Can't I do research with an MD alone?
Sure, you can do research with an MD alone. However, the path to becoming an independent researcher may actually be much more difficult.
Basic science research today has become incredibly complex; the laboratory skills, scientific knowledge, and experimental understanding required to do good work are tremendous. Medical school curricula do not do any teaching in laboratory technique or experimental design, and they do a hit-or-miss job in teaching basic science.
Thus, the degree to which medical training prepares you for a research career is field dependent: the more basic the research, the less relevant medical training is. A rough scale of research from most clinical to most basic might be:
1) clinical trials, which involve interacting with patients, lots of paperwork and administration, and applying statistical reasoning to experimental design and analysis,
2) clinical research, involves applying simple laboratory techniques developed by others to interesting pathologies or participating as a clinical liaison in experiments designed by other researchers,
3) translational work, a hot buzzword that represents the combination of basic science and clinical research and is ready to be quickly brought into clinical practice,
4) basic science research, the real nuts and bolts, cells and molecules type of research a cell biologist or immunologist might do.
After a medical education, you will decently prepared to do the first two (after a few crash courses on statistics and experimental design), and poorly prepared for the last one. There has been much frustration over the fact that we as a scientific community do the third type, translational research, so poorly, and the MD-PhD program was created partly to train researchers with the skills to improve our ability to do that kind of research.
If you're interested in basic science or translational research, you will need more training beyond an MD. This can be done formally through an MD-PhD or PhD program, or less formally later in your career following residency. The fact is, your initial research is going to be poor, and its going to take a lot of time to build up laboratory skills, acquire background knowledge, and learn how to design a good experiment. You can do this after residency by joining a laboratory as a post-doc. The advantage of training this way is that you go through medical training earlier and get paid better for doing research. However, at that point in your career people are not going to be as interested in training you, and they're going to want results quickly. You will be competing for laboratory positions with PhD's and MD-PhD's, and while researchers will be interested in hiring you to improve the clinical relevance of their work, they will not give you as much mentoring and room to learn as if you were a graduate student. That is why I say the path to research independance can actually be harder as an MD.
Now, there are plenty of researchers who are living proof that a PhD is not mandatory to do good scientific work. Bert Vogelstein, one of the most-cited scientists of all time and one of the people who first fleshed out our current molecular understanding of cancer, never had formal PhD training. However, bear in mind that just because he had no formal training does not mean that he didn't have to work just as hard after his medical training to acquire the skills necessary to run a lab. Additionally, he and other researchers of his generation come from another era, when the learning curve for doing good basic science research was not quite as steep. Today, an MD attempting to start doing basic science work will face many obstacles.
In summary, whether you need a PhD to do research depends largely on what field you want to work in. The rule of thumb is that the more basic the work, the more likely you'll need science research training. Although this level of training can be achieved within a formal graduate program or informally later in life, budding researchers may actually find it easier and more rewarding to learn research skills from a proven method of acquiring such skills: a graduate PhD program.
I'm not convinced either way myself.
