The drug titration paradox: a control engineering perspective.

PURPOSE OF REVIEW: The drug titration paradox describes that, from a population standpoint, drug doses appear to have a negative correlation with its clinical effect. This paradox is a relatively modern discovery in anesthetic pharmacology derived from large clinical data sets. This review will interpret the paradox using a control engineering perspective. RECENT FINDINGS: Drug titration is a challenging endeavor, and the medication delivery systems used in everyday clinical practice, including infusion pumps and vaporizers, typically do not allow for rapid or robust titration of medication being delivered. In addition, clinicians may be reluctant to deviate from a predetermined plan or may be content to manage patients within fixed goal boundaries. SUMMARY: This drug titration paradox describes the constraints of how the average clinician will dose a patient with an unknown clinical response. While our understanding of the paradox is still in its infancy, it remains unclear how alternative dosing schemes, such as through automation, may exceed the boundaries of the paradox and potentially affect its conclusions.

Electronic health record data is unable to effectively characterize measurement error from pulse oximetry: a simulation study.

Large data sets from electronic health records (EHR) have been used in journal articles to demonstrate race-based imprecision in pulse oximetry (SpO2) measurements. These articles do not appear to recognize the impact of the variability of the SpO2 values with respect to time ("deviation time"). This manuscript seeks to demonstrate that due to this variability, EHR data should not be used to quantify SpO2 error. Using the MIMIC-IV Waveform dataset, SpO2 values are sampled from 198 patients admitted to an intensive care unit and used as reference samples. The error derived from the EHR data is simulated using a set of deviation times. The laboratory oxygen saturation measurements are also simulated such that the performance of three simulated pulse oximeter devices will produce an average root mean squared (ARMS) error of 2%. An analysis is then undertaken to reproduce a medical device submission to a regulatory body by quantifying the mean error, the standard deviation of the error, and the ARMS error. Bland-Altman plots were also generated with their Limits of Agreements. Each analysis was repeated to evaluate whether the measurement errors were affected by increasing the deviation time. All error values increased linearly with respect to the logarithm of the time deviation. At 10 min, the ARMS error increased from a baseline of 2% to over 4%. EHR data cannot be reliably used to quantify SpO2 error. Caution should be used in interpreting prior manuscripts that rely on EHR data.

Bolus pharmacokinetics: moving beyond mass-based dosing to guide drug administration.

Despite the common approach of bolus drug dosing using a patient's mass, a more tailored approach would be to use empirically derived pharmacokinetic models. Previously, this could only be possible though the use of computer simulation using programs which are rarely available in clinical practice. Through mathematical manipulations and approximations, a simplified set of equations is demonstrated that can identify a bolus dose required to achieve a specified target effect site concentration. The proposed solution is compared against simulations of a wide variety of pharmacokinetic models. This set of equations provides a near-identical solution to the simulation approach. A boundary condition is established to ensure the derived equations have an acceptable error. This approach may allow for more precise administration of medications with the use of point of care technology and potentially allows for pharmacokinetic dosing in artificial intelligence problems.

Optimizing the use of point of care testing devices for screening patients.

Point of Care Testing (POCT) devices are regularly used to improve clinical workflows in the hospital setting despite generally having inferior performance when compared to standardized laboratory analyzers. We describe a method to improve the efficacy of using a POCT device as a screening test when the laboratory values occur over a continuum and applied this methodology to the process of International Normalized Ratios (INR) screening on day of surgery. Following IRB approval, laboratory INR values on the day of surgery were extracted from the University of Vermont Medical Center operating room's electronic health record. Two separate theoretical POCT device values were simulated from the performance characterized by two prior publications (Jacobson and Hur). The sensitivities and specificities of the two theoretical devices were calculated over a range of values, in order detect an INR value greater or equal than 1.5 and 1.8. Subsequently, the percentage of the population with an INR value over each threshold was also calculated. Laboratory data from March 2008 to December 2016 were collected, and 9320 discrete INR values were compiled ranging from 0.8 to > 20. Two POCT devices were simulated using that dataset. The sensitivities and specificities over a range of values were determined, and the optimal cutoff values were identified for each device separately. Calculating the sensitivities and specificities over a range of values can optimize the clinical efficacy of a POCT device. By optimizing the use of POCT devices, hospitals may be able to improve clinical processes and reduce costs.

Nonoperating Room Anesthesia Tardiness.

Tardiness in the operating room has been shown to decline in the day as a result of operational decisions on the day of surgery. This article studies nonoperating room anesthesia (NORA) tardiness at the University of Vermont Medical Center in cases performed in the 2015 calendar year. Tardiness was measured by subtracting actual start times from extracted scheduled start times for each NORA services line. On average, tardiness in NORA sites increased as the day progressed, with the exception of diagnostic radiology. This is likely due to limited tactical and operational opportunities to improve workflow.

Time-delay when updating infusion rates in the Graseby 3400 pump results in reduced drug delivery.

Infusion pumps are commonly used for infusion of drugs for physiologic control, and infusion rate has been demonstrated to affect the parameters of pharmacokinetic models. In attempting to develop a model that explained this behavior, we examined the behavior of the Graseby 3400 syringe pump under a range of flow conditions and with variations in syringe characteristics. Two issues were identified: start-up loss (the difference between actual and ideal delivery on initial infusion) and update loss (the difference between actual and ideal delivery when transmitting a command to change infusion rate). Under worst-case conditions, this may result in a 20-second period of zero delivery during start-up, and when updating infusion rates once per second, zero output. These effects are influenced by syringe characteristics and vary sufficiently as to make it impossible to isolate this effect from the pharmacokinetic process being controlled. The implications of this for previous published results and clinical application of target-controlled infusions are discussed.

The variability of response to propofol is reduced when a clinical observation is incorporated in the control: a simulation study.

BACKGROUND: When using a target-controlled infusion of propofol to produce sedation, the operator assumes that the individual patient's pharmacokinetic parameters match those in the control system so that the specified effect-site target is achieved, and that the specified target is appropriate for the individual patient's sensitivity. These inaccuracies cascade, and this produces error in the desired level of sedation, termed "target error." To address this issue, we designed a control system that incorporates the operator's observation of loss of responsiveness to determine patient sensitivity. Our hypothesis was that this control system would reduce the impact of pharmacokinetic parameter error and uncertainty in sensitivity on the system's target error. METHODS: A novel control system was implemented that produces a slow transition in the probability of loss of responsiveness, providing the operator with greater resolution to observe the time of this transition. The system uses the time of this transition to infer the effect-site concentration associated with loss of responsiveness, and the infusion sequence necessary to maintain this concentration. We used computer simulation to generate a population of 10,000 patients with randomly distributed pharmacokinetic parameters and sensitivity to propofol, and compared the target error of our system with that of a target-controlled infusion system targeting the effect-site concentration associated with 50% probability of loss of responsiveness. RESULTS: Our system exhibited a target error of -0.75% ± 8.96%, compared with 0% ± 27.6% for target-controlled infusion, reducing the variability in achieving the specified target by a factor of 3.1 compared with target-controlled infusion, which was significant at P < 0.0001. CONCLUSIONS: Our system reduces the impact of biological variability by including the operator in the control loop. The utility of this approach in clinical practice will require further evaluation.

Implantable drug delivery device using frequency-controlled wireless hydrogel microvalves.

This paper reports a micromachined drug delivery device that is wirelessly operated using radiofrequency magnetic fields for implant applications. The controlled release from the drug reservoir of the device is achieved with the microvalves of poly(N-isopropylacrylamide) thermoresponsive hydrogel that are actuated with a wireless resonant heater, which is activated only when the field frequency is tuned to the resonant frequency of the heater circuit. The device is constructed by bonding a 1-mm-thick polyimide component with the reservoir cavity to the heater circuit that uses a planar coil with the size of 5-10 mm fabricated on polyimide film, making all the outer surfaces to be polyimide. The release holes created in a reservoir wall are opened/closed by the hydrogel microvalves that are formed inside the reservoir by in-situ photolithography that uses the reservoir wall as a photomask, providing the hydrogel structures self-aligned to the release holes. The wireless heaters exhibit fast and strong response to the field frequency, with a temperature increase of up to 20°C for the heater that has the 34-MHz resonant frequency, achieving 38-% shrinkage of swelled hydrogel when the heater is excited at its resonance. An active frequency range of ~2 MHz is observed for the hydrogel actuation. Detailed characteristics in the fabrication and actuation of the hydrogel microvalves as well as experimental demonstrations of frequency-controlled temporal release are reported.