Methods

The study cohort was derived from the Nationwide Inpatient Sample (NIS) database from the year 2012, a subset of the Healthcare Cost and Utilization Project sponsored by the Agency for Healthcare Research and Quality. The NIS is the largest publicly available all-payer inpatient care database in the United States, including data on approximately 7 to 8 million discharges per year and is a stratified 20% sample of discharges from US community hospitals, excluding rehabilitation and long-term acute care hospitals. The details regarding the NIS data have been previously published. Annual data quality assessments of the NIS are performed, which guarantee the internal validity of the database. Ascertainment of all diagnoses and procedures was made using the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes. Peripheral vascular disease was identified by all diagnoses codes mentioned in Supplementary Table 1 as primary diagnosis codes. Patients aged <18 years were excluded and peripheral endovascular interventions were identified using ICD-9 procedural codes 39.90, 39.50, 00.55, 17.56 ( Figure 1 ; Supplementary Table 1 ). Peripheral atherectomy was identified using the ICD-9 code 17.56 (percutaneous atherectomy of other noncoronary vessels) introduced in October 2011.

Data extraction.

The subjects were divided into 2 groups on the basis of whether they underwent peripheral atherectomy or not. The coprimary outcomes were in-hospital mortality and amputation; secondary outcome was a composite of in-hospital mortality and periprocedural complications. Preventable procedural complications were identified by patient safety indicators (PSIs), version 4.4, March 2012. These indicators are based on ICD-9-CM codes and medicare severity diagnosis–related groups, and each PSI has specific inclusion and exclusion criteria. Amputation and other procedure–related complications, which included postprocedure hemorrhage requiring blood transfusion, other iatrogenic respiratory complications (which included ventilator–associated pneumonia, postprocedure aspiration pneumonia, and other respiratory complications not elsewhere classified), postprocedural stroke, or transient ischemic attack, and other vascular complications were identified using ICD-9-CM codes (listed in Supplementary Table 2 ) in any secondary diagnosis field. Vascular complications were defined as PSI code for accidental puncture or ICD-9-CM codes for injury to blood vessels, creation of arteriovenous fistula, vascular complications requiring surgery, vascular device/graft/implant complications, and other vascular complications not elsewhere classified. “Any complications” were defined as occurrence of one or more postprocedural complications listed in Supplementary Table 2 . NIS variables were used to identify patients demographic characteristics including age, gender, and race ( Table 1 ). We defined severity of co-morbid conditions using Deyo’s modification of Charlson co-morbidity index (CCI; Supplementary Table 3 ). The Healthcare Cost and Utilization Project NIS contains data on total charges that hospitals billed for services. These data were merged with the cost to-charge ratio files to get actual cost per hospital visit.

Stata 11.0 (StataCorp, College Station, Texas) and SAS 9.4 (SAS Institute Inc., Cary, North Carolina) were used for analyses. Weighted values of patient-level observations were generated to produce nationally representative estimates. Differences between categorical variables were tested using the chi-square test, and differences between continuous variables were tested using the Student t test. Hierarchical mixed effects logistic regression models were used for categoricaldependent variables like primary and secondary outcomes and hierarchical mixed effects linear regression models were used for continuous dependent variables like the cost of care. p Value <0.05 was considered significant. In all multivariate models, we included hospital level variables such as location/teaching status of hospital, hospital region, hospital bed size and patient level variables such as age, gender, Deyo modification of CCI, admission over the weekend, primary payer (with Medicare/Medicaid considered as referent), admission type (elective admission as referent), and intervention type (no atherectomy group as referent).

To control for imbalances in baseline characteristics between the 2 study (atherectomy use and no atherectomy use) groups that might influence treatment outcome, we used propensity scoring method to establish matched cohorts. A propensity score, which was assigned to each hospitalization, was based on multivariate logistic regression model that examined the impact of 10 variables (patient demographics, co-morbidities, and hospital characteristics) on the likelihood of treatment assignment. Patients with similar propensity score in 2 treatment groups were matched using a 1 to 2 scheme without replacement using greedy methods.

Furthermore, we investigated the institutional variation in atherectomy utilization by creating 3 separate hierarchical logistic regression models: model 1: unconditional model with only hospital ID intercept; model 2: model 1 + patient level variables including age, gender, Charlson score, admission day, admission type, primary payer; model 3: model 2 + hospital level variables such as hospital region, location/teaching status, and bed size. For each model, between-hospital variance was calculated along with C statistic to account for model discrimination. Interclass correlation coefficient was calculated to determine the proportion of variance attributable to between-hospital variance. The median odds ratio was also calculated to quantify the extent to which the variation in utilization of atherectomy was secondary to clustering of patients within hospitals. All supplementary material will be available online only.

Results

Table 1 lists the baseline characteristics of the 2 study cohorts (atherectomy vs no atherectomy group). Atherectomy group had 56.03% men, 60.95% were whites, and 50.81% of the patients had a CCI score of ≥2 compared with no atherectomy group with 56.23% men (p = 0.656), 66.82% whites (p <0.001), and CCI score of ≥2 in 48.33% (p <0.001). Medicare/Medicaid was the primary payer (82.53% vs 78.44%, p <0.001). A majority of procedures were done in large (57.37% vs 64.14%, p <0.001) and urban teaching (54.99% vs 57.73%, p <0.001) hospitals. The overall rate of periprocedural complications ( Table 2 ) was 15.59%. Atherectomy use was associated with a lower rate of complications (13.2%) compared with the cohort without atherectomy use (16.31%; p <0.001). The overall rate of vascular complications was 10.9% (9.45% in atherectomy group vs 11.33% in no atherectomy group, p <0.001). The rate of amputation was lower in the atherectomy group (11.51%) compared with the group without atherectomy (13.42%; p <0.001).

Table 3 lists the baseline characteristics in a propensity-matched cohort. Patient demographics, co–morbidities (CCI score), and hospital characteristics were similar among 2 groups. Atherectomy utilization was associated with a lower rate of amputation (11.18% vs 12.92%, p = 0.029), in-hospital mortality (0.71% vs 1.53%, p 0.001), any complication (13.24% vs 16.09%, p 0.001), and any vascular complication (9.51% vs 10.98%, p 0.049). Atherectomy use was however associated with higher costs ($24,790 ± 397 vs $22635 ± 251, p <0.001).

Table 3

Propensity score match (1:2 match)

Variable	Atherectomy		P-value
	NO 5046	YES 2523
Age (years)	70±0.16	70±0.23	0.991
Female	44.0%	44.3%	0.806
Charlson/Deyo comorbidity index			0.957
0	21.2%	21.0%
1	28.3%	28.5%
>=2	50.5%	50.5%
Primary Payer			0.985
Medicare / Medicaid	82.3%	82.2%
Private including HMO	13.9%	13.9%
Self pay/no charge/other	3.8%	3.9%
Admission types			0.960
Elective admission	44.8%	44.8%
Non-elective admission	55.2%	55.2%
Admission day			0.946
Weekdays	93.7%	93.8%
Weekend	6.3%	6.2%
Median household income category for patient’s zip code (percentile)			0.968
1. 0-25th	36.1%	35.8%
2. 26-50th	25.3%	25.1%
3. 51-75th	21.0%	21.4%
4. 76-100th	17.5%	17.7%
Hospital Region			0.631
Northeast	18.5%	18.2%
Midwest or North Central	27.0%	28.5%
South	41.3%	40.5%
West	13.1%	12.8%
Hospital Teaching status			0.737
Rural	4.8%	5.1%
Urban non teaching	36.8%	36.2%
Urban teaching	58.4%	58.8%
Hospital bed size			0.941
Small	10.4%	10.7%
Medium	29.8%	29.7%
Large	59.8%	59.7%
c-index	0.59%
Outcomes
Death	1.53%	0.71%	0.003
Above-knee amputation	1.74%	1.39%	0.247
Below-knee amputation	2.93%	2.46%	0.235
Minor amputation	9.16%	8.01%	0.095
Overall amputation	12.92%	11.18%	0.029
Any complication	16.09%	13.24%	0.001
Any vascular complication	10.98%	9.51%	0.049
Any complication/in hospital mortality	16.45%	13.36%	<0.001
Cost of care	22635±251	24790±397	<0.001

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related

Related posts:

1. A US Claims-Based Analysis of Real-World Lipid-Lowering Treatment Patterns in Patients With High Cardiovascular Disease Risk or a Previous Coronary Event
2. Angiographic Characteristics of Catheter-Induced Spasm of the Left Main Coronary Artery
3. Relation of Tricuspid Regurgitation to Liver Stiffness Measured by Transient Elastography in Patients With Left-Sided Cardiac Valve Disease
4. In-Hospital Outcomes in Patients With Peripheral Arterial Disease: Comparison of 2 Treatments

Stay updated, free articles. Join our Telegram channel

Join