Analysis of inpatient costs in patients with knee osteoarthritis treated by implantation of total condylar knee endoprosthesis

Background/Aim. Total knee replacement is an elective and high cost surgical procedure which is performed more frequently as a result of increasing prevalence of knee osteoarthritis. The aim of this study was to provide insight into the value and structure of inpatient costs associated with total knee replacement in Serbia. Methods. This study was conducted as an in-depth, bottom-up, retrospective, case series analysis of services consumption patterns and costs associated with inpatient treatment of patients with knee osteoarthritis by implantation of primary total condylar knee endoprosthesis from perspective of the national Republic Health Insurance Fund. We obtained data on 97 patients treated with primary unilateral or bilateral total knee replacement in 2014 at the Clinic for Orthopaedic Surgery and Traumatology of the Military Medical Academy in Belgrade, a tertiary health care university hospital. Results. Mean age of entire study sample was 67.89 years. Majority of patients (60 patients; 61.9%) had unilateral implantation of total condylar knee endoprosthesis. Bilateral implantation was performed in 37 (38.1%) patients. Mean total inpatient cost per patient for both unilateral and bilateral implantation of total condylar knee endoprosthesis was EUR 2,709.1, ranging from EUR 1,685.2 to EUR 5,356.6. Mean total inpatient cost per patient was EUR 2,093.8 for unilateral implantation and EUR 3,706.8 for bilateral implantation. Two major cost drivers were surgery specific material and surgery. Cost of implants was the highest single cost driver in all observed groups of patients. Conclusion. Our findings imply that inpatient costs associated with implantation of primary total condylar knee endoprosthesis are substantial. It seems that the most important cost drivers are surgery and surgery specific material, with implants being the highest single cost driver. Further research should be focused on analyzing factors that influence these costs in order to develop effective strategies which could contribute to substantial savings in the future.


Introduction
Assessment of economic implications of orthopaedic surgeries is gaining more attention as musculoskeletal disorders demanding such procedures are on the rise 1 .Total knee replacement (also referred to as total knee arthroplasty) is an elective, expensive, but cost-effective surgical procedure which is being performed more frequently as a result of increasing prevalence of knee osteoarthritis (gonarthrosis) [2][3][4][5][6][7] .Osteoarthritis is a degenerative disease of cartilage and surrounding tissue associated with joint pain, stiffness and limitation of movement which affects about 10% of persons over the age of 60 years [8][9][10][11] .Knee replacement involving implantation of total condylar endoprosthesis is indicated when conservative treatment can not reduce knee joint pain and dysfunction and is associated with a substantial improvement in quality of life and pain relief in patients with knee osteoarthritis 5,6,12 .This procedure involves altering the articular surfaces in a way that enables replacement of total damaged knee joint with a prosthetic implant 4,5 .Total condylar knee endoprosthesis resurfaces all three compartments of the knee (lateral, medial, and patellofemoral) 13 .
The mean utilization rate of knee replacement in Organisation for Economic Co-operation and Development (OECD) countries was 150 procedures per 100,000 people in 2011 14 .The number of performed total knee replacements has been increasing worldwide as rise in the prevalence of osteoarthritis follows the pace of acceleration of global population ageing and rising trends of obesity 12,[15][16][17][18][19][20][21][22][23][24] .The recently published Global Burden of Disease, Injuries and Risk Factors Study noted that disability-adjusted-life-years (DALYs) attributable to high body mass index (BMI) increased the most among the top five risk factors from 1990 to 2015 22 .The demand for primary total knee replacements is projected to grow to 3.48 million procedures annually in the United States by 2030 20 .However, accessibility to this procedure is inconsistent across the world and is affected by high cost and limited number of skilled personnel who can perform it 25 .Patients seeking care in publicly-funded institutions frequently spend weeks or months on the waiting list for provision of this surgery unless they are able to pay for private care 26 .For example, the mean length of time spent on waiting for this procedure within the Veterans Affairs Connecticut Healthcare System in the United States can be as long as two years 27 .These all have negative impact on patients as they experience great pain and suffer functional limitations while they await surgery 4 .
The costs associated with total knee replacement are becoming increasing concern worldwide because they put a significant financial burden on most healthcare systems 12,28 .This procedure was associated with one of the most noticeable increase for inpatient costs among all payer types in the United States 29 .Aggregate inpatient costs of total knee replacement reached United States dolar (USD) 9.2 billion in 2007, and they grew by 27.5% between 2004 and 2007 30 .Previous studies reported that cost of implants, hospital room and operating room segment of care may account over 75% of inpatient costs [31][32][33][34] .However, the cost of implants is usually the highest and it can even reach up to 87% of overall inpatient cost 6,29,35 38 .Such variations in cost of care are one of the main reasons for increasing number of patients from developed countries who travel to hospitals in emerging market countries like Taiwan, Thailand, India and Singapore where cost of this procedure can be 8 to 10 times less expensive than in the United States due to low labor and maintenance cost [39][40][41][42] .
Nearly 20,000 patients were on the waiting lists for knee and hip replacement in Serbia in 2014 with reported increase of 20% compared to previous year 43 .Average length of time that patients spent waiting for these procedure was 311 days in 2014, which was 25 days less than was reported in 2013 43 .It is likely to expect that demand for these procedures will continue to rise in Serbia with continuing population ageing 44 .The core fund in Serbia in charge for most inpatient care expenditures is Republic Health Insurance Fund which is a non-profit state owned institution subject to budget shortages [44][45][46] .Consequently, there is a need to assess major inpatient cost drivers in order to make more efficient health policy programs.Since health expenditures related to total knee replacement are substantial, it is crucial to understand cost of care provided across various settings in order to provide baseline data for pharmacoeconomic analyses in the future.
So far, there is a substantial knowledge gap on actual inpatient cost of total knee replacement in Serbia.Therefore, the aim of this study was to provide insight into the value and structure of inpatient costs associated with total knee replacement by implantation of total condylar knee endoprosthesis in Serbia.

Study design and patient selection
This study was conducted as an in-depth, bottom-up, retrospective, case series analysis of services consumption patterns and costs associated with inpatient treatment of patients with knee osteoarthritis by implantation of primary total condylar knee endoprosthesis from the perspective of the third party payer, i.e., from the national Republic Health Insurance Fund.Indirect cost and out-of-pocket patient's expenditure, as well as costs in settings other than inpatient, remained out of scope of this study.We obtained data on patients who were treated with primary unilateral or bilateral total knee replacement in 2014 at the Clinic for Orthopaedic Surgery and Traumatology of the Military Medical Academy in Belgrade, a tertiary health care university hospital.The source of data was an anonymised database consisting of electronic hospital discharge invoices.In total, 97 complete patient files were analysed.Data on age, gender and length of hospitalisation were also collected.

Structure and pricing of the used recourses
The official Republic Health Insurance Fund pricelist was applied at the time of the service provision.Average middle exchange rate for Euro (EUR) given by the National Bank of Serbia for 2014 was used to convert costs originally reported in the national currency Serbian Dinar (RSD): EUR 1 = RSD 117.2478 47 .
For the present study, total inpatient costs associated with the implantation of primary total condylar knee endoprosthesis were collected.Costs were separated into the following categories: general surgery related medical care (hospital admission day and consumables, rehabilitation services, and all other services such as social care, transport, counseling, epidemiological measures), surgery (surgical intervention and anesthesia), imaging diagnostics [classical imaging diagnostics -Roentgen, contrasts, films and consumables intended for imaging diagnostics services provision, computed tomography (CT) and ultrasound imaging diagnostics], surgery specific materials (implants, dressing material, consumables for surgical intervention and other consumables such as gloves, braunilas, tubes), laboratory analysis (general biochemistry and hemathology, coagulation status analysis, microbiology related lab), medicines (parenteral and enteral nutritive solutions and systems, blood and its deriva-tives -transfusions, antibiotics, antimicotics, antiviral and antiprotozoal drugs, analgesics, thromboprophylactic medicines and all other drugs).

Statistical analysis
Categorical variables were presented as frequencies of certain categories, while continuous variables were summarized as mean and standard deviation, as well as median and minimum and maximum values.Patients were divided into two groups based on the type of implantation: unilateral (implantation performed on only one knee) and bilateral (implantation performed on both knees).The differences in continuous variables were assessed by Mann Whitney U test because data were not normally distributed.The χ 2 test was used to assess differences in categorical variables.The differences were considered significant if probability of null hypothesis was less than 0.05.Costs are presented as mean and median cost per patient including standard deviation, minimum and maximum values.All mean and median cost values refer only to those patients that have actually used a particular service, as some services were used by few patients.Share of cost of certain category in total inpatient cost was calculated and presented graphically.Statistical analyses were performed using Microsoft Office Excel 2007 ® and IBM SPSS ® Statistics for Windows, Version 20.0 (IBM Corp, Armonk, NY, USA).

Results
Study sample consisted of 97 patients.Baseline characteristics of study sample are shown in Table 1.Mean age of entire study sample was 67.89 years, ranging from 41 to 83 years.There were 40 (41.2%)female patients and 57 (58.8%) male patients.Majority of patients (60, 61.9% patients) had unilateral implantation of total condylar knee endoprosthesis.Bilateral implantation was performed in 37 (38.1%)patients.There was no statistical difference in the mean age and mean duration of hospitalisation of patients who had unilateral implantation compared to patients who had bilateral implantation.However, fewer women had bilateral implantation compared to men.
Results of descriptive statistical analysis of cost domains are presented in Table 2. Structure and percentage ratio of mean costs per patient are shown in Figure 1.Two major cost drivers were surgery specific material and surgery.Cost of implants was the highest single cost driver in all observed groups of patients (Table 2, Figure 2).The cost associated with imaging diagnostic services was the lowest (Table 2, Figure 1).

Discussion
The value of previously reported mean inpatient cost of primary total knee replacement differs widely from country to country.When comparing results of different studies, it should be kept in mind that methods, data sources (e.g., hospital accounting system, reimbursement rates and charges, etc.) and categorization of costs may vary significantly.Mean total inpatient cost of total knee replacement estimated in our study is comparable with the mean inpatient cost in public (EUR 4,103) and private (EUR 5,226) hospitals in Portugal 36 .Studies conducted in Spain 48 , France 49 , Italy 37,50 and United Kingdom 7 reported somewhat higher values.In France, mean hospital cost was EUR 7,404 49 .In Spain, mean total inpatient cost was EUR 7,645 48 .In Italy, mean hospital cost per knee and hip replacement procedure was EUR 6,952, whereas the mean cost of the surgical procedure was EUR 3,798, while that of the inpatient care was EUR 2,924 50 .Another Italian study reported that average cost per patient (including hospital, rehabilitation and complication cost) was EUR 15,358 37 .In the United Kingdom, on average each admission costs British Pound (GBP) 6,363 according to the analysis of patient level data and services valued at 2007-2008 prices from the Knee Arthroplasty Trial 7 .In China, the total cost for unilateral procedure was USD 8,173.25,whereas for bilateral procedure it was USD 14,257.64 in 2010 51 .In Taiwan, mean total medical cost for unilateral procedure was USD 3,919 52 , whereas median cost of simultaneous bilateral procedure was USD 6,994.4 53.Early studies conducted in the United States reported that mean total inpatient cost for unilateral total knee replacement was USD 10,081 54 during 1991-1994, USD 12,561 during 1991-1992 55 and USD 15,673 during 2000-2008 56 .These costs can be even higher if patients have concomitant diagnosis of depression and anxiety 57 or venous thromboembolism and bleeding 58 .In a cohort of Australian patients, mean inpatient cost of knee replacement in the first 30 days postoperatively was Australian dolar (AUD) 21,006 for a period 2011-2012 19 .Estimated mean inpatient cost can range from Canadian dolar (CAD) 12,500 59 to CAD 14,758 60 in Canada.Higher reported costs in other countries may be attributed to the differences in the local healthcare systems, clinical, coding, administrative and costing practices in individual countries as well as patient demographics and surgeon practices.
As costs related to total knee replacement are substantial, it is important to develop strategies for their control and reduction.Orthopaedic surgeons should be the first and the most important patient advocates who will carefully evaluate hospital cost saving programs and ensure their enactment do not compromise treatment outcomes 6 .Several reports have highlighted that safe cost reduction can be achieved through the knee implant standardization process to reduce variation in implant selection and implementation of the clinical pathway programs which coordinate and standardize the activities of the physicians, nurses and other staff involved in providing care to the patients 6,32,[61][62][63][64] .One study reported that substantial cost savings can be achieved when one versus two packets of bone cement is used in combination with a hand mixing technique with no difference in clinical outcomes 28 .There are also proposals to increase number of total knee replacements in an outpatient setting 65 .However, it is more complicated to monitor recovery process as patients do not stay overnight under supervision and there is a lack of high quality evidence that directly compares outcomes of outpatient and inpatient orthopaedic procedures 65 .It has also been shown that shortening of waiting times for surgery is cost-effective and may also be cost saving 66 .
Growing body of literature has shown that the cost of implants contribute significantly to total cost of joint replacement procedures 6,29,51,67 .Large share of implant costs (64.64%) in total inpatient cost is an important finding of our study that confirms previous observations.Robinson et al. 35 reported that share of implant cost can vary from 13% to as high as 87% of total inpatient cost.Some other studies reported lower share.For example Portuguese study reported 28%-33% 36 and the US study 29%-40% 54 .Mean implant cost in our study was EUR 1,751.2, which again is similar to the cost reported in Portuguese study (EUR 1,259 -EUR 1,447) 36 and Italian study (EUR 1,850) 50 .Fixed implant cost associated with total knee replacement was CAD 3,060 in Canada 59 .An early US study reported that the average implant cost was USD 3,963 for unilateral procedure and USD 7,428 for bilateral procedure 54 .Variation in share and value of implant cost may be attributable to the patient characteristics and hospital characteristics as well as different categorization of other costs 35 .In addition, factors that may have influence on final cost of the implants are costs of design, research, development and manufacture as well as the cost of support staff such as industry representatives 68 .
The average prices of hip and knee implants have increased more than 100% over the past decade, although it would be expected to decrease with the increasing number of procedures if orthopaedic implant device companies followed conventional economies-of-scale principles 35,67 .Proposed strategies for restraining implant cost are volume-discounted vendor contracts, single-price contracts, unilateral price caps, implant standardization programs as well as surgeons cost awareness discussions 6,31,32,64,69 .Access to information on the prices of devices should be available to orthopaedic surgeons, and there should also be incentives for their participation in cost reductions programs 70 .It has also been shown that innovative implants used for total knee replacement should decrease failure of this procedure by 50%-55%, or more, compared to standard implants to be broadly cost-effective 71 .In addition, the patents on many widely used implants have recently expired and introduction of generic implants has the potential for major cost savings 68 .Companies that put generic implant replicas on the market have been established, but formal independent systems which should evaluate their absolute equivalence are yet to be founded 68 .The new generic implants are similar to the originals by a process of reverse engineering and their equivalence has been assessed in terms of geometry, but the monitoring and their independent evaluation to verify biomechanical compatibility is essential in order to prove that they are as good and safe as the originals 68 .Our findings should be interpreted in light of some potential limitations.Our analysis was restricted to the direct medical costs of procedure in inpatient setting.We did not take into account out-of-pocket patient's expenditure, indirect costs and costs associated with post-discharge period in outpatient setting when rehabilitation and complication costs may be considerable.In addition, hospital discharge invoices to Republic Health Insurance Fund may be partly unreliable in some cases as data entering is usually left to nurses or clerks who may not have sufficient comprehension of this process which can lead to incorrect data entry.Certain fraction of invoices is even disputed by the Republic Health Insurance Fund.For example, in the first half of 2016, the Republic Health Insurance Fund through the control of regularity of invoicing and demands for drug reimbursement noted that the amount of incorrect claims had a value of RSD 8,256,642.95(about EUR 67,072) 72 .In 2016, estimated expenditure of the Republic Health Insurance Fund on health care was RSD 204.3 billion (about EUR 1.66 billion) 73 .The pattern of services and materials acknowledged by the Republic Health Insurance Fund in some cases may also lead to the differences between what was invoiced and what patients really consumed, so our findings might have underestimated the true cost of some consumed services and materials.An-other limitation is rather modest sample size.However as included patients represent entire population of knee osteoarthritis patients, who were treated over a period of entire year in one of the largest university hospitals in Serbia, they certainly provide valuable insight into the value and structure of inpatient cost of total knee replacement in this region.

Conclusion
Our findings imply that inpatient costs associated with implantation of total condylar knee endoprosthesis are substantial.It seems that the most important cost drivers are surgery and surgery specific material, with implants being the highest single cost driver.Further research should be focused on analyzing factors that influence these costs in order to develop effective strategies which could contribute to substantial savings in the future.

Table 1 Baseline characteristics of study sample
Mean total inpatient cost per patient for both unilateral and bilateral implantation of total condylar knee endoprosthesis was EUR 2,709.1,ranging from EUR 1,685.2 to EUR 5,356.6.Mean total inpatient cost per patient was EUR 2,093.8(range: 1,685.2-3,358.2) for unilateral implantation and EUR 3,706.8(range: 2,195.3-5,356.6)for bilateral implantation.