How do I fix rounded shoulders from computer work?

Stand side-on to a mirror and let your arms hang. If the backs of your hands face forward and your knuckles are visible from the front, your shoulders are sitting rounded. Nearly every desk worker who tries this finds the same thing.

The standard advice is to pull your shoulders back and stretch your chest. Both are reasonable and neither is enough on its own, which is why people do them for a fortnight and give up.

Chapters 1 and 2 covered the desk and the range of positions to move through. This chapter is the one exception in the series: a specific pattern that responds to specific training, with a trial behind it that tells you how little it takes.

What rounded shoulders are, in plain terms#

A keyboard puts your hands in front of you at chest height for hours. Your shoulder blades slide forward and tip, your upper back settles into a longer curve, and your head drifts forward to keep your eyes on the screen. The whole chain moves together.

Over months, three things change. The muscles across the front of your chest sit shorter. The muscles between your shoulder blades sit longer and get less use. And the deep muscles at the front of your neck, which hold your head over your shoulders, do less of their job while the ones at the back of your neck do more of it.

The Czech neurologist Vladimir Janda described this pattern decades ago and called it upper crossed syndrome: a predictable X of tight and weak across the neck, chest and upper back. It is a clinical model rather than a proven disease, and it has been argued over since. As a map of where to put your effort, it has held up well.

Does it matter, or is this vanity?#

A fair question, given that Chapter 2 spent most of its length arguing that no single posture protects a back. The honest answer is that the neck evidence is stronger than the back evidence, and it is worth knowing exactly how strong.

A 2019 systematic review and meta-analysis in Current Reviews in Musculoskeletal Medicine, led by Ahmed Mahmoud, pooled fifteen studies covering 2,339 people. Adults with neck pain did show measurably more forward head posture than adults without it, and the degree of it correlated with their pain and disability scores.

The same review found no such association in adolescents, which is the detail that keeps this honest. Age was a confounder. A correlation in adults is not proof that the posture caused the pain, and it may well run the other way.

Why it works. Mechanically the argument is simple: the further forward your head sits, the longer the lever your neck extensors work against, and the more constant load they carry through a working day.

How you apply it at a desk. Treat this as worth ten minutes a week, not as an emergency. If your neck and shoulders are comfortable, you are training for capacity. If they ache by mid-afternoon, you are training for relief, and the trial below suggests you will get it.

The Danish trial that found the minimum effective dose#

Lars L. Andersen leads musculoskeletal research at the National Research Centre for the Working Environment in Copenhagen, and his group has spent fifteen years testing workplace exercise on people who are not athletes: office workers, laboratory technicians, slaughterhouse staff.

One of his trials, published in the journal Pain in 2011, is the one to know about. Just under two hundred office workers with frequent neck and shoulder pain, all working at least thirty hours a week, were randomised into three groups: resistance training with elastic tubing for two minutes a day, the same training for twelve minutes a day, or a control group who received general health information.

The exercise was a single movement, the lateral raise, taken to the point where the last repetitions were hard. Five days a week, for ten weeks.

Both training groups improved meaningfully against the control group. The conclusion in the paper is that as little as two minutes of daily progressive resistance training produces clinically relevant reductions in pain and tenderness.

Why it works. Two things. Load, taken close to the point of failure, is the signal that tells a muscle to adapt, and a light band held short of that gives almost none of it. And frequency puts the stimulus in front of your body five times a week instead of once, which for a small muscle group is enough.

How you apply it at a desk. Two minutes is not a metaphor. Buy an elastic band, keep it in your desk drawer, and do one hard set of lateral raises most days. Andersen's group later ran a follow-up comparing one hour a week split different ways across 447 office workers, and the short frequent version held its own. Little and often beats a weekly session you will skip.

The piece most programmes leave out#

Ann Cools, professor of rehabilitation sciences at Ghent University in Belgium, has built her research career on the shoulder blade, and specifically on how scapular muscles behave in people with shoulder pain.

Her work makes a point that changes how you should read any rounded-shoulders routine online. The shoulder blade is not a passive plate that gets dragged forward by a tight chest. It is actively positioned by muscles, chiefly the lower and middle trapezius and the serratus anterior, and in people with shoulder problems those muscles are commonly late, weak, or out-competed by an overworking upper trapezius.

Why it works. If the muscle that should pull your shoulder blade down and back is not doing its share, then stretching your chest simply gives the blade more room to sit in the wrong place. The stretch removes a restriction; something still has to do the positioning.

How you apply it at a desk. Prioritise the pulling and the low, wide movements: rows, band pull-aparts, and reaching one arm up and slightly out. Skip shoulder shrugs, which train the upper trapezius that is already working overtime. If a movement makes your neck feel tight, you are using the wrong muscle for it.

Why the screen height belongs in this chapter#

The NHS is blunt about the origin of a hunched upper back at a computer: it usually reflects a tight chest and a weak upper back, and it develops from the position the equipment forces on you.

That means the training in this chapter is competing with your setup for eight hours a day, and the setup gets far more repetitions. A laptop flat on a desk pulls your head forward every working hour, and no amount of evening band work outvotes that.

Why it works. Muscle length and motor habits adapt to the position you spend most of your time in. Change the position and you change what the training has to overcome.

How you apply it at a desk. Go back to Chapter 1 and raise your screen to eye level, then move the keyboard close enough that your elbows stay near your sides. Do that first. It is the highest-leverage part of this whole chapter and it takes ninety seconds.

The eight-week routine#

Two minutes daily, one longer session weekly, and one setup change you make once. That is all of it.

1. 1Every working day, one hard set of band lateral raises. Stand on the middle of a resistance band, hold an end in each hand, elbows softly bent, and raise your arms out to shoulder height. Slow up, slower down. Go until the last two or three repetitions are properly hard, then stop. That is the whole daily dose.
2. 2Every working day, one doorway chest stretch. Forearm flat on the door frame, elbow at shoulder height, step through gently until you feel a stretch across the front of the shoulder. Thirty seconds each side. This is the restriction-removal half, not the fix.
3. 3Two or three times a week, band pull-aparts. Hold the band in front of you at chest height, arms straight, and pull it apart until your hands are wide, squeezing between the shoulder blades. Two sets of twelve to fifteen. This is the movement that trains what Cools's work says is missing.
4. 4Once a week, add a row. A resistance band anchored in a door, a dumbbell, or a heavy bag: pull the elbow back past your ribs and let your shoulder blade travel with it. Three sets of eight to twelve, ending each set with two good repetitions left in the tank.
5. 5Once, permanently, raise the screen. Top edge at eye level, an arm's length away, keyboard close. This is the one that works while you are asleep.
6. 6Once a month, take a photo. Stand relaxed and side-on, same spot, same phone position. Eight weeks of these tell you more than how you feel on any given Tuesday.

Total commitment: about two minutes on most days and ten on one of them. Andersen's trial ran ten weeks, so give it at least eight before you judge it.

What trips people up#

Using a band that is too light. The most common error by a distance. The finding was progressive resistance training taken near failure, not gentle waving. If you could do twenty-five repetitions, the band is too easy and you are collecting minutes rather than adaptation.

Stretching only. Chest stretches feel productive because they feel like something. They remove a restriction and then hand the problem straight back to muscles that are not strong enough to hold the new position. Stretch if you like, but the training is the part that changes anything.

Shrugging instead of squeezing. If your neck is tight after a set, your upper trapezius did the work your mid-back should have. Drop the weight, keep the shoulders down and away from your ears, and think about the blade moving rather than the arm.

Expecting it in a fortnight. The trial ran ten weeks. Strength changes arrive within weeks, but the resting position of your shoulders is the sum of eight hours a day of habit, and that takes longer. Eight weeks, then look at the photos.

Stop and get it checked if you have pain that travels down an arm, pins and needles or numbness in your hands, weakness in your grip, or shoulder pain that stops you sleeping or lifting your arm overhead. Those need a physiotherapist or doctor, not a resistance band.