Disparities in the treatment of movement disorders using deep brain stimulation

In Brief

In this national database study, researchers analyzed disparities in the use of deep brain stimulation (DBS) for the treatment of Parkinson's disease and essential tremor. A multivariate regression analysis revealed significantly lower rates of DBS utilization for Black patients, female patients, patients in lower income brackets, and those without private insurance. This study identifies areas for targeted intervention to reduce disparities and establish more equitable access to neurosurgical treatment for patients with movement disorders.

Abstract

OBJECTIVE

Deep brain stimulation (DBS) is a well-established treatment for Parkinson’s disease (PD) and essential tremor (ET). Although the prevalence of PD and ET can vary by sex and race, little is known about the accessibility of neurosurgical treatments for these conditions. In this nationwide study, the authors aimed to characterize trends in the use of DBS for the treatment of PD and ET and to identify disparities in the neurosurgical treatment of these diseases based on ethnic, racial, sex, insurance, income, hospital, and geographic factors.

METHODS

Using the dates January 1, 2012, to December 31, 2019, the authors queried the National Inpatient Sample database for all discharges with an ICD-9 or ICD-10 diagnosis of PD or ET. Among these discharges, the DBS rates were reported for each subgroup of race, ethnicity, and sex. To develop national estimates, all analyses were weighted.

RESULTS

Among 2,517,639 discharges with PD, 29,820 (1.2%) received DBS, and among 652,935 discharges with ET, 11,885 (1.8%) received DBS. Amid the PD cases, Black patients (n = 405 [0.2%], OR 0.16, 95% CI 0.12–0.20) were less likely than White patients (n = 23,975 [1.2%]) to receive DBS treatment, as were Hispanic patients (n = 1965 [1.1%], OR 0.76, 95% CI 0.65–0.88), whereas Asian/Pacific Islander patients (n = 855 [1.5%]) did not statistically differ from White patients. Amid the ET cases, Black (n = 230 [0.8%], OR 0.39, 95% CI 0.27–0.56), Hispanic (n = 215 [1.0%], OR 0.39, 95% CI 0.28–0.55), and Asian/Pacific Islander (n = 55 [1.0%], OR 0.51, 95% CI 0.28–0.93) patients were less likely than White patients (n = 10,440 [1.9%]) to receive DBS. Females were less likely than males to receive DBS for PD (OR 0.69, p < 0.0001) or ET (OR 0.70, p < 0.0001).

CONCLUSIONS

The authors describe significant racial and sex-based differences in the utilization of DBS for the treatment of PD and ET. Further research is required to ascertain the causes of these disparities, as well as any differences in access to specialty neurosurgical care and referral for neuromodulation approaches.

Deep brain stimulation (DBS) is a well-established treatment for the movement disorders Parkinson’s disease (PD) and essential tremor (ET). While DBS remains an elective procedure, the majority of patients with PD develop dyskinesias and motor fluctuations within 10 years of initiating medical therapy and can require surgery.^1,2 DBS improves limb tremor, nonmotor symptoms, and overall quality of life in patients with PD.³ After a trial of relatively young patients with a shorter disease duration, the surgical inclusion criteria for DBS in PD have been expanded to include early motor fluctuations.⁴ DBS is also commonly used for medically resistant ET to treat unilateral and bilateral limb and axial (head, neck, or trunk) tremors with long-term improvements in quality of life.⁵

Prior studies have demonstrated the unequal utilization of DBS for movement disorders in the United States (US).^6–8 White patients are 5 times more likely than African American patients to undergo DBS surgery, and privately insured patients are 1.6 times more likely than Medicaid patients to undergo the procedure.⁹ Despite annual increases in the number of DBS cases, racial disparities in usage rates have persisted since at least 2010.⁹ While analyses of racial, sex, and socioeconomic disparities in DBS for ET remain limited, studies have suggested that women have substantially more severe disease when referred for DBS and have lower usage rates than men despite equal disease prevalence.^10,11

In this nationwide study utilizing the National Inpatient Sample (NIS) database, we aimed to characterize trends in the use of DBS for the treatment of PD and ET and to identify disparities in the neurosurgical treatment of these diseases based on ethnic, racial, sex, insurance, income, hospital, and geographic factors. By identifying neuromodulation treatment disparities, we hope to develop initiatives for the equitable treatment of these conditions.

Methods

Data Source

We utilized the NIS database from the Healthcare Cost and Utilization Project (HCUP) of the US Agency for Healthcare Research and Quality (AHRQ). Information from the database was queried using the dates January 1, 2012, to December 31, 2019. Patient diagnoses and procedures were determined using codes from the ICD-9 and ICD-10. The specific codes used in this study are listed in Table 1.

TABLE 1.

Codes for movement disorders and DBS

Criteria	ICD-9 (2015)	ICD-10
PD	332, 332.0: PD	G20: PD
ET	333.1: essential & other unspecified forms of tremor	G25.0: ET
DBS	02.93: implantation or replacement of intracranial neurostimulator lead(s); 01.20: cranial implantation or replacement of neurostimulator pulse generator; 86.94: insertion or replacement of single array neurostimulator pulse generator, not specified as rechargeable; 86.95: insertion or replacement of multiple array neurostimulator pulse generator, not specified as rechargeable; 86.96: insertion or replacement of other neurostimulator pulse generator; 86.97: insertion or replacement of single-array rechargeable neurostimulator pulse generator; 86.98: insertion or replacement of multiple-array (2 or more) rechargeable neurostimulator pulse generator	00H00MZ: insertion of neurostimulator lead into brain, open approach; 00H03MZ: insertion of neurostimulator lead into brain, percutaneous approach; 00H04MZ: insertion of neurostimulator lead into brain, percutaneous endoscopic approach; 00H60MZ: insertion of neurostimulator lead into cerebral ventricle, open approach; 00H63MZ: insertion of neurostimulator lead into cerebral ventricle, percutaneous approach; 00H64MZ: insertion of neurostimulator lead into cerebral ventricle, percutaneous endoscopic approach; 0JH60DZ: insertion of multiple-array stimulator generator into chest subcutaneous tissue and fascia, open approach; 0JH63DZ: insertion of multiple-array stimulator generator into chest subcutaneous tissue & fascia, percutaneous approach; 0JH70DZ: insertion of multiple-array stimulator generator into back subcutaneous tissue & fascia, open approach; 0JH73DZ: insertion of multiple-array stimulator generator into back subcutaneous tissue & fascia, percutaneous approach; 0JH80DZ: insertion of multiple-array stimulator generator into abdomen subcutaneous tissue & fascia, open approach; 0JH83DZ: insertion of multiple-array stimulator generator into abdomen subcutaneous tissue & fascia, percutaneous approach; Additional codes: 0JH60EZ, 0JH60BZ, 0JH80MZ, 0JH83MZ; Additional code modifiers: MZ = stimulator generator; BZ = single-array stimulator generator; CZ = single-array rechargeable stimulator generator; EZ = multiple-array rechargeable stimulator generator

Inclusion and Exclusion Criteria

The initial ET and PD cohorts included patients who had been discharged with a primary diagnosis of either disease, identified using the ICD codes listed in Table 1. We identified patients who had undergone DBS procedures by using the ICD codes outlined in Table 1. Patients with an unknown sex or income quartile status (< 0.1% of all patients) were excluded from all analyses.

Outcomes and Explanatory Variables

The primary outcome was DBS implantation during the hospital stay. If the principal diagnosis of the hospital admission of a patient undergoing DBS surgery was coded as PD or ET, we defined those patients as undergoing DBS for PD or ET, acknowledging that DBS can be used to treat other conditions, including dystonia¹² and epilepsy.¹³ While dystonia is a movement disorder, we did not analyze it as part of this study because of the relatively low frequency of DBS for that indication versus PD and ET; our initial queries of DBS for dystonia returned a few dozen cases per year compared to thousands for PD and ET. Patient admission information was collected, including age at admission and year of hospital admission. We calculated the AHRQ Elixhauser Comorbidity Index (ECI)¹⁴ using the diagnosis data for the discharges. To calculate the ECI, we used the AHRQ comorbidity measure variables from the Severity File from 2012 to the first three quarters of 2015. For the fourth quarter of 2015 and after, we applied the HCUP Elixhauser Comorbidity Software Refined¹⁵ for ICD-10-CM on the ICD-10 diagnosis variables. The primary characteristics collected for our disparities analysis were patient race (White, Black, Asian/Pacific Islander, Hispanic, or other/unknown), patient sex, primary payer/insurance (Medicare, Medicaid, private, other/unknown), geographic region (Northeast, Midwest, South, West), median household income quartile of the patient zip code, and hospital teaching status.

Statistical Analysis

For all admissions that met study inclusion criteria, patient and hospital characteristics were reported with appropriate summary statistics for continuous and categorical variables. Weighted national estimates were calculated for all variables accounting for the complex sample design of the NIS. The weighted frequency of disease diagnosis among all NIS discharges and the weighted frequency of DBS utilization for each disease cohort were summarized for each year from 2012 to 2019. For both disease cohorts, the weighted frequencies of patient sex, race, insurance, hospital region, income quartile, and hospital teaching status were summarized according to the use of DBS, and the trend in usage over time for each of these characteristics was plotted.

A weighted multivariable logistic regression model was created for each disease cohort to determine the odds of undergoing DBS treatment, accounting for age, sex, race, insurance, hospital region, income, hospital teaching status, and comorbidity burden (as measured by the ECI). In the regression models, the insurance statuses of self-pay, other, unknown, and no charge were combined as one category (other), and the hospital location statuses of rural and urban nonteaching institutions were combined as one category. These groupings were done because of the very small samples sizes compared to those of the other categories. The statistical significance level was set at p < 0.05. All statistical analyses were performed using SAS 9.4 (SAS Institute Inc.).

Results

Patient Cohort and Demographics

We identified a total of 2,517,639 inpatient discharges with a diagnosis of PD and 652,935 inpatient discharges with a diagnosis of ET during the study period. On average, patients in the PD cohort were 76.3 years old at admission, whereas patients in the ET cohort were 72.5 years old. The PD and ET cohorts were 42.6% (n = 1,072,060) and 54.9% (n = 358,265) female, respectively. The most common race in both cohorts was White (PD: 77.8%, ET: 84.9%). Black patients made up 6.6% of the PD cohort and 4.3% of the ET cohort; Hispanic patients made up 7.1% and 3.2%, respectively; and Asian/Pacific Islander patients made up 2.3% and 0.8%, respectively. The majority of patients used Medicare as the primary payer in both cohorts (PD: 86.9%, ER: 77.9%), whereas 3.3% of the PD cohort and 4.6% of the ET cohort used Medicaid. Private insurance was used by 7.6% of the PD cohort and 14.3% of the ET cohort. The average ECI was 12.3 in the PD cohort and 7.7 in the ET cohort. Among the PD cohort, 29,820 patients (1.2%) underwent DBS surgery; among the ET cohort, 11,885 patients (1.8%) underwent DBS procedures. Full details on the characteristics of each cohort are listed in Table 2 (PD) and Table 3 (ET).

TABLE 2.

Characteristics and DBS use of patients with PD

Category	No DBS	Received DBS	Total
No. of patients	2,487,819	29,820	2,517,639
Age at admission in yrs
Mean (SE)	76.4 (0.02)	65.0 (0.12)	76.3 (0.02)
Median (IQR)	77.2 (70.1, 83.5)	65.6 (58.9, 71.0)	77.1 (69.9, 83.4)
Sex
Female	1,062,575 (99.12%)	9,485 (0.88%)	1,072,060 (42.58%)
Male	1,425,245 (98.59%)	20,335 (1.41%)	1,445,580 (57.42%)
Race
White	1,933,935 (98.78%)	23,975 (1.22%)	1,957,910 (77.77%)
Black	166,735 (99.76%)	405 (0.24%)	167,140 (6.64%)
Asian/Pacific Islander	57,380 (98.53%)	855 (1.47%)	58,235 (2.31%)
Hispanic	175,690 (98.89%)	1,965 (1.11%)	177,655 (7.06%)
Other/unknown	154,080 (98.33%)	2,620 (1.67%)	156,700 (6.22%)
Primary insurance
Medicaid	81,160 (99.18%)	675 (0.82%)	81,835 (3.25%)
Medicare	2,167,099 (99.12%)	19,350 (0.88%)	2,186,449 (86.85%)
Private insurance	183,580 (95.39%)	8,870 (4.61%)	192,450 (7.64%)
Other	55,980 (98.37%)	925 (1.63%)	56,905 (2.26%)
Median household income of patient zip code
1st quartile	646,285 (99.21%)	5,125 (0.79%)	651,410 (25.87%)
2nd quartile	647,775 (98.91%)	7,165 (1.09%)	654,940 (26.01%)
3rd quartile	614,990 (98.69%)	8,155 (1.31%)	623,145 (24.75%)
4th quartile	578,770 (98.41%)	9,375 (1.59%)	588,145 (23.36%)
Hospital region
Midwest	603,335 (98.95%)	6,385 (1.05%)	609,720 (24.22%)
Northeast	519,920 (98.98%)	5,345 (1.02%)	525,265 (20.86%)
South	918,350 (98.92%)	10,070 (1.08%)	928,420 (36.88%)
West	446,215 (98.23%)	8,020 (1.77%)	454,235 (18.04%)
ECI
Mean (SE)	12.4 (0.02)	4.1 (0.05)	12.3 (0.02)
Median (IQR)	10.8 (3.9, 18.2)	4.1 (2.0, 4.7)	10.7 (3.8, 18.0)
Min, max	−23.0, 75.0	−14.0, 38.0	−23.0, 75.0
Hospital location/teaching status
Urban nonteaching/rural	1,035,274 (99.85%)	1,505 (0.15%)	1,036,779 (41.18%)
Urban teaching	1,452,545 (98.09%)	28,315 (1.91%)	1,480,860 (58.82%)

TABLE 3.

Characteristics and DBS utilization of patients with ET

Category	No DBS	Received DBS	Total
No. of patients	641,050	11,885	652,935
Age at admission in yrs
Mean (SE)	72.5 (0.05)	67.6 (0.22)	72.5 (0.05)
Median (IQR)	74.0 (64.9, 81.9)	68.5 (62.5, 74.0)	73.8 (64.8, 81.8)
Sex
Female	352,845 (98.49%)	5,420 (1.51%)	358,265 (54.87%)
Male	288,205 (97.81%)	6,465 (2.19%)	294,670 (45.13%)
Race
White	544,050 (98.12%)	10,440 (1.88%)	554,490 (84.92%)
Black	28,140 (99.19%)	230 (0.81%)	28,370 (4.34%)
Asian/Pacific Islander	5,405 (98.99%)	55 (1.01%)	5,460 (0.84%)
Hispanic	20,965 (98.98%)	215 (1.02%)	21,180 (3.24%)
Other/unknown	42,490 (97.82%)	945 (2.18%)	43,435 (6.65%)
Primary insurance
Medicaid	29,565 (99.06%)	280 (0.94%)	29,845 (4.57%)
Medicare	500,230 (98.35%)	8,415 (1.65%)	508,645 (77.90%)
Private insurance	90,780 (97.21%)	2,605 (2.79%)	93,385 (14.30%)
Other	20,475 (97.22%)	585 (2.78%)	21,060 (3.23%)
Median household income of patient zip code
1st quartile	144,770 (98.24%)	2,595 (1.76%)	147,365 (22.57%)
2nd quartile	175,575 (98.04%)	3,510 (1.96%)	179,085 (27.43%)
3rd quartile	170,675 (98.15%)	3,215 (1.85%)	173,890 (26.63%)
4th quartile	150,030 (98.32%)	2,565 (1.68%)	152,595 (23.37%)
Hospital region
Midwest	187,880 (98.38%)	3,100 (1.62%)	190,980 (29.25%)
Northeast	104,905 (98.71%)	1,375 (1.29%)	106,280 (16.28%)
South	215,760 (98.01%)	4,380 (1.99%)	220,140 (33.72%)
West	132,505 (97.76%)	3,030 (2.24%)	135,535 (20.76%)
ECI
Mean (SE)	7.9 (0.03)	−0.1 (0.11)	7.7 (0.03)
Median (IQR)	5.9 (−1.0, 14.1)	−1.0 (−2.9, 1.3)	5.7 (−1.1, 14.0)
Min, max	−25.0, 70.0	−18.0, 41.0	−25.0, 70.0
Hospital location/teaching status
Urban nonteaching/rural	250,250 (99.73%)	690 (0.27%)	250,940 (38.43%)
Urban teaching	390,800 (97.22%)	11,195 (2.78%)	401,995 (61.57%)

Trends in PD, ET, and DBS Utilization Over Time

DBS was used in 1.04% of all PD admissions in 2012, and this rate increased to 1.43% by 2014 and then decreased to 1.07% in 2017. Subsequently, DBS use remained steady at approximately 1.1% until 2019. In 2012, 1.39% of all patients admitted with ET underwent DBS surgery, and this rate increased steadily over the years to 1.99% in 2016. The frequency of DBS in ET patients slightly decreased to 1.88% in 2017 before increasing again to 2.24% in 2019.

Over the same period, the percentage of all discharges in the NIS with a diagnosis of PD increased from 0.77% to 1.03%, whereas the percentage of all discharges with a diagnosis of ET remained fairly steady between 0.22% and 0.25%. These trends are shown in Table 4.

TABLE 4.

Yearly frequency of DBS utilization among PD and ET patients

Year	Frequency of PD	% PD Among All NIS Discharges	Frequency of DBS for PD	Frequency of ET	% ET Among All NIS Discharges	Frequency of DBS for ET
2012	279,975	0.7674%	2,900 (1.0358%)	80,760	0.2214%	1,125 (1.3930%)
2013	270,585	0.7601%	3,750 (1.3859%)	82,225	0.2310%	1,225 (1.4898%)
2014	271,815	0.7687%	3,890 (1.4311%)	83,665	0.2366%	1,395 (1.6674%)
2015	292,815	0.8186%	3,930 (1.3421%)	80,780	0.2258%	1,510 (1.8693%)
2016	330,760	0.9271%	3,615 (1.0929%)	73,515	0.2061%	1,465 (1.9928%)
2017	347,795	0.9715%	3,720 (1.0696%)	79,125	0.2210%	1,485 (1.8768%)
2018	360,320	1.0142%	4,025 (1.1171%)	83,760	0.2358%	1,685 (2.0117%)
2019	363,575	1.0265%	3,990 (1.0974%)	89,105	0.2516%	1,995 (2.2389%)
Total	2,517,640	0.8814%	29,820 (1.1844%)	652,935	0.2286%	11,885 (1.8202%)

Prevalence of DBS Utilization by Sex, Race, and Comorbidities

Tables 2 and 3 and Figs. 1A, 1B, 2A, and 2B summarize DBS utilization by sex and race. Among the patients with PD, 1.41% of males and 0.88% of females underwent DBS. Amid those with ET, 2.19% of males underwent DBS compared to 1.51% of females. Each year from 2012 to 2019, males of both the PD and ET cohorts had higher rates of DBS surgery than the females. In the multivariable regression models, female patients had lower odds than males of undergoing DBS surgery for PD (OR 0.69, 95% CI 0.65–0.73, p < 0.0001) and ET (OR 0.70, 95% CI 0.64–0.76, p < 0.0001).

FIG. 1.

Yearly prevalence of DBS use for PD by sex (A), race (B), insurance status (C), income quartile (D), and hospital geographic region (E). Figure is available in color online only.

FIG. 2.

Yearly prevalence of DBS use for ET by sex (A), race (B), insurance status (C), income quartile (D), and hospital geographic region (E). Figure is available in color online only.

For PD, Asian/Pacific Islander patients (1.47%) were the most likely to undergo DBS, followed by 1.22% of White patients, 1.11% of Hispanic patients, and 0.24% of Black patients. For ET, 1.88% of White patients underwent DBS, followed by 1.02% of Hispanic patients, 1.01% of Asian/Pacific Islander patients, and 0.81% of Black patients. The rate of DBS for PD was higher among all other races across all years than among Black patients. For almost all years, the rate of DBS for ET was higher for White patients than for all other races.

Compared to White patients, Black patients had lower odds of undergoing DBS for PD (OR 0.16, 95% CI 0.12–0.20, p < 0.0001), as did Hispanic patients (OR 0.76, 95% CI 0.65–0.88, p = 0.0003); however, there was no significant difference in odds between White and Asian/Pacific Islander patients (OR 1.08, 95% CI 0.89–1.30, p = 0.4362). For ET, Black patients (OR 0.39, 95% CI 0.27–0.56, p < 0.0001), Hispanic patients (OR 0.39, 95% CI 0.28–0.55, p < 0.0001), and Asian/Pacific Islander patients (OR 0.51, 95% CI 0.28–0.93, p = 0.0272) were all less likely to undergo DBS than the White patients. Moreover, an increase in age was associated with significantly lower odds of undergoing DBS for PD and ET.

The mean ECI in the PD group was 12.4 among those who did not receive DBS and 4.1 among the patients who did. The mean ECIs in the ET group were 7.9 and −0.1 for the patients who did not and did undergo neuromodulation surgery, respectively. The multivariable regression model showed that with all else held constant, each 1-unit increase in the ECI was associated with 8% and 9% lower odds of undergoing neuromodulation surgery for PD and ET patients, respectively. These results are shown in Table 5.

TABLE 5.

Adjusted weighted odds ratio of DBS use among PD and ET patients

Covariate	PD		ET
	OR (95% CI)	p Value	OR (95% CI)	p Value
Age (1-yr increase)	0.939 (0.937–0.941)	<0.0001	0.988 (0.985–0.991)	<0.0001
Sex
Male	Reference	Reference	Reference	Reference
Female	0.690 (0.652–0.731)	<0.0001	0.695 (0.639–0.756)	<0.0001
Race
White	Reference	Reference	Reference	Reference
Asian/Pacific Islander	1.077 (0.894–1.298)	0.4362	0.506 (0.276–0.926)	0.0272
Black	0.155 (0.118–0.202)	<0.0001	0.387 (0.267–0.560)	<0.0001
Hispanic	0.759 (0.654–0.880)	0.0003	0.393 (0.282–0.549)	<0.0001
Primary insurance
Private insurance	Reference	Reference	Reference	Reference
Medicaid	0.108 (0.088–0.133)	<0.0001	0.283 (0.213–0.375)	<0.0001
Medicare	0.496 (0.462–0.532)	<0.0001	1.056 (0.935–1.192)	0.3808
Hospital region
West	Reference	Reference	Reference	Reference
Midwest	0.541 (0.452–0.648)	<0.0001	0.616 (0.503–0.754)	<0.0001
Northeast	0.431 (0.354–0.525)	<0.0001	0.452 (0.351–0.583)	<0.0001
South	0.655 (0.551–0.779)	<0.0001	0.822 (0.669–1.009)	0.0604
Household income
1st quartile	Reference	Reference	Reference	Reference
2nd quartile	1.307 (1.189–1.438)	<0.0001	1.052 (0.934–1.184)	0.4028
3rd quartile	1.326 (1.197–1.470)	<0.0001	0.876 (0.770–0.997)	0.0450
4th quartile	1.607 (1.439–1.796)	<0.0001	0.751 (0.646–0.874)	0.0002
Hospital location/teaching status
Urban teaching	Reference	Reference	Reference	Reference
Urban nonteaching/rural	0.079 (0.064–0.098)	<0.0001	0.089 (0.070–0.114)	<0.0001
ECI (1-unit increase)	0.916 (0.914–0.919)	<0.0001	0.912 (0.908–0.915)	<0.0001

Boldface type indicates statistical significance.

Prevalence of DBS Utilization According to Insurance and Income

Details on DBS use according to insurance and income are shown in Figs. 1C, 1D, 2C, and 2D and Tables 2 and 3. Rates of DBS for PD were highest for patients with private insurance (4.61%), followed by Medicare (0.88%) and Medicaid (0.82%). For those with ET, 2.79% of patients with private insurance received DBS, followed by 1.65% of Medicare patients and 0.94% of Medicaid patients. The multivariable regression showed that the odds of PD patients receiving DBS were lower for those with Medicaid (OR 0.11, 95% CI 0.09–0.13, p < 0.0001) or Medicare (OR 0.50, 95% CI 0.46–0.53, p < 0.0001) than for those with private insurance. For ET patients, the odds of receiving DBS, compared to those with private insurance, were lower for those with Medicaid (OR 0.28, 95% CI 0.21–0.38, p < 0.001) but not those with Medicare (OR 1.06, 95% CI 0.94–1.19, p = 0.3808; Table 5). Among the patients with PD, those in the first income quartile had significantly lower odds of undergoing DBS than each of the other three income quartiles. For patients with ET, the odds of undergoing DBS were significantly lower for those in the first quartile versus those in the fourth and third quartiles but not those in the second quartile.

Prevalence of DBS Utilization According to Region and Hospital Type

Of those patients admitted with PD, 1.77% in the West underwent DBS surgery, followed by 1.08% in the South, 1.05% in the Midwest, and 1.02% in the Northeast. Among the patients admitted with ET, 2.24% in the West received DBS, followed by 1.99% in the South, 1.62% in the Midwest, and 1.29% in the Northeast. Patients in urban teaching hospitals were more likely to undergo DBS than those in rural or nonteaching hospitals (PD: 1.9% vs 0.15%, ET: 2.78% vs 0.27%). These results are shown in Tables 2 and 3 and Figs. 1E and 2E.

Compared to those in the US West, patients in the Midwest (OR 0.54, 95% CI 0.45–0.65, p < 0.0001), South (OR 0.66, 95% CI 0.55–0.78, p < 0.0001), and Northeast (OR 0.43, 95% CI 0.35–0.53, p < 0.0001) had significantly lower odds of undergoing DBS for PD. Among the patients with ET, those in the Midwest (OR 0.62, 95% CI 0.50–0.75, p < 0.0001) and Northeast (OR 0.45, 95% CI 0.35–0.58, p < 0.0001) had significantly lower odds of undergoing DBS than those in the West, but patients in the South did not (OR 0.82, 95% CI 0.67–1.01, p = 0.0604; Table 5).

Discussion

We analyzed a national database to ascertain patterns of DBS utilization for PD and ET. Our findings indicate that there are substantial disparities in who receives DBS therapy in the US. DBS is more prevalent in the Western region of the US and among White male patients with private insurance and higher income classifications. Among minority races, Black patients have the greatest disparities in DBS utilization for both PD and ET compared to White patients. This study highlights disparities in access to DBS therapy for PD and ET.

Sample Demographics of Patients With PD and ET

While the NIS does not represent the entire US population, the cohorts extracted for this study provide a representative sample of the inpatient population from all US hospitals. The mean age of the patient cohorts with either PD or ET was upwards of 72 years. This finding matches previous epidemiology studies both globally and in North America for PD^16,17 as well as for ET.¹⁸

With regard to sex, our study demonstrated higher rates of PD in men, whereas the opposite was true for ET. Prior epidemiological research has also demonstrated a higher prevalence and incidence of PD among men than in women.¹⁹ For ET, some studies have shown that both sexes are equally likely to be diagnosed,²⁰ whereas others have reported either men¹⁸ or women²¹ to have a higher incidence.

Our cohorts of PD and ET consisted primarily of White patients (upwards of 75% for both disorder groups), whereas the proportion of minority groups (Black, Hispanic, and Asian/Pacific Islander) was much lower (all below 10%). Prior reviews on the prevalence of PD among different racial groups have shown that the disease is more common in Hispanic and White patients than in Black or Asian patients^22,23 but not to the extent shown in our study. The overwhelming prevalence of White patients in our study cohort is likely explained by the fact that PD has been historically underdiagnosed in patients of color, potentially due to underinsurance, reduced access to specialty neurologists, and cultural differences between patients and healthcare providers.²² If fewer patients of color were being seen by neurologists and then officially diagnosed with this condition, they would then appear with lower frequency in the databases that we surveyed. By extension, without a proper diagnosis, these patients would not undergo DBS, widening the disparity.

While studies on racial differences in ET are rarer, it has been reported that White patients have the highest overall prevalence of ET.^20,21 However, it is also unlikely that more than 75% of all patients with ET are White. It is not unreasonable to assume that this demographic disparity in our cohort is explained by similar factors that caused the disparity in the PD cohort.

Rising DBS Utilization for PD and ET

The utilization of DBS for PD increased in absolute number between 2012 and 2019, and a similar trend was seen in ET. The advancements likely driving the increased utilization of DBS procedures include expanded indications for younger patients with the disease,⁴ increased accessibility to specialty centers, and technological advances such as smaller batteries, rechargeable systems, better MRI compatibility, different stimulation paradigms, and closed-loop stimulation.²⁴ DBS targeting has also improved through MRI-guided procedures, allowing for more precisely placed electrodes and better subsequent symptom control.²⁴ These advances allow for more tolerable, more practical, and more efficacious stimulation, thus promoting increased acceptance among patients and providers.

Sex Disparities for DBS in PD and ET

Our study demonstrated striking sex-based differences in the use of neuromodulation for PD and ET. Such findings have become increasingly important, as recent research has suggested that there are no long-term differences between men and women in DBS for PD²⁵ or ET.²⁶ It is important to note that there was unequal sample representation between the two groups, with men having higher representation in the studies by Golfrè Andreasi et al.²⁵ and Blomstedt et al.²⁶ A recent study and literature review from the University of Virginia has revealed that among patients with PD, men were more likely to receive DBS for medication-resistant tremor symptoms, motor fluctuations, and dyskinesias.²⁷ Overall, while DBS may be equally effective for both men and women, women receive such treatment at significantly lower rates. One could argue that this is attributable to differences in prevalence among females and males; however, prior research has suggested that this may not be the case^17,20 or that the difference is minor.

Racial Disparities in DBS for PD and ET

Black patients receive DBS for both PD and ET at lower rates than other races, and this rate has remained consistently low over time. These results confirm prior studies showing that White patients are up to 5 times more likely to receive DBS for PD treatment than Black patients.⁹ Disparities among other minorities have not been well studied in the context of DBS, but our results suggested that, compared to White patients, Asian patients have similar rates and Hispanic patients have lower rates of DBS for PD. There was also an inconsistent use of DBS for ET in both Asian/Pacific Islander and Hispanic populations over the years, leading to an overall lower utilization rate than in White patients. We can speculate that there is more equivalent use of DBS for PD compared to ET among Asian and Hispanic populations because of the relative increased utilization of DBS for PD versus ET. Historically, PD has been shown to be a more prevalent indication for DBS,²⁸ is overall a more serious condition than the relatively benign ET, and thus may prompt minority patients and nonneurosurgical providers to more seriously consider a referral to neurosurgery to receive this treatment.

Even after controlling for other variables, our study revealed that racial disparities remain significant. Even though a lack of sufficient insurance coverage or socioeconomic status can explain some disparities in minority patient healthcare utilization,^29,30 our results indicated that there is a persistent disparity that cannot be explained by these factors alone. Some initiatives to reduce disparities in the care of minority patients with PD have focused on communication factors, such as increasing the number of Spanish-speaking providers and facilitating telehealth visits,²² which have increased access to neurology and subsequent neurosurgical care.

Increased access to high-volume hospitals for surgical procedures is known to reduce disparities in surgical care for minority groups.³⁰ Historically, Black and Hispanic patients have had less access to tertiary care and have undergone surgery at lower-quality, lower-volume surgical hospitals.^31–33 Even at minority-serving hospitals, physicians are less likely to perform DBS at all.²² This reflects a systemic inequality in access to specialized surgical care; thus, an effort to increase access to DBS centers is needed to equalize DBS care.

Disparities in Comorbidities

Patients who did not receive DBS for both PD and ET had a higher mean ECI. Overall, both the operative and nonoperative groups had a minimal burden of comorbidity, with ECIs under 15 indicating an in-hospital mortality risk of less than 10%.³⁴

It is possible to hypothesize that the operative groups, which had lower indices than the nonoperative groups, were less encumbered by the comorbidities of their conditions and thus were better surgical candidates. For patients with PD, cognitive impairment and neuropsychiatric symptoms are well-known comorbidities that increase the risk of perioperative morbidity.³⁵ ET can also be comorbid with cognitive decline and dementia.³⁶ The current study did not investigate which specific comorbidities were associated with increased or decreased neuromodulation surgery rates. Future research may investigate the relationship between specific comorbidities, including psychological and neuropsychiatric conditions, and surgical patients.

SES Disparities for DBS in PD and ET

It has been established that socioeconomic status influences access to healthcare, with many studies indicating that patients in lower income brackets or those who are underinsured (uninsured/self-pay or using Medicaid) have less access to surgical care, must postpone care because of costs or a lack of insurance approval, and receive overall lower-quality care.^37–40 Given the relatively specialized nature of DBS, which requires high-volume centers and functional neurosurgeons with specialized training, it is not surprising that patients with higher incomes and private insurance utilize DBS more frequently. This disparity does not appear to be resolving, as rates of DBS among Medicaid and Medicare patients did not increase significantly between 2012 and 2019 for either disease cohort.

Geographic Disparities for DBS in PD and ET

Compared to Asian nations, North America, Europe, and Australia have a high prevalence of PD, according to global epidemiology studies.¹⁷ In the US, regions with a high PD prevalence include the Midwest, the Northeast, and southern states such as Texas and Florida.^41,42 According to our study, neuromodulation use for PD was highest in the West and similar in the Midwest, South, and Northeast. This demonstrates a striking disparity between the prevalence of PD and the locations where neuromodulation procedures are conducted, highlighting the need to ensure an even geographic distribution of centers capable of performing these procedures. Although ET remains one of the most prevalent movement disorders in the US, little is known about its geographic prevalence and incidence. Therefore, it is difficult to draw conclusions as to why we found the highest use of DBS in the West and no regional variation elsewhere. Future research should investigate regional differences in the prevalence of ET across the US.

Limitations and Future Directions

Although this study identifies critical healthcare disparities for the neurosurgical community to resolve, there are a number of limitations to consider. We were restricted to using a nationally representative inpatient database and have therefore excluded any patients not included in this dataset. Errors in procedure and diagnosis coding, the exclusion of patients with PD and ET who were not hospitalized, and a lack of information on the demographic and socioeconomic details of the healthcare personnel performing procedures on patients are additional limitations of this type of study.

We also lacked the granularity with national databases to discern data points such as the length of diagnosis, dosage of prior medications, and severity of tremor. These factors help to drive the decision to pursue surgery for individual surgeons and their patients but are not available in the NIS. Future research could examine the nuances of these disparities, such as whether provider demographics or disease severity influences DBS utilization and where in the clinical workflow such disparities arise, such as a lack of patient referrals, lack of follow-up, or a lack of education about these procedures. These studies could aid in identifying the optimal intervention sites to reduce disparities in care. Especially at an institutional level, an analysis of referrals and access to specialty neurology and neurosurgical care could explain how so few Black patients are receiving DBS compared to patients of other races. Finally, NIS discharges are not associated with patient identifiers, so we may be capturing multiple discharges for the same patient. For this study, separate discharges were regarded as separate patient records.

Conclusions

We analyzed disparities in the use of DBS for the treatment of PD and ET in the US by using a national database. Significantly lower DBS utilization rates were observed in Black patients, female patients, patients outside the Western US, those without private insurance, and those from lower income brackets for both PD and ET, as well as in Hispanic and Asian patients with ET. Future research should concentrate on identifying the actionable causes of these disparities and establishing more equitable access to these neuromodulation procedures.

Acknowledgments

This work was made possible in part by grant no. TL1 TR002555 from the National Center for Advancing Translational Sciences of the National Institutes of Health (V.V.).

Disclosures

Dr. Rahimpour is a consultant for Boston Scientific and has received speaker fees from Abbott outside the submitted work.

Author Contributions

Conception and design: Rahimpour, Venkatraman, Lee, Shofty, Lad, Williamson. Acquisition of data: Venkatraman, Futch, Bode Padron, Yang, Lee, Parente. Analysis and interpretation of data: Rahimpour, Venkatraman, Yang, Lee, Parente. Drafting the article: Rahimpour, Venkatraman, Futch, Bode Padron, Yang, Lee, Seas, Parente. Critically revising the article: Rahimpour, Venkatraman, Futch, Yang, Lee, Seas, Parente, Shofty, Lad, Williamson. Reviewed submitted version of manuscript: Rahimpour, Venkatraman, Futch, Yang, Lee, Parente, Lad, Williamson. Approved the final version of the manuscript on behalf of all authors: Rahimpour. Statistical analysis: Yang, Lee. Study supervision: Rahimpour, Lad, Williamson.

Supplemental Information

Previous Presentations

Portions of this work were presented as an e-poster at the 2023 AANS Annual Scientific Meeting held in Los Angeles, California, on April 21–24, 2023.