Exploring the cost-effectiveness of 3-D-guided versus traditional 2-D corrective osteotomy in patients with a distal radius malunion

Abstract

Introduction

Distal radius malunion is a frequent complication after a distal radius fracture, impairing function and quality of life. Corrective osteotomy using conventional two-dimensional (2-D) planning often fails to achieve optimal alignment, leading to residual symptoms and revision surgery. Growing evidence shows that three-dimensional (3-D) planning with patient-specific surgical guides has better radiographic and clinical outcomes and fewer complications, but at increased upfront costs.

Methods

We developed a health state transition model to assess the cost-effectiveness of 3-D versus 2-D based surgery in the Netherlands over a five-year time horizon from a societal perspective. Clinical outcomes of 3-D, including transition probabilities, were obtained from a clinical cohort, whereas literature informed 2-D transition probabilities. Literature was used to obtain costs, and utility values. Outcomes of interests were the number of revisions, quality-adjusted life years (QALYs) and societal costs in each strategy. All 3-D analyses and guide development were performed in our in-house 3-D lab.

Results

Notably, revisions performed for nonunion or implant failure were reduced by 179 per 1,000 patients in the 3-D strategy. Overall, 3-D yielded slightly higher QALYs (+ 0.007) and lower costs (-€219), primarily due to fewer revisions performed for nonunion or implant failure and reduced productivity losses.

Conclusion

Using this model, 3-D-guided corrective osteotomy (supported by an in-house 3-D lab) is anticipated to be cost-effective versus 2-D-planned corrective osteotomy from a Dutch societal perspective; it potentially yields a modest improvement in health outcomes alongside a slight reduction in costs. However further clinical validation, including prospective collection of clinical outcome data, is needed to confirm these findings in real-world practice.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12962-026-00756-7.

Key points

Three-dimensional planning with patient-specific surgical guides for correcting distal radius malunion may lead to slightly better patient health outcomes while also reducing overall societal costs compared with conventional two-dimensional planning. The main benefits come from substantially fewer revision surgeries and lower productivity losses, despite higher upfront planning costs. Overall, this approach appears to be a cost-effective alternative, although further clinical validation is still needed.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12962-026-00756-7.

Introduction

Distal radial fractures are among the most common upper extremity injuries, accounting for 8–17% of all extremity fractures and up to 75% of all forearm fractures [1, 2]. Malunion is a frequent complication following both conservative and surgical treatment, with an incidence of approximately 17% [1–6]. It significantly impacts function, causing wrist pain, reduced grip strength, limited motion, and deformity [2–9]. Furthermore, malunion may predispose patients to early-onset osteoarthritis and may have socio-economic consequences due to work limitations [2–4, 6, 9].

Corrective osteotomy is the standard treatment for a symptomatic malunion, traditionally planned using conventional two-dimensional radiographs (2-D) to assess deformities and guide surgical correction [1–5, 7–9]. However, this 2-D planning is limited in addressing complex, multi-planar deformities, resulting in the intended radiological outcome in only 40% of cases [7]. These limitations can lead to residual pain and functional impairment, potentially leading to revision surgery [6, 8]. The latter severely impacts both quality of life (QoL) of the patient as well as costs for both hospital and society.

Advancements in three-dimensional (3-D) computer-assisted planning and usage of patient-specific surgical guides (PSG) for the intraoperative drilling and sawing have demonstrated potential to improve surgical accuracy and patient outcomes [6–9]. Studies report large improvements after this type of surgery compared to traditional 2-D-planned surgery in terms of a reduction in complication rates from 42% to 16% and better patient-reported outcomes [4–7, 9–12]. Notably, 96% of patients achieve the intended radiographic alignment with 3-D guided surgery, compared to only 40% with conventional 2-D [4, 7, 10, 11]. Furthermore, 3-D enables treatment of more complex cases [7, 9, 13].

Despite its advantages, 3-D planned and guided surgery incurs additional costs related to software licenses and PSG development. These costs must be weighed against potential improvements in health outcomes and/or economic benefits, such as shorter surgical durations, reduced complications rate, and, consequently, an improvement in QoL [4, 9, 14]. This study aims to perform an early cost-effectiveness analysis of 3-D-planned and guided corrective osteotomy versus the traditional 2-D technique for the treatment of distal radius malunions in the Netherlands.

Methods

A health state transition model was developed in R Statistical Software (v4.3.3, R Foundation for Statistical Computing, Vienna, Austria) to assess the cost-effectiveness of 3-D-guided corrective osteotomy (hereafter 3-D) compared to 2-D-planned corrective osteotomy (hereafter 2-D) from a Dutch societal perspective. In this model, simulated individuals move between health states at discrete time intervals (cycles) for a pre-specified period of time (time horizon) based on transition probabilities. Each health state is associated with health outcomes and costs, allowing to evaluate and compare 3-D versus 2-D.

In this analysis, a time horizon of five years was used with three-monthly cycles. The outcomes of interest were the number of patients who underwent a revision or hardware removal, quality-adjusted life-years (QALYs), costs per strategy (2-D or 3-D), and the incremental cost-effectiveness ratio (ICER) per QALY gained for the 3-D versus the 2-D strategy. QALYs were calculated as the product of survival and QoL. QoL was valued on a scale from 0 to 1, with 0 representing death and 1 representing perfect health. Hence, one QALY represents one year in perfect health. The ICER was calculated by dividing the difference in costs by the difference in QALYs of the 3-D strategy compared to the 2-D strategy. The analysis was performed using a template code designed for developing time-dependent cohort health state transition models [15]. This study has been approved by the Ethics Committee of the University of Twente (nr. 240188) and Institutional Review Board OCON (OCON2025001) and is reported according to the Consolidated Health Economic Evaluation Reporting Standards (CHEERS) guidelines 2022 [16].

Model structure

The health state transition model contained eleven health states and six tunnel states (see Fig. 1). The tunnel states were incorporated to implement time-dependent transition probabilities and are represented in Fig. 1 with dashed lines. Patients could not remain in tunnel states for more than one cycle.

Fig. 1.

Health state transition model structure. Transitions in red differ between the two strategies in the base-case analysis. Transitions in blue are equal for both strategies in the base-case analysis, but were adjusted in one of the scenario analyses, resulting in differences between the two strategies in this analysis

A simulated cohort of 1,000 patients entered the model in either the 2-D or 3-D state. After surgery the patients either experience ‘Acceptable pain and functional Disability’ (AD) or ‘Not-acceptable pain and functional Disability’ (ND). The cut-off value between acceptable and not-acceptable pain and functional disability was based on the patient rated wrist/hand evaluation (PRWHE) score [17]. In this study, a PRWHE score of 30 or lower indicated acceptable pain and functional disability [17].

Revision surgery was categorised into two types. Revision type 1 (R1) is a revision operation due to nonunion and/or implant failure. Such revisions do not require additional preoperative planning, and were assumed to be treated the same across both strategies. Revision type 2 (R2) is a revision in patients who remained in the ND health state after the initial surgery due to insufficient correction of the malunion after the first procedure. For this type of revision, a new 3-D analysis was conducted and new 3-D guides were developed in the 3-D strategy; in the 2-D strategy the surgery was prepared and performed similar to the initial surgery.

Patients from both groups who were in the ND health state due to symptomatic hardware, underwent hardware removal (HR). Patients could only stay one cycle in a revision or HR state. After a revision or HR, patients transitioned to one of the pain and functional disability states. Patients could stay in either AD or ND or transition between these two states. Transition to the death state was possible from all health states.

Costs were discounted with 3.0% and QALYs with 1.5% annually according to Dutch guidelines for health economic evaluations [18]. The model structure was based on the following assumptions after discussions with clinical experts (AV, CS, JoH):

1. patients could not undergo a revision after HR since it was assumed that all patients who undergo hardware removal only have complaints of the hardware, but no residual complaints of the malunion itself.

2. the probability of transitioning to R1 did not depend on the pain and functional disability of the patient;

3. patients with AD could only undergo R1;

4. patients could undergo a maximum of one revision;

5. the probabilities of transitioning between the pain and functional disability health states remained constant from 3 to 12 months post-surgery;

6. when patients experienced AD after HR, patients could not transition back to ND;

7. 12 months after initial surgery or revision, patients were assumed stable and could no longer transition between the health states.

Model inputs

Model inputs were based on data collected at OCON and from literature [6, 18–23].

Transition probabilities

Patient-level PRWHE scores and data collected at OCON informed the following transition probabilities:

1. from 3-D-guided surgery to the AD and ND health states;

2. between the AD and ND health states;

3. from AD and ND to the revisions and HR health states. (Table 1).

Table 1.

Model inputs

	Value		95% CI		Distribution	Reference
3-monthly health state transition probabilities – 3-D
Surgery to AD	0.07		0.01–0.19		Beta (α = 3, β = 39)	Internal data
AD to ND cycle 1, 2 & 6 till 12	0		Fixed		Fixed	Internal data
AD to ND cycle 3, 4 & 5	0.06		0.01–0.11		Bootstrap	Internal data
ND to AD cycle 1 & 6 till 12	0		Fixed		Fixed	Internal data
ND to AD cycle 2	0.51		0.35–0.68		Beta (α = 20, β = 19)	Internal data
ND to AD cycle 3, 4 & 5	0.19		0.11–0.28		Bootstrap	Internal data
AD to revision 1	0.01		0.003–0.02		Bootstrap	Internal data
ND to revision 1	0.01		0.003–0.02		Bootstrap	Internal data
ND to revision 2	0.01		0.002–0.02		Bootstrap	Internal data
ND to hardware removal	0.03		0.02–0.04		Bootstrap	Internal data
Revision 1 to AD	0.55		0.39–0.70		Beta (α = 23, β = 19)	Internal data
Revision 2 to AD	0.55		0.39–0.70		Beta (α = 23, β = 19)	Internal data
Hardware removal to AD	0.55		0.39–0.70		Beta (α = 23, β = 19)	Internal data
Mortality rate
Age 44	0.0012		Fixed		Fixed	[20]
Age 45	0.00122		Fixed		Fixed	[20]
Age 46	0.00133		Fixed		Fixed	[20]
Age 47	0.00143		Fixed		Fixed	[20]
Age 48	0.00134		Fixed		Fixed	[20]
Utilities
Utility AD	0.81		0.79–0.83		Beta (α = 48, β = 11)	Internal data
Utility ND	0.58		0.54–0.61		Beta (α = 25, β = 18)	Internal data
Costs
CT	€195.21		Fixed		Fixed	[22]
X-ray	€85.32		Fixed		Fixed	[22]
Outpatient visit	€124.61		Fixed		Fixed	[22]
Day outpatient care	€347.86		Fixed		Fixed	[22]
One minute of OR	€11.52		Fixed		Fixed	[22]
Anaesthesia	€132.08		Fixed		Fixed	Internal
OR materials	€83.71		Fixed		Fixed	Internal
Plate and screws	€666.17		Fixed		Fixed	Internal
OR instruments	€250		Fixed		Fixed	Internal
Cast + brace	€100.95		Fixed		Fixed	Internal
Physiotherapy appointment	€40.38		Fixed		Fixed	[22]
Software licensing costs per patient	€231.15		Fixed		Fixed	Internal
Printing 3-D guides and models	€100		Fixed		Fixed	Internal
Multidisciplinary team meeting 3-D	€42.94		Fixed		Fixed	Internal
Hourly wage technical physician	€55.85		Fixed		Fixed	[23]
Travel costs per kilometre	€0.27		Fixed		Fixed	[22]
Parking at hospital	€4.07		Fixed		Fixed	[22]
Productivity costs per hour	€41.41		Fixed		Fixed	[22]
Hourly wage surgeon	€115.91		Fixed		Fixed	[23]
Resources
Number of X-rays	2		Fixed		Fixed	Internal
Number of outpatient visits for initial surgery	2		Fixed		Fixed	Internal
Surgery duration 3-D (minutes)	81.55		73.60–89.50		Gamma	Internal
Surgery duration 2-D (minutes)	96.90		89.07-104.73		Normal	[5]
Surgery duration revision 1 (minutes)	71.13		36.18-105.31		Weibull	Internal
Surgery duration Hardware removal (minutes)	33.91		25.66–42.32		Weibull	Internal
Number of physiotherapy appointments	10		8–12		Triangle (8, 10, 12)	Internal
Number of hours 3-D analysis + guides	10		± 20%a		Normal (µ = 10, σ2 = 2)	Internal
Average number of kilometres to hospital	7.1		Fixed		Fixed	[22]
Average number of kilometres to physiotherapy	2.2		Fixed		Fixed	[22]
Number of hospital visits for initial surgery	5		Fixed		Fixed	Internal
Number of weeks absent from work following initial surgery or revision surgery	8		4–12		Triangle (4, 8, 12)	[26]
Number of hours worked during a week	23.50		Fixed		Fixed	[19, 21]
Number of hours 2-D planning	0.5		Fixed		Fixed	Internal
Relative risks 2-D
Revision 1 b	2.31		0.92–5.80		Log normal (µ = 0.10, σ2 = 0.37)	[6]
Revision 2 c	1.12		0.43–2.99		Log normal (µ = 0.84, σ2 = 0.44)	[6]
Hardware removal d	1.10		0.48–2.50		Log normal (µ = 0.12, σ2 = 0.44)	[6]
Annual discount rates
Annual discount rate costs	0.030		Fixed		Fixed	[18]
Annual discount rate effects	0.015		Fixed		Fixed	[18]

CI confidence interval, AD acceptable pain and functional disability, ND not-acceptable pain and functional disability, OR operating room, CT computed tomography

^a Sensitivity values 20% lower and 20% higher compared to base-case value

^b Relative risk for 2-D strategy of undergoing revision type 1 from all possible health states compared to 3-D strategy

^c Relative risk for 2-D strategy of undergoing revision type 2 from not-acceptable pain and functional disability compared to 3-D strategy

^d Relative risk for 2-D strategy of undergoing hardware removal from all possible health states compared to 3-D strategy

^e Relative risk for 2-D strategy of transitioning from not-acceptable pain and functional disability, revision or hardware removal to acceptable pain and functional disability in the first 3 months after surgery compared to 3-D strategy

^f Relative risk for 2-D strategy of transitioning from acceptable pain and functional disability, revision or hardware removal to not-acceptable pain and functional disability in the first 3 months after surgery compared to 3-D strategy

^g Relative risk for 2-D strategy of transitioning from not-acceptable pain and functional disability to acceptable pain and functional disability in months 3 to 12 after surgery compared to 3-D strategy

^h Relative risk for 2-D strategy of transitioning from acceptable pain and functional disability to not- acceptable pain and functional disability in months 3 to 12 after surgery compared to 3-D strategy

Only data from patients who had surgery more than one year ago were included in the data analysis. Age- and gender-specific mortality probabilities, reflecting the average age of patients at baseline of 44 years, and 76% being female, were derived from the age-specific Dutch life tables from 2022 [20].

3-D-guided transition probabilities were adjusted to 2-D-planned transition probabilities using relative risks (Table 1). In the base-case analysis, relative risks of undergoing a revision or HR were calculated from the number of complications in our 3-D cohort, for 2-D surgery this number was obtained from a systematic review by Meesters et al. [6].

Health effects

Two types of health effects were examined during this study: (1) the number of revisions or HRs in each strategy and (2) total QALYs. QoL weights for each health state were calculated using a mapping function that computes 12-month postoperative utility values based on the baseline EuroQol-5 dimensions (EQ-5D) score and 12-month postoperative PRWHE scores [24] (National Institute for Health and Care Excellence, 2016). The baseline EQ-5D score required for this mapping function was based on literature since this was not measured at OCON [8]. The mean utility value of patients with a 12-month postoperative PRWHE score of 30 or higher was applied to the ND, revisions, and HR health states. The mean utility value of patients with a 12-month postoperative PRWHE score under 30 was applied to the AD health state and tunnel states (Table 1). Total QALYs were calculated by multiplying the time spent in each health state by its utility value, then summing the products across all health states and over all cycles.

Costs

Three types of costs were incorporated:

1. healthcare sector costs, which included all medical expenses related to the malunion treatment. This included costs for hospital visits, diagnostic tests, surgery (initial, revisions and HR), medication, cast, brace, and outpatient clinical visits (surgeon and/or hand therapist).

2. costs borne by patients and family, which included travel expenses to the hospital and to hand therapy;

3. costs in other sectors, which included costs due to loss of productivity [22].

Volumes of healthcare consumptions were collected from OCON, and accompanying unit prices were retrieved from the Dutch guideline for health economic evaluations or from OCON when not available from the guideline (Table 1) [22, 25]. Unit prices have been multiplied by strategy-specific volumes to calculate costs.

Costs for R1 and HR were assumed equal for the 2-D and 3-D strategies. Costs of R2 included a new 3-D analysis and development of 3-D guides in the 3-D strategy and the standard 2-D preparation in the 2-D strategy.

Operating room (OR) costs were calculated based on the duration of the different types of surgery [22]. Surgery durations for 3-D, R1 and HR were obtained from OCON. The duration of 2-D surgery was obtained from literature and was found to be 96.9 min (SD: 32.3 min, SEM: 7.8), which was higher than the duration of 3-D surgery at OCON of 81.55 min (SD: 4.05 min) [5]. The duration for R2 surgery was assumed to be the same as that of the initial surgery.

At OCON, 3-D analysis and guide design are performed in-house in our 3D lab by a medical technician using the Mimics software (version 18, Materialise, Leuven, Belgium) and thereafter discussed with the surgeon. Printing of the surgical guides is performed at an external company due to European regulations (CE/ISO). Therefore, the additional costs for the 3-D analysis included licensing fees for the 3-D software required for planning, printing of the 3-D anatomic models and surgical guides, as well as time for preparation by the technician and multidisciplinary team meeting. Expenses for 2-D planning were calculated based on the assumption that a surgeon takes 30 min to plan the surgery, with no additional software required for this planning (Table 1).

Travel expenses were calculated based on Dutch guidelines for health economic evaluations and incorporated in the surgery, revision, and HR health states based on the number of hospital and hand therapy visits during the entire recovery period [18].

Productivity losses caused by the initial surgery and revisions were calculated using the friction costs method, according to Dutch guidelines for health economic evaluations [18]. After surgery, patients were assumed to be absent from work for eight weeks [26]. Costs due to productivity losses were calculated by multiplying this eight-week absence from work with the productivity costs per hour, employment rate and average number of hours worked per week in the Netherlands in 2023 [18, 19, 21, 26]. Hardware removal was assumed not to result in productivity losses.

All costs valued in another year than 2023 were converted to the price level of 2023 using Dutch consumer price indexes [27].

Analyses

Primary outcomes were the total costs and QALYs of each strategy, and the ICER per QALY gained of the 3-D strategy compared to the 2-D strategy. Secondary outcomes were the number of patients undergoing each type of revision or HR, per 1,000 patients.

Base-case results were obtained from a probabilistic analysis, which allows to capture the impact of joint parameter uncertainty on model outcomes. A Monte Carlo simulation with 10,000 iterations, using pre-assigned distributions for each parameter was conducted and presented in an incremental cost-effectiveness plane. Beta distributions were used for the transition probabilities between surgeries and the ND and AD health states and utility values, lognormal distributions for relative risks, and triangular and normal distributions for resource use (Table 1). Number of revisions and time to revisions were bootstrapped 10,000 times. Subsequently, these numbers of revisions and times were used to calculate the rates which were converted into 3-monthly probabilities. A combination of these 10,000 bootstrapped probabilities was randomly selected for each iteration of the probabilistic analysis. Probabilistic parameters for surgery durations were estimated using parametric distributions (Table S1).

The probability of 3-D being cost-effective compared with 2-D-surgery as a function of the willingness-to-pay (WTP) threshold, ranging from €0/QALY to €100,000/QALY, was displayed in the cost-effectiveness acceptability curve. The WTP threshold reflects the societal value of a QALY.

Deterministic one-way sensitivity analyses (DSA) were performed to assess the impact of individual parameter uncertainty on the incremental net monetary benefit (iNMB). The iNMB converts incremental QALYs and incremental costs into a single monetary value, using a WTP threshold. A positive iNMB indicates that the 3-D strategy is cost-effective compared to the 2-D strategy. For this analysis, a WTP threshold of €20,000/QALY was applied, as the long-term burden of disease of malunion treatment is expected to be limited, making this lower threshold appropriate. Parameters were varied using their 95% confidence intervals or a 20% standard error assumption when no 95% confidence interval could be calculated, according to Dutch guidelines for health economic evaluations (Table 1) [18]. Results were presented in a tornado diagram.

Scenario analyses were performed to investigate the impact of structural uncertainty associated with important model assumptions and inputs. The following scenarios were investigated:

1. Considering a hospital perspective (including healthcare sector costs only);

2. Evaluating the effect of alternative surgery durations for the 2-D and 3-D strategies, obtained from Bauer et al. [14];

3. Incorporating the effect of disutilities associated with revisions and HR. These disutilities were obtained from literature on distal radius fractures [28] and were applied for a period of two weeks, corresponding to the duration a patient needs to wear a cast after a corrective osteotomy.

Finally, a threshold analysis was performed to determine the maximum acceptable costs of 3-D at which the iNMB would become negative, indicating that 2-D would become cost-effective. In this analysis, the 3-D strategy included the costs of the CT scan, licensing fees for the 3-D software, time needed for the 3-D analysis and printing the 3-D guides and models. Table S2 in supplemental material provides an overview of the model inputs that differ from the base-case for each scenario analysis.

The model’s structure, assumptions, inputs and outputs were face validated through discussions with clinical experts, including an orthopaedic surgeon (AV), a technical physician (CS) and a hand therapist. The results of the validation were reported following the AdViSHE checklist [29].

Results

The average total QALYs per simulated individual in the 3-D and 2-D strategy were: 3.42 (95% CI: 3.00 to 3.80) for the 3-D strategy and 3.41 (95% CI: 3.00 to 3.79) for the 2-D strategy. Corresponding average costs were €14,547 (95% CI: €10,883 to €18,190) for the 3-D strategy and €14,785 (95% CI: €10,883 to €18,978) for the 2-D strategy (Table 3).

Table 3.

Base-case and scenarios results

Scenario	Strategy	Total QALYs	Total Costs	Incremental QALYs	Incremental Costs	ICER
Base-case	Soc 2-D	3.41 (3.00 − 3.79)	€14,786 (€10,749 - €18,978)	NA	NA	NA
Base-case	Soc 3-D	3.42 (3.00–3.80)	€14,567 (€10,903 - €18,210)	0.007 (-0.022–0.032)	€-219 (€-1,324 - €686)	Dominant
Health care perspective	Soc 2-D	3.41 (3.00 − 3.79)	€ 5,173 (€ 4,281 - € 6,229)	NA	NA	NA
Health care perspective	Soc 3-D	3.42 (3–3.8)	€ 5,604 (€ 4,875 - € 6,410)	0.007 (-0.022–0.032)	€ 451 (€ -115 - €918)	€ 60,696
Alternative surgery time	Soc 2-D	3.41 (3.00 − 3.79)	€15,066 (€11,043 - €19,292)	NA	NA	NA
Alternative surgery time	Soc 3-D	3.42 (3.00–3.80)	€14,873 (€11,225 - €18,556)	0.007 (-0.022–0.032)	€-192 (€-1,299–722)	Dominant
Disutilities for revisions and hardware removal	Soc 2-D	3.41 (3.00 − 3.79)	€14,786 (€10,749 - €18,978)	NA	NA	NA
Disutilities for revisions and hardware removal	Soc3-D	3.429 (3.00 − 3.79)	€14,567 (€10,903 - €18,190)	0.008 (-0.021–0.033)	€-219 (€-1,324 - €686)	Dominant

ICER - incremental cost-effectiveness ratio, NA – not applicable, Soc 2-D – Societal costs 2-D, Soc 3D – Societal costs 3D, QALYs - quality-adjusted life-years

The disaggregated costs show that most costs were incurred by the initial surgery in both strategies (Fig. 2). Although the initial costs of the 3-D strategy were higher, the decrease in type 1 revisions in the 3-D strategy compared to the 2-D strategy (-179 per 1,000 patients, 95% CI: − 329 to -51) resulted in lower total costs (Table S3 & Fig. S1). There was no significant difference in numbers of hardware removal (-11, 95% CI: -73 to 37) or type 2 revisions (1, 95% CI: -23 to 22) (Table 2).

Fig. 2.

Disaggregated costs per strategy. HR -Hardware removal, R1 - Revision, R2 - Revision 2, Soc – Societal costs 2-D, Soc 3D – Societal costs 3D

Table 2.

Revisions per 1,000 simulated individuals over a period of 5 years

Event	Strategy	Mean	95%CI	Mean difference	Difference 95%CI
HR	Soc 2-D	165	71 to 300	NA	NA
HR	Soc 3-D	154	71 to 265	-11	-73 to 37
R1	Soc 2-D	376	124 to 641	NA	NA
R1	Soc 3-D	197	64 to 352	-179	-329 to -51
R2	Soc 2-D	51	10 to 114	NA	NA
R2	Soc 3-D	52	11 to 111	1	-23 to 22

HR -Hardware removal, NA – not applicable, R1 - Revision, R2 - Revision 2, Soc 2-D –Societal costs 2-D, Soc 3D – Societal costs 3-D

At the commonly used Dutch WTP thresholds of €20,000, €50,000, and €80,000 per QALY, the 3-D strategy had a probability of being cost effective of 70%, 74%, and 76% versus the 2-D strategy (Fig. 3).

Fig. 3.

Incremental cost-effectiveness plane (a) and cost-effectiveness acceptability curves (b)

Sensitivity and scenario analyses

The deterministic one-way sensitivity analyses showed that the three parameters that most strongly affect model outcomes are: the relative risk of experiencing a revision type 1 with the 2-D-strategy versus the 3-D-strategy, the probability of experiencing a type 1 revision from the ND health state, and the duration of 2-D surgery (Fig. S2).

In all scenario analyses, except the one considering a healthcare perspective, the 3-D-strategy resulted in higher QALYs and lower costs than the 2-D-strategy (Table 3). In the healthcare perspective, the 3-D-strategy versus the 2-D-strategy resulted in higher QALYs (incremental: 0.007, 95%CI: -0.022 to 0.032) and higher costs (incremental: €432, 95%CI: €-134 to €900). This resulted in an ICER of €61,714 per QALY gained (Table 3). Based on the threshold analyses, the 3-D-strategy did not dominate the 2-D-strategy anymore when the 3-D preparation costs exceeded €1,279.

Discussion

This study explored the cost-effectiveness of 3-D-guided corrective osteotomy compared to the traditional 2-D-planned technique as treatment for a distal radial malunion, from the Dutch societal perspective, using a health state transition model.

Results of the probabilistic analysis showed that the 3-D strategy was cost-effective in 70% of model iterations at a WTP threshold of €20,000/QALY, although the incremental QALYs (+ 0.007) and cost savings from a societal perspective (-€237) of 3-D compared to 2-D were (very) small. While initial costs were higher for 3-D, these were offset by fewer type 1 revisions, resulting in lower total costs from a societal perspective. This cost reduction was primarily driven by decreased productivity losses incurred by the original and revision surgeries. Consistent with these findings, the scenario analysis showed that the 3-D strategy was consistently favorable across all scenarios with small gains in QALYs and reductions in costs, except when a hospital perspective was applied. This suggests that the higher incremental costs associated with using 3-D technologies for these surgeries cannot be fully offset by reductions in healthcare costs due to fewer revision surgeries and hardware removals.

The threshold analysis revealed that the 3-D strategy is no longer cost-effective once the preparation and material costs of the 3-D strategy exceed €1,279. This highlights the importance of ensuring that the preparation and material costs of 3-D remain below this threshold. The main costs of 3-D guided surgery are time spent by the technician to analyse the malunion, make the surgical plan and design the guides, the software license fees and the printing of the guides. In our institution, both analysis and guide design are performed in-house by a technician in our 3-D lab. The printing of the surgical guides is outsourced to a commercial provider due to the high costs associated with acquiring a 3-D printer capable of using sterilisable materials, as well as regulatory requirements related to ISO and CE certification. In hospitals without an in-house 3-D lab, the entire process is outsourced to commercial companies, resulting in higher overall costs compared to our approach. In such settings, achieving cost-effectiveness is considerably more challenging, if not impossible. This economic consideration is one of the reasons that a growing number of hospitals are establishing their own 3-D lab, rather than relying solely on external providers.

A previously conducted economic analysis into the potential cost saving of 3-D printed anatomic models and surgical guides in orthopaedic surgery showed that, despite not being reimbursed, 3-D printing has been adopted by several hospitals due to their advantages: the 3-D models and guides enable complex procedures, provide surgeons with greater confidence, and reduce surgery duration [30]. Other studies showed that it reduces surgical time and the number of required surgeries, while increasing surgeon satisfaction, lowering risk of reinterventions and infections after initial surgery, and shortening recovery time compared to conventional 2-D treatment [31, 32]. 3-D planning has also been proven to be efficient and time saving compared to conventional 2-D treatment in other joints as well as in other specialties, like craniomaxillofacial, neuro or plastic surgery, 3-D planning has also been proven to be efficient and time saving compared to conventional treatment [33–35]. Some of these aspects are incorporated as parameters in our model, whereas confidence of the surgeon and the level of complexity are other very important issues to consider which could not be incorporated in the current analysis. These non-economic factors should be taken into serious consideration as well when choosing 3-D over 2-D surgery.

Limitations

The current study has several limitations, primarily due to the lack of available evidence to inform key parameters, including QoL, relative risks, and productivity losses.

To estimate QoL values, a mapping function was applied to convert PRWHE scores into EQ-5D scores. Additionally, baseline EQ-5D scores for both strategies were derived from another study, introducing further uncertainty. Both limitations may have led to inaccurate QoL estimates, potentially resulting in either an underestimation or overestimation of incremental QALYs.

Ideally, relative risks for undergoing a revision or hardware removal would be derived from an RCT, but instead were based on a systematic review of observational studies. As this parameter was highly influential in the sensitivity analysis, its uncertainty may have a considerable impact on the robustness of the results.

Resource use, including duration of work absence to calculate productivity losses and number of hand therapy appointments, were estimated based on conversations with clinical experts rather than individually reported. This could have led to either an overestimation or underestimation of hand therapy costs and costs due to productivity losses. The latter, in particular, can significantly impact the results, as productivity losses have a substantial effect on incremental costs.

Future research

The lack of available evidence on QoL and the relative effectiveness of 3-D versus 2-D planning stresses the need for further research. Ideally, an RCT designed to capture differences in revision rate, patient’s recovery and QoL of both strategies should be performed. In such an RCT, the documentation of revisions over a period of at least five years is crucial as the sensitivity analysis indicated the influence of the relative risks of revisions on model outcomes. Additionally, the duration of work absence for each patient, including their work status before the surgery, their capacity percentage upon return, and the timeline to achieve full capacity should be documented to calculate productivity losses. We are currently performing a comparative analysis of observational data between our 3-D data and a clinic with routinely collected outcome of 2-D-planned corrective osteotomies to obtain more precise details with regards to costs and outcomes. This future comparison is expected to reflect actual practice costs more closely.

Conclusion

From a Dutch societal perspective, 3-D-guided corrective osteotomy is anticipated to be cost-effective compared with 2-D-planned corrective osteotomy, potentially yielding modest health gains and slight cost reduction. Based on these findings, adopting 3-D-guided corrective osteotomy as the standard approach, with an in-house 3D lab likely being essential from a cost perspective. However, these conclusions remain conditional on further clinical validation, as current evidence on quality of life and relative effectiveness is limited. At present, an ongoing comparison between our own 3-D cohort and a 2-D cohort from another clinic is underway to generate more robust, real-world evidence. Confirmation of long term effectiveness and cost-effectiveness will ultimately require evidence from RCTs.

Electronic Supplementary Material

Below is the link to the electronic supplementary material.

Author contributions

AV, CS, JoH, and QT collected and analyzed the data. QT, MK, and XP developed and validated the health economic model. All authors contributed to drafting the manuscript, reviewed and edited the final version, and approved the final manuscript.

Funding

No funds, grants, or other support was received.

Data availability

Data will be available upon reasonable request.

Declarations

Ethical approval

This research study was conducted retrospectively from data obtained for clinical purposes. The study was conducted in accordance with the Declaration of Helsinki and was approved by the Ethics Committee of the University of Twente (nr. 240188) and Institutional Review Board OCON (OCON2025001), with the need for written informed consent waived.

Competing interests

The authors declare no competing interests.