Direct medical costs for patients with tuberous sclerosis complex and surgical resection of subependymal giant cell astrocytoma: a US national cohort study

Abstract

Objective:

To estimate direct medical costs for patients with tuberous sclerosis complex (TSC) and surgical resection of subependymal giant-cell astrocytoma (SEGA).

Methods:

This retrospective cohort study selected patients who had SEGA surgery and TSC claims between 2000–2011 from three large US nationwide claims databases. Selected patients were age 35 or less and had continuous health insurance in the year before and the year after their first SEGA surgery claim. The study examined the patients’ demographic and clinical characteristics and estimated inpatient, outpatient, medication, and total medical costs paid by insurance companies for the pre-surgery year, post-surgery year, and other study periods, respectively. Repeated measures analysis and bootstrapping technique were used to assess the impact of the surgery on the direct medical costs.

Results:

Select patients (n = 47) had a mean baseline age of 11.6 years and 66% were male. Many had seizures (91.0%), hydrocephalus (59.6%), vision disorders (38.3%), stroke and hemiparesis (36.2%), and shunt (34.0%) in the pre-surgery year. The mean direct medical costs were $8543 (inpatient: $3770; outpatient: $3473; medication: $1300) for the pre-surgery year, and $85,397 (inpatient: $71,562; outpatient: $11,497; medication: $2338) for the post-surgery year. With the exclusion of the costs during the surgery month, the inpatient, outpatient, medication, and total costs in the post-surgery year were 1.6–4.3 times as much as the costs in the pre-surgery year (inpatient: 4.3:1; outpatient: 2.5:1; medication: 1.6:1; total: 3.1:1, p < 0.05). Repeated measures analysis with bootstrapping confirmed a link between the surgery and increases in direct medical costs (p < 0.05).

Conclusions:

SEGA surgery had a substantial impact on direct medical costs. TSC patients with the surgery experienced significant post-surgery increases in their inpatient, outpatient, and medication costs. Additional research should be conducted to examine the surgery’s cost-impact in a longer duration, or to compare the cost-effectiveness of the surgery vs other treatments.

Keywords::

• Subependymal giant cell astrocytoma
• Tuberous sclerosis complex
• Direct medical costs
• Rare disease studies

Introduction

Tuberous sclerosis complex (TSC) is a rare genetic disorder that causes benign tumors to grow in the brain, kidneys, heart, lungs, liver, eyes and skin1–10. It affects as many as 25,000–50,000 people in the US and 1–2 million people worldwide8–11 through its clinical sequelae such as seizures, developmental delay, intellectual disability, autism, renal hemorrhage, and sudden death8–21.

Among TSC patients, 5–20% have subependymal giant-cell astrocytoma (SEGA)2^,9^,22^,23, which is a glioneuronal brain tumor that arises near the foramen of Monro24–27. As it increases in size, the tumor can obstruct spinal fluid flow, increase cerebral pressure, and cause headaches, blurred vision, or sudden death from acute hydrocephalus8–11^,22–27.

Surgical resection has long been a traditional treatment for SEGA1^,2^,8–11^,22–27. Because of tumor depth, the surgery is often difficult and incomplete and results in complications1^,26^,28–35. Previously published studies on SEGA surgery have mainly been anecdotal case reports due to the rarity of the disease6^,26^,28–32^,34^,36^,37.

In 2012, we published a US nationwide retrospective study that not only confirmed the surgery’s associations with its complications and SEGA-related conditions, but also described significant numbers of repeat surgeries1^,2. However, direct post-surgery medical costs for patients with the surgery remain unknown, and the relationship between the surgery and patients’ direct medical costs continue to perplex patients, physicians, payers, and researchers. To address these issues, we conducted a US nationwide retrospective study that estimated direct inpatient, outpatient, medication, and total costs, and examined the impact of the surgery on these costs.

Methods

Data sources

Data sources for the study came from the following three large US nationwide health claims databases: MarketScan commercial claims databases (2000–2011), MarketScan Medicaid national claims databases (2004–2010) and Pharmetrics commercial claims databases (2000–2011). These databases contained electronically-encrypted inpatient, outpatient, and medication claims with enrollment records for ∼60 million patients in the US. Claims from these databases had diagnosis, procedure, and medication codes with payment and utilization information. Enrollment records from the databases gave us patients’ demographic and benefit coverage information. Together, they allowed us to determine clinical conditions and treatment, as well as healthcare utilization and costs for patients with disease.

Study design, duration, and sample selection

We used a retrospective cohort study design with TSC patients who had their first SEGA surgery claim between 2000–2011. Selected patients were 35 years of age or younger and had continuous health insurance for the year before and the year after their first SEGA surgery claim. To facilitate cross-period comparison, we divided the pre-surgery year into the first and last pre-surgery periods (each period was 6 months in length), and the post-surgery year into surgery month (the first post-surgery month), the first post-surgery period (from the 2nd through 6th post-surgery month), and the second post-surgery period (from the 7th through 12th post-surgery month). Figure 1 shows the study design, duration, and sample selection.

Figure 1. Study design, duration, and selection criteria.

Measurement

We used International Classification of Disease, 9th Revision, Clinical Modification (ICD-9CM) codes and current procedure terminology (CPT) codes to identify TSC (ICD-9CM: 759.5), SEGA (ICD-9CM: 225.0 or 237.5), and other related conditions. SEGA surgeries were identified through claims that simultaneously had TSC and SEGA diagnosis codes, as well as one of the following procedure codes: CPT code: 61510, 61512, 61518, 61519, 61520, 61521, 61526, 61530, 61545, 61546, 61548; ICD-9CM Procedure Code: 01.59, 02.2, 01.2, 01.24, 01.25.

We collected patients’ gender and baseline age as their demographic characteristics, and examined whether they had a healthcare claim with one of 21 SEGA-related conditions during any study period. These conditions included seizures, hydrocephalus, vision disorders, stroke and hemiparesis, autism, sleep disorders, and others. The Appendix presents corresponding ICD-9CM codes for these 21 conditions.

Direct medical costs were in 2010 US dollars with a 3% annual discount rate to convert costs from other years into 201038. These costs included inpatient, outpatient, and medication costs paid by insurance companies for the following study periods: the pre-surgery year, post-surgery year, last pre-surgery period (6 months), the surgery month, the first post-surgery period (from 2nd through 6th post-surgery month), and the second post-surgery period (from 7th through 12th post-surgery month).

Statistical analysis

Statistical analysis included the assessment of patients’ demographic and clinical characteristics, as well as their inpatient, outpatient, medication, and total medical costs for each study period. Analysis also involved the use of a paired sample dependent t-test for examining the pre–post differences in the prevalence rates of SEGA-related conditions. Finally, we employed repeated measures analysis and bootstrapping technique to determine the impact of the surgery on the prevalence of SEGA-related conditions and direct medical costs.

Results

Study sample and demographic characteristics

The three databases contained 10,216 TSC patients. Of these patients, 421 had SEGA (4.1%) and 117 had at least one SEGA surgery (27.8% of TSC-SEGA patients). After we had excluded patients without continuous health insurance or age greater than 35 years, 47 patients remained (Figure 2). These patients had a mean baseline age of 11.6 years, and 66% were male.

Figure 2. Sample selection scheme.

SEGA-related conditions

Percentages of patients with a healthcare claim for individual SEGA-related conditions changed throughout the study periods. In the pre-surgery year, the percentages were 91.0% for seizures, 34% for shunt, 59.6% for hydrocephalus, 38.3% for vision disorders, and 36.2% for a stroke and hemiparesis. In the last six pre-surgery months, these percentages changed to 63.8% for seizure, 34.0% for hydrocephalus, 34.0% for vision disorders, and 27.7% for stroke or hemiparesis (Table 1). After the surgery, the percentages for the first and second post-surgery period increased significantly to 89.4% and 87.2% for seizure, 59.6% and 55.3% for hydrocephalus, 36.2% and 34.0% for stroke and hemiparesis, 70.2% and 72.3% for headaches, and 23.4% and 23.4% for autisms (p < 0.05 for all conditions). Repeated measures analysis with bootstrapping suggested a link (p < 0.05) between the surgery and increases in these conditions (Table 1).

Table 1. Percentages of patients with a claim for SEGA-related conditions in three different time periods.

Clinical condition	Last pre-surgery period	First post-surgery period	Cross-period difference^a		Second post-surgery period	Cross-period difference^b		p-value of fixed surgery effect^d
			%	p Value^c		%	p Value^c
Seizure	63.8%	89.4%	25.5%	0.002	87.2%	23.4%	0.010	<0.05
Hydrocephalus	34.0%	59.6%	25.5%	0.003	55.3%	21.3%	0.002	<0.05
Vision Disorders	34.0%	36.2%	2.1%	0.415	38.3%	4.3%	0.235	0.318
Blindness	2.1%	2.1%	0.0%	ND	2.1%	0.0%	ND	ND
Coma	10.6%	14.9%	4.3%	0.291	14.9%	4.3%	0.320	0.305
Deafness	10.6%	10.6%	0.0%	ND	10.6%	0.0%	ND	ND
Speech Loss	8.5%	8.5%	0.0%	ND	8.5%	0.0%	ND	ND
Headache	53.2%	70.2%	17.0%	0.008	72.3%	19.1%	0.006	<0.05
Stroke or Hemiparesis	27.7%	36.2%	8.5%	0.038	34.0%	6.4%	0.041	<0.05
Vision Loss	2.1%	2.1%	0.0%	ND	2.1%	0.0%	ND	ND
Cognitive Difficulties	14.9%	23.4%	8.5%	0.097	21.3%	6.4%	0.159	0.101
Muscle Weakness	0.0%	2.1%	2.1%	0.000	2.1%	2.1%	0.075	<0.05
Papilloedema	8.5%	8.5%	0.0%	ND	10.6%	2.1%	0.379	0.328
Balance Disorders	4.3%	6.4%	2.1%	0.074	6.4%	2.1%	0.075	0.060
Loss of Sensation	4.3%	4.3%	0.0%	ND	4.3%	0.0%	ND	ND
Nausea & Vomiting	29.8%	38.3%	8.5%	0.391	36.2%	6.4%	0.390	0.287
Depression	6.4%	6.4%	0.0%	ND	6.4%	0.0%	ND	ND
Anxiety	8.5%	12.8%	4.3%	0.331	12.8%	4.3%	0.320	0.187
Attention Deficit Disorder	8.5%	14.9%	6.4%	0.410	14.9%	6.4%	0.477	0.299
Autism	14.9%	23.4%	8.5%	0.024	23.4%	8.5%	0.025	<0.05
Sleep Disorders	6.4%	6.4%	0.0%	ND	8.5%	2.1%	0.549	0.718

^aThis cross-period difference was between the period from the 2nd through 6th month after the surgery and the period of the last 6 months before the surgery.

^bThis cross-period difference was between the period from the 7th through 12th month after the surgery and the period of the last 6 months before the surgery.

^cThis is p Value based on paired sample dependent t-test for repeated pre–post measures with a bootstrapping technique.

^dThis is p-value based on repeated measures analysis of variance for fixed surgery effect with a bootstrapping technique over the three periods.

Direct medical costs

Mean total direct medical costs per patient increased from $8543 for the pre-surgery year to $85,397 for the post-surgery year. The majority of the increase came from inpatient care ($67,792 or 88.2% of the increase) and outpatient care ($8024 or 10.4% of the increase). Only a small portion of the increase came from medication ($1038 or 1.4% of the increase). Figure 3 presents the estimates of inpatient, outpatient, medication, and total direct medical costs per patient for the first half, second half, and full year before or after the surgery. Based on repeated measures analysis with bootstrapping technique, all pre–post comparisons of inpatient, outpatient, medication, and total direct medical costs were statistically significant (p < 0.05).

Figure 3. Cross-period comparison of direct medical costs by half and full years. All cross-period differences were statistically significant based on repeated measures analysis with bootstrapping.

Figure 4 presents the evolutionary trend in inpatient, outpatient, medication, and total direct medical costs from the last pre-surgery period through the second post-surgery period. Both inpatient and total direct medical costs per patient in the surgery month were 8–15 times as much as the costs in the last 6 pre-surgery months (inpatient: $55,486 vs $3520; total costs: $58.516 vs $6777; p < 0.05). However, outpatient and medication costs per patient in the surgery month were only slightly higher or lower than the costs in the last 6 pre-surgery months (outpatient: $2729 vs $2510; medication: $301 vs $747, p < 0.05). Finally, inpatient, outpatient, medication, and total direct medical costs per patient were 34–170% higher in the first or second post-surgery period than in the last pre-surgery period (p < 0.05). Table 2 summarizes the mean estimates and standard deviations of inpatient, outpatient, medication, and total direct costs for individual study periods.

Figure 4. Cross-period comparison of direct medical costs by study period. All cross-period variations were statistically significant based on repeated measures analysis with bootstrapping. Also, the last pre-surgery period was 6 months before the surgery, the first post-surgery period was from the 2nd through 6th post-surgery month, and the second post-surgery period was from the 7th through 12th post-surgery month.

Table 2. Direct medical costs by types and study periods.

Cost types and study period	Cost estimates (n = 47)^a
	Mean	STD
Total inpatient, outpatient, and medication costs
Pre-surgery year	$8543.1	$11,187.6
First half year = First pre-surgery period (1st through 6th month)	$1766.0	$4286.2
Second half year = Last pre-surgery period (7th through 12th month)	$6777.2	$8912.3
Post-surgery year	$85,397.0	$56,258.9
First half year (1st through 6th month)	$74,155.4	$49,790.1
Surgery month	$58,515.9	$39,164.5
First post-surgery period (2nd through 6th month)	$15,639.5	$35,485.6
Second half year = Second post-surgery period (7th through 12th month)	$11,241.6	$19,298.6
Inpatient costs
Pre-surgery year	$3770.3	$7903.5
First half year = First pre-surgery period (1st through 6th month)	$249.9	$1220.4
Second half year = Last pre-surgery period (7th through 12th month)	$3520.3	$7715.3
Post-surgery year	$71,562.1	$50,506.9
First half year (1st through 6th month)	$64,973.6	$46,476.5
Surgery month	$55,485.6	$38,541.3
First post-surgery period (2nd through 6th month)	$9488.0	$33,764.8
Second half year = Second post-surgery [eriod (7th through 12th month)	$6588.5	$17,748.7
Outpatient costs
Pre-surgery year	$3472.9	$5246.7
First half year = First pre-surgery period (1st through 6th month)	$963.3	$2507.5
Second half year = Last pre-surgery period (7th through 12th month)	$2509.6	$3205.1
Post-surgery year	$11,497.2	$9971.3
First half year (1st through 6th month)	$7843.5	$7801.5
Surgery month	$2729.1	$2906.4
First post-surgery period (2nd through 6th month)	$5114.4	$6918.4
Second half year = Second post-surgery period (7th through 12th month)	$3653.6	$3898.3
Inpatient and outpatient costs
Pre-surgery year	$7243.2	$9834.9
First half year = First pre-surgery period (1st through 6th month)	$1213.3	$3523.1
Second half year = Last pre-surgery period (7th through 12th month)	$6029.9	$8263.1
Post-surgery year	$83,059.2	$55,045.1
First half year (1st through 6th month)	$72,817.1	$48,798.7
Surgery month	$58,214.7	$39,112.8
First post-surgery period (2nd through 6th month)	$14,602.5	$34,797.9
Second half year = Second post-surgery period (7th through 12th month)	$10,242.1	$19,470.2
Medication costs
Pre-surgery year	$1300.0	$2818.1
First half year = First pre-surgery period (1st through 6th month)	$552.7	$1480.7
Second half year = Last pre-surgery period (7th through 12th month)	$747.3	$1549.6
Post-surgery year	$2337.7	$4028.2
First half year (1st through 6th month)	$1338.3	$2268.0
Surgery month	$301.2	$425.0
First post-surgery period (2nd through 6th month)	$1037.0	$1978.0
Second half year = Second post-surgery period (7th through 12th month)	$999.5	$1821.5

^aCross-period differences and annual cost differences before and after the surgery were all statistically significant at p < 0.05 using repeated measures analysis and bootstrapping techniques.

Discussion

SEGA with TSC is a very rare disease. This rarity imposes considerable challenges in conducting empirical studies that examine clinical and economic outcomes of SEGA surgery among TSC patients. By combining three large US national claims databases, the current study was able to identify a sufficiently large sample of TSC patients who had the surgery. By using repeated measures analysis with bootstrapping, we estimated and compared the direct inpatient, outpatient, medication, and total costs for and across different periods before and after the surgery.

The current study confirmed the link between SEGA surgery and significant increases in direct medical costs. Different from traditional pre–post medical cost comparison studies, our study not only compared direct medical costs between pre- and post-surgery years, but also compared the costs between non-surgery periods so that the effect of cost spike during the surgery month can be removed. Surprisingly, even with the exclusion of the costs during the surgery month, the inpatient, outpatient, medication and total direct medical costs in the first post-surgery year were still 1.6–4.3 times as much as the costs in the pre-surgery year (inpatient: 4.3:1; outpatient: 2.5:1; medication: 1.6:1; total: 3.1:1, p < 0.05). We surmise that both short-term and long-term complications associated with the surgery might have contributed to the cost increase, or that post-operative increases in clinical conditions such as seizure, hydrocephalus, headache, stroke or hemiparesis, and autism might have contributed to these increased costs. Although emerging evidence suggested the former1, additional research is needed to identify underlying drivers of the cost increases, and to assess how tumor size, surgical complications, or co-morbid conditions affected the post-operative costs. Further, it is necessary to point out that some complications occurred more frequently in our study than studies in other countries23.

Like most studies using healthcare claims databases, this study has limitations. First, the sample selection criteria have excluded patients without continuous health insurance over the study periods. Second, the claims databases used in the study were mainly from large US employers, state-specific Medicaid programs and small commercial health plans; therefore, the results may not be applicable to patients with other types of health insurance, such as Medicare and federal employee health insurance programs. Third, all clinical conditions, procedures and medications identified in the study were based on ICD-9CM diagnosis codes, CPT codes or NDC codes noted on healthcare claims and were not validated through medical chart review. Therefore, clinical conditions, procedures, or medications not recorded on healthcare claims were not included in these studies. Fourth, even with adequate study design and robust analytic methods, the study was not a randomized, controlled trial and did not adjust for many possible confounding factors. Hence, robust causality between the SEGA surgery and the cost increases has yet to be established. Finally, because of data source limitation, the current study only compared the direct medical costs in the year before and after the surgery, it did not take account of long-term costs, indirect costs, or the quality adjusted life years. Therefore, future research should be conducted to examine the long-term costs, indirect costs, quality adjusted life years, and the relationship between the cost increases and surgery-related morbidity.

Conclusions

Although SEGA surgery remains a useful treatment option for TSC patients with SEGA, it was associated with significant increases in inpatient, outpatient, medication, and total direct medical costs.

Transparency

Declaration of funding

The funding of this project was provided by Novartis Pharmaceutical Corporation.

Declaration of financial/other relationships

PS is employed by Kailo Research Group, which received unrestricted research grants from Novartis. ZL is employed by Novartis and owns company stocks. MK & DK are consultants to Novartis.

Acknowledgments

The authors acknowledge the feedback from reviewers.

Appendix

Appendix Diagnosis codes for identifying SEGA-related conditions.

SEGA-related condition	ICD9-CM (4th or 5th digit may be omitted)
Seizure	3459, 7803
Hydrocephalus	3313, 3314, 7410, 7423
Vision Disorders	360, 361, 362, 363, 364, 365, 367, 368, 369, 370, 371, 372, 373, 374, 375, 376, 377, 378, 379, 921, 918, 930, 940, 870, 871, v720
Blindness	369
Coma	78001, 78002, 78003
Deafness	38843, 389
Speech Loss	4381, 31531, 3079, 7843, 78441, 7845
Headache	339, 784
Stroke or Hemiparesis	342, 431, 432, 4330, 43311, 43321, 43331, 43381, 43401, 43411, 43491, 4382, 4385, V171, V1254, 99702
Vision Loss	3699
Cognitive Difficulties	317, 318, 319, 33183
Muscle Weakness	72887
Papilloedema	377
Balance Disorders	7804
Loss of Sensation	7820
Nausea & Vomiting	7870
Depression	2962, 2963, 2980, 311, V790
Anxiety	30000, 30002, 29384, 313
Attention Deficit Disorder	3140
Autism	299
Sleep Disorders	3074, 327, 7805