Unexpected deaths and cardiac arrest (CA) episodes occurring in hospitals1,2 are often preceded by alarm signs.3,4 Likewise, unprogrammed admissions of hospitalized patients to the Critical Care Department (CCD) are usually preceded by alterations in vital signs that have not been detected and/or registered, and which have not received medical intervention.5,6

It is clear that not all the technology of a CCD can be found in conventional hospital wards, but Medical Information Systems (MIS) currently allow the early detection of certain parameters that identify patients at risk of suffering serious adverse events.

The continuous work burden of Emergency Departments and the increased complexity of patients admitted to hospital are causing a rise in the proportion of admissions of patients at a high risk of developing serious disease susceptible to ulterior admission to the CCD.

Correctly understanding this problem allows us to consider possibilities for interprofessional and multidisciplinary collaboration not only with the medical specialties but also with other professional areas such as communications engineering. In fact, collaboration and coordinated clinical management constitute the basis for guaranteeing care continuity in a patient-centered structure or system. The development of computer programs offers a fast and easy way to detect variables indicative of physiological and laboratory test instability, with a view to detecting possible patients at risk.

The use of Medical Information Systems (MIS) in hospitals aims to improve the outcomes of the healthcare process, placing emphasis on safety and quality. The adoption of tools that allow us to manage the care of such patients can lead to a more efficient distribution of the existing healthcare resources, since they provide us in time with the information needed for optimizing decision making.7

In Son Llàtzer Hospital (Palma de Mallorca, Spain) we attempt to secure the integration of communication technologies for use in healthcare activities, with a view to increasing the quality of patient care and thereby improving survival and lowering the costs.

The main objective of the present study was to describe the implementation of an automatic MIS for detecting rapid care alerts (RCA) in hospitalized patients. Likewise, an analysis was made of the characteristics of the patients with CA and unexpected death.

A total of 17,991 patients were admitted by the Emergency Department during the period between 1 January 2010 and 31 December 2010. A total of 507 patients died in hospital (2.8%).

An analysis was made of the alerts transmitted online. During the study period, the MIS registered 9647 rapid care alerts (RCA) in a total of 4020 patients–this being equivalent to 26.4 alerts a day. Of these alert episodes, 1100 (11.4%) proved to be false (e.g., wrong automated arterial pressure recordings, or wrong respiratory frequency or oxygen saturation measurements). A final total of 8547 alerts were thus reviewed. Sixty-five patients (8 patients/1000 interventions) presented episodes with criteria of RCA requiring early intervention by the CCD either because adequate treatment had not been provided or because detection by the intensivist preceded action on the part of the ward physician (Table 2).

During the study period we registered 45 cardiac arrests (2.5 CA/1000 admissions); of these cases, 29 (64%) presented prior criteria of RCA that were either not treated or were not evaluated adequately by the CCD or ward physician. Of the patients who survived CA (12 cases), 58.3% (7 patients) had not received prior care referred to the observed RCA. Table 3 describes the demographic characteristics, establishing comparisons according to survival. In the multivariate analysis of the predictors of survival at hospital discharge among the patients who had suffered CA, including in the analysis those factors of statistical significance and of clinical relevance, we identified monitorization (odds ratio [OR]: 0.226; 95% confidence interval [95%CI]: 0.050–0.903; p<0.045) and cardiac rhythm amenable to defibrillation (OR: 0.5; 95%CI 0.009–0.134; p<0.002) as independent predictors of a good prognosis.

A total of 507 deaths were evaluated. We registered 35 unexpected deaths (6.9% of the total mortality), of which 32 occurred in the ward (UWD) and three in the CCD. Of the UWD, 24 (75%) were preceded by criteria of RCA not subjected to treatment or evaluated by either the CCD or the ward physician; two were preceded by criteria of RCA subjected to treatment; and 6 presented no criteria of RCA. A total of 34.4% of the deaths (11 patients) were evaluated as avoidable deaths. In 9 patients with criteria of RCA not subjected to treatment (13.8%) and in 10 unexpected deaths (31.3%), the nurse had expressed concern to the ward physician regarding the evolution of the patient.

On comparing the two semesters of the year regarding the incidence of mortality and CA since the start of the project, a significant reduction in unexpected deaths was noted (OR: 0.404; 95%CI 0.194–0.842; p=0.012) in the second semester (Table 4).

In the present study, a large percentage of cardiac arrests and unexpected deaths were preceded by criteria of RCA not subjected to treatment. These results coincide with those obtained by other authors.9–12A priori, it is logical to assume that an intervention during this period of time can improve the patient prognosis. However, different studies evaluating rapid response systems have yielded contradictory results that may be a consequence of the application of different methodological designs.13–15

There are not data supporting the idea that teams led by physicians are more effective than those supervised by nurses. Such systems make use of apparently very costly resources.14,15 In a special article published in 2010 in this journal, Murias et al.16 conducted an exhaustive review of the advantages of Telemedicine, regarding it as a complementing tool capable of improving the medical care offered by intensivists. Recently, Gómez-Tello et al.7 underscored the importance of the introduction of Medical Information Systems in the CCD. Our study is new in that it involves utilization of the MIS for rapidly responding to alarm situations. In Son Llàtzer Hospital, implementation of the MIS for the detection of RCA in hospitalized patients has given rise to a considerable cultural change as regards patient safety, thanks to the training program for hospital nurses and physicians (including intensivists). Launching of the project did not imply increased resource utilization for the hospital, and made it possible to increase the activity of the CCD, due to the large number of patients detected by the system on a daily basis, which otherwise would not have been systematically reviewed. This increase in intensive care activity (with the assessment of 25 alert activations a day) has not required the assignment of added material or human resources giving rise to associated cost increments.

There is scientific evidence that early identification and intervention in clinical situations characterized by physiological instability can prevent cardiac arrest, deaths and readmissions to the SMI.17,18 During the year of our study we recorded no decrease in the incidence of cardiac arrest. Monitorization and rhythms amenable to defibrillation were associated to an improved prognosis. These results are similar to those obtained by other investigators.19 A large percentage of patients with criteria of RCA not subjected to treatment or who were not adequately evaluated, and a large proportion of unexpected deaths, occurred during the first and second nursing shifts. These data suggest that we must intensify and improve our efforts to secure early care for such patients, offering monitorization and admission to the CCD where necessary, and must improve nursing and medical personnel training in the detection of high risk patients in the ward.

In the present study, the intensivists considered it appropriate to decide admission for treatment or monitorization in the case of those patients with criteria of RCA requiring early intervention on the part of the CCD. This observation points to the importance of the adoption of new strategies by the Extended Intensive Care Services (EICS),20,21 with the purpose of improving morbidity–mortality through the early detection of high risk patients outside the ICU. In our study, the most frequently detected criteria of RCA leading to admission to the CCD were respiratory failure and diminished consciousness. Other authors have reported similar findings.12,19,22,23 An important observation is the fact that nursing alerts represent an important percentage of the alerts referred to patients with criteria of RCA not subjected to treatment and unexpected deaths. Recently, the studies of Mitchel et al.24 and of the investigators of the MERIT and ANZICS studies25 concluded that it is of benefit to establish regular training programs for physicians and nurses, with a view to detecting these patients at risk and to secure early action. In our hospital, and in reference to all patients without life support limitations who require high oxygen concentrations (FiO2>50%) or suffer diminished consciousness, we have advised the physicians to examine each case jointly with the supervising physician, and to evaluate the advisability of early admission to the ICU. Future investigations in our field should attempt to unify the alert criteria and response times of the intensivists in an adequate cost-effectiveness scenario.

There is sufficient evidence that early intervention in patients with sepsis avoids progression to severe sepsis and septic shock.26,27 The clinical data suggest that those patients who develop sepsis in hospital wards may suffer a delay in treatment (fluid therapy, vasoactive drugs, antibiotic treatment, delayed admission to the CCD).28 In our study, sepsis was the etiology most commonly detected by the MIS. Although not the objective of our study, early transmission of the criteria of RCA through the MIS, and rapid action on the part of the CCD, could contribute to the early diagnosis and treatment of these patients, as evidenced by other authors.29

Among the main limitations of our study, mention must be made of the lack of data referred to the period prior to introduction of the MIS. The availability of such information, and evaluation of the patients treated by the ward physicians, would have allowed us to assess the impact of the intervention. Mention also must be made of the lack of measurement of the response times on the part of the intensivists. Other limitations in turn may be related to divulgation of the project among all the hospital nursing and medical personnel members, which in some cases undergo constant renewal. A study involving a larger number of patients is needed, with prolongation of the study period, in order to corroborate and consolidate our conclusions.

The present study shows that the MIS can be a promising complementary tool in the daily work of the CCD that may contribute to improve the prognosis of patients admitted to hospital wards.

The authors declare that they have no conflicts of interest.