November 2021


January 2022

H.G. Kreeftenberg (1,2), J.T. Aarts (2), C. Kerkhoven (2), J. van Rosmalen (3), A.J.G.H. Bindels (1,2), P.H.J. van der Voort (4,5)  Departments of 1 Internal Medicine and 2 Intensive Care, Catharina Hospital, Eindhoven, the Netherlands 3 Department of Intensive Care, Maxima Medical Center, Veldhoven, the Netherlands 4 TIAS school for Business and Society, Tilburg University, Tilburg, the Netherlands 5 Department of Critical Care Medicine, University Medical Centre Groningen, University of Groningen, Groningen, the Netherlands


H.G. Kreeftenberg -
Original Article

A survey on implementation of physician assistants in ICUs in the Netherlands


Background: In several countries, the advanced practice provider (APP) is positioned as a clinician in an acute care setting. In the Netherlands, physician assistants (PAs), the equivalent of the APP, are increasingly being employed in ICUs. It is unknown to what extent PAs are present in Dutch ICUs, what they do and how they are appreciated. To provide a general overview, we conducted a survey.

Methods: The survey was sent to intensivists and PAs working in ICUs in the Netherlands. The survey focused on familiarity with the PA, the underlying reasons for implementation and their tasks.

Results: Sixty-five intensivists (representing 85% of all hospitals) and 43 of the 55 (78%) Dutch PAs responded. Twenty ICUs employed PAs and eight were considering doing so. In these ICUs, both intensivists and PAs were satisfied about their performance. Quality improvement and continuity of care were the main reasons mentioned for implementation. In 60% of the ICUs, the PAs performed medium complex tasks such as inserting central venous catheters, in 20% PAs performed complex tasks like treatment of unstable patients with intubation. The main reasons for not implementing PAs were: ‘the intensivist performed the care without residents’, ‘ICU care is too complex for PAs’, ‘no need for additional personnel, enough residents available’ and ‘too expensive’.

Conclusion: ICUs in the Netherlands are recognising the potential of a PA on the ICU. Especially larger ICUs are already implementing PAs. Despite this, the majority of ICUs do not yet employ PAs. Presumptions about the profession sometimes inhibit acceptance of these professionals, although PAs are highly appreciated in the ICUs where they do work.

Physician assistants (PAs) are increasingly being implemented in Dutch intensive care units (ICUs). Internationally, these PAs, often designated as advanced practice providers, have been trained to operate in the medical domain of intensive care and to facilitate and evaluate the daily medical problems together with an intensivist and ICU nurses. Because PA is a relatively new profession, the current status of implementation and the tasks delegated to PAs in Dutch ICUs is unclear. Several international studies about implementation have been performed and it is clear that the Dutch ICUs are frontrunners.[1] We conducted a survey among intensivists and PAs working in ICUs to gain insight into the current status of the acute care PA in the Netherlands.

The survey was sent to all 73 Dutch ICUs by email via the Netherlands Intensive Care Society (NVIC). A reminder for response was sent after four weeks. Eventually, the nonresponders from the Dutch survey were approached by telephone. In addition, all PAs working on the ICUs in the Netherlands were asked by email to complete the survey. A reminder was sent to the PAs via Whatsapp-messenger after six weeks.

The survey
The survey consisted of questions about the baseline characteristics of the hospital and ICU such as number of beds, type of hospital and ICU, and number of residents and PAs. The second part consisted of questions about acquaintance with and the position of PAs in the ICU. Additionally, the survey addressed the underlying reasons for either implementing or not implementing this profession and the extent of the tasks of the PA. A flowchart representing the survey is presented in figure 1. In questions about amounts, a number had to be filled, in questions about motivation of choices, up to seven predetermined options could be filled in, and for binary questions yes or no could be filled.

Volume 30 - issue 4 - article A survey on implementation of physician assistants in ICUs in the Netherlands_1

Based on the Dutch national guidelines of the Central Committee on Research Involving Human Subjects (the CCMO), ethical approval was not necessary because the survey involved nonpatient data and the participation of intensivists and PAs was voluntary.

The analysis was conducted with Excel for Windows 2013 (©2013 Microsoft Corporation) and SPSS 25 (IBM Corporation, Armonk, NY, USA). Results are presented in absolute numbers and percentages. Non-parametric continuous data were analysed with the Mann-Whitney test. The distribution was given with the interquartile range (IQR). Correlation was calculated by the Spearman correlation coefficient for nonparametric data. Dichotomous data were analysed with the Chi-square test with Yates continuity correction and when distributions were less than five, the Fisher’s exact test
was used.

Survey among intensivists
Intensivists from 65 hospitals in the Netherlands responded (85%). Twenty of the 65 participating hospitals had already implemented or are implementing PAs on the ICU. In eight ICUs, use of a PA was under consideration. Thirty-seven hospitals reported not implementing PAs on the ICU. Together, the ICUs employed 51 full time equivalent (FTE) PAs (55 persons).

Of the 20 hospitals working with PAs, 13 were large teaching hospitals and university medical centres (UMCs), which is half of the total number of responding large teaching hospitals. The other seven hospitals that used PAs were general hospitals. Eight hospitals are in the process of or considering using PAs, five of these are general hospitals.

Of the UMCs, three of the eight were already using PAs on the ICU, and three were starting to employ PAs. Two UMCs reported having enough residents to guarantee continuity of care without PAs (figure 2).

There was no significant difference in implementation of PAs between UMCs and other large teaching hospitals (p=1.00) or general hospitals (p=0.11). The number of PAs implemented in large teaching hospitals compared with general hospitals was significantly higher (p=0.03).

There was a significant correlation between the size of an ICU and implementation of PAs (r=0.361, p<0.001) which implies that the larger ICUs employ more PAs than the smaller ICUs.

Figure 3 summarises the reasons for ICUs to consider working with PAs. The reasons were: solving physician shortage (n=12), improvement in continuity of care (n=20), improvement in quality of care (n=15), career perspective for ICU nurses (n=20) and workload relief by delegating tasks from physicians to the PA (n=7).

Nine of the participating ICUs were unfamiliar with PAs and did not work with them. The reasons for not having PAs were: no need because there were no shortages in residents (n=3); care provided by non-physicians is inferior to care provided by physicians (n=2); no reason (n=1); not yet discussed within group (n=1). Two ICUs worked with intensivists only, so without residents or PAs.

Volume 30 - issue 4 - article A survey on implementation of physician assistants in ICUs in the Netherlands_2
Volume 30 - issue 4 - article A survey on implementation of physician assistants in ICUs in the Netherlands_3
Volume 30 - issue 4 - article A survey on implementation of physician assistants in ICUs in the Netherlands_4

Figure 4 summarises the reasons for not working with PAs. Thirty ICUs were familiar with PAs but did not actually work with them. Their considerations for not working with PAs: no need for PAs because there is no shortage in residents (n=11); no need for working with residents or PAs (n=11); too expensive (n=2); time investment is too great (n=1); ICU care is too complex for PAs (n=2); have not considered it yet (n=3). Not all respondents who did not implement PAs stated a reason.

Survey among PAs
Of the 55 PAs working on Dutch ICUs, 43 responded to the survey (78%). Eleven of these PAs are still in training. The previous job of most PAs was ICU nurse (n=41, 95%), one was a physiotherapist and one was an anaesthesist nurse. The mean age of the group was 44 (IQR 12) years. The median time of employment was six years (IQR 7). The PAs are predominantly employed in larger hospitals with the minimum number of beds being 10 (median 15.5 beds). These hospitals implemented a mean of 2.0 (IQR 2) FTE PAs. The main reason for working with PAs was providing continuity of care (n=42, 95%) and high quality of care (n=28, 65%). Other reasons were providing career opportunities for ICU nurses (n=25, 58%) and alleviating the workload of the intensivists (n=14, 33%).

Communication of PAs as assessed by intensivists
Three out of the 23 intensivists who reported on communication of the PA thought that their communication with patients was as good as that of the intensivist (13%). Six intensivists reported they considered the communication of the PA less proficient than both resident and intensivist (26%) and 13 intensivists considered the communication of the PA less proficient than the intensivist but more proficient than residents (57%). One intensivist considered the communication of the PA more proficient than both resident and intensivist.

Regarding communication with other attending physicians, one intensivist considered communication with PAs to be equal to intensivists, nine intensivists considered it less proficient than the intensivist and resident (39%) and 13 intensivists considered it less than the intensivist but better than a resident (57%).

Communication assessed by PAs
All 32 certified PAs felt that their communication toward patients and family was adequate. Eighteen respondents felt that their communication with patients was equal in quality compared with residents and intensivists (56%). Twelve respondents felt that they communicated less proficiently than the intensivist but more adequately than the resident (38%) and two PAs graded themselves better in communication than both intensivist and resident (6%).

Eleven of the certified PAs felt that their communication towards the attending physicians was as good as the communication of the intensivists themselves (34%). Twenty felt that their communication was less effective than intensivists but more effective than residents (63%) and one felt that the communication of the PA was more proficient than both intensivist and resident (3%).

Shifts and other tasks assessed by intensivists
Of the 20 ICUs with certified PAs, the PAs did not work night shifts in five ICUs (25%). In eight of the ICUs the PA did not participate in the rapid response team (40%). All 20 ICUs reported that PAs performed low-complex tasks such as arterial catheter placement or treatment of the uncomplicated postoperative patient. Twelve (60%) reported that the PAs performed medium-complex tasks such as placement of central venous catheters or supervising electrical cardioversion performed by residents. Four of the 20 ICUs (20%) reported that PAs performed high-complex tasks such as consultation in the emergency department or unsupervised treatment of unstable patients.

Shifts and other tasks assessed by PAs
All 32 certified PAs responded to have supervision assignments and responsibilities, 26 supervised inexperienced residents (81%) and 20 supervised invasive procedures (63%). Twentynine (81%) PAs reported they participated in evening, night and weekend shifts.

Interventions assessed by intensivists
All 20 ICUs reported that certified PAs inserted arterial lines (100%). Eighteen of the 20 ICUs (80%) reported that PAs inserted central venous catheters and performed electrical cardioversions unsupervised. Seventeen ICUs (75%) reported that the PA performed complex intra-hospital transport. Intubations performed by PAs were reported in 12 of the 20 ICUs (60%).

Interventions assessed by PAs
All of the certified PAs reported inserting both arterial catheters and central venous catheters. Twenty-five of the 32 certified PAs reported performing intubations (78%), 28 reported performing electrical cardioversion and intra-hospital transportation with mechanically ventilated patients (88%) and 14 reported inserting thoracostomy tubes (44%). Twenty-seven PAs experienced a shift of the performance of invasive procedures from residents towards PAs (84%).

Employment and international cooperation
Seven intensivists expected the PA to stay on the same ICU for more than 10 years, six of them expected them to stay for five to ten years and one ICU expected them to stay two to five years. Seventeen ICUs want to expand the number of PAs. Moreover, 17 ICUs think that international cooperation would benefit the implementation and expansion of the profession. Eight ICUs do not think that international cooperation would benefit the integration of the PAs on Dutch ICUs.

Thirty-one out of 32 of the responding certified PAs think they would benefit from a more ICU-centred training or acute care training after their graduation to PA. Of the 32 certified PAs, three of them felt that they continuously had to live up to expectations and three PAs felt they were limited in their performance because their expertise was underestimated. One PA would have liked to expand research activities but did not have the opportunity to do so.

The results of this Dutch survey provide a general overview about the use of PAs in the ICUs in the Netherlands. It shows that PAs are predominantly working in high-volume ICUs. Of the Dutch ICUs, 42% are considering or have already implemented PAs to improve the quality and continuity of care. When the smaller ICUs with only intensivists and without additional personnel are excluded, up to 58% have implemented PAs in their ICUs, either on top of existing formation or by replacing some of the residents by PAs. Next to quality and continuity of care, also career opportunities for employees is considered an argument to implement PAs.

When focusing on the specific tasks of the PA, such as communication or procedures, not much has been reported in the literature about evaluation of communication of acute care PAs. One study described perceptions of nurses, PAs and physicians regarding each other’s communication. PAs communicated adequately, which is in line with the results from our survey.[2] Three intensivists thought that the communication skills were equal between PAs and intensivists. Intensivists also agreed that the communication skills of the PA were better than those of the residents. However, PAs tend to assess their skills on patient and attending physician communication better than intensivists do. Intensivists and PAs both reported that PAs were able to perform the most common procedures. More difficult procedures such as insertion of central venous catheters or intubation but also participation in a rapid response team are not always undertaken by PAs, although there is evidence that these procedures can safely be performed by other care providers than intensivists.[3-9] Moreover, the legal foundations in the Netherlands allow them to do so.[10]

The majority of ICU’s do not yet work with PAs (58% vs 42%). The main reported reason is a policy decision, mainly in the smaller ICUs, to employ only intensivists and no other physicians or non-physician care providers. This budget containment is in contrast with a recent review about the cost-effectiveness of the PA and discusses financial and quality advantages.[11] If residents are readily available, there is also no incentive to explore the profession of the PA. Of the ICUs, 10% presume that quality issues will arise when employing PAs, although they did not actually use PAs. ICUs which do implement PAs, however, do not recognise these quality issues. These two contrasting opinions may partly be explained by misconceptions about the capabilities of PAs, possibly amplified by the time it takes to train a PA.

Alongside this Dutch survey, a European survey was distributed (unpublished results) which returned very limited results and showed that most European countries are unacquainted with this profession. This confirms the finding in the literature that the profession of PAs working in critical care has matured significantly in only a few countries: the United States of America, the Netherlands, Australia and the United Kingdom.[12]

This study has strengths and weaknesses. The high response rate from the Dutch survey together with agreement with the limited available literature and the accordance between intensivists and PAs, provides credibility and paints the picture of a relatively unknown profession with diverse potential. However, a weakness of every survey is the number of unanswered questions when zooming in on opinions of respondents. This may reflect unfamiliarity with the PA or differences in understanding and interpretation of the questions. In addition, the survey is not a validated questionnaire which implies that the questions might be multi-interpretable. Therefore, an acquiescence bias could be present: although the answers did provide negative answers and the survey was distributed by the NVIC and not by an individual person, bias could have been introduced because of the tendency to answer questions positively to avoid conflict. Furthermore, because some questions could be answered with more than one item, the response order effect could play a role. The number of questions with multiple items was, however, limited.

International cooperation will be helpful to expand acquaintance with PAs and to improve general awareness. Also, in the Netherlands, qualitative research into this profession will generate evidence which will standardise implementation and possibly convince the critics that PAs are a respected additional workforce in the critical care setting.

We have shown that familiarity of working with PAs in the critical care setting is increasing but not extensive in Dutch ICUs. In ICUs which have implemented PAs both intensivists and PAs are satisfied about their performance. PAs perform most tasks that are usually done by residents. Presumptions about the profession are barriers to the implementation of PAs.

All authors declare no conflict of interest. No funding or financial support was received.


  1. Kreeftenberg HG, van Rosmalen J, Aarts JT, van der Voort PHJ. Physician assistants in intensive care units in the Netherlands: a narrative review with recommendations. Neth J Crit Care.2020;28:200-4.
  2. Aslakson RA, Wyskiel R, Shaeffer D, et al. Surgical intensive care unit clinician estimates of the adequacy of communication regarding patient prognosis. Crit Care. 2010;14:R218.
  3. Bevis LC, Berg-Copas GM, Thomas BW, et al. Outcomes of tube thoracostomies performed by advanced practice providers vs trauma surgeons. Am J Crit Care. 2008;17:357-63.
  4. Sirleaf M, Jefferson B, Christmas AB, Sing RF, Thomason MH, Huynh TT. Comparison of procedural complications between resident physicians and advanced clinical providers. J Trauma Acute Care Surg. 2014;77:143-7.
  5. Kreeftenberg HG, Aarts JT, Bindels A, van der Meer NJM, van der Voort PHJ. Procedures Performed by Advanced Practice Providers Compared With Medical Residents in the ICU: A Prospective Observational Study. Crit Care Explor.2020;2:e0101.
  6. Gershengorn HB, Xu Y, Chan CW, Armony M, Gong MN. The Impact of Adding a Physician Assistant to a Critical Care Outreach Team. PLoS One. 2016;11:e0167959.
  7. Gupta S, Balachandran M, Bolton G, et al. Comparison of clinical outcomes between nurse practitioner and registrar-led medical emergency teams: a propensity-matched analysis. Crit Care. 2021;25:117.
  8. Scherr K, Wilson DM, Wagner J, Haughian M. Evaluating a new rapid response team: NP-led versus intensivist-led comparisons. AACN Adv Crit Care. 2012;23:32-42.
  9. Stempek S, Wozniak J, Janz D, et al. Outcomes of ICU intu-bations performed by advanced practice providers compared to physicians. Crit Care Med.2019;474:479
  10. Gnirrep I, van Tilburg-van Hedel SW, Kreeftenberg HG, Lasut W, van Rosmalen J, Kerkhoven C, Aarts JT. Consensusdocument taakherschikking Intensivist – Physician Assistant-IC. Accessed 01 11 2022.
  11. van den Brink G, Hooker RS, Van Vught AJ, Vermeulen H, Laurant MGH. The cost-effectiveness of physician assistants/associates: A systematic review of international evidence. PLoS One. 2021;16:e0259183.
  12. Kreeftenberg HG, Pouwels S, Bindels A, de Bie A, van der Voort PHJ. Impact of the Advanced Practice Provider in Adult Critical Care: A Systematic Review and MetaAnalysis. Crit Care Med. 2019;47:722-30.