We used the database from the Project to Collect Medical Near-Miss/Adverse Event Information of the Japan Council of Quality Health Care [24]. Figure 2 is a flow diagram of the searched results using this database. In April, 2018, the keyword of “thoracic surgery” was used, and 211 cases were searched from the database. Eighty-six cases were hit by searching the word of “hemorrhage” or

“bleeding” in 211 cases and the remaining 125 cases were excluded. The searching word of “pulmonary artery” hit 29 cases and another searching word of “pulmonary vein” hit 14 cases. After the exclusion due to the duplicated cases and the nonconformance reports, 25 cases were finally selected as the “bleeding/hemorrhage cases in thoracic surgery”.

Between July 2002 and March 2020, we retrospectively selected 13 experienced cases of intraoperative and postoperative hemorrhagic incidents in our department of thoracic surgery. This study was approved by the University of Miyazaki Hospital Clinical Research Ethics Board (No. O-0736 and No. O-0737) in July 2020.

We analyzed the searched cases (n = 25) and the experienced cases (n = 13) based on the items of surgical sites, procedures, site of bleeding, perioperative period, causes of bleeding, response and treatment, support by cardiovascular surgeon, circulation auxiliary equipment, and prognosis. Table 1 shows the case details of the bleeding accidents (n = 25) from the database of the Project to Collect Medical Near-Miss/Adverse Event Information and Table 2 shows our experienced cases (n = 13) of bleeding accidents in our department.

Dr. Erik Hollnagel provided a deep insight into the ideal way of safety management in social technology systems, focused on “failure”, and called safety management aimed at “eliminating failure” as “Safety-I”. Focusing on “how daily work is done under disturbance and constraints”, “making things go well”, it is called “Safety-II”, and the difference in approach is clearly shown [1].

Table 3 displays the features and differences of Safety-I and Safety-II [9]. Safety-I recognizes the limitations of Safety-II and has evolved from a critical standpoint. However, Safety-I and Safety-II are complementary to each other, and safety management of social technology systems also requires therapy, and Safety-II is not a substitute for Safety-I [9].

In order to develop resilience engineering theory into surgical safety, it is necessary to clarify the specific methods for the following four items.

1) Observe the daily work performed at the surgical site,

2) Understand various “adjustments” performed by the surgeon according to the surgical situation,

3) Predict “functional resonance” that causes unexpected situations,

4) Take preventive measures.

In this study, we observed the surgical work during thoracic surgery from the viewpoint of surgical safety, collected adjustment cases especially for bleeding cases, and analyzed the background factors.

Surgical care is a complex system and is characterized by uncertainty. The

Bleeding cases were accidental and should be handled with care. Patient age, gender, and clinical background were anonymized and only short-term outcomes were described upon survival and death. L: Left, R: Right, Pn: Pneumonectomy, Seg: Segmentectomy PA: Pulmonary artery, PV: Pulmonary vein, UL: Upper lobectomy, LL: Lower lobectomy, CPB: Cardiopulmonary bypass, PCPS: Percutaneous cardiopulmonary support

Bleeding cases were accidental and should be handled with care. Patient age, gender, and clinical background were anonymized and only short-term outcomes were described upon survival and death. L: Left, R: Right, PA: Pulmonary artery, UL: Upper lobectomy, LL: Lower lobectomy, ML: Middle lobectomy, CP: Completion pneumonectomy, CPB: Cardiopulmonary bypass.

situation at the site of surgery is constantly changing, and the people working there are making “various adjustments according to the situation” rather than “as prescribed”.

To improve the quality of surgery, we have to learn from failures and incidents, such as bleeding cases (Safety-I), and developed our skills and practice of surgical safety management.

In order to form a system to support surgical safety, we looked back on our experience in dealing with bleeding cases and developed a method to apply a new approach (Safety-II) called resilient healthcare assuming a complex system to surgery.

Safety-I recognizes the limitations of Safety-II and has evolved from a critical standpoint. However, Safety-I and Safety-II are complementary to each other, and safety management of social technology systems also requires therapy, and Safety-II is not a substitute for Safety-I. This table is quoted in English from the book that Professor Kazue Nakajima authored. (Introduction to Resilient Health Care. Ability to respond flexibly with disturbances and constraints) (Japanese), Igakushoin, ISBN976-4-26-02828-8, page 14, 2019) [9].

Table 1 shows the case details of the bleeding accidents (n = 25) searched in the medical accident information collection project database. Bleeding cases were accidental and should be handled with care. Patient age, gender, and clinical background were anonymized and only short-term outcomes were described about survival and death. Table 4 shows the summary of the bleeding accidents from Table 1. Regarding the surgical sites, the left side (68%) was common. Regarding the procedure, the upper lobectomy cases (56%) were commonly observed. Regarding the sites of bleeding, there were the pulmonary artery (56%), the pulmonary vein (12%), and the aorta (12%). Bleeding was observed in the intraoperative (76%) and postoperative periods (20%).

The causes of bleeding included damage of adhesion detachment, damage of automatic suture insertion, detachment of ultrasonic scalpel, damage during vascular taping, damage when forceps are inserted, when vascular forceps are applied, when the resected lung is removed, lymph node dissection, exfoliation of the infiltrated adventitia of the pulmonary artery, and when one lung ventilation is released.

The circulation auxiliary equipment was required in 8 cases (32%), the breakdown was 5 cases of “cardiopulmonary bypass” (20%), and 3 cases of “percutaneous cardiopulmonary system” (12%). There were 9 cases of the support by the cardiovascular surgeon (36%).

The corresponding procedures included suture stasis and total resection, that is, there were 4 cases of a pneumonectomy for life-saving (16%), that was a change in procedure, and 9 cases (36%) involved death.

Thoracic surgery is a very high-risk operation that can lead to life-threatening conditions that can lead to increased bleeding and death if the procedure or operation is incorrect. We can learn about various types of bleeding events and apply these findings to normal routine surgery. It is important to learn, prepare and have a large repertoire.

Table 2 shows the case details of our experience (n = 13) searched in our database. Bleeding cases were accidental and should be handled with care. Patient age, gender, and clinical background were anonymized and only short-term outcomes were described regarding survival and death. Table 5 shows a summary of the bleeding accidents from Table 2. Regarding the surgical sites, the left side (62.9%) was common. About the procedure, the upper lobectomy cases (53.8%) were commonly observed. Regarding the sites of bleeding, it was the pulmonary artery (100%). The bleeding was observed in the intraoperative

(84.6%) and postoperative periods (15.4%).

The causes of bleeding included injury of adhesion peeling operation, pull-out injury during lung traction, injury due to forceps insertion, gauze attachment with staple line, thoracoscopic collision, infectious vascular rupture, detachment of vascular ligature, removal of automatic suture, and suction tube hit and sucking of the transected end.

These cases were observed within or outside the expected range. The circulation auxiliary equipment was required in 3 cases (21.4%), i.e., the breakdown was 3 cases of “cardiopulmonary bypass” (21.4%). There were 3 cases of support by the cardiovascular surgeon (23.1%).

The corresponding procedures included suture stasis and total resection, that is, there was 1 case of pneumonectomy for life-saving (7.7%), which was a change in procedure, and there were no deaths.

For the life-threatening conditions, we avoided a patient’s death by changing the surgical procedure from a lobectomy to pneumonectomy, and with the cooperation of the cardiovascular surgeons with multi-professional experience and the formation and cooperation of highly-reliability surgical teams in the hospital organizations. We could do it and learn the practical experience. It was a successful case by implementation of resilient healthcare and a balanced combination of Safety-I and Safety-II. By learning from the successful cases, that is, the bleeding can be stopped or life can be saved by appropriate measures, that is, the successful cases. Crisis response and management methods can be surely practiced in the event of a life crisis.

The conventional safety management method (Safety-I) of “learning from failure, identifying the cause, and taking measures to prevent recurrence” is premised on the linear system industry, such as automobile production lines, and is complex like as medical care. It has been pointed out that there is a limit to its usefulness in systems (complex systems).

Against this criticism, resilience engineering theory said, “Because success and failure have the same origin in complex systems, we focused on what was done well, not failure, and predicted the future without waiting for an accident. Take the opposite approach of the conventional safety management method of “corresponding to the target” (Safety-II).

With the knowledge and coping strategies for thoracic surgery incidents and accidents revealed, we divided them into four surgical situations and considered surgical safety in the practice of resilient healthcare: 3.3.1. daily routine work in normal surgery, 3.3.2. unusual patient’s individual factors, 3.3.3. massive bleeding, and 3.3.4. unusual crisis of the unexpected major bleeding.

The bleeding of the pulmonary artery branch is performed using the lung tissue itself, gauze, cotton compression, attaching a sheet of hemostatic tissue adhesion, and suturing the blood vessels. In the first place, it is important to prevent bleeding, and to avoid bleeding.

Based on the protective manipulation of the blood vessels, intravascular sheath detachment is recommended. Although a safe detachment, detachment with force control without tension on the fragile pulmonary artery wall is fundamentally an important operation. The foresight of peripheral branching of blood vessels at the time of exfoliation is to expose the branch vessels, lung grasping method for expanding visual field, and direction of counter traction. The balance between gender and power is important, and prevention of pulmonary artery intimal detachment and withdrawal of blood vessel roots is important.

Surgeons have to learn how to dissect blood vessels with infiltration and inflammatory adhesions. Bleedings of the A³ and A¹⁺² branches in the upper lobes are a higher risk than the bleedings of the lower lobes A⁶ and A⁷⁸⁹¹⁰ branches, because the difficulty of repair is high, and emergency simulation becomes important. Depending on whether the site of bleeding of the pulmonary artery has to be resected or not, the flexibility of the hemostatic operation method changes.

From the viewpoint of patient safety, bleeding events must be learned from the standpoint of the surgeon, assistants, supervising physician, and training physician. As the total strength of the surgery team, everyday intraoperative communication becomes very important, for example, an easy-to-speak atmosphere, that is, non-technical skills resilience, and resilience ability for unexpected accidents. We have to consider the range of possible bleeding treatments based on the presence of inflammatory adhesions, a history of thoracotomy, post-chemoradiation, and the presence of a cardiovascular team.

In the case of unexpected, uncontrollable, massive bleeding, or life-threatening bleeding that is not controllable, the following items should be noted. It is necessary to simulate in advance from the usual.

1) To check and stabilize the vitality in cooperation with the anesthesiologist and surrounding nurse.

2) Confirmation of the bleeding point and compression of bleeding site because the bleeding from the upper lobe pulmonary artery increased the fatal risk. This will reduce the amount of bleeding and prevent secondary disasters (increased vascular damage).

3) To prioritize hemostasis (compression), add the thoracotomy, secure the intrathoracic visual field, remove massive hematomas, peel off, secure and taping of the main trunk of the pulmonary artery, superior pulmonary vein and inferior pulmonary vein.

4) Do not release the pressure on the bleeding site, check the bleeding area, or proceed with recovery of the patient’s vitals until the blood transfusion is secured and the blood transfusion occurs.

5) To consider the option to change to a life-saving procedure such as a pneumonectomy.

6) To consider support from a cardiovascular surgeon and prepare for a cardiopulmonary bypass.

7) Efforts should be made to understand blood loss, check the amount of blood transfusion, minimize blood loss, and pump blood during the transfusion to maintain the patient’s vitals.

8) After preparation, remove massive hematoma, release compression of the bleeding, taking into account the determination of damage, the size of the lesion, consider the amount of blood transfusion, consider the hemostasis methods and blood vessel repair methods.

9) Explanation to patient’s family, report to the department of patient safety management, and act as a hospital organization.

10) Efforts to improve the surgical technical skills (vascular suture method) of the surgical procedure from normal times.

11) Simulation during critical bleeding.

12) Non-technical skills during crisis (situation awareness, decision making, leadership and followership, communication, and stress relief) are important.

We must be aware of organizing the surgical team and operating room team from the usual one.

In order to simulate in a crisis situation and to demonstrate the resilience of the surgical team, all the teams can perform non-technical skills.

Figure 3 shows a general model for the performance fluctuations in the stability of the operating team by responding to disturbances in thoracic surgery. The vertical axis represents the stability and performance, and the horizontal axis represents the time course of the surgery. In the case of lung cancer surgery, a lobectomy is performed, there are three important scenes, that is, dissection of the pulmonary artery, pulmonary vein, and lobular bronchus. However, a surgeon sometimes encounters unexpected situations, such as adhesion and vascular bleeding events, which means disturbances in the normal surgical course.

These disturbances are well responded by the individuals, operating team, and hospital organization. Compared to the normal flow, it consumes a lot of time and performance drops during the overall course of surgery. The surgical team becomes less stable and unstable.

In the event of a life-threatening massive bleeding, the surgeon responds with an operating room staff team or hospital organization to prepare and administer blood transfusions, or with the help of a cardiovascular surgeon, save lives using a heart-lung machine and the bleeding site. The surgeon repairs the hemorrhage and the anesthesiologist stabilizes the hemodynamics. The patient’s life is then maintained, recovered, and successfully saved.

Depending on the magnitude and situation of the disturbance, the performance of the surgical team will fluctuate and the stability will change. It is Safety-II that changes the performance of the surgical team according to the disturbance, stabilizes the place, and allows life to continue.

Figure 4 shows the reconciling Work-As-Imagined (WAI: How to do the work that you think in your head) and Work-As-Done (WAD: How the actual work is done) in Surgery. In surgery, the WAI is manual and the surgical procedure plan.

Surgery is performed according to preoperative complications, the WAD is

the actual surgery, that is, the demonstration of surgical practice. However, in reality, the results of surgery appear as surgical outcomes due to time and member constraints and disturbances such as major bleeding. Surgical outcomes are recorded and evaluated in texts, pictures, photographs, videos, etc. These valuable experiences will be shared at team conferences to reflect, learn and improve, which are stored as a large “repertoire”. The surgeons’ feedback for the next surgery is to improve technical skills. It is important to reconcile the WAI and WAD in surgery.

Figure 5 shows a system to support surgical safety and implementation of the

resilient healthcare theory. A balanced combination of Safety-I and Safety-II is important for building a system that supports surgical safety. Work-As-Imagined (WAI) is “how to do the work that you think in your head”. Work-As-Done (WAD) is “how the actual work is done”. There is a gap between the WAI and WAD, which needs to be reconciled. By using the PDCA cycle (plan, do, check, and act), the quality of surgery improves.

For implementation of the resilience healthcare theory in surgery, it is important to go through a mechanism of anticipating, monitoring, responding and learning. By increasing these four potentials, the ability of resilience healthcare will improve, but for this purpose, it is important to do it while learning and to learn while doing.

Lastly, in order to reduce the number of surgical bleeding incidents, the following items are important.

1) We have to know the many variations of bleeding cases that we have experienced and learned.

2) We should know how to deal with various bleeding cases, and regularly practice so that we can practice successful methods.

3) During surgery, it is important to predict that bleeding may occur at any time.

4) We should monitor the operative situation at the time of bleeding and the response.

5) We should know performance fluctuations during the operative time course due to various disturbances.

6) We have to adjust the various disturbances under constraints and stabilize the operative field.

To learn bleeding cases that we had responded well in the past and to build it as a system to support surgical safety, for that purpose, reconciling the WAI and WAD in the practice of surgical resilient healthcare, from the perspective of Safety-I and Safety-II, it is important to consider surgical patient safety measures.