Full-Time Intensivists, Patient Care, and Teaching

ABSTRACT & COMMENTARY

Synopsis: At one community teaching hospital, medical ICU length-of-stay and mortality declined and resident-in-training examination scores improved during the year after a full-time intensivist was added to the staff.

Source: Manthous CA, et al. Mayo Clin Proc 1997;72: 391-399.

Although critical care is provided mainly by full-time intensivists in many countries, in the United States, most hospitals do not have dedicated medical staff who spend all of their time in the ICU. Ever since the special qualification examination in critical care medicine was initiated by the American Board of Internal Medicine more than a decade ago, there has been vigorous debate about the clinical and economic effects of intensivists in American hospitals. This report describes what happened in one community teaching hospital in Connecticut during the year before and the year following the addition of a full-time intensivist to its medical staff as director of critical care.

In July 1993, the hospital hired an internist-intensivist as full-time medical ICU director. His primary role was to provide education and guidance for the residents who managed all patients in that unit; previously, instruction in critical care had consisted of a syllabus and ad hoc teaching by the hospitals pulmonologists. The hospital had previously established guidelines for ICU admission and discharge, and these were retained during the new intensivists first year. For the year prior to and the year following the arrival of the intensivist, data were requested from the hospital's automated data-processing system for all patients admitted to the unit on the MICU service, and the results of this process form the basis for the present report.

Data available on the patients included coded diagnoses, length of hospital and ICU stay, and mortality; prospectively recorded APACHE II scores from the second study year were compared to retrospectively calculated scores on a randomly selected 20% sample of patients from the first year. Residents were given a 40-item exam from a published text (Hall JB, et al. Principles of Critical Care: Pretest Self-Assessment and Review. New York: McGraw-Hill; 1991) in July 1993 and again in June 1994; the intensivist's teaching during that year was from the same authors' companion study guide (Hall JB, et al. Principles of Critical Care: Companion Handbook. New York: McGraw-Hill; 1993).

The 459 patients admitted during the first year (pre-intensivist) had similar diagnoses and APACHE II scores to the 471 patients admitted during the second year. All-cause mortality for patients admitted to the MICU was 34% in the first year and 25% in the second year (P = 0.002); deaths in the MICU were 21% and 15%, respectively (P = 0.02). Mean durations of MICU stay (5.0 vs 3.9 days) and of hospital stay (22.6 vs 17.7 days) both fell significantly in the second year (P < 0.05 for both). The 22 residen's scores on the standardized examination increased from a mean of 54% correct to a mean of 67% correct (an increase of 24%; P < 0.01) over the one-year study period. Although Manthous and colleagues acknowledge that their study lacked the ability to ascribe cause and effect to the observed changes in patient outcomes and resident inservice scores, they conclude by stating that "the aggregate of available data may justify the inclusion of hospital-based intensivists in community hospitals."

COMMENT BY DAVID J. PIERSON, MD

It would be most disappointing to everyone involved if adding a full-time teaching intensivist to the staff of an eight-bed ICU, in which there had previously been none, had no identifiable effect. It is therefore heartening that Manthous et al were able to demonstrate decreased mortality and length-of-stay among patients admitted to the unit during the intensivist's first year, and also an increase in the resident's scores on a critical care inservice exam. Care should be taken, however, not to make more of the reported results than is justified by the study design and other circumstances of this report.

Although the demographics and diagnostic categories of the patients admitted to the unit before and after the arrival of the intensivist were similar, and the APACHE II scores of a subset of them were comparable, we cannot be sure that some factor other than better management could not account for the observed differences. Manthous et al state that the same general admission and discharge guidelines were in force during both years; however, the new intensivist was responsible for their implementation during the second but not the first study year, raising the possibility of selection bias, conscious or otherwise.

The residen'ts scores on taking the examination a second time after one year showed that their knowledge had increased. However, ascribing the improvement to the presence of the intensivist per se is problematic. In addition to having the benefit of a full years teaching (2-3 hours didactic instruction per day plus twice-monthly seminars), using the same syllabus on which the exam was based, the residents also had an additional year of experience in caring for ICU patients. Whether the same material presented by a physician with different credentials, or simply the additional clinical experience supplemented by independent study, would have produced the same results, cannot be determined from the data presented.

Does this report demonstrate that the addition of a full-time intensivist to a given community hospital's staff would decrease mortality, reduce length-of stay, and improve clinical teaching? No. What it does show, however, is that, at the hospital in question, care in the MICU improved by the measures used during the same period that an intensivist was added to the staff. Possible explanations for the observed changes are numerous, but one of them is that the knowledge, skills, and enthusiasm of the new staff member truly did lead to improved patient care and better house staff teaching.

Care maps and patient management protocols are supposed to even out clinical decision making and render the individual carrying out the management less important. These assumptions have not been directly tested. It would be interesting to compare patient outcomes in a closed unit staffed by trained intensivists to those in an open unit in which adherence to established clinical practice guidelines was enforced. Unfortunately, a randomized control trial under such circumstances would be practically impossible to perform. In the meantime, this provocative report lends support to the argument that a dedicated, well-trained intensivist can have a positive impact on patient care and clinical teaching in an ICU.