants who were fully retired at baseline, though depression and smoking predicted viewing change among the full sample but not among the retired. Our findings thus testify to the potential for, and magnitude of, naturally occurring increases in TV viewing among older adults over time, even where accounting for the retirement transition.
Our data support previous studies by demonstrating associations between TV viewing and behavioural and psychosocial variables . For example, a large Belgian cross-sectional sample found higher levels of TV viewing among functionally limited, less educated, widowed, and (semi-)urban dwelling older adults . In a community sample of older Japanese adults, TV time was associated with not being in full-time employment, lower educational attainment, increased weight, living in regional areas and low moderate-to-vigorous physical activity . Yet, factors associated with variation in static TV viewing scores need not be associated with changes in TV viewing over time; indeed, though our supplementary analysis showed that most predictors of TV viewing time also predicted changes in viewing time, chronic illness was associated with greater TV viewing time at follow-up but not increases in viewing time between waves. Our results showed that older adults with lower SES, depression, overweight and obese, physical inactivity, and smokers were more likely to increase their TV viewing. These subgroups may therefore require especial attention, with a particular focus on cardiovascular and respiratory disorders and mental health problems, when developing sedentary behaviour reduction interventions.
Our findings call for the development of effective behaviour change interventions to reduce sedentary behaviour among older adults. That more physically active participants tended to decrease their TV viewing time over the two waves concurs with previous research showing that engagement in physical activity can reduce time spent in sedentary activity ,. Sedentary behaviour change interventions might most usefully seek to displace TV viewing minutes with physical activities. Calls have been made for interventions to encourage older adults to get ‘out and about’ as a means of displacing sedentary home-based activities such as TV viewing . Yet, seasonal changes, and a perceived lack of safety and security in the local neighbourhood, can limit the effectiveness of non-home-based interventions among the elderly . TV viewing is often driven by enjoyment of TV , which may constrain the acceptability of interventions to replace TV viewing with alternative, non-sedentary activities. Given growing evidence of the positive health impacts of minimal-intensity physical activity relative to sedentary behaviour -, it may be feasible and beneficial to health to incorporate light-intensity physical activities into TV viewing patterns. Several potentially low-intensity activities have been proposed for insertion into otherwise sedentary TV viewing periods so as to promote physical activity and reduce sitting time, such as marching on the spot during commercial breaks , doing chores , operating a foot pedal device , or merely standing up, as a balance activity . Frequent performance of such activities has the potential to maintain the muscle power, balance and confidence required to stay physically active in older adulthood . Additionally, behavioural psychology suggests that adding activity into stable TV viewing routines would be particularly conducive to the formation of physical activity habits: consistent repetition of physical activities while watching TV should lead, through associative learning, to the activities becoming automatically activated with minimal mental effort when watching TV ,. In this way, physical activity should become an ingrained part of the TV viewing routine .
Limitations of this study should be acknowledged. First, the self-report TV viewing measure used has not been validated. However, many objective accelerometry devices cannot reliably differentiate sitting from other forms of light ambulatory movement, precluding true validation of self-reported sedentary behaviour measures. In the absence of such data, it is notable that the TV measure used in the present dataset has demonstrated convergent validity with various psychosocial, physical and biochemical risk factors hypothesised to be linked with sedentary behaviour ,,. Second, given the paucity of data on changes in TV viewing duration over time, it is unclear to what extent TV viewing patterns that we have reported are likely to replicate to other samples and settings. Participants reported watching TV for an average of 4.7 hours per day at baseline and 4.9 at follow-up. This exceeds the average 4 daily hours of TV time observed among older adults in a 2005 UK survey , and the 3.0-3.5 daily hours observed in UK EPIC-Norfolk data . Levels of sedentary behaviour may differ considerably both across and within nations , making it difficult to estimate the generalisability of our data. Relatedly, our analytic sample was healthier and more active than those who were lost to follow-up. Inactive people must be engaged in research for findings to have external validity. Third, all measures were self-reported, and so actual TV viewing time may have been underestimated. In addition, participants may have reported the amount of time the TV was turned on but not necessarily for how long they watched it. Population-based data should ideally employ both self-report and objective measures for verification purposes . It is also unclear whether TV viewing self-reports were consistently accurate over time. Some of the observed variation in TV viewing duration may be attributable to a lack of measurement stability, and it is not possible to isolate true changes in TV time from method error due to inconsistent reporting. More data is needed to evaluate the reliability, replicability and generalisability of our findings to other samples. Fourth, we focused on TV viewing duration, but not the times at which TV viewing occurred. It is unclear whether TV viewing minutes were mostly accrued in prolonged bouts of sitting, or are dispersed over multiple shorter viewing periods throughout the day. Complementing reports of total TV time with accelerometry, so as to obtain time-stamped data on when TV viewing most typically occurs, could aid the development of time-appropriate sedentary reduction intervention strategies .