There's an old joke about med students believing they have whatever disorder they learned about that week, and there's some truth in it.
When a doctor evaluates a patient, he/she asks all the relevant questions, looks at all the available information, and makes as objective of a diagnosis as possible.
When a person self-diagnoses, he/she looks through disorders, finds one that kind-of-sort-of matches the symptoms, and then molds his or her symptoms to match the diagnosis.
The problem is that the latter happens even to professionals, and they have the background knowledge and clinical experience to know what questions to ask and what the subtle differences are between diagnoses. A lay person is working without those things, and has a much higher chance of error.
On the flip side, once someone self-diagnoses, they now go to the doctor with that in their heads. So, they consciously or subconsciously report their symptoms in a way that exaggerates what matches their disorder while minimizing the parts that don't, making it harder for the doctor to make an objective clinical judgement. A consequence of this could be a missed diagnosis and improper treatment.
Meanwhile, people have a nasty habit of getting their self-diagnoses into their heads. It's not uncommon for a patient to claim they have a diagnosis, the doctor to ask who diagnosed it, and for the patient to start with, "Oh, well, I um was looking online..." If a doctor fails to ask that, it starts appearing in medical record, and gets copied from history to history as doctors look up old visits. The family doctor faxes records to the hospital, the admitting doctor sticks it on the admission note, it gets faxed to the follow-up specialist after discharge...
Barring all that, it's often difficult for a doctor to get a self-diagnosis out of a patient's head. He or she may have to spend an inordinate amount of time trying to convince you that you don't have cancer, or you're not schizophrenic, etc. Use the Internet to get informed, but Google is not a doctor.