There are a lot of forces in play when one falls on a screamer. A lot of new studies indicate that they are of no advantage and in some instances put higher loads on an ice screw. As it has been stated before there is no substitute for a properly placed screw. Most studies that show the benefits of their use are done by the companies that manufacture them. Most independent studies show that in general they produce a neutral effect. This was discussed in great detail on Mountain Project. In particular Jim Titt and R Gold who both have considerable knowledge and credibility state that the benefits of a screamer are dubious at best.
If it provides you with the emotional confidence to climb ice in an efficient manner then I say use them. I believe that there probably aren't any negative issues with using screamers (just my humble opinion).
Below is a quote from RGold on Mountain Project:
"Unless and until somebody can provide credible tests, I think screamers are an exercise in wishful thinking. To the extent that this is true, Dolgio (and Will Gadd's) position that they are just lipstick on an inadequate protection pig seem like the only rational approach.
The various testimonials are worthless, since they do not involve repeating the same situation without the screamer to evaluate what effect the screamer did or did not have. If screamers make you feel better about your pro, go ahead and use 'em, but understand that there is no good experimental evidence for their effectiveness on anything except very short falls, and elementary physics considerations guarantee that they cannot reduce the loads significantly when long falls are involved.
The most interesting experimental results about screamers in the the climbing context are the drop tests done by the CAI (www.caimateriali.org/homepage.html
, but I can't find the article at this point) [Edit: the article is at www.caimateriali.org/index.php?id=27
], tests that suggest that screamers are not effective in reducing peak loads when their full deployment is not enough by itself to absorb all fall energy. In other words, if the screamer fully deploys and then the rope has to stretch further to stop the fall, there appears to be little benefit to having had the screamer present at all. As far as I know, the CAI tests are the only ones that used a real belayer rather than just tying off the rope to an anchor and so providing a totally static belay.
Beware of the pseudo-physics arguments in places like the Yates site, which, after admitting that there does not seem to be any good explanation for so-called "observations" about screamer performance, advance the idea that screamers work by making the fall arrest take more time. The fall isn't going to stop until all of its potential energy has been absorbed. Energy is the area under the graph of tension vs. stretch, not tension vs. time (impulse). So, for example, it takes more time to stop a long factor-1 fall than it does to stop a short factor-1 fall, but the peak loads are the same. Time isn't the issue, except in the secondary sense that it takes longer for more stretching to happen.
In the case of long falls, it is obvious without experiment that screamers will have a negligible effect, because the fraction of the fall's potential energy the screamer can absorb is small and so the majority of energy absorbtion still falls to the rope. But the CAI tests suggest that even in more moderate falls the screamer is not much help. The most optimistic perspective on screamers says that they will reduce the load on the anchor to the level obtained by a fall that is 3 feet shorter. This is good news for, say, a five foot fall, but insignificant for a 20 foot fall (waist 10 feet above the pro), and ignores the CAI testing, with real belayers, that suggests that screamers might, in low-friction conditions, short-circuit load-reducing slippage in the belay device and so end up doing nothing or even increasing peak load.
The BD tests cited upthread seem to me to confirm this, as well as reveal some fundamental misunderstandings on the part of the BD engineers involved. Gauging the seriousness of a fall by using the fall factor is appropriate when the mode of energy absorbtion is the rope, which can be treated, to a first approximation, as an elastic medium that is scalable (more or less rope can be involved in the fall arrest). A screamer is neither elastic nor scalable. It is capable of absorbing a fixed amount of fall energy, basically its activation load times the length of the extended screamer. If this fixed amount of energy is a significant fraction of the total fall energy, the screamer will be effective, otherwise not, and so the only thing that matters for screamer effectiveness is the fall height, not the fall factor. The BD tests are almost a joke, because in order to "detect" screamer effectiveness they have to use falls so short you could basically grab the pro. Those tests tell you nothing about the performance of screamers in any situation beyond the shortest aid fall, and in particular provide no evidence about what would happen for longer falls of the same fall factor.
From Jim Titt:
"With screamers there are three camps amongst the manufacturers. There are those who promote them as the answer to everything and make claims which even the most deluded find a bit dubious. Then there are those who hum and haw around the benefits with plenty of "cans" and "mays" (the Petzl website is a masterpiece for this). Then there are the ones who never got them to work, got pissed off with the whole business and donīt make them, who are incidentally in the majority.
From my chats with the technicians in industry the most positive thing you will hear is "they might help" which is not exactly a positive endorsement. Some, who have spent a long time trying different sewing patterns and cold days on the test rigs are less charitable.
Since there is no UIAA/CE test for these things all we can fall back on is our intuition, a bit of private testing, some back of the envelope calculations, information leakage from the industry and the Italian testing which was about as complete and credible as we can hope for at this stage. From all these the manufacturers claims donīt come out well and some braver souls such as myself would be prepared to brand them as hype or worse to sell an effectively worthless product.
Since the issue has been open for many years and no manufacturer has ever published any verified test results from an independent test laboratory to clearly establish where the undeniable benefits of screamers stop and the undeniable negative aspects start, then I feel the public can say or do what they think and so they should.
Time perhaps for one of the manufacturers to step up to the plate?"