Over the past 10 yr, three-dimensional (3D) printing has grown considerably in the fields of medicine and surgery. Using data from patient CT or MRI scans, this technology allows anatomical models and patient-specific instruments (PSIs) or implants to be developed, and thus finds its full place in the personalization of surgery (Reference Hodgdon, Danrad, Patel, Smith, Richardson and Ballard1). 3D objects based on patient images are designed by computer-aided design/computer-aided manufacturing (CAD/CAM) technologies and the final product is obtained by adding materials layer by layer. This technology is used in all surgical domains, particularly in orthopedic and craniomaxillofacial surgery, to carry out preoperative planning, to provide intraoperative assistance, or to train students (Reference Tack, Victor, Gemmel and Annemans2). The use of 3D printing technology makes it possible either to manage more complex cases that could not be handled by standard medical devices available on the market, or to improve the outcomes of routine interventions.
The wider use of this technology has made it possible to initiate its clinical evaluation. The first results of initial comparative studies have tended to demonstrate a reduction in operating time and an improvement in aesthetic results (Reference Serrano, van den Brink, Pineau, Prognon and Martelli3;Reference Diment, Thompson and Bergmann4). In 2018, a review of the clinical trial databases identified 92 ongoing clinical trials whose first results are expected within the next 2 yr and which should fill the gaps in clinical evaluation (Reference Witowski, Sitkowski, Zuzak, Coles-Black, Chuen and Major5).
Nevertheless, one of the major limitations cited as a barrier to its expansion is the extra cost associated with its use (Reference Chow and Cheung6–Reference Murray, Edwards, Mainprize and Antonyshyn9). In the case of in-house production within the healthcare facility, these additional costs may come from the purchase of 3D printers, materials, or even the subscription to software required for device design. In the case of production outsourced to a third-party provider, the additional costs may also be linked to the provision of associated services, such as virtual preoperative simulation. These costs may be partially or entirely funded by the institutions and therefore limit the use of this technology. Consequently, decision-makers in hospital settings, and in the context of limited resources, need information on this topic to support their decisions.
Thus, in parallel with the clinical evaluation, economic evaluation of 3D printing is essential to assess the impact of this health technology and is also a valuable tool to support decision-making. We, therefore, carried out a systematic literature review to assess its economic evaluation, identify the costs associated with its use, and highlight the first quantitative data available.
Material and Methods
Study Inclusion
In order to identify the relevant economic data on 3D printing, a systematic literature review was conducted in the PubMed and Embase databases and in the National Health Service Economic Evaluation Database (NHS EED) at the University of York, following the PRISMA recommendations for reporting (Reference Moher, Liberati, Tetzlaff and Altman10) (Supplementary file 1, PRISMA Checklist). The search strategy was developed using the PubMed database and then applied to other databases (Supplementary file 2, Study Protocol). The keywords used were “3D printing AND costs and surgery,” “Cost-Benefit Analysis AND 3D printing,” “Economics” AND 3D printing,” “Costs AND Cost Analysis [Mesh] AND 3D printing,” “Health Care Economics AND Organizations [Mesh] AND Printing, Three-Dimensional”[Mesh].” Initially, titles and abstracts were screened by two independent reviewers (CS and NM) to identify studies that reported an assessment of the costs associated with the use of 3D printing for surgical applications according to the PICOS (Population, Intervention, Comparators, Outcomes, and Study type) framework (Reference Schardt, Adams, Owens, Keitz and Fontelo11) (Supplementary file 1, Study Protocol). No inclusion limits were set on the printing technology used or on the place of production (hospital, third-party provider). All printed medical devices were considered (anatomical models, surgical guides, or implants) intended for use in any surgical domain. All items with a cost calculation related to 3D printing were included. Only articles published in English and French between 2009 and 2019 were considered. The exclusion criteria were: studies reporting the use of 3D printing in dental surgery or for the production of external prostheses, studies related to fundamental research or without hospital application or not applied in humans, literature reviews and studies presented at conferences. In a second step, the eligible studies were then selected after a full-text reading by CS and NM, and the exclusion criteria were the same as in the first step. In both steps, in cases of discordant screening or selection, the two researchers discussed the discrepancy until a consensus was reached.
For the included studies, organizational data (place of production, professionals involved), technical data (the type of medical device printed, technology, and materials), and economic data (cost or savings items) were collected. When costs were not in U.S. dollars (USD), local currencies were converted to USD using the exchange rate relating to the study period (the period during which the cost data were collected). When the study period was not clearly stated, we used the exchange rate from the article's publication year. The values were then inflated using USD inflation rates to the base year of the analysis (i.e., 2019) (Reference Turner, Lauer, Tran, Teerawattananon and Jit12). Exchange and inflation rates from the World Bank were used as a source. Consequently, to improve data interpretation, all costs presented in this article are expressed in USD 2019 values.
Quality Assessment
In order to assess the quality of articles screened, we used two established checklists to appraise the reporting and methodological quality of studies included. These tools are all qualitative instruments. In cases of discordant classifications, two researchers (CS and NM) discussed discrepancies until a consensus was reached. Firstly, we used the Consolidated Health Economic Evaluation Reporting Standards (CHEERS) checklist, which is a practical tool used to assess the reporting quality of health economics studies (Reference Husereau, Drummond, Petrou, Carswell, Moher and Greenberg13). This instrument includes twenty-four items addressed in six categories (title and abstract, introduction, methods, results, discussion, and other); we assigned one point if the item was complete, a half-point for the partial answer, and no points if the information was absent. The maximum score reachable with this tool is normally 24. Nevertheless, some items were not applicable, and so the maximum score reachable was calculated for each article. We also evaluated the level of evidence using the fivefold scale of Sackett et al. (Reference Sackett14). This rating system allows rapid identification of the potential clinical quality of the study (Supplementary file 3, Quality Assessment).
Results
Study Selection
After excluding duplicates, 473 studies were identified, of which 429 were excluded on the basis of title and abstract. The reasons for exclusion are presented on the PRISMA Flow Chart (Figure 1). Of the remaining forty-four studies, thirty-five were excluded after the full-text analysis. Thus, nine studies met the eligibility criteria and were included in the systematic review. The characteristics of the included studies are summarized in Table 1. The nine studies were published between 2015 and 2019. The printed devices were mainly anatomical models (n = 6). The other printed devices were surgical instruments (n = 2), a simulator (n = 1), and a surgical template (n = 1). No studies have evaluated the production costs of an implant. The printing of the devices occurred in-house in six (67 percent) studies.
Figure 1. PRISMA flow chart of included studies.
Table 1. Characteristics of the Included Studies
CAD, computer aided design; 3D, three-dimensional; USD, U.S. dollars.
Quality of the Included Studies
According to the CHEERS checklist, the quality of the studies included was poor (Supplementary file 3, Quality Assessment) (Reference King, Park, Christensen and Danrad15–Reference Yang, Li, Li, Zhao, Wei and Zhang23). The number of items and recommendations gained by each study ranged from 1.5 to 9 with an average of 4 (the maximum reachable score was 22 here). According to the fivefold scale of Sackett et al., six studies were classified as level IV, two as level III, and one as level II.
Cost Evaluation
Nine types of costs were identified in these studies (Table 2): costs in printing materials (n = 6), staff costs (n = 3), operating room costs (n = 3), software subscription (n = 2), purchase and maintenance of a 3D printer (n = 2), purchase of the printed device from an external supplier (n = 2), electricity costs (n = 1), an ancillary component for the simulator (n = 1), and ancillary services such as preoperative simulation (n = 1).
Table 2. Derivation of Costs (Costs Converted to USD 2019 Values)
a Cost not included in the cost derivation.
The material cost was the most frequently reported cost and varied from less than USD 1 to USD 146 (Reference King, Park, Christensen and Danrad15;Reference Legocki, Duffy-Peter and Scott16;Reference Liu, Gao, Du, Chen, Fu and Chen18;Reference Rankin, Giovinco, Cucher, Watts, Hurwitz and Armstrong19;Reference Scerrati, Trovalusci, Albanese, Ponticelli, Tagliaferri and Sturiale21;Reference Witowski, Pedziwiatr, Major and Budzynski22). Of these studies, four involved an anatomical model, two involved surgical instruments, and one involved a simulator. The printing cost does not appear to be higher for an instrument (USD 2.99 according to Rankin et al., with the purchase cost of the printer) than for an anatomical model (up to USD 146 according to Witowski et al., in material costs only).
The second most reported cost in the studies was the staff cost (Reference Legocki, Duffy-Peter and Scott16;Reference Li, Copeland-Halperin, Kaminsky, Li, Lodhi and Miraliakbari17;Reference Resnick, Inverso, Wrzosek, Padwa, Kaban and Peacock20). In the study by Legocki et al. (Reference Legocki, Duffy-Peter and Scott16), the cost of a model was calculated from material costs and staff costs and was estimated at around USD 97. Resnick et al. (Reference Resnick, Inverso, Wrzosek, Padwa, Kaban and Peacock20) focused on staff costs by accurately measuring the time spent on each task and estimated the staff cost at between USD 2,916 and USD 3,114.
The cost of the operating room (OR) was taken into account in three studies (Reference King, Park, Christensen and Danrad15;Reference Li, Copeland-Halperin, Kaminsky, Li, Lodhi and Miraliakbari17;Reference Yang, Li, Li, Zhao, Wei and Zhang23). The definition of what this cost included was not clearly stated in the included studies. For King et al. (Reference King, Park, Christensen and Danrad15), this cost was based on the OR cost/min in their institution (USD 108/min). For Li et al. (Reference Li, Copeland-Halperin, Kaminsky, Li, Lodhi and Miraliakbari17) this cost was USD 4,695/hr and included anesthesia and all related costs. In the Yang study (Reference Yang, Li, Li, Zhao, Wei and Zhang23), this OR cost was included in the hospital costs without any further detail.
The purchase and maintenance of the 3D printer are mentioned in four studies but taken into account in the cost calculation in only two studies (Reference King, Park, Christensen and Danrad15–Reference Li, Copeland-Halperin, Kaminsky, Li, Lodhi and Miraliakbari17;Reference Rankin, Giovinco, Cucher, Watts, Hurwitz and Armstrong19). This cost of printer purchasing varied between USD 2,285 and USD 6,706.
Three studies compared the costs between standard surgery and surgery with a 3D printer (Reference King, Park, Christensen and Danrad15;Reference Li, Copeland-Halperin, Kaminsky, Li, Lodhi and Miraliakbari17;Reference Yang, Li, Li, Zhao, Wei and Zhang23). Two studies concluded that the use of 3D printing reduces costs by reducing operating time (Reference King, Park, Christensen and Danrad15;Reference Li, Copeland-Halperin, Kaminsky, Li, Lodhi and Miraliakbari17).
Three studies also compared the costs of internalized and externalized production (Reference Legocki, Duffy-Peter and Scott16;Reference Li, Copeland-Halperin, Kaminsky, Li, Lodhi and Miraliakbari17;Reference Resnick, Inverso, Wrzosek, Padwa, Kaban and Peacock20). Resnick et al. (Reference Resnick, Inverso, Wrzosek, Padwa, Kaban and Peacock20) compared the production of a surgical template by outsourced 3D printing with an internalized manual production. Only staff time was taken into account in the cost calculation and the authors found a significant difference in favor of outsourced production. Li et al. (Reference Li, Copeland-Halperin, Kaminsky, Li, Lodhi and Miraliakbari17) determined that there is an economic interest in internalized production when more than twenty-seven cases are performed per year.
Discussion
In the early years of the development of 3D printing technology in the medical field, the additional costs associated with this innovative technology were prohibitive for many users. With the standardization of this technology, a cost assessment has become essential to determine its possible integration into patient care. With this in mind, we were somewhat surprised to find so few studies on this topic and that most were of poor quality.
Of the nine studies, four have as their primary or secondary objective the evaluation of these costs. In other cases, the costs are only additional data, briefly evaluated. This literature review thus highlights a lack of robustness in the methodology of these studies.
In total, in these nine studies, only nine items of expenditure were evaluated. No comparison is possible among these studies because the costs taken into account are too heterogeneous, ranging from the simple cost of the material to the labor costs. If we only consider the material cost, the printing cost of a device is low—less than USD 150 (Reference King, Park, Christensen and Danrad15–Reference Li, Copeland-Halperin, Kaminsky, Li, Lodhi and Miraliakbari17;Reference Rankin, Giovinco, Cucher, Watts, Hurwitz and Armstrong19;Reference Scerrati, Trovalusci, Albanese, Ponticelli, Tagliaferri and Sturiale21;Reference Witowski, Pedziwiatr, Major and Budzynski22). The cost differences can be explained by the type of material used and the amount of material required, depending on the complexity of the device to be printed (Reference Witowski, Pedziwiatr, Major and Budzynski22). On the other hand, as soon as equipment and staff costs are considered, the cost is much higher and can reach several thousand dollars. It appears that certain items of expenditure, such as the purchase of a printer or a software subscription, are almost never taken into account, even though they can have a significant impact on the cost of a device.
Similarly, modification of logistics may generate additional costs related to the need to deploy or increase specific activities in the healthcare facility, such as sterilization or virtual preoperative planning. Thus, the personal time spent on these new activities must be taken into account in the economic evaluation. This is the purpose of the study conducted by Resnick et al. (Reference Resnick, Inverso, Wrzosek, Padwa, Kaban and Peacock20). They measured the time spent at each stage of the production cycle and deducted the costs in terms of staff. In their center, virtual planning with 3D printing was less expensive than the traditional manual method. It, therefore, seems very important to clearly define the production cycle and the people involved to obtain an accurate cost evaluation. However, of the nine studies included, only two cite at least one of the professionals involved, and only one details the role of each of them (Reference Liu, Gao, Du, Chen, Fu and Chen18;Reference Resnick, Inverso, Wrzosek, Padwa, Kaban and Peacock20;Reference Witowski, Pedziwiatr, Major and Budzynski22).
Initially, the purchase of a 3D printer, whose cost could exceed USD 500,000, was not an option for healthcare facilities, which then had to turn to third-party providers. Since these suppliers charge not only for the purchase of a medical device but also for the associated services, the cost of a medical device varies from a few hundred dollars to more than USD 20,000, depending on the complexity of the device to be printed. With the advent of low-cost printers, internalized production has increased and therefore makes it possible to obtain medical devices at a lower cost. In the present literature review, three studies compare the costs of internalized versus outsourced production (Reference Legocki, Duffy-Peter and Scott16;Reference Li, Copeland-Halperin, Kaminsky, Li, Lodhi and Miraliakbari17;Reference Resnick, Inverso, Wrzosek, Padwa, Kaban and Peacock20). Legocki et al. (Reference Legocki, Duffy-Peter and Scott16) estimated the cost of their model at less than USD 100, while the same device purchased commercially would cost more than USD 2,000. However, this study does not take into account the purchase and maintenance costs of a 3D printer or labor costs. Li et al. (Reference Li, Copeland-Halperin, Kaminsky, Li, Lodhi and Miraliakbari17) show that by taking maintenance costs into account, centers that have little use of 3D printing technology will have a greater incentive to turn to an external supplier.
The studies included in this review focus mainly on the extra costs associated with 3D printing use. Three studies evaluate the potential savings generated by this technology (Reference King, Park, Christensen and Danrad15;Reference Li, Copeland-Halperin, Kaminsky, Li, Lodhi and Miraliakbari17;Reference Resnick, Inverso, Wrzosek, Padwa, Kaban and Peacock20). King et al. (Reference King, Park, Christensen and Danrad15) estimated a cost savings of USD 1,773 per patient, only based on OR time reduction, but they did not factor in the production cost of the device. According to Li et al. (Reference Li, Copeland-Halperin, Kaminsky, Li, Lodhi and Miraliakbari17), based on a 2-hr reduction in operating time and four cases per year, the use of 3D printing would save USD 4,195 with in-house computer-aided surgical simulation and USD 21,122 with commercial production. Taking into account 3D printer purchase and maintenance costs, the authors estimate that twenty-seven cases per year must be carried out in order to have an economic interest in internalizing production. Ballard et al. (Reference Ballard, Mills, Duszak, Weisman, Rybicki and Woodard24) also estimated from a literature review the savings achieved through reduced operating time thanks to 3D printing technology and calculated a saving ranging from USD 1,835 to USD 11,094 per surgical case for an anatomical model and from USD 681 to USD 4,115 per surgical case for a PSI. They also estimate that sixty-three models and/or PSIs are the minimum number of cases required per year to have an economic interest in internalizing production.
We identified several reasons that may explain why the evaluated costs vary so much between the included studies. First, although most of the 3D printed devices were anatomical models, they were designed for different purposes in various surgical domains. For example, it is difficult to draw general conclusions from a cost comparison between an anatomical model of a liver and an anatomical model of a jaw. Second, material and staff costs are obviously not the same in the various countries in which the evaluations took place. In addition, we showed that the total costs taken into account ranged from the simple cost of the material to labor costs. This does not allow a direct comparison between studies. Third, we showed that three studies compared internalized and outsourced production of 3D objects; these studies highlighted a significant difference in costs between the two options. These data are rather informative, underlining that the method of production chosen greatly influences the costs. This also explains the variability in the costs observed between studies with differing methods of production.
In light of the present study, there is no economic evidence that clearly shows the cost-effectiveness of 3D printing in surgery. In certain conditions, some studies showed the potential savings that can be made using the technology, but these are very context-dependent models and the generalizability of these results seems limited. Despite the limitations of the studies included, these findings increase the knowledge of the economic impact of 3D printing in surgery. We think that this information is valuable for decision-makers, especially in a hospital setting, because it highlights the sources of the costs of this technology and could help to anticipate and plan the introduction of 3D printing. With ever-growing demands for innovative and costly technologies such as 3D printing, hospital-based health technology assessment seems essential for guiding decisions and helping hospital managers to select the best strategies for their healthcare facility (Reference Sampietro-Colom, Lach, Pasternack, Wasserfallen, Cicchetti and Marchetti25). In addition, the present study also underlines the dramatic organizational impacts of such technology. From our point of view, economic evaluations are not sufficient to fully capture the potential impact of 3D printing, and specific organizational evaluations are needed to understand the multidimensional aspects involved in implementing this technology (Reference Roussel, Carbonneil and Audry26).
The present work has some limitations that should be highlighted. First, the small number of studies included in the review does not enable us to draw strong conclusions concerning the economic evaluation of 3D printing in surgery. Second, three different databases were used to perform the systematic review, but it is possible that some studies on this topic may have been published in data sources other than scientific journals. We only focused here on articles published in scientific journals and did not include studies from grey literature, because we expected to collect data of good quality. Third, we used the CHEERS checklist to assess the quality of the articles retrieved, but some of the included studies cannot be considered full economic evaluations. Consequently, the CHEERS checklist may not have been a suitable instrument for checking all the articles included. Nevertheless, we thought that, as a reporting guideline, this checklist was the most appropriate tool that we could use here. Finally, our review was also limited by the relatively poor quality of the articles retrieved.
Conclusion
This literature review highlights the lack of reliable economic data on 3D printing technology. In addition, no studies involving implants were found in the literature, although they are widely used, particularly in maxillofacial surgery. Nevertheless, this review makes it possible to identify expenditures or items that will have to be taken into account in order to carry out more robust studies, such as cost-effectiveness analyses, and to collect information that is useful for decision-makers in a hospital setting. Further economic and organizational studies will be essential to determine the future of this technology in surgery, which is competing with other innovative technologies such as virtual reality (Reference Gregory, Gregory, Sledge, Allard and Mir27).
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S0266462320000331.
Acknowledgments
This research received no specific funding from any agency in the public, commercial, or not-for-profit sectors.
