The study linked here is one of many examples of the correlation between regular moderate exercise and mortality rate found in human epidemiological data. In animal studies it can be proven that exercise causes reduced mortality, but that is very hard to demonstrate directly in human populations - researchers can't just set up the same experimental groups and wait. So statistical methods are used, and the combination of those and the animal studies gives a good level of confidence to suggest that yes, it is a matter of exercise providing benefits rather than a matter of people more likely to live longer regardless also being more likely to exercise.
Fifteen minutes of daily exercise is associated with a 22% lower risk of death and may be a reasonable target for older adults. The authors studied two cohorts. A French cohort of 1011 subjects aged 65 in 2001 was followed over a period of 12 years. An international cohort of 122,417 subjects aged 60 was included from a systematic review and meta-analysis, with a mean follow up of 10 years. Physical activity was measured in Metabolic Equivalent of Task (MET) minutes per week, which refers to the amount of energy (calories) expended per minute of physical activity. One MET minute per week is equal to the amount of energy expended just sitting. The number of MET minutes an individual clocks up every week depends on the intensity of physical activity. For example, moderate intensity activity ranges between 3 and 5.9 MET minutes while vigorous intensity activity is classified as 6 or more. The recommended levels of exercise equate to between 500 and 1000 MET minutes every week. The authors looked at the associated risk of death for four categories of weekly physical activity in MET minutes, defined as inactive (reference for comparison), low (1-499), medium (500-999) or high (≥1000).
During the follow up there were 88 (9%) and 18,122 (15%) deaths in the French and international cohorts, respectively. The risk of death reduced in a dose response relationship as the level of exercise increased. Compared to those who were inactive, older adults with low, medium and high activity levels had a 22%, 28% and 35% lower risk of death, respectively. "These two studies show that the more physical activity older adults do, the greater the health benefit. The biggest jump in benefit was achieved at the low level of exercise, with the medium and high levels bringing smaller increments of benefit. We think that older adults should progressively increase physical activity in their daily lives rather than dramatically changing their habits to meet recommendations. Fifteen minutes a day could be a reasonable target for older adults. Small increases in physical activity may enable some older adults to incorporate more moderate activity and get closer to the recommended 150 minutes per week."