Background
Since the publication of the report “To err is human” in the year 2000, the readiness for structured analysis of adverse events (AEs) in medicine has been increasing [
1]. AEs are defined as treatment outcomes that are below the currently expected medical standard and result in temporary or permanent harm to patients [
2].
The incidence of AEs in hospitalized patients ranges from 3 to 17%. Up to 50% of AEs are classified as preventable [
3,
4]. As early as the 1980s, the Harvard Medical Practice Study showed that 4% of patients in a New York hospital were harmed by AEs. As a result, human and organizational factors became the focus of risk management, as these factors, together with insufficient medical knowledge, are seen as triggers of preventable adverse events (pAEs) that can be reduced through interventions [
5,
6].
Intra- and interprofessional communication of health care workers (HCW) is a significant human element in health care. Faulty communication behavior is exacerbated by rapid team changes, high workloads, and lack of mutual respect. Lack of communication training and awareness of its importance leads to loss of essential information and compromises patient safety [
7,
8]. PAEs can be attributed to inadequate communication in up to 70% of cases [
9]. Error analyses in surgical departments showed that causal factors for adverse events were 65% due to human factors and only 4% due to technical problems [
3].
PAEs also occur in obstetrics. In a study in Sweden 12.2% adverse events were detected of which 73.7% were assessed as preventable [
10]. The special feature of obstetrics as a medical discipline lies in the coincidence of unpredictable emergencies, as well as the simultaneous responsibility for the expectant mother and the unborn child. However, the general expectation is that the birth of a child is a joyful and unencumbered event, and oftentimes possible complications are ignored [
11]. This challenge must be met by an interdisciplinary team facing frequently changing staff, shift work, and different levels of competence. The consequences of pAEs in obstetrics are far-reaching and affect not only the future family but also the obstetric team involved. The obstetric team usually must deal with the possible medico-legal consequences of an incident in the long term and often suffers from the psychological burden of being "second victims" themselves [
12,
13]. In obstetrics, too, feelings of guilt, anxiety, frustration, and self-doubt are reported as frequent consequences of pAEs [
11]. The consequences are on the individual side with personal suffering, costs for the institution due to fines or compensation payments up to the loss of a job or an employee, resp., and thus further work intensification. It may even accelerate susceptibility to errors [
14].
A reduction in pAEs can be achieved by improving interdisciplinary communication in the HCW team in obstetrics [
15]. In a review of communication and teamworking trainings in obstetrics, an improvement in professional and empathic communication was observed in different programs. In 56%, communication was trained as the main construct and in 44%, communication was integrated into a broad teamwork training, including crew resource management or emergency simulation approaches. Most teamwork approaches have a clear communication component, as the allocation of resources and effective management of obstetric teams can only be achieved through effective interdisciplinary communication. Of the 71 studies that were included in the review, most focused on technical aspects of communication, e.g., effective handovers. 16 studies also reported clinical outcomes of communication, e.g., behavior in emergency simulation. Only three studies investigated teamwork in relation to the distal outcome of patient safety and medical outcomes, e.g., measured by the Adverse Outcome Index (AOI), of which only two showed moderate effects on patient safety [
16]. This result can be partly explained by the fact that the preventability of AEs cannot be determined in a standardized way and often only serious "sentinel events" are reported.
An explicit definition and monitoring system for the risk factors for pAEs is needed to reduce the occurrence of pAEs [
12]. This was done in an exemplary approach within the project TeamBaby. The research approach was developed for identifying pAEs, which could be used to evaluate a communication intervention. An interdisciplinary communication training for medical staff in obstetrics was designed and implemented. It focused on increasing patient safety through improved interprofessional communication of the obstetric team. These measures aimed to promote effective communication to reduce pAEs. To evaluate the effectiveness of the measures, categories of different pAEs after implementation of the TeamBaby communication training (year 2020) were compared to a reference year before the intervention (year 2018). The research hypothesis was that the frequency of pAEs would decrease in the year following the implementation of the training [
15].
Discussion
This study tested the research hypothesis whether the frequency of preventable adverse events (pAEs) would decrease in the year following the implementation of a communication training addressing health care workers in obstetrics. Overall, 2,865 of all 3,351 births in the year 2018 were analyzed as reference year and 2,846 of all 3,302 births in the year 2020 after the training. A significant reduction in pAE was observed in terms of 13.35% in the year 2018 vs. 8.83% in 2020. Subjective ratings were in favor of the effectiveness of the intervention but as there was no active control groups or a randomized control trail, the effectiveness can only be concluded from this pre-experimental design. For the reduction in pAE, there are several background variations that could account for the reduction in pAE. Future research is needed to validate the findings accordingly.
One should keep in mind that AEs in the treatment of patients will not be completely avoidable despite technical progress and sufficient training. Within the medical disciplines, obstetrics plays a special role. In the case of complications, fatal and possibly life-determining consequences can occur for mother and child, although the expectant mother is initially not a patient, and giving birth is a physical process and not a disease requiring treatment. However, the rise in maternal risk factors has increased the challenges for obstetric staff, who are additionally burdened by medico-legal aspects with a simultaneous increase in work pressure and reduced staff.
AEs, responsible for morbidity and mortality in hospitalized patients, are reported to be between 3 and 17%, and up to 50% are considered preventable [
3]. In obstetrics, an incidence of up to 5.9% is reported [
10]. In statistics, obstetrics is often not evaluated separately but subsumed under the specialist discipline of gynecology, which makes it difficult to state the exact number of AEs in obstetrics. Studies that explicitly consider obstetrics give incidences between 0.4% and 3.6% with a preventability of more than 50% [
10,
22]. In our evaluation, the incidence of AEs is 23% and 29%, respectively. However, with the help of a very extensive trigger list, we did not only filter for the typical sentinel events and thus carried out a differentiated evaluation of 56 triggers. The evaluation according to preventability showed similar results as in the existing literature, here the incidence was 3% and 2.6%, respectively.
However, in many studies, the data are evaluated retrospectively from different documentation systems, which in turn leads to difficult comparability and distortions. Often, the mostly complex and situational conditions in the documentation systems can neither be identified nor precisely assigned retrospectively.
One method for standardized retrospective analysis of AEs is the "Harvard method" and the "Global Trigger Tool (GTT)". In a two-stage procedure, they retrospectively evaluated patient files for defined medical treatment errors. An additional 6-point scale was used to classify the preventability of the adverse events [
23]. In a retrospective analysis of 311 births, 38 (12.2%) AEs were identified using the GTT. Of these, 28 (73.7%) were classified as preventable. Most of the events were grade 3/4 perineal tears, bladder voiding disorders, and anesthesia-related complications. In addition, a distinction was made between prolonged hospitalization (63.2%) and temporary harm (31.6%) [
10].
Another tool for standardized recording and comparability of adverse events is the Adverse Outcome Index (AOI) [
24]. This includes 10 items of obstetric adverse events, so-called sentinel events. However, the AOI does not allow differentiation between preventable and unavoidable adverse events. It also lacks adjustment for pre-existing risk factors of the patient.
One paper that used this index to measure the effect of teamwork training failed to achieve the study objective of reducing the overall number of adverse events. Only the decision to deliver time in case of an emergency section was significantly reduced [
25]. Another study also used the AOI as a measurement tool to demonstrate the effectiveness of team training. Here, the score was only used with 5 parameters as a measurement tool for the primary outcome. There was also no effect on the AOI demonstrated by the training. Overall, an adverse event according to the AOI 5 was identified in 11.3% of cases [
26]. However, pAEs do not always result from medical errors and can therefore not only be measured by scores like the AOI that records only sentinel events. Often, they can be traced back to so-called human factors such as insufficient communication [
3,
9]. The human factor can impact patient safety via so-called “non-technical skills” (NTS). NTS consist of a variety of cognitive skills (e.g., situational awareness), social skills (e.g., ability to work in a team), behavioral skills (e.g., effective communication), and personal skills (e.g., individual stress management). Especially cognitive skills are connected to patient safety incidents because they are related to impaired situational management or delayed treatment decision [
27]. Another practical example in which the human factor of both medical staff and obstetric patients can lead to pAE, are language barriers. Language barriers can lead to the poor comprehension of diagnosis and treatment options, thus affecting effective information flow and decision-making [
28]. Patient safety can therefore be improved by strengthening the human factors and addressing language barriers. One point of action could be respecting the patient perspective by for example involving patient representatives. An essential contribution to this is the teaching of clear and structured communication [
29].
The present study shows a statistically significant reduction in the total number of pAEs after a communication training was introduced. The communication training was substantially more focused on communication than on teamwork compared to other interventions. In most other interventions, teamwork training is implemented; but it requires sufficient preparation and resources at the site, such as time capacity or operating rooms [
30]. In the categories of diagnostic errors and birth position, on which clear and structured communication has a great influence, a significant reduction could be achieved. These aspects were specifically trained in the communication lessons and the subsequent microteaching sessions. The training contents of more structured handovers with the ISBAR strategy as well as the introduction of the closed-loop communication technique were intended to contribute to the avoidance of errors and ambiguities. The closed-loop technique, which was developed in emergency medicine, shows small to medium effects in the literature [
31,
32], where it was mostly evaluated in the context of broader training on teamwork in critical situations [
31‐
33]. The training sessions in the TeamBaby project focused on a short intervention aiming at communication and behavior change. Afterwards, healthcare workers perceived a lower risk of potential triggers for pAEs.
Another important aspect of the training was the role of hierarchies. Role-plays were used to practice giving timely warnings of possible avoidable mistakes in critical situations, regardless of the professional and hierarchical position of the person acting. The so-called "Speaking-Up" should help to overcome hierarchical levels to address safety-relevant concerns and thus increase patient safety [
34]. The effectiveness of trainings using Speaking-Up has been demonstrated in some studies, but the results are mixed [
35]. The increase in self-efficacy beliefs in difficult situations in the present study is an indication of the effect of Speaking Up training to object to those superior in hierarchy.
In general, staff reported that the training increased their awareness of safe and effective communication. The general awareness of the role of communication was strengthened in the training using an educational film [
36]. Positive effects of the training may therefore also be because the awareness of possible negative effects of communication and the awareness of one's role improved. A conceivable indication here is the result, paradoxical at first glance (marginally significant), that the staff's assessment of their general self-efficacy decreased after the training. It is possible that the training has increased the sensitivity for the susceptibility to errors in communication so that the participants assess their communication skills more critically than before the training as demonstrated by the self-reported data. Although these results indicate the overall effectiveness of the training and explain potential mechanisms, it must be borne in mind that the hypothesis, namely that the reduction of pAE can be achieved by improving interdisciplinary communication, cannot be answered without alternative explanations. Due to the large population and the time passed between 2018 and 2020, a specific reduction in pAE cannot be attributed to the intervention. It is possible that changes in the background or a heightened awareness for patient safety risks accounted for the change in pAE.
An extensive review of the effectiveness of communication training in obstetrics examined 71 studies, which showed large differences in the quality and design of communication training and studies [
16]. In most cases, broader teamwork training with a focus on communication was implemented and had an impact on proximal outcomes such as communication skills or behavior. In total, however, only three studies looked at the direct impact on patient safety. In two of these three studies, a positive effect of communication training on patient safety could be demonstrated. However, the constructs examined were still very heterogeneous; only one study used the AOI as an outcome parameter. Especially in the comparison of these three studies, it can be stated that both the concrete and change-sensitive identification and recording of (avoidable) adverse events are of high value, as well as training aimed at concrete behavioral changes. If only very serious adverse events ("sentinel events") or insufficient risk factors are collected to determine the preventability of corresponding events, positive effects of communication training may not be identified. On the other hand, it is also conceivable that an intervention was not sufficiently theory-based or comprehensive to achieve positive effects. Accordingly, both the interventions and the data evaluation should be planned in an interdisciplinary team with the stakeholders on-site and adapted to the circumstances.
To reduce pAEs through training, it is particularly important to understand their key active components making the training work. As early as the 1990s, a key role was attributed to the human factors. Even though AEs caused by human factors will never be 100% preventable, effective risk management should always include communication training to strengthen the human factors in the team [
5]. However, it seems to be the combination of different aspects, but further research is needed to test this in more detail.
In the future, in addition to strengthening human factors, it may also be useful to include addressing risk factors in a structured risk management. In the present work, risk factors were filtered in the cases of an AE. If risk factors are differentiated into those that can be controlled and those that cannot be or controlled, this can be included in individualized risk management. In this way, risk factors that can be controlled can be clearly communicated to the team and appropriate prophylactic measures can be taken. Checklists, adapted for the most important obstetric risk factors, are useful tools for risk-conscious individual birth management [
37].
Limitations must be considered when interpreting the present results. In the evaluation of the individual categories of pAEs, no statistically significant reductions were found in peripartum therapy delay. This can be explained, among other reasons, by an insufficient differentiation of the categories so that our data cannot reflect pAE according to whether they occurred before, during or after birth. In the future, it may be useful to differentiate these categories into antepartum, peripartum, and postpartum birth management. In addition, management of special obstetric emergencies such as emergency cesarean section, shoulder dystocia, or postpartum hemorrhage should be considered. There was no statistically significant improvement in inadequate fetal monitoring, indicating that not all pAE improved over the course of the analysis. However, this does not impact the study findings negatively because effects of improvement in this category are more likely to come from special training that teaches the interpretation of CTGs during labor than from general communication training [
38].
The main limitation is the lack of a control group. A pre-post comparison of the period after the training without the possibility of controlling the results for the communication training was conducted. Therefore, it cannot be ruled out that the reduction in pAEs is at least partly due to the further development in obstetric care or to other influences for which no controls could be conducted. It is possible that a heightened awareness had an unsystematic impact on pAE instead of systematic training effects. Moreover, the background variations must be considered, as the sample size comprises a large number of cases and this complicates the specific evaluation of the reduction in pAEs. In addition, the impact of the COVID-19 pandemic must be considered. Theoretically, the training helped to mitigate the negative effects of the pandemic, but we also do not have a comparator with this regard to test for the mechanisms. Accordingly, we cannot link the reduction of pAE to the intervention without alternative explanations and future research is needed to establish whether communication trainings can actively reduce pAE.
In addition, only half of the training participants completed the questionnaires in a form that could be evaluated. Thus, the evaluation results must be interpreted with caution due to the potentially selective drop-out and lower power. It is possible that mostly HCW who experienced the intervention as helpful answered the questionnaire at the second time point while the more critical ones did not. The hospital where the intervention and analyses were conducted is a teaching hospital with a corresponding high staff turnover. In the period between the measurements before and after the intervention, people may have left the obstetrics department and accordingly were no longer available to respond. In collecting the data, the present study used retrospective routine data that can be extracted and analyzed at any time point. In addition, the conditions for pAEs, especially the communication aspect, cannot always be retrieved from the available documentation systems with perfect validity and reliability. This reduces the quality and consequently the internal validity of the extracted routine data. Thus, this opens many further questions, which should be researched in the future.