Description of Network 5 Surgeon Data Analytic Methods

Medicare Part B Vascular Access Procedure Reports were prepared for all surgeons in the Network 5 region, which includes the District of Columbia, Maryland, Virginia and West Virginia. Each report is confidential and shared only with the surgeon performing the vascular access procedures reported therein. The surgeon is free to further distribute the report.  Click here to download a printable version of the analytic methods applied.

Methods

Data: Medicare Part B claims

• Contained a procedure code found in Table 1;
• Contained a diagnosis code found in Table 2;
• Were for a service occurring on or between January 1, 2009 and December 31, 2010;
• Were either paid or denied payment by CMS;
• Were performed by a surgeon classified as a general, thoracic, or vascular surgeon in the 2009 or 2010 claims; for 2009, peripheral vascular disease and cardiac surgeons were also included;
• Were submitted by a surgeon having a practice address in the District of Columbia, Maryland, Virginia or West Virginia, per the National Provider Identifier (NPI) Registry, as of January 2012; and
• Were submitted by a surgeon who performed at least one vascular access placement procedure.

Table 1: Vascular Access Procedure Codes Used to Select Medicare Part B claims, 2009 and 2010

Table 2: ICD-9-CM Diagnosis Codes Used to Select Medicare Part B claims, 2009 and 2010

Definitions

Permanent vascular access placement patterns

The following measures are reported for the surgeon, the state, and the Network in 2009 and 2010; United States measures are for 2010 only. The Network region includes the District of Columbia, Maryland, Virginia and West Virginia.

AVF % is the percent of permanent vascular access placements that were arteriovenous fistulas. AVF % = (AVF Total Placed / (AVF Total Placed + AVG Total Placed)) * 100.

AVG Total Placed is the count of vascular access placements that were arteriovenous grafts.

AVG % is the percent of permanent vascular access placements that were arteriovenous grafts. AVG % = (AVG Total Placed / (AVF Total Placed + AVG Total Placed)) * 100.

Note: Column percentages in the Report sum to 100%. Any deviations from 100% are due to rounding. For example, a surgeon may have placed 20 AVFs and 10 AVGs. In this case, the AVF % = (20/(20+10))*100 = 66.7%; AVG % = 33.3%.  Summing: 66.7% + 33.3% = 100.0%.

Numbers of vascular access procedures

The following are reported for the surgeon, the state median, and the Network median in 2009 and 2010. 

Patients is the number of Medicare beneficiaries with kidney disease for whom Medicare Part B claims were submitted for either placement of a vascular access or secondary procedure, as given by the codes in Table 1.

Important considerations

• The Medicare Part B Vascular Access Procedure Reports were produced for surgeons in District of Columbia, Maryland, Virginia and West Virginia who submitted Medicare Part B claims as described through the inclusion criteria above. There were 312 surgeons in this geographic region meeting inclusion criteria in 2009 and 296 in 2010. The state, Network, and United States (in 2010 only) averages for AVF and AVG, as well as the state and Network medians for patients, AVF + AVG, secondary procedures and CVC were calculated from claims submitted by the 312 and 296 surgeons in the respective years. Reports were prepared for the 296 surgeons who performed placements in 2010.

• The effectiveness of placed accesses cannot be ascertained with the claims data.

• Whether a vascular access is the patient’s first, fourth, or sixth placement cannot be determined with the claims data.

• Patients move or travel for services and may use more than one surgeon.  This is a surgeon driven analysis.  The data were divided using surgeon practice locations, not patient addresses.  CMS drew the 2009 claims data in 2010 and partitioned the data set per provider practice location at that time. Thereafter, CMS sent each Network data files for only those providers with practice locations within the Network’s geographic boundaries.  The same procedure was repeated with the 2010 claims. In the latter case, the data sets were constructed in 2011, partitioned per provider practice location in January 2012 and distributed to each Network, as was previously done. This means that a surgeon who moved his/her practice location from one Network geographic location to another within the relevant timeframe may not see his/her 2009 data in the Vascular Access Feedback Report produced by the Network. This happens because the data do not follow the surgeon.

• The data may be used in combination with other data for internal quality improvement initiatives.