Offshore telementored ultrasound: a quality assessment study

Historically, practising medicine has always been a combination of skills for diagnosis and commencing treatment based on patients’ signs and symptoms [1, 2]. As one would expect, doing so in an austere environment is even more challenging where the clinical scenarios and patterns of illness and injury vary widely, and the access to medical equipment and qualified healthcare personnel are limited or even lacking [3]. Norway is an important supplier of oil and gas to the global market and employs approximately 21,000 people running between 80 and 90 oil installations [4]. These platforms and ships are spread along the Norwegian continental shelf and have limited medical and logistic support. Operating in such remote locations and challenging climatic environments requires a well-functioning health service [5]. Today, this remote medical practice is run by offshore nurses in hospital units onboard the installations and search-and-rescue (SAR) personnel working as part of medical evacuation (medevac) teams onboard helicopters [6]. The nurses, often working alone, are trained to carry out focused clinical examinations, perform certain medical tests, including recording vitals and electrocardiograms (ECGs), and provide simple blood and urine tests (e.g., haemoglobin, CRP, glucose). The nurse also has the opportunity to consult with a physician onshore, either by phone or videoconference [7]. However, this remote offshore healthcare service is in many cases insufficient for determining a final diagnosis [8, 9]. Whereas patients hospitalized onshore are referred for further medical imaging, such as X-ray, computerized tomography (CT), magnetic resonance imaging (MRI) or ultrasound (US), to determine their diagnosis [1, 10], this equipment has not, to date, been available for offshore workers. The result is an extended use of medevacs with SAR helicopters to bring the patients off the platforms and admit them to onshore hospitals. This process is a costly affair and not without risks, especially in challenging weather conditions [6]. It can take several hours from alerting the SAR team until the patient reaches definitive care, or in the worst-case scenario, no evacuation is possible due to restricted flying conditions [7]. Therefore, there is a need for extended medical practice with provision of more advanced diagnostic and management advice via telecommunication (i.e., telemedicine [11, 12]). A solution could be to connect US machines to the already installed videoconference units (medical units) found onboard most oil installations [13]. The development of lightweight, battery-powered, and easily transportable devices has made US in the field (i.e., outside hospitals) possible in contrast to bulkier and power-demanding X-ray machines and CT and MRI scanners [14]. Continued improvement in both US and telecommunication technology may open opportunities for improved clinical decision-making [12, 15]. US imaging can be obtained instantly and correlated to the patient’s presenting signs and symptoms and repeated if the condition changes [16]. Combining the offshore nurses’ physical examination with a focused goal-directed US scan [i.e., point-of-care ultrasound (PoCUS)] can confirm and refute life-threatening diagnoses, thereby assisting the onshore physicians in the initial evaluation and management of critically ill and injured patients [16, 17]. However, the acquisition and interpretation of PoCUS examinations are highly user-dependent [17, 18]. The operator experience amongst offshore nurses and SAR personnel thus appears to be a limitation to widespread use. We believe that this limitation can be compensated through real-time assistance using videoconferencing to link novice operators to geographically separate US experts to enable remote guidance of PoCUS examinations, an activity recognized as telementored US, tele-US or remote telementored US (RTMUS) [17, 19]. This utility of US has been evident in various extreme expeditions, such as to Mount Everest [20]. In the last decade, experience with remote applications for PoCUS has continued to increase, and research has been carried out in a variety of locations, such as rural areas in third world countries and even in space (International Space Station) [21‐24]. However, telementored US between offshore nurses and an onshore on-call physician service has never been tested and evaluated. Thus, our study describes the initial experiences from a telementored US trial sending real-time video and US images from an offshore installation.

The objectives of this research are as follows:

All offshore workers that volunteered to be scanned were male with a mean age of 48.7 ± 10.3. The mean weight was 86.1 kg ± 13.4 kg. The image quality score of the cineloops was higher with the FATE protocol than the FAST protocol with medians of 4 and 3, respectively. When comparing the percentage of images suitable for interpretation, 96.4% of the cardiac views (i.e., FATE) using the phased-array transducer and 79.1% of the thoraco-abdominal views (i.e., FAST) using the curved-array transducer had an image score equal to or above 3 (%^≥3). No image scores were calculated for the lungs (i.e., e-FAST) as bilateral lung sliding at the pleural line was easily identified in all participants with 100%^≥3. The median quality scores and fractions of interpretable images for both protocols are outlined as box plots in Fig. 4a, b. Further details about the images’ scores and %^≥3 by acoustic window when performing telementored FATE and FAST are available in Tables 1 and 2.

The mean scan time to complete a full telementored FATE scan comprised six cardio-thoracic views with stored cineloops was 4 min 15 s. Similarly, the mean scan times for FAST comprising four thoraco-abdominal views and bilateral lung scans were 1 min 20 s and 0 min 32 s, respectively. Combined, the mean total scan time to perform a telementored e-FAST was 1 min 52 s. The time measurements for each position were limited to probe-on-skin until image capture. The results for total scan time per protocol, defined as the sum of the mean scan times of each position, are shown in Table 3.

It is demonstrated in this study that US examinations combined with audio and camera views can be streamed in real-time via a standard internet connection found onboard offshore installations alongside the Norwegian continental shelf. Furthermore, the results show that an expert user can guide an offshore nurse, novel to US, through PoCUS protocols with a high degree of precision and within an acceptable scanning time. Finally, the vast majority of the stored focused heart, lung and abdominal cineloops are of high-quality images suitable for clinical interpretation. These findings are relevant and may enable telementored US to increase diagnostic accuracy with patients in remote locations.

Although it may not be reasonable to consider PoCUS in all clinical settings, a common theme in austere environments is limited access to healthcare resources and imaging capabilities [3, 10]. As one would expect, many of the US applications found useful in-hospital are equally applicable out-of-hospital. Despite the versatile range of diagnostic and procedural applications, implementation of PoCUS in prehospital care has been slow due to both technical and educational barriers. It is thus important that any equipment carried to a remote location is lightweight, durable and rugged [32]. Many of the current machines are portable and battery-powered and can be brought to the patient regardless of location [31]. They are increasingly cheap, robust and produce high-quality images. Technological advances have led to truly hand-carried PoCUS devices [33]. Another important technical feature of these machines is the capacity for the connections to different platforms, systems and applications [34, 35]. A systematic review by Gopaul et al. [36] lists the lack and cost of formal US training as the main barriers to implementing US in resource-limited settings. Current World Health Organization (WHO) recommendations for US training are quite extensive and involve 300–500 US scans per physician to achieve an acceptable skill level [37]. Our study shows that this educational barrier can be lowered by having experts connect to novice users and guide them through PoCUS protocols. By simply knowing how to operate the on/off button, gain and depth adjustments and save function on the M-Turbo machine, the offshore nurse procured cineloops and images of diagnostic value (score ≥ 3) in 96.4% of the cardiac views and 79.1% of the thoraco-abdominal views. These results clearly indicate that telementored US has the potential to augment clinical decision-making when used as an adjunct to the standard physical examination and perhaps being the only realistic imaging modality available to offshore patients.

In trauma care, it is important to detect the presence of major thoracic and abdominal fluid collection in patients with haemorrhagic shock. The FATE [38] and FAST [39] protocols were developed to efficiently diagnose or rule out a variety of life-threatening conditions. FAST is the basic building block for PoCUS and one of the most studied US tools for haemodynamically unstable patients. Not only it is an effective scan to diagnose free fluid in the thoraco-abdominal cavities and pericardial sac, but it can also be performed quickly and at bedside [39]. The results show that telementored FAST took only 1 min and 20 s, with an additional 32 s to scan the lungs (i.e., extended version of FAST). Although more comprehensive (4 min and 15 s), FATE is one of the most valuable scans for cardiac function, potentially revealing the causes of cardiac arrest and circulatory shock. This protocol has been developed to define the aetiology of circulatory collapse and more specifically to identify reversible causes during ongoing resuscitation (i.e., diagnosing fluid in the pericardial sac or pleural space and evaluating the size, shape and function of the ventricles) [25, 28]. For critically ill and injured patients, time is always critical, and many deaths are preventable if reversible causes are recognized and treated expeditiously [40, 41]. In remote settings, it is important that patients are triaged and transported to the correct level of care. PoCUS examinations performed on-site within a few minutes are thus clinically extremely attractive [3]. The mean scan time for the FATE protocol was longer than FAST but involved obtaining six views rather than four (Table 3). The time recorded in our study was from skin contact to cineloop or image capture, not the time in between the different scanning positions. The reason for this was that the volunteers in this study were on duty when being scanned and consequently had to be available for phone calls and urgent matters if needed. As a result, several of the scanning sessions were disjoined and for this reason the total scan time was not recorded. Total scan time would, therefore, be slightly longer, but both protocols can be performed in a time frame fully comparable to many other diagnostic procedures (e.g., CT, ECG, heart and lung auscultation, arterial and venous blood sampling).

The diagnostic quality of the cineloops obtained from the FATE protocol exceeded those from the FAST protocol, which was a surprise to us, as cardiac scans have been perceived as more cumbersome and difficult to acquire. When looking at the image score by acoustic window for FAST (Table 2), the subcostal 4-chamber view stands out with a much lower %^≥3 score (57%) than the other three views (84%, 78% and 97%, respectively), bringing the total %^≥3 score (79%) for this protocol down considerably. The subcostal 4-chamber view is also part of the FATE protocol (Table 1), but the ≥ 3 score was 92%. This discrepancy could be explained by the different transducers used for FAST and FATE. The transducer for the FATE protocol uses phased-array ultrasonics where the US beam can be swept electronically without moving the probe (i.e., beam steering technology). This difference makes it better suited for visualizing a beating heart compared to the more fixed US beam emitted from the curved-array transducer used for the FAST protocol. As an alternative, the phased-array transducer can also be used for FAST. The total FAST %^≥3 score in our data (Fig. 4b) could, therefore, theoretically increase from 79.1 to 88% if the probes were switched for the subcostal 4-chamber view. However, looking at all data combined, diagnostic information can be extracted from 87.8% of all images (%^≥3). Furthermore, our study shows that evaluating the lungs is rapid and feasible in all patients. This finding is important as evaluating the pleural line for horizontal sliding (i.e., lung sliding) could distinguish between a normal lung with lung sliding and a pneumothorax without lung sliding [29, 42]. This diagnostic amendment has been added to the standard version of FAST, hence the name e-FAST [43]. In our study, scores were not given for the lung scans (video for lung sliding and M-mode image of the pleural line), but these images and videos were evaluated for whether presence of lung sliding and seashore sign (M-mode) were visible. The expert user found that the pleural line and lung sliding were visible in all images and videos obtained (100%). This high feasibility is likely because the pleural line is located superficially (i.e., close to the anterior chest) in most people, regardless of body state, thus making it easy to visualize with US. However, there are some limitations to lung US, such as the presence of dressings or subcutaneous empyema [44].

Previous literature on telementored US discuss feasibility, technicality and outcomes of having a remote expert guiding novice users in different PoCUS protocols. A selection of relevant studies is shown in Table 4. The research includes a variety of locations where US images had been transferred. The majority of the publications are feasibility studies on healthy volunteers, whereas studies looking at real patients seem to, for practical reasons, take place at hospitals [17, 22, 45]. A study by McBeth et al. [46] also used a phantom to assess the participants’ ability to identify simulated pathology. FAST and e-FAST are the most commonly studied protocols, but in a few publications, the US examinations were adapted to meet specific patient needs [22, 47]. The communication equipment used is described as basic and can be bought off-the-shelf, such as laptops with a web camera, mobile phones with FaceTime™ or similar software apps and head-mounted cameras (e.g., Go-Pro™) [17, 22, 46‐49]. One exception, similar to our study, is a publication by Olivieri et al. [45] where an already available advanced telecommunication system was used. In all studies, a recurrent theme is that data transmission (i.e., real-time transfer of US and video) is limited by network availability and internet access [3]. Therefore, telementored US may not be suitable for all geographic areas. A systematic review by Marsh-Failey et al. [50] suggests a minimum bandwidth for image transfer of 500 kbps, but the studies described in Table 4 use either a 3G or 4G data network or public wireless network (WiFi), with no significant delay in image transfer reported. In our study, we used a fibre-optic internet line that was part of a deep-sea cable stretched out on the seabed from Statfjord C to an onshore relay station. We did not experience any problems with the transfer of US images from the M-Turbo machine or the roof camera from the hospital bay with both split-screens displayed on the onshore computer (Fig. 2). In contrast, the onshore US expert perceived all PoCUS examinations in real-time with no delay in audio or video signals. The nurse performing the scans could, as described, at all time see the US expert in real-time on the Medical unit screen. The possible impact this might have on the scan performance has not been investigated or analysed. As the videoconference setup allowed for bidirectional video, disabling this function was not considered. It was not crucial for the nurse being able to see the US expert as all instructions given were verbally. The US expert has however stated that he considered it a positive element in guiding process, that both pars could see each other real-time.

Remote oil and gas operations present a multitude of health risks. However, they can be kept as low as reasonably practicable by implementing a number of strategies, such as health risk assessments, medical emergency response planning including medevac, healthcare practitioner competency requirements, remote medical support and telemedicine [5, 6, 12]. To our knowledge, we are the first to explore the use of telementored US in an offshore setting. A few oil companies have, since the data collection period in 2012, installed US in their hospitals, but these installations have almost exclusively been in extremely remote locations (i.e., sites where medical evacuation to a hospital can never be achieved within 4 h, even in the best of circumstances). Increasingly, the energy industry is operating in these environments with the goal of protecting the health of the offshore workers at an equal standard as their non-remote counterparts. As part of their remote healthcare strategy, Equinor ordered an investigation of their offshore medical healthcare systems to describe the use of their SAR helicopters and register the patients’ diagnoses or symptoms leading to emergency medevac. This process resulted in a prospective study by Østerås et al. [13] looking at baseline characteristics of all patients (n = 382) evacuated by SAR from three Equinor offshore installations in the North Sea over a time period of 2 years. The top three diagnoses or symptoms were chest pain (n = 102), abdominal pain (n = 75) and trauma (n = 68), making up 65% of all patients. PoCUS may be used as a diagnostic tool for all these conditions and some others on the list [e.g., cardiac arrhythmia (n = 10), breathing difficulties (n = 7), cardiac arrest (n = 4) and obstetrics (n = 1)]. Altogether, telementored US could possibly be useful in up to 71% of patients being evacuated from an offshore platform. Furthermore, out of all SAR missions, 71% were a code yellow. This code is used in situations where a patient’s condition is not immediately life-threatening (i.e., code red) but often undetermined and may deteriorate. These patients are often evacuated if the offshore nurse and onshore physician feel that further examination and early treatment are necessary. Some patients are even evacuated as a preventive action or simply because of limited access to diagnostic capabilities at the offshore hospitals. There are high costs involved in using the SAR service for medical evacuations, and by adding PoCUS, the nurses might be able to identify and monitor numerous conditions that safely can be treated on-site. Hence, oil and gas companies may prevent health complications, minimize unnecessary medical evacuations (code yellow), facilitate necessary ones (codes yellow and red) and optimize care during transfer (codes yellow and red) if they implement telementored US offshore. Figure 5 shows the setup of an US machine connected to the Medical unit at an offshore hospital today.

From a more global perspective, the World Health Organization (WHO) states that diagnostic imaging is a necessary procedure for accurately treating at least 25% of the worlds’ population and that X-ray and US alone can meet over 90% of imaging needs [51]. Therefore, they recommend X-ray and US be available for all patients in primary healthcare settings [52]. However, US technology has several advantages compared to X-ray. It is cheaper, portable, battery-powered, dynamic and does not involve ionizing radiation [16]. For all these reasons, telementored US should be part of a global commitment to provide clinical support and overcome geographical barriers with the overall aim of improving health outcomes.

First, this study was a small quality assessment study with a limited number of healthy participants (n = 37) where all volunteers were male. Second, the telementored US examinations were between the same offshore nurse and onshore expert in all cases which most likely improved the scan time and image quality over the course of the study. However, the data collection was done over a short period of time and for that reason we chose not to analyse the progression potentially reflected in the improved scan time and image quality. Third, the transmission times for audio and video signals (both US and roof camera) were not calculated, but only perceived by the onshore expert to be in real-time without any delays in synchronous streaming. Finally, we used an advanced medical communication unit (Medical unit) connected to the internet via a high-speed fibre-optic cable and did not encounter any issues with the technical equipment during the 2 days of scanning. For these four reasons, the generalizability of our results may be diminished. However, our data analysis is based on a total of 518 telementored US scans, all performed in exactly the same way with the same collaborating nurse and physician. Previous studies have not described the need to use advanced communication systems but rather various cheap and commercially available communication hardware and software (Table 4). Furthermore, a systematic review shows that even a bandwidth speed as low was 500 kbps could be sufficient for synchronous streaming with good image quality [50]. In our study, the US images and videos were retrieved from the US machine to a USB stick and reviewed retrospectively; hence, image degradation due to bandwidth speed was not an issue when evaluating image quality scores. Last, we want to comment on the fact that the data collection was carried out several years ago. There has not been much change with regard to the use of offshore telementored ultrasound and the equipment used in this study is still in use today. We therefore believe that the data are still valid and applicable.

Telementored US using existing communication and network infrastructure available at offshore oil and gas installations in the North Sea is feasible and allows real-time sharing of US cineloops and images. Remotely located experts onshore can guide inexperienced offshore nurses through different PoCUS examinations with ease, such as FATE and e-FAST. Finally, the vast majority of telementored US images and cineloops procured by novices are of high enough quality to visualize relevant anatomical landmarks and to extract diagnostic information. Future research should focus on clinical outcomes of implementing telementored US in remote locations.