Evaluation of compliance with a computerized protocol: weaning from mechanical ventilator support using pressure support

Artificial intelligence-based clinical decision support in pediatrics

Machine learning models may be integrated into clinical decision support (CDS) systems to identify children at risk of specific diagnoses or clinical deterioration to provide evidence-based recommendations. This use of artificial intelligence models in clinical decision support (AI-CDS) may have several advantages over traditional "rule-based" CDS models in pediatric care through increased model accuracy, with fewer false alerts and missed patients. AI-CDS tools must be appropriately developed, provide insight into the rationale behind decisions, be seamlessly integrated into care pathways, be intuitive to use, answer clinically relevant questions, respect the content expertise of the healthcare provider, and be scientifically sound. While numerous machine learning models have been reported in pediatric care, their integration into AI-CDS remains incompletely realized to date. Important challenges in the application of AI models in pediatric care include the relatively lower rates of clinically significant outcomes compared to adults, and the lack of sufficiently large datasets available necessary for the development of machine learning models. In this review article, we summarize key concepts related to AI-CDS, its current application to pediatric care, and its potential benefits and risks. IMPACT: The performance of clinical decision support may be enhanced by the utilization of machine learning-based algorithms to improve the predictive performance of underlying models. Artificial intelligence-based clinical decision support (AI-CDS) uses models that are experientially improved through training and are particularly well suited toward high-dimensional data. The application of AI-CDS toward pediatric care remains limited currently but represents an important area of future research.

Performance of an Electronic Decision Support System as a Therapeutic Intervention During a Multicenter PICU Clinical Trial: Heart and Lung Failure-Pediatric Insulin Titration trial (HALF-PINT)

Background

The use of electronic clinical decision support (CDS) systems for pediatric critical care trials is rare. We sought to describe in detail the use of a CDS tool (Children’s Hospital Euglycemia for Kids Spreadsheet [CHECKS]), for the management of hyperglycemia during the 32 multicenter Heart And Lung Failure-Pediatric Insulin Titration trial.

Research Question

In critically ill pediatric patients who were treated with CHECKS, how was user compliance associated with outcomes; and what patient and clinician factors might account for the observed differences in CHECKS compliance?

Study Design and Methods

During an observational retrospective study of compliance with a CDS tool used during a prospective randomized controlled trial, we compared patients with high and low CHECKS compliance. We investigated the association between compliance and blood glucose metrics. We describe CHECKS and use a computer interface analysis framework (the user, function, representation, and task analysis framework) to categorize user interactions. We discuss implications for future randomized controlled trials.

Results

Over a 4.5-year period, 658 of 698 children were treated with the CHECKS protocol for ≥24 hours with a median of 119 recommendations per patient. Compliance per patient was high (median, 99.5%), with only 30 patients having low compliance (<90%). Patients with low compliance were from 16 of 32 sites, younger (P = .02), and less likely to be on inotropic support (P = .04). They were more likely to be have been assigned randomly to the lower blood glucose target (80% vs 48%; P < .001) and to have spent a shorter time (53% vs 75%; P < .001) at the blood glucose target. Overrides (classified by the user, function, representation, and task analysis framework), were largely (89%) due to the user with patient factors contributing 29% of the time.

Interpretation

The use of CHECKS for the Heart And Lung Failure-Pediatric Insulin Titration trial resulted in a highly reproducible and explicit method for the management of hyperglycemia in critically ill children across varied environments. CDS systems represent an important mechanism for conducting explicit complex pediatric critical care trials.

Clinical Trial Registration

ClinicalTrials.gov Identifier: NCT01565941, registered March 29 2012; https://clinicaltrials.gov/ct2/show/NCT01565941?term=HALF-PINT&draw=2&rank=1

Assessing respiratory mechanics using pressure reconstruction method in mechanically ventilated spontaneous breathing patient

BACKGROUND

Respiratory system modelling can aid clinical decision making during mechanical ventilation (MV) in intensive care. However, spontaneous breathing (SB) efforts can produce entrained "M-wave" airway pressure waveforms that inhibit identification of accurate values for respiratory system elastance and airway resistance. A pressure wave reconstruction method is proposed to accurately identify respiratory mechanics, assess the level of SB effort, and quantify the incidence of SB effort without uncommon measuring devices or interruption to care.

METHODS

Data from 275 breaths aggregated from all mechanically ventilated patients at Christchurch Hospital were used in this study. The breath specific respiratory elastance is calculated using a time-varying elastance model. A pressure reconstruction method is proposed to reconstruct pressure waves identified as being affected by SB effort. The area under the curve of the time-varying respiratory elastance (AUC Edrs) are calculated and compared, where unreconstructed waves yield lower AUC Edrs. The difference between the reconstructed and unreconstructed pressure is denoted as a surrogate measure of SB effort.

RESULTS

The pressure reconstruction method yielded a median AUC Edrs of 19.21 [IQR: 16.30-22.47]cmH2Os/l. In contrast, the median AUC Edrs for unreconstructed M-wave data was 20.41 [IQR: 16.68-22.81]cmH2Os/l. The pressure reconstruction method had the least variability in AUC Edrs assessed by the robust coefficient of variation (RCV)=0.04 versus 0.05 for unreconstructed data. Each patient exhibited different levels of SB effort, independent from MV setting, indicating the need for non-invasive, real time assessment of SB effort.

CONCLUSION

A simple reconstruction method enables more consistent real-time estimation of the true, underlying respiratory system mechanics of a SB patient and provides the surrogate of SB effort, which may be clinically useful for clinicians in determining optimal ventilator settings to improve patient care.

Computerized Decision Support Systems for Mechanical Ventilation in Children

Mechanical ventilation is an effective treatment in the ICU but can have significant adverse effects. Approaches from adult research have been adopted in pediatric critical care despite known differences in respiratory physiology and ICU processes. There continues to be considerable variation in how ventilators are managed. Computerized decision support systems implement explicit protocols, and are designed to make mechanical ventilation management safer, more consistent, and more lung protective. Variable results and low or unknown compliance with protocols and CDSS tools have been reported. To date, there has been limited research regarding CDSS for mechanical ventilation in children.

Clinical decision support systems in the pediatric intensive care unit

OBJECTIVE

To review the use of clinical decision support systems (CDSS) available in the pediatric intensive care unit (PICU).

DATA SOURCES

Relevant English language publications indexed in Medline, as well as CDSS-related white papers and texts.

STUDY SELECTION AND DATA EXTRACTION

Studies related to CDSS were considered.

DATA SYNTHESIS

CDSS are operationally defined as computer software programs that aid healthcare providers in their clinical decision making. Once used solely for diagnostic support, many CDSS now have the ability to transform clinical practice through interactive assistance with therapeutic best practices. The recent emphasis on improving quality and patient safety through the incorporation of electronic health records as supported by Leapfrog and other agencies has encouraged advancements in the use of CDSS tools that leverage the capabilities of stand-alone electronic health records. CDSS are of particular interest in the PICU where rapid decision-making benefits from tools that can improve patient safety. CDSS have been described in the PICU with varying effects on healthcare outcomes. A growing consensus indicates that the success of such interventions depends as much or more on how they are implemented and used in such complex environments as on their programming. In the current review, the types and features of various CDSS tools and the supporting evidence are discussed. Factors such as liability, human factors engineering, alert fatigue, and audit trails are also covered.

CONCLUSION

CDSS have the potential to improve clinical practice in PICU settings. Care should be taken when selecting and implementing such systems to achieve the goal of improved clinical practice while avoiding potential adverse impacts sometimes associated with the implementation of new technologies in complex healthcare settings.

Rate of reintubation in mechanically ventilated neurosurgical and neurologic patients: evaluation of a systematic approach to weaning and extubation

OBJECTIVE

To assess whether a systematic approach to weaning and extubation (intervention) is superior to the sole physician's judgment (control) in preventing reintubation secondary to extubation failure in patients with neurologic disorders.

DESIGN

Randomized controlled trial.

SETTING

Intensive care unit of a large teaching hospital.

PATIENTS

Three hundred eighteen intubated patients who had been receiving mechanical ventilation for at least 12 hrs and were able to trigger the ventilator.

INTERVENTIONS

Patients were randomized to the intervention (n = 165) or control group (n = 153).

MEASUREMENTS

Rate of reintubation after extubation failure occurring within 48 hrs (primary end point). Duration of mechanical ventilation, length of intensive care unit stay, mortality, rate of tracheotomy (secondary end points). The perception of the research protocol by the intensive care unit staff was also assessed.

MAIN RESULTS

The rate of reintubation was lower in the intervention (5%) than in the control (12.5%) group (p = 0.047). There was no difference in any of the other outcome variables (secondary end points). Simplified Acute Physiologic Score II (adjusted odds ratio 1.042 per unit; 95% confidence interval 1.006-1.080; p = 0.022) and inclusion in the control group (adjusted odds ratio 2.393; 95% confidence interval 1.000-5.726; p = 0.05) were the only two independent predictive factors for the risk of extubation failure. The protocol was felt by the staff to determine an improvement in patients' clinical outcome, but to increase intensive care unit workload; nurses and physiotherapists considered its impact on their professional role more positively than physicians.

CONCLUSIONS

In patients with neurologic diseases, a systematic approach to weaning and extubation reduces the rate of reintubation secondary to extubation failure without affecting the duration of mechanical ventilation, and is overall positively perceived by intensive care unit professionals.

Improving Adherence to a Mechanical Ventilation Weaning Protocol for Critically Ill Adults: Outcomes After an Implementation Program

• Background Despite multiple reminders, education sessions, and multidisciplinary team involvement, adherence to an evidence-based mechanical ventilation weaning protocol had been less than 1% in a general systems intensive care unit since implementation.

• Objective To assess the effectiveness of using an implementation program, the Model for Accelerating Improvement, to improve adherence and clinical outcomes after restarting a mechanical ventilation weaning protocol in an adult general systems intensive care unit.

• Methods A prospective comparative design, before and after implementation of the Model for Accelerating Improvement, was used with a consecutive sample of 129 patients and 112 multidisciplinary team members. Clinical outcomes were rate of unsuccessful extubations, rate of ventilator-associated pneumonia, and duration of mechanical ventilation; practice outcomes were staff’s understanding of the mechanical ventilation weaning protocol, perceptions of the practice safety climate, and adherence to the weaning protocol.

• Results After the intervention, the rate of unsuccessful extubations decreased, and staff’s understanding of and adherence to the weaning protocol increased significantly. The rate of ventilator-associated pneumonia, duration of mechanical ventilation, and staff’s perceptions of the practice safety climate did not change significantly.

• Conclusion Implementing the Model for Accelerating Improvement improved understanding of and adherence to protocol-directed weaning and reduced the rate of unsuccessful extubations.

Computers in the ICU: where we started and where we are now

The first use of computers in critical care units were described in the mid 1960s. They reported the use of very large mainframe computers that filled entire rooms yet had very limited memory and processing capacities by today's standards. These were limited to only a few institutions until microprocessors were developed increasing computation speed and expanding memory capacity by many magnitudes. This allowed smaller more affordable stand alone systems to be developed and the inclusion of microprocessors into bedside devices. As the capacity expanded uses broadened. Simple results review developed into a more complete electronic medical record. Databases were created allowing population analysis for research and systems quality improvement activities. Decision support started as simple alerting of potential errors and dangers and expanded into more sophisticated clinical decision-making support. With this came problems that needed solutions. As the amount of information became overwhelming to the bedside clinician, methods to filter and display data made it more useful. Security and confidentiality became major concerns. Data input solutions had to be found including interfaces between computers, bedside devices and instruments designed to automate data input like scanners, bar coders, and other devices. The biggest issue of all however, was developing acceptance among clinicians and creating the cultural change required for successful implementation of electronic medical records. This paper will explore these issues.

A practical approach to evidence-based medicine: lessons learned from developing ventilator management protocols

Development of robust clinical protocols is a time-consuming process that requires allocation of sufficient time and resources. As outlined for the mechanical-ventilation protocol, the goals of the protocol must be clear before embarking on this journey. Critical appraisal of existing evidence is an essential first step to developing a draft protocol. The inclusion of a multidisciplinary team of clinicians, including nurses and respiratory therapists, is required to refine the logic and to ensure protocol acceptance. Extensive bedside testing of the protocol across the spectrum of patients in which it will be applied is essential for ensuring that the protocol works as intended. An effective, ongoing training program must be implemented to ensure protocol success. The final protocol should be carefully monitored in 50 to 100 patients to ensure that clinicians are complying with the protocol rules and that adverse events related to the protocol are not occurring. Finally, protocols must be tended to and updated when new evidence becomes available.

Outcomes in post-ICU mechanical ventilation: a therapist-implemented weaning protocol

BACKGROUND

Therapist-implemented protocols have been used to extubate or wean patients in the ICU setting. Barlow Respiratory Hospital (BRH) functions as a center for weaning patients from prolonged mechanical ventilation (PMV) in the post-ICU setting of a long-term acute-care (LTAC) facility. A therapist-implemented patient-specific (TIPS) weaning protocol was developed at BRH to standardize weaning from PMV.

STUDY DESIGN

Prospective cohort study with historical control.

METHODS

A weaning protocol incorporating the procedures and pace of LTAC weaning was developed using available scientific evidence and expert consensus. After training of staff, collection and analysis of pilot data, and revisions and refinement of the protocol, the TIPS protocol was implemented hospital-wide. It was monitored for outcome, variance, and respiratory care practitioner (RCP) and physician compliance.

RESULTS

Forty-six RCPs worked with eight pulmonologists treating 271 consecutive patients admitted for weaning from PMV during an 18-month period. Nineteen patients were excluded from weaning attempts by any method after initial physician evaluation. The remaining 252 patients (9,135 total ventilator days) were compared with a group of 238 patients treated by the same physicians in the 2 years before instituting protocol weaning. Median time to wean declined significantly from 29 days in historical control subjects to 17 days for TIPS protocol patients (p < 0.001). Outcomes (scored at discharge) were comparable for the two groups (TIPS group vs control group): weaned, 54.7% vs 58.4%; ventilator-dependent, 17.9% vs 10.9%; died, 27.4% vs 30.7% (p = 0.10). Variances incurred by physicians and RCPs were 324 and 136, respectively, for the 9,135 ventilator days.

CONCLUSIONS

Patients weaned from PMV using a new therapist-implemented protocol at BRH, an LTAC facility specializing in weaning, had significantly shorter time to weaning than historical control subjects, with comparable outcomes. The weaning outcome data collected after the implementation of the TIPS protocol are in fact attributable to its use, as we found a high degree of compliance with the protocol.

Enhancement of mechanical ventilation of neonates by computer technology

Mechanical ventilation is often required in the extremely small preterm infant because of transient failure of the immature control of breathing system, chest wall instability, weakness of the respiratory pump, and chronic lung disease. To better address some of these issues, computer algorithms have been developed for continuous automated control of the arterial hemoglobin oxygen saturation (by changing the fraction of inspired oxygen), of the tidal volume (by varying peak inspiratory pressure), and of minute ventilation (by changing respirator rate). Feasibility studies have shown that closed-loop computer control of physiologic target variables may stabilize oxygenation, decrease the transpulmonary pressure cost of ventilation, and expedite weaning from mechanical ventilation. In addition to synchronizing the upstroke in ventilator pressure with the onset of the spontaneous respiratory effort, computer technology may render the changeover from inspiration to expiration patient-cycled in a mode called pressure support ventilation. Proportional assist ventilation enhances ventilation in near perfect synchrony with the patient's spontaneous respiratory efforts. With this mode, the clinician selects gains for elastic and/or resistive unloading to tailor the ventilator pressure contour to the specific derangement in pulmonary mechanics (restrictive and/or obstructive). Most of these newer ventilatory strategies are as yet unproven and require randomized, controlled clinical study. Similarly, although on-line pulmonary mechanics monitoring during mechanical ventilation has become available with the adent of computer technology, its clinical utility is still not well established.