Draft

185  Locke Ctsi Translational Research Trainee Symposium Talk

185.1 Summary

  • Improved Recognition of Inpatient Hypercapnic Respiratory Failure
  • Hypercapnic Respiratory Failure
  • E.g.: Home PAP improves mortality in ↑CO2-COPD
  • By the time patients are caught: ↑morbidity
  • Prior work:
  • Probabilistic Modeling Likelihood of Hypercapnia
  • Aim 1: Optimize (and externally validate) local performance of statistical model
  • Aim 2: To estimate the prevalence of unrecognized hypercapnia in inpatients
  • Aim 3: Estimate the potential impact of early & reliable identification
  • Aim 3: To estimate the potential impact of early & reliable identification
  • OVERALL: Patients with Consequentially Missed Hypercapnia can be Identified Using Health Record Elements

185.2 Slide outline

185.2.1 Slide 1

  • Improved Recognition of Inpatient Hypercapnic Respiratory Failure
  • Brian Locke, MD
  • T32 Fellow
  • Division of Pulmonary, Critical Care, and Occupational Pulmonary Medicine
  • Department of Internal Medicine
  • University of Utah School of Medicine
  • This research was supported by the National Institutes of Health under Ruth L. Kirschstein National Research Service Award 5T32HL105321 from the NIH
  • and The American Thoracic Society through the Academic Sleep and Pulmonary Integrated Research/Clinical Fellowship (ASPIRE) Program ### Slide 2
  • Hypercapnic Respiratory Failure
  • Hypoxemic (Low O2) Respiratory Failure
  • Hypercapnic (High CO2) Respiratory Failure
  • Low arterial blood oxygen tension
  • High arterial blood carbon dioxide tension
  • (Relatively) Reliably Identified by Pulse Oximetry
  • No routine non-invasive testing
  • Build up here CO2 in blood rises
  • Patients with hypercapnia are identified unreliably and late in their illness ### Slide 3
  • E.g.: Home PAP improves mortality in ↑CO2-COPD
  • Weight loss
  • Medication Adjustment
  • Pulmonary Rehab
  • Hypercapnia is common, is probably becoming more common, and is treatable.
  • Common contributors to Hypercapnic Respiratory Failure:
  • Advanced Age
  • Opiate Use
  • Obesity
  • Advanced Lung Dz
  • Multimorbidity
  • Population-standardized prevalence PaCO2 > 45 mmHg after excluding iatrogenic causes:
  • 150 per 100,000 person/year
  • Only inpatients, only those with blood gasses checked
  • Comparison: Decompensated Cirrhosis 94.9 (US, 2017) per 100,000 person year ### Slide 4
  • By the time patients are caught: ↑morbidity
  • Inpatient ICD code for Hypercapnic Respiratory Failure
  • 23% 30d readmission rate
  • Same as CHF, more than MI
  • 66% with recurrence
  • Could we catch them earlier? Would it help?
  • Decompensated Cirrhosis, for reference ### Slide 5
  • Prior work:
  • Tool
  • Health Record-Based Modeling
  • Transcutaneous CO2
  • Purpose
  • Refine Estimation of Pre-test Odds of Hypercapnia
  • Allow non-invasive CO2 assessment (patients unlikely to consent for research ABGs)
  • Limitation
  • Approach only externally validated in outpatient OHS assessment
  • Prior 95% agreement +/- 6 mmHg insufficient for unstratified clinical use
  • Total
  • No ABG Obtained
  • ABG Obtained
  • N
  • 879,019
  • 675,620
  • 203,399
  • Ambulatory
  • 39% (343,699)
  • 50% (334,457)
  • 5% (9,242)
  • Emergency
  • 20% (175,792)
  • 22% (146,066)
  • 15% (29,726)
  • Inpatient
  • 41% (359,528)
  • 29% (195,097)
  • 81% (164,431)
  • Candidate Predictors
  • (LASSO)
  • Age
  • Sex
  • Serum HCO3-
  • Serum Potassium
  • Serum Creatinine
  • Admitting Dx ### Slide 6
  • Probabilistic Modeling Likelihood of Hypercapnia
  • Validation & Tuning
  • Aim 1: Integration of NLP/LLM
  • Aim 2: Prospective Validation of PPV
  • Aim 3: Evaluation of Prognostic Significance ### Slide 7
  • Aim 1: Optimize (and externally validate) local performance of statistical model
  • Apply TriNetX model to retrospective data from the University of Utah.
  • Reference standard: patients who had arterial blood gasses obtained.
  • Add natural language processing using established tools / pipelines (CLAMP, LLM )
  • Adjust model weights
  • Generate measures of model performance (discrimination and calibration) to inform other aims ### Slide 8
  • Aim 2: To estimate the prevalence of unrecognized hypercapnia in inpatients
  • Methods:
  • Deferred consent (CO2 Narcosis)
  • Verifying Positive Predictive Value (precision) in high-risk patients is more efficient than stratified sampling of the whole population.
  • Aim demonstrate possible benefit in highest risk, rathern than delineate scope of the probably (initially)
  • Sample Size Calculation:
  • 25% ‘prevalence’ (Nowbar); model AUC; Se/Sp TcCO2 for true reference standard 85%/85%
  • How will this fail?
  • Aim 1 results will influence needed sample size
  • Can I access EHR data fast enough?
  • Hawthorne effect? Due to hypoxemia (elevation), do they get identified earlier? ### Slide 9
  • Prior work:
  • Statistical Insights:
  • Accuracy needed for clinical care ≠ Required accuracy for model verification
  • Pre-test probability estimate allows stratified sampling → efficiency
  • …
  • 49: Correct 93%, Wrong 7%: weight: 1.98%
  • 48: Correct 88%, Wrong 12%: weight: 2.39%
  • 47: Correct 80%, Wrong 20%: weight: 2.47%
  • 46: Correct 69%, Wrong 31%: weight: 2.92%
  • 45: Correct 57%, Wrong 43%: weight: 3.11%
  • SENSITIVITY
  • SPECIFICITY
  • 44: Correct 57%, Wrong 43% 3.56%
  • 43: Correct 69%, Wrong 31% weight 3.73
  • 42: Correct 80%, Wrong 20% weight 4.16%
  • 41: Correct 88%, Wrong 12% weight 4.25%
  • 40: Correct 93%, Wrong 7% weight 4.58%
  • ….
  • Expected test characteristics:
  • sensitivity 89%
  • specificity 92%
  • if
  • agreement ~6 mmHg
  • same distribution of CO2
  • Nowbar et al. 2004: All w/ BMI ≥35 kg/m2 admitted to inpatient medicine service at the Denver VA. Everyone (n277) gets an ABG. 31% had hypercapnia. Only 1/3 of those had therapy started (teams were told) ### Slide 10
  • Aim 3: Estimate the potential impact of early & reliable identification
  • Might clinicians already identify intervenable cases?
  • Prior (limited) research:
  • Marik et al. 2013: patient diagnosed with OHS are repeatedly admitted and missed diagnosed prior to their ultimate recognition
  • Meservey et al. 2020: most readmissions are recurrence
  • Hypothesis: Patients predicted to have had hypercapnia that was not recognized have a similar or excess risk of subsequent readmission as those identified (by blood gas and/or ICD code)
  • Analogy to Cancer Screening
  • The diagnostic barnyard ‘fence’:
  • Birds won’t get caught
  • Turtles won’t benefit form being caught
  • Rabbits are the ones to focus on ### Slide 11
  • Aim 3: To estimate the potential impact of early & reliable identification
  • Study Design:
  • Retrospective EHR-based review
  • Dataset: Longitudinal, ideal self-contained system
  • Exposures of interest:
  • Hypercapnic Respiratory Failure by ICD code
  • Hypercapnic Respiratory Failure by ABG (or by VBG) & not ICD code
  • Predicted Risk of Hypercapnic Respiratory Failure by commonly-obtained data-elements
  • Effect modifier: PAP/NIV therapy
  • Outcome: Hazard of death or rehospitalization with hypercapnia
  • Threats to validity:
  • Is it the hypercapnia risk, per-se? (negative controls)
  • Risk might be ameliorated by treatment
  • Risk among patients with hypercapnia is not binary [ie risk PaCO290 >>> 50]
  • Those who likely had hypercapnia, but were not sampled
  • Model predicted to have hypercapnia
  • Comparison in consequences ### Slide 12
  • OVERALL: Patients with Consequentially Missed Hypercapnia can be Identified Using Health Record Elements
  • OVERALL: There is a significant burden of unrecognized hypercapnia among hospitalized patients.
  • 3 aims
  • Aim 1: Model Refinement and Local Optimization
  • Aim 2: Prospective Validation
  • Aim 3: Assess Potential Impact
  • Hypothesis
  • Fine-tuning and refining previously derived model will improve local performance
  • Patients with hypercapnia that are not recognized are common can be identified using health record elements with high positive-predictive value.
  • Inpatients who were predicted, but not confirmed, to have hypercapnic respiratory failure are at high risk of adverse outcomes.
  • Rationale
  • The model should be optimized for local performance
  • Must be verified that model identifies unrecognized hypercapnia
  • Need to confirm that the identified patients face elevated risk (ABG PaCO2 is an imperfect reference standard).
  • Approach
  • Optimizing model performance +/- use of NLP tools) applied to U of U data
  • Non-invasively sample (with TcCO2) patients predicted to have hypercapnia that teams have not recognized
  • Among patients with an admission for hypercapnia that is identified, how many had preceding admissions likely to have hypercapnia?
  • Design
  • Retrospective, U of U
  • Prospective, U of U
  • Retrospective, VA
  • Payoff: This work will evaluate the potential for improved recognition of hypercapnia to help a particularly high-risk, enlarging, and understudied group of patients.
  • Career Development: Prospective validations of bioinformatic tools require a unique skillset ### Slide 13
  • Research Team: Wayne Richards1,2 MSJoseph Finkelstein2 MD PhD MAJeanette Brown1 MD PhDRamkiran Gouripeddi2 MBBS MSKrishna M Sundar1 MDDivisions of 1Internal Medicine and 2Biomedical InformaticsThanks for advice from: Robert Paine, MD; Barb Jones MD MSCI; Ken Kawamoto MD PhD MHS; MDCRC6450
  • Additional Questions
  • Handling of partial verification
  • Incorporation of Venous BG
  • Prospective validation logistics
  • Other modeling end-points
  • Realtime availability of data
  • VA aata access time/effort
  • PPV vs Se/Sp/Prevalence?
  • Socio-technological barriers
  • Sources of provider/unit/institution variation?
  • Brian.Locke@hsc.utah.edu ### Slide 14
  • What is expected to come from this research?
  • Validation of a method of stratification that can be used to:
  • Facilitate the identification of patients with high CO2 → improved patient care
  • Identify patients with hypercapnia (± a verification method) for study inclusion → more representative cohorts of patients.
  • Enable future studies assessing:
  • Overall burden of disease from hypercapnic respiratory failure
  • Effect of exposures (medications, environmental, policy) on risk of hypercapnia
  • Test the effect of treatment interventions on broader types of high-risk patients.
  • Career-development:
  • Development of models (statistical, ML, etc.) exploding in healthcare
  • Bottle-neck is in prospective validation of uses, which takes a unique skillset to evaluate.

185.3 Learning objectives

  • Improved Recognition of Inpatient Hypercapnic Respiratory Failure
  • Hypercapnic Respiratory Failure
  • E.g.: Home PAP improves mortality in ↑CO2-COPD
  • By the time patients are caught: ↑morbidity
  • Prior work:

185.4 Bottom line / summary

  • Improved Recognition of Inpatient Hypercapnic Respiratory Failure
  • Hypercapnic Respiratory Failure
  • E.g.: Home PAP improves mortality in ↑CO2-COPD
  • By the time patients are caught: ↑morbidity
  • Prior work:

185.5 Approach

  1. TODO: Outline the initial assessment or decision point.
  2. TODO: Outline the next diagnostic or management step.
  3. TODO: Outline follow-up or escalation criteria.

185.6 Red flags / when to escalate

  • TODO: List red flags that require urgent escalation.

185.7 Common pitfalls

  • TODO: Capture common errors or missed steps.

185.8 References

TODO: Add landmark references or guideline citations.

185.9 Slides and assets

185.10 Source materials