Tuesday January 12 2010
The associations in this study were weak
Every hour per day spent watching television "increases risk of heart disease death by a fifth”, according to a report in The Daily Telegraph.
This news story is based on an Australian study that looked at an association between sedentary behaviour, estimated by number of hours spent watching television, and the risk of death. The research found a very small association between the length of time spent watching television and mortality rates. There was also a very small association with deaths specifically related to heart disease.
The small significance of these associations, along with limitations in the study design, mean this research should be interpreted cautiously. However, it is well known that regular exercise has beneficial effects on health and that people should try to maintain an active lifestyle.
Where did the story come from?
This research was conducted by Professor David Dunstan and colleagues from the Baker IDI Heart and Diabetes Institute of Melbourne, and other Australian universities. The study was funded by the National Health and Medical Research Council in Australia and was published in the peer-reviewed medical journal Circulation.
The research was generally reported well, but the press has tended to exaggerate the results, which were only of borderline significance.
What kind of research was this?
This was a cohort study looking at the relationship between television viewing habits and the risk of death, including mortality related to heart disease.
The researchers were interested in the time spent watching television as a way of measuring how sedentary a person was. There is already evidence to suggest that regular exercise is good for health, but in this case researchers wanted to see how staying inactive for long periods of time affected health.
A cohort study is good for looking for associations between specific factors (in this case viewing habits) and outcomes such as death. However, researchers need to ensure that they consider all possible confounders (factors that may influence the association being studied).
There are several limitations in the design of this particular study, such as the use of various self-reported measures, which are likely to involve some inaccuracy. A further drawback is that these factors were measured at only one period in time. Participants were asked to record the number of hours they had spent watching television for the previous seven days. Collecting this type of data at only one point in time can be problematic as many lifestyle habits change over time.
What did the research involve?
The study recruited 8,800 participants from regions all over Australia between 1999 and 2000.
The participants had a medical examination and tests after fasting overnight for over nine hours. During these tests the researchers measured the participants’ glucose, fat levels, cholesterol levels, resting blood pressure and waist circumference. The study excluded people who had reported a previous history of stroke or heart disease, or were pregnant at the start of the study.
The researchers asked the participants to report the number of hours that they had spent sitting down watching television in the previous seven days. They divided the respondents into three categories: people who watched less than two hours per day, people who watched between two and four hours, and people who watched more than four hours of television each day. The researchers excluded any participants that did not have complete data on the number of hours spent watching television.
The researchers gave the participants a questionnaire to gather information on their demographics, educational level and exercise habits. They also asked whether the participants used any type of cholesterol medication, whether they smoked or if they had a parental history of diabetes. The participants were also asked to fill in a food frequency questionnaire to estimate their calorific intake. If the participants reported total energy intake between 500 and 3,500 calories per day (kcal/d) for women and 800 and 4,000 kcal/d for men then they were included in analysis.
The study was followed up until 2006, with an average follow-up of six years. The researchers then assessed the number of mortalities in the study group and the causes of these deaths.
What were the basic results?
Over the study period there were 284 deaths.
In initial basic analysis the researchers found that there were borderline associations between more hours watching television per day and death due to any cause (hazard ratio [HR] per one hour increase per day, 1.11, 95% CI, 1.03 to 1.20). There was also a borderline association with death related to heart disease (HR,1.18; 95% CI, 1.03 to 1.35).
However, there were several factors that the researchers had measured that could contribute to heart disease and mortality. Adjusting for factors such as age, sex, leisure-time exercise and waist circumference reduced the association between the number of hours spent watching television and mortality.
The researchers compared people who watched less than two hours of television a day, and people who watched more than four hours a day. They found that, after adjusting their analysis for the influence of age and sex, those watching more television had an increased risk of death due to any cause and death due to heart disease (hazard ratios 1.67, 95% CI 1.20 to 2.33, and 2.12, 1.20 to 3.77, respectively).
After further adjustments to account for additional factors this effect was reduced to a 46% increase in risk for all-cause mortality and an 80% increase in risk for heart-disease-related death. However, these results had borderline statistical significance, if any.
How did the researchers interpret the results?
The researchers suggest that “these findings from a large population-based cohort of Australian men and women indicate that prolonged television viewing time is associated with an increased risk of all-cause and CVD (heart-disease-related) mortality”.
This study looked at associations between the length of time spent watching television and the risk of death, both due to any cause and specifically to heart disease. Although the study did find some associations, these were relatively weak and were of borderline statistical significance.
Studying an outcome such as mortality rates can be very complex due to the numerous contributory factors involved. The interdependence of these influences means that it is difficult to isolate one contributory factor such as television-watching habits.
While this study had the benefit of a large sample size it has various limitations, some of which the researchers highlighted:
- They only took a single measure of television-watching habits at the start of the study. These habits may have changed over the course of the six-year study period.
- Many of the measures asked participants to self-report. Sometimes individuals can respond differently to questions, resulting in inaccurate data.
- The researchers were using television-watching habits as an indication of how sedentary a person was. Although they say this may be a reasonable way of measuring sedentary behaviour, another measure of how much time a person spent sitting down per day might have been useful.
This study has only shown a small association between watching television and mortality, and it should not be taken as conclusive evidence. Also, although health measures were taken, it may be the case that people in poorer health are likely to be sedentary rather than sedentary behaviour causing poor health. Despite the limitations of this research, it does not undermine the importance of physical activity, and it is recommended that people take regular exercise of 30 minutes at least five days a week as part of a healthy lifestyle.