Humans are inevitably biased, and these biases can also affect use of the electronic health record (EHR), data collection and analysis, algorithm development and deployment, machine learning and artifical intelligence, and other informatics work. Today, Dr. Sara Murray and Dr. Michelle Lin highlight the importance of understanding the limitations of data applications in routine healthcare delivery and patient care and why diversity and inclusion in informatics matters.   

Show Notes

​

[0:00] - Intro
  • ​Clinical informatics is defined as the application of informatics and information technology to deliver healthcare services. It is also referred to as applied clinical informatics and operational informatics.
  • As data collectors, humans are biased, and our biases are perpetuated in informatics work ​

[01:51] - Introduction to Dr. Sara Murray and Dr. Michelle Lin

[03:06] - A Step in Their Shoes
  • Taste of home: Dr. Lin shares her COVID pandemic cravings for Taiwanese comfort dishes
  • Reality > expectations: Dr. Murray shares her previous misconceptions about food choices available in San Francisco--these biases are similar to biases in EHR databases!

[05:50] - Representation is Key!
  • Dr Murray emphasizes how informatics need to account for differences in patient populations related to the quality of patient care, and diversity in informatics helps keep biases to a minimum

[07:45] - Unheard Primary Sources
  •  Dr. Lin voices her concerns about how patients are often not asked to report data--rather, assumptions are made--to facilitate the efficiency of patient care
  • ​Completely and accurately recording patient information is especially difficult during COVID-19
  • Barriers to healthcare access are higher among patients of color, complicating predictions regarding their future needs

[11:56] - Racial Bias in Algorithms
  • Dr. Lin talks about how data can give an inadequate representation of patients’ utilization of healthcare services, mentioning previous work published by Dr. Ziad Obermeyer on racial bias in algorithms.

[12:20] - Predictive Overbooking
  • Dr. Murray highlights the ways in which sensitive patient demographic information can be used in predictive overbooking to negatively impact patient care through discrimination Discrimination by Artificial Intelligence (AI) in a Commercial Electronic Health Record—a Case Study
  • Predictions focus on the patient ‘no-show’-issue that can be detrimental to clinics relying on revenue
  • Double-booking patients likely more deeply impacts the patient in greater need of care in the event patients assigned to the same time slot both show up

[15:12] - Introduction to Trustworthy Artificial Intelligence (AI) 
  • Dr. Murray delineates the seven guidelines published by the European Commission that aims to address the issue of transparency in clinical algorithms
  • Ethics Guidelines for Trustworthy AI
  • Knowing algorithm inputs and the components of a score help identify limitations in predictions
  • Dr. Murray shares the steps UCSF has taken to validate algorithms to ensure they are deployable to providers identifying the corresponding recommendations on appropriate patient care

[23:00] - Effectively Engaging Populations 
  •  Dr. Lin points out how the barriers to technology access has a disproportionate impact across patient populations, including, for example, limited internet access. 

[24:38] - Responding to COVID-19 
  • Dr. Murray and Dr. Lin illuminate the thoughtful work being carried out to identify disparities in EHR data and assess whether an in-person visit is appropriate for a patient
  • Patients of color--especially those chronic comorbid conditions--are at higher risk during the pandemic
  • Telehealth technology is rapidly expanding to ensure populations in greater need of patient care have convenient access to it

[29:35] - Improving Telehealth
  • Dr. Lin and Dr. Murray acknowledge that telehealth technology, although accompanied by complications, nonetheless helps vulnerable patient populations bypass challenges associated with in-person visits

[33:20] - Ethical by Design
  • Dr. Leung recalls a talk from Dr. Katleen Gabriel's a Women in Data Science event about this concept, where even at early stages of prototyping any technology, health information technology included, should consider important ethical considerations and engage diverse design team members as a product is developed.

[34:55] -  Smart Application
  •  Dr. Murray stresses how the applications of models, such as the ‘no-show’ model, are even more important than the existence of underlying bias in models
  • Models’ ability to solve problems equitably and ethically are central to facilitating patient care
[35:29] - Women in Informatics
  •  Dr. Lin and Dr. Murray share sentiments on the importance of diversity in leadership to better address bias
  • Women are becoming more involved in informatics work and have much to offer

[37:22] - Continue the Conversation Online!
[37:47] - Outro

Credits:
Co-hosts/Producers: Tiffany Leung, Sarah Takimoto 
​Assistant Producer: Joanna Jain
Editors: Joanna Jain, with special thanks to Deepti Yechuri
Production Assistant: Emily Han
Website/Art design: Ann Truong
Social Media: DJ Gaines
Music: Chris Dingman

Special thanks to Dr. Moises Auron for introducing us to Dr. Mohit Gupta.

Disclaimer

.The DEI Shift podcast and its guests provide general information and entertainment, but not medical advice. Before making any changes to your medical treatment or execution of your treatment plan, please consult with your doctor or personal medical team. Reference to any specific product or entity does not constitute an endorsement or recommendation by The DEI Shift. The views expressed by guests are their own, and their appearance on the podcast does not imply an endorsement of them or any entity they represent. Views and opinions expressed by The DEI Shift team are those of each individual, and do not necessarily reflect the views or opinions of The DEI Shift team and its guests, employers, sponsors, or organizations we are affiliated with.