Background
Medical errors have always constituted an important problem in medical services. The frequent occurrence of medical errors can seriously threaten the lives and safety of patients and result in substantial economic losses [
1,
2]. Research has shown that more than 170,000 people die of medical errors every year, representing the third most common cause of death in the United States [
3]. Regarding the use of pacemakers and implantable cardioverters, FDA device malfunction reports revealed that 61 deaths annually were attributed to device malfunction from 1990 to 2002 [
4]. The Implant Files report showed that 2 million injuries and 80,000 deaths were associated with medical devices from 2008 to 2017 [
5]. As medical devices are widely used in the diagnosis, cure and treatment of disease, adverse medical device events and medical device safety surveillance have attracted widespread attention [
6,
7]. A survey revealed that the unexpected failure of medical devices could cause adverse events in radiology departments [
8]. The risk of failures of medical devices must be analyzed and assessed, and measures regarding the quality control of devices are essential for ensuring and improving patient safety [
9].
With the advantages of cost effectiveness and real-time imaging [
10], ultrasound methods, such as transesophageal echocardiography [
11], the detection of pulmonary congestion [
12], guidance for live biopsy [
13], the cannulation of the subclavian vein [
14] and the establishment of vascular access in critically ill patients [
15], have been widely used in hospitals. However, the risk associated with these methods increases substantially if device failure occurs. Many efforts have been made to analyze the failure of ultrasound devices and adopt efficient safeguard measures. The failure modes and rates of various failures were studied, and routine quality control plans were recommended based on one or several radiological assessments [
16,
17]. The causes of transducer failures were further investigated to improve the detection of failures in advance [
18‐
20]. However, there is a lack of quantitative risk assessments of different types of failure modes for medical ultrasound devices. In recent years, failure mode, effect and criticality analysis (FMECA) has been used as a risk assessment approach for medical devices [
21,
22]. Conventional FMECA is used to directly score the failure mode [
1,
23], and it is not very accurate for risk assessment, as identical failure mode may be induced by different causes. Moreover, due to a lack of engineering experience, physicians cannot determine the causes of failure, which may delay the effective resolution of failures, which could have further serious effects.
To identify the risks of various failure modes and determine the failure causes, we utilized an FMECA approach for the quantitative risk evaluation of different failure modes of ultrasound devices. The 4216 failure records for 2096 sets of ultrasound devices at various hospitals were collected and analyzed. First, the failure modes and causes were marked and classified for every record. Cluster analysis was performed separately for various failure modes and causes, and visual correlations for different categories of failure modes and causes were established. Then, by applying the analysis model, the risk priority number (RPN) for each failure mode was back-calculated based on the severity (S), occurrence (O) and detectability (D) of failure causes. Given that the final effect of failure is directly related to the cause, this quantitative risk assessment method is more accurate than traditional methods. Preventive strategies are discussed for high-risk failure modes.
Discussion
The higher RPN is, the greater the risk. The RPN can be used to assess the necessity of measures taken and which measure should be prioritized to reduce loss after a failure and improve the reliability of the system, thereby reducing the associated cost. On the basis of the results, control panel failure is the most frequent failure type. Unclear images are associated with the maximum severity and are the most difficult to detect. Overall, unclear images pose the greatest RPN, which means that such issues can potentially have the greatest impact on ultrasonic diagnostic procedures. Thus, for a failure mode with a high RPN, appropriate quality control measures should be implemented to reduce the effects of failures.
Unclear images, dark shadows on images and lack of image information, which all occur in the display unit, are mainly caused by probe crystal failure, the attenuation of probe performance and probe circuit failure. In addition, electromagnetic interference, improper parameter setting, software system failure, display hardware failure and other issues also cause abnormal displays. In view of the above problems, some measures should be regularly taken to reduce the risk of display unit failures including training users more frequently in how to hold the probe, keep the probe clean and wipe the coupling agent with soft paper after use and using dedicated power supplies for ultrasound devices to reduce the interference caused by other electrical devices.
Power failure, system crashes and automatic restarts are the three most serious failure modes in a host system. These issues are mainly caused by software system failure, board failure, hard disk failure and poor contact faults. The preventive strategies for such failures are as follows: reduce the use of portable storage devices to avoid software viruses, back up the system and reinstall the software whenever necessary, perform regular device maintenance to determine whether there are bad sectors in the hard disk, keep the device clean and reduce poor contacts.
Control key malfunction is the most frequent failure mode due to the large number of panel knobs, frequent use of devices, improper operation, accumulation of dust and other reasons. Unrecognized probe issues are mainly caused by probe circuit faults, poor contact faults and interface failures. Therefore, it is necessary to perform maintenance and keep the device clean in clinical practice.
The major highlights of this study included a quantitative assessment of the risks of different failure modes based on a large number of medical ultrasound devices at hospitals in multiple districts. Additionally, a FMECA method was adopted to determine the level of risk for different failure modes. Because one failure type may be caused by various factors, a visual failure mode tracing diagram was created, from which the classification of failures and the correlations between the failure modes and failure causes were identified. The RPN of each failure mode was obtained based on the RPN of the cause and the corresponding relationship between the phenomenon and cause.
A quantitative FMEA approach was utilized for linear accelerator quality assurance [
22]. The SOD was assessed for potential risks. Few previous studies have investigated the risk associated with different types of failure modes of medical ultrasound devices. To analyze the failure mode of ultrasound device, researcher once categorized the failures of ultrasound devices were into 5 types based on 108 cases [
16]. To the best of our knowledge, no previous research has analyzed such a large amount of ultrasound device failure data or the correlations between failure modes and causes. Additionally, quantitative assessments of the potential impacts of failure modes in various categories have never been conducted. Given that the final effect of failure is directly related to the cause, this FMECA risk assessment approach is more accurate than traditional methods.
Through quantitative scoring of each failure mode, it was found that the three failure modes with the highest RPN value were unclear image, unable to power on and dark shadow on the image. Ultrasound equipment users could make corresponding plans according to the risk score of failure mode to avoid adverse events. In addition, according to the corresponding diagram of failure mode and cause obtained from the analysis, it was helpful for the physicians and clinical engineers to find the cause of the failure, which can reduce the impact on the clinical diagnostic process, the potential risks to patient safety and economic loss. In addition, the frequency of failures with high-risk values can be used as an index to evaluate the reliability of ultrasound devices, thus providing direction regarding technical enhancements to manufacturers and improving device selection and purchasing methods.
Our study had some limitations. It is worth noting that due to the diversity of ultrasound device maintenance service providers, it was impossible to guarantee coverage for all failure records for such a large number of ultrasound devices. We collected as many samples as possible in a variety of areas. Second, although we performed a quantitative evaluation of the risks of different failure modes, potential safety hazards for patients and physicians should be further discussed.
Conclusions
This study was conducted to quantitatively evaluate the risk of different failure modes of ultrasound devices based on a FMECA approach. Through a retrospective analysis of a large sample of maintenance records, the corresponding relationships between the failure modes and failure causes were established. By applying the analysis calculation model, the RPN for each failure mode was finally obtained. The failure mode of image unclear had the maximum risk priority number, with the high frequency of occurrence, severity and detectability value. According to the risk level of different failure modes, the device administrator can make corresponding preventive plans. And the correspondence analysis of failure modes and causes can be conducive to determining the cause of failure, thus reducing the occurrence of adverse events and economic losses.
Future research should evaluate the threshold of the RPN of the failure mode of ultrasound device. When this threshold is exceeded, the device needs to be stopped immediately for repair or replacement, or further measures should be taken to protect the patient’s safety. Moreover, the failure risk rating of ultrasound device applied in different fields should not be the same. For example, the severity of transesophageal cardiac probe failure is much higher than that of conventional cardiac probe during the disease diagnosis. The safety risk caused by ultrasound failure in surgical intervention is much greater that in routine physical examination. Therefore, research on the quantitative assessment of ultrasonic failure risk in specific application fields will be more targeted to reduce the safety risk.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.