Comparative cost-effectiveness analysis of the subacromial spacer for irreparable and massive rotator cuff tears

The model was built and analyzed with the software TreeAge Pro Healthcare 2017. For each IRCT treatment strategy, the course of treatment and outcomes are represented (Fig. 2). A probabilistic subset of those patients will have success with a treatment or course of action, while the remaining will require revision or another treatment, which also has a probability of success or failure, based on the prior treatment and outcome. As shown in Fig. 2, if a patient begins with non-operative care but symptoms worsen, they will proceed to either rTSA or InSpace™. For patients undergoing partial repair, they may either have a successful procedure or require a revision. As with patients receiving non-operative care, if they require revision, they will either undergo rTSA or receive InSpace™. For patients initially undergoing rTSA, their procedure will either be successful or will require a revision procedure. If they require a revision, the revision will either be successful or fail. If the revision fails, they will remain in that health state. If a patient initially utilizes InSpace™, they will either have no complications or require a revision (rTSA) which will either be successful or fail. If the revision fails, the patient will remain in that health state. Thus, a weighted average of the possible returns for each alternative (costs and QALYs), with probabilities used as weights, is calculated.

In the model, we operated under several assumptions regarding the patient treatments and costs: (1) all patients entering the model are diagnosed with IRCT; (2) prior to entering the respective treatment arms, patients underwent equivalent prior non-operative care of 6 weeks with equivalent costs; (3) costs included direct medical care costs; (4) post-surgery rehabilitation costs were not included due to a lack of studies containing that information; and (5) patients who experience complications following surgical care require revision surgery. Additionally, model parameters, displayed with their corresponding data sources in Table 1, consist of costs, health utilities, and transition probabilities. Costs represent direct medical costs to the health care system and were sourced from published literature and based on US Medicare cost of surgical costs, including the cost of the procedure, implant costs, and costs of hospital care. All costs were normalized to 2017 US dollars using a cumulative inflation rate based on the consumer price index (CPI) for health care services (use Sundar source). They were then translated into Italian health care Euros using the purchasing power parity (PPP) conversion factor, which is the number of units of the country’s currency required to buy the same amounts of goods and services in the domestic market as the US dollar would buy in the USA, currently 0.75 for Italy. PPPs show the ratio of prices in national currencies of the same good or service in different countries. They were then translated into Italian health care system Euros using a purchasing power parity (PPP) of 0.75. The cost of InSpace is assumed to be less than the cost of rTSA and similar in cost to arthroscopic repair. This input was validated by a review of Italian hospital charges. As reduced quality of life (QoL) is associated with both disease treatment and disease health outcomes (Whitehead 2010), we incorporated QoL for each health outcome of interest using estimates of utilities reported in the peer-reviewed literature. In this analysis, health utilities measure quality of life on a 0–1 scale anchored by death (0) and perfect health (1) and are associated with each IRCT treatment outcome. The primary patient outcome of interest for this analysis is the total quality-adjusted life years (QALYs) experienced by the patient over the course of a 24-month period. The timeframe was chosen based on the length of follow-up time in recently published studies. We assume that treatment decisions (i.e., revision) are made within one year, allowing a year of follow-up time. Utilities are used to calculate quality-adjusted life years (QALYs) experienced by a patient by decrementing the amount of time that has passed by the health utility experienced during that time period. For example, a patient experiencing perfect health (i.e., health utility = 1) for a period of two years will experience 2 QALYs during that 2-year period. A patient experiencing a health utility of 0.5 will experience 2 × 0.5 = 1 QALY during that two year period.

Movement through the decision tree, shown in Fig. 2, is governed by (1) the success or failure rates of the various IRCT management strategies; (2) the likelihood of revision procedures; and (3) the success or failure rates of the revision procedures. These model transition probabilities were derived from the literature and their values and data sources are also displayed in Table 1. As hypothetical patients traverse the model, applicable costs and outcomes (health utilities) associated with the treatment alternatives for IRCTs accrue. Using the accrued cost and accrued utilities as the effectiveness parameter for each potential path through the model, the model evaluates the expected cost and expected QALYs based on the transition probabilities for each treatment strategy.

All analyses were conducted by following the recommendations from the ISPOR Panel on Cost-effectiveness in Health and Medicine. Comparisons of strategies were made using the incremental cost-effectiveness ratio (ICER), which is the ratio between the difference in costs and QALYs of each strategy compared to the previous strategy, when strategies are ordered by cost, e.g., incremental dollars per QALY. A willingness to pay (WTP) threshold for the additional cost incurred for each change in QALYs is established. Typically, strategies with an ICER of less than 50,000 US dollars (45,483 €) are considered cost-effective. A strategy is dominated when another strategy is both less costly and more effective and is dominant when it costs less and is more effective than another alternative. A strategy with a negative ICER is dominated, reflecting the additional cost per loss in QALYs. Incremental analysis is critical for decisions regarding the allocation of scarce resources. The “base case” analysis is the scenario which involves evaluating the model with each parameter at its estimated value. One- and two-way sensitivity analyses were performed to explore the impact of varying key parameter values from the base case on the model results.

The expected costs, QALYs, and ICERs associated with each treatment alternative are presented in Table 2. After conducting the base case analysis, the study finds that InSpace™ is the dominant strategy, meaning that it is both less costly and more effective than all other treatment alternatives (Fig. 3). InSpace™ dominates both arthroscopic partial repair and reverse shoulder arthroplasty since it costs less than both options and increases effectiveness by 0.06 and 0.10 QALYs, respectively. While non-operative care is the least costly management strategy, InSpace™ results in a gain of 0.05 QALYs for the additional cost of 522 €, resulting in an ICER of 10,440 €/QALY gain, which is below the standard willingness to pay ratio of $50,000 US dollars.

Typically, strategies with an ICER of less than 50,000 US dollars (45,483 €) are considered cost-effective. Thus, InSpace™ is the most cost-effective strategy and may be a preferable treatment strategy for the management of irreparable rotator cuffs.

A one-way sensitivity analysis was conducted on key variables including cost of InSpace™, probability of InSpace™ revision, health utility associated with InSpace™ success, probability of success of non-operative care, cost of reverse total shoulder arthroplasty, cost of arthroscopic partial repair, and cost of non-operative care. In the sensitivity analysis, the parameter range utilized was based on a range of values established in the literature, where applicable, or was allowed to vary (one at a time) up or down by 20%. Notably, even if the cost of InSpace™ is increased by 20% or the probability of success of the non-operative strategy decreases to 87.5%, it is not dominated by an alternative treatment option. Further, if the cost of InSpace™ decreases by 20%, the probability the success of InSpace™ increases to 96%, or if the cost of non-operative care increases to $17,327 US dollars, then InSpace™ dominates all strategies. However, if the utility associated with InSpace™ success drops by 20%, then non-operative care provides higher QALYs than InSpace™ at lower costs. The results of the one-way sensitivity analysis are shown in Table 3.

Figure 4 shows the results of the two-way sensitivity analysis between the utility of successful InSpace™ and successful rTSA in a comparison of the three surgical treatments. Across the grid, the colours represent the preferred strategy based on the net monetary benefit (NMB) using a WTP strategy of $50,000 US dollars (45,483 €). InSpace™ was preferred with a utility greater than 0.55 and a difference no greater than 0.20 between it and rTSA, while rTSA was only preferred if the utility difference was larger than 0.20. Partial arthroscopic repair was only preferred when the utility of InSpace™ was below 0.55 and utility of rTSA below 0.75.