Development and Maintenance of a Pleural Disease Service

Role of the 'Pleurologist'

Kelsey Cameron, PA; Dana Teodoro, PA; Azam Kasis, MBBS, MRCP; Matthew Evison, MD, MRCP


Semin Respir Crit Care Med. 2019;40(3):297-304. 

In This Article

Proceduralist and Outcomes

The human body is complex and the systems are inextricably intertwined. Similarly to caring for the human body there is no way to separate medical training, care provision, and patient safety. Therefore, in discussing pleural procedures we have to consider who is currently performing pleural procedures, who should be performing them for the best patient care, and how do we train and maintain the skill of the operators.

The main pleural procedures are thoracentesis, insertion of an intercostal drain, tunneled catheter placement, thoracoscopy, and pleural biopsy. The latter three are usually performed by medicine physicians with specialty interest and training in pleural procedures, radiologists, or surgeons, but thoracentesis and intercostal drain placement are often performed by many different health care providers in many different places, such as the emergency department, intensive care unit, and radiology and general medicine units.


In examining the literature on thoracentesis, we can see that a great deal of effort has been spent on identifying risk factors for complications. Some of this looks at who is performing the procedure and what effect that has on complications.

A systematic review and meta-analysis identified that thoracentesis done with ultrasound (US) significantly reduces the risk of a pneumothorax (odds ratio [OR]: 0.3; 95% confidence interval [CI]: 0.2–0.7).[1] The data is robust enough around the risk reduction US provides that some guidelines[2] and expert reviews endorse its use and the question has been posed whether absence of it is medicolegally defensible.[3] Therefore, regardless of the operator, US should be used. If the institution or proceduralist does not want to rely on calling another specialist, then the proceduralist performing a thoracentesis should be someone with access to an US and training in its use. This may limit the proceduralists because many do not get training on thoracic US as demonstrated by a survey of 117 general physicians in England.[4] The survey found that 19% of higher specialty trainees and 0% of core medical trainees had formal qualifications in thoracic US, 29 and 26% had informal training, and 48 and 68%, respectively, did not have training and did not expect to have access to any training.

It may be that US use is now serving as a defining characteristic for who performs the procedure. In a study of invasive bedside procedures on patients admitted to a general medicine service, they found that 87% of thoracenteses were referred to radiology or the hospitalist procedure service instead of being done by the primary team.[5] This may be because these services have US access and training. However, the study did not find the percent of thoracentesis referred differed from the number of paracentesis and lumbar puncture, so maybe the referral pattern is dictated by something different. When Duszak et al reviewed Medicare Physician/Supplier Procedure Summary master files from 1993 to 2008, they found that radiologists surpassed pulmonary and critical care physicians as the main provider of thoracentesis by 2003 and all other specialty groups substantially decreased in volume over the 15-year period as demonstrated in Figure 1.[6] They surmised that the safety provided by image guidance is a factor in the referral patterns to those with accessibility and training, as well as economic factors, citing low reimbursement being a reason providers refer a thoracentesis. Ault et al reported only 57 iatrogenic pneumothoraces in 9,320 procedures (0.61%), an extremely low rate, and all of these procedures were done with US.[7] There may be more than just US guidance responsible for such a low complication rate. The 9,320 thoracenteses were supervised or done by a single clinician who is the medical director of the procedure service. This extremely low complication rate may be attributable to the expertise of the physician. Aside from ensuring US use, we must examine other procedural factors and how they influence outcomes. Is there a difference in outcomes based on the level of experience of the operator, specialty, or location?

Figure 1.

Trends in the specialty performing thoracentesis from 1993 to 2008. Graphic demonstration of volume by radiologists, pulmonary critical care medicine (PCCM) physicians, primary care physicians, surgeons, and all other specialties.

In a study by Kay et al looking at patient characteristics that predict referral and procedural outcomes based on procedural characteristics, they found no difference in the risk of complication based on location of a procedure (radiology: n = 13 vs. bedside: n = 4; p = 0.62).[5] They also did not find a difference in complication rates for those referred (n = 23) versus those not referred (n = 4; p = 0.62); however, the complication rate was too low to properly power the study. Few studies comment on the outcomes based on the service performing the procedure. One study to help inform us is a retrospective cohort analysis of all bedside medical procedures performed by the medical procedure service (MPS) or primary service at a large teaching hospital.[8] In this study they tabulated attending participation in 99.7% of those procedures done by the MPS with resident participation compared with 47.0% of procedures performed by the internal medicine teaching service with resident participation. They also found that the MPS was more likely to use best practice safety measures (96.8 vs. 90.0%, p = 0.0004) and appropriate use of US guidance (96.8 vs. 90.0%, p = 0.00004). With supervision and best practice safety measures, including US use, one would expect the complication rate to be lower for the MPS. However, they found that 4 out of 149 thoracentesis procedures done by the MPS had a pneumothorax (2.7%) compared with 1 out of 72 done by the primary service (1.4%; p = 1.0). They were surprised by the lack of difference and offered the theory that residents who are uncomfortable and less experienced, and thus more likely to cause complications, are the ones that may preferentially refer the procedures. They also commented that patients with characteristics that may make complications more likely may have preferentially been referred. Also, they did find that both their MPS and primary service documented high senior clinician participation and use of best practice safety models and adherence to these may account for the low complication rates for bedside procedures.

As mentioned previously, thoracentesis procedures are being performed by specialty and subspecialty services more and more, and this may lead to attending and fellow-level clinicians performing the procedure. There are some studies that look at the training level of the proceduralist in relation to outcome measures. In a systematic review by Gordon et al, they found a lower pneumothorax rate with experienced operators (3.9 vs. 8.5%, p = 0.04).[1] However, when they restricted it to four studies that made direct comparisons, the relationship was nonsignificant (OR: 0.7; 95% CI: 0.2–2.3). When they specifically looked at the difference between experienced radiologists and experienced pulmonologists, there was no significant difference (4.3 vs. 4.4%, p = 0.81). These results leave us unsatisfied as to whether we truly get better results with senior-level clinicians. Perhaps the concentration of the skill, and not the level of training, is playing an important role. For example, in the Annals of Hepatology, there is a report of the outcomes of patients having a large-volume paracentesis by a nurse practitioner compared with the outcomes of patients having a large-volume paracentesis by a first- or second-year gastroenterology fellow.[9] They found no statistical difference in the number of complications (6.1 vs. 6%).

A thoughtful study out of the Mayo clinic provides great insight on whether concentrating a procedural skill impacts outcomes. Duncan et al retrospectively found a higher risk of complications after a thoracentesis in their pulmonary department compared with their radiology department (8.7 vs. 1.9%; p = 0.099).[10] They did not find a significant difference in complication rates in their department based on the level of training, physician, or physician-in-training (2 of 27 physicians [7.4%] vs. 3 of 31 physicians-in-training [9.6%]; p > 0.9). In response to their increased pneumothorax rate compared with their radiology colleagues, they implemented a quality-improvement intervention. As a part of their initiative they required the operator to demonstrate competency by attending a half-day workshop and have direct supervision on their first 10 thoracenteses. Ultimately, they ended up limiting the privileges to five proceduralists, out of 44 physicians in their department, who obtained the required competency and would maintain proficiency. After their quality-improvement project, their rate of pneumothoraces went from 8.8 to 1.1% (p = 0.0034). One cannot pinpoint the degree to which limiting the number of proceduralists affected the decrease in complication rates because it was a multipronged approach; however, the findings are compelling and argue for concentrating the skill and volume in the hands of a set of experts. Also, regardless of the exact contribution that the decision to limit the number of proceduralists plays in attaining these outcomes, they offer a method for creating a "zero-risk experiential training environment" that is generalizable and offers great potential for improving outcomes.

Chest Tubes

A National Patient Safety Agency report has drawn attention to the possible serious risks of intercostal chest drain (ICD) insertion and management.[11] Over a 3-year period they reported 27 cases of serious harm or death related to ICDs. Following this report a national survey was conducted by Harris et al to tabulate adverse incidents related to ICDs between 2003 and 2008.[12] They had 68% of the 148 trusts they queried, reply. They received reports of 31 cases of intercostal drain misplacement resulting in seven deaths. There were 47 reports of serious lung or chest-wall injury resulting in eight deaths and six cases of wrong-side placement with two deaths. The questionnaire also inquired about the local policies surrounding ICD placement, including the training required to obtain privileges to place them. They found that 30% allowed any physician, 27% required at least 1 year of training, 32% required at least 2 and 11% at least 4 years of training. Another questionnaire asked physicians-in-training at the Royal Liverpool Hospital and the Liverpool Heart and Chest Hospital if they observed or performed ICD placement and if they had performed it, were they supervised or unsupervised.[13] Lastly, they asked the participants to mark on a diagram the optimal location to place an ICD. Forty-four percent of the participants chose the optimal location within the safe triangle. When divided by specialty they found 25% (n = 3) of those working in cardiology or pulmonology chose a location within the safe triangle compared with 58% (n = 11) of those working in cardiothoracic surgery. Unfortunately, 16 (48%) who reported placing ICDs unsupervised chose a location outside the recommended area, and, of those who had been supervised, only 25% chose the correct location. These results lead us to believe that significant education and training needs to be implemented, trainees should be supervised, and those working in surgery are more likely to choose the correct location.

Intercostal drain placement is a necessary procedure in many locations, including the intensive care units, medical floors, and emergency departments. However, the frequency with which they are required is less than that of a thoracentesis. If experiential training, level of training, US use, and limiting the operators to concentrate on proficiency resulted in less complications after thoracentesis, the same is likely true for intercostal drain placement. In the next section, let us examine the literature on training methodology and what elements of a training program are likely to yield the greatest risk reduction.