The most appropriate answer to this issue must first match the size and the nature of the business. A fifty person business, or smaller, may have one IT person. A business with 50,000 employees may have thousands of IT employees. A financial institution absolutely should practice their disaster recovery monthly. A car repair shop could probably practice their disaster recovery once a year.
I believe that disaster recovery should be practiced frequently. Everyone in IT support should know exactly what they would do when the disaster happens. Failing to do this will result in people standing around with their thumbs up their asses saying 'What should I do?' if a disaster happens.
The fire drills can be managed in such a way as to minimize the impact on daily activity. For instance you don't have to shut down the entire IT department at one time. You can have individuals practice their part of the disaster recovery while the remaining people continue with their daily work. This would have the benefit of more closely immitating real life disaster recovery. Each person should be able to do their part without supervision when the time comes. If a disaster actually happens some people will arrive sooner than others and some people will be on vacation or otherwise unavailable to help. If each person is used to working on recovering from a disaster independently then they can begin their part as soon as they are available when the real thing happens.
My response to the person who suggested a fire drill of once a year is that people cannot be fully prepared to perform their part of disaster recovery unless they practice the procedure more frequently than once per year. I think that monthly fire drills ensure that people are fully trained at all times and that those people will be confident enough to make good decisions quickly when required.
The person that suggests that 'if it ain't broke then don't fix it' is taking the least responsible and least sensible approach imaginable. How can you know if something is or isn't going to work unless you test it? Having some pretty document describing a disaster recovery plan isn't enough. There may be practical problems with the plan that you won't know about unless you test it.
One of the most important characteristics of system administrators is responsibility. It isn't enough to think that you probably know what to do when a disaster happens and that you can just make up a recovery procedure as you go. You have to know every step that you are going to take before the disaster happens. You have to know the phone numbers that you are going to call and the people that you are going to speak to when the time comes. You have to know when you can expect to receive your off site backups. You have to know where your alternate site equipment is located and who to speak to in order to gain access to it. The only thing that you can be sure of is that if disaster happens you will be confronted with dozens of issues that you couldn't foresee unless you practice your procedure.
While the car repair shop won't have an alternate site I believe that all businesses of any size can afford to maintain off site backups. Any business can afford to rent a safe deposit box at a local bank. That approach would work very well for 99% of businesses.
In conclusion I believe that disaster recovery has to be as familiar to IT support personnel as are their passwords. There is no room for playing it by ear. That is irresponsible. Failing to practice existing disaster recovery procedures is also irresponsible. Disaster recovery plans have to be scaled to the size and nature of the business, but all businesses are capable of keeping off site backups of their business data.
