Introduction

Colonoscopy is an essential diagnostic and therapeutic tool in gastroenterology for the evaluation and treatment of a wide spectrum of colorectal diseases. However, it is a complex procedure that requires skilled endoscopy professionals and may sometimes result in adverse patient experiences during its performance1,2.

In recent times, innovative handle has attracted considerable attention in the realm of endoscopy equipment research and development. The design of novel colonoscopy handles that are intended to enhance the efficiency and improve the comfort of both doctors and patients. Some examples include the integration of real-time feedback, labour-saving and artificial intelligence3,4,5. The advent of technological advances has led to the development of gastrointestinal endoscopy robots such as MASTER, EndoSAMURAI and Endoscopic Operation Robot6,7,8. These sophisticated devices facilitate more precise and accurate operations during procedures9,10.

It is an irrefutable fact that the majority of contemporary advancements in endoscopic devices are centred on the objective of reducing patient discomfort and minimising the risk of bowel injury. However, the clinical issues of reducing muscle fatigue, increasing grip comfort and reducing occupational injury for endoscopists cannot be avoided11,12,13. It is evident that endoscopists are susceptible to muscle strain and joint injury during colonoscopy14. The prevalence of musculoskeletal injuries, particularly those affecting the left thumb, has been well documented. Numerous studies have highlighted the need for enhanced handle design to address ergonomics, with specific recommendations pertaining to handle size, button position and switch design15,16. Conventional handle designs have also been associated with unintuitive steering mechanisms, which have been shown to prolong the learning curve and affect the overall operator experience17. Consequently, there is a necessity to enhance the design of colonoscopy handles with a view to reducing operator fatigue and injury.

In this study, a novel colonoscopy handle was developed. The handle was designed to facilitate hand operation and featured a rocker and buttons, replacing the conventional two-gear system. The objective of this innovation was to enhance flexibility, comfort, and efficiency. It is important to acknowledge the extensive use of surface electromyography (sEMG) signals in numerous studies to evaluate muscle load and fatigue18,19. This has provided researchers with a robust analytical instrument with which to compare muscle engagement between the novel and traditional handle. Moreover, we administered a questionnaire to assess subjective experiences of fatigue and efficiency.

Materials and methods

Equipment and consumables

The equipment and consumables for this study principally comprise a digestive endoscopy system, a human colorectal teaching model, an innovative design endoscopic handle, a sEMG data acquisition device and lubricant. The details are shown in Table 1.

Table 1 Specifications and manufacturers of equipment and consumables.
Full size table

Handle design

The assistant device is composed of two main components: a handle and a base. The base serves as the support system, featuring a metal structure that provides stability for both the handle and the doctor’s arm. The handle is designed using the "drive-by-wire" method. A comprehensive schematic representation of the colonoscope handle assist device is presented in Fig. 1. The device incorporates a base plate, a button control unit, a wheel unit, and a rocker control unit.

Fig. 1
figure 1

Structure diagram of the colonoscope handle assist device. (a) Overall three-dimensional structure diagram; (b) Three-dimensional structure diagram without the shell; (c) Left view; (d) Front view and top view of traditional colonoscope handle. The structural descriptions corresponding to the numbers are as follows:1. Remote control handle; 2. Shell; 3. Base plate; 4. Bracket; 5. Colonoscope handle; 5-1. Large wheel; 5-2. Small wheel; 5-3. Button; 6. Remote control mounting plate; 7. Support column; 8. Servo motor 1; 9. Drive wheel 1; 10. Servo motor 1 mounting plate; 11. Servo motor 2 mounting plate; 12. Servo motor 2; 13. Drive wheel 2; 14. Button control unit; 14-1. Stand plate; 15. Controller mounting plate; 16. Controller; 17. Servo motor fixed plate; 18. Top piece 1; 19. Servo motor 3; 20. Top piece 2; 21. Wheel unit; 21-1. Sleeve; 21-2. Bearing 1; 21-3. Small groove wheel; 21-4. Bearing 2; 21-5. Bearing 3; 21-6. Large groove wheel; 21-7. Bearing 4.

Full size image

The base plate functions as the foundational support for all other components. It is securely connected to a casing and a bracket, which provide protection and additional support. The button control unit is attached to the base plate and positioned on the side of the colonoscope handle, corresponding to the handle’s buttons. This configuration enables the physician to manipulate the functions of the colonoscope handle, such as the water supply and insufflation, by actuating the buttons on the control rocker.

The wheel unit is located beneath the colonoscope handle, ensuring proximity to the base plate for optimal control of the colonoscope’s movement and direction. This configuration enhances operational flexibility and convenience. The colonoscope handle is fixed within the wheel unit and can be readily detached. The wheel unit comprises two small groove wheels and two large wheels, which collaborate to facilitate unimpeded movement of the colonoscope’s distal end. This configuration is designed to enhance ease of operation and maneuverability while ensuring high stability and durability.

The rocker control unit is located at the rear of the colonoscope handle and comprises a mounting plate, a support column, and the handle itself. The support column is securely attached to the base plate, with the mounting plate fixed to the support column. The mounting plate is rectangular with a hollow circular center, allowing for straightforward placement of the handle and ensuring precise transmission of handle movements to the colonoscope. The handle is mounted on this plate, and the rocker control slider enables the doctor to adjust the colonoscope’s state, direction, and angle, facilitating precise examinations and enhancing the procedure’s flexibility and convenience. This design is highly adaptable, allowing for adjustments in the handle’s position and orientation as required. The hollow circular design of the mounting plate also reduces the device’s weight, improving stability and maneuverability. The support column is engineered to firmly secure the mounting plate to the base plate, ensuring stability during operation and preventing any unwanted movement or instability. The handle has been ergonomically designed to facilitate straightforward control of the colonoscope’s movement and operation.

Additionally, the handle incorporates a digital recording function for the operation process. The digitization of input instructions from the slider enables the handle assist device to record the operation process. This feature is advantageous for subsequent data analysis and storage of expert operating techniques.

The novel handle is produced by 3D printing technology

Based on the structure diagram in Fig. 1, 3D printing technology was used to create the novel handle. The physical diagram is shown in Fig. 2.

Fig. 2
figure 2

The prototype system of the colonoscope handle assist device under study. (a) Physical diagram of the device; (b) Application of the device for endoscopic examination in a colon model.

Full size image

Subjects participating in this study

To evaluate the functionality and ergonomics of conventional and novel colonoscopy handles, we recruited volunteers from Beijing Tsinghua Changgung Hospital affiliated with Tsinghua University. Initial recruitment included a subset of volunteers for preliminary experiments to optimize the experimental protocol. Subsequently, formal validation experiments were conducted with the finalized protocol, and the generated data were utilized for statistical analysis.

Formal study participants were stratified into two cohorts: novices and experienced endoscopists. Prior to study initiation, all participants received comprehensive briefings detailing the study objectives and methodologies. This ensured participant understanding and cooperation throughout the experimental procedures.

Task description

Experimental flow

The experimental flow of this study is divided into four sections: the first section is the training of subjects on the upper GIE training model; the second section is the debugging and installation of the sEMG acquisition device; the third section is the validation of the performance of the innovatively designed handles by subjects on the colonoscopy training model; and the fourth section is the measurement of subjects’ body dimensions and filling out of questionnaires. The detailed steps are illustrated in Fig. 3.

Fig. 3
figure 3

The four sections of the experimental process of this study.

Full size image

Endoscopic handle training

The following essay will provide a comprehensive overview of endoscopic handle training. Conventional handle training is primarily designed for novice endoscopic operators. In contrast, training on the novel colonoscopy handle is mandatory for all individuals, irrespective of their colonoscopy experience. The training protocol is detailed as follows.

Participants received standardized training from experimental staff utilizing an upper gastrointestinal model with a selected gastroscope. Training focused on three core competencies: endoscopic lens directional control, landmark anatomical position exposure, and lesion probing. Directional control training emphasized precise maneuvering of the lens in four cardinal directions. Anatomical exposure training followed a sequential protocol: esophageal intubation via the esophageal inlet (Fig. 4a), gastric body entry via the cardia (Fig. 4b), observing the fundus (Fig. 4c), exposing the angle of the stomach (Fig. 4d), and arriving at the sinus (Fig. 4e). Lesion probing training concentrated on detecting prepositioned gastric angle ulcer.

Fig. 4
figure 4

Demonstration of the endoscopic handle training process. (a) 1 point each for being able to expose the oesophageal inlet and enter the oesophagus; (b) 1 point each for being able to identify the cardia and enter the gastric body; (c) 1 point for being able to observe the gastric fundus; (d) 1 point for being able to visualise the angle of the stomach; (e) 1 point for being able to reach the gastric sinus.

Full size image

The training program was administered in 30 min sessions, followed by a structured assessment. Participants meeting established criteria advanced to validation experiments. Training program evaluation utilized a structured 12-point assessment tool. Scoring allocated 4 points for directional control maneuvers, 7 points for anatomical landmark identification, and 1 point for successful gastric angle ulcer detection. Training to standard was defined as achieving a cumulative score of ≥ 10 points.

Following the completion of the standardised training programme, the subjects assessed the effectiveness of the training programme by means of a questionnaire aimed at determining the subjects’ familiarity with the two handles prior to the experiment.

The training was approved by the Ethics Committee of Beijing Tsinghua Changgung Hospital, it was performed in accordance with relevant guidelines and regulations, and informed consent was obtained from all subjects prior to training.

Handle validation protocol

Subjects executed colonoscopy procedures on a training model utilizing both traditional and novel handle designs. The sequence of handle validation was randomized to mitigate potential order effects on participant fatigue and model familiarity.

The experimental endpoint was defined as successful cecum intubation within 750 s, followed by initiation of withdrawal maneuvers. During withdrawal, participants were instructed to manipulate the colonoscope’s anterior lens using both handle configurations to identify prepositioned lesions, including polypoid and superficial bulging lesions. To preserve experimental integrity, investigators withheld specific lesion details from participants. Subjects were permitted up to three attempts per handle configuration, with a brief resting period between failed attempts.

An experienced endoscopist meticulously documented procedural metrics, including loop formation incidence, loop reduction success rates, and lesion detection efficacy during withdrawal. Documentation adhered to standardized endoscopic reporting protocols to ensure data validity and reliability for subsequent comparative analysis.

The handle validation protocol was approved by the Ethics Committee of Beijing Tsinghua Changgung Hospital, it was performed in accordance with relevant guidelines and regulations, and informed consent was obtained from all subjects prior to implementation.

Evaluation of surface electromyography as an objective indicator

sEMG data acquisition site and equipment debugging

In addition to the objective of documenting the quality of subjects’ colonoscopy performance, the present study sought to provide an objective assessment of the level of muscle load experienced by subjects during colonoscopy using different handles. In the preliminary experiment, we measured sEMG signals from as many muscles as possible in the left upper extremity during the utilisation of two types of handles. Subsequently, we evaluated the stability of the signals recorded from various muscles during the pilot experiment and identified the primary effector muscles during actual colonoscopy operations. This process led us to select seven muscle groups as the final measurement targets (Fig. 5). These muscles comprise the left middle deltoid (M1), the left lateral head of the triceps brachii (M2), the left biceps brachii (M3), the left brachioradialis (M4), the left flexor carpi radialis (M5), the left abductor pollicis brevis (M6), and the left extensor pollicis longus (M7).

Fig. 5
figure 5

Location of left upper limb electrode paste and device debugging.

Full size image

The sEMG measurement equipment (Ultrium 8, Noraxon) was provided by Dr. Ma’s laboratory at the School of Aeronautics and Astronautics, Tsinghua University. Electrode placement references were obtained from the equipment manuals and the Seniam.org project website (http://www.seniam.org/).

Equipment calibration

Prior to data collection, the Noraxon Ultrium 8 sEMG system underwent standardized calibration procedures according to the manufacturer’s specifications. Electrode impedance was maintained below 5 kΩ through thorough skin preparation involving shaving, abrading with alcohol swabs, and applying proper electrode gel. The system was configured with a sampling rate of 1500 Hz and a bandpass filter (10–500 Hz) to optimize signal quality while minimizing movement artifacts.

Signal normalization followed established SENIAM recommendations for isometric voluntary contractions (IVC). Participants performed three 5-s maximum voluntary contractions (MVC) for each muscle group, with 2-min rest intervals between trials and real-time visual feedback provided to monitor contraction quality. The root mean square value from the central 3-s segment of each MVC trial was averaged to establish reference values. During experimental trials, sEMG signals were normalized as a percentage of MVC (%MVC) to enable inter-subject comparisons.

Signal processing included:

  1. (1)

    50 Hz notch filtering to eliminate power line interference

  2. (2)

    Full-wave rectification and linear envelope detection using a 4th-order Butterworth low-pass filter (6 Hz cutoff)

  3. (3)

    Amplitude normalization based on MVC reference values

  4. (4)

    Temporal normalization aligned with the total procedure duration

Signal-to-noise ratio (SNR) was maintained above 3 dB across all recordings. Trials with SNR < 3 dB were excluded from analysis (7 out of 83 trials). The complete processing pipeline was implemented in OriginPro 2021 software.

Methods of analysing sEMG data

Following data collection, sEMG signals from the seven groups of muscles were filtered, and the signal envelopes of sEMG were obtained using the OriginPro 2021 software. Within the software, the “Envelope Type” was designated as “Upper Envelope”, and the “Smooth Point” was calibrated to 2. The definite integral of the signal envelope was subsequently calculated to reflect the total work (W) of a muscle during a given process. The mean rate of work (P) was then determined by dividing the total work (W) within a process by the total time of the process.

MVC is widely regarded as a significant metric for evaluating the extent of muscle fatigue; however, the present study identified certain drawbacks associated with the MVC. The majority of subjects encountered difficulties in achieving their MVC for each muscle, and the exertion required for MVC led to fatigue, which may have influenced subsequent test results. To circumvent these issues, the testing discrepancy ratio (DR) emerged as a viable alternative.

The statistical analysis of sEMG is chiefly concerned with the comparison of the total work (W) and mean rate of work (P) generated by different handles during the intubation process. Additionally, the data of P for both intubation and withdrawal were aggregated for statistical analysis to capture the handle’s overall impact on muscle load throughout the entire colonoscopy operation. However, it is important to note that the data of W for intubation and withdrawal were not combined due to the inherent difference in time between these two processes, which could potentially confound the statistical analysis. Furthermore, the DRs of inexperienced and experienced subjects were analysed separately for subgroup analysis.

Subjective experience assessment

To investigate participants’ subjective experiences with different colonoscopy handles, a comprehensive questionnaire was administered to volunteers. The questionnaire was divided into four sections:

  1. (1)

    Basic Information: This section included questions about gender, age, professional title, colonoscopy procedure frequency, colonoscopy experience duration, etc.

  2. (2)

    Fatigue Self-Assessment: This segment assessed musculoskeletal fatigue experienced when operating both handle types, focusing on specific body regions including fingers, wrists, arms, shoulders, and waist.

  3. (3)

    Work Efficiency Evaluation: This section covered aspects such as learning curve, movement precision, intubation/withdrawal efficiency, and safety considerations.

  4. (4)

    Design Satisfaction: This part collected feedback on handle size, grip ergonomics, and button design.

Most questions utilized a Likert scale format. Participants rated statements such as "I feel soreness in my left wrist" on a five-point scale: 5 (Strongly agree) to 1 (Strongly disagree). Participants were also encouraged to provide open-ended feedback on their experiences with both handle designs to facilitate a comprehensive subjective assessment.

Data processing and analysis

Statistical analyses were conducted using SPSS software, with GraphPad Prism utilised for figure processing. Initially, the Shapiro–Wilk test was employed to assess the conformity of the data with parametric test assumptions. The t-test was used to compare two groups of continuous variables, such as the height and weight of subjects in distinct groups, when the data adhered to a normal distribution. Conversely, the Mann–Whitney U test was employed for continuous variables that did not adhere to a normal distribution, as exemplified by the age of subjects. In the analysis of sEMG, the selection of the Wilcoxon signed-rank test or the paired t-test was contingent on the normal distribution conformity of the data difference. The Wilcoxon signed-rank test was also deployed for the purpose of comparing feedback gathered from subjects in the questionnaire survey regarding various aspects of the novel and traditional handles, including subjective ratings of fatigue, efficiency, and related parameters. After the Wilcoxon signed-rank test, if the result shows significant difference (p < 0.05), post-hoc pairwise Wilcoxon signed-tank tests will be done to confirm validity of the p-value. Fisher’s exact test was utilised to compare categorical variables, such as subjects’ glove size, familiarity with different handles, and satisfaction with handle design. For the analysis of intubation time, a mixed-effects model was implemented to simultaneously evaluate the impact of two factors: handle type and prior colonoscopy experience.

Results

Characteristics of subjects

A total of 23 volunteers were selected to assess the functionality and ergonomics of the colonoscope handle auxiliary equipment, hereafter referred to as the novel handle. Four volunteers participated in the preliminary experiment, which enabled the refinement of the experimental protocol. Subsequently, 19 volunteers participated in the formal validation experiment.

The volunteers who participated in the formal experiment were divided into two groups: the first group consisted of novices (10 individuals), including medical students, nurses, and non-medical personnel with no previous exposure to colonoscopy procedures. The second group consisted of experienced individuals (9 individuals), specifically gastroenterologists from Beijing Tsinghua Changgung Hospital, who had independently performed endoscopic procedures on patients as part of their clinical work.

In the experiment, contributing a total of 76 processes. Each subject underwent four processes, including insertion and withdrawal of the colonoscopy using both the novel and traditional handle. The characteristics of subjects is provided in detail in Table 2.

Table 2 Demographic and anthropometric characteristics of subjects.
Full size table

Endoscopic handle training results

All subjects successfully completed and assessed to standard within the 30-min training time, which is incredible in a real clinical work environment. This may be due to the fact that the training in this study was conducted only in the upper GI model. The training scores of each subject are displayed in Table 3.

Table 3 Operational training scores for subjects in the novice and experienced groups.
Full size table

The results of the questionnaire showed that experienced subjects were more familiar with the traditional handle, while their familiarity with the novel handle was comparatively lower (33.3% vs. 66.7%) (Fig. 6). This may be attributable to the fact that the unfamiliarity of the procedure poses an inherent limitation to the adoption of novel handles by experienced endoscopists.

Fig. 6
figure 6

Subjects’ familiarity with traditional and novel handles after training. (a) The absolute number of of responses to the question “How familiar are you with the novel colonoscopy handle compared to the traditional handle?” is presented, with error bars showing means and 95% confidence intervals. (b) The percentage distribution of responses to the question “How familiar are you with the novel colonoscopy handle compared to the traditional handle?” is presented. Fisher’s exact test was utilized for statistical analysis. *p < 0.05.

Full size image

In contrast, 70% of novice subjects reported that their familiarity with the new colonoscopy handle was at least comparable to that of the traditional handle (30% “equivalent”, 20% “higher”, 20% “much higher”) (Fig. 6). This finding suggests that the novel handle is user-friendly and can be readily grasped with the theoretical knowledge imparted during training.

The work efficiency of the novel handle was comparable to that of the traditional handle

However, the observed variance in proficiency with the two handle types had no impact on the colonoscopy quality performed by the subjects. In the experienced group, the success rate, defined as the probability of reaching the ileocecal region within 750 s after scope insertion, was 100% for both types of handles (Fig. 7a). It is noteworthy that there was no statistically significant difference in the time taken to reach the ileocecal region between the two types of handles (91.47 ± 86.08 s for the traditional handle vs. 124.73 ± 124.33 s, p = 0.092) (Fig. 7b). Conversely, within the novice group, a total of 15 colonoscopies were conducted using the traditional handle, resulting in 6 successes (40.00% success rate). In contrast, the novel handle was employed on 18 occasions, resulting in 8 successes (44.44% success rate) (Fig. 7a). The success rate and time to reach the ileocecal region (574.12 ± 171.46 s for the traditional handle vs. 524.23 ± 186.23 s for the novel handle, p = 0.618) showed no significant differences (Fig. 7b). Subsequent analysis using a mixed-effects model indicated that prior colonoscopy experience (novice group vs. experienced group) significantly influenced the time taken to reach the ileocecal region. However, the type of handle utilised did not demonstrate a significant effect on this parameter (Fig. 7b).

Fig. 7
figure 7

Comparison of the work efficiency of different handles. (a) Success rate in reaching the ileocecal region of the lower gastrointestinal model within 750 s following colonoscopy insertion; (b) Time taken from the insertion of the colonoscope to reaching the ileocecal region in instances of successful operations; (c) Quality of the subjects’ operations, assessed by an experienced endoscopist; (d) The percentage distribution of responses pertaining to subjects’ subjective overall experiences of work efficiency while employing traditional and novel handles. (e) The detailed subjective evaluation regarding work efficiency/safety in different aspects. Higher scores mean greater satisfaction. For (a) and (c), Fisher’s exact test was utilized for statistical analysis. For (b), a mixed-effects model was employed to simultaneously evaluate the impact of two factors: handle type and prior colonoscopy experience. For (e), the Wilcoxon signed-rank test was used. T, the traditional handle; N, the novel handle; ns, not significant; *p < 0.05; ****p < 0.0001.

Full size image

The quality of the subjects’ operations was assessed by a physician with extensive endoscopy experience, based on the anticipated risk and actual occurrence of colonic loop formation during colonoscopy. Scores were assigned according to the following criteria: Score 5: no risk of colonic loop formation during operation; Score 4: low risk of colonic loops; Score 3: high risk of colonic loop formation, but no loops actually formed; Score 2: presence of colonic loop(s) with successful untangling; Score 1: presence of colonic loop(s) with failure to untangle. While a marked discrepancy in the overall quality of operations was evident between the experienced and novice groups due to differences in endoscopy experience, no such discrepancy was observed among subjects using different handles in either group (Fig. 7c). This finding indicates that the quality of operations with the novel handles was comparable to that achieved with the traditional handle.

Furthermore, the probability of detecting a pre-installed lesion during the withdrawal process was found to be 100% in both groups of subjects, irrespective of the handle used. This finding indicates that the novel handle is objectively comparable to the traditional handle in terms of colonoscopy operation efficiency and quality.

Subjective evaluations of work efficiency using different handles were collected via questionnaires. In the novice group, the majority of subjects (70.00%, 7 individuals) believed that utilising the novel handle would enhance work efficiency compared to the traditional handle. Conversely, only a few subjects believed that using the novel handle would result in no change (20.00%, 2 individuals) or a decrease in efficiency (10.00%, 1 individual). In contrast, within the experienced group, more than half of the subjects (55.56%, 5 individuals) expressed the opinion that the novel handle would not be as efficient as the traditional handle, one individual (11.11%) considered the two handles to be equally efficient, and three individuals (33.33%) believed that the novel handle would be more efficient. The discrepancy in the proportion of responses selected for this question between the novice and experienced groups was found to be statistically significant (Fig. 7d).

In the subsequent phase of the study, participants were asked to provide a more detailed evaluation of their experiences concerning the efficiency, convenience, and safety of the novel handle in comparison to the traditional handle. The results indicated that the novel handle received a significantly higher satisfaction score for “operational simplicity”, suggesting that the subjects found the operation of the novel handle to be more straightforward to master (Fig. 7e). Conversely, the novel handle received a significantly lower score for “preparation convenience”, indicating that the subjects perceived the novel handle as being more cumbersome to set up, potentially due to the extra step of attaching the handle to the auxiliary equipment before operation. No statistically significant differences were observed between the two handle types in the evaluation of “operational accuracy”, "efficiency of intubation", "efficiency of withdrawal", "coordination between left and right hand", and “operational safety”. In the subgroup analysis, no statistically significant differences were found between the novel handle and the traditional handle in the novice group. In the experienced group, the novel handle was found to be significantly inferior in terms of “preparation convenience” when compared to the traditional handle. However, no significant differences were observed in the scores for the other items (Fig. 7e).

The overall muscle load was reduced with the adoption of the novel handle

The data from the 76 processes were divided into two segments: the novel handle segment and the traditional handle segment. The DR was then calculated by dividing the datum in the traditional handle segment by the corresponding datum in the novel handle segment. This approach yielded a total of 38 groups of DRs, with each group comprising 14 values (i.e., DRs of total work for 7 muscles and DRs of mean rate of work for 7 muscles). However, due to the presence of invalid data from the withdrawal process with the novel handle for 3 subjects and data from the withdrawal process with the traditional handle for 4 subjects, possibly due to defective contact between the skin and electrodes, only 33 valid groups of DRs were obtained instead of the expected 38. The waveforms of the left upper limb sEMG data during the experiment are shown in Fig. 8.

Fig. 8
figure 8

Waveforms of sEMG data from seven muscle surfaces of the left upper limb.

Full size image

The sEMG data demonstrate substantial variations in the total work (W) of intubation across various muscles when operating with the novel handle in comparison to the traditional handle, upon analysis of data from all subjects. Specifically, muscles M3 (left biceps brachii), M5 (left flexor carpi radialis), and M6 (left abductor pollicis brevis) exhibited significantly lower W when using the novel handle (see Fig. 9a). Subgroup analysis further corroborates these observations, demonstrating substantial reductions in W for M3 and M6 in the novice group, and for M3 in the experienced group (see Fig. 9b and c). However, an unexpected finding emerges in the experienced group, where a significant increase in the W of M1 (left middle deltoid) is noted when using the novel handle, suggesting that experienced users exert even more effort on M1 when using the novel handle than the traditional handle, contrary to the original hypothesis (Fig. 9b).

Fig. 9
figure 9

Comparison of the sEMG signals using different handles. The sEMG results reflecting the total work (a–c) or the rate of work (d–f) of intubation are presented. The Wilcoxon signed-rank test or the paired t-test was used depending on the normal distribution conformity of the paired data difference. M1, left middle deltoid; M2, left lateral head of triceps brachii; M3, left biceps brachii; M4, left brachioradialis; M5, left flexor carpi radialis; M6, left abductor pollicis brevis; M7, extensor pollicis longus. Lines indicate data from the same subject. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.

Full size image

A similar observation was made in the analysis of the rate of work (P) during the intubation process, where significant differences were observed in muscles M3, M5, and M6, indicating less power exerted when employing the novel handle compared to the traditional one. Conversely, a significant difference is found in M1, suggesting an increased rate of work using the novel handle (Fig. 9d). These differences persist in subgroup analyses, with differences in M3 and M6 observed in both experienced and novice groups, while differences in M5 are significant only in the experienced group (Fig. 9e and f).

No obvious differences in W and P during the intubation process are observed in muscles M2 (left lateral head of the triceps brachii) and M4 (left brachioradialis) between the different handle types (Fig. 9).

In the questionnaire, subjects reported that using the novel handle for colonoscopy led to reduced overall fatigue compared to the traditional handle. A significant proportion of subjects, specifically 84.21% (16 individuals), reported a reduction in overall fatigue levels when utilising the novel handle. In contrast, only 15.79% (3 individuals, 2 from the novice group and 1 from the experienced group) reported no difference in fatigue levels between the two handles. It is noteworthy that no subjects reported increased fatigue levels with the novel handle (Fig. 10a).

Fig. 10
figure 10

Participants’ subjective assessment of fatigue level in the context of using different handles. (a) The percentage distribution of responses pertaining to subjects’ subjective overall experiences of fatigue while employing traditional and novel handles. (b) The detailed subjective evaluation regarding fatigue level in different body parts. Higher scores represent higher levels of fatigue. For (b), the Wilcoxon signed-rank test was used. *p < 0.05; **p < 0.01; ***p < 0.001.

Full size image

Furthermore, subjects provided subjective ratings of fatigue for specific body parts. The results indicated significantly lower fatigue levels in various body parts, including the fingers, wrist, forearm, upper arm, and shoulder on the side holding the handle (i.e., the left side) during the use of the novel handle compared to the traditional handle. Furthermore, a significant reduction in fatigue levels was observed in other areas, such as the neck, back, waist, legs, and overall mental state, when the novel handle was employed. However, no significant differences were observed in terms of eye discomfort between the two handles (Fig. 10b).

A subsequent subgroup analysis revealed that, among the novice group, due to the limited sample size, significant differences in fatigue levels were observed only in the left upper limb (including fingers, wrist, forearm, upper arm, and shoulder) and mental state between the two handle types. Conversely, within the experienced group, utilisation of the novel handle elicited a substantial decline in fatigue levels, particularly in the left shoulder, in contrast to the traditional handle, with no significant disparities observed elsewhere.

Subjects expressed a heightened level of satisfaction with the novel handle’s mechanical design

The mechanical aspects of a handle are of paramount importance in shaping the operator’s overall experience. The present investigation entailed a comprehensive evaluation of diverse handles, encompassing considerations such as grip comfort, dimensions, button configuration, and associated parameters. A notable finding was that the novel handle demonstrated a significantly higher level of grip comfort compared to the traditional handle. However, this difference was only statistically significant within the novice group, as revealed by subgroup analysis (Fig. 11b). Addressing the issue of prolonged usage pain, both novice and experienced groups reported a significantly higher level of discomfort when using the traditional handle for an extended period, compared to the novel handle (Fig. 11d). In summary, the novel handle was found to be more comfortable to grip than traditional handles.

Fig. 11
figure 11

Participants’ subjective evaluation of the handle design. (a) Participants’ subjective evaluation of the sizes of different handles; (b) Participants’ subjective evaluation of the grip comfort of different handles. A higher score represents higher satisfaction in the grip sensation; (c) Participants’ agreement with the statement “I think the buttons of the traditional (or novel) handle are difficult to reach”; (d) The subjects’ assessments of the pain level induced by prolonged gripping of the handles. A higher score represents a higher pain level after prolonged gripping of the handles. For (a) and (c), Fisher’s exact test was used. For (b) and (c), the Wilcoxon signed-rank test was used. T, the traditional handle; N, the novel handle; ns: not significant; *p < 0.05; **p < 0.01; ***p < 0.001.

Full size image

In evaluating handle size, a substantial proportion of subjects (47.4% and 21.1%) perceived the traditional handle as either "a little big" or “too big”, respectively. In contrast, only 31.6% considered the traditional handle to be "just the right size" (Fig. 7a). Conversely, a significant majority (84.2%) of subjects regarded the novel handle as an ideal size, with negligible proportions expressing opinions of it being either slightly smaller (1 subject) or slightly larger (2 subjects) (Fig. 11a).

With regard to the design of directional buttons on the handle, 63.2% and 84.2% of subjects expressed overall satisfaction with the traditional and novel handles’ directional button designs, respectively. No significant difference was observed between the two designs in terms of overall satisfaction. However, a closer examination of specific reasons for dissatisfaction revealed that 47.4% of subjects found the buttons on the traditional handle difficult to reach, while only 5.3% reported similar difficulties with the novel handle (Fig. 11c). This discrepancy was found to be statistically significant in the novice group.

Comparison of two colonoscopy handles

It is clear that the development of the novel colonoscopy handle is still at the prototype stage and has not yet entered animal and clinical trials. However, it offers significant advantages such as no need to modify existing endoscopic equipment and future applications in telemedicine and robotic colonoscopy. We compared the devices with traditional and novel handles as shown in Table 4.

Table 4 Comparison of two enteroscopic handles in clinical applications.
Full size table

Discussion

This study demonstrates that enhancing the design of the colonoscopy operating handle results in a novel handle that is not only non-inferior to the traditional handle in terms of work efficiency but also exhibits significant advantages in reducing endoscopic muscle fatigue and joint injury. The ergonomic configuration of the novel handle renders it particularly suitable for novice colonoscopists. Empirical evidence derived from the analysis of multiple indicators substantiates that the novel handle holds considerable potential for clinical application.

A substantial body of research has underscored that the conventional mode of operation of the colonoscopy handle frequently culminates in substantial muscle fatigue and wear, giving rise to occupational injuries among endoscopists who operate colonoscopes for protracted durations14,16,20,21. The findings of this study indicate that the novel handle, equipped with a booster device, effectively diminishes muscle fatigue without necessitating modifications to existing endoscopic equipment or compromising procedural efficiency. This outcome underscores the clinical viability and practical applicability of the new handle for integration into routine endoscopic procedures, potentially streamlining clinical workflows while maintaining high standards of patient care.

A previous literature review revealed that research and development of endoscopic devices have predominantly centred on mechanical and robotic flexible multitasking platforms22. Devices like the IOP, DDES, ANUBIScope, and MASTER systems have primarily enhanced the scope’s insertion or lens front, necessitating major equipment modifications8,23,24. In contrast, improvements to colonoscopy handles have seldom been studied, with even fewer evaluations of handle operation ergonomics. The novel endoscopic handle was specifically developed to address practical clinical challenges, facilitating seamless integration into existing clinical workflows. Unlike previous studies, it offers several distinct advantages: effortless installation and manoeuvrability, reduced occupational injury risk, and experimentally confirmed efficacy. Furthermore, the handle is fabricated using 3D printing technology, ensuring rigorous electrical and mechanical safety standards. Its design simplifies disassembly, and it presents no additional complexity in cleaning and sterilisation protocols when compared to traditional colonoscopy handles.

Operators engaged in performing colonoscopies are susceptible to a range of occupational injuries, including muscle strain, joint damage, varicose veins, visual fatigue, and psychological stress14. Previous and ongoing studies, like that by Pullens17, have aimed to improve the mode of operation of endoscopic handles, by utilizing robotic steering to enhance ease of use and intuitiveness. As for the areas of muscle and joint damage, previous studies have found that common areas of pain when using traditional endoscopy handle are the shoulder (9–19%), elbow (8–15%) and hand/fingers (14–82%)16,20,25, that the evidence mostly relied on survey documentation, with few employing objective methods like quantification of muscle fatigue and performance.

Conversely, this study innovatively used a surface EMG acquisition device to quantify the work and efficiency of left upper limb muscles during colonoscopy. This not only provides an objective basis for evaluating colonoscopy handles but also establishes a data foundation for ergonomic handle design. When analysing the sEMG data, we found that the left middle deltoid work was significantly higher when using the novel handle. The left biceps brachii, left flexor carpi radialis and left abductor pollicis brevis was significantly higher when using the traditional handle. It is noteworthy that these muscles are susceptible to occupational injuries among colonoscopy operators, underscoring the importance of the novel handle in alleviating muscle strain and reducing such injuries14,15. Furthermore, in addition to reduced fatigue in the left upper extremity, as indicated by sEMG results, subjects reported improvements in fatigue levels in other body parts, including the neck, waist, back, lower extremities, and mental state. This enhancement may be ascribed to the adoption of an overall relaxed posture by subjects during utilisation of the novel handle.

The questionnaire also collected subjective assessments of the different handles regarding work efficiency and mechanical design. Notably, the evaluation of work efficiency was significantly influenced by the operator’s experience. A considerable proportion of participants in the novice group believed that the novel handle could enhance work efficiency, whereas the experienced group exhibited a contrasting trend. However, both groups provided positive feedback on the ease of handle operation, indicating that the novel handle’s operational mode is more intuitive. Regarding mechanical design, participants expressed favourable opinions about the novel handle’s size and grip comfort. A significant number of participants reported that the traditional handle had issues with buttons being difficult to reach, a problem that the novel handle effectively resolved.

This study demonstrated that the novel colonoscopy handle offers significant advantages and requires no modifications to existing endoscopic equipment. We propose that this device has potential for clinical translation and future applications in telemedicine and robotic colonoscopy. However, as the device remains in the prototype stage, further refinement is necessary to address its inherent limitations and facilitate clinical application. Future research will focus on enhancing the responsiveness of the novel handle’s button servos to ensure reliable suction, water and air delivery, and locking functions. Additionally, we will integrate a force feedback mechanism into the handle to improve operational safety. Preclinical validation will involve rigorous animal experiments to assess the handle’s operability and safety before proceeding to clinical trials.

Moreover, the handle now includes a preliminary data recording function, which enables the translation of endoscopists’ manoeuvres into electronic signals. This advancement establishes a foundation for developing smart handles via machine learning, a key future research direction. We have also recorded expert manoeuvre data for future deep learning and expert system model training. Combining this with existing image AI recognition technology provides a solid basis for subsequent robotic colonoscopy research.

This study has several limitations. Firstly, the experienced and novice groups differed in demographic and anthropometric factors such as gender, age, height, and weight, due to differences between the two populations and limited available subjects. These factors can influence colonoscope operation and the learning curve. Future studies should refine the classification of “experienced subjects” to enhance result validity and generalizability. One approach is to create subgroups based on the number (> 3000 cases) of colonoscopies performed. Secondly, limited training time prevented equal familiarity with both handles in the experienced group, potentially affecting comparison fairness. The study’s small sample size also limits result generalizability, necessitating further validation in larger cohorts. Additionally, the pre-installed lesions in the lower GI model were easily visible, leading to 100% detection rates in both groups. This precluded effectively evaluating how different handles affect bowel observation quality during withdrawal. Future studies should use GI tract models with adjustable lesions. Lastly, colonoscope function validation focused only on lens rotation due to experimental constraints. Other functions like air and water delivery weren’t assessed.

In summary, we have developed a novel handle with the primary advantage of reducing muscle fatigue, while maintaining efficiency comparable to traditional handles. The mechanical design has been well-received by subjects and demonstrates novice-friendliness, making it suitable for learning by novice colonoscopists. The findings of this study lay the foundation for the future development of smarter handles.