Trends in Utilization of Upper Extremity Reconstructive Surgery Following Traumatic Brain Injury and Stroke

Abstract

Background: Spasticity resulting from traumatic brain injury (TBI) or stroke can lead to debilitating sequelae, including deformities from joint subluxation and spasticity, causing a loss of functional independence. Despite the effectiveness of surgery to address these issues, it is unclear how often these procedures are performed. The objective of the study was to determine the rate of, and trends associated with, reconstructive upper extremity surgery in patients following TBI or stroke. Methods: The National Inpatient Sample was queried for International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) diagnosis codes for TBI and stroke as well as procedural codes representing functional upper extremity reconstruction from 2001 to 2012. Temporal trends were assessed for case volume, patient demographics, financial considerations, and hospitalizations. Results: A total of 2132 reconstructive procedures were performed in patients with TBI or stroke during the study period, with fewer than 230 cases conducted in any given year and no appreciable increase in case volume over time. This represented less than 1% of eligible, appropriate candidates undergoing surgery. Middle-aged, white females were the most common patients to have such surgery. Medicare was the primary payer for reconstruction, and the cost of surgery increased substantially over time. There was a trend toward longer hospital stays, and the inpatient mortality was approximately 0.5%. Conclusions: There is a substantial underutilization of upper extremity reconstructive surgery for patients with spasticity following TBI or stroke. Increasing costs and limited access to appropriate care may be contributing to differences in use among specific patient subgroups.

Introduction

Neurological and cerebrovascular accidents, such as traumatic brain injury (TBI) and stroke, can be debilitating maladies. Collectively, these conditions are considered the leading cause of disability among adults in the United States.^13,16 The sequelae of TBI and stroke vary in severity and breadth depending on the location and extent of the neurological insult. Cognitive deficits, aphasias, dysphagia, and depression, among others, can result and lead to other medical morbidities, such as aspiration pneumonia, deep vein thrombosis, urosepsis, etc.^5,16 While these complications often occur in a period of weeks following TBI or stroke, the musculoskeletal impairments typically manifest months to years later. These sequelae include muscle contractures and spasticity, joint subluxation and flexion deformities, heterotopic ossification, and several others related to spasticity.^8,16 Patients with upper extremity impairments from these conditions generally perceive bimanual tasks necessitating fine motor dexterity to be most challenging, limiting their ability for self-care and participation in work and leisure activities.⁶ Consequently, the primary goal in managing patients with these morbidities is to restore their functional independence and, thus, enable them to reintegrate into their community.

Nonoperative management of the musculoskeletal complications of TBI and stroke emphasizes progressive rehabilitation and prevention of contracture formation. Specific treatments, such as the use of orthotics and physical therapy, attempt to orient the affected limb/joints in functional positions and encourage range of motion exercises. Pharmaceutical and other modalities, including nerve blocks and electrical stimulation, can be effective adjuncts.¹⁶ However, physical and occupational therapy cannot affect a permanent change in motor control. Systemic drug therapy for increased muscle tone has generalized effects and cannot be targeted to specific offending muscles. Moreover, botulinum toxin injections provide only temporary modulation of muscle tone. Therefore, surgery is often the only treatment that can correct a limb deformity or improve function. From an operative standpoint, the mainstays of treatment to correct upper extremity deformities and improve function involve tendon transfers, muscle lengthening or releases, and joint fusions.^9,15 The particular procedure(s) are dictated by the nature of the deformity and the presence or absence of volitional motor function in the extremity. Surgical interventions have proven to be successful in patients with various neurological injuries, such as hemiplegia, tetraplegia, and cerebral palsy.^4,7,14,17

Despite the effectiveness of operative management, it is unclear how often upper extremity surgery for patients with TBI or stroke is performed. Curtin et al demonstrated substantial underutilization of appropriate surgery for patients with tetraplegia.³ Given the much greater prevalence of spasticity and hemiplegia, and considering the increasing life expectancy of those suffering from neurological insults, there is a marked need for surgery to enable these patients to lead a functional life. Thus, the objectives of the present study were to determine the rate of upper extremity reconstructive surgery following TBI or stroke, as well as analyze trends in the patient demographics, financial considerations, and hospitalizations of those undergoing such surgery.

Materials and Methods

Temporal trends in the utilization of upper extremity reconstruction in patients with late effects of TBI or stroke from 2001 through 2012 were reviewed utilizing the National Inpatient Sample (NIS). The NIS database is the largest all-payer hospital discharge database in the United States. It was developed as a part of the Healthcare Cost and Utilization Project, which is sponsored by the Agency for Healthcare Research and Quality and falls under the Department of Health and Human Services of the US government. The database consists of an approximately 20% stratified sample of US hospitals. Previous studies have employed the NIS database to determine and analyze national trends of several surgeries.^2,3,12 All data contained therein are de-identified; thus, this study was exempt from review by the institutional review board.

To identify patients for the present study, we used the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM). The NIS database was queried for ICD-9-CM diagnosis codes representing ischemic or hemorrhagic stroke and codes for TBI, including late effects of cerebrovascular disease with hemiplegia/hemiparesis (438.2X) or monoplegia of upper limb (438.3X) and late effect of intracranial injury without mention of skull fracture (907.0). Patients were ultimately included if they had an ICD-9-CM procedural code representing functional upper extremity reconstruction (see Supplemental Table S1). These procedure codes primarily encompass a large breadth of tendon surgeries in both the hand and arm, as well as other procedures commonly performed in combination with upper extremity reconstruction including tendon autograft harvest, tenodesis, and joint arthrodesis.

To determine an accurate estimate of the annual rate of reconstructive procedures necessary for patients with spasticity from a TBI or stroke, approximately 730 000 new strokes occur each year within the United States.¹⁶ Of these, about 80% survive the initial event, and 76% of those in the surviving cohort develop spasticity with significant upper extremity dysfunction.¹⁶ This suggests that there are 443 840 new cases of spasticity secondary to a stroke or other TBI within the United States each year. Prior studies regarding the utility of upper extremity reconstruction in the setting of a neurological event estimated that approximately 50% of these patients would benefit from surgery.³ Thus, an estimated 221 920 reconstructive procedures should be conducted on these patients each year.

Statistical analysis was performed using SPSS for statistical computing (IBM Corp., Armonk, New York). Patient demographics (including age, sex, and race), primary payer, length of hospital stay (LOS), mortality (based on death during hospitalization), and total charges were collected for each patient. Temporal trends for each of these variables, as well as surgical volume, were analyzed with Pearson correlation coefficients (r). All P values are 2-tailed with P < .05 indicating statistical significance.

Results

A total of 2132 patients underwent an upper extremity reconstructive procedure following TBI or stroke during the 12-year period. The overall demographics, payer, and hospitalization data for this population are shown in Table 1. The mean age of this patient cohort was 55.6 years, with the majority of patients below the age of 65 years. There was a female predominance (60.5%), and the most common race was white (76.2%) followed by black (12%). Nearly 60% of the procedures were paid for primarily by Medicare, but 27.2% were financed through self-pay. Patients typically remained in the hospital for approximately 5 days with a low inpatient mortality of less than 1%. Consequently, the typical patients who underwent reconstructive surgery during this period were middle-aged white females whose procedures were covered largely by Medicare.

Table 1.

Overall Demographics, Payer, and Hospitalization Data for Entire Study Population (n = 2132).

Characteristic	Value
Age (years)
Mean age (SD)	55.6 (22.3)
Age group (%)
0-44 years	23.8
45-64 years	34.3
65-74 years	22
75+ years	19.9
Sex (%)
Male	39.5
Female	60.5
Race (%)
White	76.2
Black	12
Hispanic	7.6
Asian/Pacific Islander	2.9
Native American	0.3
Other	1
Insurance (%)
Medicare	58.8
Medicaid	10
Self-pay	27.2
No charge	1.2
Other	2.8
Other trends (SD)
LOS (days)	5.4 (7.9)
Mortality rate (%)	0.47
Total charges ($, per case)	44 133 (70 883)

Note. SD = standard deviation; LOS = length of stay.

Patient demographics (Table 2), financial assessments (Table 3), and hospitalization details (Table 4) were also analyzed for each calendar year. With specific regard to demographics, the mean patient age generally increased over the study period, with the lowest age of 49.4 years in 2002 to a high of 63.3 years in 2011. Corroborating this, as noted in Table 5, age had a moderate positive correlation with time (r = 0.59). In addition, while female patients predominated during the study period, there was a general decline in their representation (and a converse increase in the percentage of males). This is reflected in a moderate negative correlation between the percentage of females and time (r = −0.35) and a moderate positive correlation in the percentage of males (r = 0.55). Moreover, with respect to race, while white patients predominated, the percentage of black patients increased over time (as evidenced by a strong moderate correlation, r = 0.59) at a correspondingly higher rate than their white counterparts (r = 0.31). The percentage of the remaining races had only negligible or weak correlations with time.

Table 2.

Patient Demographics Per Calendar Year.

Parameter	Calendar year
	2001	2002	2003	2004	2005	2006	2007	2008	2009	2010	2011	2012
Age (years)	54.1	49.4	56.6	55.6	52.7	52.9	56.3	50.6	55.4	61	63.3	57.1
Female (%)	61.6	75	70.6	60.7	66	62.7	74.2	48.1	64.6	48.3	62.6	43.6
White (%)	82.8	85.4	80.5	82.7	65.8	66.7	88.3	65	69.4	81.4	80	71.9

Table 3.

Financial Assessments Per Calendar Year.

Parameter	Calendar year
	2001	2002	2003	2004	2005	2006	2007	2008	2009	2010	2011	2012
Total charges ($)	15 960	15 733	35 111	28 976	43 231	28 075	29 672	45 462	72 371	66 216	58 943	70 074
Medicare (%)	53.1	55.1	75.8	60	48.8	49.1	62.9	53.5	47.8	61.1	72.7	68.4
Medicaid (%)	15.2	2.6	2	6.7	15.6	13.6	17.7	10.2	15.8	7.2	7.6	5.3
Self-pay (%)	24.6	39.1	19.6	30.7	31.9	31.4	12.1	33.7	31.5	31.7	19.7	18.4

Table 4.

Hospital Assessments Per Calendar Year.

Parameter	Calendar year
	2001	2002	2003	2004	2005	2006	2007	2008	2009	2010	2011	2012
LOS (days)	4.5	3.1	5	5.7	6.6	3.3	4.7	5.4	7.2	7.9	4.7	5.7
Mortality (%)	0	0	0	0	3.1	0	0	0	0	0	0	2.6

Note. LOS = length of stay.

Table 5.

Correlation Between Each Parameter With Respect to Time (Calendar Year).

Parameter	Pearson correlation coefficient (r)	P value
Case volume	0.29	.37
Age	0.59	.04
Male	0.55	.06
Female	−0.35	.27
White	0.31	.32
Black	0.59	.05
Hispanic	0.18	.57
Asian/Pacific Islander	0.27	.39
Native American	0.13	.68
Other (race)	−0.04	.89
Total charges	0.89	< .01
Medicare	0.46	.13
Medicaid	0.01	.97
Self-pay	−0.06	.85
No charge	0.04	.9
Other (payer)	−0.21	.52
LOS	0.47	.12
Mortality	0.26	.42

Note. LOS = length of stay.

Figure 1 depicts the number (volume) of upper extremity reconstructive surgeries performed during each calendar year for patients with TBI/stroke. As illustrated in the graph, the case volume ranged from a low of 124 to a high of 229 procedures per year. Thus, there was a weak positive correlation between the case volume and time (r = 0.29). With only an average of 178 reconstructive surgeries being performed in this patient population yearly, and considering that surgery is the preferred treatment in an estimated 221 920 annual cases, less than 1% of patients are being ideally managed with surgery.

Figure 1.

Case volume per calendar year.

Temporal changes in the financial aspects of the reconstructive procedures were also identified. Table 3 demonstrates that there was a substantial increase in the total cost per surgery, with the largest and sustained increase occurring in 2009 onward with a nearly $27 000 per case surge from 2008. This manifested as a strong positive correlation between total (per case) charges and time (r = 0.89). With regard to the expected primary payer, Medicare remained the most common source every calendar year, covering approximately 50% or more of all cases per annum. This corresponded with a moderate positive correlation between Medicare coverage and time (r = 0.46). Notably, self-pay was the second most common primary payer, financing 12.1% to 39.1% of all cases. Unlike Medicare coverage, however, self-pay (and the other expected primary payers) had only negligible or weak correlations with time (r < 0.21).

Details of the hospitalizations were also recorded (Table 4). The typical LOS following upper extremity reconstructive surgery ranged from approximately 3 to 8 days. There was a trend toward longer hospital stays, with a moderate positive correlation between LOS and time (r = 0.47). In addition, of the 12 years analyzed in the present study, inpatient deaths were found in only 2 years (2005 and 2012), with a 0% mortality in the intervening years.

Discussion

Surgery is a powerful rehabilitation tool. Various surgical treatments, including tendon transfers, arthrodesis, and musculotendinous releases, have been shown to be effective in treating the deformities associated with spastic patients with familiar patterns of upper motor neuron dysfunction.^1,9,10,16 Nonetheless, despite this evidence, the present study demonstrates that upper extremity reconstructive surgeries in this patient population are extremely underutilized. Specifically, less than 1% of patients who suffered TBI or stroke with residual spasticity (and were appropriate operative candidates) had a reconstructive procedure. In addition, there was no appreciable increase in the volume of such appropriate cases over the 12-year study period, suggesting that, even by conservative estimates, there is no indication that more surgeries will be performed in the near future. An investigation performed by Curtin et al similarly concluded that surgery is substantially underutilized in the tetraplegic population with relatively stagnant case volume numbers over time.³

The current study also shows that, of those patients who do undergo upper extremity reconstruction, the “typical” patient is a white, middle-aged female. Evaluating particular temporal trends more closely, the present investigation demonstrated that the demographics of this surgical population are changing. Notably, the mean age of those patients undergoing upper extremity reconstruction is increasing over time. This is likely attributable to the fact that TBI/stroke often occurs at older ages and these patients are tending to live longer.¹¹ In addition, while females are still the predominant sex, the number of males undergoing reconstruction is steadily increasing (with a resultant decrease in the relative percentage of females). Because these neurological injuries, especially stroke, occur more often in males, this trend suggests that the surgical profile is beginning to more closely approximate the characteristics of those actually affected by these conditions.¹¹ Similarly, while the majority of surgical patients are white, the percentage of black patients is increasing with time. This could indicate slightly improved access to these reconstructive surgeries between races, although the overall rates among minorities remain low. In addition to demographic changes, certain financial trends were also observed. Perhaps the most striking of these was the substantial increase in total charges over time. The lowest mean cost per case of $15 733 in 2002 ballooned by approximately 360% to $72 371 in 2009, and this trend toward higher costs has been largely sustained. This imposes a massive financial burden on both the patient and the greater health care system as a whole, increasing well beyond the rate of inflation. Such substantial costs may make patients and health care providers reluctant to pursue surgical interventions, thereby contributing to their minimal utilization. Moreover, the current study highlights the large role that Medicare plays as the primary payer, with a moderate increase in the percentage of surgeries covered by Medicare. This increase is likely a reflection of the aforementioned increasing age of the surgical population. Furthermore, the study illustrated that the mean hospital LOS is increasing over time, which may be due to the complexity of the reconstructive surgeries being performed, the need to manage other underlying comorbidities, and coordination of additional postoperative home care services. The increase in LOS is likely also contributing to the rising total costs. Last, the overall mortality during hospitalization for upper extremity reconstruction remains low, perhaps due to the relatively safe nature of the procedures and effective medical management of these patients’ comorbidities (which are the primary factor with respect to mortality).

As previously mentioned, Curtin et al conducted a review of the use of reconstructive procedures in patients with tetraplegia.³ The current study shares certain similarities with the findings found by their group. The present study and theirs both found profound underutilization of these surgeries among hemiplegic and tetraplegic patients, respectively. Although not as dramatically, Curtin et al also noted that the mean age of the surgical population increased with time and that the predominant race was white. There are, however, notable differences between the 2 studies. Specifically, while Curtin et al assessed 4 distinct years in the late 20th century, the current study assessed a much larger, 12-year period within the 21st century to enable a more robust analysis. In addition, the hemiplegic surgical population was found to be largely female, while the tetraplegic population reported by Curtin et al was overwhelmingly male. Furthermore, Curtin et al noted that Medicaid was the primary payer initially, but its contribution diminished with time as Medicare ultimately became the majority payer. In the present study, the percentage of cases covered by Medicare was also increasing, but contributions by Medicaid had no appreciable change. Also, while there was no evidence of self-pay among the tetraplegic population, a fair percentage of spastic patients elected self-pay, possibly because not as many patients had insurance as noted by the relatively low Medicaid/non-Medicare payer numbers. Unlike the Curtin et al study, the current investigation delved deeper into the data to provide new insights on total cost, LOS, and mortality.

This study, however, is not without its limitations. Principally, the data presented here are estimates based on the NIS database. Consequently, the actual case numbers may vary slightly from those reported. Nonetheless, prior studies have successfully utilized the NIS database to determine nationwide trends.^3,12 Moreover, the database only accounts for inpatient, not outpatient, procedures. Thus, it is possible that the data underestimated the overall case volume rates. However, given the massive underutilization noted in the present study, and considering that Curtin et al found very few outpatient cases performed for their tetraplegic cohort, this is likely negligible and would not have changed the overall trends.³ Furthermore, the correlations analyzed in this study represent trends in the data but do not definitively indicate a cause-and-effect relationship.

Ultimately, the present investigation found substantial underutilization of upper extremity reconstructive surgery in patients with spasticity from TBI or stroke. The data highlight the massive, increasing financial burden of these surgeries. Therefore, costs must be modulated (perhaps by decreasing hospital stays) to help mitigate the financial barrier to these effective surgeries. This is even more important considering that a notable percentage of patients elect self-pay. Physician education and awareness about the effectiveness of reconstructive procedures to increase familiarity, coordination with neurologists and primary care providers regarding the surgical services that are available, and communication with patients is crucial to their future widespread adoption and implementation.