|
|
| Research
>
Vascular Access |
Trend Analysis
Introduction
Vascular access services were examined for the years 1991 through 1996. Data sources included Inpatient Institutional SAFs, Outpatient Institutional SAFs, and Part B Physician Supplier SAFs. Vascular access services were identified by CPT-4 codes (Part B, Outpatient Institutional) and a combination of Principal ICD9-CM diagnosis codes, DRG codes, and Principal ICD9-CM procedure codes (Inpatient Institutional). The codes used are listed in Table I and Table II. Part B and Outpatient codes were grouped according to type of service, while an effort was made to identify "pure" vascular access admissions (i.e., hospitalizations that were specifically related to vascular access) for Inpatient services.
Methods
Part B was analyzed directly, by accumulating all of the claims that were coded with the CPT codes found in Table I. Net payment amounts, place of service, and specialty codes of the providers were obtained from the Part B SAF. No restrictions regarding modality or Medicare as secondary payer were applied.
Inpatient vascular access costs were identified using a stepwise procedure. Qualifying admissions (those with DRGs listed in Table II) were further examined to determine the principal diagnosis code. If a vascular access diagnosis code was found, the admission was included in the analysis. Admissions with non-vascular access diagnosis codes were then examined for the presence of vascular access principal procedure codes. Finally, if a vascular access principal procedure code was located, the further restriction that the procedure date must have been within 5 days of the admission date was applied to exclude those cases where an admission for a reason other than vascular access included a vascular access procedure. If all conditions were met, the admission was considered a "pure" vascular access admissions.
Outpatient vascular access costs were evaluated in a fashion similar to the Part B analysis. Outpatient claims with a vascular access CPT coded service in any of the revenue lines were included. Outpatient claims with vascular access CPT codes often included non-vascular access related services, so an attempt was made to estimate the vascular access related costs. Since the revenue lines do not contain payment amounts, costs were apportioned using the Milliman and Robertson method (the claim payment amount was multiplied by the ratio of submitted charges for the vascular access revenue line to submitted charges for the entire claim to arrive at an estimate for the vascular access related service).
Table I: Vascular Access CPT codes
CPT-4 Code Description Group
36800 Insertion of cannula (venous-venous) Insertion Permanent Access
36810 Insertion of cannula (AV, Scribner type) Insertion Permanent Access
36820* AV anastomosis (Cimino type) Insertion Permanent Access
36821 AV anastomosis (Cimino type) Insertion Permanent Access
36825 Creation of AV fistula, autogenous graft Insertion Permanent Access
36830 Creation of AV fistula, nonautogenous graft Insertion Permanent Access
36835 Insertion of Thomas Shunt Insertion Permanent Access
36011 Selective catheter placement, first order branch Insertion Temporary Access
36012 Selective catheter placement, second order branch Insertion Temporary Access
36488 Placement of central venous catheter, percutaneous(Age 2 years or under) Insertion Temporary Access
36489 Placement of central venous catheter, percutaneous(Over age 2 years ) Insertion Temporary Access
36490 Placement of central venous catheter, cutdown(Age 2 years or under) Insertion Temporary Access
36491 Placement of central venous catheter, cutdown(Over age 2 years ) Insertion Temporary Access
36533 Insertion of implantable venous access port Insertion Temporary Access
35190 Repair AV fistula, extremities Revision/Removal
35876 Thrombectomy of graft, with revision Revision/Removal
35900* Excision of infected graft extremities Revision/Removal
35903 Excision of infected graft extremities Revision/Removal
35910* Excision of infected graft extremities Revision/Removal
36534 Revision of implantable venous access port Revision/Removal
36535 Removal of implantable venous access port Revision/Removal
36815* Revision of AV fistula Revision/Removal
36832 Revision of AV fistula Revision/Removal
36834 Plastic repair of AV aneurysm Revision/Removal
37190* Plastic repair of AV aneurysm Revision/Removal
37607 Ligation or banding of AV fistula Revision/Removal
M0900* Excision without graft Revision/Removal
35460 Transluminal angioplasty, venous Angioplasty
35476 Transluminal angioplasty, percutaneous, venous Angioplasty
75978 Transluminal angioplasty, radiological supervision and interpretation Angioplasty
35875 Thrombectomy of arterial or venous graft Declot
36860 Cannula declot, without balloon catheter Declot
36861 Cannula declot, with balloon catheter Declot
37201 Transcatheter therapy, infusion for thrombolysis Declot
75896 Transcatheter therapy, infusion for thrombolysis, radiological supervision and interpretation Declot
37205 Transcatheter introduction of stents Stents
75960 Transcatheter introduction of stents, radiological supervision and interpretation Stents
36005 Injection procedure for contrast venography Angiography
36145 Introduction of needle or intracatheter, AV shunt Angiography
75790 Angiography, radiological supervision and interpretation Angiography
75820 Venography, radiological supervision and interpretation Angiography
00532 Anesthesia, central venous catheter Anesthesia
01784 Anesthesia, congential or acquired AV fistula Anesthesia
01844 Anesthesia, shunt or revision Anesthesia
* Deleted code
Table II: Vascular Access Inpatient codes
DRG Description
112 Percutaneous cardiovascular procedures
120 Other circulatory system OR procedures
315 Other kidney and urinary tract OR procedures
442 Other OR procedures for injuries with comorbid condition
443 Other OR procedures for injuries without comorbid condition
478 Other vascular procedures with comorbid condition
479 Other vascular procedures without comorbid condition
ICD9-CM Description
Diagnosis
996.1 Mechanical complications of vascular device, implant, and graft
996.62 Infection and inflammatory reaction due to vascular device, implant and graft
996.73 Other complications due to renal device, implant, and graft
ICD9-CM Description
Procedure
38.95 Venous catheterization for renal dialysis
39.27 Arteriovenostomy for renal dialysis
39.42 Revision of arteriovenous shunt for renal dialysis
39.43 Removal of arteriovenous shunt for renal dialysis
39.93 Insertion of vessel-to-vessel cannula
39.94 Replacement of vessel-to-vessel cannula
-------------------------------------------------------------------
New Access Analysis
Patient selection
All patients included in Part B (PB) SAF for 1996 were considered for an incident-based new access analysis.
Patients with a vascular access insertion CPT code during the first six months of 1996 were included (CPT codes: 36533, 36800, 36810, 36821, 36825, 36830, 36835).
Patients were followed through 12/31/96.
Data construction
The first occurrence of a vascular access insertion code before 7/1/96 was considered an inclusion event. Any vascular access complication CPT codes occurring before the inclusion event were excluded.
All vascular access CPT codes (complication or insertion) occurring after the inclusion event and before 12/31/96 were included in the analysis.
Once Part B vascular access CPT codes were identified, an attempt was made to match Inpatient institutional (IP) and Outpatient institutional (OP) claims to the Part B claims in such a way as to consider only "pure" vascular access hospitalizations and outpatient claims.
Place of service code was used to determine where to look for a matching claim, and date of service was used to match claims. For OP claims, a match was defined as a claim with a date range inclusive of the PB service date, a matching CPT code, and a payment amount greater than zero. For IP claims, two sets of matching criteria were adopted. For insertion CPT codes, a match was defined as a claim with a date range inclusive of the PB service date, a vascular access related DRG code (112, 120, 315, 442, 443, 478, 479), and a vascular access related Principal Procedure ICD9 code (38.95, 39.27, 39.42, 39.43, 39.93, 39.94). For complication CPT codes, a match was defined as a claim with a date range inclusive of the PB service date, and a Principal Diagnosis ICD9 code indicating vascular access complications (996.1, 996.62, 996.73) In addition, the admission date for any vascular access hospitalization was required to be no more than 10 days before the PB service date of the vascular access CPT code in order to qualify as a "pure" vascular access hospitalization. (Since the DRG and Procedure Code combination is not entirely foolproof, we assume that if the vascular access procedure has not been performed within 10 days of admission then the admission was not a "pure" vascular access admission.) Both sets of IP criteria also require a payment amount greater than zero.
Once matching IP and OP claims were collected, an attempt was made to consolidate PB
claims, with their matching IP and/or OP claims, into "vascular access events". A vascular access event is a collection of PB claims whose service dates occurred during the pre-determined event duration. An event duration was defined as:
The date range of an IP claim matching the current PB claim, or
The date range of an OP claim matching the current PB claim, or
The date range of the current PB claim
Vascular access events were categorized as :
Insertion only only insertion CPT codes were located during the event
Insertion + complications both types of CPT codes were located during the event (no distinction was made regarding the order of CPT codes)
Complications only only complication CPT codes were found during the event
Once an event was defined, submitted charges, allowable, and payment amounts of all Part B claims were totaled for the event. While a vascular access event may have had multiple PB claims, only a single dollar amount for submitted charges, allowable, and payment amounts for each matching IP and/or OP claim was included to avoid double counting.
Data analysis
General totals of events were computed by event type , place of service , IP events with matching IP claims and OP events with matching OP claims , as well as a distribution of total events per patient and procedures per event.
Financial Event totals for submitted charges, allowable and payment amounts were computed for each event type. IP analysis only considered those events that had matching IP claims (33.0% of all IP events). OP analysis was limited to only those events which had matching OP claims (40.7% of all OP events). Finally, a Part B total was computed for all events.
CPT code analysis occurrence of specific CPT codes was analyzed. The top ten CPT codes by count and by total allowable dollars were identified for insertions and complications. However, this analysis was carried out at the PB claim level rather than the vascular access event level, as each event had the possibility of multiple CPT codes. The same analysis was carried out for IP matched events and for OP matched events.
Specialty Analysis the top ten specialties performing vascular access procedures were identified both by count (Table 11) and by total allowable dollars (Table 12) for each event type. This analysis was also carried out at the PB claim level. The same analysis was conducted for IP matched events and for OP matched events.
Graft Survival:
Defining the Patient Cohort
Hemodialysis patients with any simple fistula (CPT-4 code 36821), autogenous graft (CPT-4 code 36825), and synthetic graft (CPT-4 code 36830) insertion from 1/15/94 to 6/30/94 were included. We only counted once if a patient had multiple graft insertion billing codes on the same day. An entry period of four weeks was defined as 14 days before and after the date of synthetic graft insertion. All patients who had a graft removed, a new graft placed, a change of modality, or who died during this two-week period were excluded. Remaining patients were categorized into 4 groups:
graft only
graft + permanent catheter(s) (CPT-4 codes: 36533 and 36800) placed in the two-week entry period
graft + temporary catheters (CPT codes: 36489 and 36491)
graft + both permanent catheter(s) + temporary catheter(s).
All vascular access or catheter interventions were considered to be outcome events (i.e., time to first intervention, time to final access removal, or new access replacement). The final group included a total of 24,471 patients.
Defining Time to Events
The analysis of vascular access failure was performed by looking at the time to first event, which was based on the date associated with a CPT-4 procedure code captured from the HCFA Part B claims file. These codes and dates defined the primary outcome of intervention-free use of the device during the period of 1/15/94 to 12/31/94. The study start date for a given patient was defined as the graft insertion date.
Time to first event (total event-free access survival) was defined as the number of days from the study start date to the earliest of the following CPT -4 coded services:
the first revision (35832, 35190, 36534, 36834, or 37607)
the first declotting (35875, 35876, 36860, 36861, 37201, or 75896)
the first use of Urokinase (J3364, or J3365)
the first angioplasty (35460, 35476, or 75978)
the first stent (37205 or 75960)
the first excision of infected graft (35903)
the removal of the 1st graft (37799)
the insertion of the 2nd graft (36821, 36825, or 36830)
after the first "qualifying" bridge catheter, any additional hemodialysis catheter placed (36489, 36491, 36533, or 36800)
Patients were censored for modality change (transplant or peritoneal dialysis), death, or end of the study (December 31, 1994). We chose to censor at death, instead of treating death as a graft failure, in order to focus specifically on the mechanisms of access failure rather than on patient outcome.
Time to final event (regarding total patency, i.e., total graft life) was defined as the number of days from the study start date to the removal of the 1st graft or the insertion of the 2nd graft (whichever came first), with censoring for modality change, death, or end of the study.
Time to death (regarding patient survival) was defined as the number of days from the study start date to the earliest date of death, censor for modality change, or end of study.
Survival Analysis
Access failure rates, which are expressed in terms of first events or final events per 1000 patient years, are specific to patient age, race, gender, primary renal diagnosis, prior ESRD exposure time (01, 12, 25, >5 years, and no prior exposure), ESRD network, and comorbid conditions. Comorbid conditions are defined from ICD-9 codes extracted from Part A and Part B claims prior to the internal access insertion date. These include atherosclerotic heart disease (ASHD), congestive heart failure (CHF), peripheral vascular disease (PVD), cerebrovascular accident/transient ischemic attack (CVA/TIA), other cardiovascular disease, chronic obstructive pulmonary disease (COPD), and cancer (CA), The "Diabetic" disease category consisted of those patients whose primary cause of ESRD was reported to be diabetes, while the "Non-Diabetic" category included patients diagnosed as having hypertension, glomerulonephritis, or "other."
Time to first event and time to final event were both right censored. Kaplan-Meier survival curves were generated for the four study groups and compared using log-rank and Wilcoxon tests. A Cox proportional hazards analysis was performed to assess the joint impact of demographic characteristics, underlying renal disease, prior ESRD exposure time, and comorbid conditions on vascular access survival. Finally, we adopted the Andersen-Gill multiplicative hazards model to evaluate multiple event outcomes. All types of events and their corresponding occurrence time were counted within the period of total graft life time. We also adjusted for the same set of covariates as in the Cox regression in this Andersen-Gill model.
Financial Analysis
In our financial analysis we calculated the average Medicare payments per member month related to the four groups. We estimated the time and aggregate cost for each element of Part A and Part B claims based on the total patency of graft for each patient. Medicare reimbursements for insurance claims that spanned the study start date or stop date of the analysis were prorated using the average cost ($) per patient month between the from date and the to date on the claim. We defined the from date as 5 days before the graft insertion date and the to date as 10 days post implant. We assumed that hospital stays beyond these -5 and +10 days were for reasons beyond the access related procedure, since most hospital stays with access procedures (e.g., revisions) were within this period. We therefore prorated reimbursements in the claim bills according to the above from or to date, based on the average cost per patient month over the full period covered by the claim.
|
| <
Back |
|
|
