For decades, public health guidelines around the world have recommended adults get at least 30 minutes of moderate-intensity physical activity most days of the week. This is roughly equivalent to 150 minutes of moderate-intensity activity per week, or 75 minutes of vigorous-intensity activity.
So it surprised, and perhaps dismayed, many to read headlines last week claiming “we need to do five times as much exercise as we’ve been told”. These claims came from a study published in the British Medical Journal (BMJ), that concluded:
People who achieve total physical activity levels several times higher than the current recommended minimum level have a significant reduction in the risk of the five diseases studied.
Indeed, the more exercise you do the better. However, the important issue that has been missed is that the BMJ study based its findings on self-reported physical activity time accumulated across multiple areas of life. These include such incidental physical activity as lifting boxes at work and cleaning windows, as well as going for jogs and playing tennis. But most studies that explore associations between disease risk and exercise are based on specifically designated physical activity.
By using a self-reported multiple-domain measure of physical activity, this study has generated estimates – of around 13 to 16 hours of brisk walking or six to eight hours of running per week – that are significantly higher than studies using measures focused on designated exercise activities, such as jogging or lifting weights.
How was the study conducted?
The BMJ study was a systematic review, where researchers analyse a number of existing studies and papers on a particular topic. In this case, authors reviewed studies between 1980 and 2016 that examined the association between physical activity and risk of five common diseases: breast cancer, colon cancer, diabetes, heart disease and stroke.
In the 174 papers analysed, authors extracted data relating to total minutes of physical activity per week. The activity in these studies was performed across multiple domains that included during work time, such as carrying light loads; transport, such as walking to work or to shops; and leisure, such as walking the dog or playing tennis.
Many of these studies reported only leisure-time activity. For these, authors used statistical methods to estimate what the equivalent of total physical activity would have been across the several domains.
These estimates were based on data from nationally representative surveys that had used the World Health Organisation’s Global Physical Activity Questionnaire (GPAC), which also measures physical activity across multiple domains.
What were the results?
The study confirmed that physical activity, even when modest and equivalent to 150 minutes per week at moderate intensity, is associated with risk reductions for some of the major causes of illness and premature death in Australia (diabetes, heart disease and stroke).
The study also showed that additional health benefits can be gained by engaging in higher levels of physical activity, which includes reducing the risk of breast and colon cancer.
In the case of diabetes, heart disease and stroke, the authors showed reasonably large risk reductions (14-16%) for adults who performed 30 to 99 minutes of activity per day, compared to less than 30 minutes per day.
However, risk reduction in the vicinity of 15% was only seen in colon cancer in adults who accumulated 100-199 minutes of activity per day. An even higher amount of activity – more than 200 minutes per day – was associated with a risk reduction of 15% in breast cancer.
How should we read the findings?
This research specifically examined associations of total physical activity across multiple domains with chronic disease risk. The use of this multiple-domain approach resulted in an overestimation of activity levels. For example, the highest levels of physical activity in the paper equate to more than 19 hours of moderate-intensity activity per day. This is clearly unrealistic.
Multiple-domain physical activity questionnaires routinely overestimate physical activity, often showing higher numbers than those that use questionnaires focusing on leisure time only. This is for several reasons.
Activity performed during leisure time tends to be purposeful and more vigorous than the incidental activity typically done in other domains. Because of this, it’s easier for study participants to recall and report their leisure-time activity.
Leisure is also the domain where adults have most discretion about how they spend their time and is therefore more amenable to change.
So as a reminder, the BMJ study used the World Health Organisation’s questionnaire, which measures physical activity across multiple domains as the basis for their estimates of physical activity.
This questionnaire may be a reasonable tool for ranking participants from least active to most active – across the multiple domains in their life – or for monitoring physical activity in populations. But estimates of minutes per week derived from this questionnaire can’t be taken as an accurate representation of the volume of physical activity associated with disease risk.
All we can really conclude from this study is that modest amounts of physical activity provide health benefits and that there are additional benefits to be gained with additional physical activity.
What else should we consider?
When health agencies first prescribed levels of physical activity for the public, the focus was on vigorous exercise and heart fitness. The recommendation of 150 minutes of moderate-intensity activity per week emerged in the 1990s in response to evidence demonstrating the benefits of physical activity at lower intensities.
The Australian government revised its national physical activity guidelines in 2014, which are sufficient for conferring many health benefits. Australians are now encouraged to aim for 300 minutes of moderate (or 150 minutes of vigorous) activity per week, to incorporate strength training at least twice a week, and to try and reduce the amount of sedentary time they accrue each day.
Only 56% of Australian adults and 25% of older Australians met the Australian government’s physical activity guidelines in 2014-15.
Reports stating people need to exercise five times as much risk discouraging the majority from engaging in achievable behaviours that are clearly good for their health. – Brigid Lynch and Paul Gardiner
I generally agree with this Research Check in that there are methodological problems with how physical activity was measured in the BMJ study.
The most significant issue with this study is the use of self-report measures that measure all domains of physical activity and are known to seriously overestimate the amount of physical activity done.
This overestimation dilutes the dose-response effects of physical activity, suggesting we need to do a lot more to achieve the same health outcomes. The study results would have been more robust if the authors had used objective measures of total physical activity (assessed through accelerometers, which are instruments measuring body acceleration) in their meta-analysis.
Several other eminent physical activity researchers have also pointed at this, and other more technical, methodological issues in their response published in the BMJ. – Corneel Vandelanotte