The Journal of Arthroplasty xxx (2019) 1e8

Contents lists available at ScienceDirect

The Journal of Arthroplasty

journal homepage: www.arthroplastyjournal .org

Logistical and Economic Advantages of Sterile-Packed, Single-UseInstruments for Total Knee Arthroplasty

Tyler D. Goldberg, MD a, John A. Maltry, MD b, Mukesh Ahuja, MBBS, MS, CPI c, *,Jason A. Inzana, PhD d

a Texas Orthopedics, Sports and Rehabilitation Associates, Austin, TXb Tucson Orthopaedic Institute, Tucson, AZc Medacta USA, Inc, Chicago, ILd Telos Partners, LLC, Denver, CO

a r t i c l e i n f o

Article history:Received 18 October 2018Received in revised form25 February 2019Accepted 4 March 2019Available online xxx

Keywords:total knee arthroplastysterilizationsingle-use instrumentspostdischarge complicationsefficiency and economic benefits

One or more of the authors of this paper have disconflicts of interest, which may include receipt of payinstitutional support, or association with an entity imay be perceived to have potential conflict of intedisclosure statements refer to https://doi.org/10.1016/

https://doi.org/10.1016/j.arth.2019.03.0110883-5403/© 2019 Published by Elsevier Inc.

a b s t r a c t

Background: Total knee arthroplasty (TKA) is well established as a clinically successful and cost-effectiveprocedure. The transition of the US healthcare system from a fee-for-service model to a value-based caremodel requires careful examination of patient care to ensure both quality and efficiency. Sterile-packed,single-use instruments have been introduced as a tool to help streamline the operating room (OR) logisticswhile reducing sterilization requirements. The aim of this study was to examine the potential logistic andeconomic benefits of single-use instruments compared to traditional, reusable instruments for TKA.Methods: Four variables related to TKA costs and logistics were modeled in this study: OR turnover timetray sterilization, tray management time, and 90-day infection rates. Model input data for traditionalinstruments and single-use instruments were based on peer-reviewed literature. A total of 200 sites and500 cases per site were simulated using the Monte-Carlo-Technique.Results: The median total cost savings with single-use instruments was $994 per case. The largest driverfor cost savings was tray sterilization. Sites with higher staff wages and sterilization costs had a largerprobability of realizing greater cost savings with adoption of single-use instruments. In cases usingsingle-use instruments, up to 51% of operating days could have accommodated an additional proceduredue to the time savings in OR turnover.Conclusion: This cost modeling study observed significant potential for logistical and economic im-provements in TKA with single-use vs reusable instruments. Although few studies have been conductedto measure the impact of single-use instruments in practice, the results of these simulations motivatefurther investigation.

© 2019 Published by Elsevier Inc.

Total knee arthroplasty (TKA) is well established as a clinicallysuccessful and cost-effective procedure for alleviating pain,restoring function, and enabling a higher quality of life for patientswith advanced osteoarthritis [1,2]. These positive outcomes havedriven significant growth in the number of TKA procedures per-formed globally. In the United States alone, the number of pro-cedures is projected to increase from 693,400 cases in 2010 to 3.5million cases by 2030 [3,4]. At an average hospital cost of $16,000

closed potential or pertinentment, either direct or indirect,n the biomedical field whichrest with this work. For fullj.arth.2019.03.011.

per case, the annual cost of knee arthroplasties in the United Statesexceeded $11 billion in 2010 [3].

As a response to the escalating costs in the United States, thehealthcare system is evolving from a fee-for-service system to avalue-based system that incentivizes quality care delivered undercontrolled costs [5,6]. For total joint arthroplasty, the Centers forMedicare and Medicaid Services is currently trialing the Compre-hensive Care for Joint Replacement model, which holds hospitals

Funding: Medacta USA, Inc. provided funding for the preparation of the studymanuscript.* Reprint requests: Mukesh Ahuja, MBBS, MS, CPI, Department of Medical Affairs,

Medacta USA, Inc, Chicago, IL 60608.

T.D. Goldberg et al. / The Journal of Arthroplasty xxx (2019) 1e82

financially accountable for the quality and global cost of care frompatient admission through 90 days postdischarge [7]. With thesechanges in reimbursement, providers must evaluate all aspects ofTKA procedures to ensure financial viability while delivering high-quality care to this rapidly growing patient population. Manyhealthcare facilities and research studies are focused on controllingoperating costs, minimizing postdischarge complications, andoptimizing the utilization and scheduling of operating rooms (ORs)[8e12]. New techniques and technologies that foster greater effi-ciency, reduce unanticipated spending, and minimize patientcomplications will be critical to accommodate the continuinggrowth in healthcare utilization along with the associated resourcerequirements. For example, streamlining the surgical instrumen-tation is commonly emphasized for improving efficiency in the ORas well as hospital operations and would greatly support TKA in theoutpatient setting by reducing the need for handling, storage, andsterilization of instruments [8,9,13].

The utilization of single-use, sterile-packed instrumentation is agrowing strategy that could help control costs while improvingquality and efficiency. The risk of surgical site infections couldpotentially be reduced through single-use instrumentationcompared to traditional, reusable instruments. Multiple studieshave suggested that reusable instrumentation could becomecontaminated following resterilization, before use in surgery, andmay be associated with surgical site infections [14e18]. One studyobserved significantly more surgical site infections from TKA pro-cedures performed with traditional, reusable instrumentscompared to single-use instruments [19]. Furthermore, usingsingle-use instruments for TKA significantly reduces the number oftrays that need to be sterilized for each case, which can translate tosubstantial cost reductions and logistical improvements [19e21].The OR staff responsible for setup and cleanup of the instrumentscould also realize significant time savings for case turnover [19],which has been an important focus of many studies for improvinghospital efficiencies and OR utilization [9e11,22]. Finally, tradi-tional instrument sets for TKA are typically loaned to the facility by

Fig. 1. (A) Efficiency instruments and MyKnee blocks for a total knee arthrop

a manufacturer, and the logistics associated with managing theintake, processing, and transport can be a substantial burden[13,23].

The GMK Efficiency single-use instruments (Medacta Inter-national, Castel San Pietro, Switzerland) come terminally sterileand can streamline the instrumentation logistics and OR turnoverfor each TKA case, as all of the opened Efficiency instruments andtrays are disposed of at the end of the case [24]. One Efficiencyset includes all of the instruments for preparation, resection, andsizing of the tibia, femur, and patella as well as the trials and thesystem-specific gap spacers, alignment aids, and recut guides(Fig. 1). The general use instruments, such as retractors andosteotomes, are not included as single-use instruments, so atleast 1 tray of reusable instruments is still sterilized and used ineach case. Efficiency instruments can be used in combinationwith traditional or patient-specific cutting blocks, such asMyKnee Patient Matched Instrumentation (Medacta Interna-tional; Castel San Pietro, Switzerland). The economics and clinicaladvantages of patient-specific instrumentation have been inves-tigated by many recent studies, with a variety of conclusionsabout the potential benefits for different surgeons, facilities, andsystems [20,25e28]. In contrast, there have only been a couple ofsingle-site studies on traditional vs single-use instruments, whichboth observed significant advantages for the single-use in-struments [19,29].

There are many factors associated with TKA costs and logisticalefficiencies that can be highly variable between different facilities,surgeons, and cases [30e32]. Cost modeling through probabilisticsimulations can be an effective means of gaining insights into therange of potential savings and better understanding the key fac-tors that may drive variations in savings. This study aimed atinvestigating the range of potential cost savings for TKA proced-ures performed with GMK Efficiency single-use instrumentationcompared with traditional, reusable instruments through a prob-abilistic cost model that incorporates variance around multiplekey cost-drivers.

lasty (TKA) case vs (B) traditional instruments required for a TKA case.

Table 1Model Inputs and Literature Sources for Resource Variables.

Variable Model Sampling Level Instruments Normal DistributionModel Input (Mean ± SD)

References

90-d Infection rate (%) Site Traditional 0.58% ± 0.07% [12,32e34]Single-use ¼ 0.6 � TIIR Parameterb [19]

Trays used per case (n) Case Traditional 8 ± 2 [19,20,35e38]Single-use 1 ± 0 Parameterb [24]

OR turnover time (min) Case Traditional 38.4 ± 7.7 [9]Single-use Time savings

17.5 ± 3.5[19,29,37,38]

Tray management time (min) Case Traditional 355 ± 71 [13]Single-use 80 ± 16 [24]

TKA procedure time (min) Case Traditional 79 ± 15 [19,44]a

Single-use 79 ± 15

OR, operating room; SD, standard deviation; TIIR, traditional instrument infection rate; TKA, total knee arthroplasty.a Procedure time not affected by single-use instruments.b Parameters were also examined in a sensitivity analysis.

T.D. Goldberg et al. / The Journal of Arthroplasty xxx (2019) 1e8 3

Methods

Cost Modeling Approach

The Monte Carlo technique was used for this study to simulate awide variety of potential scenarios and outcomes related to per-forming a primary TKA procedure with traditional, reusable in-struments vs single-use instruments. Four variables were identifiedthat would potentially be affected by the type of instrumentation,which included sterilization costs, logistics for tray management(receiving, packing, storing, etc.), OR turnover time (instrumentsetup and cleanup), and treating surgical site infections during thefirst 90 days postdischarge. These variables were subdivided into aresource variable (eg, number of trays to sterilize per case; Table 1)and a unit cost variable (eg, cost of sterilizing 1 tray; Table 2). A totalof 200 sites (eg, hospitals) and 500 cases per site were simulated(100,000 total cases).

Briefly, the Monte Carlo simulation technique draws a randomvalue from the probability distribution of each variable for eachsimulation and repeats this process for thousands of simulationsto gain insight into the range of potential outcomes. In this study,some of the variables were sampled at the site level (eg, cost ofsterilizing 1 tray) so that they would be fixed across the 500cases performed at that site. The rest of the variables wereindividually sampled at the case level (eg, number of trays usedin each case). The value of each resource variable (eg, number oftrays to sterilize) was multiplied by its respective unit cost var-iable (eg, cost to sterilize each tray) to estimate the costs for eachcase (Fig. 2).

The probability distributions were constructed for each of theinput variables based on literature reports of measurements fromclinical studies or estimates from expert panels (Tables 1 and 2). Insome cases, variances were imputed due to a lack of available data.Details on the construction of each probability distribution aredescribed in Appendix A. Briefly, all cost variables were modeled aslognormal distributions [39], and the activity variables were

Table 2Model Inputs and Literature Sources for Unit Cost Variables.

Variable Model Sampling Level Median

Infection treatment ($/case) Case $30,020Tray sterilization ($/tray) Site $57.87OR turnover time ($/min) Site $5.72Tray management time ($/min) Site $1.03

IQR, interquartile range; OR, operating room.

assumed to follow normal distributions. Costs were adjusted forinflation to 2018 US dollars based on the USmedical care ConsumerPrice Index [40]. The OR turnover timewith traditional instrumentsand turnover time savings with single-use instruments wereassumed to be positively correlated (r ¼ 0.8), so multivariatesampling techniques were used for that part of the simulation.

Modeling the Effect of Time Savings During OR Turnover

Two approaches were used to examine the effect of reducing ORturnover time. In the primary model described above, the costsrelated to each of the variables were modeled based on the conceptof time-driven activity-based costing (TDABC). This accountingapproach measures the personnel time and consumables requiredfor each activity and multiplies those resource requirements by theunit costs (personnel wages, consumable costs, etc.). TDABC hasbeen suggested as a more conservative and accurate accountingapproach compared to traditional accounting methods for TKA[41,42].

The second modeling technique examined the effect of ORturnover time on underutilization and overutilization of the OR.Single-room OR blocking was assumed where a surgeon and thesupport teamwould be allotted the space for a full 8-hour, 10-hour,or 12-hour day. The mean surgical procedure time was assumed tobe equal for traditional and single-use instruments [19]. A simplescheduling technique was assumed where the cases were sched-uled back-to-back with 2 hours allotted for each case, based on amean turnover time of 38 minutes and mean procedure time of79 minutes (total ¼ 117 minutes; Table 1). For example, an 8-hourOR block had 4 cases scheduled each day. The Monte Carlo tech-nique was used to determine the length of each procedure andturnover time. If the scheduled cases for the day finished with atleast 2 hours to spare, an additional case could have beencompleted that day. If the scheduled cases took longer than theallotted time for the OR over the course of the full day, the extratime was considered as staff overtime.

(IQR) Lognormal Model Input(Location, Scale)

References

($19,104-$47,033) (10.31, 0.67) [12]($44.50-$77.23) (4.07, 0.41) [19,26,37,38]($4.46-$7.60) (1.75, 0.40) [30]($0.71-$1.52) (0.03, 0.56) [19]

Fig. 2. Schematic diagram of the probabilistic economic model for comparing the costs of traditional (Td) and Efficiency (Ef) instruments in TKA. A total of 200 sites (eg, hospitals)and 500 cases per site were simulated using Monte Carlo techniques. Four variables were sampled at the site-level (green) and the rest of the variables were sampled for eachindividual case (case-level; blue). The probability distributions were based on values from literature or imputed when data were not available. Costs were modeled using lognormaldistributions while the rest of the variables (except infection, which was binomial) were modeled with normal distributions. Operating room (OR) turnover time and turnover timesavings with Efficiency instruments were assumed to be positively correlated, so multivariate sampling techniques were used for that variable. Costs were calculated separately fortraditional and Efficiency instruments based on the unique distributions for each case-level variable.

T.D. Goldberg et al. / The Journal of Arthroplasty xxx (2019) 1e84

Parametric Sensitivity Analysis

For single-use instruments, the 90-day infection rate for a givensite was multiplied by a reduction factor of 0.6, which assumes thatsingle-use instruments will reduce the infection rate by 40% rela-tive to traditional instruments at a given site. With Efficiencysingle-use instruments, only 1 tray of instruments needs to besterilized to complete a TKA case [24]. Considering that surgeonsoften have specific preferences for customized instrument sets, anadditional 1 or 2 trays of reusable instruments could be required. Aparametric sensitivity analysis was carried out to examine the ef-fect of changing the infection rate reduction factor and the numberof trays sterilized with single-use instruments. The risk of infectionwith single-use instruments vs traditional instruments was variedfrom 60% (base case) to 40%, 80%, or 100% (equal infection risk). Thenumber of trays sterilized in conjunction with single-use in-struments was increased from 1 (base case) to 2 or 3.

Analysis of Model Outcomes

Themean costs per casewere calculated for each of the 200 sitesfor traditional and single-use instruments and the distribution ofcosts, or cost savings, was examined across the 200 sites. Costcomparisons between traditional and single-use instruments werecalculated using Wilcoxon’s matched-pairs signed-rank test bypairing costs within each site. The top and bottom 10% of sites,based on total cost savings with single-use instruments, werequeried to determine which variables significantly differed be-tween those sites (eg, do sites with savings in the 90th percentilehave greater OR costs per minute compared to sites in the 10th

percentile of savings?). Statistical comparisons between the topand bottom 10% of sites were made using 2-tailed t-tests with theHolm-Sidak method to correct the P values for multiple compari-sons. The parametric analysis parameters were analyzed throughrepeated measures (within each site) 2-factor analysis of variancethat also accounted for the difference between traditional andsingle-use instruments. Differences were deemed statistically sig-nificant for multiplicity-adjusted P values < .05.

Results

Cost Savings With Single-Use Instruments

The median and interquartile range (IQR) of total costs per casewas $1391 (IQR ¼ $1128-$1687) for traditional instruments and$406 (IQR ¼ $308-$515) for single-use instruments (P < .0001;Fig. 3). The dispersion of total costs observed with single-use in-struments was dramatically reduced compared to traditional in-struments, as demonstrated qualitatively by the frequencydistributions (Fig. 3) as well as the IQRs (single-use: $207 per casevs traditional: $559 per case). Within each site, these cost differ-ences between instrument types translated to a median total costsavings of $994 per case (IQR ¼ $759-$1231). The smallest meancost savings observed by a site was $4.50 per case and the largestwas $1829 per case (Fig. 4).

Cost-Drivers for Traditional vs Single-Use Instruments

For traditional instruments, the largest cost-driver was traysterilization (median ¼ $480/case), followed by tray management

10030

050

070

090

011

0013

0015

0017

0019

0021

000

5

10

15

20

25

30

Mean Cost per Case ($)

Prop

ortio

nof

Site

s(%

)Traditional

Efficiency

Fig. 3. Frequency distribution of the mean total costs per case for traditional, reusableinstruments vs single-use Efficiency instruments across the 200 simulated sites.

800

600

400

200 0

200

400

600

800

Infections

OR Turnover

Tray Management

Tray Sterilization

Mean Cost per Case ($)

TraditionalEfficiency

Fig. 5. On average, tray sterilization was the greatest cost-driver for traditional in-struments, followed by tray management time, OR turnover time, and treating in-fections within the first 90 days postdischarge. For Efficiency instruments, treatment ofinfections was the largest cost-driver on average due to the significant cost reductionsachieved for tray sterilization, tray management, and OR turnover.

T.D. Goldberg et al. / The Journal of Arthroplasty xxx (2019) 1e8 5

logistics ($372/case), OR turnover time ($225/case), and 90-dayinfections ($199/case). For single-use instruments, the greatestcost-driver was OR turnover time (median ¼ $123/case), followedby 90-day infections ($101/case), tray management logistics($85/case), and sterilization ($60/case; Fig. 5). Cost distributions foreach of the metrics tended to have a smaller dispersion for single-use instruments compared to traditional instruments (Fig. S1).

Site-Dependent Factors Affecting Cost Savings With Single-UseInstruments

On average, the sites that reduced costs themost by switching tosingle-use instruments had significantly higher sterilization costsand tray management costs compared to the sites saving the least(Table 3). There was an average of 2.5 more infection cases withtraditional instruments among the �90th percentile of cost savingsites and an average of 0.7 more infection cases with single-use

025

050

075

010

0012

5015

0017

5020

000

3

6

9

12

15

Mean Cost Savings per Case ($)(Traditional – Efficiency)

Prop

ortio

nof

Site

s(%

)

Cost Savings with Efficiency Instruments

75th25th 50th10th 90thPercentiles:

Fig. 4. Frequency distribution of cost savings by site (paired analysis). Approximately75% of sites saved at least $750 per case, 50% saved at least $1000 per case, and 25%saved over $1200 per case. The minimum savings was $4.50 and the maximum savingswas $1829.

instruments in the �10th percentile of sites. This observation isactually a random effect attributable to the probabilistic nature ofthese simulations, considering that the overall infection ratesacross the sites were similar (�10th percentile: 0.59% vs �90thpercentile: 0.58%). The OR time cost, the OR turnover time, theaverage number of trays used per case, and the tray managementtime were not significantly different between the sites saving themost and the least (Table 3).

Underutilization and Overutilization of the OR

Assuming an 8-hour OR day, where 4 cases were scheduled perday, an extra case could have been completed (2þ hours ofunderutilized OR time) on 11% of the operating days when usingsingle-use instruments compared to 0.1% of operating days whenusing traditional instruments. The proportion of days that couldaccommodate an additional case increased to 29% and 51% for 10-hour and 12-hour OR days with single-use instruments comparedto 0.3% and 0.6%, respectively, for traditional instruments (Table 4).

Similarly, the proportion of operating days where the OR wasoverutilized, resulting in staff overtime pay, was 35%-38% fortraditional instruments and 0.1%-0.7% for single-use instrumentsacross 8-hour, 10-hour, or 12-hour days. This translated to overtimepay costs averaging up to $1326 ± $568 per 100 cases for traditionalinstruments and $11.00 ± $19.00 per 100 cases for single-use in-struments (Table 4).

Parametric Sensitivity Analysis

Varying the relative risk of infection for single-use vs traditionalinstruments from 0.4 up to 1.0 (equal infection risk), with 0.6 as thebase case, did not significantly affect the total costs of single-useinstruments vs traditional instruments (P ¼ .14; Fig. S2A). Evenwhen equal infection risk was assumed, the mean total costs oftraditional vs single-use instruments was $1412 ± $386 vs $509 ±$170, respectively (P < .0001). Varying the number of trays thatneeded to be sterilized with the use of single-use instruments from1 (base case) to 3 also did not significantly affect the total costs (P ¼.31; Fig. S2B). The mean total cost with single-use instrumentsranged from $424 ± $153 (1 tray) to $522 ± $160 (3 trays; P ¼ .31)compared to $1427 ± $374 for traditional instruments (P < .0001).

Table 3Site Metrics for the Top and Bottom 10% of Cost Savings With Single-Use Instruments.

Site Metric �10th Percentile of Cost Savings �90th Percentile of Cost Savings Comparison (�90th vs �10thPercentile)

Mean ± SD Mean ± SD Mean Difference P Value

Cost savings ($/case) $456 ± $135 $1600 ± $118 $1144 <.0001Site variablesSite infection rate (%) 0.59% ± 0.08% 0.58% ± 0.06% �0.01% .98Site sterilization cost ($/tray) $48 ± $19 $92 ± $43 $44 .001a

Site OR turnover cost ($/min) $5.23 ± $2.20 $7.01 ± $3.18 $1.79 .25Site tray management cost ($/min) $0.67 ± $0.24 $1.95 ± $1.02 $1.28 <.0001a

D Counts and costs (traditional � single-use)Number of infections (n/site) �0.7 ± 1.5 2.5 ± 2.2 3.2 <.0001a

Tray count (n/case) 7.0 ± 0.1 7.0 ± 0.1 0.0 .84OR turnover time (min/case) 17.5 ± 0.1 17.5 ± 0.1 0.0 .98Tray management time (min/case) 275 ± 6 275.3 ± 7.1 0.4 .98Infection cost ($/case) �$159 ± $190 $295 ± $214 $454 <.0001a

Sterilization cost ($/case) $338 ± $132 $647 ± $300 $309 .001a

OR turnover cost ($/case) $91 ± $39 $123 ± $56 $31 .25Tray management cost ($/case) $185 ± $68 $537 ± $274 $351 <.0001a

OR, operating room; SD, standard deviation.a Indicates P < .05 based on 2-tailed t-tests. Holm-Sidak method was used to correct for multiple comparisons.

T.D. Goldberg et al. / The Journal of Arthroplasty xxx (2019) 1e86

Discussion

Improving operational efficiencies both inside and outside of theOR and reducing the dispersion of costs are critical initiatives formany healthcare providers. Single-use instrumentation has thepotential to facilitate time and cost efficiencies while bringinggreater reliability and predictability to multiple steps in the globaldelivery of TKA compared to traditional, reusable instruments. Bymodeling the keycost-drivers that could be affected by single-use vsreusable instruments, this study observed that the majority of sitescould potentially realize substantial cost savings, reduce opportu-nity costs of OR underutilization, and bring improved predictabilityto budgeting through less variance in the operating costs.

In a single-site clinical study, Siegel et al [19] reported costsavings of $480-$600 with single-use instruments vs traditionalinstruments for TKA, based on measurements of the resource re-quirements and costs associated with OR turnover and tray steril-ization. Furthermore, they reported significantly fewer infectionswith single-use instruments (0.2% vs 3%; P ¼ .006). In closeagreement with those results, the present study estimated a me-dian cost savings of $522 based on OR turnover and tray steriliza-tion alone. Multiple studies have emphasized the importance ofcase turnover in the OR for optimizing productivity and achievingan increase in number of cases. Through interdisciplinary work-flows, Cendan and Good [43] reported reductions in turnover time

Table 4Summary of Results on Underutilization and Overutilization of OR Time.

Length of OR Day Single-Use Instruments Traditional Instruments

Proportion of operating days when an extra case could have been completed(mean ± SD)8 h 10.8% ± 2.7% 0.1% ± 0.2%10 h 29.3% ± 4.5% 0.3% ± 0.5%12 h 51.0% ± 5.1% 0.6% ± 0.8%

Proportion of operating days requiring staff overtime (mean ± SD)8 h 0.7% ± 0.7% 37.8% ± 4.1%10 h 0.2% ± 0.5% 36.8% ± 4.6%12 h 0.1% ± 0.4% 35.0% ± 5.6%

Average overtime pay per 100 cases (mean ± SD)8 h $11.00 ± $19.00 $1326 ± $56810 h $3.00 ± $9.00 $1173 ± $62812 h $1.00 ± $5.00 $967 ± $449

OR, operating room; SD, standard deviation.

of 16 minutes, which translated to an increase of approximately 0.5cases per day on average. This is very consistent with the presentstudy that observed an additional case could be scheduled on up to51% of operating days when using single-use instruments.

Tray sterilization and tray management were not only thelargest contributors to traditional instrument costs in this study butwere also the primary areas for cost savings. Furthermore, thesesteps tend to be a significant impediment for efficient, multi-casesurgical days, particularly in outpatient settings such as ambula-tory surgical centers. Studies of single-use instrumentation andpatient-specific instrumentation consistently conclude that signif-icantly fewer trays need to be sterilized compared to traditional,reusable instruments (range, 4-12) [19,20,35e38]. However, thecosts reported in the literature of sterilizing 1 tray are highly var-iable (range, $31-$100/tray) [19,26,37,38], which can result fromoperational and staff wage differences between sites and the ac-counting methods for estimating those costs. Consistent with thisconcept, Haas and Kaplan demonstrated that the total standardizedpersonnel costs associated with TKA can vary from 68% of themedian at the 10th percentile to 131% of the median at the 90thpercentile across sites [30]. Based on the results of the presentstudy, the large variations in tray sterilization costs as well as thewages of hospital staff could substantially impact the potential forcost savings with single-use instruments. The sites that saved themost had tray sterilization costs nearly twice that of sites saving theleast (�90th percentile mean: $92/tray vs �10th percentile mean:$48/tray). The aggregated wages of staff responsible for tray man-agement also had a significant effect, with the sites in the �90thpercentile of cost savings paying staff an average of $117/h vs only$40.20/h among the sites in the �10th percentile.

An important limitation to consider, which was not accountedfor in the model, is the probability that necessary instruments orcomponents will not be available at the time of the surgery. Forreusable instruments, a set may not have been repacked properlybefore sterilization. This would cause delays in the case to acquirean additional sterilized set with the missing instruments or theuse of flash (immediate-use) steam sterilization. It is also possiblethat a reusable or single-use instrument could be accidentallycontaminated beyond the sterile field, requiring a sterile backup orreplacement, but the probability of such an event would likely beequal for reusable and single-use instruments.

This study focused on a specific single-use instrument set thatthe authors were familiar with; however, these findings are likely

T.D. Goldberg et al. / The Journal of Arthroplasty xxx (2019) 1e8 7

to be generalizable to other single-use instrument sets with asimilar design (eg, included instruments). Additionally, the costsestimated in the present study were not intended to compre-hensively represent the global cost of a TKA procedure. Rather thisstudy focused on cost components that single-use instrumentscould affect. Based on TDABC studies, the cost of a global TKAprocedure ranges from $9800 to $14,000 [41,42], so the costsstudied herein account for approximately 10%-20% of the total TKAcosts. The price of the single-use instruments was also notincluded in the model. Instead, the potential cost savings could beused by healthcare providers or manufacturers to estimate break-even pricing. Siegel et al [19] reported a quote of $490 for onesingle-use instrument system. In this study, 95% of the simulatedsites realized cost savings exceeding $500 per case. Consideringthat cost savings were primarily driven by reductions in traysterilization and logistical management costs, these variablesshould be accurately understood for the most reliable modelingresults. The number of trays used along with tray sterilizationcosts were well reported in the literature but were quite variable.In contrast, estimates for the logistical management of traditionalinstrument loaner trays and the associated costs were not wellmeasured or reported in the literature. The limitations of theliterature data pose inherent limitations for any modeling study.To further enhance confidence in model outcomes, future studiesthat use a shadowing technique for TDABC would be important tomore accurately estimate the resource requirements for loaner setmanagement and elucidate the sources for variability in traysterilization [42]. Based on the results of this economic model,clinical and economic studies at a variety of sites (large-volume vssmall-volume hospitals vs ambulatory surgery centers) could bedesigned to measure the resource savings and validate or updatethe model studied herein.

Conclusions

This study used probabilistic modeling techniques to examinethe range of potential logistical and economic advantages thathealthcare providers may realize by using single-use instrumentsfor TKA procedures compared to traditional, reusable instruments.Of the 200 simulated sites, 95% saved at least $500 per case and48% saved at least $1000 per case. Sites with higher staff wagesand sterilization costs had a greater probability of realizing sub-stantial cost savings with the adoption of single-use instrumentscompared to traditional reusable instruments. Single-use in-struments also dramatically reduced overtime in the OR andenabled the opportunity for a substantial increase in the numberof cases. In conclusion, this modeling study suggests that single-use instruments have a compelling potential to help improve thequality and efficiency of delivering TKA procedures, warrantingfuture prospective studies to measure the actual resource and costsavings observed in practice.

References

[1] Ponnusamy KE, Sumerville L, McCalden RW, Marsh JD. Cost-effectiveness oftotal knee arthroplasty vs nonoperative management in normal, overweight,obese, severely obese, morbidly obese, and super-obese patients: a Markovmodel. J Arthroplasty 2018;33:S32e8.

[2] Kamaruzaman H, Kinghorn P, Oppong R. Cost-effectiveness of surgical in-terventions for the management of osteoarthritis: a systematic review of theliterature. BMC Musculoskelet Disord 2017;18:183.

[3] Williams S, Wolford M, Bercovitz A. Hospitalization for total knee replacementamong inpatients aged 45 and over: United States, 2000e2010. NCHS databrief, no 210. Hyattsville, MD: National Center for Health Statistics; 2015.https://www.cdc.gov/nchs/data/databriefs/db210.pdf [Accessed 23.05.18].

[4] Kurtz S, Ong K, Lau E, Mowat F, Halpern M. Projections of primary and revisionhip and knee arthroplasty in the United States from 2005 to 2030. J Bone JointSurg Am 2007;89:780e5.

[5] Piccinin MA, Sayeed Z, Kozlowski R, Bobba V, Knesek D, Frush T. Bundlepayment for musculoskeletal care: current evidence (Part 1). Orthop ClinNorth Am 2018;49:135e46.

[6] Piccinin MA, Sayeed Z, Kozlowski R, Bobba V, Knesek D, Frush T. Bundlepayment for musculoskeletal care: current evidence (Part 2). Orthop ClinNorth Am 2018;49:147e56.

[7] (CMS), C.f.M.a.M.S. Comprehensive care for joint replacement model. https://innovation.cms.gov/initiatives/cjr; 2018 [Accessed 24.05.18].

[8] Bert JM, Hooper J, Moen S. Outpatient total joint arthroplasty. Curr RevMusculoskelet Med 2017;10:567e74.

[9] Ngu JC. Improving OR efficiency in a university medical center arthroplasticsurgery service. AORN J 2010;92:425e35.

[10] Bender JS, Nicolescuet TO, Hollingsworth SB, Murer K, Wallace KR, Ertl WJ.Improving operating room efficiency via an interprofessional approach. Am JSurg 2015;209:447e50.

[11] Bhatt AS, Carlson GW, Deckers PJ. Improving operating room turnover time: asystems based approach. J Med Syst 2014;38:148.

[12] Cram P, Lu X, Li Y. Bundled payments for elective primary total kneearthroplasty: an analysis of Medicare administrative data. Geriatr Orthop SurgRehabil 2015;6:3e10.

[13] Guedon AC, Wauben LS, van der Eijk AC, Vernooji AS, Meeuwsen FC, van derElst M, et al. Where are my instruments? Hazards in delivery of surgical in-struments. Surg Endosc 2016;30:2728e35.

[14] Dancer SJ, Stewart M, Coulombe C, Gregori A, Virdi M. Surgical site in-fections linked to contaminated surgical instruments. J Hosp Infect2012;81:231e8.

[15] Al-Jandan BA, Ahmed MG, Al-Khalifia KS, Farooq I. Should surgical burs beused as single-use devices to avoid cross infection? A case-control study. MedPrinc Pract 2016;25:159e62.

[16] Waked WR, Simpson AK, Miller CP, Magit DP, Grauer JN. Sterilization wrapinspections do not adequately evaluate instrument sterility. Clin Orthop RelatRes 2007;462:207e11.

[17] Mont MA, Johnson AJ, Issa K, Pivec R, Blasser KE, McQueen D. Single-useinstrumentation, cutting blocks, and trials decrease contamination duringtotal knee arthroplasty: a prospective comparison of navigated and non-navigated cases. J Knee Surg 2013;26:285e90.

[18] Mobley KS, Jackson 3rd JB. A prospective analysis of clinical detection ofdefective wrapping by operating room staff. Am J Infect Control 2018;46:837e9.

[19] Siegel GW, Patel NN, Milshteyn MA, Buzas D, Lombardo DJ, Morawa LG. Costanalysis and surgical site infection rates in total knee arthroplastycomparing traditional vs. single-use instrumentation. J Arthroplasty2015;30:2271e4.

[20] Noble Jr JW, Moore CA, Liu N. The value of patient-matched instrumentationin total knee arthroplasty. J Arthroplasty 2012;27:153e5.

[21] Rodrigues AST, Gutierres MAP. Patient-specific instrumentation in total kneearthroplasty. Should we adopt it? Rev Bras Ortop 2017;52:242e50.

[22] Dexter F, Abouleish AE, Epstein RH, Whitten CW, Lubarsky DA. Use of oper-ating room information system data to predict the impact of reducing turn-over times on staffing costs. Anesth Analgesia 2003;97:1119e26.

[23] Stockert EW, Langerman A. Assessing the magnitude and costs of intra-operative inefficiencies attributable to surgical instrument trays. J Am CollSurg 2014;219:646e55.

[24] Goldberg T, Seaveyet R, Kuse K, Domyahn M, Torres A. Value in single useinstruments for total knee arthroplasty: patient outcomes and operating roomefficiency. White paper; 2017. https://media.medacta.com/media/drgoldberg-gmkefficiency-whitepaper.pdf.

[25] Sassoon A, Nam D, Nunley R, Barrack R. Systematic review of patient-specificinstrumentation in total knee arthroplasty: new but not improved. ClinOrthop Relat Res 2015;473:151e8.

[26] Barrack RL, Ruh EL, Williams BM, Ford AP, Foreman K, Nunley RM. Patientspecific cutting blocks are currently of no proven value. J Bone Joint Surg Br2012;94(11 Suppl A):95e9.

[27] Nunley RM, Ellison BS, Ruh EL, Williams BM, Foreman K, Ford AD. Are patient-specific cutting blocks cost-effective for total knee arthroplasty? Clin OrthopRelat Res 2012;470:889e94.

[28] McLawhorn AS, Carroll KM, Blevins JL, DeNegre ST, Mayman DJ, Jerabek SA.Template-directed instrumentation reduces cost and improves efficiency fortotal knee arthroplasty: an economic decision analysis and pilot study.J Arthroplasty 2015;30:1699e704.

[29] Mont MA, McElroy MJ, Johnson AJ, Pivec R. Single-use instruments, cuttingblocks, and trials increase efficiency in the operating room during total kneearthroplasty: a prospective comparison of navigated and non-navigated cases.J Arthroplasty 2013;28:1135e40.

[30] Haas DA, Kaplan RS. Variation in the cost of care for primary total kneearthroplasties. Arthroplast Today 2017;3:33e7.

[31] Kadhim M, Gans I, Baldwin K, Flynn J, Ganley T. Do surgical times and effi-ciency differ between inpatient and ambulatory surgery centers that are bothhospital owned? J Pediatr Orthop 2016;36:423e8.

[32] Cram P, Ravi B, Vaughan MS, Lu X, Li Y, Hawker G. What drives variation inepisode-of-care payments for primary TKA? An analysis of Medicare admin-istrative data. Clin Orthop Relat Res 2015;473:3337e47.

[33] Nichols CI, Vose JG. Clinical outcomes and costs within 90 days ofprimary or revision total joint arthroplasty. J Arthroplasty 2016;31:1400e1406.e3.

T.D. Goldberg et al. / The Journal of Arthroplasty xxx (2019) 1e88

[34] Middleton A, Lin YL, Graham JE, Ottenbacher KJ. Outcomes over 90-day epi-sodes of care in Medicare fee-for-service beneficiaries receiving jointarthroplasty. J Arthroplasty 2017;32:2639e2647.e1.

[35] Hamilton WG, Parks NL, Saxena A. Patient-specific instrumentation does notshorten surgical time: a prospective, randomized trial. J Arthroplasty2013;28(8 Suppl):96e100.

[36] Renson L, Poilvache P, Van den Wyngaert H. Improved alignment and oper-ating room efficiency with patient-specific instrumentation for TKA. Knee2014;21:1216e20.

[37] Watters TS, Mather RC, Browne JA, Berend KR, Lombardi Jr AV, Bolognesi MP.Analysis of procedure-related costs and proposed benefits of using patient-specific approach in total knee arthroplasty. J Surg Orthop Adv 2011;20:112e6.

[38] Tibesku CO, Hofer P, Portegies W, Ruys CJ, Fennema P. Benefits of usingcustomized instrumentation in total knee arthroplasty: results from anactivity-based costing model. Arch Orthop Trauma Surg 2013;133:405e11.

[39] Canes A. How is healthcare cost data distributed? Using proc univariate to drawconclusions aboutmillions of different customers. Bethesda, MD: SESUG; 2016.

[40] CPI Latest Numbers. U.S. Bureau of labor statistics, division of consumer pricesand price indexes. 2018. https://www.bls.gov/cpi/data.htm. [accessed 25.05.18].

[41] Akhavan S,Ward L, Bozic KJ. Time-driven activity-based costingmore accuratelyreflects costs in arthroplasty surgery. Clin Orthop Relat Res 2016;474:8e15.

[42] DiGioia 3rd AM, Greenhouse PK, Giarrusso ML, Kress JM. Determining the truecost to deliver total hip and knee arthroplasty over the full cycle of care:preparing for bundling and reference-based pricing. J Arthroplasty 2016;31:1e6.

[43] Cendan JC, Good M. Interdisciplinary work flow assessment and redesigndecreases operating room turnover time and allows for additional caseload.Arch Surg 2006;141:65e9. discussion 70.

[44] Gadinsky NE, Manuel JB, Lyman S, Westrich GH. Increased operating roomtime in patients with obesity during primary total knee arthroplasty: conflictsfor scheduling. J Arthroplasty 2012;27:1171e6.

T.D. Goldberg et al. / The Journal of Arthroplasty xxx (2019) 1e8 8.e1

Appendix A. Details of Model Inputs and ProbabilityDistributions

Probability Distributions for Each Activity Variable

A normal distributionwas assumed for all activity variables. The90-day infection rate was selected from the distribution and heldconstant for each case performed at a given site. The distribution ofinfection rates was based on a mean and standard deviationcalculated from the reported rates for traditional instruments[12,32e34]. For single-use instruments, the 90-day infection ratefor a given site was multiplied by a reduction factor of 0.4, 0.6, or0.8. This technique assumes that single-use instruments will reducethe infection rate to 40%, 60%, or 80% of the current rate at a givensite. The concept of this reduction is based on the report of Siegelet al [19], who reported a reduction in infection rate from 3% withtraditional instruments to 0.2% with single-use instruments.Furthermore, many other studies support the concept that reusableinstruments can become contaminated after sterilization and in-crease the risk of surgical site infection [14e18]. Therefore, areduction factor of 60% was used as a conservative base case and aparametric sensitivity analysis examined the effect of changing thatreduction factor to 40% or 80%. The infection rate for each site wasused to sample from a binomial distribution to determine whethereach specific case acquired an infection.

The reported mean or median number of traditional instrumenttrays used per case was used to calculate a mean and standard de-viation of trays used [19,20,35e38]. The number of trays selectedfrom the distribution during the Monte Carlo simulation wasrounded to the nearest integer. The manufacturer’s white papersuggests that only 1 tray needs to be sterilized for each TKA caseusing single-use instruments [24]. Some surgeons have specific in-struments that they prefer, which could increase the number of traysrequiring sterilization even with single-use instruments. Therefore,the number of trays sterilized with single-use instruments wasstudied as a sensitivity analysis parameter using either 1, 2, or 3 trays.

The mean monthly OR turnover time data reported for 3 sur-geons over 2 years was used to calculate the mean and standarddeviation [9]. For single-use instruments, the time saved was mostoften reported rather than total turnover time. The reported timesavings from 4 articles was used to calculate a mean and thestandard deviation was imputed with a 20% coefficient of variation[19,29,37,38]. The amount of turnover time with traditional in-struments and the potential for time savings with single-use in-struments were assumed to be positively correlated with a

coefficient of 0.8 (ie, cases with larger turnover times have greaterpotential for larger time savings). Therefore, a multivariate normalrandom number tool was used to draw values from the correlateddistributions simultaneously during the Monte Carlo simulations.

The tray management time involves everything outside of theOR and central sterilization department. Guedon et al [13] esti-mated 355 minutes dedicated to set management per case. Novariance was provided, so a 20% coefficient of variation wasimputed for a standard deviation of 71 minutes. The average timeper tray was calculated by dividing 355 minutes by the meannumber of traditional trays used (8) and the standard deviationwascalculated through propagation of error. The manufacturer’s whitepaper on Efficiency single-use instruments estimated 1.3 hours(80 minutes) for tray management time [24]. A 20% coefficient ofvariation was imputed for a standard deviation of 16 minutes.

The TKA procedure time (incision to closure) was assumed tofollow the same distribution for traditional and single-use in-struments based on the findings of Siegel et al [19]. The average andstandard deviation of procedure times were based on all patientsreported by Gadinsky et al [44].

Probability Distributions for Each Cost Variable

All cost variables were modeled as lognormal distributions [39].The location and scale metrics of the lognormal distributions wereselected by iterative optimization to fit the median and inter-quartile range (IQR) values derived from the literature. Costs wereadjusted for inflation to 2018 US dollars based on the US medicalcare Consumer Price Index [40]. The median and IQR for treatmentof an infection were estimated from the data published by Cramet al [12] as the difference in postdischarge costs for patients withan infection vs patients with no complications. Themean ormediancost of sterilizing 1 tray reported by 4 articles was used to calculatea median cost and IQR [19,26,37,38].

For the cost of OR turnover time, Haas et al [30] cited that the ORcost per minute without the surgeon is at least $5 per minute.Therefore, $5 per minute was used as a conservative median value.The IQR data reported by Haas et al for standardized personnelcosts during surgery was used to estimate the 25th and 75th per-centiles [30]. For tray management personnel costs, Siegel et alreported $0.84 per minute for central sterilization departmentpersonnel [19]. This cost was used as the median and the 25th and75th percentile proportions reported by Haas et al for “Standard-ized personnel costs day of surgery, before surgery” were used toestimate the IQR [30].

010

020

030

040

050

060

070

00

5

10

15

20

25

Mean Cost of Infection per Case ($)

Rel

ativ

efre

quen

cy(%

)

Cost of Infection

0 50 100

150

200

250

300

350

400

450

0

5

10

15

20

25

Mean Cost of OR Turnover Time per Case ($)

Prop

ortio

nof

Site

s(%

)

Cost of OR Turnover Time

010

020

030

040

050

060

070

080

090

00

20

40

60

80

Mean Cost of Tray Sterilization per Case ($)

Prop

ortio

nof

Site

s(%

)

Cost of Tray Sterilization

010

020

030

040

050

060

070

080

090

010

000

10

20

30

40

50

Mean Cost of Tray Management per Case ($)

Prop

ortio

nof

Site

s(%

)

Cost of Tray Management

Traditional

Efficiency

Traditional

Efficiency

Traditional

Efficiency

Traditional

Efficiency

Fig S1. Frequency distributions of costs related to each of the 4 modeled variables across all 200 simulated sites for traditional and Efficiency instruments. In addition to having ahigher probability of lower costs for Efficiency instruments, the dispersion of costs tends to be much smaller.

40% 60% 80% EqualRisk

0

500

1000

1500

2000

Infection Risk ofSingle-Use InstrumentsRelative to Traditional

Mea

nTo

talC

osts

($)

Infection Risk

1 2 30

500

1000

1500

2000

Number of Trays toSterilize with

Efficiency Instruments

Mea

nTo

talC

ost s

($)

Number of Single-use Trays

Traditional

Single-use

Fig S2. Varying the infection risk reduction factor from 40% to 100% (equal risk) did not significantly affect the total costs or cost savings of using Efficiency instruments (P ¼ .14). Theaverage total cost with Efficiency instruments ranged from $384 ± $144 (at 40%) to $509 ± $170 (at equal risk). The number of trays that needed to be sterilized with the use ofEfficiency instruments also did not significantly affect the total costs (P ¼ .31). The average total cost with Efficiency instruments ranged from $424 ± $153 (1 tray) to $522 ± $160 (3trays). The effect of these parameters was analyzed through repeated measures (within each site) 2-factor analysis of variance that also accounted for the difference betweentraditional and Efficiency instruments, which was statistically significant in each case (P < .0001).

T.D. Goldberg et al. / The Journal of Arthroplasty xxx (2019) 1e88.e2