Does intermediate care improve patient outcomes or reduce costs?

It is well recognized that, although numbers of ICU beds are limited, many patients
are admitted to the ICU who do not really need full ‘intensive care’ 1]. Proponents of the IMCU suggest that such units could, therefore, be used to free
up ICU beds for patients most in need of full ICU facilities and expertise. In one
large multicenter study in the US, 20% of all ICU admissions were considered ‘low
severity’ and potentially ‘clinically unnecessary or inappropriate’ 4]. Other more recent studies have reported that 20 to 30% of all ICU patients are admitted
for less than 24 hours for routine surveillance/monitoring 5],6]; this percentage may be considerably higher on surgical ICUs than on medical or mixed
units. Such patients are at low risk of developing complications and hence unlikely
to require invasive therapy during their ICU stay 1],4],7],8], making them potentially appropriate candidates for IMCUs with higher staff:patient
ratios and more specialized equipment, notably for monitoring, than on the general
ward but less than on the ICU, theoretically making these units a more cost-effective
option for patients not needing full ICU facilities.

As well as direct admissions to the IMCU for low-risk patients requiring short-term
intensive monitoring, these units are also used as a ‘step-down’ facility for patients
who no longer need intensive therapy, but are perhaps not ready to return to the general
ward 9]. Without the presence of an IMCU, such patients would, perhaps unnecessarily, be
kept on the ICU, thus occupying a bed and preventing its use by a patient who may
benefit more. Alternatively, they may be discharged to the ward early, especially
in small ICUs with limited bed availability, carrying the risk of increased readmission
rates, which have been associated with increased mortality rates 10]. IMCUs are also used as ‘step-up’ units, admitting patients who require more intensive
management than is available on the general floor but do not yet need full intensive
care, in the hope that this move would prevent a later ICU admission.

Effects on outcomes?

Despite the theoretical reasons to support the use of IMCUs and the many such units
now in existence, there are relatively few published data directly assessing their
value and the studies that have been published are mostly retrospective in nature.
In an early study, Franklin and colleagues 11] reported that the opening of an IMCU was associated with a decrease in mortality
rates across the medical service, largely because of a decrease in the numbers of
deaths on general medical wards. The authors suggested that this was in part because
more ICU beds had been made available as lower-risk intensive monitoring patients
were admitted to the IMCU so that high-risk patients who would otherwise have been
managed on the general ward because of bed shortages were being more appropriately
managed on the ICU. Beck and colleagues 12] reported that patients with high severity scores who were discharged to hospital
wards had a higher risk of in-hospital death compared with patients discharged to
a high-dependency unit, suggesting that the IMCU helped prevent ‘premature’ discharges
to the ward. However, Campbell and colleagues 13] reported that discharge to a high-dependency unit was an independent risk factor
for early ICU re-admission, suggesting that these patients had been discharged prematurely;
importantly, as mentioned earlier, ICU readmission is associated with increased mortality
rates 10]. In a before-after study in a surgical ICU, opening an IMCU was associated with an
increase in the overall severity of illness of the patients admitted to the ICU, but
without increased mortality 14]. The creation of a step-up ‘subintensive care unit’ within an acute care for the
elderly department was associated with improved patient outcomes compared with an
historical cohort of patients with similar disease severity admitted to the general
acute care ward 15]. However, in a retrospective cohort study of data from 28 ICUs in the Netherlands,
the presence of an IMCU was associated with higher in-hospital mortality than if no
such unit was available 16]. In a 14-year observational study, Teli and colleagues 17] reported a significant decrease in ICU admissions for routine vascular surgical patients
after creation of a high-dependency unit. Finally, Ranzani and colleagues 18] performed a propensity-matched analysis of 160 patients discharged from the ICU to
an IMCU over a 5-year period in a teaching hospital in Brazil. Ninety-day mortality
rates and unplanned ICU readmissions were similar in patients discharged to the IMCU
and those discharged to the general ward.

Few studies have prospectively collected mortality data for comparisons of intermediate
and intensive care. In one study, Bellomo and colleagues 19] found that the opening of a four-bed high-dependency unit in their department had
no effect on mortality rates or hospital length of stay and was associated with an
increase in the number of patients requiring re-intubation. In a prospective before-after
study, Solberg and colleagues 20] reported that introduction of an IMCU, for use as a step-down unit from the ICU,
was associated with improved ICU utilization and more appropriate use of ICU beds,
such that a sicker population of patients, with higher mortality rates, was admitted
to the ICU during the IMCU period. However, there were no differences in the numbers
of ICU referrals, readmissions to the ICU, or ICU length of stay before and after
the IMCU was opened.

In a recent multinational observational study performed in 167 ICUs from 17 European
countries, the presence of an IMCU in the hospital was associated with a significantly
reduced adjusted hospital mortality for adults admitted to the ICU (odds ratio 0.63,
95% confidence interval 0.45 to 0.88, P?=?0.007), notably those admitted for full intensive care therapy rather than just
for monitoring 21]. However, in this study less than 25% of the ICU patients actually received treatment
in an IMCU either before or after their ICU admission. Hence, if IMCUs do reduce mortality,
they must do so by somehow improving the performance of the ICU in caring for the
sickest patients who never receive care in the IMCU. The mechanism for this effect
on care outside the IMCU is not clarified in the study. In addition, the apparent
‘harm’ of not having an IMCU was derived from data on a minority (16%) of the participating
ICUs that were smaller and less likely to be academic centers than ICUs in hospitals
with IMCUs.

Effects on costs?

The theoretical increased efficiency associated with availability of an IMCU and the
reduced staff:patient ratios on such units have often been promoted as likely to be
associated with reduced costs. In an early systematic review of the literature, only
three studies were identified that had conducted economic analyses and they all concerned
respiratory IMCUs. The authors concluded that there were insufficient data to support
‘the viewpoint that the addition of a [transitional care unit] to an institution with
ICU and general ward beds is cost-effective’ 22]. In a more recent prospective study, Bertolini and colleagues 23] reported that, for patients with exacerbation of chronic obstructive pulmonary disease,
the total cost per patient was lower in a respiratory IMCU than on an ICU. However,
Solberg and colleagues 24] noted that, although the costs of an IMCU day were less than those of an ICU day,
total hospital costs per patient increased significantly after introduction of an
IMCU. The authors suggested that this was due to a greater severity of illness in
the patients admitted after introduction of the IMCU and hence longer ICU stays and
was not related to the introduction of the IMCU per se. Moreover, they suggested that their IMCU may have been running at less than optimal
efficiency, because only 85% of beds were occupied during the study period as a result
of staffing constraints. In a Canadian hospital in which the IMCU was closed for budgetary
reasons, Byrick and colleagues 25] reported an increase in the numbers of admissions to the ICU with low severity of
illness and with short stays and an increase in the numbers of patients discharged
with low nurse workload requirements; these findings were interpreted as an increase
in inefficient use of staff and resources and the IMCU was reopened.

Assessing cost differences occurring as a result of introducing IMCU beds can be difficult,
with the existing literature using the typical bed-day cost algorithm, which does
not factor in the reduced acuity and cost of patients as they move through the ICU
and hospital, such that an ICU day cost before they move to the ward is very similar
to the first day on the ward cost 26]. The cost savings achieved by shifting patients from one bed to another (from an
ICU to a long-term acute care facility or an ICU to the ward a few days earlier) need
to be calculated as the total cost of care, not just the ICU costs.