I am not sure that there is a consensus about their use but here is my opinion about tourniquets in remote and hostile environments.
1. Learn how to use one and practice with it.
2. Apply to stop bleeding not controlled by well-aimed direct pressure.
3. Use something wide and firm (but not hard) that can apply circumferential pressure. The pressure should be sufficient to stop bleeding. Make sure that it is in good shape and not a knock-off.
4. Place proximally (upstream) and as close to the wound as possible.
5. Don’t release in the field if the patient is in shock, has an an amputated limb, or has a wound site that cannot be monitored for re-bleeding.
6. For a long evacuation, wait an hour before trying to release it. If bleeding starts again, re-secure. Note the time and leave it in place until definitive care is reached or arrives.
7. Under dangerous circumstances, one may be applied before a thorough evaluation is possible. These should be applied to the proximal thigh or arm if there is any question about the location and/or number of wounds. Carefully check the wound when it is safe and feasible. As indicated, leave, reposition, or release it or add a second one proximally.
The following is an explanation of my above opinion. None of this should be misconstrued as a blanket endorsement to buy and carry one on all trips.
Tourniquets have a checkered history and hyperbolic claims continue to muddy the water. Past and current combat experience in the SW Asian theaters has drawn renewed attention to them because injuries to limbs have been a major source of life-threatening bleeding. There, they are being used successfully to control obvious and potentially serious bleeding. In the later case, they are applied before a proper assessment is possible e.g., multiple casualties, continued live fire. The tourniquets used are relatively cheap and can be lifesaving if used properly. As with anything in medicine, nothing works 100% of the time.
In civilian practice, it is relatively rare for death from limb bleeding to occur because properly applied, well-aimed direct pressure failed. Still, tourniquets have their use outside of theater (e.g., mass casualty), so knowing how to use one is important. The relevant questions include what, where and for how long.
A good tourniquet ought to be soft (but not mushy) and wide. Within limits, wider is better. To be effective, the circumferential pressure needs to be sufficient to stop bleeding. A sphygmomanometer (BP cuff) might be ideal except that they usually will not maintain adequate pressure for a long enough period of time. They and similarly designed devices are also bulky and fragile. The gauges break easily and the fabric, bladder and tubes are vulnerable to sharp objects. Cordage, like a rope or 550 cord (parachute), is not a good choice either because of the potential for direct skin and neurovascular injury.
There are a variety of more serviceable versions. Two of them, the CAT (combat application tourniquet) and SOFTT (special operations forces tactical tourniquet), have worked reasonably well in combat. They are compact, inexpensive and easily applied, even by the patient. Their advantages are a tradeoff for effectiveness.
One needs to have enough remaining limb to hold the tourniquet. I have heard intelligent people argue that they should never be applied to forearms and legs (lower). Generally, I disagree and experience would seem to bear that opinion out. They should be applied as close to the wound as possible. When circumstances prevent a proper assessment for location and number of wounds, some recommend using only the proximal arm (upper) and/or thigh as default positions.
People fear tourniquets because prolonged use can lead to neurovascular damage and tissue death. We know that tissue death from impaired circulation can occur in as little as two hours. We also know that tourniquets have been left on for over 16 hours without any notable harm.
Releasing a tourniquet has its own risks and there are circumstances where removal never makes sense. These later would include limb amputation, shock, the inability to monitor the wound or continued bleeding. Intermittently releasing them to temporarily restore circulation has been reported to lead to unrecognized, ongoing blood loss and patient death. On a long evacuation, if the conditions seem otherwise safe, waiting 1 hour before attempting a removal seems like a reasonable time interval. If bleeding starts again, resecure, note the time and leave it in place.
Improper application is an important cause of failure. They can also fail when they breakdown from environmental exposure or from poor construction (e.g., older version knockoff). Always check your equipment before heading out and replace anything questionable. Practice with any tool before you need it for a real emergency.
There are plenty of good resources online that cover step-by-step application and the identification of knockoffs (e.g., date printed on webbing, red tip on the end of webbing).