What does the evidence show?
What does the evidence show?
When the question of how to use data to guide your training came up during a podcast discussion last week, I dusted off a familiar bromide. The output from GPS watches, power meters, heart rate monitors, and other forms of modern wearable tech is great for describing your training, I said, but not so good for prescribing it. If your pace is quicker than usual or your heart rate is lower, that tells you you’re getting fitter, but it doesn’t necessarily mean you should, say, start training harder.
I figured that was a pretty deep and meaningful statement—until I got an email from a longtime journalistic and athletic mentor, Amby Burfoot, who won the Boston Marathon in 1968 and went on to edit Runner’s World for many years. He agreed with most of my points, he said. But he had a question:
What is all our new running data good for if we can’t prescribe?
I mean, if my 10x400 workout gets slower every week, my morning heart rate increases, my hours of sleep decrease, I can prescribe from this data: It’s time to take two weeks off.
But I could assess all the above in 1970 with a $7.99 Timex wristwatch with a sweep second hand. Have we made any advances in running and training data since then?
I’ve been thinking about Amby’s question for a few days now, because it’s a big and controversial one. To some extent, I’m inclined to say that we haven’t actually made many significant practical advances in training data since 1970. Things like heart rate monitors and power meters may be helpful for beginning exercisers who haven’t yet internalized the feeling of different training zones. Elite athletes with high-tech monitoring and sophisticated scientific teams helping them interpret their data may be able to extract some useful insights. But is there any mass-market tool that offers useful and accessible training insights to enthusiastic but nonprofessional endurance athletes?
The most promising candidate I could come up with is heart rate variability (HRV), a formerly arcane measure of how regularly your heart beats that is now easily accessible to recreational athletes. If your heart is beating 60 times per minute, that doesn’t mean there’s exactly one second between each beat. Sometimes it’s 0.99, sometimes it’s 1.01, and so on. That variation in the interval between beats tells us something about balance between the sympathetic (“fight or flight”) and parasympathetic (“rest and recover”) branches of your nervous system. In a nutshell, higher variability is a sign of parasympathetic drive, indicating that you’re more recovered, while lower variability corresponds to sympathetic drive, a sign that your body is still under stress. For a more detailed primer, check out this post from Marco Altini, who developed the HRV4Training app.
Over the past decade or so, HRV has emerged as a way of monitoring training and recovery status from day to day. It can be easily measured with a chest strap or wrist-based heart rate monitor or by using the camera on your phone with an appropriate app like HRV4Training. In theory, you wake up in the morning and your HRV indicates whether you’re ready to train harder, if you should maintain your current training level, or if you need to back off. In other words, HRV prescribes your training. But does it actually work?
As it happens, I was at a conference last year where Grégoire Millet, a French researcher at the University of Lausanne in Switzerland, offered a sneak peek at some results from a test of HRV-guided training with cross-country skiers from the French national team. That study has since been published in the European Journal of Applied Physiology and offers the most rigorous real-world test of the concept that I’ve seen so far.
The protocol involved 18 skiers who took part in a 15-day training camp at the French national training center in Prémanon, sleeping in altitude chambers that simulated the thin air at 9,000 feet above sea level. Half of them followed a preassigned training plan designed by the national team coach, while the other half altered their training each day based on their morning HRV measures. Before and after the training camp, they did a VO2max test and a 10K roller-ski trial. (Another six athletes did the same training without sleeping in the altitude chambers, but we can ignore them for our purposes.)
To adjust their training, the skiers in Millet’s study followed these rules:
If their HRV stayed the same or increased, they increased their training load; for example, by skiing longer that day.
If their HRV decreased by more than 30 percent, they decreased their training load.
If their HRV decreased two days in a row, they took a rest day.
In the end, the two groups ended up doing relatively similar overall training. On average, the skiers in the HRV group modified their training due to HRV readings an average of 3.3 times. But since the modifications could be either increases or decreases, they ended up with pretty much the same total training volume.
After two weeks of heavy training, the “normal” group showed significant changes in their heart rate variability corresponding to increased “fight or flight” and reduced “rest and recover” responses. These changes, the researchers suggest, would translate in the long term to an increased chance of overtraining or getting sick. The HRV-guided group, on the other hand, kept these parameters stable. It’s an intriguing finding, even if the benefits remain mostly hypothetical.
As for race performance, the picture is less encouraging. There was no difference in VO2max changes and little difference in final 10K roller-ski times. The abstract says that the HRV-guided group was the only one to “significantly” improve performance—but their average improvement of 2.7 percent (p<0.05) is hardly different from the control group’s borderline-significant improvement of 2.5 percent (p=0.07). To be fair, it’s really hard to improve VO2max and race performance in athletes who are already elite, so finding a big difference after just two weeks was always going to be a stretch. There is one earlier study from 2007, in nonathletes, that did find improvements in performance with a similar HRV-guided training protocol.
So, after all that, the verdict on HRV-guided training is…a big maybe? I would love to see more studies like this one, trying different training approaches and following athletes for a much longer period of time. But in the real world, I doubt the best approach is a rigid algorithm that says “you must increase (or decrease) your training based on this morning’s HRV reading and nothing else.” Instead, it’s one more piece of data that you incorporate into your decision-making process, alongside all your other objective data, subjective feelings, and accumulated experience. It’s a thumb on the scale, not a yes/no decision-maker.
Altini has a bunch of practical advice (for example, here and here) on how to use HRV effectively. You should measure your HRV daily at the same time and under the same conditions (ideally right after you wake up, while lying down). Because day-to-day readings can fluctuate based on all sorts of life stresses (travel, sleep, work, etc.), watch for trends that deviate from longer-term averages (Altini’s app compares to a baseline from the previous 30 days) rather than overinterpreting a single reading. And perhaps most important: Start with a well-designed training plan. HRV data may suggest occasional deviations from that plan, but it’s not a replacement for it.
That’s a point echoed by elite triathlon scientist-coaches Dan Plews and Paul Laursen: “It’s difficult to imagine using HRV to change training on a daily basis (micro level) by itself, without other variables alongside it, and presently we wouldn’t recommend it,” they write. “However, when you use it together with other subjective measures and within the context of the training plan, it’s extremely helpful, and takes much of the guesswork away during those heavy training periods when you’re questioning whether or not to adjust the load.”
I’ve spoken to a bunch of people who work with elite athletes who, like Millet, Plews, and Laursen, believe that HRV data adds real value. I’ve also spoken with plenty who believe that while the underlying physiology may be sound, the inherent variability in daily HRV measurements makes it essentially useless for training prescription. In other words, it’s great to describe how the training has gone—to look back and, with hindsight, see exactly when things were clicking and where you went off the rails. But the evidence that you can really use it to alter your training in real time and measurably enhance performance is still thin at best.
Which brings me back to Amby’s initial question: Have we gained anything since 1970? In the recent round of interviews marking the 50th anniversary of his Boston win, someone asked Amby about the most significant new technological changes since his heyday. His answer? “Socks that wick sweat.” (For me, it was moisture-wicking tech T-shirts that rocked my world in the mid-1990s.) As prosaic as it sounds, it’ll take some doing to surpass those innovations—but HRV seems worth watching.
My new book, Endure: Mind, Body, and the Curiously Elastic Limits of Human Performance, with a foreword by Malcolm Gladwell, is now available. For more, join me on Twitter and Facebook, and sign up for the Sweat Science email newsletter.