“Cost Assessment of Plating versus Tension Band Wiring Constructs for the Treatment of Mayo Type 2A Olecranon Fractures”

Abstract

Purpose:

For adults with displaced Mayo 2A olecranon fractures, tension band wiring (TBW) or plating may be used for fixation with the expectation of similar clinical outcomes. The primary hypothesis was that surgical encounter total direct costs (TDC) are lower for TBW than plating. Our secondary hypothesis was that combined surgical TDC are lower for TBW even with a 100% rate of subsequent tension band hardware removal and a 0% rate of plate removal.

Methods:

Patients who underwent TBW or plating of an isolated unilateral Mayo 2A olecranon fracture between July 2011 and January 2020 at a single academic medical center were identified. Using information technology cost tools provided by our institution, TDC for each surgical encounter pertaining to plate fixation, TBW, and hardware removal were obtained and converted to 2020 US dollars. Relative TDC were compared between plate fixation and TBW groups using univariate and multivariable generalized estimating equations with log-link.

Results:

Of the 97 included patients, the mean age was 50 ± 21 years and 48% were female. TBW and plate fixation were performed on 18% (17/97) and 82% (80/97) of patients, respectively. Demographics were similar between groups, although the finding that plate fixation cost 2.6 times the cost of TBW within the index surgery was significant in the multivariable model, independent of potential confounders (coefficient 2.55, 95% confidence interval: 2.09 to 3.10). Mean TDC remained significantly greater for plate fixation even under the hypothetical situation where 100% TBW were removed, and the plate removal rate was 0% (cost difference 181%).

Conclusions:

The use of tension band wiring, relative to plate fixation, may improve the cost of care for operative Mayo 2A olecranon fractures. This finding was robust to the rate of hardware removal.

Level of Evidence:

III: Economic and Decision Analyses

Introduction:

Olecranon fractures are common in adults, comprising 10% of all upper extremity fractures, with an incidence of 12 per 10,000 individuals.^1,2 Numerous methods to treat olecranon fractures have been proposed, dependent on patient demographics and fracture characteristics. Studies have suggested that for all age groups, and in the absence of other clinical indications, nonoperative treatment is preferred for isolated, non-displaced olecranon fractures.^3,4 Authors have even suggested that nonoperative treatment may be preferable in certain simple displaced olecranon fracture cases,⁵ especially among elderly patients.^6–8 Despite this evidence, the current standard of treatment for displaced olecranon fractures is surgical with open reduction and internal fixation (ORIF).⁹ Various techniques have been developed for these fractures such as intramedullary nailing,¹⁰ suture fixation,¹¹ tension band wiring (TBW) with Kirschner and cerclage wires, and plate fixation. Traditionally, in cases of simple displaced (Mayo type 2A) fractures, TBW is an option. In cases of comminuted articular surfaces or oblique displaced (Mayo type 2B) fractures, plate fixation is the most frequently used method.¹² However, recent work has suggested that postoperative union rates, functionality, and patient-reported outcomes are no different between these two methods in the treatment of simple or oblique displaced fractures.^9,13–18 These comparisons, however, can only be drawn when fracture classification is correctly diagnosed, because with more severe, comminuted fractures, utilizing the incorrect fixation method could lead to inadequate construct stability and unfavorable outcomes.⁴

Although the clinical efficacy of TBW and plate fixation for displaced olecranon fractures has largely been found to be similar, a notable criticism of TBW is its relatively high rate of hardware removal, which increases overall treatment costs through the requirement for additional surgery.^19,20 This rate has been reported to be almost 2-times higher following TBW than plate fixation.²⁰ Some common reasons contributing to this high rate of removal include pain, implant prominence, functional impairment, implant breakage, and hardware migration.^15,20 Therefore, with similar outcomes in the treatment of Mayo type 2A fractures, and in light of potentially differing rates of hardware removal, the optimal surgical technique remains in question. Further knowledge of the costs related to index surgeries (TBW and plating) and subsequent hardware removal may inform this discussion.

Cost payment models have gained increasing popularity in the last decade and have relevant implications for upper extremity surgery.²¹ Our institution possesses an internal, value-driven outcomes (VDO) database that contains specific patient and item-level total institutional direct cost (TDC) and payment data for a variety of health care services, including surgery. This validated VDO resource has effectively identified areas of large variability in cost, leading to improved value of care.^21,22–29 This VDO tool was utilized to test two hypotheses. Our primary hypothesis was that the TDC associated with TBW are significantly less than for plating. Our secondary hypothesis was that the TDC of the combined surgical encounters are lower for tension band wiring even under the hypothetical situation in which 100% of tension band hardware is subsequently removed and 0% of plates are removed.

Methods:

Our IRB-approved retrospective cohort study consisted of all adult patients (≥ 18 years) who underwent ORIF of an isolated, unilateral, displaced, non-comminuted olecranon fracture (Mayo 2A)³⁰ between July 2011 and January 2020 at a single tertiary academic institution. Using our institution’s electronic data warehouse, patients were identified by current procedural terminology code (CPT 24685). Patients who underwent any simultaneous procedures at the time of olecranon fixation were excluded to avoid confounding the surgical total direct cost. Manual chart review was performed to verify demographic factors, treating specialty (hand surgery versus orthopaedic trauma surgery), fixation type (TBW versus plating), injury-specific variables, intraoperative variables, and whether hardware was subsequently removed. Injury radiographs were also reviewed to assign the fracture pattern using the Mayo Classification.³⁰ Patient inclusion, data extraction, and fracture classification were performed by two authors (J.N.S and A.R.S) and validated for accuracy with the corresponding author (N.H.K.). TBW and plate fixation were performed as previously described.²²

TDC data for each index surgery and hardware removal surgery were obtained using our institution’s value-driven outcomes (VDO) tool, which tabulates item-level direct costs for all clinical encounters at our institution.²¹ Specifically, prospectively collected cost data is linked to individual patient surgical encounters and is available upon an internal query. TDC is comprised of various granular sub-categories including supply, facility, implant, operating room (OR) utilization, post-anesthesia care unit (PACU) utilization, intraoperative imaging, pharmacy, and lab costs. Professional personnel costs associated with the patient encounter, and indirect costs such as lost wages for patient recovery, are not included the TDC. As we noted a mixture of inpatient and outpatient surgeries for olecranon ORIF, we defined the surgical encounter TDC as the sum of OR and PACU utilization, implant, and supply costs for this study. This was done because the VDO tool records the costs of laboratory studies, imaging, and medications for an entire inpatient stay – it is not possible to determine the costs of those services that are specific to the surgical encounter. Nonetheless, prior studies utilizing this VDO tool in orthopedics have reported that pharmacy, laboratory, and imaging costs comprise less than 5% of the surgical encounter TDC,^23–25 and at our institution there is little variation in these parameters given the care pathways in place for olecranon ORIF. Patients universally receive a single dose of prophylactic antibiotics, a miniature c-arm is used intraoperatively, and no laboratory studies are routinely obtained. To comply with institutional policies that preclude reporting of actual cost figures related to non-publicly disclosed contractual agreements, all reported cost data was reported as a relative cost by dividing the cost of each patient’s encounter by the median total direct cost of the data set.

In addition to determining surgical encounter TDC associated with TBW, a secondary analysis was performed to evaluate the “best/worst case scenario” regarding hardware removal rates between these two fixation methods. This secondary analysis considered the hypothetical situation in which TBW is associated with a 100% rate of hardware removal, and plating is associated with a 0% hardware removal rate. To do so, the TDC of the subsequent hardware removal surgeries were calculated using the VDO tool for included patients who subsequently underwent hardware removal. “worst case” total treatment costs for TBW were then calculated as the sum of the median hardware removal cost and the cost of the index TBW surgery for each patient. The “best case” total treatment cost for plating was the median cost for all patients who underwent plating with no additional costs related to hardware removal.

Statistical Methods

Patient and clinical variables, including postoperative complications, were summarized descriptively, and compared between fixation methods. Continuous variables were summarized as mean (standard deviation: SD), median (interquartile range: IQR) and range, and compared using t-test or exact Wilcoxon rank sum test depending on data normality. Categorical variables were summarized as N (%).

Since data was available from July 2011 and January 2020, we compared costs from different years on equal footing by first converting costs to 2020 dollars using the Consumer Price Index. Each cost type was divided by the median cost of tension bands to facilitate comparisons between plate and tension bands. Given that the distribution of our cost data was skewed right, we used a modified Park test to select an appropriate outcome model.³¹ A gamma distribution provided optimal fit, and a log-link was used so that exponentiated cost results represented ratios. We compared total cost with fixation methods and within the plate group using univariable generalized estimating equations (GEE) models to account for clustering of patient outcomes within providers. Predictors with univariable p-values <0.1 were selected for multivariable analysis. Multicollinearity within the multivariable model was evaluated by calculating the variance inflation factor (VIF) for each predictor. To assess a worst-case scenario for the tension band group, we calculated the median removal cost for patients within the tension band group and added it to the total costs for these patients, and compared the costs with the plate group using an exact Wilcoxon rank sum test. Statistical significance was evaluated at the 0.05 level using two-sided tests.

An a priori power analysis was performed that accounted for the higher proportion of plate fixation relative to TBW cases historically seen at our institution (5:1 ratio). With the large cost effect size previously reported between these two fixation types (mean ± SD of $14,160 ± 3876.02 for plate and $5,171.06 ± $2,367.54 for TBW),¹³ we would require 18 total patients (15 plate fixation and 3 TBW) to achieve 80% power with a two-sided 0.05 significance level using a two-sample t-test. Given that we anticipated additional analyses, we planned to collect a minimum of 10 TBW.

Results:

Of the 97 patients included, mean age was 50 ± 21 years at the time of surgery and 48% were female. Plating was performed in 82% (80/97) of the cohort, and TBW was performed in 18% (17/97). Baseline characteristics and surgical details are described in Table 1. Plating and TBW groups were similar in terms of demographic variables. Both groups were similar regarding surgical factors except for differences in surgery time (mean of 0.6 hours greater for plating) and the distribution of the treating subspecialty using tension bands (82% of TBW performed by hand surgeons versus 18% by trauma/other surgeons). Postoperative union and complication rates are described in Table 2 and were found to be similar the two fixation methods.

Table 1:

Descriptive Summary of Baseline Patient Characteristics and Surgical Details

Variable	Total	Plate/Screw	Tension Bands
	N= 97	N= 80	N= 17
Demographic Factors
Age (year)	50 ± 21	50 ± 22	48 ± 20
Female	47 (48%)	40 (50%)	7 (41%)
Body Mass Index	26 ± 7	26 ± 7	27 ± 8
Race
Not White or Caucasian	8 (8%)	5 (6%)	3 (18%)
White or Caucasian	89 (92%)	75 (94%)	14 (82%)
Risk Factors
Osteopenia/Osteoporosis	15 (16%)	13 (17%)	2 (12%)
Tobacco Use	16 (17%)	13 (17%)	3 (18%)
Insurance ^
Commercial	46 (47%)	35 (44%)	11 (65%)
Government Other	14 (14%)	13 (16%)	1 (6%)
Medicaid	3 (3%)	3 (4%)	0 (0%)
Medicare	23 (24%)	19 (24%)	4 (24%)
Workers Compensation	5 (5%)	5 (6%)	0 (0%)
Self-Pay	6 (6%)	5 (6%)	1 (6%)
Surgical Details
Treating Subspecialty
Hand	45 (46%)	31 (39%)	14 (82%)
Trauma	39 (40%)	37 (46%)	2 (12%)
Other Orthopaedic	13 (13%)	12 (15%)	1 (6%)
Operative Information
Injury to Surgery (days)	7 ± 14	5 ± 12	13 ± 23
Surgery Time (hours)	1.6 ± 0.6	1.7 ± 0.6	1.1 ± 0.2
Bone Graft	16 (16%)	15 (19%)	1 (6%)
Triceps Suture	20 (21%)	18 (22%)	2 (12%)
Post-operative Block	35 (36%)	31 (39%)	4 (24%)
Fo11ow-up (months)	5.0 ± 5.5	5.3 ± 5.9	3.4 ± 2.9
Plating Details
# of Plates	1.0 ± 0.3	1.0 ± 0.3	N/A
# of Screws	7.0 ± 2.4	7.0 ± 2.4	N/A
# of Locking Screws	2.9 ± 2.3	2.9 ± 2.3	N/A

Data represents mean ± standard deviation, or count (percent of total). Bolded p-values indicate a significant difference. Missing values by group: Follow up (months)=5/1, Weeks to hardware removal=64/12, BMI=3/0, Osteopenia/Osteoporosis=4/0, Tobacco=5/0, #Plates=1/0, #Screws=1/0, #Locking screws=2/0, Plate manufacture=3/17, Fracture classification=3/3, Days from injury to surgery=2/1.

Table 2:

Postoperative Union and Complication Rates

Variable	Total	Plate/Screw	Tension Bands
	N= 97	N= 80	N= 17
Fracture Union	96 (99%)	79 (99%)	17 (100%)
Complication
Hardware Irritation	24 (25%)	19 (24%)	5 (29%)
Limited Range of Motion^	12 (12%)	10 (13%)	2 (12%)
Infection	1 (1%)	1 (1%)	0 (0%)
Contracture	2 (2%)	2 (3%)	0 (0%)
Nerve Impingement	3 (3%)	3 (4%)	0 (0%)
Ulceration	1 (1%)	1 (1%)	0 (0%)
Bursitis	2 (2%)	2 (3%)	1 (6%)
Reoperation
Hardware Removal	21 (21%)	16 (20%)	5 (29%)
Irrigation and Debridement	1 (1%)	1 (1%)	0 (0%)

^{^}Limited range of motion is defined as a lack of return to pre-injury range of motion in expected time frame, requiring longer follow-up and therapy consultation.

Regarding index olecranon ORIF, use of plate fixation was associated with significantly higher mean total direct costs, as compared to TBW, by a factor of 2.38 (95% confidence interval [CI]: 2.16 to 2.63, p < 0.05) on univariate analysis. Median and interquartile values are reported in Figure 1. Furthermore, except for PACU costs, all other cost categories (OR, implant, and supply) were significantly lower for TBW than plating (Table 3). A univariate gamma regression analysis identified relevant variables that had a significant effect on the TDC among the two groups (Table 4). Controlling for these identified potential confounders, a multivariable generalized estimating equation gamma regression analysis also revealed that plating was associated with significantly higher surgical encounter TDC than TBW when accounting for other potential confounders (2.55, 95% CI: 2.09 to 3.10, p < 0.05). This may be interpreted as a 155% cost increase when using plate fixation. Further, each additional 60 minutes of surgery time increased TDC by 41% (coefficient 1.41, 95% CI: 1.13 to 1.77, p < 0.05).

Figure 1:

Total Direct Costs of Plate and Tension Band Wiring Fixation in Two Scenarios

Box and whisker plots illustrating the total direct costs of the two surgical fixation methods in two scenarios. A) Initial implantation of plate fixation and TBW and B) a hypothetical case where 0% of plates were removed and 100% of tension bands were removed. In both comparisons, TBW median TDC were standardized as 1, although the standardized values are equal, they represent different actual dollar amounts. *Significantly different TDC relative to paired TBW subgroup (p < 0.05). TDC: total direct cost. TBW: tension band wiring.

Table 3:

Surgical Encounter Total Direct Costs

Variable	Tension Band Wiring	Plate Fixation	p-value
Initial Implantation Surgery
PACU	1.00 (0.78 to 1.24)	0.85 (0.61 to 1.21)	0.56
Supply	1.00 (0.14 to 2.88)	3.09 (2.46 to 4.28)	<0.05
Operating Room	1.00 (0.85 to 1.08)	1.37 (1.11 to 1.58)	<0.05
Implant	1.00 (0.16 to 5.65)	33.02 (29.51 to 36.95)	<0.05
Total	1.00 (0.78 to 1.20)	2.36 (2.21 to 2.85)	<0.05
Hypothetical Scenario with 100% tension band and 0% plate removal
Total	1.00 (0.85 to 1.14)	1.66 (1.56 to 2.01)	<0.05

Data represents median (interquartile range). Bolded p-values indicate a significant difference. In both comparisons, tension band median total direct costs of each category were standardized as 1. PACU: post anesthesia care unit.

Table 4:

Univariate Gamma Regression Analysis

Variable	Ratio	p-value
Weeks since injury	0.98 (0.96,1.01)	0.27
Age (per 10 year increase)	1.01 (0.98,1.04)	0.56
Female Sex (versus Male)	1.03 (0.91,1.16)	0.68
White Race (versus Other)	0.99 (0.70,1.40)	0.96
Body Mass Index	1.01 (0.99,1.02)	0.50
Osteopenia/Osteoporosis	0.94 (0.80,1.11)	0.46
Tobacco User (versus Non-User)	1.41 (1.06,1.88)	<0.05
Insurance (Commercial)	--Reference--	-
Other Government	1.31 (0.99,1.73)	0.06
Medicaid	1.16 (0.87,1.55)	0.32
Medicare	1.02 (0.87,1.20)	0.81
Self-Pay	1.06 (0.93,1.21)	0.39
Workers Compensation	0.97 (0.81,1.16)	0.73
Anesthesia Status	1.09 (0.93,1.28)	0.30
Surgery time	1.41 (1.13,1.77)	<0.05
Number of Plates	2.27 (1.90,2.72)	<0.05
#Screw	1.12 (1.09,1.15)	<0.05
#Locking screws	1.07 (1.04,1.11)	<0.05
Plate Manufacturer 1	0.91 (0.80,1.04)	0.16
Plate Manufacturer 2	0.82 (0.70,0.95)	<0.05
Bone graft	1.23 (1.05,1.44)	<0.05
Subspecialty -Trauma	1.15 (0.91,1.47)	0.25
Subspecialty-Other	1.20 (0.96,1.51)	0.11
Postop block	1.00 (0.90,1.12)	0.96
Suture	1.08 (0.90,1.29)	0.43

Data represents estimate (95% confidence interval) of the mean total direct cost ratio among plate fixation and tension band wiring groups. Bolded p-values indicate a significant difference. However, all but tobacco use were inherent characteristics of the fixation methods.

Although the rate of hardware removal did not differ considerably among the two groups, an analysis was performed that investigated the difference in median TDC between the “worst case” for TBW patients where hypothetically all patients had their hardware removed, and the “best case” for plate fixation patients where hypothetically no patients had their hardware removed. Even in this extreme scenario, the difference in total direct costs between the two groups was statistically significant (p < 0.05; Table 3). After creating this worst- and best-case scenario, to facilitate cost comparisons between the two groups, we divided costs in both groups by the median cost in the tension band group. Thus, by design, the tension band group had a combined standardized median total direct cost of 1.00 (IQR: 0.85 to 1.14) while the plating group, only considering the index surgery, demonstrated a standardized median total direct cost of 1.66 (IQR: 1.56 to 2.01; Figure 1).

Of the patients who underwent hardware removal, the median time from initial surgery to removal was 27 weeks (Table 5). Neither time to removal nor rate of removal were considerably different among the two groups. However, the median TDCs associated with removal of both hardware demonstrated large variation. Median standardized TDC of 1.00 (IQR: 0.84 to 1.17) and 1.95 (IQR: 1.46 to 2.10) were found for the TBW and plate fixation groups, respectively (Table 5).

Table 5:

Hardware Removal Information

Variable	Total	Tension Band Wiring	Plate Fixation
Number of patients	21 (21%)	5 (29%)	16 (20%)
Time to removal (weeks)	27 (22 to 49)	24 (18 to 28)	28 (22 to 51)
Total TDC of removal*	-	1.00 (0.84 to 1.17)	1.95 (1.46 to 2.10)

Data represents count (percent of group total), or median (interquartile range). Tension band wiring median total direct costs (TDC) were standardized as 1.

^*TDC includes direct costs of the operating room and post-anesthesia care unit usage, implants, and supplies.

Discussion:

The main finding of this study is that surgical encounter total direct costs for tension band wiring are significantly less when compared to plate fixation upon initial implantation. Specifically, tension band surgery is 39% the mean direct cost of plate fixation when potential confounders are controlled. Additionally, even in the hypothetical extreme scenario where all tension bands are removed in a second surgery, and no plates are removed, the cumulative TDC associated with both tension band surgeries is still appreciably lower than the singular plate fixation index surgery. Considering these findings, we were able to confirm both our primary and secondary hypotheses that there would be a difference in total direct cost within these two scenarios.

The findings of this study are consistent with previous studies that have similarly investigated cost data associated with olecranon fracture repair.^13,20 In their study of Mayo 2A olecranon fractures, Amini et al. also showed in their cohort of 20 patients divided equally between TBW and plate fixation groups that the charges to patients of the index surgery associated with plate fixation were significantly greater at a value of $14,160.26 compared to the cost of tension band implantation at $5,171.06.¹³ Similar to our study, this cost ratio between fixation types is roughly 40%. An analysis on hardware removal was also performed in this article. Although our studies relate in this capacity, they differ given that Amini et al. evaluated the surgical charges billed to the patient found in their medical records whereas we were able to evaluate the internal, unique, direct cost of each surgical encounter.³² This ability, provided by our institution’s VDO tool, allowed us to demonstrate that the TDC subcategories of supply, implant, and OR utilization were all significantly less costly for tension band wiring compared to plate fixation. When the results of our study, evaluating institutional costs, and those of Amini et al., evaluating patient costs, are considered together, they strongly support the notion that tension band wiring is significantly less costly. These findings are further verified by the work of Francis et al. who constructed a Monte Carlo simulation consisting of 100,000 trials assessing the cost-effectiveness of TBW and plate fixation considering critical variables such as implant removal rate (34.4% for TBW, and 17.5% for plate fixation) that was gathered systematically from available literature. They similarly concluded that TBW is a more cost-effective strategy in treating appropriate displaced olecranon fractures, even with a higher rate of removal and reoperation.²⁰

The benefits stemming from the financial advantage of TBW over plate fixation would be negligible if the performance of plate fixation was found to be greater in the correction of displaced olecranon fractures. However, numerous studies have reported that union rates, functional scores, and patient-reported outcomes are similar between the two fixation types in the treatment of displaced fractures.^9,13–18,33 The nonunion and complication rates, were measured in this investigation and were found to be similar between fixation types.

Concerning patient reported outcomes, Duckworth et al. performed a prospective randomized trial that found no differences in patients treated for a displaced olecranon fracture with either fixation type in range of motion, Disability of the Arm, Shoulder and Hand (DASH) score, Broberg and Morey score, and Mayo elbow score at all postoperative assessment points up to one year.⁹ Gathen et al. likewise demonstrated no functional difference between the two fixation types irrespective of fracture classification and also showed no difference in time off work following surgery.¹⁸ Biomechanical studies have also concurred that both methods provide similar compression to the fracture site, contributing to the similar clinical outcomes reported.^34,35Taken together, it appears that TBW yields a similar postoperative clinical, functional, and biomechanical outcome at a lower cost than plating, and therefore probably better value, for patients with operative Mayo 2A fractures. A formal cost-utility analysis may be useful to further elucidate this.

This study has limitations that should be considered when interpreting the findings. First, the reported results are unique to the pricing contracts made between our institution and applicable suppliers. Therefore, the generalizability of our results should be understood within this context and may differ for other health institutions. Second, our study was a retrospective assessment of patient information and may be vulnerable to selection bias. Third, the cost assessment in this study is unique to the surgical encounter itself and only direct institutional costs from the index surgery were studied. Direct costs of any complication other than hardware removal as well as any indirect costs such as time off work were not individually calculated. Future cost-utility studies that include these factors are required to test the observations in the current study. Fourth, there is a possibility that patients who underwent their index surgery within our institution, underwent hardware removal outside of our institution. This is unlikely to have an impact on the qualitative nature of the findings, but it represents a common limitation of retrospective studies. Lastly, a cost-effectiveness analysis would be the best way to evaluate the effect of complications on overall costs, potential for decrease in function or quality of life, or indirect costs related to lost wages.

Our study demonstrated that the use of tension band wiring was significantly less costly than the use of the plate fixation in the operative correction of displaced olecranon fractures. Given the similar efficacy previously identified between these two surgical fixation types, the reduced cost of tension band wiring, even with subsequent removal, is a potential area to reduce operative care costs.