2.1. Study design
Six analogue astronauts from crew 227 of the Tharsis mission (2022) at the MDRS gave written informed consent to participate in this research into bone fracture surgery. The methods were performed in accordance with relevant guidelines and regulations and approved by the hospital-faculty ethics committee of the Cliniques Universitaires Saint-Luc, Belgium (N°B403201523492).
2.1.1. Knowledge group
Astronauts were divided into 3 groups according to their educational background. The first group called “Anat” was somewhat skilled in human anatomy (studies in the medical or biomedical field but not in surgery). The second group called “Meca”, had knowledge of mechanics, stability, forces movements and constraints (civil engineering degree). The third group, “Others”, had no prior knowledge of either anatomy or mechanics. In Belgium, an experienced orthopedic and trauma surgeon performed the same surgery three times to provide a reference to serve as the “surgical control”. He was experienced with the classical Hoffmann® external fixators25–27 but not the EZExFix. He therefore received the same pre-surgery theoretical information as the astronauts would receive before performing the surgery.
2.1.2. Experimentations
On day one of the mission, the three groups first attended a quick one-hour theoretical course on indications, anatomical landmarks and steps to attach the EZExFix followed by a practical demonstration.
Subsequently, each astronaut had to perform all eight tasks one after the other, sometimes as the operator placing the EZExFix on the broken leg (referred to as “Anat 1” or “Anat 2” depending on the person in the “Anat” group), sometimes as the assistant, helping to maintain the fracture reduction. Two surgeries were performed simultaneously by two astronauts during which they had to set up the EZExFix to repair an artificial broken leg, in the most efficient and quickest way, similar to a time trial. Twelve different combinations, or blocks, were therefore needed to crossmatch all operators (Fig. 1). Each astronaut performed the surgery four times as the operator and four times as the assistant leading to a total of 24 surgeries for all six astronauts. The different groups can also be compared in terms of skills to assess the need for prior basic skills in anatomy or mechanics. The figure 1 illustrates how the study design can integrate as well as the background knowledge (name of the group), the confrontation with every different operator and three different progressive learning conditions (different colors).
2.1.3. Learning conditions
Given a fracture occurring in space could induce stress in the crew, different conditions were applied to detect possible differences in performance. Carrying out each surgery as a competitive timed trial highlighted the importance of time and potentially induced a stress on the two operators competing. The stress level was also changed during the mission by scheduling the surgeries under three different conditions (Fig. 1). Standard conditions (blue) implied that all equipment was already prepared laid out on a table. All astronauts had to carry on at least one surgery under standard conditions before performing it under stressful conditions. Stressful conditions were obtained by executing the surgery either during an extravehicular activity (EVA) wearing space suits (green) or at an unexpected moment, such as during the night or at mealtimes, with nothing prepared (yellow). For the sake of homogeneity, the figure 1 allowed to organize the time trials so that each astronaut can perform the surgeries against each member of the other two groups and so that they can perform them twice under standard conditions and two times under stress conditions (EVA and unexpected moment).
2.1.4. Operations scheduling
In summary, all operational constraints included four surgeries for each astronaut, all possible combinations of operators, two standard conditions and one of both stressful conditions for each astronaut, with at least one trial in standard conditions before any stressful condition. Yet, the analogue mission actually involved eight different scientific projects in addition to this EZExFix project. In such a context, finding a schedule that fits all the operational restrictions of all scientific projects, under shared limited time and resources, ended up with a particularly complex combinatorial problem. In order to solve it and to schedule the 12 timed trials, an artificial intelligence system called Romie28 has been used all along the mission to first, find the best possible a priori schedule of the entire mission, and then, as the mission went on, adapt and reoptimize the decisions about the remaining sols (martian days of simulation), based on what actually happened (or didn't), in order to maximize the probability of the mission success.
2.2. Fractured leg model
The bone model used to reproduce a fractured leg was that of a left tibia with cortical hard density and low cancellous bone structure, a pre-drilled intramedullary canal and a distal opening (LSH1385, Synbone SDN BHD, Malaysia). A simple oblique fracture line was created in the middle of the tibial diaphysis by using a laser and diamond bandsaw so as to always reproduce the same AO/OTA 42A2 fracture type following AO classification (Fig. 2). Soft tissues made from foam rubber sheet (22320, Komprex®, Lohmann & Rauscher, Germany) were fixed around the bone to produce the shape of the leg and were covered by a sock to mimic the skin. The distal opening was used to attach a prosthetic foot in order to provide an indication of the axis in the event of further realignment surgeries. Each surgery required a new leg model.
2.3. External fixator
The EZExFix concept considers both the mechanical stability needed to treat a long bone fracture and the ease of execution. A new unilateral biplanar EZExFix was developed in collaboration with orthopedic surgeons from the Cliniques Universitaires Saint-Luc (Brussels) and engineers from ECAM (High Industrial Institute, Brussels). This new technology’s core focus was to be a low-cost, fast and easy-to-use fixator to extend the use of this treatment to hostile environments23. Figure 3 illustrates both the parts needed to build the EZExFix and the final assembly. This device can be used to treat all types of tibial shaft fractures from the simplest to more complex and is also suitable for treatment of significant soft tissue lesions. Mechanical properties were previously validated and are similar to the Hoffmann® 3 fixator (Stryker Trauma AG, Selzach, Switzerland)24.
A quick guide containing the key steps needed to place the EZExFix was written to help astronauts carry out their tasks (Supplementary figure 1). Two EZExFix were created to allow face to face surgeries. Soft tissues can then be removed from the bone while keeping in place the EZExFix, which can be disassembled after data harvesting and reused for subsequent surgeries.
2.4. Analysis parameters
2.4.1. Safety
Patient safety is paramount and astronaut operators must not damage blood vessels, nerves or tendons of their crewmates. Safe zones, described for the classical external fixation technique29,30 were assessed with a scorecard including five analysis criteria (Supplementary figure 2) that were recorded during each experiment and for each operator. These safe zones are the tibial crest and the anteromedial side of the tibia, as well as the pins position with respect to joints and fracture lines. Range of movement has to be preserved; this was done by not directly inserting pins into the joint whilst also avoiding any areas that are required to move. For this reason, pins could only be inserted in the bone diaphysis. The depth of pins was also evaluated because an over-screwing by 5 mm can damage structures behind the second cortex, such as blood vessels or nerves. Each safe zone was registered as correct or not, and the sum of safe zones respected for each operation was calculated.
2.4.2. Procedure steps
To devise an efficient set up, steps had to be performed in a specific order (Supplementary figure 1). Four main steps were determined, and sub-criteria for each step were established in order to determine the failure or success of each step. Since the EZExFix is limited to certain degrees of freedom, the triangulation and the frame had to be built in the first step. Then, the correct positioning of the EZExFix on the broken leg was essential in order to check the pins’ orientations, the respect of safe zones and the absence of compression point on the skin. The width of the incision in the skin, and the order in which pins were inserted were evaluated in the third step. The total number of pins and the stability of fracture reduction were assessed in the fourth step. Skin compression and stability were both defined as main criteria directly influencing patient comorbidities and outcomes. These criteria were considered as a condition sine qua non to ensure a healthy evolution of the broken leg. The scorecard (Supplementary figure 2) allowed these steps to be validated and to identify potential mistakes or shortcomings. Each step was defined as a success or a failure, and the sum of step outcomes was calculated for each surgery.
2.4.3. Time
The total time taken by each operator to complete the task was noted on the scorecard, as well as intermediate times for each step, expressed in minutes. A video of each timed trial was recorded in parallel if needed.
2.5. Statistical analysis
Quantitative variables were analyzed in terms of central tendency (mean and median) and dispersion by the range (minimum-maximum) due to the small sample size and normality not always respected. Comparisons between educational background and learning conditions were evaluated by one-way ANOVA for a quantitative variable (time). The homogeneity of variances was examined by Levene’s test. The same comparisons were evaluated by a Poisson regression for discrete variables (safe zones and steps) in order to detect a main effect and/or an interaction between both factors. The absence of overdispersion was verified by a Chi-squared test. In order to compare the different durations of each step within the astronaut group, a one-way ANOVA for repeated measures was performed. The sphericity of the variance-covariance matrix was evaluated by Mauchly’s test. The normality of residuals was verified by QQ plots for both ANOVA tests. All generalized linear models integrated multiple comparisons if justified, adjusted with a Bonferroni correction. The time comparison of each step between astronauts and the surgical control was performed with non-parametric two-tailed Mann-Whitney tests because of the small number of samples. The level of significance was always set to 0.05 in order to reject the null hypothesis. All statistics were performed using SPSS software (V.27, IBM SPSS, Inc., Chicago, IL, USA).