Disparities in Access to Video-Assisted Thoracic Surgical Lobectomy for Treatment of Early-Stage Lung Cancer

Abstract

Background:

Video-assisted thoracic surgical (VATS) lobectomy was originally described in 1992 and has slowly gained increasing acceptance. This study documents current utilization and outcomes for VATS versus open lobectomy in three Mid-Atlantic states.

Subjects and Methods:

Inpatient discharge claims data were queried for all lobectomies performed for cancer in the Mid-Atlantic United States (New Jersey, New York, and Pennsylvania) between October 2007 and December 2008. Associations between sociodemographic factors and likelihood of undergoing VATS were tested with multiple logistic regressions. Outcomes were compared using two-sample t tests.

Results:

There were 5489 lobectomies included. Forty-seven percent of patients were men, and 62% were ≥65 years old. Thirty-two percent (1741 of 5489) were VATS lobectomies. Men were less likely to undergo VATS (odds ratio 0.86, P=.004); those ≥65 years old were more likely to undergo VATS (odds ratio 1.34, P=.009). Uninsured patients underwent VATS less often (odds ratio 0.46, P=.01). Patients from rural areas were less likely to undergo VATS (odds ratio 0.67, P=.001), although travel distances for both groups were similar. A total of 44% of these hospitals (125 of 284) performed only open lobectomies. Patients without private insurance or Medicare were less likely to undergo a VATS lobectomy (P=0.01). In general, the likelihood of VATS lobectomy increased as hospital lobectomy volume increased. VATS patients had a shorter median length of stay (5 versus 6 days, P<.001) and a higher rate of discharge to home (92.5% versus 89.5%, P<.001). Mean total hospital costs were less for VATS patients ($49,948 versus $56,280, P<.001). In-hospital mortality was less for VATS patients (1.2% versus 2.1%, P=.013).

Conclusion:

For the period studied, patients in rural areas, in low-volume hospitals, who had Medicaid, or with lower median incomes underwent VATS lobectomy less often.

Introduction

Since the first reported case by Roviaro et al.¹ in 1992, the treatment of lung cancer by video-assisted thoracic surgical (VATS) lobectomy has slowly gained acceptance among thoracic surgeons. A recent review² of pulmonary resections submitted by participants in the Society of Thoracic Surgeons general thoracic surgery database reported that the percentage of lobectomies performed by VATS had increased for three consecutive years (21.6% in 2004, 28.6% in 2005, and 32% in 2006). These data were contributed by 225 surgeons, of whom only 1 was neither board-certified in thoracic surgery nor trained in a thoracic surgery training program approved by the Accreditation Council for Graduate Medical Education. The purpose of this study was to obtain a broader analysis of the evolving utilization of VATS lobectomy compared with open lobectomy for the treatment of cancer by also including patients who underwent surgery in community settings and presumably by general surgeons as well as cardiothoracic surgeons who do not participate in the Society of Thoracic Surgeons database. For this analysis, we queried a population-based dataset made up of inpatient hospital discharge claims. The designation of a new specific International Classification of Diseases (ICD-9) procedure code³ for VATS lobectomy in 2007 allowed us to identify these cases in the dataset and to compare them with open lobectomies.

Subjects and Methods

Inpatient discharge claims data were queried for all lobectomies performed for cancer in the Mid-Atlantic United States (New Jersey, New York, and Pennsylvania) between October 2007 and December 2008. Each state mandates reporting of claims in a de-identified format from all non-federal hospitals and surgery centers (in New Jersey, freestanding surgery centers are not required to participate). Databay Resources (Warrendale, PA) acquires the claims data from the individual state agencies (New Jersey Department of Health and Senior Services, New York State Department of Health Statewide Planning and Research Cooperative System, and Pennsylvania Healthcare Cost Containment Council) and produces a database product suitable for secondary analysis. The reported claims data were validated by either the individual state agencies or the participating hospitals prior to data collection by Databay Resources. Available data include demographic information (including race as documented in the medical record), clinical diagnoses, procedure codes, length of stay, disposition at discharge, and hospital charges.

Only patients undergoing procedures at non-federal hospitals within the tristate service area from October 2007 to December 2008 were included. Place of residence outside the studied tristate area was not used to exclude patients, provided the patient underwent the procedure in the service area. Patients were retrospectively identified based on the 2005 ICD-9 diagnosis and procedure codes.³ The ICD-9 procedure code for VATS lobectomy (324.1) was introduced in the third quarter of 2007. Prior to that time, patients undergoing VATS lobectomy in the dataset could not be easily distinguished from patients undergoing VATS wedge resection for diagnosis and subsequent open lobectomy because the procedure may have been coded as a thoracoscopy and lung resection. Therefore, this analysis includes only cases performed after September 2007. The study population was limited to cases with both a primary procedure code for lung lobectomy (ICD-9 procedure codes 324, 324.1, and 324.9) and a diagnosis code for neoplasm (ICD-9 diagnosis code 140-239).

Cases were geocoded to zip code centroid. Zip code of residence was available for every case. Using a publicly available geocoding tool (Excel [Microsoft, Redmond, WA] Geocoding Tool v2 [Juice Analytics, Herndon, VA] referencing Yahoo! Maps Web Services—Geocoding API [Yahoo, Sunnyvale, CA]), zip codes were assigned a latitude and longitude. Once geocoded, patients were matched to publicly available area-based sociodemographic data, specifically year 2000 census tract poverty rate (percentage of census tract at or below 100% poverty line) and county-level U.S. Department of Agriculture rural–urban continuum codes. Hospitals were geocoded to an actual street address. Using the great-circle distance formula: \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland,xspace}\usepackage{amsmath,amsxtra}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6} \begin{document}\begin{align*}d ={\rm acos}({\rm sin}({\rm lat1}) \ \cdot \ {\rm sin}({\rm lat2}) + {\rm cos}({\rm lat1}) \cdot {\rm cos}({\rm lat2}) \cdot \hbox{cos(long2} - \hbox{long1})). R \end{align*}\end{document}

a straight-line distance was calculated between a patient's place of residence and the hospital at which the procedure was performed.

Hospital lobectomy volume was calculated as the total number of open and VATS lobectomies performed for neoplasm at a particular hospital during the study period. Although all cases were included for determination of hospital procedure volume, inward border crossers (cases for which the patient's address was outside of the service area [New Jersey, New York, and Pennsylvania]) were excluded from the distance analyses in order to minimize the impact of extreme outliers. Five roughly equal-sized groups of patients were created based on lobectomy volume of the treating hospital. Integer cutpoints were chosen that allowed for the most even distribution of cases among the five groups.

Using Pearson's chi-squared tests and multiple logistic regressions, we investigated the association between sociodemographic factors (census tract poverty rate, county rurality, gender, age, race, and primary payer) and the odds of undergoing a VATS procedure. Nonparametric equality-of-medians tests were used to compare outcomes (cost, length of stay, disposition, and mortality) for open versus VATS procedures.

The study was approved by the Institutional Review Board of Fox Chase Cancer Center, Philadelphia, PA.

Results

Of the cases examined, 5489 lobectomies met the criteria for inclusion. Forty-seven percent of cases (2601 of 5489) were in men, and 62% of cases (3400 of 5489) were ≥65 years old. Thirty-two percent (1741 of 5489) of lobectomies were VATS procedures. Table 1 lists clinical and demographic differences between patients who underwent open versus VATS lobectomies. Older patients and women were slightly more likely to undergo a VATS procedure. Patients from areas of lower socioeconomic status were less likely to be treated with VATS. In the multivariate analysis, all of these factors remained significantly associated with likelihood of a VATS lobectomy (Table 2). Men, younger patients, patients without private insurance, patients from lower socioeconomic areas, and patients from rural areas were less likely to undergo a VATS lobectomy. After controlling for these other factors, there was no relationship between race and likelihood of VATS.

Table 1.

Clinical and Demographic Patient Characteristics by Lobectomy Approach

	VATS		Open
	n	%	n	%	Univariate P
Gender					.004
Male	776	30	1825	70
Female	965	33	1923	67
Age (years)					.009
≥65	1122	33	2278	67
18–64	619	30	1470	70
Race					.01
Asian	38	35	70	65
Black	107	28	280	72
Native American	21	53	19	48
Other or unknown	67	27	184	73
White	1508	32	3195	68
Rurality					<.001
Rural	127	23	425	77
Metropolitan or micropolitan	1612	33	3313	67
Payer					.01
Private	631	33	1281	67
Medicare	1019	32	2177	68
Medicaid	58	24	187	76
Other government insurance	10	29	24	71
Uninsured	9	16	45	84
Unknown payer	14	30	33	70
Travel distance (median miles)	9.3		8.4		.025
Household income (median for patient zip code)	$49,083		$44,973		<.001
Percentage poverty (median for patient zip code)	6.7		7.5		.001
Hospital lobectomy volume (median number of cases)	58		38		<.001

VATS, video-assisted thoracoscopic surgery.

Table 2.

Odds Ratio Likelihood for Video-Assisted Thoracoscopic Surgery Approach

		95% confidence interval
	Odds ratio	Upper	Lower
Gender
Male	0.86	0.76	0.96
Female	1.00
Age (years)
≥65	1.34	1.12	1.60
18–64	1.00
Race
Asian	1.17	0.77	1.76
Black	0.82	0.64	1.04
Native American	—	—	—
Other or unknown	0.78	0.58	1.05
White	1.00
Rurality
Rural	0.67	0.54	0.83
Metropolitan or micropolitan	1.00
Payer
Private	1.00
Medicare	0.78	0.65	0.92
Medicaid	0.69	0.50	0.95
Other government insurance	0.83	0.38	1.81
Uninsured	0.46	0.22	0.96
Unknown payer	0.86	0.46	1.62
Median household income for zip code
≥$50,000	1.20	1.06	1.36
<$50,000	1.00

Median hospital lobectomy volume was much higher for patients undergoing VATS, suggesting that hospital experience also influenced the type of procedure performed. Among hospitals performing lobectomies, the median number of VATS lobectomies performed was 1 (range, 0–101); 44% (125 of 284) did not perform any VATS lobectomies. Only 27% of hospitals performing lobectomies (76 of 284) performed ≥5 VATS lobectomies over the 15-month study period. The VATS-to-total lobectomy percentage (ratio of VATS lobectomies to the sum of [VATS lobectomies + open lobectomies]) varied widely among the hospitals within each volume quintile. However, the VATS-to-total lobectomy percentage increased as hospital volume increased (Table 3). This relationship then diminished for the highest-volume quintile group. Individual hospital VATS-to-total lobectomy percentages in this quintile group varied from 4.7% to as high as 70.0%.

Table 3.

Data by Hospital Volume Quintiles

Hospital lobectomy volume ^a	Number of hospitals	Number of patients	VATS/total lobectomies (%)
1–16	197	1164	20.0
17–32	44	1055	28.1
33–61	25	1064	34.6
62–139	13	1087	45.9
≥140	5	1119	30.6

Hospital lobectomy volume is presented as number of lobectomies per hospital during the study period (15 months).

VATS, video-assisted thoracoscopic surgery.

Median length of stay was 5 days (range, 1–89 days) for VATS lobectomies versus 6 days (range, 1–302 days) for open lobectomies (P<.001). Overall, mean in-hospital mortality was low (1.8%). Patients undergoing open procedures were more likely to die during the hospitalization than patients undergoing VATS procedures (2.1% versus 1.2%, P=.013). Patients undergoing open procedures were also more likely to require transfer or placement in a skilled nursing facility or other intermediate-care facility upon discharge (10.5% versus 7.5%, P=.001). Although the median cost per day was lower for open cases ($9055.70 versus $11,321.22; P<.001), the mean total hospital cost was about 13% greater for open cases than for VATS cases ($56,280.46 versus $49,948.73, P<.001).

Conclusions

We examined hospital discharge claims for the Mid-Atlantic United States from late 2007 through 2008 in order to learn more about the adoption of VATS lobectomy as compared with open lobectomy for the treatment of lung cancer. Unlike previous reports,^2,4,5 this study examines practice patterns during a more recent time period using data drawn from a broad patient population of all non-federal hospitals in the specified states.

With respect to the adoption of VATS lobectomy for the treatment of lung cancer, we found that VATS lobectomy was performed at a majority (56%) of hospitals during the time period examined. However, only 32% of all lobectomies were performed using VATS. Although the current data come from New York, New Jersey, and Pennsylvania, whereas the Surveillance Epidemiology and End Results (SEER) dataset is reported to be representative of the entire U.S. population, significant numbers of rural and community hospitals were included, so that comparison with SEER data may be valid. Farjah et al.⁴ documented an increase in the percentage of lobectomies performed by VATS from 1% to 9% in their analysis of the SEER Medicare database from 1994 to 2002. The higher percentage of lobectomies for cancer performed by VATS in the current dataset from late 2007 through 2008 may reflect increased acceptance of the procedure among surgeons and patients in the 5 years since 2002.

Boffa et al.² reported that 32% of lobectomies submitted by surgeons (99% board-certified in thoracic surgery) to the Society of Thoracic Surgeons general thoracic surgical database for the year 2006 were performed by VATS. It is not possible to reliably distinguish board-certified and/or academic general thoracic surgeons and cardiothoracic surgeons from general surgeons in the current dataset. Therefore, we could not determine if the relatively high percentage of lobectomies performed by VATS that we observed represents wider adoption at the community hospital level, an increase in the proportion of VATS lobectomy performed by thoracic surgeons at specialty centers, or both. The use of VATS lobectomy varied by hospital lobectomy volume, in that the percentage of lobectomies performed by VATS demonstrated an increasing trend until the highest quintile of hospital lobectomy volume was reached. Surgeons who work at hospitals where a high volume of lobectomies for lung cancer are performed may have an advantage when it comes to learning or adopting a new surgical technique. It may be safer to try a new technique for lobectomy in a hospital at which a high volume of lobectomies is performed and an experienced team is available to care for the patient. This impact of hospital volume on outcomes was described by Birkmeyer et al.⁶ However, it is still interesting to note that the highest volume quintile demonstrated a breakdown in the trend and instead had a significantly lower percentage of VATS lobectomies. There were only five hospitals in this quintile, and that may account for the variability observed. There are several other possible explanations for this observation. Surgeons in the highest-volume hospitals may have greater operating time constraints and might hesitate to perform VATS lobectomy because it might take longer.⁷ A different patient mix in high-volume hospitals may also explain this trend. Higher-stage patients or more complex cases (e.g., chest wall invasion, superior sulcus tumors, patients with mediastinal lymph node metastases) may be more likely to be referred to high-volume centers. The use of VATS lobectomy may be less applicable in these patients. As we emphasized previously, clinical staging data were not available from this particular database, and we cannot make comparisons about patient stage between both groups from these data.

Although the difference in the percentage of lobectomies performed by VATS based on gender (33% for women and 30% for men) was statistically significant, its clinical importance is less clear. The difference in the percentage of lobectomies performed by VATS in the elderly (33% for ≥65 years old versus 30% for those 18–64 years old) is also of unclear clinical importance. However, studies have shown that the advantages of VATS lobectomy compared with open lobectomy may be greater in the elderly.^8,9 Surgeons may be more likely to offer VATS lobectomy to elderly patients because of those published reports.

In-hospital mortality and median length of hospital stay were both observed to be lower for the VATS group than for the open lobectomy group. Although other reported studies^7,10–12 have not demonstrated an improvement in perioperative mortality, these studies did demonstrate that VATS is associated with a shorter length of hospital stay. In our dataset, the likelihood of postdischarge placement in inpatient rehabilitation programs or skilled nursing facilities was likewise lower for VATS lobectomy patients. Interpretation of these data is limited by the lack of preoperative clinical cancer staging information. This study also did not take into account the associated co-morbidities of the patients in the two groups. Without a risk factor analysis, it is not valid to conclude that the VATS approach itself confers better perioperative outcomes. It may be that patients undergoing open lobectomy had larger, more advanced cancers and had more co-morbidities than VATS lobectomy patients. This may explain why total hospital costs were less for VATS lobectomy patients.

The main findings of this study related to the disparities in the utilization of VATS lobectomy that were seen based on demographic and socioeconomic factors. The likelihood that a patient would undergo VATS lobectomy for cancer was significantly less when lobectomy was performed in a rural hospital, when the patient was uninsured or had Medicaid or other government insurance (other than Medicare), or for those with a lower median household income.

Rural hospitals may have fewer board-certified thoracic surgeons serving them, or they may have lower lobectomy volumes. Unlike the finding for many complex cancer operations where centralization of care has been demonstrated,¹³ the travel distance for patients undergoing either VATS or open lobectomy was similar, suggesting that the availability or lack of availability of VATS lobectomy had not influenced the patient's choice of hospital at the time these data were reported.

There are a few possible explanations for the decreased likelihood of VATS lobectomy seen in those with lower median income or lack of private insurance. One explanation may be cost, in that the cost of equipment used to perform VATS lobectomy determines whether VATS lobectomy is offered at hospitals that treat these patients. It is possible that patients with lower incomes may be more likely to live in rural areas, where hospitals seem to offer VATS lobectomy less often. Conversely, it is possible that patients with private insurance or higher incomes may be more likely to live in urban areas, where VATS lobectomy is more available. Another explanation is based on patient education. It is possible that patients with higher degrees of education have higher incomes and live in urban areas and are more likely to seek out and request a VATS lobectomy approach. In a recent review of a nationwide database, Gopaldas et al.¹⁴ identified this same socioeconomic disparity between VATS and open lobectomy patients to be statistically significant, whereby higher-income patients were more likely to undergo VATS lobectomy than lower-income patients. Our analysis of more recent data confirms their findings and validates some of the explanations listed above.

Some of the limitations of this analysis have already been mentioned. The reported claims data were validated by either the individual state agencies or the participating hospitals prior to data collection by Databay Resources. Errors in coding are possible. However, these errors might be randomly distributed at any given hospital, so that VATS lobectomy cases and open cases may have an equal chance of being miscoded. Similarly, procedures with an initial diagnostic intent without a preoperative established cancer diagnosis may have unintentionally excluded some patients because their cancer diagnosis may not have been coded during their hospitalization. However, again, these errors likewise might be randomly distributed, so that diagnostic VATS cases and diagnostic open cases may have an equal chance of unintentional patient exclusion or coding error. As with other administrative databases, the lack of available data on preoperative clinical cancer staging is a major limitation in the comparison of VATS lobectomy with open lobectomy with respect to mortality and other postoperative outcomes, including discharge outcome (home versus skilled nursing facility) and cost. Another limitation is that the mortality data reflect in-hospital mortality only, that is, patients who were discharged but died before 90 days were not captured in this analysis.

This analysis describes the recent practice of lobectomy for cancer in the U.S. Mid-Atlantic region. It provides a baseline description against which future data regarding practice patterns can be compared. Observed changes in procedure utilization over time may provide information regarding the adoption of new technologies in general, the impact of technologies that compete with VATS (e.g., as robotic surgery), the effects of policymakers'changes to the healthcare system, and the attitudes of patients and surgeons toward VATS lobectomy. Despite the advantages of minimally invasive surgery, we have documented disparities within our healthcare system regarding the access to VATS lobectomy. Although the etiologies of these disparities cannot be completely determined by our current analysis, further research may help to identify specific causes and possible solutions to these inequalities.

Footnotes

Disclosure Statement

W.J.S. has disclosed financial interest with Johnson & Johnson, as a private shareholder. K.B.S., P.C.S., and J.A.S. have no conflicts of interest or financial ties to disclose.

References

Roviaro

, Rebuffat

, Varoli

, Vergani

, Mariani

, Maciocco

. Videoendoscopic pulmonary lobectomy for cancer. Surg Laparosc Endosc, 1992; 2:244–247.

Boffa

, Allen

, Grab

, Gaissert

, Harpole

, Wright

. Data from the Society of Thoracic Surgeons General Thoracic Surgery database: The surgical management of primary lung tumors. J Thorac Cardiovasc Surg, 2008; 135:247–254.

International Classification of Diseases, 9th Revision, Clinical Modification, 6th. Washington, DC: U.S. Public Health Service, U.S. Department of Health and Human Services, 2005.

Farjah

, Wood

, Mulligan

, Krishnadasan

, Heagerty

, Symond

, Flum

. Safety and efficacy of video-assisted versus conventional lung resection for lung cancer. J Thorac Cardiovasc Surg, 2009; 137:1415–1421.

Rocco

, Internullo

, Cassivi

, Van Raemdonck

, Ferguson

. The variability of practice in minimally invasive thoracic surgery for pulmonary resections. Thorac Surg Clin, 2008; 18:235–247.

Birkmeyer

, Siewers

, Finlayson

, Stukel

, Lucas

, Batista

, Welch

, Wennberg

. Hospital volume and surgical mortality in the United States. N Engl J Med, 2002; 346:1128–1137.

Luketich

, Meehan

, Landreneau

, Christie

, Close

, Ferson

, Keenan

, Belani

. Total videothoracoscopic lobectomy versus open thoracotomy for early-stage non small-cell lung cancer. Clin Lung Cancer, 2000; 2:56–60.

Demmy

, Plante

, Nwogu

, Takita

, Anderson

. Discharge independence with minimally invasive lobectomy. Am J Surg, 2004; 188:698–702.

Cattaneo

, Park

, Wilton

, Seshan

, Bains

, Downey

, Flores

, Rizk

, Rusch

. Use of video-assisted thoracic surgery for lobectomy in the elderly results in fewer complications. Ann Thorac Surg, 2008; 85:231–236.

10.

McKenna RJ

, Houck

, Fuller

. Video-assisted thoracic surgery lobectomy: Experience with 1,100 cases. Ann Thorac Surg, 2006; 81:421–425.

11.

Villamizar

, Darrabie

, Burfeind

, Petersen

, Onaitis

, Toloza

, Harpole

, D'Amico

. Thoracoscopic lobectomy is associated with lower morbidity compared with thoracotomy. J Thorac Cardiovasc Surg, 2009; 138:419–425.

12.

Yan

, Black

, Bannon

, McCaughan

. Systematic review and meta-analysis of randomized and nonrandomized trials on safety and efficacy of video-assisted thoracic surgery lobectomy for early-stage non-small-cell lung cancer. J Clin Oncol, 2009; 27:2553–2562.

13.

Stitzenberg

, Sigurdson

, Egleston

, Starkey

, Meropol

. Centralization of cancer surgery: Implications for patient access to optimal care. J Clin Oncol, 2009; 27:4671–4679.

14.

Gopaldas

, Bakaeen

, Dao

, Walsh

, Swisher

, Chu

. Video-assisted thoracoscopic versus open thoracotomy lobectomy in a cohort of 13,619 patients. Ann Thorac Surg, 2010; 89:1563–1570.