Blog

ProjectBlog: We’re Not There Yet: Why New ICD-10 ‘Z-Codes’ on Seeking Work are a Bad Idea

Blog: We’re Not There Yet: Why New ICD-10 ‘Z-Codes’ on Seeking Work are a Bad Idea


Our Public Comment to the Centers for Disease Control (CDC)

The following is the public comment Patchwise Labs submitted to the Centers for Disease Control regarding the newly proposed ICD-10 Z-codes to document social determinants of health (SDOH).


May 10, 2019
ICD-10 Coordination and Maintenance Committee 
nchsicd10CM@cdc.gov

To Whom it May Concern,

We were delighted to hear the news that US healthcare industry leadership has proposed new codes to document social determinants of health (SDOH) in clinical practice through expansion of the ICD-10 Z-codes. As a creative strategy firm dedicated to promoting innovation to better address the SDOH, we’ve been tracking the development of emerging policy levers, technology platforms, and business strategies for the last few years.

It’s long been evident that a major bottleneck to progress in “moving upstream” towards more holistic, preventative value-based care delivery is the ability to measure SDOH. These new codes represent an important and timely step in the right direction, and we applaud the Committee for your leadership in considering these codes and offering the public an opportunity to comment.

We understand you must be receiving a large volume of input; the remainder of our comment centers on a single issue: We would urge you to remove two codes from subcategory Z56.8, ‘Other problem related to employment,’ for final approval:

  • Z56.83 — Unemployed and seeking work
  • Z56.84 — Unemployed but not seeking work

The reason boils down a single issue: Work requirements in the Medicaid program.

The intent of the new codes as a whole appears well-intentioned, and we would agree with the submitted commentary from United Healthcare (UHC) that the proposed z-codes would “increase specificity needed to appropriately capture, analyze, and act on SDOH data to improve outcomes for both consumers and populations.”

Yet, it has become clear in the last few years that, as a nation, the United States of America is not in agreement as to the intent of Medicaid as laid out in the 1965 Social Security Act Amendments: To “furnish medical assistance on behalf of families with dependent children and of aged, blind, or disabled individuals, whose income and resources are insufficient to meet the costs of necessary medical services.”

We submit there is ample evidence that suggests these two new codes would be too easily and readily weaponized to deny coverage for individuals and families based on a simple checkbox datapoint that could easily be abstracted from the rest of their medical and social record.

To be clear, we would agree there is potential value in using Z56.83–84 to help identify individuals in need of employment assistance, job placement, or financial assistance programs, or as part of a broader effort to document the socioeconomic factors occurring in patients’ lives or at the community level. However, we believe the potential for harm outweighs the benefit at this stage. This is less of a “yes or no” question, and more a matter of whether this is the right timing to incorporate these specific codes into federal policy.

The issue goes deeper than the political intent of work requirements, whose federal approval U.S. District Judge James Boasberg recently referred to as “arbitrary and capricious because it did not address … how the project would implicate the ‘core’ objective of Medicaid: the provision of medical coverage to the needy.”

The Pandora’s Box issue here is the risk of creating and codifying a data hazard that undermines the goals of these new z-codes: Digitizing documentation of whether individuals are seeking employment or not would make it frighteningly easy to roll out eligibility algorithms that kick a family off of their medical coverage, based on a single datapoint.

This may sound like dystopian science fiction, but it’s already become reality: In her book Automating Inequality, Virginia Eubanks provides numerous examples of how public systems have further disenfranchised the vulnerable in areas of welfare programs, homeless services, SNAP and EBT, and Medicaid eligibility.

She argues that automation tools codify the human bias in public assistance programs: “One of my greatest fears in this work is that we’re actually using these systems to avoid some of the most pressing moral and political challenges of our time — specifically poverty and racism.”

In closing: If the goal of the new codes is to “capture these social diagnoses and barrier situations to assist providers and consumers in obtaining routine care, medications, and preventive services,” as UHC suggests, then we submit that at this juncture, these two codes would create an irreversible path to undermining that goal. In the public’s interest, we’d urge you to wait on approving Z56.83–84 until the industry has the opportunity to better understand, debate, and discuss possible downstream implications.

Thank you for your consideration and for this opportunity to comment.

Sincerely,

Naveen Rao

Managing Partner
Patchwise Labs


Source: Artificial Intelligence on Medium

Leave a Reply

Your email address will not be published. Required fields are marked *

Back To Top
a

Display your work in a bold & confident manner. Sometimes it’s easy for your creativity to stand out from the crowd.

Social