Aggregate analysis of sentinel events as a strategic tool in safety management can contribute to the improvement of healthcare safety

Aggregate analysis of sentinel events as a strategic tool in safety management can contribute to... Abstract Objective To examine if clustering of root causes of sentinel events (SEs) can contribute to organisational improvement of healthcare and patient safety by providing insight into organisational risk factors, patterns and trends. Design Retrospective, cross-sectional review of SEs from a hospital database reported to the Board of directors in 2016. Setting A regional teaching hospital in the Netherlands. Intervention(s) Clustering of characteristics and variables of SEs to establish vulnerabilities and patterns of failure factors of the organisation. Main Outcome Measure(s) Characteristics and contributory causes of failure of SEs identified via root cause analysis (RCA). Outcomes reported using descriptive statistics. Results A total of 21 events were included involving 21 patients. Mean age was 56.7 years (SD 24.4), 71.4% were above 50 years of age. In 81.8%, the care was multi-disciplinary and in 76.2% the event resulted in permanent harm or injury. Of the 132 identified contributory root causes, most were related to human factors (53.8%) and organisational factors (40.2%). Technical and patient-related factors were identified in 3.0%. Organisational improvement strategies focused on the care of elderly patients, patients subjected to multi-disciplinary care and on improving knowledge, protocols and coordination of care. Conclusion Clustering variables of SEs and contributory factors of failure through RCA helps to delineate a hospital-specific profile by providing a detailed insight into risk factors, patterns and trends in an organisation and to determine the best strategies for improvement by drawing lessons across events. root cause analysis, adverse events, quality of healthcare, incident reporting and analysis, risk management, safety management Introduction Safety and quality is essential for the delivery of optimal healthcare and remains an important issue. The first step to improve safety is to gain insight into the frequency and seriousness of unsafe situations. A systematic review of the literature conducted in 2018 showed that the incidence of in-hospital adverse events was 9.2% and in 7.4% the event was lethal. Almost half of the reported events were considered preventable [1]. The burden of adverse events on healthcare resources is significant; it is estimated to amount to 1% of the national healthcare budget [2]. Most attention has been directed at hospitals because of the high-risk procedures and hazardous environment [3]. Several reports have highlighted the need for healthcare organisations to monitor and learn from adverse events [4]. The importance of developing effective systems for learning from failure is still growing [5]. An incident reporting system is a tool that has the potential of providing insight into the occurrence of events, making it possible to identify causes and risk factors [6]. Incident reporting and analysis originated in safety-critical industries and in the transport sector [6]. Identifying and learning from events helps in continuously making the healthcare system safer [6, 7]. To protect and improve public healthcare, the safety and quality of the Dutch healthcare system is monitored by the Dutch Health and Youth Care Inspectorate (IGJ). The IGJ oversees and regulates all Dutch healthcare providers and professionals. It is mandatory for all Dutch healthcare organisations to report a sentinel event (SE), i.e. an unintended and unexpected event related to the quality of care which caused death or serious harm to a patient. This is an important tool contributing to the improvement of quality and safety [8]. Appropriate response in case of an SE includes conducting a timely and reliable root cause analysis (RCA), followed by developing improvements, implementing these improvements and finally monitoring the effectiveness of the improvements [9]. RCA has been widely adopted as a method to help clinicians and healthcare organisations to analyse and learn from SEs by systematically investigating events, correcting causes and preventing reoccurrence [10, 11]. As RCA of an SE is based on an in-depth analysis of a single event, it often provides tools to improve care in specific situations. Although RCA is a powerful tool for developing and maintaining awareness of risks, it remains challenging to effectively detect organisational and management factors contributing to SEs within an organisation [12, 13]. Organisations tend to approach each RCA independently, rather than drawing lessons across investigations [12]. It has been hypothesised that clustering of data of various SEs can help in capturing organisational risk factors and defects across events, thus supporting system improvement in organisations [3]. The objective of this study was to examine if clustering of root causes of SEs can contribute to improvement of healthcare and patient safety by identifying vulnerabilities and patterns of failure factors of an organisation. Methods Study design and setting This study was conducted in the Zaans Medical Center (ZMC), a regional teaching hospital in the Netherlands with 293 beds, 1300 employees, 120 medical specialists and 27.000 admissions every year. Employees of the ZMC are expected to report suspected SEs to the Board of directors by means of a reporting system. The Board members determine, after consulting experts, whether the incident should be classified as an SE and should be reported to the inspectorate. In the current study, we retrospectively analysed all suspected SEs reported to the board of directors in 2016. The reason for conducting an extensive analysis focusing on identified root causes was that the number of SEs reported in 2016 was significantly higher than in previous years. According to the policy of the ZMC, this study met the criteria for organisational improvement of patient and healthcare safety and was exempt from ethics review, as all data were anonymized and de-identified. All reported SEs were analysed by means of the Systematic Incident Reconstruction and Evaluation (SIRE) method, a Dutch prototype RCA method. SIRE, adapted from a method developed by the National Center for Patient Safety of the Department of Veterans Affairs, was developed to examine and analyse adverse events [14]. The process focuses primarily on systems and not on individuals [10]. Basic and contributing factors are discovered in a process similar to the diagnosis of a disease, i.e. by answering the following questions: What happened? Why did it happen? How can we prevent it happening again? How will we know if the actions we took made a difference? This method also includes the perspectives of the patient and the patient’s family. RCAs are performed in the ZMC by an experienced multi-disciplinary team consisting of clinicians, nurses, staff members and a coordinator. If necessary, internal or external experts are consulted as well. The team does not include persons who were involved in the incident under investigation. RCAs and reports are considered to be confidential and are treated accordingly, but improvement measures are broadly disseminated within the organisation. The following procedure was conducted in the ZMC. The first step in the RCA process is to identify and reveal underlying dominant risk and failure factors, so-called root causes. Root causes are used to provide a more realistic view of how systems work, as well as to create effective and lasting solutions [15–17]. In the second step, the identified root causes are classified by linking them to one of the five main categories and subcategories of the medical version of the Eindhoven Classification Model (ECM) of system failure, Table 1 [17]. Table 1 Description of the main and subcategories of the Eindhoven Classification Model Main category    Subcategory  Code  Description  Technical    External  T-ex  Technical failures beyond the control and responsibility of the investigating organisation.      Design  TD  Failures due to poor design of equipment, software, labels or forms.      Construction  TC  Correct design, which was not constructed properly or was set up in inaccessible areas.      Materials  TM  Material defects not classified under TD or TC.  Organisational    External  O-ex  Failures at an organisational level beyond the control and responsibility of the investigating organisation, such as in another department of area (address by collaborative systems).      Transfer of knowledge  OK  Failures resulting from inadequate measures taken to ensure that situational or domain-specific knowledge or information is transferred to all new or inexperienced staff.      Protocols  OP  Failures relating to the quality and availability of the protocols within the department (too complicated, inaccurate, unrealistic, absent, or poorly presented).      Management priorities  OM  Internal management decisions in which safety is relegated to an inferior position when faced with conflicting demands or objectives. This is a conflict between production needs and safety. Example: decisions that are made about staffing levels.      Culture  OC  Failures resulting from the collective approach and its attendant modes of behaviour to risks in the investigating organisation.  Human    External  H-ex  Human failures originating beyond the control and responsibility of the investigating organisation. This could apply to individuals in another department.    Knowledge-based behaviour  Knowledge-based behaviour  HKK  The inability of an individual to apply their existing knowledge to a novel situation. Example: a trained blood bank technologist who is unable to solve a complex antibody identification problem.    Rule-based behaviour  Qualifications  HRQ  The incorrect fit between an individual training or education and a particular task. Example: expecting a technician to solve the same type of difficult problems as a technologist.      Coordination  HRC  A lack of task coordination within a healthcare team in an organisation. Example: an essential task not being performed because everyone thought that someone else had completed the task.      Verification  HRV  The correct and complete assessment of a situation including related conditions of the patient and materials to be used before starting the intervention. Example: failure to correctly identify a patient by checking the wristband.      Intervention  HRI  Failures that result from faulty task planning and execution. Example: washing red cells by the same protocol as platelets.      Monitoring  HRM  Monitoring a process or patient status. Example: a trained technologist operating an automated instrument and not realising that a pipette dispenses reagents is clogged.    Skill-based behaviour  Slips  HSS  Failures in the performance of highly developed skills. Example: a technologist adding drops of reagents to a row of test tubes and then missing the tube or a computer entry error.      Tripping  HST  Failures in whole body movements. These errors are often referred to as ‘slipping, tripping, or falling’. Examples: a blood bag slipping out of one’ s hands and breaking or tripping over a loose tile on the floor.  Patient-related    Patient-related factor  PRF  Failures related to patient characteristics or conditions, which are beyond the control of staff and influence treatment.  Main category    Subcategory  Code  Description  Technical    External  T-ex  Technical failures beyond the control and responsibility of the investigating organisation.      Design  TD  Failures due to poor design of equipment, software, labels or forms.      Construction  TC  Correct design, which was not constructed properly or was set up in inaccessible areas.      Materials  TM  Material defects not classified under TD or TC.  Organisational    External  O-ex  Failures at an organisational level beyond the control and responsibility of the investigating organisation, such as in another department of area (address by collaborative systems).      Transfer of knowledge  OK  Failures resulting from inadequate measures taken to ensure that situational or domain-specific knowledge or information is transferred to all new or inexperienced staff.      Protocols  OP  Failures relating to the quality and availability of the protocols within the department (too complicated, inaccurate, unrealistic, absent, or poorly presented).      Management priorities  OM  Internal management decisions in which safety is relegated to an inferior position when faced with conflicting demands or objectives. This is a conflict between production needs and safety. Example: decisions that are made about staffing levels.      Culture  OC  Failures resulting from the collective approach and its attendant modes of behaviour to risks in the investigating organisation.  Human    External  H-ex  Human failures originating beyond the control and responsibility of the investigating organisation. This could apply to individuals in another department.    Knowledge-based behaviour  Knowledge-based behaviour  HKK  The inability of an individual to apply their existing knowledge to a novel situation. Example: a trained blood bank technologist who is unable to solve a complex antibody identification problem.    Rule-based behaviour  Qualifications  HRQ  The incorrect fit between an individual training or education and a particular task. Example: expecting a technician to solve the same type of difficult problems as a technologist.      Coordination  HRC  A lack of task coordination within a healthcare team in an organisation. Example: an essential task not being performed because everyone thought that someone else had completed the task.      Verification  HRV  The correct and complete assessment of a situation including related conditions of the patient and materials to be used before starting the intervention. Example: failure to correctly identify a patient by checking the wristband.      Intervention  HRI  Failures that result from faulty task planning and execution. Example: washing red cells by the same protocol as platelets.      Monitoring  HRM  Monitoring a process or patient status. Example: a trained technologist operating an automated instrument and not realising that a pipette dispenses reagents is clogged.    Skill-based behaviour  Slips  HSS  Failures in the performance of highly developed skills. Example: a technologist adding drops of reagents to a row of test tubes and then missing the tube or a computer entry error.      Tripping  HST  Failures in whole body movements. These errors are often referred to as ‘slipping, tripping, or falling’. Examples: a blood bag slipping out of one’ s hands and breaking or tripping over a loose tile on the floor.  Patient-related    Patient-related factor  PRF  Failures related to patient characteristics or conditions, which are beyond the control of staff and influence treatment.  Table 1 Description of the main and subcategories of the Eindhoven Classification Model Main category    Subcategory  Code  Description  Technical    External  T-ex  Technical failures beyond the control and responsibility of the investigating organisation.      Design  TD  Failures due to poor design of equipment, software, labels or forms.      Construction  TC  Correct design, which was not constructed properly or was set up in inaccessible areas.      Materials  TM  Material defects not classified under TD or TC.  Organisational    External  O-ex  Failures at an organisational level beyond the control and responsibility of the investigating organisation, such as in another department of area (address by collaborative systems).      Transfer of knowledge  OK  Failures resulting from inadequate measures taken to ensure that situational or domain-specific knowledge or information is transferred to all new or inexperienced staff.      Protocols  OP  Failures relating to the quality and availability of the protocols within the department (too complicated, inaccurate, unrealistic, absent, or poorly presented).      Management priorities  OM  Internal management decisions in which safety is relegated to an inferior position when faced with conflicting demands or objectives. This is a conflict between production needs and safety. Example: decisions that are made about staffing levels.      Culture  OC  Failures resulting from the collective approach and its attendant modes of behaviour to risks in the investigating organisation.  Human    External  H-ex  Human failures originating beyond the control and responsibility of the investigating organisation. This could apply to individuals in another department.    Knowledge-based behaviour  Knowledge-based behaviour  HKK  The inability of an individual to apply their existing knowledge to a novel situation. Example: a trained blood bank technologist who is unable to solve a complex antibody identification problem.    Rule-based behaviour  Qualifications  HRQ  The incorrect fit between an individual training or education and a particular task. Example: expecting a technician to solve the same type of difficult problems as a technologist.      Coordination  HRC  A lack of task coordination within a healthcare team in an organisation. Example: an essential task not being performed because everyone thought that someone else had completed the task.      Verification  HRV  The correct and complete assessment of a situation including related conditions of the patient and materials to be used before starting the intervention. Example: failure to correctly identify a patient by checking the wristband.      Intervention  HRI  Failures that result from faulty task planning and execution. Example: washing red cells by the same protocol as platelets.      Monitoring  HRM  Monitoring a process or patient status. Example: a trained technologist operating an automated instrument and not realising that a pipette dispenses reagents is clogged.    Skill-based behaviour  Slips  HSS  Failures in the performance of highly developed skills. Example: a technologist adding drops of reagents to a row of test tubes and then missing the tube or a computer entry error.      Tripping  HST  Failures in whole body movements. These errors are often referred to as ‘slipping, tripping, or falling’. Examples: a blood bag slipping out of one’ s hands and breaking or tripping over a loose tile on the floor.  Patient-related    Patient-related factor  PRF  Failures related to patient characteristics or conditions, which are beyond the control of staff and influence treatment.  Main category    Subcategory  Code  Description  Technical    External  T-ex  Technical failures beyond the control and responsibility of the investigating organisation.      Design  TD  Failures due to poor design of equipment, software, labels or forms.      Construction  TC  Correct design, which was not constructed properly or was set up in inaccessible areas.      Materials  TM  Material defects not classified under TD or TC.  Organisational    External  O-ex  Failures at an organisational level beyond the control and responsibility of the investigating organisation, such as in another department of area (address by collaborative systems).      Transfer of knowledge  OK  Failures resulting from inadequate measures taken to ensure that situational or domain-specific knowledge or information is transferred to all new or inexperienced staff.      Protocols  OP  Failures relating to the quality and availability of the protocols within the department (too complicated, inaccurate, unrealistic, absent, or poorly presented).      Management priorities  OM  Internal management decisions in which safety is relegated to an inferior position when faced with conflicting demands or objectives. This is a conflict between production needs and safety. Example: decisions that are made about staffing levels.      Culture  OC  Failures resulting from the collective approach and its attendant modes of behaviour to risks in the investigating organisation.  Human    External  H-ex  Human failures originating beyond the control and responsibility of the investigating organisation. This could apply to individuals in another department.    Knowledge-based behaviour  Knowledge-based behaviour  HKK  The inability of an individual to apply their existing knowledge to a novel situation. Example: a trained blood bank technologist who is unable to solve a complex antibody identification problem.    Rule-based behaviour  Qualifications  HRQ  The incorrect fit between an individual training or education and a particular task. Example: expecting a technician to solve the same type of difficult problems as a technologist.      Coordination  HRC  A lack of task coordination within a healthcare team in an organisation. Example: an essential task not being performed because everyone thought that someone else had completed the task.      Verification  HRV  The correct and complete assessment of a situation including related conditions of the patient and materials to be used before starting the intervention. Example: failure to correctly identify a patient by checking the wristband.      Intervention  HRI  Failures that result from faulty task planning and execution. Example: washing red cells by the same protocol as platelets.      Monitoring  HRM  Monitoring a process or patient status. Example: a trained technologist operating an automated instrument and not realising that a pipette dispenses reagents is clogged.    Skill-based behaviour  Slips  HSS  Failures in the performance of highly developed skills. Example: a technologist adding drops of reagents to a row of test tubes and then missing the tube or a computer entry error.      Tripping  HST  Failures in whole body movements. These errors are often referred to as ‘slipping, tripping, or falling’. Examples: a blood bag slipping out of one’ s hands and breaking or tripping over a loose tile on the floor.  Patient-related    Patient-related factor  PRF  Failures related to patient characteristics or conditions, which are beyond the control of staff and influence treatment.  The ECM coding system has been tested and validated in medical settings [18]. In this model, latent failures are classified first to increase the likelihood of discovering all causes underlying the event. The main categories of the ECM system are latent factors (technical and organisational), active failures (human), patient-related factors and unclassifiable factors, Table 1 [19, 20]. The RCA results in a report with an action plan that identifies causes and intended strategies for improvement to reduce the risk of similar events occurring in the future. The plan addresses responsibilities for implementation and strategies for measuring the effectiveness of the actions. Data collection The SEs reported in 2016 were obtained from the hospital database. In this database, all reported events are recorded anonymously. Events were included if they were classified as an SE and if the ZMC was primarily involved. For each recruited SE report, the following variables were extracted and analysed: baseline characteristics (patient age, gender), event characteristics (day and time of the incident, number and type of healthcare professionals involved, incident category, degree of harm, number of healthcare organisations involved and incident location), root causes as identified through RCA, and improvement measures that were developed and implemented. Types and locations of the events were categorised according to the Harvard Medical Practice Study classification [21]. Statistical analysis Data of the identified and recruited SEs were first summarised using descriptive statistics and frequency tables. We used the identified RCAs to extrapolate the detected root causes. Statistical analyses were performed using SPSS (version 20). The appropriate analyses were performed, including mean, median, percentages and standard deviation (SD). Results The total number of suspected SEs reported to the Board of Directors in 2016 was 28. After consultation of experts, 23 events were considered to be SEs. The 23 SEs were reported to the IGJ and subsequently evaluated by a multi-disciplinary team. For the current analysis, two of these events were excluded. One event was excluded because according to the Dutch Health Care inspectorate (IGJ) it was not an SE but a complication of a surgical procedure. One other event was excluded because it occurred in another healthcare organisation and was not related to the procedures of or the care provided by the ZMC. As a result, 21 SEs were included in the current analysis. Characteristics of the reported sentinel events The 21 included SEs involved a total of 21 patients. The characteristics of the patients and of the events are shown in Table 2. In 14 cases (66.7%) the patient was female. The mean age was 56.7 years (SD 24.4), and in 71.4% the patient was above 50 years of age. In seven cases (33.3%), at least one other healthcare organisation was involved. Table 2 Baseline and sentinel events characteristics Characteristics  Number (%)  Age (year)     <1  2 (9.5)   20–49  4 (19.0)   50–59  3 (14.3)   60–69  4 (19.0)   70–79  6 (28.6)   >80  2 (9.5)  Gender     Male  7 (33.3)   Female  14 (66.7)  Time     Daytime (7 am–5 pm)  14 (66.7)   Evening and night (5 pm–7 am)  7 (33.3)  Event location     Floor units  9 (42.9)   Operating rooms  6 (28.6)   Emergency department  3 (14.3)   Other  3 (14.3)  Event type     Operations  7 (33.3)   Diagnostic  5 (23.8)   Medications  3 (14.3)   Therapeutic  2 (9.5)   Peripartum  2 (9.5)   Other  2 (9.5)  Professional involved     Medical specialist  58 (43.6)   Nurse  23 (17.3)   Resident/registrars  17 (12.8)   Operation personnel  14 (10.5)   Management  10 (7.5)   Other  11 (8.3)   Total  133 (100)  Number of professionals involved in one event     3–4  8 (38.1)   5–6  6 (28.6)   7–8  5 (23.8)   ≥9  2 (9.5)  Number of medical specialists involved in one event     1  5 (23.8)   2  6 (28.6)   3  4 (19.0)   4  3 (14.3)   5  1 (4.8)   6  2 (9.5)  Consequences     No permanent inconvenience/injuries  2 (9.5)   Permanent injuries/harm  16 (76.2)    Death  6 (28.6)   Unknown  3 (14.3)  Characteristics  Number (%)  Age (year)     <1  2 (9.5)   20–49  4 (19.0)   50–59  3 (14.3)   60–69  4 (19.0)   70–79  6 (28.6)   >80  2 (9.5)  Gender     Male  7 (33.3)   Female  14 (66.7)  Time     Daytime (7 am–5 pm)  14 (66.7)   Evening and night (5 pm–7 am)  7 (33.3)  Event location     Floor units  9 (42.9)   Operating rooms  6 (28.6)   Emergency department  3 (14.3)   Other  3 (14.3)  Event type     Operations  7 (33.3)   Diagnostic  5 (23.8)   Medications  3 (14.3)   Therapeutic  2 (9.5)   Peripartum  2 (9.5)   Other  2 (9.5)  Professional involved     Medical specialist  58 (43.6)   Nurse  23 (17.3)   Resident/registrars  17 (12.8)   Operation personnel  14 (10.5)   Management  10 (7.5)   Other  11 (8.3)   Total  133 (100)  Number of professionals involved in one event     3–4  8 (38.1)   5–6  6 (28.6)   7–8  5 (23.8)   ≥9  2 (9.5)  Number of medical specialists involved in one event     1  5 (23.8)   2  6 (28.6)   3  4 (19.0)   4  3 (14.3)   5  1 (4.8)   6  2 (9.5)  Consequences     No permanent inconvenience/injuries  2 (9.5)   Permanent injuries/harm  16 (76.2)    Death  6 (28.6)   Unknown  3 (14.3)  Table 2 Baseline and sentinel events characteristics Characteristics  Number (%)  Age (year)     <1  2 (9.5)   20–49  4 (19.0)   50–59  3 (14.3)   60–69  4 (19.0)   70–79  6 (28.6)   >80  2 (9.5)  Gender     Male  7 (33.3)   Female  14 (66.7)  Time     Daytime (7 am–5 pm)  14 (66.7)   Evening and night (5 pm–7 am)  7 (33.3)  Event location     Floor units  9 (42.9)   Operating rooms  6 (28.6)   Emergency department  3 (14.3)   Other  3 (14.3)  Event type     Operations  7 (33.3)   Diagnostic  5 (23.8)   Medications  3 (14.3)   Therapeutic  2 (9.5)   Peripartum  2 (9.5)   Other  2 (9.5)  Professional involved     Medical specialist  58 (43.6)   Nurse  23 (17.3)   Resident/registrars  17 (12.8)   Operation personnel  14 (10.5)   Management  10 (7.5)   Other  11 (8.3)   Total  133 (100)  Number of professionals involved in one event     3–4  8 (38.1)   5–6  6 (28.6)   7–8  5 (23.8)   ≥9  2 (9.5)  Number of medical specialists involved in one event     1  5 (23.8)   2  6 (28.6)   3  4 (19.0)   4  3 (14.3)   5  1 (4.8)   6  2 (9.5)  Consequences     No permanent inconvenience/injuries  2 (9.5)   Permanent injuries/harm  16 (76.2)    Death  6 (28.6)   Unknown  3 (14.3)  Characteristics  Number (%)  Age (year)     <1  2 (9.5)   20–49  4 (19.0)   50–59  3 (14.3)   60–69  4 (19.0)   70–79  6 (28.6)   >80  2 (9.5)  Gender     Male  7 (33.3)   Female  14 (66.7)  Time     Daytime (7 am–5 pm)  14 (66.7)   Evening and night (5 pm–7 am)  7 (33.3)  Event location     Floor units  9 (42.9)   Operating rooms  6 (28.6)   Emergency department  3 (14.3)   Other  3 (14.3)  Event type     Operations  7 (33.3)   Diagnostic  5 (23.8)   Medications  3 (14.3)   Therapeutic  2 (9.5)   Peripartum  2 (9.5)   Other  2 (9.5)  Professional involved     Medical specialist  58 (43.6)   Nurse  23 (17.3)   Resident/registrars  17 (12.8)   Operation personnel  14 (10.5)   Management  10 (7.5)   Other  11 (8.3)   Total  133 (100)  Number of professionals involved in one event     3–4  8 (38.1)   5–6  6 (28.6)   7–8  5 (23.8)   ≥9  2 (9.5)  Number of medical specialists involved in one event     1  5 (23.8)   2  6 (28.6)   3  4 (19.0)   4  3 (14.3)   5  1 (4.8)   6  2 (9.5)  Consequences     No permanent inconvenience/injuries  2 (9.5)   Permanent injuries/harm  16 (76.2)    Death  6 (28.6)   Unknown  3 (14.3)  The types and locations of the events are shown in Table 2. Seven events (33.3%) were operation-related, five (23.8%) were related to the diagnostic procedure and three (14.3%) were medication-related. Almost half of the events occurred on the ward (42.8%) and 28.6% occurred in the operating theatre. In 18 of the 21 events (81.8%), at least two medical specialists or departments were involved. In 16 cases (76.2%), the event resulted in permanent harm or injury and in 6 cases (28.6%), the patient died. Involvement of healthcare professionals A total of 133 healthcare professionals were involved in the 21 events. The number of professionals involved in an event ranged from 3 to 11, with a mean of 6.3 (SD 2.5). Medical specialists were predominantly related to the SE in 43.6% of the cases, followed by nurses in 17.3% and resident physicians in 12.8%. The number of medical specialists related to the treatment at the time of the event ranged from 1 to 5, with a mean of 3 (Table 2). Causes of sentinel events Analysis of the 21 SEs resulted in the identification of 132 contributory root causes, ranging between 3 and 15 in a single event, with a mean of 6.3 (SD 3.8). When the root causes were classified into the five main categories according to the ECM taxonomy, 71 (53.8%) were human, 53 (40.2%) organisational, 4 (3.0%) technical and 4 (3.0%) patient-related. None of the root causes was unclassifiable. Technical, human and organisational root causes were subdivided into one or more specific subcategories (Table 1). Human factors were the most frequently identified root causes (n = 71). More than half of the human factors (53.5%) were related to knowledge-based behaviour (HKK), followed by verification (HRV) in 19.7% (Fig. 1). Of the organisational factors (n = 53), 66.0% were related to protocols and procedures (OP) followed by management priorities (OM) in 13.2%. In the category technical factors (n = 4), three were related to design (TD) and one to material defects (TM). External and patient-related factors were hardly identified, and hence an in-depth analysis was not possible. Figure 1 View largeDownload slide The distribution of the 132 identified root causes in the 21 SE analysed. Root causes were divided into the following categories: Human, Organisational, Technical and Patient-related. Data are presented as numbers (n) or percentage (%). Abbreviations: HKK, knowledge-based behaviour; HRV, verification; HRC, coordination; H-ex, external; OC, external; OP, protocols and procedures; OM, management priorities; OK, transfer of knowledge; TD, design; TM, material defects. Figure 1 View largeDownload slide The distribution of the 132 identified root causes in the 21 SE analysed. Root causes were divided into the following categories: Human, Organisational, Technical and Patient-related. Data are presented as numbers (n) or percentage (%). Abbreviations: HKK, knowledge-based behaviour; HRV, verification; HRC, coordination; H-ex, external; OC, external; OP, protocols and procedures; OM, management priorities; OK, transfer of knowledge; TD, design; TM, material defects. Improvement strategies Prevention strategies in case of a single event tend to focus on that event, aiming to improve local work condition and processes, improvement strategies on the organisational level were rarely proposed. Based on the identified root causes in our analysis, various interventions were carried out. A hospital-based protocol was designed in collaboration with medical specialists, nurses and management, describing the important aspects of patient care, including the processes of admission, diagnosis, treatment and discharge. Furthermore, the protocol draws attention to the special requirements necessary for treating elderly and complex patients requiring multi-disciplinary care. The protocol describing the different roles of a medical specialist was updated, describing and clarifying the cooperation between departments and healthcare workers and the coordination of multi-disciplinary care and formalising the role of the medical specialist who must guide and coordinate the care provided for the patient. To tackle human and organisational factors, which occurred most frequently, an organisation-wide programme has been constructed to improve the knowledge, quality and availability of protocols and procedures and to enhance safety behaviour by means of training programmes. Regular audits were planned to examine the safety behaviour of departments. Discussion General findings and interpretation Healthcare should be available to people when necessary and should be effective and safe. Incident reporting and analysis is widely recognised as an important tool for improving reliability and patient safety in the care delivery process [23]. An increase in safety can only be achieved if interventions tackle the right underlying causes. This study reports on accumulated data of SEs and examined if clustering of root causes of SEs, i.e. events with significant harm to patients, helps in identifying the weaknesses of an organisation. In 21 included SEs, 132 root causes were identified, illustrating the multi-factorial nature of events. Events predominantly occurred on wards and the operation theatre and involved elderly patients and patients receiving multi-disciplinary care. The events had permanent consequences in 76.2% of the cases. Root causes were predominantly related to human factors (knowledge-based behaviour, verification and coordination) and organisational factors (protocols and procedures, management priorities and transfer of knowledge). On the basis of our analysis, several improvement strategies were carried out on both local and organisational level. A hospital-based protocol was designed regarding the care of patients, especially the elderly and patients receiving multi-disciplinary care. An organisational programme was developed to improve knowledge and safety behaviour by training healthcare professionals. The coordinating role of medical specialists was formalised to improve the provided multi-disciplinary care. The quality and availability of protocols and process descriptions were evaluated and improved. Proposed prevention strategies tackle both local, organisational and management factors. However, a clear framework is lacking to identify which types of incidents can be prevented with which type of interventions [12]. Clustering of RCAs seems valuable for identifying risk factors, undesirable patterns and trends as it focusses on systems and processes on an organisational level; a hospital-specific profile can be delineated. As hospitals are complex organisations, practically every step in the precarious process can have a negative effect on patient safety and healthcare safety. A safety management approach is crucial to identify factors and preconditions affecting the organisation [24, 25]. Prevention strategies of a single event tend to focus on the surroundings and circumstances of that event, aiming to improve local work procedures and processes, making it difficult to detect organisational vulnerabilities and patterns of failure and subsequently recommend improvement on that level [22]. Improvements strategies on the organisational level were rarely proposed. Aggregate analysis of SEs is a systematic process; priorities for improvement are based on multiple cases, rather than just one case. The importance of this approach is that it recognises that organisations have inherent weaknesses and that failures will manifest in local working conditions; focusing on individual responsibility is likely to be ineffective [11]. A systemic, rather than the individual view is likely to support continuous improvement and promote future reporting of failures in healthcare [13]. Comparison with previous studies After extensively reviewing the literature, we did not encounter other studies reporting on data of SEs of a single organisation. Nevertheless, the identified root causes in our study, predominantly human and organisational, are in accordance with previous studies [26–28]. The identified factors in the study of Smits et al. [26] were predominantly labelled as organisational and human, considered preventable in 66%–93% of the cases and related to permanent disability in 13–20%. Failures related to the quality and availability of protocols were frequently related to permanent disability. The root causes of failure of 522 unintended events at ten emergency departments in the Netherlands were predominantly human and organisational [25]. In 25%, the event was related to the cooperation between departments and healthcare workers within an organisation. Nuckols et al. [28] reported the characteristics of 2246 identified incidents in the USA. The events occurred primarily in floor units (50%) and were related to medication, operations and procedures. Similar to the improvement strategies in our study, most reported strategies were quality assurance/peer review, training, improvement of organisational procedures, improvement of sources of information and evaluation [1, 22, 27]. Prevention strategies at the organisational level were least often considered. Although, it seems worthwhile to direct interventions on organisational causes because organisational causes are nearly always believed to be preventable [26]. Strengths and limitations Our study has various advantages. We used an existing reporting system for identifying events, and 96% of the identified SEs were included. RCA was performed directly with the assessment of the SEs; the data were therefore original and accurate, not based on summaries of compiled information. Root causes were classified via ECM, which is a well-established and evidence-based framework. A limitation of the current study is that we were unaware of the intentions and attitudes of professionals regarding reporting of events and were not able to correct for possible underreporting. All reporting systems rely on professionals submitting events occurring within routine work situations, but many events will go unreported, with doctors being less likely than nurses to report events [13, 29]. Besides, although an SE is clearly defined, there is still uncertainty whether events should be classified as SEs. Unintended events that occur more frequently are believed to share the same underlying failure factors [30, 31]. As we were not able to accurately identify unintended incidents, we did not include them in our analysis. In the current analysis, all identified root causes were considered equal, while in clinical practice this is not the case. Contributory factors can vary according to the level of proximity to the event [11]. Caution must be exercised, as only 21 SEs were included. To produce reliable results, it has been suggested to include at least 50 events, preferably across organisations [27]. The advantage of the current method is that a hospital-specific profile can be delineated to determine where best to focus strategies for improvement. The reliability and generalisability of this approach can be improved by applying this method in several organisations on a regular basis. Implications for practice and future research Clustering data of events helps to target interventions and to increase organisational learning; it is a process of creating and applying valid knowledge to improve patient safety of an organisation. Aggregation of RCA can help to build organisational and technical defences. Clear information and action feedback are fundamental in supporting continuous improvement in the healthcare processes [25]. Information concerning root causes of failures and corrective actions should be widely disseminated across organisations at the regional and the national level to increase awareness, prevent recurrence and improve safety. Effective safety feedback systems represent an ongoing cyclical process of reporting, evaluation and corrective actions [25]. Formulating corrective actions is more difficult than finding failure causes [23]. How organisations should implement corrective actions may vary in practice and is an important topic for further research. We want to encourage others to apply this framework to help improve the healthcare system. The advantage is that organisations can learn from each other, especially from implemented improvements. To be able to achieve this, it is imperative to use the same taxonomy of failure factors in order to enable regional, national and international comparisons. The future lies in targeted incident reporting, robust analysis with the engagement of healthcare professionals and evaluation of the effectiveness of improvements. Incident reporting must be coupled with visible, sustainable improvements and, most importantly, dissemination of the acquired knowledge. Conclusion Incident reporting and analysis is crucial for improving healthcare safety. RCA has been adopted as a method to analyse and identify root causes of failure. The current study describes a strategic tool in safety management: by aggregating data of SEs, a framework of contributing root causes of failure is produced. A hospital-specific profile can be delineated to provide detailed insight into organisational patterns and risk factors, but most importantly to determine where best to focus organisational and management changes to improve patient safety and healthcare safety in the entire organisation, by drawing lessons across events. Acknowledgements The authors want to thank all those involved in the reported events. Furthermore, we thank the members of the multi-disciplinary team that conducted the analysis of the SEs of the ZMC. Special thanks to the patients and families for their contributions. Finally, we thank the hospital and Board of directors for making this report and its publication possible. Funding None. Data sharing We would be happy to share the available data of this study. Ethical approval According to the policy of the ZMC, this work was exempt from ethics review, as all data were anonymised and de-identified. References 1 de Vries EN, Ramrattan MA, Smorenburg SM et al.  . The incidence and nature of in-hospital adverse events: a systematic review. Qual Saf Health Care  2008; 17: 216– 23. Google Scholar CrossRef Search ADS PubMed  2 Hoonhout LH, de Bruijne MC, Wagner C et al.  . Direct medical costs of adverse events in Dutch hospitals. BMC Health Serv Res  2009; 9: 27. Google Scholar CrossRef Search ADS PubMed  3 Harmsen M, Gaal S, van Dulmen S et al.  . Patient safety in Dutch primary care: a study protocol. Implement Sci  2010; 5: 50– 8. Google Scholar CrossRef Search ADS PubMed  4 Hutchinson A, Young TA, Cooper KL et al.  . Trends in healthcare incident reporting and relationship to safety and quality data in acute hospitals: results from the National Reporting and Learning System. Qual Saf Health Care  2009; 18: 5– 10. Google Scholar CrossRef Search ADS PubMed  5 Williams SK, Osborn SS. The development of the National Reporting and Learning System in England and Wales, 2001–2005. Med J Aust  2006; 184: 65– 68S. 6 Barach P, Small SD. Reporting and preventing medical mishaps: lessons from non-medical near miss reporting systems. Br Med J  2000; 320: 759– 63. Google Scholar CrossRef Search ADS   7 Macrae C. The problem with incident reporting. BMJ Qual Saf  2016; 25: 71– 5. Google Scholar CrossRef Search ADS PubMed  8 Leistikow I, Mulder S, Vesseur J et al.  . Learning from incidents in healthcare: the journey, not the arrival, matters. BMJ Qual Saf  2017; 26: 252– 6. Google Scholar CrossRef Search ADS PubMed  9 Battles JB, Kaplan HS, Van der Schaaf TW et al.  . The attributes of medical event reporting systems: experience with a prototype medical event-reporting system for transfusion medicine. Arch Pathol Lab Med  1998; 122: 231– 8. Google Scholar PubMed  10 Iedema RA, Jorm C, Long D et al.  . Turning the medical gaze in upon itself: root cause analysis and the investigation of clinical error. Soc Sci Med  2006; 62: 1605– 15. Google Scholar CrossRef Search ADS PubMed  11 Leape LL. Why should we report adverse incidents? J Eval Clin Pract  1999; 5: 1– 4. Google Scholar CrossRef Search ADS PubMed  12 Wu AW, Lipshutz AK, Pronovost PJ. Effectiveness and efficiency of root cause analysis in medicine. JAMA  2008; 299: 685– 7. Google Scholar CrossRef Search ADS PubMed  13 Benn J, Koutantji M, Wallace L et al.  . Feedback from incident reporting: information and action to improve patient safety. Qual Saf Health Care  2009; 18: 11– 21. Google Scholar CrossRef Search ADS PubMed  14 Root cause analysis tools . Ann Arbor, MI: Department of Veterans Affairs, National Center for Patient Safety, 2007. www.va.gov/ncps/CogAids/RCA/index.html. 15 Mers TM. Medical event reporting system for transfusion medicine reference manual version 3.0 . Columbia University, New York, 2001. http://www.mers-tm.net. 16 van Vuuren W, Shea CE, van der Schaaf TW. The development of an incident analysis tool for the medical field . Eindhoven: Eindhoven University of Technology, 1997. http://alexandria.tue.nl/repository/books/493452.pdf. 17 Schaaf van der TW, Habraken MMP. PRISMA-Medical a brief description . Eindhoven: Eindhoven University of Technology, Faculty of Technology Management, Patient Safety Systems, 2005. 18 Smits M, Janssen JCJA, de Vet R et al.  . Analysis of unintended events in hospitals: inter-rater reliability of constructing causal trees and classifying root causes. Int J Qual Health Care  2009; 21: 292– 300. Google Scholar CrossRef Search ADS PubMed  19 Kaplan HS, Battles JB, Schaaf Van der TW et al.  . Identification and classification of the causes of events in transfusion medicine. Transfusion  1998; 38: 1071– 81. Google Scholar CrossRef Search ADS PubMed  20 Rasmussen J. Outlines of a hybrid model of the process operator. In: Sheridan TB, Johanssen G (eds). Monitoring behaviour and supervisory control . London: Plenium Press, 1976, 371– 384. Google Scholar CrossRef Search ADS   21 Harvard Medical Practice Study. Patients, doctors and lawyers: medical injury, malpractice litigation and patient compensation in New York. The Report of the Harvard Medical Practice Study to the State of New York . New York, NY: Department of Health, 1990. 22 Zegers M, de Bruijne MC, de Keizer B et al.  . The incidence, root-causes, and outcomes of adverse events in surgical units: implication for potential prevention strategies. Patient Saf Surg  2011; 5: 13. Google Scholar CrossRef Search ADS PubMed  23 Jansma JD, Zwart DL, Leistikow IP et al.  . Do specialty registrars change their attitudes, intentions and behaviour towards reporting incidents following a patient safety course? BMC Health Serv Res  2010; 10: 100– 9. Google Scholar CrossRef Search ADS PubMed  24 Lipshitz R, Popper M, Friedman VJ. A multifacet model of organizational learning. J Appl Behav Sci  2002; 38: 78– 98. Google Scholar CrossRef Search ADS   25 Anderson JE, Kodate N. Learning from patient safety incidents in incident review meetings: organisational factors and indicators of analytic process effectiveness. Safety Sci  2015; 80: 105– 14. Google Scholar CrossRef Search ADS   26 Smits M, Zegers M, Groenewegen PP et al.  . Exploring the causes of adverse events in hospitals and potential prevention strategies. Qual Saf Health Care  2010; 19: e5. Google Scholar PubMed  27 Smits M, Groenewegen PP, Timmermans DR et al.  . The nature and causes of unintended events reported at ten emergency departments. BMC Emerg Med  2009; 9: 16. Google Scholar CrossRef Search ADS PubMed  28 Nuckols TK, Bell DS, Liu H et al.  . Rates and types of events reported to established incident reporting systems in two US hospitals. Qual Saf Health Care  2007; 16: 164– 8. Google Scholar CrossRef Search ADS PubMed  29 Evans SM, Berry JG, Smith BJ et al.  . Attitudes and barriers to incident reporting: a collaborative hospital study. Qual Saf Health Care  2006; 15: 39– 43. Google Scholar CrossRef Search ADS PubMed  30 Kanse L, Schaaf van der TW, Vrijland ND et al.  . Error recovery in a hospital pharmacy. Ergonomics  2006; 49: 503– 16. Google Scholar CrossRef Search ADS PubMed  31 Wright L, van der Schaaf TW. Accident versus near miss causation: a critical review of the literature, an empirical test in the UK railway domain, and their implications for other sectors. J Hazard Mater  2004; 111: 105– 10. Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018. Published by Oxford University Press in association with the International Society for Quality in Health Care. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png International Journal for Quality in Health Care Oxford University Press

Aggregate analysis of sentinel events as a strategic tool in safety management can contribute to the improvement of healthcare safety

Loading next page...
 
/lp/ou_press/aggregate-analysis-of-sentinel-events-as-a-strategic-tool-in-safety-Sm2jwUxoYT
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press in association with the International Society for Quality in Health Care. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com
ISSN
1353-4505
eISSN
1464-3677
D.O.I.
10.1093/intqhc/mzy116
Publisher site
See Article on Publisher Site

Abstract

Abstract Objective To examine if clustering of root causes of sentinel events (SEs) can contribute to organisational improvement of healthcare and patient safety by providing insight into organisational risk factors, patterns and trends. Design Retrospective, cross-sectional review of SEs from a hospital database reported to the Board of directors in 2016. Setting A regional teaching hospital in the Netherlands. Intervention(s) Clustering of characteristics and variables of SEs to establish vulnerabilities and patterns of failure factors of the organisation. Main Outcome Measure(s) Characteristics and contributory causes of failure of SEs identified via root cause analysis (RCA). Outcomes reported using descriptive statistics. Results A total of 21 events were included involving 21 patients. Mean age was 56.7 years (SD 24.4), 71.4% were above 50 years of age. In 81.8%, the care was multi-disciplinary and in 76.2% the event resulted in permanent harm or injury. Of the 132 identified contributory root causes, most were related to human factors (53.8%) and organisational factors (40.2%). Technical and patient-related factors were identified in 3.0%. Organisational improvement strategies focused on the care of elderly patients, patients subjected to multi-disciplinary care and on improving knowledge, protocols and coordination of care. Conclusion Clustering variables of SEs and contributory factors of failure through RCA helps to delineate a hospital-specific profile by providing a detailed insight into risk factors, patterns and trends in an organisation and to determine the best strategies for improvement by drawing lessons across events. root cause analysis, adverse events, quality of healthcare, incident reporting and analysis, risk management, safety management Introduction Safety and quality is essential for the delivery of optimal healthcare and remains an important issue. The first step to improve safety is to gain insight into the frequency and seriousness of unsafe situations. A systematic review of the literature conducted in 2018 showed that the incidence of in-hospital adverse events was 9.2% and in 7.4% the event was lethal. Almost half of the reported events were considered preventable [1]. The burden of adverse events on healthcare resources is significant; it is estimated to amount to 1% of the national healthcare budget [2]. Most attention has been directed at hospitals because of the high-risk procedures and hazardous environment [3]. Several reports have highlighted the need for healthcare organisations to monitor and learn from adverse events [4]. The importance of developing effective systems for learning from failure is still growing [5]. An incident reporting system is a tool that has the potential of providing insight into the occurrence of events, making it possible to identify causes and risk factors [6]. Incident reporting and analysis originated in safety-critical industries and in the transport sector [6]. Identifying and learning from events helps in continuously making the healthcare system safer [6, 7]. To protect and improve public healthcare, the safety and quality of the Dutch healthcare system is monitored by the Dutch Health and Youth Care Inspectorate (IGJ). The IGJ oversees and regulates all Dutch healthcare providers and professionals. It is mandatory for all Dutch healthcare organisations to report a sentinel event (SE), i.e. an unintended and unexpected event related to the quality of care which caused death or serious harm to a patient. This is an important tool contributing to the improvement of quality and safety [8]. Appropriate response in case of an SE includes conducting a timely and reliable root cause analysis (RCA), followed by developing improvements, implementing these improvements and finally monitoring the effectiveness of the improvements [9]. RCA has been widely adopted as a method to help clinicians and healthcare organisations to analyse and learn from SEs by systematically investigating events, correcting causes and preventing reoccurrence [10, 11]. As RCA of an SE is based on an in-depth analysis of a single event, it often provides tools to improve care in specific situations. Although RCA is a powerful tool for developing and maintaining awareness of risks, it remains challenging to effectively detect organisational and management factors contributing to SEs within an organisation [12, 13]. Organisations tend to approach each RCA independently, rather than drawing lessons across investigations [12]. It has been hypothesised that clustering of data of various SEs can help in capturing organisational risk factors and defects across events, thus supporting system improvement in organisations [3]. The objective of this study was to examine if clustering of root causes of SEs can contribute to improvement of healthcare and patient safety by identifying vulnerabilities and patterns of failure factors of an organisation. Methods Study design and setting This study was conducted in the Zaans Medical Center (ZMC), a regional teaching hospital in the Netherlands with 293 beds, 1300 employees, 120 medical specialists and 27.000 admissions every year. Employees of the ZMC are expected to report suspected SEs to the Board of directors by means of a reporting system. The Board members determine, after consulting experts, whether the incident should be classified as an SE and should be reported to the inspectorate. In the current study, we retrospectively analysed all suspected SEs reported to the board of directors in 2016. The reason for conducting an extensive analysis focusing on identified root causes was that the number of SEs reported in 2016 was significantly higher than in previous years. According to the policy of the ZMC, this study met the criteria for organisational improvement of patient and healthcare safety and was exempt from ethics review, as all data were anonymized and de-identified. All reported SEs were analysed by means of the Systematic Incident Reconstruction and Evaluation (SIRE) method, a Dutch prototype RCA method. SIRE, adapted from a method developed by the National Center for Patient Safety of the Department of Veterans Affairs, was developed to examine and analyse adverse events [14]. The process focuses primarily on systems and not on individuals [10]. Basic and contributing factors are discovered in a process similar to the diagnosis of a disease, i.e. by answering the following questions: What happened? Why did it happen? How can we prevent it happening again? How will we know if the actions we took made a difference? This method also includes the perspectives of the patient and the patient’s family. RCAs are performed in the ZMC by an experienced multi-disciplinary team consisting of clinicians, nurses, staff members and a coordinator. If necessary, internal or external experts are consulted as well. The team does not include persons who were involved in the incident under investigation. RCAs and reports are considered to be confidential and are treated accordingly, but improvement measures are broadly disseminated within the organisation. The following procedure was conducted in the ZMC. The first step in the RCA process is to identify and reveal underlying dominant risk and failure factors, so-called root causes. Root causes are used to provide a more realistic view of how systems work, as well as to create effective and lasting solutions [15–17]. In the second step, the identified root causes are classified by linking them to one of the five main categories and subcategories of the medical version of the Eindhoven Classification Model (ECM) of system failure, Table 1 [17]. Table 1 Description of the main and subcategories of the Eindhoven Classification Model Main category    Subcategory  Code  Description  Technical    External  T-ex  Technical failures beyond the control and responsibility of the investigating organisation.      Design  TD  Failures due to poor design of equipment, software, labels or forms.      Construction  TC  Correct design, which was not constructed properly or was set up in inaccessible areas.      Materials  TM  Material defects not classified under TD or TC.  Organisational    External  O-ex  Failures at an organisational level beyond the control and responsibility of the investigating organisation, such as in another department of area (address by collaborative systems).      Transfer of knowledge  OK  Failures resulting from inadequate measures taken to ensure that situational or domain-specific knowledge or information is transferred to all new or inexperienced staff.      Protocols  OP  Failures relating to the quality and availability of the protocols within the department (too complicated, inaccurate, unrealistic, absent, or poorly presented).      Management priorities  OM  Internal management decisions in which safety is relegated to an inferior position when faced with conflicting demands or objectives. This is a conflict between production needs and safety. Example: decisions that are made about staffing levels.      Culture  OC  Failures resulting from the collective approach and its attendant modes of behaviour to risks in the investigating organisation.  Human    External  H-ex  Human failures originating beyond the control and responsibility of the investigating organisation. This could apply to individuals in another department.    Knowledge-based behaviour  Knowledge-based behaviour  HKK  The inability of an individual to apply their existing knowledge to a novel situation. Example: a trained blood bank technologist who is unable to solve a complex antibody identification problem.    Rule-based behaviour  Qualifications  HRQ  The incorrect fit between an individual training or education and a particular task. Example: expecting a technician to solve the same type of difficult problems as a technologist.      Coordination  HRC  A lack of task coordination within a healthcare team in an organisation. Example: an essential task not being performed because everyone thought that someone else had completed the task.      Verification  HRV  The correct and complete assessment of a situation including related conditions of the patient and materials to be used before starting the intervention. Example: failure to correctly identify a patient by checking the wristband.      Intervention  HRI  Failures that result from faulty task planning and execution. Example: washing red cells by the same protocol as platelets.      Monitoring  HRM  Monitoring a process or patient status. Example: a trained technologist operating an automated instrument and not realising that a pipette dispenses reagents is clogged.    Skill-based behaviour  Slips  HSS  Failures in the performance of highly developed skills. Example: a technologist adding drops of reagents to a row of test tubes and then missing the tube or a computer entry error.      Tripping  HST  Failures in whole body movements. These errors are often referred to as ‘slipping, tripping, or falling’. Examples: a blood bag slipping out of one’ s hands and breaking or tripping over a loose tile on the floor.  Patient-related    Patient-related factor  PRF  Failures related to patient characteristics or conditions, which are beyond the control of staff and influence treatment.  Main category    Subcategory  Code  Description  Technical    External  T-ex  Technical failures beyond the control and responsibility of the investigating organisation.      Design  TD  Failures due to poor design of equipment, software, labels or forms.      Construction  TC  Correct design, which was not constructed properly or was set up in inaccessible areas.      Materials  TM  Material defects not classified under TD or TC.  Organisational    External  O-ex  Failures at an organisational level beyond the control and responsibility of the investigating organisation, such as in another department of area (address by collaborative systems).      Transfer of knowledge  OK  Failures resulting from inadequate measures taken to ensure that situational or domain-specific knowledge or information is transferred to all new or inexperienced staff.      Protocols  OP  Failures relating to the quality and availability of the protocols within the department (too complicated, inaccurate, unrealistic, absent, or poorly presented).      Management priorities  OM  Internal management decisions in which safety is relegated to an inferior position when faced with conflicting demands or objectives. This is a conflict between production needs and safety. Example: decisions that are made about staffing levels.      Culture  OC  Failures resulting from the collective approach and its attendant modes of behaviour to risks in the investigating organisation.  Human    External  H-ex  Human failures originating beyond the control and responsibility of the investigating organisation. This could apply to individuals in another department.    Knowledge-based behaviour  Knowledge-based behaviour  HKK  The inability of an individual to apply their existing knowledge to a novel situation. Example: a trained blood bank technologist who is unable to solve a complex antibody identification problem.    Rule-based behaviour  Qualifications  HRQ  The incorrect fit between an individual training or education and a particular task. Example: expecting a technician to solve the same type of difficult problems as a technologist.      Coordination  HRC  A lack of task coordination within a healthcare team in an organisation. Example: an essential task not being performed because everyone thought that someone else had completed the task.      Verification  HRV  The correct and complete assessment of a situation including related conditions of the patient and materials to be used before starting the intervention. Example: failure to correctly identify a patient by checking the wristband.      Intervention  HRI  Failures that result from faulty task planning and execution. Example: washing red cells by the same protocol as platelets.      Monitoring  HRM  Monitoring a process or patient status. Example: a trained technologist operating an automated instrument and not realising that a pipette dispenses reagents is clogged.    Skill-based behaviour  Slips  HSS  Failures in the performance of highly developed skills. Example: a technologist adding drops of reagents to a row of test tubes and then missing the tube or a computer entry error.      Tripping  HST  Failures in whole body movements. These errors are often referred to as ‘slipping, tripping, or falling’. Examples: a blood bag slipping out of one’ s hands and breaking or tripping over a loose tile on the floor.  Patient-related    Patient-related factor  PRF  Failures related to patient characteristics or conditions, which are beyond the control of staff and influence treatment.  Table 1 Description of the main and subcategories of the Eindhoven Classification Model Main category    Subcategory  Code  Description  Technical    External  T-ex  Technical failures beyond the control and responsibility of the investigating organisation.      Design  TD  Failures due to poor design of equipment, software, labels or forms.      Construction  TC  Correct design, which was not constructed properly or was set up in inaccessible areas.      Materials  TM  Material defects not classified under TD or TC.  Organisational    External  O-ex  Failures at an organisational level beyond the control and responsibility of the investigating organisation, such as in another department of area (address by collaborative systems).      Transfer of knowledge  OK  Failures resulting from inadequate measures taken to ensure that situational or domain-specific knowledge or information is transferred to all new or inexperienced staff.      Protocols  OP  Failures relating to the quality and availability of the protocols within the department (too complicated, inaccurate, unrealistic, absent, or poorly presented).      Management priorities  OM  Internal management decisions in which safety is relegated to an inferior position when faced with conflicting demands or objectives. This is a conflict between production needs and safety. Example: decisions that are made about staffing levels.      Culture  OC  Failures resulting from the collective approach and its attendant modes of behaviour to risks in the investigating organisation.  Human    External  H-ex  Human failures originating beyond the control and responsibility of the investigating organisation. This could apply to individuals in another department.    Knowledge-based behaviour  Knowledge-based behaviour  HKK  The inability of an individual to apply their existing knowledge to a novel situation. Example: a trained blood bank technologist who is unable to solve a complex antibody identification problem.    Rule-based behaviour  Qualifications  HRQ  The incorrect fit between an individual training or education and a particular task. Example: expecting a technician to solve the same type of difficult problems as a technologist.      Coordination  HRC  A lack of task coordination within a healthcare team in an organisation. Example: an essential task not being performed because everyone thought that someone else had completed the task.      Verification  HRV  The correct and complete assessment of a situation including related conditions of the patient and materials to be used before starting the intervention. Example: failure to correctly identify a patient by checking the wristband.      Intervention  HRI  Failures that result from faulty task planning and execution. Example: washing red cells by the same protocol as platelets.      Monitoring  HRM  Monitoring a process or patient status. Example: a trained technologist operating an automated instrument and not realising that a pipette dispenses reagents is clogged.    Skill-based behaviour  Slips  HSS  Failures in the performance of highly developed skills. Example: a technologist adding drops of reagents to a row of test tubes and then missing the tube or a computer entry error.      Tripping  HST  Failures in whole body movements. These errors are often referred to as ‘slipping, tripping, or falling’. Examples: a blood bag slipping out of one’ s hands and breaking or tripping over a loose tile on the floor.  Patient-related    Patient-related factor  PRF  Failures related to patient characteristics or conditions, which are beyond the control of staff and influence treatment.  Main category    Subcategory  Code  Description  Technical    External  T-ex  Technical failures beyond the control and responsibility of the investigating organisation.      Design  TD  Failures due to poor design of equipment, software, labels or forms.      Construction  TC  Correct design, which was not constructed properly or was set up in inaccessible areas.      Materials  TM  Material defects not classified under TD or TC.  Organisational    External  O-ex  Failures at an organisational level beyond the control and responsibility of the investigating organisation, such as in another department of area (address by collaborative systems).      Transfer of knowledge  OK  Failures resulting from inadequate measures taken to ensure that situational or domain-specific knowledge or information is transferred to all new or inexperienced staff.      Protocols  OP  Failures relating to the quality and availability of the protocols within the department (too complicated, inaccurate, unrealistic, absent, or poorly presented).      Management priorities  OM  Internal management decisions in which safety is relegated to an inferior position when faced with conflicting demands or objectives. This is a conflict between production needs and safety. Example: decisions that are made about staffing levels.      Culture  OC  Failures resulting from the collective approach and its attendant modes of behaviour to risks in the investigating organisation.  Human    External  H-ex  Human failures originating beyond the control and responsibility of the investigating organisation. This could apply to individuals in another department.    Knowledge-based behaviour  Knowledge-based behaviour  HKK  The inability of an individual to apply their existing knowledge to a novel situation. Example: a trained blood bank technologist who is unable to solve a complex antibody identification problem.    Rule-based behaviour  Qualifications  HRQ  The incorrect fit between an individual training or education and a particular task. Example: expecting a technician to solve the same type of difficult problems as a technologist.      Coordination  HRC  A lack of task coordination within a healthcare team in an organisation. Example: an essential task not being performed because everyone thought that someone else had completed the task.      Verification  HRV  The correct and complete assessment of a situation including related conditions of the patient and materials to be used before starting the intervention. Example: failure to correctly identify a patient by checking the wristband.      Intervention  HRI  Failures that result from faulty task planning and execution. Example: washing red cells by the same protocol as platelets.      Monitoring  HRM  Monitoring a process or patient status. Example: a trained technologist operating an automated instrument and not realising that a pipette dispenses reagents is clogged.    Skill-based behaviour  Slips  HSS  Failures in the performance of highly developed skills. Example: a technologist adding drops of reagents to a row of test tubes and then missing the tube or a computer entry error.      Tripping  HST  Failures in whole body movements. These errors are often referred to as ‘slipping, tripping, or falling’. Examples: a blood bag slipping out of one’ s hands and breaking or tripping over a loose tile on the floor.  Patient-related    Patient-related factor  PRF  Failures related to patient characteristics or conditions, which are beyond the control of staff and influence treatment.  The ECM coding system has been tested and validated in medical settings [18]. In this model, latent failures are classified first to increase the likelihood of discovering all causes underlying the event. The main categories of the ECM system are latent factors (technical and organisational), active failures (human), patient-related factors and unclassifiable factors, Table 1 [19, 20]. The RCA results in a report with an action plan that identifies causes and intended strategies for improvement to reduce the risk of similar events occurring in the future. The plan addresses responsibilities for implementation and strategies for measuring the effectiveness of the actions. Data collection The SEs reported in 2016 were obtained from the hospital database. In this database, all reported events are recorded anonymously. Events were included if they were classified as an SE and if the ZMC was primarily involved. For each recruited SE report, the following variables were extracted and analysed: baseline characteristics (patient age, gender), event characteristics (day and time of the incident, number and type of healthcare professionals involved, incident category, degree of harm, number of healthcare organisations involved and incident location), root causes as identified through RCA, and improvement measures that were developed and implemented. Types and locations of the events were categorised according to the Harvard Medical Practice Study classification [21]. Statistical analysis Data of the identified and recruited SEs were first summarised using descriptive statistics and frequency tables. We used the identified RCAs to extrapolate the detected root causes. Statistical analyses were performed using SPSS (version 20). The appropriate analyses were performed, including mean, median, percentages and standard deviation (SD). Results The total number of suspected SEs reported to the Board of Directors in 2016 was 28. After consultation of experts, 23 events were considered to be SEs. The 23 SEs were reported to the IGJ and subsequently evaluated by a multi-disciplinary team. For the current analysis, two of these events were excluded. One event was excluded because according to the Dutch Health Care inspectorate (IGJ) it was not an SE but a complication of a surgical procedure. One other event was excluded because it occurred in another healthcare organisation and was not related to the procedures of or the care provided by the ZMC. As a result, 21 SEs were included in the current analysis. Characteristics of the reported sentinel events The 21 included SEs involved a total of 21 patients. The characteristics of the patients and of the events are shown in Table 2. In 14 cases (66.7%) the patient was female. The mean age was 56.7 years (SD 24.4), and in 71.4% the patient was above 50 years of age. In seven cases (33.3%), at least one other healthcare organisation was involved. Table 2 Baseline and sentinel events characteristics Characteristics  Number (%)  Age (year)     <1  2 (9.5)   20–49  4 (19.0)   50–59  3 (14.3)   60–69  4 (19.0)   70–79  6 (28.6)   >80  2 (9.5)  Gender     Male  7 (33.3)   Female  14 (66.7)  Time     Daytime (7 am–5 pm)  14 (66.7)   Evening and night (5 pm–7 am)  7 (33.3)  Event location     Floor units  9 (42.9)   Operating rooms  6 (28.6)   Emergency department  3 (14.3)   Other  3 (14.3)  Event type     Operations  7 (33.3)   Diagnostic  5 (23.8)   Medications  3 (14.3)   Therapeutic  2 (9.5)   Peripartum  2 (9.5)   Other  2 (9.5)  Professional involved     Medical specialist  58 (43.6)   Nurse  23 (17.3)   Resident/registrars  17 (12.8)   Operation personnel  14 (10.5)   Management  10 (7.5)   Other  11 (8.3)   Total  133 (100)  Number of professionals involved in one event     3–4  8 (38.1)   5–6  6 (28.6)   7–8  5 (23.8)   ≥9  2 (9.5)  Number of medical specialists involved in one event     1  5 (23.8)   2  6 (28.6)   3  4 (19.0)   4  3 (14.3)   5  1 (4.8)   6  2 (9.5)  Consequences     No permanent inconvenience/injuries  2 (9.5)   Permanent injuries/harm  16 (76.2)    Death  6 (28.6)   Unknown  3 (14.3)  Characteristics  Number (%)  Age (year)     <1  2 (9.5)   20–49  4 (19.0)   50–59  3 (14.3)   60–69  4 (19.0)   70–79  6 (28.6)   >80  2 (9.5)  Gender     Male  7 (33.3)   Female  14 (66.7)  Time     Daytime (7 am–5 pm)  14 (66.7)   Evening and night (5 pm–7 am)  7 (33.3)  Event location     Floor units  9 (42.9)   Operating rooms  6 (28.6)   Emergency department  3 (14.3)   Other  3 (14.3)  Event type     Operations  7 (33.3)   Diagnostic  5 (23.8)   Medications  3 (14.3)   Therapeutic  2 (9.5)   Peripartum  2 (9.5)   Other  2 (9.5)  Professional involved     Medical specialist  58 (43.6)   Nurse  23 (17.3)   Resident/registrars  17 (12.8)   Operation personnel  14 (10.5)   Management  10 (7.5)   Other  11 (8.3)   Total  133 (100)  Number of professionals involved in one event     3–4  8 (38.1)   5–6  6 (28.6)   7–8  5 (23.8)   ≥9  2 (9.5)  Number of medical specialists involved in one event     1  5 (23.8)   2  6 (28.6)   3  4 (19.0)   4  3 (14.3)   5  1 (4.8)   6  2 (9.5)  Consequences     No permanent inconvenience/injuries  2 (9.5)   Permanent injuries/harm  16 (76.2)    Death  6 (28.6)   Unknown  3 (14.3)  Table 2 Baseline and sentinel events characteristics Characteristics  Number (%)  Age (year)     <1  2 (9.5)   20–49  4 (19.0)   50–59  3 (14.3)   60–69  4 (19.0)   70–79  6 (28.6)   >80  2 (9.5)  Gender     Male  7 (33.3)   Female  14 (66.7)  Time     Daytime (7 am–5 pm)  14 (66.7)   Evening and night (5 pm–7 am)  7 (33.3)  Event location     Floor units  9 (42.9)   Operating rooms  6 (28.6)   Emergency department  3 (14.3)   Other  3 (14.3)  Event type     Operations  7 (33.3)   Diagnostic  5 (23.8)   Medications  3 (14.3)   Therapeutic  2 (9.5)   Peripartum  2 (9.5)   Other  2 (9.5)  Professional involved     Medical specialist  58 (43.6)   Nurse  23 (17.3)   Resident/registrars  17 (12.8)   Operation personnel  14 (10.5)   Management  10 (7.5)   Other  11 (8.3)   Total  133 (100)  Number of professionals involved in one event     3–4  8 (38.1)   5–6  6 (28.6)   7–8  5 (23.8)   ≥9  2 (9.5)  Number of medical specialists involved in one event     1  5 (23.8)   2  6 (28.6)   3  4 (19.0)   4  3 (14.3)   5  1 (4.8)   6  2 (9.5)  Consequences     No permanent inconvenience/injuries  2 (9.5)   Permanent injuries/harm  16 (76.2)    Death  6 (28.6)   Unknown  3 (14.3)  Characteristics  Number (%)  Age (year)     <1  2 (9.5)   20–49  4 (19.0)   50–59  3 (14.3)   60–69  4 (19.0)   70–79  6 (28.6)   >80  2 (9.5)  Gender     Male  7 (33.3)   Female  14 (66.7)  Time     Daytime (7 am–5 pm)  14 (66.7)   Evening and night (5 pm–7 am)  7 (33.3)  Event location     Floor units  9 (42.9)   Operating rooms  6 (28.6)   Emergency department  3 (14.3)   Other  3 (14.3)  Event type     Operations  7 (33.3)   Diagnostic  5 (23.8)   Medications  3 (14.3)   Therapeutic  2 (9.5)   Peripartum  2 (9.5)   Other  2 (9.5)  Professional involved     Medical specialist  58 (43.6)   Nurse  23 (17.3)   Resident/registrars  17 (12.8)   Operation personnel  14 (10.5)   Management  10 (7.5)   Other  11 (8.3)   Total  133 (100)  Number of professionals involved in one event     3–4  8 (38.1)   5–6  6 (28.6)   7–8  5 (23.8)   ≥9  2 (9.5)  Number of medical specialists involved in one event     1  5 (23.8)   2  6 (28.6)   3  4 (19.0)   4  3 (14.3)   5  1 (4.8)   6  2 (9.5)  Consequences     No permanent inconvenience/injuries  2 (9.5)   Permanent injuries/harm  16 (76.2)    Death  6 (28.6)   Unknown  3 (14.3)  The types and locations of the events are shown in Table 2. Seven events (33.3%) were operation-related, five (23.8%) were related to the diagnostic procedure and three (14.3%) were medication-related. Almost half of the events occurred on the ward (42.8%) and 28.6% occurred in the operating theatre. In 18 of the 21 events (81.8%), at least two medical specialists or departments were involved. In 16 cases (76.2%), the event resulted in permanent harm or injury and in 6 cases (28.6%), the patient died. Involvement of healthcare professionals A total of 133 healthcare professionals were involved in the 21 events. The number of professionals involved in an event ranged from 3 to 11, with a mean of 6.3 (SD 2.5). Medical specialists were predominantly related to the SE in 43.6% of the cases, followed by nurses in 17.3% and resident physicians in 12.8%. The number of medical specialists related to the treatment at the time of the event ranged from 1 to 5, with a mean of 3 (Table 2). Causes of sentinel events Analysis of the 21 SEs resulted in the identification of 132 contributory root causes, ranging between 3 and 15 in a single event, with a mean of 6.3 (SD 3.8). When the root causes were classified into the five main categories according to the ECM taxonomy, 71 (53.8%) were human, 53 (40.2%) organisational, 4 (3.0%) technical and 4 (3.0%) patient-related. None of the root causes was unclassifiable. Technical, human and organisational root causes were subdivided into one or more specific subcategories (Table 1). Human factors were the most frequently identified root causes (n = 71). More than half of the human factors (53.5%) were related to knowledge-based behaviour (HKK), followed by verification (HRV) in 19.7% (Fig. 1). Of the organisational factors (n = 53), 66.0% were related to protocols and procedures (OP) followed by management priorities (OM) in 13.2%. In the category technical factors (n = 4), three were related to design (TD) and one to material defects (TM). External and patient-related factors were hardly identified, and hence an in-depth analysis was not possible. Figure 1 View largeDownload slide The distribution of the 132 identified root causes in the 21 SE analysed. Root causes were divided into the following categories: Human, Organisational, Technical and Patient-related. Data are presented as numbers (n) or percentage (%). Abbreviations: HKK, knowledge-based behaviour; HRV, verification; HRC, coordination; H-ex, external; OC, external; OP, protocols and procedures; OM, management priorities; OK, transfer of knowledge; TD, design; TM, material defects. Figure 1 View largeDownload slide The distribution of the 132 identified root causes in the 21 SE analysed. Root causes were divided into the following categories: Human, Organisational, Technical and Patient-related. Data are presented as numbers (n) or percentage (%). Abbreviations: HKK, knowledge-based behaviour; HRV, verification; HRC, coordination; H-ex, external; OC, external; OP, protocols and procedures; OM, management priorities; OK, transfer of knowledge; TD, design; TM, material defects. Improvement strategies Prevention strategies in case of a single event tend to focus on that event, aiming to improve local work condition and processes, improvement strategies on the organisational level were rarely proposed. Based on the identified root causes in our analysis, various interventions were carried out. A hospital-based protocol was designed in collaboration with medical specialists, nurses and management, describing the important aspects of patient care, including the processes of admission, diagnosis, treatment and discharge. Furthermore, the protocol draws attention to the special requirements necessary for treating elderly and complex patients requiring multi-disciplinary care. The protocol describing the different roles of a medical specialist was updated, describing and clarifying the cooperation between departments and healthcare workers and the coordination of multi-disciplinary care and formalising the role of the medical specialist who must guide and coordinate the care provided for the patient. To tackle human and organisational factors, which occurred most frequently, an organisation-wide programme has been constructed to improve the knowledge, quality and availability of protocols and procedures and to enhance safety behaviour by means of training programmes. Regular audits were planned to examine the safety behaviour of departments. Discussion General findings and interpretation Healthcare should be available to people when necessary and should be effective and safe. Incident reporting and analysis is widely recognised as an important tool for improving reliability and patient safety in the care delivery process [23]. An increase in safety can only be achieved if interventions tackle the right underlying causes. This study reports on accumulated data of SEs and examined if clustering of root causes of SEs, i.e. events with significant harm to patients, helps in identifying the weaknesses of an organisation. In 21 included SEs, 132 root causes were identified, illustrating the multi-factorial nature of events. Events predominantly occurred on wards and the operation theatre and involved elderly patients and patients receiving multi-disciplinary care. The events had permanent consequences in 76.2% of the cases. Root causes were predominantly related to human factors (knowledge-based behaviour, verification and coordination) and organisational factors (protocols and procedures, management priorities and transfer of knowledge). On the basis of our analysis, several improvement strategies were carried out on both local and organisational level. A hospital-based protocol was designed regarding the care of patients, especially the elderly and patients receiving multi-disciplinary care. An organisational programme was developed to improve knowledge and safety behaviour by training healthcare professionals. The coordinating role of medical specialists was formalised to improve the provided multi-disciplinary care. The quality and availability of protocols and process descriptions were evaluated and improved. Proposed prevention strategies tackle both local, organisational and management factors. However, a clear framework is lacking to identify which types of incidents can be prevented with which type of interventions [12]. Clustering of RCAs seems valuable for identifying risk factors, undesirable patterns and trends as it focusses on systems and processes on an organisational level; a hospital-specific profile can be delineated. As hospitals are complex organisations, practically every step in the precarious process can have a negative effect on patient safety and healthcare safety. A safety management approach is crucial to identify factors and preconditions affecting the organisation [24, 25]. Prevention strategies of a single event tend to focus on the surroundings and circumstances of that event, aiming to improve local work procedures and processes, making it difficult to detect organisational vulnerabilities and patterns of failure and subsequently recommend improvement on that level [22]. Improvements strategies on the organisational level were rarely proposed. Aggregate analysis of SEs is a systematic process; priorities for improvement are based on multiple cases, rather than just one case. The importance of this approach is that it recognises that organisations have inherent weaknesses and that failures will manifest in local working conditions; focusing on individual responsibility is likely to be ineffective [11]. A systemic, rather than the individual view is likely to support continuous improvement and promote future reporting of failures in healthcare [13]. Comparison with previous studies After extensively reviewing the literature, we did not encounter other studies reporting on data of SEs of a single organisation. Nevertheless, the identified root causes in our study, predominantly human and organisational, are in accordance with previous studies [26–28]. The identified factors in the study of Smits et al. [26] were predominantly labelled as organisational and human, considered preventable in 66%–93% of the cases and related to permanent disability in 13–20%. Failures related to the quality and availability of protocols were frequently related to permanent disability. The root causes of failure of 522 unintended events at ten emergency departments in the Netherlands were predominantly human and organisational [25]. In 25%, the event was related to the cooperation between departments and healthcare workers within an organisation. Nuckols et al. [28] reported the characteristics of 2246 identified incidents in the USA. The events occurred primarily in floor units (50%) and were related to medication, operations and procedures. Similar to the improvement strategies in our study, most reported strategies were quality assurance/peer review, training, improvement of organisational procedures, improvement of sources of information and evaluation [1, 22, 27]. Prevention strategies at the organisational level were least often considered. Although, it seems worthwhile to direct interventions on organisational causes because organisational causes are nearly always believed to be preventable [26]. Strengths and limitations Our study has various advantages. We used an existing reporting system for identifying events, and 96% of the identified SEs were included. RCA was performed directly with the assessment of the SEs; the data were therefore original and accurate, not based on summaries of compiled information. Root causes were classified via ECM, which is a well-established and evidence-based framework. A limitation of the current study is that we were unaware of the intentions and attitudes of professionals regarding reporting of events and were not able to correct for possible underreporting. All reporting systems rely on professionals submitting events occurring within routine work situations, but many events will go unreported, with doctors being less likely than nurses to report events [13, 29]. Besides, although an SE is clearly defined, there is still uncertainty whether events should be classified as SEs. Unintended events that occur more frequently are believed to share the same underlying failure factors [30, 31]. As we were not able to accurately identify unintended incidents, we did not include them in our analysis. In the current analysis, all identified root causes were considered equal, while in clinical practice this is not the case. Contributory factors can vary according to the level of proximity to the event [11]. Caution must be exercised, as only 21 SEs were included. To produce reliable results, it has been suggested to include at least 50 events, preferably across organisations [27]. The advantage of the current method is that a hospital-specific profile can be delineated to determine where best to focus strategies for improvement. The reliability and generalisability of this approach can be improved by applying this method in several organisations on a regular basis. Implications for practice and future research Clustering data of events helps to target interventions and to increase organisational learning; it is a process of creating and applying valid knowledge to improve patient safety of an organisation. Aggregation of RCA can help to build organisational and technical defences. Clear information and action feedback are fundamental in supporting continuous improvement in the healthcare processes [25]. Information concerning root causes of failures and corrective actions should be widely disseminated across organisations at the regional and the national level to increase awareness, prevent recurrence and improve safety. Effective safety feedback systems represent an ongoing cyclical process of reporting, evaluation and corrective actions [25]. Formulating corrective actions is more difficult than finding failure causes [23]. How organisations should implement corrective actions may vary in practice and is an important topic for further research. We want to encourage others to apply this framework to help improve the healthcare system. The advantage is that organisations can learn from each other, especially from implemented improvements. To be able to achieve this, it is imperative to use the same taxonomy of failure factors in order to enable regional, national and international comparisons. The future lies in targeted incident reporting, robust analysis with the engagement of healthcare professionals and evaluation of the effectiveness of improvements. Incident reporting must be coupled with visible, sustainable improvements and, most importantly, dissemination of the acquired knowledge. Conclusion Incident reporting and analysis is crucial for improving healthcare safety. RCA has been adopted as a method to analyse and identify root causes of failure. The current study describes a strategic tool in safety management: by aggregating data of SEs, a framework of contributing root causes of failure is produced. A hospital-specific profile can be delineated to provide detailed insight into organisational patterns and risk factors, but most importantly to determine where best to focus organisational and management changes to improve patient safety and healthcare safety in the entire organisation, by drawing lessons across events. Acknowledgements The authors want to thank all those involved in the reported events. Furthermore, we thank the members of the multi-disciplinary team that conducted the analysis of the SEs of the ZMC. Special thanks to the patients and families for their contributions. Finally, we thank the hospital and Board of directors for making this report and its publication possible. Funding None. Data sharing We would be happy to share the available data of this study. Ethical approval According to the policy of the ZMC, this work was exempt from ethics review, as all data were anonymised and de-identified. References 1 de Vries EN, Ramrattan MA, Smorenburg SM et al.  . The incidence and nature of in-hospital adverse events: a systematic review. Qual Saf Health Care  2008; 17: 216– 23. Google Scholar CrossRef Search ADS PubMed  2 Hoonhout LH, de Bruijne MC, Wagner C et al.  . Direct medical costs of adverse events in Dutch hospitals. BMC Health Serv Res  2009; 9: 27. Google Scholar CrossRef Search ADS PubMed  3 Harmsen M, Gaal S, van Dulmen S et al.  . Patient safety in Dutch primary care: a study protocol. Implement Sci  2010; 5: 50– 8. Google Scholar CrossRef Search ADS PubMed  4 Hutchinson A, Young TA, Cooper KL et al.  . Trends in healthcare incident reporting and relationship to safety and quality data in acute hospitals: results from the National Reporting and Learning System. Qual Saf Health Care  2009; 18: 5– 10. Google Scholar CrossRef Search ADS PubMed  5 Williams SK, Osborn SS. The development of the National Reporting and Learning System in England and Wales, 2001–2005. Med J Aust  2006; 184: 65– 68S. 6 Barach P, Small SD. Reporting and preventing medical mishaps: lessons from non-medical near miss reporting systems. Br Med J  2000; 320: 759– 63. Google Scholar CrossRef Search ADS   7 Macrae C. The problem with incident reporting. BMJ Qual Saf  2016; 25: 71– 5. Google Scholar CrossRef Search ADS PubMed  8 Leistikow I, Mulder S, Vesseur J et al.  . Learning from incidents in healthcare: the journey, not the arrival, matters. BMJ Qual Saf  2017; 26: 252– 6. Google Scholar CrossRef Search ADS PubMed  9 Battles JB, Kaplan HS, Van der Schaaf TW et al.  . The attributes of medical event reporting systems: experience with a prototype medical event-reporting system for transfusion medicine. Arch Pathol Lab Med  1998; 122: 231– 8. Google Scholar PubMed  10 Iedema RA, Jorm C, Long D et al.  . Turning the medical gaze in upon itself: root cause analysis and the investigation of clinical error. Soc Sci Med  2006; 62: 1605– 15. Google Scholar CrossRef Search ADS PubMed  11 Leape LL. Why should we report adverse incidents? J Eval Clin Pract  1999; 5: 1– 4. Google Scholar CrossRef Search ADS PubMed  12 Wu AW, Lipshutz AK, Pronovost PJ. Effectiveness and efficiency of root cause analysis in medicine. JAMA  2008; 299: 685– 7. Google Scholar CrossRef Search ADS PubMed  13 Benn J, Koutantji M, Wallace L et al.  . Feedback from incident reporting: information and action to improve patient safety. Qual Saf Health Care  2009; 18: 11– 21. Google Scholar CrossRef Search ADS PubMed  14 Root cause analysis tools . Ann Arbor, MI: Department of Veterans Affairs, National Center for Patient Safety, 2007. www.va.gov/ncps/CogAids/RCA/index.html. 15 Mers TM. Medical event reporting system for transfusion medicine reference manual version 3.0 . Columbia University, New York, 2001. http://www.mers-tm.net. 16 van Vuuren W, Shea CE, van der Schaaf TW. The development of an incident analysis tool for the medical field . Eindhoven: Eindhoven University of Technology, 1997. http://alexandria.tue.nl/repository/books/493452.pdf. 17 Schaaf van der TW, Habraken MMP. PRISMA-Medical a brief description . Eindhoven: Eindhoven University of Technology, Faculty of Technology Management, Patient Safety Systems, 2005. 18 Smits M, Janssen JCJA, de Vet R et al.  . Analysis of unintended events in hospitals: inter-rater reliability of constructing causal trees and classifying root causes. Int J Qual Health Care  2009; 21: 292– 300. Google Scholar CrossRef Search ADS PubMed  19 Kaplan HS, Battles JB, Schaaf Van der TW et al.  . Identification and classification of the causes of events in transfusion medicine. Transfusion  1998; 38: 1071– 81. Google Scholar CrossRef Search ADS PubMed  20 Rasmussen J. Outlines of a hybrid model of the process operator. In: Sheridan TB, Johanssen G (eds). Monitoring behaviour and supervisory control . London: Plenium Press, 1976, 371– 384. Google Scholar CrossRef Search ADS   21 Harvard Medical Practice Study. Patients, doctors and lawyers: medical injury, malpractice litigation and patient compensation in New York. The Report of the Harvard Medical Practice Study to the State of New York . New York, NY: Department of Health, 1990. 22 Zegers M, de Bruijne MC, de Keizer B et al.  . The incidence, root-causes, and outcomes of adverse events in surgical units: implication for potential prevention strategies. Patient Saf Surg  2011; 5: 13. Google Scholar CrossRef Search ADS PubMed  23 Jansma JD, Zwart DL, Leistikow IP et al.  . Do specialty registrars change their attitudes, intentions and behaviour towards reporting incidents following a patient safety course? BMC Health Serv Res  2010; 10: 100– 9. Google Scholar CrossRef Search ADS PubMed  24 Lipshitz R, Popper M, Friedman VJ. A multifacet model of organizational learning. J Appl Behav Sci  2002; 38: 78– 98. Google Scholar CrossRef Search ADS   25 Anderson JE, Kodate N. Learning from patient safety incidents in incident review meetings: organisational factors and indicators of analytic process effectiveness. Safety Sci  2015; 80: 105– 14. Google Scholar CrossRef Search ADS   26 Smits M, Zegers M, Groenewegen PP et al.  . Exploring the causes of adverse events in hospitals and potential prevention strategies. Qual Saf Health Care  2010; 19: e5. Google Scholar PubMed  27 Smits M, Groenewegen PP, Timmermans DR et al.  . The nature and causes of unintended events reported at ten emergency departments. BMC Emerg Med  2009; 9: 16. Google Scholar CrossRef Search ADS PubMed  28 Nuckols TK, Bell DS, Liu H et al.  . Rates and types of events reported to established incident reporting systems in two US hospitals. Qual Saf Health Care  2007; 16: 164– 8. Google Scholar CrossRef Search ADS PubMed  29 Evans SM, Berry JG, Smith BJ et al.  . Attitudes and barriers to incident reporting: a collaborative hospital study. Qual Saf Health Care  2006; 15: 39– 43. Google Scholar CrossRef Search ADS PubMed  30 Kanse L, Schaaf van der TW, Vrijland ND et al.  . Error recovery in a hospital pharmacy. Ergonomics  2006; 49: 503– 16. Google Scholar CrossRef Search ADS PubMed  31 Wright L, van der Schaaf TW. Accident versus near miss causation: a critical review of the literature, an empirical test in the UK railway domain, and their implications for other sectors. J Hazard Mater  2004; 111: 105– 10. Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018. Published by Oxford University Press in association with the International Society for Quality in Health Care. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

International Journal for Quality in Health CareOxford University Press

Published: May 19, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off