Electrodiagnostic (EDX) testing, including electromyography (EMG) and nerve conduction studies (NCS) can provide invaluable information to ensure correct diagnosis of a variety of musculoskeletal and neurologic complaints. Therapy guided by such a diagnostic evaluation can improve the function and quality of life for veterans with such common conditions as carpal tunnel syndrome (CTS) and lumbosacral radiculopathy (LSR), while avoiding unnecessary and potentially dangerous surgery for persons with similar complaints with other etiologies. Past VHA experience with quality improvement efforts such as the National Surgical Quality Improvement Project has demonstrated that high quality, validated, clinical data is necessary (though not sufficient) to motivate active participation among clinicians.
Our previous work suggests that use of EDX procedures varies widely across VISNs in VHA, analogous to the marked geographic variation in provision of EDX by specific provider types that we found in our earlier study of private sector data. Moreover, we have found that EDX encounters at low volume facilities and those performed by non-physician providers are less likely to adhere to evidence based national guidelines published by the American Association of Electrodiagnostic Medicine (AAEM). This suggests that there may be important opportunities for improving the quality of EDX services in VHA.
In this project we have three primary objectives:
1)Examine the accuracy of administrative data for identifying electrodiagnostic procedures and whether they are guideline adherent.
2)Demonstrate that facility and provider variation in adherence to authoritative guidelines for EDX procedures is not due to clinical differences among patients - e.g., procedures being performed for subtly different clinical indications, which might not be identified using administrative data.
3)Identify and/or confirm specific patterns of guideline nonadherence, for example, the failure to perform electromyogram (EMG) studies corresponding to the nerve groups of interest.
Using our established administrative data algorithms, we identified all persons who had an EDX encounter during fiscal year 2009, characterized their indication for the procedure and determined if the testing adhered to guidelines. Using VistaWeb, we reviewed the electronic medical record (EMR) for a sample of these encounters. We selected our sample to ensure adequate numbers of persons in important subgroups defined by facility volume, provider type, and indication. Using the EMR, we determined the indication for the procedure and the specific muscles and nerves that have been studied. This EMR review was carried out by trained abstractors under close supervision by physicians with experience requesting and performing EDX studies. We are now converting this raw data into binary codes of adherent versus non-adherent to national guidelines. Next, we will determine whether facility volume is associated with guideline adherence in this smaller dataset. Finally, we will describe specific patterns of non-adherence.
We have learned several lessons about the reliability of certain VA administrative data fields regarding use of these specialized and expensive procedures. Learning these lessons has slowed our progress but will provide valuable guidance for future work by our group and others. In particular we note:
1)Almost all electrodiagnostic procedures are performed (or closely supervised by) a trained electrodiagnostic specialist - either a physiatrist or a neurologist. However, the administrative record for FY 2009 EDX procedures suggests that 2.5% are performed by non-physicians and a similar number by non-specialist physicians. Upon chart review we found that these cases were generally misattributed.
a.This pattern is related to facility. In a few facilities, such miscoding is common, in others it does not occur.
b.This miscoding appears to have decreased over time. In an analysis of FY 2005 data we have previously found that over 6% of EDX procedures were attributed to non-specialist physicians and over 5% were attributed to non-physicians.
c.This does not reflect the inclusion of trainees.
2)A substantial number of EDX procedure codes are used when no EDX procedure was in fact performed.
a.This is responsible for some of the errors identified in #1. That is, when a procedure code is assigned in error, the provider is often not a physiatrist or neurologist
b.This in turn is responsible for the apparent association of provider type with guideline non-adherence. When a provider is miscoded as a non-physician, the EDX code is often wrongly assigned. When this is the case, it is not surprising that the overall pattern of EDX codes does not meet guideline recommendations. Thus, our prior finding of an association of provider type with guideline adherence was spurious.
3)Documentation of the EDX procedures performed and their results varies greatly among facilities. In some cases, it is not possible to identify the specific nerves and muscle groups studied, and/or no quantitative data regarding the results is available in the VA electronic medical record. We are unable to determine if this adversely affects care. It does not appear to be the case that the data is simply not available in the location where we expect to find it.
We have previously demonstrated considerable variation among facilities in the extent to which administrative data document that the use of EDX procedures conforms to authoritative guidelines developed by national societies. Our current studies suggest that relying on this data may be misleading. However, our chart review documents considerable variation in how EDX are performed and documented. We anticipate that our results will provide important guidance regarding how much effort should be directed at reducing variation in this area.
None at this time.