The efficiency of your decision process is the degree to which it actually succeeds in choosing therapies that pass your cutoff threshold. The efficiency is controlled by the accuracy of your estimation of the promise of the various therapies. In this case (and the next two slides as well), I will assume you've selected the same a reasonable cutoff value. Yet the results of the filter will vary according to the efficiency or accuracy of your estimates of the promise of the various options.
The accuracy of your estimation of promise depends on your thinking, and what kinds of things you consider as well as how much evidence is available to support the therapy. Since alternative therapies generally have less evidence available to use to estimate their promise, the efficiency of decision for alternative therapies naturally tends to be less - and it's important to keep this in mind if making life and death decisions on alternative therapies. Some alternatives have more evidence than others and it will be easier to accurately guess the chance that these might help than others - while such therapies may not actually work any better than those for which the evidence is less, the fact that you have a better chance to pick them out of the sea of options makes them better bets!
In this case, I assume that your estimates of promise are completely random with respect to the actual promise of the therapies (technically having the same frequency distribution, but not correlated to the actual promise). In this case, despite choosing a reasonable cutoff value, you will pick a random selection of therapies. This makes sense: if your guesses as to the promise of therapies are essentially useless, your selection of therapies will be just random. That may sound bad, but it's actually possible to do worse!