Re-engineering for Accurate, Timely, and Communicated Diagnosis of Cardiovascular Disease in Women (DREAM Lab)

Principal Investigator: Kristen E. Miller, Dr.P.H., CPPS, MedStar Health Research Institute, Hyattsville, MD 
Co-PI: John Yosaitis, M.D., MedStar Health, Washington, DC 
AHRQ Grant No.: HS027280 
Project Period: 09/12/19-09/29/24 
Description: DREAM Lab addressed systemic factors contributing to delayed or missed cardiovascular disease (CVD) diagnoses in women. Women are nearly twice as likely as men to receive an incorrect diagnosis following a heart attack and 30 percent more likely to have stroke symptoms misdiagnosed in emergency departments.^1,2 The lab used a systems engineering framework and human-centered design to identify contributing factors, codesign interventions, and evaluate their impact across ambulatory care settings.

The specific aims were to:

1. Identify factors contributing to diagnostic errors and inappropriate clinical management of CVD in women.
2. Codesign and prioritize human-centered solutions to reduce diagnostic risk.
3. Evaluate the structure, process, and outcome effects of interventions in simulated and real-world clinical settings. 

To achieve these aims, the lab implemented nine interwoven workstreams, including environmental scanning, data analysis, stakeholder engagement, design of clinical decision support (CDS) tools, and patient-facing interventions. Using a variety of data sources—EHRs, patient safety event reports, and patient experience surveys—researchers analyzed more than 88,000 CVD cases, reviewed over 12,000 safety event reports, and applied natural language processing and machine learning to identify patterns in diagnostic error risk.^1,2

The lab developed a taxonomy for patient-perceived diagnostic errors and validated it through free-text responses in national patient satisfaction surveys.³ The team found that diagnostic breakdowns were more commonly identified in paper-based than web-based surveys, especially in domains such as information gathering and patient-provider communication. Stakeholder interviews with women patients revealed consistent reports of dismissal, delays in care, and lack of diagnostic clarity, often influenced by gender and racial biases. These findings informed the development of patient-facing tools, including discharge materials and clinical communication prompts, tailored to women at risk for CVD.¹

Additional innovations included a scoping review of patient question prompt lists for use during diagnosis, which revealed that most prompts focused on treatment rather than earlier diagnostic stages—highlighting the need for more proactive patient engagement strategies.⁴ The lab also created a revised version of the DEER (Diagnostic Error Evaluation and Research) taxonomy to more accurately reflect patient experiences of diagnostic breakdowns.

In the health IT domain, the team reviewed 51 CDS tools for CVD and found that most lacked social determinant or gender-specific variables. These gaps informed a roadmap for designing more inclusive, bias-resistant CDS tools in the future. Built environment audits and geographic information system mapping further informed checklists and environmental design recommendations for improving diagnostic communication, particularly in underserved communities.

The lab contributed to the advancement of diagnostic safety science through system-level interventions that integrate patient engagement, clinical redesign, and advanced analytic tools. The team’s work supports national efforts to reduce disparities in cardiovascular care for women and address systemic gaps outlined in the Improving Diagnosis in Health Care report by the National Academy of Medicine.^1,5

To date, this PSLL’s work has resulted in at least five peer-reviewed publications, 10 citations in other publications, and several national presentations.

Publications

2024

• Baker KM, et al. Using patient experience surveys to identify potential diagnostic safety breakdowns: A mixed methods study. J Patient Saf 2024;20(8):556-563.
• Hill MA, et al. “What else could it be?” A scoping review of questions for patients to ask throughout the diagnostic process. J Patient Saf 2024;19(7):529-534.
• Tabaie A, et al. Evaluation of a natural language processing approach to identify diagnostic errors and analysis of safety learning system case review data: Retrospective cohort study. J Med Internet Res 2024;26:e50935.

2022

• Giardina TD, et al. Defining diagnostic error: A scoping review to assess the impact of the National Academies’ report Improving Diagnosis in Health Care. J Patient Saf 2022;18(8):770-778.

2021

• Smith KM, et al. Using patient experience surveys to assess diagnostic safety in urgent care. Health Serv Res 2021;56(S1):53-54.

References

1. Miller K. Final Report: Re-engineering for Accurate, Timely, and Communicated Diagnosis of Cardiovascular Disease in Women (DREAM Lab). 2024, MedStar Health Research Institute: Hyattsville, MD.
2. Tabaie A, et al. Evaluation of a natural language processing approach to identify diagnostic errors and analysis of safety learning system case review data: Retrospective cohort study. J Med Internet Res 2024;26:e50935.
3. Baker KM, et al. Using patient experience surveys to identify potential diagnostic safety breakdowns: A mixed methods study. J Patient Saf 2024;20(8):556-563.
4. Hill MA, et al. “What else could it be?” A scoping review of questions for patients to ask throughout the diagnostic process. J Patient Saf 2024;19(7):529-534.
5. Giardina TD, et al. Defining diagnostic error: A scoping review to assess the impact of the National Academies' report Improving Diagnosis in Health Care. J Patient Saf 2022;18(8):770-778.