A point-of-care thoracic ultrasound protocol for hospital medical emergency teams (METUS) improves diagnostic accuracy

Medical Emergency Teams (METs) are called to the bedside when patients on hospital wards deteriorate. [1, 2] METs use various algorithms to assess the patient’s condition, most frequently the ABCDE method. The MET physician will use a combination of history, physical examination and point-of-care laboratory tests to assess the patient. The addition of point-of-care ultrasound (POCUS) could potentially improve diagnostic accuracy. [3] METs are associated with a reduction in patient mortality [4], but few data exist on differences in operation procedures by METs. The role of POCUS for instance during MET calls has not been investigated extensively yet, even though its role in the emergency room (ER) and intensive care unit (ICU) is well established [5‐13]. Recently, Zieleskiewicz et al. published the first prospective observational study on the effect of the use of a multi-organ POCUS protocol during MET calls [14]. In this study, the use of a multi-organ POCUS protocol improved diagnostic accuracy of the MET significantly. We also designed a prospective trial hypothesizing that multi-organ POCUS would increase MET diagnostic ability and postulated that the use of POCUS would also increase the diagnostic certainty of the MET physician. Influence on certainty by the use of POCUS has been found in a study on ER patients [5] and improving certainty might also be important to attending MET physicians.

Because most MET calls are requested for respiratory and or hemodynamic deterioration [15], we designed a POCUS protocol consisting of cardiac and lung (thoracic) ultrasound.

We included a total of 100 patients, 52 patients in the US+ group and 48 patients in the US− group. In 5 patients the independent expert could not determine a reliable definite diagnosis (2 in the US+ and 3 in the US− group). These patients were excluded from the comparison of initial/second diagnosis with the definite diagnosis, but were included in other analysis. In total, there were 310 MET calls during the study period.

Flowchart of study enrollment and exclusion reasons are in listed in Fig. 1.

Patients characteristics are described in Table 1.

In this single-center prospective observational study the use of a thoracic POCUS protocol improved the number of correct diagnosis significantly. We also found that the number of correct diagnosis increased in the US+ group after the use of POCUS and that POCUS increased MET physician’s certainty significantly. We did not find differences in mortality, first MET treatments or supervisor attendance.

There are multiple studies evaluating the use of POCUS in the ICU and ER department, until now only two studies focused on the use of POCUS during MET calls. Zieleskiewicz et al. published a paper [14] in which they evaluated the effect on diagnostic adequacy of thoracic POCUS during MET calls. As in our study, they found a significant increase in the number of correct diagnosis when POCUS was used (80% versus 94%). Furthermore, the time to first treatment was significantly lower in the POCUS group and there was an association with outcome parameters such as mortality, but the latter was not confirmed in the propensity score. Although this study has many similarities with ours, there are also some important differences. Both studies are single centered and prospective observational studies. The inclusion criteria are comparable as is the inclusion rate over time and the number of inclusions out of the total MET calls (34% Zieleskiewicz et al. versus 32% Blans et al.). In both studies, the protocol consisted of cardiac and lung POCUS, Zieleskiewicz et al. also used vascular POCUS to rule out lower extremity thrombosis, but the latter was used infrequently. Both studies found that the use of POCUS during MET calls improved the number of correct diagnosis made by the MET physician (primary endpoint). In both studies, chart review was used to establish a definite diagnosis, but there are some differences in the exact way in which this process was carried out. In contrast to the Zieleskiewicz et al. study, we used the term “partly” correct diagnosis if not all elements of the definite diagnosis made by the external expert were scored by the MET physician. If we add the partly correct to the completely correct numbers, our results would increasingly be comparable to the Zieleskiewicz results (US+ group 94% correct and US− group 69% correct).

Another important methodological difference with our study is the fact that Zieleskiewicz et al. used two MET teams, one which used POCUS and one did not. The two MET’s alternated every other day. Both MET deployments (POCUS and not using POCUS) were considered standard therapy and therefore no consent was deemed necessary. We, however, choose to use the same MET but asked to use POCUS only during even weeks and discouraged the use of POCUS during odd weeks. Because Zieleskiewicz et al. used two separate METs, theoretically the difference in the number of correct diagnosis was not only the result of the use of POCUS, but also due to differences between the achievements of the two separate teams although a large number of seniors and juniors randomly composed each MET. In our study, the MET physician was the same during both weeks (US+ and US−).

Furthermore, it is unclear in the Zielskiewicz et al. study who exactly performed the POCUS protocol, their METs are staffed by junior and senior physicians (with minimally a level 2 in thoracic ultrasound). We also found a significant impact on diagnostic accuracy, but in our study POCUS was done by residents only, this aspect is worth emphasizing; POCUS can be of an extra value to the less experienced physician during MET calls. We found that without POCUS the number of correct diagnosis was relatively low (51.1%). Although this has been found before in studies in which POCUS was done by more experienced staff [11, 20], we think that this low percentage can also be partially explained by the fact that in our study less experienced physicians (residents) included the patients.

Several studies show that training programs for residents in multi-organ POCUS have satisfactory results in terms of acquiring adequate ultrasound skills and increasing diagnostic abilities. [21‐28]. Our study supports the evidence that residents may obtain clinically relevant POCUS skills in a relatively short period of time, including making the right diagnosis during a MET call. Our findings are of interest for other hospitals in which the MET is also staffed by residents; POCUS training will improve their diagnostic ability and certainty also during acute situations like MET deployments. Also important to stress is the fact that the use of the POCUS protocol never resulted in a change towards a completely incorrect diagnosis.

Because we asked the MET physician to note to a diagnosis before the use of POCUS and one after the use of POCUS, we could measure more precisely the effect of the use of POCUS and this was significant in the US+ group and also positive though not significant due to small numbers in the US− group. This also supports the fact that POCUS was the reason for more diagnostic accuracy.

The other study on the use of ultrasound by MET’s was published by Sen et al. [29]. In this small study of 50 patients on the effect on diagnosis of a combined lung and lower extremity vascular POCUS protocol was evaluated. They showed that lung POCUS was feasible, but due to the small number of patients there was no statistically significant effect on the number of correct diagnosis. They therefore concluded that the use of lung ultrasound was non-inferior to MET clinical assessment. It could well be that by the addition of cardiac POCUS and the inclusion of more patients Zieleskiewicz et al. and we were able to prove a beneficial effect of POCUS on diagnosis during MET calls.

We also evaluated the effect of POCUS on physician’s certainty. We showed that in the US+ and US− groups baseline certainty was quite high (eight on a 10 point VAS scale), but certainty increased significantly in the US+ group after the use of POCUS indicating that POCUS did not only improve certainty because certainty levels were low to begin with. There is one other study on the impact of POCUS on diagnostic certainty [5]. Shokoohi et al. looked at 118 ER patients and found that the use of a multi-organ POCUS protocol lowered uncertainty for a diagnosis. We are aware of the relatively small increase in certainty that was found in our study, but this finding could be psychologically important for residents during stressful clinical encounters such as MET calls. Junior physicians are less certain in the diagnostic process compared to more experienced colleagues [30].

We could not find differences in other outcome parameters (28-day mortality, initial treatment or supervisor attendance) possibly due to the small number of included patients. It will be difficult to design a study on the use of POCUS during MET calls large enough to detect differences in outcome parameters such as mortality. Preferably, a multicenter trial could help in including large enough numbers of patients. On the other hand, one could argue that the existing evidence is sufficient to support the incorporation of POCUS in MET protocols.

Our study has several limitations. This study was conducted in a single center and focused on ward patients with respiratory and or hemodynamic deterioration. Therefore, the presented results may not necessarily apply to other clinical settings. There were some differences in baseline characteristics between the US+ and US− groups. Significantly more patients in the US+ group had a history of myocardial infarction and heart failure. We have no reason to believe that these baseline differences had a significant impact on the study results. Our study design is prone to selection bias, but also has several advantages as discussed above. We decided to exclude the MET physicians without sufficient POCUS training from the trial. This resulted in a substantial number of non-included patients (87).

Our POCUS protocol consisted out of a combination of 5 basic cardiac questions and for the lung the Blue protocol was used which is consistent with current international POCUS practice. No further prespecified POCUS flow chart was used which makes our findings perhaps difficult to validate by others.

The diagnoses of the MET physicians were compared to the final diagnosis made by one blinded experienced intensive care consultant 2 weeks afterwards (independent expert) on the basis of a thorough chart review. This method remains challenging although often used in POCUS studies [12, 13]. In our study, only one independent expert reviewed the charts and was blinded to the ultrasound findings and diagnosis made by the attending MET physician. He had absolutely no other role in the study and because he is a member of the hospital mortality committee he is experienced in extracting official diagnoses from chart review.

We found that the use of a thoracic (cardiac and lung) POCUS protocol during MET calls due to respiratory and or hemodynamic deterioration has significant positive impact on establishing the correct diagnosis and a small but significant impact on MET physician’s diagnostic certainty.