Is evaluating ICU efficiency an important step in the management of these units?

The monitoring and analysis of efficiency indicators allow the evaluation, over time, of an ICU and its comparison with other units, besides helping to identify modifiable factors for quality improvement, cost reduction and efficiency gains.

Among the main indicators are clinical outcomes, especially mortality rates and ICU length of stay. For the correct interpretation of these parameters, adjusted risk assessments using predictive scores and models are essential.

Especially for patients in the ICU, safely reducing hospitalization time is a strategy that has been widely recommended, in a scenario where there is increasing pressure for increased efficiency in the hospital sector. It is essential to estimate the length of stay reliably to implement adequately such a strategy.

Estimating ICU length of stay can be useful for a number of measures that lead to unit efficiency, including:

• Identify high-risk long-stay patients and implement specific care plans for them;
• Estimate and plan the capacity of attendance and ICU patient flow (vacancies for elective surgeries and transfers of patients);
• Allow benchmarking through risk-adjusted measures, thus comparing, in a fair way, the average length of stay amongst different ICUs.

The Epimed Performance module of predictive analysis estimates the length of stay and the risk of long-stay of patients hospitalized in ICU, providing relevant information for decision making and institutional, strategic and operational benefits.

The current definitions of long-stay vary (7,10 or 14 days) according to the profile of the units and the patients. However, we believe that this logic does not lend itself to correctly evaluating what is, in fact, a long-stay, since the main diagnosis of the patient influences substantially the definition of what would be adequate or normal.

As an example, we can imagine that an 8-day ICU stay is long for a post-operative patient with myocardial revascularization surgery (MRS) whereas the mean is 3 days but not for a patient with community-acquired pneumonia (where the average is 8 days).

Thus, from a database of more than 1.4 million patients, we used as a long-stay definition the duration that is above the 90th percentile. With this, returning to the previous example, a patient in the postoperative period of MRS would have a long-stay when it exceeded 6 days in the ICU, while severe community pneumonia exceeded 15 days.

The system can be used via the web or on mobile devices, facilitating the implementation at the bedside. In general, the unit interface (bed panel) and the individual data of each patient can and should be used as guides and qualifiers for the discussion of cases in the multi-professional round daily.

The unit and hospital reports, in turn, have a more advisory role to be used in management, performance evaluation, and periodic planning.

When obtaining the early identification of patients with moderate or high risk of a long stay in the ICU, a series of measures can be taken, among them we highlight:

• Maximum implementation of prevention protocols that can reduce hospitalization time (e.g., superficial sedation, early mobilization, prevention of infections)
• Communication with family, assistant staff, hospital management, discussion with the operator about hospitalization expectations
• Preparation of care transition (e.g., semi-intensive unit, the transition to rehabilitation or home-care)

Exceeding up to 10-20% can be considered reasonable considering that we speak of estimates. However, if the patient significantly exceeds, it is an opportunity for a more detailed assessment to identify the possible causes. It is common in these cases to identify adverse events (many of them preventable), lack of adherence to best practices (e.g., deep sedation, inadequate thromboprophylaxis), or non-medical (social/economic) situations that not only explain the discrepancy but may also serve for improvement in subsequent patients or to alert to the need for team trainings, educational measures or protocol implementation/review.

Another important aspect here refers to logistical issues (no vacant bed or absence of semi-intensive unit beds). Such factors, when identified (along with the metric "mean time of hospital discharge decision till the effective decision time" generated by the Epimed ICU Monitor system) can assist in the institution planning for the required flow and dimensioning of these beds.

Today, most ICUs have difficulty implementing all available care protocols. The identification of high-risk patients can be a way to optimize the implementation of protocols of high complexity or cost, such as early mobilization, subglottic aspiration, among others.

In addition, although not a goal but an estimate, the expected duration of hospitalization may be an additional factor for the ongoing re-evaluation of discharge viability as well as for early discharge planning (hospital accreditation goal). The use of the tool in mobile devices (tablets and smartphones) allows the introduction of this element in bedside discussions and in the multi-professional round.

When identifying high-risk patients or with high hospital stay lengths, managers can optimize resources and determine that these are the priority patients to be evaluated.

In fact, when we work the extremes (that is, prioritize each day that more qualified professionals and seniors immediately evaluate patients whose estimate of length of stay is too short or long), we have a greater opportunity to optimize resources by releasing more quickly ICU of those patients at low risk, and making the interventions more precocious in the more complex patients.

Early identification of high-risk of long-stay patients can provide concrete support so that as soon as these patients are stable, they begin rehabilitation processes in the hospital or in post-acute or transitional care hospitals.

It may also be decisive in the earlier prioritization of the use of a semi-intensive unit vacancy or even a request for home hospitalization.

The ICU Length of Stay Estimate Plug-in can provide concrete and less subjective elements that can improve the prognostic predictive capacity of ICU professionals. With these elements you can improve the dialogue with patients and their families and thus adjust their expectations.

Today, discharge planning is part of the goals proposed by international accreditation models. By estimating the length of stay and the risk of long-stay, the ICU Stay Time Estimate Plug-in can be a crucial element in meeting this goal currently required by JCI and international accreditations.