Many "Instruments" or Questionnaires, have been developed and utilized over the last 30 years.  One of the first Outcomes Measures which has had widespread use as a General Health Measure to date, is the SF36.  This was created by the Rand Corporation many years ago, in order to follow the General Health Status of populations.  It is not highly specific nor highly selective in demonstrating particular disease conditions, because it was developed to evaluate general health status.  It has, however, been widely utilized because of its reproducibility and statistically scrutinized method of development. 
Many specific tools have been developed since then, in an effort to demonstrate the effectiveness of interventions for more specific disease conditions.  In Orthopaedics, these include the IKDC, the Tegner, and the Lysholm knee scales as examples.  Some of these have been found to be very helpful in measuring specific functional  improvements.  This is important in determining the effect of disease specific treatment.  Unfortunately, several measures have also been used extensively over the years which have actually been found later to correlate poorly with functional improvement.  These measures were often developed more by intuition than by valid statistical means, but have been incorporated in the body of medical literature through use. 
The collection of Validated Outcomes Measures is one necessary prerequisite to improving medical care.  This along with Patient Demographics, and Physician treatment data create a complete Outcomes Database.
In addition, it should be stressed that because of the many "instruments" available, it is common for physicians to connect several "instruments" into one, by sharing similar questions, and report on multiple scores.  Unfortunately, this is not a valid means to collect data.  Each Questionnaire must be completed individually from start to finish.  Combined Questionnaires could be proven valid if statistic validation were performed utilizing the combined measure on a large test population.  Hopefully, with improved data collection efforts, the specificity and sensitivity of the instruments themselves will be improved.  In addition, utilizing "criteria" based questionnaires, the same or more information may be garnered utilizing fewer questions.