1 INTRODUCTION

Cost-efficiency balancing is one of the areas of greatest concern in healthcare policy [1]. To achieve this, economic analysis is needed to assess the value of treatment and assist policymakers in allocating medical resources. Factors to be taken into consideration include the cost of treatment, as well as the quality and quantity of the patient's life. Cost-effective analysis (CEA) is generally recommended as an important tool for the economic analysis of health care [2-4].

Computer-assisted orthopedic systems (CAOS) such as robotics, computerized navigation and other computer-assisted devices have burgeoned in all orthopedic subspecialties, especially in arthroplasty and spinal surgery [5-7], over the past two decades [8-10]. The use of CAOS offers surgeons convenient and precise operative manipulations [11-14] and benefits patients with better clinical outcomes and lower failure rates of operations [15-19]. The development and application of CAOS are rapidly increasing [20, 21]. Given that the cost of CAOS is high, it is necessary to determine whether the accessional expenditures incurred by this new technique live up to the incremental outcomes as expected. Indeed, numerous studies have demonstrated the promising upsides of CAOS compared to traditional procedures [14, 22, 23]. However, recent studies on the efficacies of CAOS have shown controversial results. Bush et al. [24] found that an experienced surgeon could perform unicompartmental knee arthroplasty (UKA) by freehand even more precisely than using robotics. Passias et al. [25] reported the relatively higher costs of CAOS in lumbar fusion did not lead to a decrease in failure rate. Hickey et al. [26] pointed out that a myriad of comparative trials were underpowered to detect total knee arthroplasty (TKA) revision rate reductions due to CAOS. The authors subsequently conducted a simulated analysis. Although the precision of CAOS was set higher than that of the conventional method, the 15-year survivorship of simulated TKA showed a difference of 2% in revision rate between the CAOS group and the conventional method group, and the sample sizes are required to include 5000 or more, which was barely practical for real randomized controlled trials (RCTs) [26]. In addition, results regarding CAOS should be cautiously interpreted because bias probably exists. A recent review of joint arthroplasty found that almost all studies (91%) comparing robotic-assisted versus conventional techniques involved authors with financial conflicts, and studies with conflicts were more likely to report favorable outcomes of robotics than studies not associated with conflicts [27]. It is, therefore, arbitrary to state that CAOS is cost-effective.

CEAs in different disciplines of orthopedics, such as traumatology, arthroplasty, spine surgery, and sports medicine [28-31], have been fully discussed in previous systematic reviews; however, those in CAOS, the new-generation devices, have been poorly studied. The purpose of this systematic review was to identify, summarize, and evaluate published economic analyses assessing the cost-effectiveness of CAOS and to answer whether CAOS is cost-effective.

2 MATERIALS AND METHODS

3 RESULTS

There were 1288 studies found from the initial retrieval. After reading through the titles and abstracts, we excluded 1260 studies and reviewed the full texts of the remaining 28. Thirteen studies were then excluded and no other research meeting the inclusion criteria was identified in the reference lists. As a result, there were 15 eligible articles in this systematic review (Figure 1).

3.3 Quality of CEA

The overall mean QHES score was 86.1 (range: 70–92) and the mean quality scores for joint arthroplasty and spinal surgery were 86 and 86.7, respectively. Figure 2 showed the frequency with which the included studies met each QHES criterion. Study perspective (p = 0.112), industry-funded or conflict (p = 0.411), and authors’ degree (more than 2 authors with a degree higher than MD) (p = 0.183) were not significantly predictive of obtaining a high-quality CEA.

4 DISCUSSION

Our review summarized the CEA publication in CAOS and those studies were generally of high quality. Based on a paucity of available studies, the cost-effectiveness of CAOS was certificated in the field of joint arthroplasty. In spinal surgery, however, the answer remains controversial.

To our knowledge, this is the first systematic review of CEA in CAOS. In the era of precision medicine, CAOS is developing rapidly, with the expectation of precise manipulation, better clinical outcomes, and longer survival [57]. However, recent studies question its additional effectiveness compared to traditional methods. Therefore, CEAs are appealed [24, 25]. This review reflected the increasing interest in CEAs in CAOS and the requirement for prudent evidence.

Our current systematic review highlighted the evidence of health utility clarified by CEAs and also estimated the qualities of analyses to find aspects where the methodological promotions of CEAs could achieve. There is only one previous publication that examined the cost-effectiveness of CAOS in general, which reported four CEAs of navigation-assisted TKA without quality assessments [58]. In particular, we noticed that none of the CEAs in the present review were conducted from the societal perspective recommended in the United States, as they included both direct and indirect costs [59].

In this review, there is no certain consensus but a sketchy range regarding the ultimate question of analyzing cost-effectiveness, the WTP threshold. Studies using the US dollar system have generally used a threshold of $50,000–$100,000/QALY and those using the GBP system of £20,000–£30,000/QALY [60]. A procedure that is considered costly usually costs over $100,000 to gain a QALY. Nevertheless, researchers have questioned the scientific basis of the traditional value for this threshold, as it was determined several decades ago [61, 62]. In 2008, Braithwaite et al. [62] established a mathematical model based on insurance and mortality data and appealed that the acceptable lower and upper bounds for cost-effectiveness were $109,000 per QALY and $297,000 per QALY, respectively. The World Health Organization recommended three times the gross domestic product per QALY [63], but the applied WTP threshold still varied within a limited range between previous values, even in the most recent publications. The WTP threshold is crucial for determining cost-effectiveness and should be set prudently rather than crudely in future studies.

Previous CEAs and related reviews concerning manual joint arthroplasty have proved its cost-effectiveness compared with no surgery [64, 65]. Our findings suggested that CAOS could be cost-effective under several assumptions, such as the high surgical volume mentioned in almost all studies. The costs associated with the use of CAOS, ranging from US$400,000 (£290,000) to US$700,000 (£502,000), are a major concern, and the annual maintenance fees can reach £140,000 [66]. In the analysis models, a high number of surgical procedures could reduce the average cost per patient incurred by CAOS. In addition, the high volume could make the distinction of outcomes between the CAOS and conventional groups visible. Besides, the high volume helps surgeons overcome the learning curve. The annual cases suggested by these studies were over 250 for CAOS TKA and 100 for CAOS UKA. Our review seemed to reveal a dilemma for low-volume institutes where the improved clinical outcomes from CAOS might not be equated with the increased economic investments in current settings. Although the employment of this new device can probably attract more patients, hospital administrators should balance the investments, volumes, and clinical outcomes. Alternatives other than purchase, such as renting or cost-sharing strategies for low-volume institutes, could also be considered to reduce the expenditures for employing CAOS.

Three CEAs about CAOS for spinal surgery were identified, and based on their results, the answer remains uncertain. The studies by Passias et al. [25] and Garcia et al. [55] showed completely contradictory results, even in the same procedure, which is MIS lumbar interbody fusion. We believed that the contradiction was due to the difference between their economic evaluation methods. In the study by Passias et al. [25], the purchase costs of CAOS were integrated and allocated to the cost per patient based on surgical volume, whereas in the study by Garcia et al. [55], the purchase costs of CAOS were not taken into account, resulting in a sharp decrease in the total costs of CAOS procedure. This might explain why the results of Garcia et al. [67] were positive and the results of Passias et al. [25] were negative. Additionally, a previous study demonstrated that CAOS might fail to reduce the postoperative complications in spinal surgery [67], which also have a negative impact on health and economics. With the increasing application of CAOS in spinal surgery, further studies are needed to explore their cost-effectiveness and facilitate resource allocation decisions.

One of the most important goals of CAOS is to achieve a lower failure rate in surgery. In our review, CAOS arthroplasty demonstrated a longer survival than manual operation in the simulation models [43, 47, 51, 53]. As for spinal surgery, CAOS could also increase the accuracy of pedicle screw insertion and reduce the risk of revision [56]. Most of these studies attributed the better longevity of CAOS to the ability of CAOS that could obtain a better alignment of implants. The revision rate of CAOS also significantly influenced the cost-effectiveness. Moschetti et al. [48] stated that the 2-year revision rate should be lower than 1.2% to achieve a cost-effective CAOS UKA [48]. Burn et al. [68] studied the cost-effectiveness of CAOS TKA and THA at a given WTP threshold of £20,000 to £30,000 per QALY [68]. In their analysis, the revision rate was set at 5.3% for CAOS TKA and 8.2% for CAOS THA. The authors found that a 50% relative reduction in revision rate could improve the threshold with £1094 for knee replacement and £1347 for hip replacement [68]. In the CEAs for CAOS, especially when using economic models, researchers should carefully estimate the revision rate of CAOS as the economic models are extremely sensitive to these values. There is also a call for studies to improve the survival of computer-assisted surgeries.

The majority of the studies involved were of high quality, but there were also weaknesses. Most studies used data from registries or publications. Only one study employed prospective data [56], which is considered the best available source [42]. Future studies on CEAs of CAOS would benefit from an RCT design and data collection. The potential factors related to the quality of CEAs defined in previous studies were also evaluated in this review, but none of them showed a statistical relationship. It should be noted that over half of the CEAs for CAOS in this review were funded by industries or conducted by authors who had a conflict of interest, although we are not inferring that these studies are biased.

Currently, published CEAs for CAOS are limited to the disciplines of primary joint arthroplasty and spinal surgery. However, CAOS has been used in complex revision arthroplasty [69], traumatology [70], and bone tumor surgery [71]. MacAskill et al. [69] and Zhang et al. [72] reported two cases of revision TKA and a case of revision THA using the MAKO robotic system (Stryker). Several studies employed the navigation system to facilitate internal fixation in patients with hip fractures [70, 73, 74]. However, no CEA has evaluated their economic outcomes. Thus, further studies are needed in these areas.

We noted several limitations in this review. As this is a systematic review, it may be unavoidable that relevant studies are missing. The paucity of available studies and heterogeneity of the reported outcomes limited the conclusions. However, we provided an overview of CAOS where there are few CEAs, so future studies are needed. We also used QHES to evaluate the quality of the articles. Although this scale has been employed for evaluation in many economic analyses, its primary focus is the quality of reporting [31]. The assessment of the relationship between quality and associated factors should also be interpreted with caution.

5 CONCLUSION

There are a small number of CEAs of CAOS, and our current review of available studies found that CAOS in joint arthroplasty may be cost-effective under various assumptions, particularly in the case of high surgical volumes. Its cost-effectiveness in spinal surgery needs more research effort to be clearly presented. Although current CEAs are of high quality, requiring more attention to their methodology. Further studies are needed to assess the cost-effectiveness of CAOS in different areas of orthopedics.

AUTHOR CONTRIBUTIONS

Hua Li: Data curation (equal); formal analysis (equal); methodology (equal); writing–original draft (lead); writing–review and editing (equal). Tengfeng Zhuang: Methodology (equal); writing–original draft (equal); writing–review and editing (equal). Wenrui Wu: Formal analysis (supporting); investigation (equal); methodology (equal); resources (equal); writing–original draft (supporting); writing–review and editing (supporting). Wenyi Gan: Formal analysis (supporting); methodology (equal); resources (supporting); writing–original draft (supporting); writing–review and editing (supporting). Chongjie Wu: Formal analysis (supporting); writing–original draft (supporting); writing–review and editing (supporting). Sijun Peng: Methodology (supporting); resources (supporting); writing–original draft (supporting); writing–review and editing (supporting). Songwei Huan: Conceptualization (equal); methodology (equal); writing–original draft (equal); writing–review and editing (equal). Ning Liu: Conceptualization (lead); supervision (lead); writing–review and editing (lead).

CONFLICT OF INTEREST

The authors declare no conflict of interest.

ETHICS STATEMENT

This is a systematic review and no ethical approval is required.

INFORMED CONSENT

Not applicable.