Respiratory support is required in most children in the pediatric intensive care unit. Decision-support tools (paper or electronic) have been shown to improve the quality of medical care, reduce errors, and improve outcomes. Computers can assist clinicians by standardizing descriptors and procedures, consistently performing calculations, incorporating complex rules with patient data, and capturing relevant data. This article discusses computer decision-support tools to assist clinicians in making flexible but consistent, evidence-based decisions for equivalent patient states.
Key points
- •
Despite the proven effect of decision-support tools, there is no uniformity or agreement on ventilator management decisions.
- •
Most critical care practitioners believe they are being lung protective, but it is likely that consistent, replicable decisions are not made to minimize ventilator support across the duration of mechanical ventilation for patients with lung injury.
- •
Because respiratory support is required in most children in the pediatric intensive care unit and because complications may occur with its use, it is essential to develop strategies to improve patient outcome and reduce medical errors related to mechanical ventilation.
Introduction
Pediatric and neonatal critical care has been practiced formally for more than 50 years and invasive mechanical ventilation is among the most common and dramatic intensive care unit (ICU) procedures, with about 30% (range 20%–60%) of patients in a pediatric ICU (PICU) supported by this technique. Mechanical ventilation is a life-saving method for thousands of patients who cannot breathe on their own and is administered until patients resume independent (spontaneous) breathing. In particular, patients with acute respiratory distress syndrome (ARDS) and acute lung injury, now collectively termed pediatric ARDS (PARDS), typically require days to weeks of mechanical ventilation for respiratory failure. Despite experience with mechanical ventilation, little is known about how best to ventilate patients with specific disease entities or syndromes such as PARDS. There is uncertainty about the best choice of ventilator type, mechanical ventilation mode, and the therapeutic goals of mechanical ventilation support, including best practices regarding weaning from mechanical ventilation, allowing resumption of spontaneous breathing, and removing the endotracheal tube. Although the pediatric age range is wide (newborn to 18 years), normalization of important respiratory monitoring and control parameters to body weight (eg, tidal volume and compliance), allow ventilator management approaches to be consistent across the entire pediatric age spectrum.
Thirty years ago, Pollack and colleagues reported that at a single US hospital, long-stay patients were 7% of the population but used 50% of the PICU care days and 48% of the technology resources. They also had a much higher mortality rate compared with short-stay patients. Five years later, a prospective cohort Australian study reported complications in 24% of 500 patients requiring respiratory support. Although mortality has improved markedly, these overall observations about long-stay patients persist and continue to be reported in several more recent studies. In addition, respiratory support increases the risk of complications attributable to the endotracheal tube, positive pressure ventilation, and/or sedation. A recent prospective cohort, single-center study found that nearly half of ventilated PICU patients were found to be at risk for ventilator-related volutrauma. Although the basic principles of lung protective ventilation have been embraced by pediatric intensive care physicians, there is still great variability in ventilator management. Most critical care practitioners believe they are being lung-protective, but it is likely that consistent, replicable decisions are not made to minimize ventilator support across the duration of mechanical ventilation for patients with lung injury. Because respiratory support is required in most children in PICU, and because complications may occur with its use, it is essential to develop strategies to improve patient outcome and reduce medical errors related to mechanical ventilation.
Need for Mechanical Ventilation Protocols
The observed variation in clinical practice is likely due in part to clinicians’ low adherence to guidelines, and this is compounded because many facets of caring for a mechanically ventilated patient in the PICU lack high levels of evidence, or evidence is conflicting. There is good evidence, however, that clinical decision-making with a protocol decreases practice variation between clinicians, standardizes patient care, and improves research and patient outcomes. Management of mechanical ventilation is an iterative intervention encompassing many individual decisions that must be made over the course of a patient’s treatment, usually by multiple practitioners. Replicable ventilator management decisions should help decrease practice variability and directly shorten the length of mechanical ventilation for children in PICUs. Written protocols for management of respiratory support decreased ventilator days in adults when compared with usual physician orders. Similar findings have been observed in pediatrics and neonates. The Pediatric Acute Lung Injury Consensus Conference (PALICC) recommendations support use of a goal-directed protocol to guide ventilator management.
Paper Versus Computer-Based Protocols
Decision-support tools, both paper and electronic, have been demonstrated to improve medical care, reduce errors, and improve patient outcomes. However, paper-based protocols depend on caregiver availability, are often written in broad terms so they remain dependent on clinician judgment and local context for interpretation, and are, therefore, difficult to transfer from a PICU or NICU to another ICU. Paper-based tools can be difficult to follow accurately, leading to low adherence rates. There was 30% compliance in a pediatric ventilator study with a paper protocol compared with 94% in an adult ventilator management study with a computer-based protocol and greater than 97% in a recent pediatric insulin-glucose randomized controlled trial.
Computer decision-support (CDS) tools, such as computer-based protocols, were cited by the Institute of Medicine as a method to reduce medical errors. Such tools aim to ensure replicable, evidence-based clinician decisions for equivalent patient states and to improve protocol compliance. The tools can generate explicit recommendations that can be carried out with little variability between clinicians while remaining responsive to the patient’s unique situation. The tools can assist clinicians by standardizing descriptors and procedures, by consistently performing calculations, by incorporating complex rules with patient data, and by capturing data relevant to decision-making. Computer-based protocols can contain more extensive detail than textual guidelines or paper-based flow diagrams while protecting the user from complexity and information overload.
Pediatric Versus Adult Computer-Based Protocols
Much of the research on computer protocols in the ICU has been conducted in adult ICUs. Research from adult ICU is commonly extrapolated for practice in the PICU, but PARDS has distinct epidemiology, outcomes, and likely pathophysiology compared with adult ARDS. Little is known about the use of adult-derived CDS tools in pediatric medicine. An early application, an anti-infective tool, seemed to be beneficial to children and allowed cost savings. A second application, a glucose-insulin protocol, used a common (shared) protocol that seemed to be as applicable to children as to adults. Additionally, this study demonstrated excellent protocol compliance (>97%) compared with a paper protocol and tighter adherence to the target blood glucose range. Although children are developmentally and physiologically not little adults, it seems that, at least in certain domains, CDS protocols can be used with little modification for both adults and children. Nonetheless, in general, adult protocols seem likely to need modification in the granularity (detail) of the decision rules to be usable in children, and research needs to be done to investigate this issue.
Although a computer-based mechanical ventilation protocol has been developed for adult ARDS, large differences between adult and pediatric critical care bring into question the practice of extrapolating adult evidence regarding ventilator management with computer protocols, to children with PARDS. In particular, pediatric clinicians may be more comfortable with smaller (more granular) changes of 0.05 in the fraction of inspired oxygen, rather than the 0.1 increments recommended in the adult ARDSNet protocol, as suggested by analysis of large datasets from both Vanderbilt Children’s Hospital and Children’s Hospital Los Angeles. In addition, the ranges of pH (<7.15; 7.15–<7.30; 7.30–7.45; >7.45) used in the ARDSNet protocol to guide changes in ventilator rate and tidal volume may seem too broad for many pediatric intensivists, particularly those with significant postoperative cardiac surgery background in their training. There are also differences in how predicted body weight from height or length is calculated for determination of tidal volumes in lung-protective volume-controlled modes, with pediatricians needing to deal with failure to thrive, contractures, and obesity, with only the latter being of major consideration for adult practitioners. It is probable that measurement of ulna length will prove to be the best and safest predictor of height and the subsequent percentile for age used to calculate weight and thence tidal volume over the entire pediatric age range. The adult protocol calls for recording the set volume delivered by the ventilator, whereas pediatric practice is to measure the tidal volumes directly, but this has not been standardized as either in the ventilator or at the endotracheal tube, with large differences recorded between them. Finally, adult protocol management of ARDS is based on Assist Control volume control, a mode rarely used by pediatric intensivists who seem to prefer pressure-regulated modes.
Introduction
Pediatric and neonatal critical care has been practiced formally for more than 50 years and invasive mechanical ventilation is among the most common and dramatic intensive care unit (ICU) procedures, with about 30% (range 20%–60%) of patients in a pediatric ICU (PICU) supported by this technique. Mechanical ventilation is a life-saving method for thousands of patients who cannot breathe on their own and is administered until patients resume independent (spontaneous) breathing. In particular, patients with acute respiratory distress syndrome (ARDS) and acute lung injury, now collectively termed pediatric ARDS (PARDS), typically require days to weeks of mechanical ventilation for respiratory failure. Despite experience with mechanical ventilation, little is known about how best to ventilate patients with specific disease entities or syndromes such as PARDS. There is uncertainty about the best choice of ventilator type, mechanical ventilation mode, and the therapeutic goals of mechanical ventilation support, including best practices regarding weaning from mechanical ventilation, allowing resumption of spontaneous breathing, and removing the endotracheal tube. Although the pediatric age range is wide (newborn to 18 years), normalization of important respiratory monitoring and control parameters to body weight (eg, tidal volume and compliance), allow ventilator management approaches to be consistent across the entire pediatric age spectrum.
Thirty years ago, Pollack and colleagues reported that at a single US hospital, long-stay patients were 7% of the population but used 50% of the PICU care days and 48% of the technology resources. They also had a much higher mortality rate compared with short-stay patients. Five years later, a prospective cohort Australian study reported complications in 24% of 500 patients requiring respiratory support. Although mortality has improved markedly, these overall observations about long-stay patients persist and continue to be reported in several more recent studies. In addition, respiratory support increases the risk of complications attributable to the endotracheal tube, positive pressure ventilation, and/or sedation. A recent prospective cohort, single-center study found that nearly half of ventilated PICU patients were found to be at risk for ventilator-related volutrauma. Although the basic principles of lung protective ventilation have been embraced by pediatric intensive care physicians, there is still great variability in ventilator management. Most critical care practitioners believe they are being lung-protective, but it is likely that consistent, replicable decisions are not made to minimize ventilator support across the duration of mechanical ventilation for patients with lung injury. Because respiratory support is required in most children in PICU, and because complications may occur with its use, it is essential to develop strategies to improve patient outcome and reduce medical errors related to mechanical ventilation.
Need for Mechanical Ventilation Protocols
The observed variation in clinical practice is likely due in part to clinicians’ low adherence to guidelines, and this is compounded because many facets of caring for a mechanically ventilated patient in the PICU lack high levels of evidence, or evidence is conflicting. There is good evidence, however, that clinical decision-making with a protocol decreases practice variation between clinicians, standardizes patient care, and improves research and patient outcomes. Management of mechanical ventilation is an iterative intervention encompassing many individual decisions that must be made over the course of a patient’s treatment, usually by multiple practitioners. Replicable ventilator management decisions should help decrease practice variability and directly shorten the length of mechanical ventilation for children in PICUs. Written protocols for management of respiratory support decreased ventilator days in adults when compared with usual physician orders. Similar findings have been observed in pediatrics and neonates. The Pediatric Acute Lung Injury Consensus Conference (PALICC) recommendations support use of a goal-directed protocol to guide ventilator management.
Paper Versus Computer-Based Protocols
Decision-support tools, both paper and electronic, have been demonstrated to improve medical care, reduce errors, and improve patient outcomes. However, paper-based protocols depend on caregiver availability, are often written in broad terms so they remain dependent on clinician judgment and local context for interpretation, and are, therefore, difficult to transfer from a PICU or NICU to another ICU. Paper-based tools can be difficult to follow accurately, leading to low adherence rates. There was 30% compliance in a pediatric ventilator study with a paper protocol compared with 94% in an adult ventilator management study with a computer-based protocol and greater than 97% in a recent pediatric insulin-glucose randomized controlled trial.
Computer decision-support (CDS) tools, such as computer-based protocols, were cited by the Institute of Medicine as a method to reduce medical errors. Such tools aim to ensure replicable, evidence-based clinician decisions for equivalent patient states and to improve protocol compliance. The tools can generate explicit recommendations that can be carried out with little variability between clinicians while remaining responsive to the patient’s unique situation. The tools can assist clinicians by standardizing descriptors and procedures, by consistently performing calculations, by incorporating complex rules with patient data, and by capturing data relevant to decision-making. Computer-based protocols can contain more extensive detail than textual guidelines or paper-based flow diagrams while protecting the user from complexity and information overload.
Pediatric Versus Adult Computer-Based Protocols
Much of the research on computer protocols in the ICU has been conducted in adult ICUs. Research from adult ICU is commonly extrapolated for practice in the PICU, but PARDS has distinct epidemiology, outcomes, and likely pathophysiology compared with adult ARDS. Little is known about the use of adult-derived CDS tools in pediatric medicine. An early application, an anti-infective tool, seemed to be beneficial to children and allowed cost savings. A second application, a glucose-insulin protocol, used a common (shared) protocol that seemed to be as applicable to children as to adults. Additionally, this study demonstrated excellent protocol compliance (>97%) compared with a paper protocol and tighter adherence to the target blood glucose range. Although children are developmentally and physiologically not little adults, it seems that, at least in certain domains, CDS protocols can be used with little modification for both adults and children. Nonetheless, in general, adult protocols seem likely to need modification in the granularity (detail) of the decision rules to be usable in children, and research needs to be done to investigate this issue.
Although a computer-based mechanical ventilation protocol has been developed for adult ARDS, large differences between adult and pediatric critical care bring into question the practice of extrapolating adult evidence regarding ventilator management with computer protocols, to children with PARDS. In particular, pediatric clinicians may be more comfortable with smaller (more granular) changes of 0.05 in the fraction of inspired oxygen, rather than the 0.1 increments recommended in the adult ARDSNet protocol, as suggested by analysis of large datasets from both Vanderbilt Children’s Hospital and Children’s Hospital Los Angeles. In addition, the ranges of pH (<7.15; 7.15–<7.30; 7.30–7.45; >7.45) used in the ARDSNet protocol to guide changes in ventilator rate and tidal volume may seem too broad for many pediatric intensivists, particularly those with significant postoperative cardiac surgery background in their training. There are also differences in how predicted body weight from height or length is calculated for determination of tidal volumes in lung-protective volume-controlled modes, with pediatricians needing to deal with failure to thrive, contractures, and obesity, with only the latter being of major consideration for adult practitioners. It is probable that measurement of ulna length will prove to be the best and safest predictor of height and the subsequent percentile for age used to calculate weight and thence tidal volume over the entire pediatric age range. The adult protocol calls for recording the set volume delivered by the ventilator, whereas pediatric practice is to measure the tidal volumes directly, but this has not been standardized as either in the ventilator or at the endotracheal tube, with large differences recorded between them. Finally, adult protocol management of ARDS is based on Assist Control volume control, a mode rarely used by pediatric intensivists who seem to prefer pressure-regulated modes.
Computerized protocols: development and application
Decision-support tools vary in terms of how dynamic they are, the degree of specificity of their recommendations, and the level of integration into workflow. One end of the spectrum includes general guidelines that consist of a set of broad, static recommendations. At the other end of the spectrum are computerized protocols, which are CDS tools that function as a set of standardized orders, with detailed explicit instructions based on dynamic patient-specific parameters, available at the point-of-care. The latter protocol has been called an explicit computerized protocol (ECP).
ECPs need to represent best-available evidence in a computable format. These typically contain validated mathematical formulae resulting from physiologic studies (eg, alveolar gas equations ) or rules with the formulation: if…then… (eg, if PRVC mode and pH >7.45 and tidal volume >6 mL/kg and PIP <30 cm H 2 0, then decrease tidal volume by 1 mL/kg). Other formats, such as Bayesian probability computations, may also be used to represent the knowledge and associated uncertainty. All potential paths that could be taken by the child during his or her PICU course need to be accounted for and should result in a specific, explicit instruction.
Simplicity in the rules and in the use of the application are essential. Data entered manually should be the exception to simplify the caregiver’s work, with decision support integrated into the workflow as much as possible. In the example of hemodynamic instability, data could theoretically be acquired directly from the patient’s infusion pump or the electronic medical record. To get data directly, ventilator terminology must be agreed on and matched to the context of a specific computer-based decision-support tool. Standard terminologies, such as SNOMED-CT, ICD-10, and LOINC have been mandated for use in electronic health records as part of the Affordable Care Act and are used to help integrate data across different data sources. The National Library of Medicine coordinates across these terminology services but, unfortunately, pediatric-specific terms have been generally lacking. There are several ongoing national initiatives to create sharable, comparable medical data. Moreover, ventilator mode descriptions and terminology are often highly variable between and within ventilator manufacturers.
Users must believe that they can count on the system to be available whenever they need it. The response time must be fast, data integrity must be maintained, and data redundancy minimized. It is important that systems function at several sites for a period of time so that major problems or software bugs have been eradicated, decreasing downtime and improving acceptance. It is also essential to assess the amount of training required for users to feel comfortable with the ECP. If users become frustrated with the ECP, its performance will be suboptimal as a consequence.
The instructions should be delivered at bedside computer terminals, or even in the monitor or ventilator. ECPs are initially developed in an open-loop manner, in which a recommendation is displayed on screen and an active intervention by the clinician is required to apply this recommendation. It is important to capture specific reasons for clinician refusal to follow instructions, analyze them, and determine if iterative refinement of the rules is needed both during the initial development phase and as part of ongoing surveillance of protocol effectiveness and compliance. Capturing clinician decisions to accept or decline the recommendation, and the reason for declines, is the sine qua non to develop a robust ECP.
Ultimately, after a validation phase with refinement of protocols and with clinician acceptance of safety, some or all the ECP recommendations can potentially be implemented in a closed-loop ECP, in which ventilator settings are dynamically adjusted to a patient’s condition according to the ECP recommendations without caregiver intervention but still under caregiver supervision. Displaying data specific to the decision process is strongly recommended. Closed-loop mechanical ventilation can be simple, with a single-output variable managed based on a single-input variable, such as in pressure support ventilation, in which flow is automatically adjusted to maintain pressure. Closed-loop ventilation can also be complex, with multiple inputs controlling multiple outputs. Some simple forms of closed-loop ventilation are quite common, whereas less is known about more complex forms.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
