# Towards a National Diagnostic Excellence Dashboard - Partnering with Stakeholders to Construct Evidence-Based Operational Measures of Misdiagnosis-Related Harms

> **NIH AHRQ R01** · JOHNS HOPKINS UNIVERSITY · 2021 · $372,122

## Abstract

Project Summary (Abstract)
This four-year project combines the assets of a leading academic medical center (Johns Hopkins Medicine)
with those of a recognized international leader in diagnostic excellence (Society to Improve Diagnosis in
Medicine [SIDM]) and a major physician specialty society (American College of Emergency Physicians
[ACEP]) to break new ground in operational measurement of patient harms linked to diagnostic error.
To achieve the goal of improved patient outcomes through diagnostic excellence, it is essential to be able to
measure diagnostic performance. Diagnostic errors are the largest cause of preventable harms in US medical
care, affecting an estimated 12 million people each year, causing permanent disability or death in at least 0.5
million. Diagnostic safety is a priority research area for AHRQ and the National Academy of Medicine (NAM).
A key impediment to “moving the needle” on reducing harms from diagnostic error is the lack of measures that
matter to both patients and clinicians, yet can be fully operationalized (i.e., routinely monitored in the existing
workflow). Impactful diagnostic outcome measures would assess serious morbidity and mortality in clinical
contexts where diagnostic errors are known to occur. Ideal measures would be specific, valid, precise, and
comparable across institutions to facilitate benchmarking that identifies both low and high outlier performers.
This proposal uses a novel approach to constructing evidence-based diagnostic outcome measures with
readily-available administrative and claims data sets. The Symptom-disease Pair Analysis of Diagnostic Error
(“SPADE”) method first identifies a clinically-plausible relationship between a common presenting symptom
and a dangerous underlying disease (e.g., chest pain-heart attack, fever-sepsis, dizziness-stroke). It then
searches for a statistically-valid pattern of unexpected adverse events (e.g., observed greater than expected
short-term inpatient hospitalization following a treat-and-release emergency department [ED] visit). Once such
patterns are confirmed, they can be monitored to assess the impact of interventions to improve diagnosis.
This proposal seeks to mature a partially-developed SPADE measure (for dizziness-stroke, a frequent cause of
serious misdiagnosis-related harms) to the point of readiness for use in national benchmarking of hospital-level
diagnostic performance for quality improvement. This SPADE pair has been validated through detailed chart
review and statistical testing using data from four Johns Hopkins hospitals, and the National Quality Forum
(NQF) has named this measure as a top priority for immediate development. This project will advance the
measure towards broad adoption with two Specific Aims: (1) engage key national stakeholders to optimize
attributes of the missed stroke measure (via expert panel and emergency physician survey) and (2) measure
diagnostic performance of US hospital EDs using the refined missed stroke me...

## Key facts

- **NIH application ID:** 10201710
- **Project number:** 5R01HS027614-02
- **Recipient organization:** JOHNS HOPKINS UNIVERSITY
- **Principal Investigator:** DAVID NEWMAN-TOKER
- **Activity code:** R01 (R01, R21, SBIR, etc.)
- **Funding institute:** AHRQ
- **Fiscal year:** 2021
- **Award amount:** $372,122
- **Award type:** 5
- **Project period:** 2020-07-01 → 2024-05-31

## Primary source

NIH RePORTER: https://reporter.nih.gov/project-details/10201710

## Citation

> US National Institutes of Health, RePORTER application 10201710, Towards a National Diagnostic Excellence Dashboard - Partnering with Stakeholders to Construct Evidence-Based Operational Measures of Misdiagnosis-Related Harms (5R01HS027614-02). Retrieved via AI Analytics 2026-05-22 from https://api.ai-analytics.org/grant/nih/10201710. Licensed CC0.

---

*[NIH grants dataset](/datasets/nih-grants) · CC0 1.0*