Depicting adverse events in cardiac theatre: the preliminary conception of the RECORD model

Behavioral outcomes in the operating theatre

Optimal outcome following an operation is a result of multimodal harmony of variables. Attitude, team work, organization and protocol implementation as well as optimal knowledge of the task referred to, are important as simply seen when one defines the difference between mistake an error: A mistake which means “wrongly taken” is a wrong response that if you thought about it you would realize is wrong. An error which means “to wander or stray” is a wrong response because you have no knowledge about what the right answer is. With other wards a person if given a second chance has the potential to correct a mistake, whereas a person has no potential to correct an error until he learns what is correct.

In such a volatile environment mistakes and their precursors are waiting to occur; More than 50% of surgical adverse events occur in theatre [2] whereas half could have been prevented with simple-measures. As per Wiegmann [3] about 60% of surgical errors were picked up and managed immediately by the surgical team. It is interesting that of these errors, 32% were events during which the team experienced difficulty performing a specific technique or procedure.

Adverse events are caused by decisions taken by persons at the delivery end of a system or are results of deficiencies in the organizational and management level of the system.

Surgical error may occur when lack of skill or ability is combined with behavioral deficiency. What we mean with this statement? Well, when the knowledge or the skills required for a given procedure are lacking and despite this, for example the leading surgeon is deciding to undertake the procedure with no safety backup. With other wards lack of appreciation of ones knowledge limitations can lead to misjudgment and adversity.

Skill-based errors were thought to be the most common type of unsafe acts in theatre [4] followed by changing or bending the predetermined rules in under pressure circumstances.

Reasons “swiss cheese” model illustrates that the organizational influence (climate, resource management, and policies) impact supervisory processes (scheduling, training, and oversight), which in turn establish the preconditions (technological and teamwork related) that produce errors.

The Human Factor Analysis Classification System (HFACS) framework was developed using over 300 naval aviation accidents obtained from the US Naval Safety Center.

Based on HFACS, questions were developed to target each of the causal categories within the HFACS by ElBardissi et al. [4]. With other words this group have tested the application of HFACS in Cardiac surgery. When asked in an open-ended manner to characterize surgeon personality and its reflection in the operating theatre environment, 35% of individuals stated that surgeons are arrogant or demeaning, negatively affecting team function, 24% stated that surgeons have a strong or intense personality, also negatively affecting team function. Therefore, in this report with regards to teamwork issues, 59% of responders indicated that a surgeon’s personality (by means of surgeons misbehavior in theatre), negatively affects the way the surgical team functions, Because the surgeon’s leadership and team-ness are hugely important [5], many institutions have attempted to address this issue of occurred behavior by sending surgeons to charm school in order to improve their attitude in the operative room; this is an interesting point and has to be taken seriously from our senior colleagues. However, the exact role of teamwork in error causation is unclear, emphasizing the importance of additional research in this area before implementing interventions that have no scientific basis.

Although it is unlikely that one’s innate personality can be changed it is undoubtedly possible to alter aspects of behavior, which impact negatively on colleagues and on the team in the work place. With appropriate training, individuals can improve and thus function more effectively as part of a multidisciplinary team.

Major and minor events represent various types of human errors. Those events have been elegantly defined and extensively studied [6]. Minor events are errors that alone were not expected to have serious consequences for the patient; these included instrument handing, errors by the scrub nurse, or communication problems in the theater team. Major events are more serious errors, such as for example, accidental injury of a coronary artery and severe laceration of the aorta during canulation.

The surgeon’s diagnostic skill, knowledge of the various maneuvers to rectify problems, and communication with the team are important prerequisites of compensation. Uncompensated major events have a detriment outcome. If each of them is appropriately compensated, however, death can be avoided. Minor events are different. They are less apparent and insidious, and many of them are not even noticed by the operators. They are independent of case complexity and correlate with “lack of flow”, distractions and lack of communication in the operative theatre. More specifically Barach et al. [6] in a prospective observational study of one hundred two patients undergoing pediatric cardiac surgery, they analyzed minor and major events; the surgeon wore a video head-camera so that observers could also follow the surgical steps. The observers created a handwritten observational tool to document the surgical flow. A median of 1.1 major events (range 0–6) occurred per case; with 43% of the major events observed to occur when the aortic cross clamp is release and is followed by weaning and separation from CPB. A median of 18.3 minor events (range 2–54) occurred per case, most often during CPB. This study revealed that adverse events occurred regularly during cardiac surgery and moreover that case complexity and duration were significant predictors of major adverse events, which in turn correlates with adverse outcome. On the other hand, reactive than pre-emptive compensatory mechanisms were detected in 98% of major events and 90% of minor events.

The RECORD model

In order to identify the failed or absent defenses on the chain that leads to unsafe acts and sentinel events we hypothesize and presenting here a model that maybe have a practical application in safe delivery of surgical care.

We need to outline here that this is a preliminary report of the model, which recently started to evaluate its feasibility in practice.

The model proposed is based in the abundance literature that suggests that surgical practice should be reported, evaluated and one should reflect on it.

Furthermore the model takes into consideration the visual grading of a procedure; either with the use of “human observers” specialists to the procedure in question able to grade it, or the use of a video camera recording.

Lastly the model aim to use two more tools: audit forums and morbidity and mortality meetings and the use of questioners with a five point Likert scale based on frequency of occurrence (always= 5, very often= 4, sometimes= 3, rarely= 2, never= 1).

We are currently enrolling 100 CABG cases in an attempt to answer two aspects:

The last step of the model would be to use a Video camera recording randomly and human factor specialists.

Implementation all the steps of the RECORD model may be cumbersome and preliminary results are not out yet however we believe that this multifaceted approach may allow us to capture and study “ adverse events in cardiac theatre” in order to improve HF. In Figure 1 an algorithm to stepwise investigations of adverse effects in theatre is given. In Table 1, the principles of the RECORD model are depicted.

Limitations of applying “various systems” in clinical practice

Under-reporting [19] of incidents is a limiting factor that unfortunately it relates to the old notion of “a blame culture” in medicine; it is inappropriate, when junior doctors do not report incidents [20] because they fear they will be blamed. Equally it is wrong when they are blamed by senior colleagues; under the circumstances the team should be united to investigate and learn from the errors.

The obligation to name and define the role of the person reporting the incident can be a restrictive factor for reporting; it is apparent that confidentiality encourages an easier reporting procedure and probably should be implemented when possible. Moreover, the retrospective nature of reporting incidences further necessitates their need for validation and accuracy.

Clear definitions of [8] near misses in cardiac surgery is complicated by the need to distinguish between those events against serious peri-operative and post- operative complications. This can only be achieved when the human factors observers are trained and familiar with the procedures under investigation. With other words standardized training and calibration of observers would improve the data collection.

The capture of adverse actions and events are subjective and depend on the observers training and education; therefore, depicting accurately data in this respect could be limiting by interpreter reliability. Moreover, there are many technical challenges when video taping surgical teams such as logistics, ethics and interpretation [12]. Weaknesses in using video for data include lengthy video review processes, poor audio, and the inability to adequately analyze events outside the field of view [21].

The development and implementation of a system for measuring technical performance in the operating theatre is difficult. The challenges of grading the technical expertise, has been addressed by Karamichalis et al. [22]. An individual practitioner’s surgical performance includes the technical domain and other nontechnical skills such as cognitive flexibility, decision-making etc.

Technical performance, is defined as “the adequacy of the surgical anatomic repair intended,” and as per Larrazabal et al. [23] intraoperative technical performance is one of the most important, if not the most important, parts of the therapeutic process and determines postoperative outcomes. None of the currently available quality monitoring tools measures technical performance; the authors created and validated a scoring tool in this respect; the limitations of their tool is however due to the fact that in their model “rating surgical adequacy” is based on ECHO, which creates the biases of the operator dependent technique.

Lastly, using models for reducing errors may be part of the solution; as per Auroy et al. [24] risk assessment and control require analysis of both outcomes and process of care.

We admit into several limitations of this report. The model reported here is more of a zealous and enthusiastic project, in an attempt to capture more accurately, the already well defined many aspects of human factors in the cardiothoracic environment.

The first attempt of implementing a robust model was conceived by our team, following few years of using Briefing and debriefing in the cardiac operating room, in a manner similar to the model reported by Papaspyros et al. [25] with satisfactory outcomes (unpublished results).

We need to clarify here, that we conceived the idea of the RECORD model in an attempt to put more variables together in a continuous fashion. With other words, we would try to compile a “feedback approach” by the staff involved in the theatre environment together with the “specialists reporting approach” of the human factor observers in a stepwise fashion.

We would also like to confess, that we are still working to make the questioners simple and reproducible, for easy use.

So overall, this is a preliminary idea that we are working to materialize.

Lastly, in this editorial we hypothesized that putting more variables together we will be able to overcome some of the limitations to human factor research and analysis of performance assessment; although we just initiate implementing our model in clinical practice we hope that we will be back with real world data analysis in order to be able to depict and act upon clusters of latent errors before they become active. We are hoping to show in a subsequent study by piloting this model, that we will be able to support the basic idea of the model (more variables, more feedback, less bias) for others to emulate.