Computerized decision support for mechanical ventilation of trauma induced ARDS: results of a randomized clinical trial.

BACKGROUND: Variability and logistic complexity of mechanical ventilatory support of acute respiratory distress syndrome, and need to standardize care among all clinicians and patients, led University of Utah/LDS Hospital physicians, nurses, and engineers to develop a comprehensive computerized protocol. This bedside decision support system was the basis of a multicenter clinical trial (1993-1998) that showed ability to export a computerized protocol to other sites and improved efficacy with computer- versus physician-directed ventilatory support. The Memorial Hermann Hospital Shock Trauma intensive care unit (ICU) (Houston, TX; a Level I trauma center and teaching affiliate of The University of Texas Houston Medical School) served as one of the 10 trial sites and recruited two thirds of the trauma patients. Results from the trauma patient subgroup at this site are reported to answer three questions: Can a computerized protocol be successfully exported to a trauma ICU? Was ventilator management different between study groups? Was patient outcome affected? METHODS: Sixty-seven trauma patients were randomized at the Memorial Hermann Shock Trauma ICU site. "Protocol" assigned patients had ventilatory support directed by the bedside respiratory therapist using the computerized protocol. "Nonprotocol" patients were managed by physician orders. RESULTS: Of the 67 trauma patients randomized, 33 were protocol (age 40 +/- 3; Injury Severity Score [ISS] 26 +/- 3; 73% blunt) and 34 were nonprotocol (age 38 +/- 2; ISS 25 +/- 2; 76% blunt). For the protocol group, the computerized protocol was used 96% of the time of ventilatory support and 95% of computer-generated instructions were followed by the bedside respiratory therapist. Outcome measures (i.e., survival, ICU length of stay, morbidity, and barotrauma) were not significantly different between groups. Fio2 > or = 0.6 and Pplateau > or = 35 cm H2O exposures were less for the protocol group. CONCLUSION: A computerized protocol for bedside decision support was successfully exported to a trauma center, and effectively standardized mechanical ventilatory support of trauma-induced acute respiratory distress syndrome without adverse effect on patient outcome.

Efficacy of computerized decision support for mechanical ventilation: results of a prospective multi-center randomized trial.

200 adult respiratory distress syndrome patients were included in a prospective multicenter randomized trial to determine the efficacy of computerized decision support. The study was done in 10 medical centers across the United States. There was no significant difference in survival between the two treatment groups (mean 2 = 0.49 p = 0.49) or in ICU length of stay between the two treatment groups when controlling for survival (F(1df) = 0.88, p = 0.37.) There was a significant reduction in morbidity as measured by multi-organ dysfunction score in the protocol group (F(1df) = 4.1, p = 0.04) as well as significantly lower incidence and severity of overdistension lung injury (F(1df) = 45.2, p < 0.001). We rejected the null hypothesis. Efficacy was best for the protocol group. Protocols were used for 32,055 hours (15 staff person years, 3.7 patient years or 1335 patient days). Protocols were active 96% of the time. 38,546 instructions were generated. 94% were followed. This study indicates that care using a computerized decision support system for ventilator management can be effectively transferred to many different clinical settings and significantly improve patient morbidity.

Clinical informatics: 2000 and beyond.

Healthcare has begun to flounder in the mounting flood of data available from automated monitoring equipment, microprocessor controlled life-support equipment, such as ventilators, ever more sophisticated laboratory tests, and the myriad of minor technological wonders that every hospital and clinic seem to collect. It is no longer enough to merely display the data in a large spreadsheet or on a complex, colorful time-sequence graph. The next generation of healthcare information systems must help the clinician to assimilate the myriad of data and to make fast and effective decisions. The following is a list of features that the next generation of computer systems will have to include if they are to have a significant impact on the quality of patient care: data acquisition, data storage, information display, data processing, and decision support. By automating or streamlining repetitive or complex tasks, correlating and presenting complex and potentially confusing data, and tracking patient outcomes, the computer can augment clinicians' skills to improve patient care.

Testing and validation of computerized decision support systems.

Systematic, through testing of decision support systems (DSSs) prior to release to general users is a critical aspect of high quality software design. Omission of this step may lead to the dangerous, and potentially fatal, condition of relying on a system with outputs of uncertain quality. Thorough testing requires a great deal of effort and is a difficult job because tools necessary to facilitate testing are not well developed. Testing is a job ill-suited to humans because it requires tireless attention to a large number of details. For these reasons, the majority of DSSs available are probably not well tested prior to release. We have successfully implemented a software design and testing plan which has helped us meet our goal of continuously improving the quality of our DSS software prior to release. While requiring large amounts of effort, we feel that the process of documenting and standardizing our testing methods are important steps toward meeting recognized national and international quality standards. Our testing methodology includes both functional and structural testing and requires input from all levels of development. Our system does not focus solely on meeting design requirements but also addresses the robustness of the system and the completeness of testing.

Medical informatics academia and industry: a symbiotic relationship that may assure survival of both through health care reform.

There are often clear lines drawn identifying the demilitarized zone between medical informatics academics and industry. Academics were "pure" intellectuals sequestered in ivory towers that effectively shielded them from the realities of the world. Industry has historically focused on creating effective products that produce financial return to the corporation. Both the paradigms of academia and industry are quickly becoming dinosaurs in the era of health care reform where both medical informatics academia and industry are under increasing pressure to develop and prove that medical informatics has a positive impact on health care both in terms of the quality of care as well as cost. Unfortunately, neither academia or industry alone are going to be able to successfully complete this task. The purpose of this paper is to describe such a collaborative effort that has produced a computerized decision support system for the management of mechanical ventilation in patients with the Adult Respiratory Distress Syndrome (ARDS) that is now installed and supported on three different commercial CIS platforms. This collaborative effort has allowed us to successfully mount a large multi-center clinical trial designed to determine efficacy.

A successful protocol for the use of pulse oximetry to classify arterial oxygenation into four fuzzy categories.

Pulse oximetry is widely used in critical care medicine to noninvasively estimate arterial hemoglobin oxygen saturation. Despite the obvious benefits of using pulse oximetry to detect life threatening desaturations, it is unknown how well pulse oximetry is able to predict the finer graduations of arterial oxygenation needed for clinical decision making. A computerized protocol was developed for the use of pulse oximetry to classify arterial oxygenation into four fuzzy categories and tested in a prospective clinical trial which compared the oxygenation category assigned by the protocol to one assigned by a respiratory therapist. In 3,742 classifications from 15 patients over a seven month period, the protocol showed 96% agreement with the therapists in the direction of therapy and 75% agreement with the oxygenation classes assigned by the therapists.

A model-based simulator for testing rule-based decision support systems for mechanical ventilation of ARDS patients.

A model-based simulator was developed for testing rule-based decision support systems that manages ventilator therapy of patients with the Adult Respiratory Distress Syndrome (ARDS). The simulator is based on a multi-compartment model of the human body and mathematical models of the gas exchange abnormalities associated with ARDS. Initial testing of this system indicates that model-based simulators are a viable tool for testing rule-based expert systems used in health-care.