Cost-effectiveness of follow-up treatment after knee operations

Orthopedic operations are among the most common operations in German hospitals and, according to initial forecasts, will continue to gain in importance as society ages [1, 2]. In order to avoid immobilization damage and to specifically support the healing process, mobilization measures should be started as early as possible during operations on knee joints, which must be carried out (almost) exclusively passively in the early postoperative phase. This is particularly true if the patient cannot or is not allowed to carry out any active self-exercises [3-7]. In addition to physiotherapeutic treatments, the use of motor-driven movement splints (Continuous Passive Motion - CPM splints) has become established in practice for passive movement of the joints. In the knee area, CPM movement splints are often used after endoprosthetic joint replacements, cartilage, tendon and ligament reconstructions as well as for exercise-stable osteosynthesis [8-13]. Numerous studies deal with the rehabilitative aftercare of patients after joint operations, but only rarely in the context of outpatient follow-up treatment in Germany. This work focuses on the actual care situation of patients in Germany in the sense of real world evidence. The starting point is a process cost analysis that documents the postoperative follow-up treatment of patients after knee joint surgery. The focus is on evaluating the cost-effectiveness of different treatment alternatives. Postoperative mobilization can be accomplished through the use of CPM motion splints, physical therapy, or a combination of both. The use of physiotherapy is regulated via the Therapeutic Products Directive; CPM movement splints are listed in the list of aids according to Section 139 SGB V as externally powered movement splints for the knee joint, among other things [14].

>> It can be assumed that from the perspective of the payers, but also from a societal perspective, an operation on the knee joint has significant financial effects. Since there are only a few analyzes on the costs of rehabilitation treatment in the literature [15], the present health economic modeling examines the economic viability of the following three treatment alternatives: treatment with a CPM movement splint, physiotherapy and the combination of both alternatives. First, an evaluation of the direct costs is carried out, then the financial expenditure for society as a whole for the follow-up care of knee operations in Germany is evaluated.
methodology
Since the aim of this work is to determine the costs from both the perspective of statutory health insurance (GKV) and from a societal perspective, the analysis was divided into two steps.
In a budget impact analysis (BIA), the direct financial consequences of the individual therapy alternatives for rehabilitative aftercare were first evaluated from the perspective of the statutory health insurance. The proportion of patients who receive a treatment option, the distribution of the therapy in the German healthcare system and the costs associated with the reimbursement of the individual treatment options must be taken into account [16, 17]. The robustness of the results was also checked through sensitivity analyzes and visualized using tornado diagrams.
In a second part, the total social costs were then evaluated, taking effectiveness parameters into account. For this purpose, indirect costs in the form of productivity losses as well as intangible effects, operationalized through quality of life parameters, were also taken into account.
The following three treatment alternatives after knee operations are compared in detail as part of the BIA and the comments on social costs:
• Treatment arm 1: Patients who received CPM splint treatment (CPM group)
• Treatment arm 2: Patients who received only physiotherapy treatment (physiotherapy group)
• Treatment arm 3: Patients who received both CPM splint treatment and physiotherapy treatment (combination group)

The calculations carried out were calculated using the programs Microsoft Excel, R and TreeAge Pro. The time horizon for the calculation covered a period of six weeks from the time of the first visit (maximum ten days after the operation) and was supplemented by retrospective information from the examiners on inpatient postoperative treatment.
Process cost analysis
As part of a recently carried out process cost analysis, a total of 243 patients after knee operations were surveyed at two points in time six weeks apart regarding their postoperative follow-up treatment. Statistically significant differences in socio-demographic information on age and gender as well as in everyday functionality could not be determined between the groups during the first visit. This ensured the basic requirement for the statistical comparability of the study groups.
The aim of the process cost analysis was to describe the supply constellation and the corresponding cost variables. Among other things, patient information on the duration of the inpatient stay, the type of operation and detailed information on the rehabilitative aftercare process were collected. In addition, the everyday effectiveness of the rehabilitation procedures used was determined, taking into account the patient-relevant parameters of pain, functionality, quality of life and stability.
Since the literature did not provide sufficient information on the transition probabilities of actual care in Germany, the survey served as the basis for the BIA.

Cost parameters
The present BIA was carried out from the perspective of the statutory health insurance company in Germany. Accordingly, the direct costs of the patient's postoperative rehabilitative aftercare are relevant. This includes rehabilitation treatment, physiotherapy and the use of other therapeutic services, aids and medicines. Co-payments and co-payments by the insured are deliberately excluded from the calculations because the patients in question were partially exempt from co-payments. The operation and any doctor's visits that may arise as part of the follow-up treatment were also not included in the calculation. The recording of the relevant cost parameters begins immediately after the knee joint operation and thus covers postoperative aftercare for both the inpatient and outpatient sectors. All assumed in the model
Cost parameters are shown in Figure 1.
The average duration of the inpatient stay was determined using the data from the process cost analysis. This was then offset against the base case value for the inpatient stay of the Institute for the Hospital Remuneration System (InEK) for the respective therapy options. The federal base case value (BBFW) represents the starting point for calculating these costs assigned to the postoperative inpatient stay. The cost range per day of inpatient stay was therefore set at 90-385 euros (mean value: 235 euros). These costs also include inpatient physiotherapy and inpatient splint treatment via the corresponding DRG, as hospitals purchase CPM movement splints once and therefore the treatment does not have to be reimbursed additionally.
Outpatient physiotherapeutic treatments are included in the BIA according to the physical therapy remuneration list (vdek) with a basic value of 16.62 euros per 20-minute treatment. For the BIA, an average value was taken from the data from the process cost analysis. The information on other therapeutic services varied significantly within the arms considered. Here, an average value corresponding to the therapy services received for the individual arms was included in the calculation based on the physical therapy reimbursement list (vdek) [18].
The costs for the six-week outpatient splint therapy are based on medical aid contracts with various health insurance companies. Depending on the service content of the contract, the costs for a six-week supply of CPM knee movement splints are between 300-450 euros. In order to reflect the variability of these costs, the average value was calculated. As part of the six-week treatment, the average cost for the knee indication is 375 euros.
The other aids were given average market prices; the costs were included in the calculations of the different groups according to the frequencies mentioned. The most frequently used aids were crutches and orthoses or bandages.
The same applies to the list of medications taken due to the operation. Here, an average price was determined for the individual arms based on the 2017 Red List according to the frequency mentioned [19]. The most commonly documented active ingredients were ibuprofen, diclofenac, metamizole, oxycodone, paracetamol and tilidine.

Total population
Since the process cost analysis is clearly aimed at depicting the actual reality of care for the patient (real world evidence), the patients were included in the evaluation regardless of the underlying disease or type of operation. The information on the type of knee operation varied accordingly (see Table 2).
In order to create a uniform calculation basis, all relevant knee operations were determined using the corresponding ICD-10 codes. In detail, the patient numbers were derived from the hospital statistics for 2016 (hospital statistics - diagnosis data of patients in hospitals) [24]. For the projection carried out as part of the BIA, this estimate amounted to 352,599 knee operations.
Results
Budget Impact Analysis
After completion of the six-week observation period, direct therapy costs of 820.41 euros per patient (3) were generated for patients as part of the inpatient and outpatient follow-up treatment with CPM movement splints after knee operations. In comparison, direct costs of 960.71 euros per patient were incurred if the follow-up treatment only consisted of the use of physiotherapeutic measures. At 1,396.48 euros, the combination treatment of CPM splint treatment and physiotherapy resulted in the highest treatment costs for patients after knee operations.
From the perspective of the statutory health insurance, the CPM splint treatment is the most cost-effective of the three treatment alternatives. In particular, the costs for postoperative inpatient stays were cheaper than for patients with physiotherapeutic treatment alone or combination therapy. This is particularly noticeable because in many cases only passive movements are permitted following the operation and meaningful physiotherapeutic treatment is only possible much later. Therefore, even higher costs for physiotherapy can be expected for long-term considerations (> 6 weeks). The higher costs for assistive devices in the CPM treatment arm are due to a larger proportion of patients who received an orthosis in addition to the CPM splint treatment (around 80% and thus significantly more than in the other two groups).
Sensitivity analyses
In order to examine the robustness of the cost parameters and the results based on them, various sensitivity analyzes were carried out. As part of this, the following cost parameters were varied:
I. Duration of inpatient stay
II. Physiotherapeutic treatments (up to 6 weeks)
III. outpatient CPM treatment
IV. other aids
V. other therapy services
VI. Painkillers/medicines
The results of the sensitivity analyzes can be found in Table 4. The standard deviations for the first two model variables from the process cost analysis were adopted1. The costs for outpatient splint treatment vary depending on the service content of the medical aid contract.
The statistical uncertainty of the cost data for aids, therapy services and medicines was depicted with a deviation of ±20% from the base case value.
As part of the analysis of the knee indication, the direct treatment costs vary depending on the treatment arm between a minimum value of 102.09 euros and a maximum value of 3,030.21 euros. When looking at the maximum values, it can be observed that the CPM splint treatment remains the most cost-effective even in the sensitivity analyses. Overall, it can be said that the stability of the results is maintained in the sensitivity analyzes despite the sometimes large cost ranges. These cost ranges (and in this context also the minimum costs of 0 euros in some cases) are due to the standard deviations of the data from the care study. Of course, no costs only arise if no physiotherapeutic or inpatient treatment has been given (see 4).
In Table 5, the cost parameters from the case-centered analysis were extrapolated to the population of all knee operations in Germany. The table therefore shows the direct total costs for a six-week follow-up treatment of all operations that take place in a year. With direct costs of 289,274,188.36 euros, the total annual expenditure for patients with CPM splint treatment on the knee joint is the lowest. Assuming a six-week follow-up treatment, there is potential for savings of up to 50 million euros compared to physiotherapy. The total costs for combination therapy are significantly higher (492,398,130.98 euros). The tornado diagram in 1 further illustrates the cost ranges of the sensitivity analyses. For patients after knee surgery, it is clearly evident that the parameters of inpatient stay and physiotherapy have the largest cost ranges in the present data set. These two parameters have the greatest influence on the uncertainty of the health economic model and should be taken into account as a limitation when interpreting the results.
Costs from the perspective of society
From the perspective of society, in addition to the direct treatment costs, indirect costs in the form of lost productivity and intangible effects (quality of life parameters) must also be taken into account. First, the indirect costs are considered.

Indirect costs
Since the average age of the patients from the process cost analysis was 51.59 years, an additional consideration of the indirect costs due to illness-related loss of productivity seems sensible. It was shown that patients after knee surgery with combination therapy had a similarly high number of days lost from work for all indications (CPM splint therapy: 32.63 days; physiotherapy: 33.30 days; combination therapy: 33.06 days). At 74%, the employment rate for patients with combination therapy was significantly above the comparative values for CPM splint therapy (59%) and physiotherapy (54%). No information was available regarding possible early retirement in the context of the process cost analysis.
The indirect costs consist of the duration of incapacity for work of the working subpopulation and the average basic value according to the Hanover Consensus (90 euros per day of incapacity for work) [23]. The amount of these indirect costs can be found in 6. For patients after knee operations, the costs resulting from loss of production are lowest when treated with physiotherapy alone (1,631.52 euros). Slightly higher costs are incurred for CPM splint treatment (1,743.40 euros); the highest indirect costs were generated by patients with combination therapy (2,209.88 euros) due to the higher employment rate.

total cost
Direct and indirect costs together form the total costs from a societal perspective. These can also be seen in Figure 6. This shows that the total costs for patients with CPM splint therapy are lowest (2,563.80 euros). Physiotherapy is somewhat more expensive (2,592.22 euros). The highest total costs were incurred by patients after knee surgery for the combination therapy (3,606.36 euros). Losses in productivity, inpatient stays and physiotherapeutic treatment each make up the largest cost components.

Intangible effects (everyday functionality)
From a societal perspective, a pure evaluation of the (total) costs is in many cases not meaningful enough and, in the context of this work, is logically supplemented by the effectiveness parameters from the data from the process cost analysis. These effectiveness parameters, which were additionally collected as part of the process cost analysis, were determined for patients after knee surgery using a validated version of the Oxford Knee Score (OKS) [25].
This Oxford Score is used as an assessment
Instrument used in the context of empirical studies to evaluate patient-related everyday functionality and quality of life after knee operations. As part of the decision tree analysis, a treatment is defined as successful if the Oxford Score has a total score of 30 or more points. The overall value range is between 0-48, with increasing values corresponding to better everyday functionality for the patient. According to Dawson [26, 27], the patient only has a mild to moderate limitation, which can be treated by a family doctor. If the Oxford score is ≥40 points, no further treatment steps are necessary.
Figure 2 illustrates the results of the process cost analysis using a decision tree model. The CPM splint treatment leads
for patients after knee operations among the largest proportions of successful treatments (57%). However, follow-up treatment using physiotherapeutic measures resulted in the lowest proportion of successful treatments (19%).
The results for combination therapy, in which physiotherapeutic treatment is supplemented by CPM splint treatment, are significantly more positive than for physiotherapy. The consideration of quality of life parameters is not insignificant for the evaluation of the treatment arms, which is why intangible effects in the form of effectiveness-adjusted costs were also used in addition to the BIA. To calculate these intangible effects, the difference in the Oxford Knee Score between the two visits was determined as part of the process cost analysis. The results are recorded in Table 7. If one takes into account the improvement in the Oxford Scores between the two times of the investigation, the trend in the evaluation of the total costs is confirmed and shows a clear picture with regard to the effectiveness-adjusted costs in Figure 6.
An improvement in the OKS by one score point is the cheapest way to achieve the CPM splint treatment (163.82 euros). Combination therapy already has a significantly worse cost-effectiveness ratio (350.81 euros). Physiotherapy performs worst when taking into account the parameters of everyday effectiveness (495.64 euros). The results on social costs clearly show that CPM splint treatment is the most cost-effective and cost-effective of the three treatment options. Overall, combination therapy, which achieves similarly good effectiveness with higher costs, is considered to be somewhat more cost-intensive. The cost-effectiveness ratio of the physiotherapeutic treatment strategy is the least favorable compared to the other treatment options.
discussion
The present BIA examined the costs of three rehabilitation procedures (CPM splint therapy, physiotherapy, combination therapy) for outpatient, postoperative follow-up care after knee joint surgery from two different perspectives. The perspective of the GKV was reflected through the budget impact analysis. In addition, an evaluation of the total costs from the perspective of society was carried out. For CPM splint therapy, direct costs of 820.41 euros for six weeks of postoperative treatment were observed after knee operations. This made it the most cost-effective of the three treatment alternatives. These results remained even after the total costs had been evaluated and were further strengthened when intangible effects were taken into account. It was shown here that treatment with CPM movement splints on the knee joint is the most cost-effective and cost-effective of the three treatment options from both a statutory health insurance and a societal perspective.
The results of the BIA and the overall cost analysis provide additional empirical evidence as to what economic postoperative aftercare following knee surgery can look like. The claim to care of CPM splint therapy, which has been anchored in the directory of aids within the orthopedic care landscape since 1996, is further underlined by the present results. The actual supply situation in Germany is extremely complex and diverse. However, the results on economic efficiency confirm initial trends from comparable studies. A recently published work as part of a cost-effectiveness analysis showed the superiority of combination therapy (CPM + PT) compared to inpatient follow-up treatment according to the rehabilitation standard with physiotherapy alone for knee prosthesis operations [15]. The present work resolves the uncertainty as to the extent to which these results are also confirmed in a cross-group outpatient setting. However, the present BIA also has some limitations that should not go unmentioned in the course of the discussion. These are partly in the health economics
Model assumptions, but above all in the underlying database. The quality of life parameters additionally collected as part of the process cost analysis were indicative.
determined cross-functionally and cannot replace a comprehensive quality of life study on the respective indications.
The extrapolations of the total costs for the individual indications are based on values from hospital statistics [24]. It must be emphasized that the information provided by patients for a specific indication was not always consistent with the common ICD-10 coding. Therefore, the frequency of operations and their influence on the costs for society as a whole should only be viewed as model assumptions. Because physiotherapy is almost obligatory in the process cost analysis, a precise examination of the necessity, benefits and cost-effectiveness of physiotherapeutic measures would be a basic prerequisite for their widespread use. The lack of standardization of physiotherapeutic treatments makes such analyzes comparatively difficult compared to CPM splint treatment. Overall, however, it can be confirmed that outpatient follow-up treatment of patients after knee surgery with CPM splint therapy is the most cost-effective and cost-effective treatment option. <<