The QRDA specification for quality measures relies heavily on the Start time of every piece of patient data. Without a precise start time there is no way to associate a specific piece of information (e.g. "Medication, Active") with a particular inpatient visit or any quality measure.

The problem is that many clinical situations arise where the precise (or even approximate) time of a particular event is unknown. For example, when the patient comes to an emergency room, they often have no idea when they started on a particular medication. It could be a month, a year or a decade. Previously, the clinician would enter "unknown", "more than a week", "some time last year", etc. These are all perfectly meaningful from a clinical perspective and often all that is really necessary. Now that the QM/QRDA requires a precise time, the EHR is going to have to require a precise time from the user, when they may not have one. This will effectively force the clinical user to enter an inaccurate time (e.g. 1/1/2012 12:00am instead of "some time last year") which might be read my a clinician downstream and misinterpreted as an accurate precise time. Paradoxically, this could result in a decrease in the quality of care.

Also, in the clinical environment, exact times are hard to come by, even when the events occur during the ED or Inpatient encounter. Documentation the the EHR may be done hours after the event and clinicians may not remember the exact time and so enter an approximate exact time. This rarely has significant repercussions clinically, but may miss the mark when calculating complex quality measure calculations that depend on exact timing.

My request to those writing and implementing QMs would be to try to minimize reliance on exact times as much as possible and realize that many of the times are going to be approximate at best.