... Numerous problems have resulted. [From the CPT codes, Current Procedural Terminology codes used by physicians in billing, covering evaluation and management (E&M) services - ed.] The detailed guidelines often cause clinicians to overdocument, making the medical record an ineffective source of communication.
... A fundamental concern is that the office-visit descriptors and interpretive guidelines emphasize often-irrelevant elements of patients’ clinical histories and examinations, rather than decisionmaking and care-management activities. This is particularly problematic in the case of clinicians caring for patients with multiple chronic conditions.
Now EHR experts argue that the priority placed on documentation has diverted software designers’ focus from more important activities that would improve the quality and efficiency of care.  The current focus produces EHR-generated data dumps, including repetitive documentation of elements of patients’ histories and physical examinations, that merely result in electronic versions of clinically cumbersome, uninformative patient records. 
Then why are they popular? Here's why:
Studies show that EHRs pay for themselves within a few years and then generate profit partly because of facilitated coding, not greater practice efficiency. [EHR's saving the government money? A pipe dream - ed.]
Partial list of references cited in the excerpts above:
 Park T, Basch P. A historic opportunity: wedding health information technology to care delivery innovation and provider payment reform. Washington, DC: Center for American Progress, 2009. (http://www.americanprogress.org/issues/2009/05/health_it.html.)
 Hartzband P, Groopman JG. Off the record — avoiding the pitfalls of going electronic. N Engl J Med 2008;358:1656-8.
In fact, I'd made similar observations about an ED EHR in a hospital where a relative was treated. I wrote:
... I reviewed a printout from the ED system myself, and found a collection of what I call “legible gibberish” (a mass of information as if the EMR system is just a warehouse for clinical data) but no diagnosis of her problem. A nonspecific and non-useful diagnoses of “abdominal pain” was all I could find – and that was on page 8 of an 12 page printout.
... These observations and events cause me to believe your electronic medical records systems are not serving the patients and the physicians properly and could result in patient harm.
The outputs of inpatient EMR's are far worse. See my Feb. 2011 post "Two weeks, two reams." The thousands of pages of data-dumped legible gibberish I've seen has been stunning, both in terms of the ignorance of basic information science and the wasted effort and dangers to patients represented when other physicians need to refer to these old records in caring for sick people.
Berenson et al. add to our understanding of these phenomena.