In the HUS and the HKI, a total of 195,593 HaiPro-reports were filed in 2012–2017. In HUS, the total number of these reports was 93,176, of which 1.614 (about 2%) were related to medical imaging and only 359 (about 0.4%) were related to dentistry and in HKI the total amount of the reports was 102,417, of which only 61 (about 0.06%) concerned dentistry (Fig. 4). In the HKI-data, medical imaging-related reports were not available separately categorized comparable to the HUS-data. Altogether, 420 dentistry-related PSI and 2464 PSIs that matched our keyword search, from both HUS and HKI, were evaluated individually not to miss any DMFR-related reports. We found 43 reports of DMFR-related PSIs, of which 35 were from the HUS-data and 8 from the HKI-data (Fig. 5).
Reported PSI during 2012–2017 (HUS, HKI)
The number of DMFR-related reports in HUS was almost fourfold compared to HKI during the evaluated period
In our study, the amount of reported PSIs compared to the number of dental radiographs taken in their corresponding organizations stand for approximately 0.02% in HUS (35/201,113) and 0.001% in HKI (8/729,981).
Event type and contributing factors
The most prevalent event type of all 43 DMFR-related PSIs was laboratory-, medical imaging-, or other patient examination-related events (14 cases, 33%). The second most common event type of PSI was related to flow or control of information (11 cases, 26%). Other event types included accidents (4 cases), medical devices or their usage (3 cases), medication or fluid therapy (1 case), other treatment or patient monitoring (1 case), and invasive operation (1 case). The event type was defined as unknown on one occasion and left undefined for eight cases. Out of these 14 cases, three reports revealed a mix-up in patients: once a referral to a panoramic image was prescribed to the wrong patient and the error only became clear after the image was obtained. Once a 3D-model for orthognathic surgical planning was delivered under the name of another patient, but this was noticed by the clinician before the event led to any harm to the patient. The third case included a patient receiving another patient’s information regarding the upcoming dates for a scheduled control visit and radiographic examination, but this error was also noticed shortly after it happened, and no harm occurred.
The most common contributing factors linked to the aforementioned event type were perceived as shortcomings related to communication and flow of information (4/14, 29%). Once the contributing factor was reported to be related to medical devices and equipment and once to work procedure, but in more than half of the cases (8/14, 57%), the contributing factors remained unclear as they were either classified as unknown (3 cases) or left completely empty (5 cases).
The second most common event type, total of 11 cases, was defined to be related to flow or control of information and in almost half of these cases, the contributing factor was related to limitations of communication and flow of information (5/11, 45%). Other contributing factors that were identified were related to work procedure (2 cases); education, instruction about the workplace, and competence (2 cases); and to work environment (incl. utensils) and resources (2 cases).
Even though on eight occasions, the event type was not reported, only one of these cases was also left empty regarding information about possible contributing factors. Of the seven remaining cases without a defined event type, the commonest contributing factor was related to medical devices and equipment (3/8, 38%). Other perceived contributing factors were communication and flow of information (2 cases), patient/customer or relatives (1 case), and work environment (incl. utensils) and resources (1 case). Event types revealed contributing factors that were related to communication and flow of information (1 case), patient/customer or relatives (2 cases), work procedure (1 case), medical devices and equipment (1 case), no recognizable contributing factors, normal situation (2 cases), unknown (1 case), or undefined (2 cases).
Risk classification
Evaluation of the risk classification showed only one report where the risk was assessed to be moderate (risk level III); in all other reported AE, the risk was considered irrelevant (risk level I) or minor (risk level II). On 8/43 occasions, the risk level remained unassessed. This is consistent with the findings stating that mostly no harm or only slight harm had been caused to the patient regarding the outcome of the adverse event. Moderate harm to the patient was reported only once, but the risk level of the event was assessed to be minor (II) even in this case. This report was made by a nurse after a prolonged surgical tooth extraction, pointing out the need for the clinician to better plan the operation, for example using CBCT imaging.
The outcome for the patient and the organization
When categorized by the contributing factors, the most frequently reported events (12/43) were due to communication problems or varying information gaps that delayed patient care and increased the workload of healthcare professionals. Despite the delay, the AE was evaluated to not having caused any harm to the patient for 75% of the cases (9/12). Out of the remaining 25% occasions, the outcome to the patient was reported as “unknown” once, and on the remaining two occasions, this information was not provided at all. These communication-related AEs were common in the HKI-data as half of all incidents in the HKI-data revealed shortcomings in communication (4/8). Out of all the communication-based problems or information gaps, 25% were related to the failure of identifying the patient correctly and, thus, the radiological examinations were stored under the wrong patient’s data (3/12), but these errors were identified before any harm occurred. While most of the communication problems or information gaps (7/12) were caused by confusion due to insufficient or unclear oral or written communication, almost half of the cases (5/12) revealed that there was a shortcoming in using, finding, or accessing the relevant information by the healthcare workers that would have resulted in avoiding the reported PSI.
The contributing factors for the most reported adverse events included device malfunction (5/43) and organization policies (5/43). Malfunction-related reports were only found in the HUS-data. The malfunction of the device led on all occasions to image loss during radiological examination and thus to the need to renew the examination, which ultimately increased patients’ radiation dose. Most of these device-related reports (80%) were related to a CBCT device. In addition to these five malfunction-related AE, we found two additional CBCT-related PSI in the HUS-data. In one case, the radiographer forgot to change the CBCT imaging protocol settings prior to the examination, leading to the need to renew the radiological examination and the other case was a software-related PSI, which was reported immediately after a scheduled CBCT software update. The update brought about a change in the proportions of the acquired images leading to abnormal measurement results in the acquired data. This problem was quickly noticed on site by the staff of the radiology department and corrected by the device manufacturer.
During ICT-related problems (4/43), radiological images of patients were unobtainable, reportedly delaying patient care. These problems showed no recognizable common variable. Patient-related incidents (4/43) were as common as ICT-related PSI in our data. These were reported due to a highly active child almost running against the panoramic device, while it was being positioned for the examination and due to nausea, dizziness, or fainting of adult patients.
Classified under different event types, we found a total of 8/43 cases where either the radiographs were unattainable by the clinician for various reasons (6/8) or that the report of the patient’s radiograph was unavailable (2/8) at the time of the patient’s appointment with the clinician.
Assessing the outcome for the organization is also part of the HaiPro-report and, in our data, the most common organizational outcome was an increase in the healthcare professionals’ workload (14/43), followed by a decline in the organization’s reputation (9/43). On eight occasions, there was no harm to the organization regarding the AE, and on nine occasions, no assessment of this information was performed. An increase in expenses (2/43) and treatment duration (2/43) was also reported as an adverse outcome for the organization.
