Rehabilitation utilization in malignant primary brain tumors compared to stroke and traumatic brain injury: Analysis using a large claim database

Abstract

Background

There is concern regarding the underutilization of rehabilitation services for the malignant primary brain tumor (MPBT) population following hospitalization. Our aim is to assess physical therapy (PT), occupational therapy (OT), and speech-language pathology (SLP) use after an MPBT diagnosis, evaluate the trend from 2001 to 2018, and compare to traumatic brain injury (TBI) and stroke.

Methods

Adult cases of MPBT, TBI, and stroke were extracted from MarketScan database. Inpatient and outpatient data were screened for inpatient rehabilitation use at the time of diagnosis and postdischarge outpatient PT, OT, and SLP over 12 months. Generalized linear regressions were used for analysis.

Results

The cohort was composed of 3381 MPBT, 205 366 stroke, and 24 825 TBI patients. After diagnosis, 1% of MPBTs were discharged to skilled nursing facilities (SNF) and 3% to inpatient rehabilitation facilities (IRF). Rehabilitation use at 12 months was 19% PT, 8% OT, and 6% SLP. These percentages were lower than stroke and TBI; stroke: 8% SNF, 8% IRF, 22% PT, 10% OT, and 8% SLP; TBI: 7% SNF, 7% IRF, 22% PT, 8% OT, and 6% SLP. Outpatient therapies increased from 2001 to 2018, with PT use consistently higher than OT and SLP. MPBT had the greatest increases in OT (7.95 times) and PT (3.89 times) compared to stroke and TBI, while stroke had the greatest increase in SLP (0.98 times).

Conclusions

MPBT patients had the highest increase in OT and PT utilization when compared to stroke and TBI. However, there remains a utilization gap which demonstrates the need for improvement.

Strokes, traumatic brain injuries (TBI), and neoplasms are 3 of the most prevalent health conditions associated with disability, globally and in the United States.¹ Primary brain tumors, consisting of benign (70%) and malignant (30%), account for 85–90% of primary central nervous system (CNS) tumors.² Given the impact brain tumors have on neurological and cognitive function, they have shown similar overall rehabilitation care needs to those with stroke and brain trauma.³ Although these conditions have different etiologies, they share similar complications and have shown similar functional improvements after rehabilitation discharge^4–6

Among individuals with brain tumors who receive acute rehabilitation, the most common neurologic complications are cognitive deficits (80%), weakness (78%), and visual-perceptual deficits (53%). Of this brain tumor population, 75% present 3 or more deficits and 39% have 5 or more.⁷ In contrast, outpatient rehabilitation utilization is less studied.^8,9 The size and anatomical location of brain tumors create functional impairments, which create a direct impact on the need for rehabilitation. Additionally, underlying comorbidities and side effects from the required treatment interventions, such as surgery, radiation, chemotherapeutics, and/or additional medical interventions may contribute to or create a need for rehabilitation.⁸ Rehabilitation for brain tumors is primarily focused on functional recovery and/or management of deficits in mobility, gait, transfers, cognition, speech, swallowing, activities of daily living (ADLs), and instrumental or independent activities of daily living. These impairments impact daily function, independence, and quality of life for patients and caregivers, and can put a strain on psychosocial, financial, emotional, and vocational health.¹⁰

The rehabilitation needs in brain tumor patients may occur upon diagnosis, during initial or later treatments, or late effects. However, the functional impairments and deficits are similar to those found in other more prevalent brain pathologies, including stroke and traumatic brain injury patients. Regardless of their diverse needs for rehabilitation, literature shows similar functional improvements between these 3 patient populations. O’Dell et al.¹¹ compared individuals with brain tumors undergoing inpatient rehabilitation with TBI individuals. Within the brain tumor population, greater functional gains were made for persons with meningioma, left-sided cerebral lesions, and those not receiving radiation therapy; yet they found similar efficiency of functional gains when comparing brain tumors versus TBI. Another study matched brain-tumor individuals by location of lesion and age with TBI individuals and found that brain tumor individuals admitted to inpatient rehabilitation can achieve comparable functional status compared to TBI individuals, although they had less total change, they achieved similar rates of functional rates, shorter LOS and rehabilitation needs.¹² Additionally, a retrospective study showed that regardless of malignancy, individuals with brain tumors who received intensive rehabilitation after surgery improved motor, cognition, and ADL function; these improvements were similar when compared to individuals with stroke.⁵

Despite rehabilitation being shown to improve function and quality of life after brain tumor diagnosis similar to TBI and stroke,¹³ there is concern regarding its utilization in both the inpatient and outpatient settings.^1,14 Existing literature suggests potential causes of an underutilization of rehab services including lack of awareness of its benefits for the cancer population and the secondary lack of screening for rehabilitation needs in the oncology setting.^15,16 Other reasons to consider are the gradual onset of impairments leading to delayed recognition of a need for intervention and the lack of functional reports of decline by the patients. There are also access barriers such as the geographic location of cancer rehabilitation speciality services, financial costs of rehabilitation, health literacy, adherence to rehabilitation appointments and exercise, and lack of insurance.¹⁶

In a small single institutional study, Burgess et al.¹⁷ demonstrated variability in utilization, as 100% of patients with brain tumors received PT, 57% received OT, and 13% received SLP. PT addresses deficits related to mobility and transfers; OT addresses the performance and participation in daily occupations (ADL, IADL, health management, work, leisure, etc.); and SLP addresses communication, swallowing, and cognition.¹⁸

Compared to the literature for brain tumors, there is a wider breadth of literature demonstrating the benefits of multidisciplinary rehabilitation for patients with stroke and traumatic brain injury. This is likely due to smaller incidence and prevalence of brain tumors. In fact, stroke, and traumatic brain injury are the most common types of acquired brain injuries that occur with an annual incidence of 15 million people worldwide for stroke¹⁹ and approximately 1.7 million people in the United States for TBI^20–22 There are publications reporting brain tumor patients to make functional gains in the acute and postacute care arenas, similar to the gains made by stroke patients.^23–26

This study aims to describe the current rehabilitation utilization of all 3 patient populations with a national database and to review and understand the rehabilitation utilization for brain tumor patients from a national perspective and to evaluate the trend of utilization over the years. We also wanted to describe rehabilitation utilization amongst the various disciplines, including PT, OT, and SLP, as we recognize the importance of multidisciplinary rehabilitation utilization for brain tumor patients. Specifically, we evaluated the rehabilitation use of PT, OT, and SLP over a 12-month timeframe after the initial hospitalization for patients with malignant primary brain tumors (MPBT) in comparison to those with a TBI or stroke. It is important to note that rehabilitation needs may occur at variable times for patients with all 3 brain pathologies evaluated. Patients of all 3 diagnoses may have immediate functional impairments that resolve within hours to days with medical management and do not require rehabilitation services initially. Some patients have rehabilitation needs that occur later in their ongoing medical surveillance and management, while others may have immediate rehabilitation needs and may persist for a prolonged period or a lifetime. We predicted an increase in rehabilitation services usage for brain tumor patients from 2001 to 2018 due to increasing research and publications on the needs of brain tumor patients during this timeframe.

Materials and Methods

Data Source

The Merative MarketScan Research Database was used for this project. This data is built from paid claims by employer-sponsored insurance, COBRA, private insurance managed Medicaid and Medicare Supplemental (also known as Medigap). It is a comprehensive database in which more than 100 payers contribute data. Insurance holders and their dependents enter this database when enrollment starts, and they exit when that coverage ends. While enrolled, their continuum of care is recorded along with corresponding payment and captured in different de-identified datasets such as inpatient, outpatient, medications, and others with a link through a unique encrypted ID.²¹ This data has been extensively used for medical, public health and epidemiology research for decades.²¹ We have a custom subset of this database with individuals who have experienced a neurological/neurosurgical condition at some point from 2000 to 2021. For this study, we restricted to records of 2000–2019, strategically excluding COVID-19 years of 2020 and 2021 as they have impacted healthcare utilization and may affect the real-world rehabilitation use that was of interest in this study.

Ethical Considerations

MarketScan is a large claims de-identified database and therefore informed consent was not needed. Institutional Review Board approval was obtained (IRB #10.0559).

Sample Selection and Time on the Study

Individuals 18 years and older with a primary diagnosis of traumatic brain injury, stroke, or brain tumor were extracted from inpatient tables. The claim codes used for extraction are in Table 1. To ensure an incidence cohort, we restricted to those who had at least 12 months look-back continuous enrollment. The prediagnosis look-back enrollment time was calculated as prediagnosis look-back time = diagnosis hospitalization admission date –start enrollment date (or first claim date in the data set). For each condition, retained individuals were to not have had any of the other conditions in the previous year or for 1 year follow-up. For example, those with TBI who had a claim of either stroke or brain tumor in the year prior or post to the hospitalization with TBI were excluded. Similarly, those with stroke who had TBI or brain tumors in the year leading to or following the stroke, and those with brain tumors who had TBI or stroke in the previous or following year were all excluded. For brain tumors, those with claims of metastasis to the central nervous system in that period were also excluded to focus on primary brain tumors only. An inclusion and exclusion flowchart are portrayed in Figure 1. If the patient had multiple hospitalizations satisfying the above inclusion/exclusion criteria, the first hospitalization one was flagged as the index diagnosis hospitalization and was set as the beginning of follow-up time.

Table 1.

Claim Codes Used to Define Traumatic Brain Injury (TBI), Stroke, and Brain Tumor, and the Codes Used Define Physical Therapy (PT), Occupational Therapy (OT), Speech-Language Pathology Therapy (SLPT)

	Claim Codes
Conditions	Diagnosis Codes
Traumatic Brain Injury (TBI)	ICD-9: 800, 801, 803, 804, 850-854, 950.1-950.3
	ICD-10: S02.0, S02.1, S02.8, S02.91, S04.02, S04.03, S04.04, S06, S07.1
Stroke	ICD-9: 362.3, 430, 431, 433, 434, 435, 436, 437.0, 437.1
	ICD-10: H34, I60, I61, I63, I65, I63, I66, G45, I67.89, I67.2, I67.81, I67.82, I67.89
Brain tumor	ICD-9: 191
	ICD-10: C71
Metastasis to the Central nervous system	ICD-9: 198.3 ICD-10: c79.3, C79.4
Rehabilitation use	Procedure codes
Physical Therapy (PT)	CPT-4: 97161, 97162, 97163, 97110, 97530, 97164
Occupational Therapy (OT)	CPT-4: 97165, 97166, 97167, 97168
Speech-Language Pathology (SLP)	CPT-4: 92523, 92610, 92522, 97129, 92507, 92526

Figure 1.

Inclusion and Exclusion flowchart.

Patient Characteristics

Characteristics at the time of index diagnosis hospitalization were noted as baseline. These include age, sex, year of index hospitalization, insurance type (commercial, Medicaid, and Medicare) and comorbidities. For comorbidities, the Elixhauser comorbidity score²⁷ computed using an adaptation to ICD-9-CM and ICD-10 codes developed by Quan et al.²⁸ was used. The Elixhauser comorbidity score is a count of comorbidities ranging from 0 (no comorbidities) to 31 (the total number of chronic conditions accounted for by the score). For this study, we categorized the Elixhauser into 4 categories: 0, 1, 2, and 3 or more.

Study Outcome

The main outcome of this study is rehabilitation use which was evaluated at discharge from index hospitalization (to skilled nursing facility (SNF), inpatient rehabilitation facility (IRF), or postdischarge outpatient rehabilitation only) and within 12 months following discharge from the index diagnosis hospitalization. For this reason, those who did not have 12 months or more of continuous insurance coverage after discharge were excluded. Postdiagnosis enrollment time was calculated as: postdiagnosis follow-up time = end enrollment date (or last claim date in the data set) – diagnosis hospitalization discharge date. For this, 12 months of postindex hospitalization discharge claims were screened for physical therapy (PT), occupational therapy (OT), and speech-language pathology (SLP). Claim codes used to search are in Table 1. These were evaluated at 3-, 6-, and 12-months. A secondary outcome was the trend of rehabilitation use over the years and a comparison of trends for different therapy types across different types (PT, OT, and SLP) and conditions (TBI, stroke, and brain tumor).

Statistical Analysis

Age was summarized with median and interquartile range (25th–75th percentiles). It was not normally distributed per the Smirnov–Kolmogorov test, a quantile regression to the median with the group as the only factor was used to obtain the 3-group comparison as well as the 2-by-2 comparisons (stroke vs brain tumor, TBI vs brain tumor, and TBI vs stroke) which were obtained through linear contrasts estimates. Sex, insurance type, Elixhauser category, and rehabilitation use were summarized using counts and percentages, and compared across the 3 groups using a logit regression model with group as the only factor. Two-by-two comparisons were obtained by linear contrasts on the group factor. The trend analysis was performed with a generalized linear regression model with a binary distribution and a logit link. The model included the condition group, the year, and their interaction.

Since the characteristics were statistically different across the 3 conditions, a sensitivity analysis was performed by comparing rehabilitation use on matched groups. The matching process was performed by matching those with stroke to those with tumor, then those with TBI to those with tumors separately, and then combining those matched to the same person with a brain tumor in 1 match triplet unit. An exact match was used. This consists of finding a match with the same age, sex, insurance, and number of comorbidities (0, 1, 2, 3+). Since many people with stroke or TBI can be matched to the same person, a match was selected randomly from a set of exact matches. The matched groups had exactly the same distributions on all characteristics. Rehabilitation use outcomes were compared in the matched data to evaluate whether the results were affected by the difference in characteristics.

All tests were 2-sided, and the significance level was set to a multiple comparison Bonferroni adjusted P-value of .003 (0.05/17 models). The 17 models include 4 rehabilitation-use category comparisons (discharge to SNF, discharge to IRF, postdischarge rehabilitation only, and no rehabilitation within 12 months), the 12 PT/OT/SLP/overall comparisons (4 at 3 months, 4 at 6 months, and 4 at 12 months), and 1 regression model for trend. Data preprocessing and analyses were performed in SAS 9.4 (SAS Institute, Inc).

Results

Study Cohort

The study cohort was composed of 3381 individuals with brain tumors, 205 366 stroke, and 24 825 TBI patients (Figure 1, Table 2). Those with brain tumors were younger than stroke and TBI (median 47 vs 52 and 67 years, respectively, P < .0001); the majority were on commercial insurance (81% compared to 30% in stroke and 48% in TBI, P < .0001); and all of them had at least 1 comorbidity (100% vs 90% for stroke and 65% for TBI). Females made up 47% of the tumor group compared to 54% stroke and 41% TBI.

Table 2.

Characteristics of the Study Cohort

		Brain Tumor	Stroke	TBI	P-value	2-by-2 Comparisons P-Values
		n = 3381	n = 205366	n = 24825		Stroke vs Brain Tumor	TBI vs Brain Tumor	TBI vs Stroke
Age	Median (IQR)	47 [35, 56]	67 [57, 78]	52 [34, 67]	<.0001	<.0001	<.0001	<.0001
Sex	Male, n (%)	1799 (53%)	94 396 (46%)	14 571 (59%)	<.0001	<.0001	<.0001	<.0001
	Female, n (%)	1582 (47%)	11 0970 (54%)	10 254 (41%)
Insurance	Commercial, n (%)	2726 (81%)	61 056 (30%)	11 993 (48%)	<.0001	<.0001	<.0001	<.0001
	Medicaid, n (%)	427 (13%)	53 835 (26%)	7266 (29%)	<.0001	<.0001	<.0001	<.0001
	Medicare, n (%)	228 (7%)	90 475 (44%)	5566 (22%)	<.0001	<.0001	<.0001	<.0001
Elixhauser Index	0, n (%)	0 (0%)	20 317 (10%)	8595 (35%)	<.0001	<.0001	<.0001	<.0001
	1, n (%)	992 (29%)	40 948 (20%)	6879 (28%)	<.0001	<.0001	0.0474	<.0001
	2, n (%)	1275 (38%)	45 837 (22%)	4529 (18%)	<.0001	<.0001	<.0001	<.0001
	3+, n (%)	1114 (33%)	98 264 (48%)	4822 (19%)	<.0001	<.0001	<.0001	<.0001

Rehabilitation Use Immediately After Discharge and Postdischarge

From the index hospitalization, 1% of brain tumor patients were discharged to SNF, 3% to IRF, and 21% had only postdischarge outpatient rehabilitation (Table 3). These percentages were significantly lower compared to stroke and TBI (P < .0001). Discharge to SNF and IRF were individually 8% for stroke and 7% for TBI. Postdischarge outpatient rehabilitation rates were not significant when compared between groups. The brain tumor group had the highest rate of no rehabilitation within 12 months after discharge compared to stroke and TBI (75%, 65%, and 67%, respectively; P < .0001). Within 12 months following the hospitalization of brain tumor, 19% had at least 1 PT service, 8% had OT, and 6% had SLP. For stroke the percentages were 22% PT, 10% OT, and 8% SLP. Similarly, in TBI, there was higher use of PT compared to OT and SLP (22%, 8%, and 6%, respectively). See Figure 2. These observed differences in outcomes were independent of individual characteristics. In cohorts matched on age, sex, insurance, and comorbidities (Supplementary Table 1), similar significances were observed (Supplementary Table 2).

Table 3.

Rehabilitation Use Comparisons Between TBI, Stroke, and Brain Tumor Within 12 Months of Diagnosis

Outcomes		Brain Tumor	Stroke	TBI	P-value	2-by-2 comparisons p-values
		n = 3381	n = 205366	n = 24825		Stroke vs Brain Tumor	TBI vs Brain Tumor	TBI vs Stroke
Overall	Rehabilitation use
	Discharged to SNF	45 (1%)	16 318 (8%)	1782 (7%)	<.0001	<.0001	<.0001	<.0001
	Discharged to IRF	109 (3%)	17 410 (8%)	1770 (7%)	<.0001	<.0001	<.0001	<.0001
	Postdischarge rehabilitation only	694 (21%)	38 073 (19%)	4551 (18%)	.0085	.0032	.0021	.4282
	No rehabilitation within 12 months	2533 (75%)	133 565 (65%)	16 722 (67%)	<.0001	<.0001	<.0001	<.0001
Postdischarge Rehab	3 months post index
	Physical Therapy (PT)	332 (10%)	25 833 (13%)	3550 (14%)	<.0001	<.0001	<.0001	<.0001
	Occupational Therapy (OT)	156 (5%)	13 763 (7%)	1530 (6%)	<.0001	<.0001	.0004	.0013
	Speech-Language Pathology (SLP)	142 (4%)	11 322 (6%)	1062 (4%)	<.0001	.0009	.8333	<.0001
	At least 1 therapy type	444 (13%)	31 256 (15%)	3986 (16%)	<.0001	.0008	<.0001	.0005
	6 months post index
	Physical Therapy (PT)	459 (14%)	34 255 (17%)	4514 (18%)	<.0001	<.0001	<.0001	<.0001
	Occupational Therapy (OT)	199 (6%)	17 243 (8%)	1819 (7%)	<.0001	<.0001	.0023	<.0001
	Speech-Language Pathology (SLP)	181 (5%)	13 925 (7%)	1266 (5%)	<.0001	.0011	.5304	<.0001
	At least 1 therapy type	583 (17%)	40 406 (20%)	4993 (20%)	<.0001	.0004	<.0001	.1016
	12 months post index
	Physical Therapy (PT)	633 (19%)	44 596 (22%)	5394 (22%)	.0002	<.0001	<.0001	.9634
	Occupational Therapy (OT)	254 (8%)	20 821 (10%)	2095 (8%)	<.0001	<.0001	.0675	<.0001
	Speech-Language Pathology (SLP)	219 (6%)	16 054 (8%)	1423 (6%)	<.0001	.0040	.0827	<.0001
	At least 1 therapy type	762 (23%)	51 261 (25%)	5907 (24%)	<.0001	.0012	.1067	<.0001

Figure 2.

Overall 12-month Post-discharge Rehabilitation Use.

Rehabilitation Use Within 3, 6, and 12 Months After Discharge

Three months after discharge from a brain tumor hospitalization, the overall outpatient rehabilitation use was 13% which increased to 17% at 6 months and 23% at 12 months timepoint (Table 3). The highest type of rehabilitation use was PT, which was 9% at 3 months, 13% at 6 months, and 19% at 12 months. OT and SLP were respectively 5% and 4% at 3 months and increased 1–2% at 6 and 12 months. For stroke and TBI, the rate of therapy was higher at each time point and for each therapy type. For these conditions also, there is a higher use of PT (13% and 14% respectively at 3 months increasing to 22% at 12 months) compared to OT and SLP (7% at 3 months increasing to 6–10% at 12 months).

Rehabilitation Use Over the Years

Annually, the use of all therapies has increased from 2001 to 2018 (P < .0001, Table 4). The use of PT has remained consistently higher than OT and SLP (Figure 3A–C). However, the trends (change rates) of brain tumor and stroke were significantly higher than the trend of TBI (Table 4) with stroke having a higher trend than brain tumor in earlier years while brain tumor trend was higher in later years (Figure 3D). Stroke had the highest trends of all 3 therapies (Table 4). Comparing those diagnosed in 2018 to those diagnosed in 2001, rehabilitation use has increased for all conditions (Table 5). Brain tumor had the highest increases in OT (7.95 times) and PT (3.89 times) compared to stroke and TBI. Stroke had the highest increase in SLP (0.98 times).

Table 4.

Annual Rate of Change and Standard Error of Rehabilitation Use

		Brain Tumor [1]	Stroke [2]	TBI [3]	P-values
					[2] vs [1]	[3] vs [1]	[3] vs [2]
12-Month rehabilitation	PT (1)	1.1% ± 0.06%*	1.2% ± 0.06%*	0.8% ± 0.06%*	.1479	<.0001	<.0001
	OT (2)	0.69% ± 0.06%*	0.79% ± 0.06%*	0.39% ± 0.06%*	.1479	<.0001	<.0001
	SLP (3)	0.49% ± 0.06%*	0.59% ± 0.06%*	0.18% ± 0.06%*	.1479	<.0001	<.0001
P-values	(2) vs (1)	<.0001	<.0001	<.0001
	(3) vs (1)	<.0001	<.0001	<.0001
	(3) vs (2)	.0027	.0027	.0027

Figure 3.

Trend of Rehabilitation Use.

Table 5:

12-Month Rehabilitation Use Change for People Diagnosed in 2018 Compared to 2001

Year	Brain Tumor			Stroke			TBI
	PT	OT	SLP	PT	OT	SLP	PT	OT	SLP
2001	6.56%	1.64%	4.92%	10.62%	3.76%	3.82%	12.54%	4.53%	3.48%
2018	32.11%	14.68%	6.42%	30.02%	15.93%	11.4%	28.94%	11.45%	5.51%
% Difference	389%	795%	30%	183%	324%	198%	131%	153%	58%

Discussion

The results of this study, using MarketScan Research Database data over a 2-decade timespan, provide insight into national utilization of rehabilitation services for patients with brain tumors in comparison to those with brain pathologies stroke and TBI which have similar rehabilitation needs due to intracranial insults. The data shows a lower utilization of rehabilitation services for patients with brain tumors in SNF and IRF when compared to those with stroke or TBI, though near equal for outpatient services amongst the 3 patient populations. The data also shows a lower utilization of rehabilitation services 3, 6, and 12 months after hospital discharge for brain tumor patients, with these patients receiving significantly more PT than OT and SLP. The data demonstrates rehabilitation utilization has steadily increased since 2001 for all patient populations studied.

Potential benefits of multidisciplinary comprehensive inpatient and outpatient rehabilitation services include a reduction in disability along with improvements in functional outcomes and quality of life.⁹ Our study shows a combined 25% of brain tumor patients receiving rehabilitation services in inpatient and outpatient settings during a 12-month timeframe.²⁹ Our study results are comparable to the utilization found in a brain tumor population in Italy where 12.8% of patients received inpatient rehabilitation and 11.8% received outpatient rehabilitation within 12 months of diagnosis.¹⁴ We have also found that 4% of patients with brain tumors received inpatient rehabilitation services and stroke and TBI patients received 16% and 14% respectively. There are several possible explanations for this difference in rehabilitation use among brain tumors, stroke, and TBI. Brain tumor patients may require urgent outpatient oncologic care that cannot be provided in an inpatient rehabilitation facility and oncologic care is a priority in those scenarios. Patients who require surgical intervention for brain tumors may have complete or near-complete recovery of deficits postoperatively due to the successful removal of tumor and/or reduction in intracranial swelling.³⁰ If functional impairments are present postoperatively, the severity may not require immediate inpatient or outpatient rehabilitation. The functional impact of a brain tumor condition can evolve and potentially progress over time rather than immediately upon diagnosis and hospitalization. While we hope this not the case, some functional impairments may go unrecognized and/or underreported and thus outpatient rehabilitation referrals are not initiated. If inpatient rehabilitation is required for patients with PBTs, studies have shown that, with comparable lengths of inpatient rehabilitation stays, patients with brain tumors have higher or comparable functional independence measure score improvements compared to stroke and brain injuries.^4,12,31,32

In this study, we found that outpatient rehabilitation interventions were utilized for primary malignant brain tumors (23% over 12 months after diagnosis), which is lower than that of both stroke and TBI patients. SLP utilization was lower for brain tumor compared to stroke but either higher or comparable to TBI. OT (8%) and SLP (6%) services were utilized less than PT (19%) in brain tumor at 12 months. The difference of use found in our study may be due to higher rates of impairments with gait and general transfers requiring PT. However, these patients may have concurrent difficulty with ADL concerns, including toilet transfers and hygiene which would require OT services. Also, it has been reported that up to 80% of brain tumor patients experience some cognitive impairment requiring SLP.⁷ Therefore, our findings suggest that, even in cases where rehabilitation is used, OT and SLP may be underutilized compared to PT. Multidisciplinary rehabilitation efforts are critical for functional improvement optimization or even education for patient/caregiver adjustments to impairments acquired from the tumor location and/or the treatment effects. These findings are supportive of the development and implementation of programs geared to increase the use of these therapies in brain tumor patients.

Despite the low utilization of rehabilitation services for patients with primary malignant brain tumors, this study shows an increasing utilization over the years, and this is a similar pattern for patients with stroke or TBI. Ideally, this improved rehabilitation utilization over the years is due to improved healthcare provider awareness of rehabilitation needs in the brain tumor population and improved access to rehabilitation care. However, we are unable to delineate the reasons for an upward trend through our research. Additional research is needed to explore the driver of this positive change. Furthermore, research and advocacy are needed to further improve rehabilitation utilization since brain tumor patients remain significantly underserved with 75% of brain tumor patients not receiving any rehabilitation interventions within 12 months of diagnosis.

Strength and Limitations

MarketScan is a large database, with data from all over the United States. As a claims data source, it represents real-world practice of rehabilitation uses and it is longitudinal. Moreover, having commercial, Medicare, and Medicaid portions, this data spans distinct stages of life. All these are strengths of this study. However, there are some noteworthy limitations. First, MarketScan data is a convenient sample of insured individuals. It does not include the uninsured. Second, the use of ICD-9/10 diagnosis codes limits our ability to get a granular classification of tumors³³-for example, ICD-9 does not distinguish metastasis to the brain and spinal cord- and neither ICD-9 nor ICD-10 provides information on tumor grade or histology. This is important because different brain tumors have varied rehabilitation needs. Therefore, for this study, we grouped all the malignant brain tumors together and did not include metastasis to the brain. Third, brain tumor patients may not have rehabilitation needs immediately after their initial diagnosis but may have higher rehabilitation needs later after the 12-month timeframe of our study due to the later progression of the disease. Future studies should evaluate long-term rehabilitation use. Fourth, this data does not allow us to identify the reason why an individual did not receive rehabilitation. Despite these limitations, this project contributes to the conversation of rehabilitation use after brain tumors in comparison to TBI and stroke.

Conclusions

Rehabilitation utilization for patients with MPBT in the inpatient and outpatient setting over a 12-month timeframe after diagnosis is significantly lower than that of patients with stroke or TBI. However, in recent years the trend for MPBT patients receiving rehabilitation has increased faster than TBI and stroke. In those with MPBT who do receive rehabilitation services, PT is utilized at a higher rate than both OT and SLP. Fortunately, over the years, rehabilitation utilization for all 3 disciplines (PT, OT, and SLP) has increased; however, there remains a significant utilization gap which demonstrates the need to improve. Future research is indicated to further study the neurologic deficits at diagnosis as related to location and type of brain tumors, as well as deficits occurring and remaining within and after the 12-month timeframe demonstrated in this study.

Supplementary material

Supplementary material is available online at Neuro-Oncology Practice (https://academic.oup.com/nop).

Funding

This work was supported by University of Louisville; internal funding.

Conflict of interest statement

None declared.

Authorship statement

S.G. and M.N. conceived the idea. B.U. performed the statistical analyses. S.G., B.U., E.A., and M.N. wrote different sections of the paper. D.K. and M.B. reviewed the draft and provided critical feedback. All the authors helped shape the research and contributed to the final manuscript.

Data availability

No new data were generated in support of this research.