Substituting 8 of 89 targeted zip codes can reduce bias

State response to making program fair to Black Ohioans pending

Harm Reduction Ohio and the Columbus Kappa Foundation proposed a simple solution to the racial and Appalachian bias in the state’s new $2.5 million naloxone distribution program: change 8 of the 89 targeted zip codes to more accurately target where overdose death is occurring in Ohio and the demographic groups who are dying.

Last week, Gov. Mike DeWine and the Ohio Department of Mental Health and Addiction Services (OMHAS) proposed rushing the delivery of 30,000 Narcan kits to 89 targeted zip codes in an effort to reduce overdose death, which reached a record high of more than 5,000 overdose deaths in 2020.

However, the selected zip codes showed obvious racial bias, targeting many affluent, suburban zip codes with predominantly White residents and lower overdose death rates while ignoring many zip codes with horrifically high overdose death rates and mostly Black residents.

This is important because naloxone is a drug that reverses opioid overdoses, including those for fentanyl, even when it’s found in cocaine, meth or counterfeit pills. Narcan is the brand name for naloxone in nasal spray form.

Offer to reduce racial bias 

Nathaniel Jordan of the Columbus Kappa Foundation

Harm Reduction Ohio, the state’s largest naloxone distributor, documented the racial bias in the state’s plan here. The Columbus Kappa Foundation, the state’s largest naloxone distributor in the African American community, joined the outcry against a racially biased Narcan distribution plan. The zip code where the foundation is headquartered — in the Mount Vernon neighborhood of Columbus — was among the high mortality, predominantly Black neighborhoods ignored in the governor’s plan so that naloxone could be directed to affluent Columbus suburbs such as Hilliard.

Together, Harm Reduction Ohio (HRO) and the Columbus Kappa Foundation developed a plan to reduce most of the racial bias in the state’s plan while also helping poor, predominantly White Appalachian counties who were also inexplicably ignored in DeWine’s plan despite having some of the highest overdose death rates in Ohio.

Lori Criss, director, OMHAS

The offer was made in a politely worded email sent to OMHAS director Lori Criss Tuesday afternoon. The email is shown below and reads:

HRO-Columbus Kappa proposal to reduce racial bias

Zip codes to ADD

The HRO-Columbus Kappa plan identifies 8 high-mortality zip codes that can be substituted for eight suburban zip codes with lower overdose death rates. The zip codes to be added are:

• 44304 – Akron.
• 45206 – Cincinnati.
• 44507 – Youngstown.
• 45416 – Dayton.
• 43203 – Columbus.
• 45223 – Cincinnati.
• 44307 – Akron.
• 45631 – Gallia County.

Detailed demographic and overdose information for these zip codes are provided in this chart:

Zip codes to REPLACE 

The zip codes to be replaced are:

• 45014 – Fairfield.
• 45238 – Cincinnati.
• 43068 – Reynoldsburg.
• 44077 – Painesville/Fairport Harbor..
• 45342 – Miamisburg.
• 45103 – Batavia.
• 45244 – Cincinnati.
• 43026 – Hilliard.

Detailed demographic and overdose information for these zip codes are provided in this chart:

If this chart is hard to read, click here for a bigger, close-up look.

The HRO-Columbus Kappa proposal is based on a zip-code-level analysis of overdose deaths in the Ohio Department of Health mortality database from January 2018 through October 2020. A total of 295 of Ohio’s 1,200 zip codes had 100 or more overdose deaths in the period studied and were used in the analysis.

As the chart of “Zip Codes to be Replaced “shows, the state decided to direct extra life-saving naloxone to zip codes with some of the lowest overdose death rates in all of Ohio. At the same time, the state ignored other zip codes with mostly Black or Appalachian residents and some of the highest overdose death rates in all of Ohio .

This makes no sense. A targeted naloxone distribution program should target only zip codes with high overdose rates.

Oops: The state’s methodological misfire

The cause of the poor targeting was a methodological mistake made by the state: It used overdose death counts rather than overdose death rates. As a result, fast-growing suburban areas with big population — but low death rates — were chosen over inner city and rural areas that have fewer residents by high rates of overdose death.

The average population in the low-mortality suburban districts chosen by the state is 45,242, according to Census Bureau’s 2019 estimates. The high mortality districts that were ignored have an average population of 8,487.

The low-mortality zip codes targeted by the state had a total population of 361,932 and 218 deaths during the period studied. The high-mortality zip codes that would replace them had a total population of 67,893 and 217 overdose deaths during the same time.

The state’s methodological error caused the targeted zip codes to be an almost random collection of zip code numbers that reflect, more than anything, population growth in zip codes over time.

What the state missed

The real-world results the state missed:

• The four zip codes with the worst overdose death rates in Ohio.
• Eight of the 12 zip codes with the worst overdose death rates in Ohio.
• The zip code with the worst overdose death rate among Black Ohioans.
• Five of the nine zip codes with the worst overdose death rates in Ohio among Black residents.
• Four of the five Appalachian counties (Gallia, Meigs, Vinton, Pike) that suffered the Ohio’s worst overdose death rates in 2020.
• The zip code with the worst methamphetamine overdose death rate.
• The three zip codes with the worst cocaine overdose death rates.

Bad choices

But driven by population change (rather than overdose deaths in recent years), the state did target:

• 32 zip codes that have overdose death rates below the state average.
• 10 counties that have overdose death rates below the state average.

The HRO-Columbus Kappa Foundation proposal doesn’t ask the state to fix its mistake. Instead, it asks that the state add the zip codes with overdose death rates that rank 3, 4, 12, 23, 24, 28, 29 and 33 in the state.

Racial bias

Seven of the zip codes proposed to be added to the state’s program have large African American populations. Overdose death rates for Black Ohioans now exceed those for White Ohioans, so the racial injustice of favoring White residents in a targeted naloxone distribution effort is obvious and unacceptable.

The eighth zip code to be added is for Gallipolis in Gallia County, which has the state’s second high overdose death rate but was left out of the plan for targeted naloxone distribution.

The proposed fix asks the state to remove the zip codes ranking towards the bottom of of the 295 zip codes for which overdose death rates were computed, specially zip codes ranking 185, 208, 229, 241, 260, 268 and 284.

It should go without saying that an overdose reduction plan that targets a zip code that ranks 284 out of 295 in overdose death rate is poorly thought out, especially when dispersing extra Narcan, an expensive drug that costs $75 per kit.

State’s response?

Harm Reduction Ohio and the Columbus Kappa Foundation hope to have a response from the state soon on the HRO-Columbus Kappa proposal to mitigate the racial bias in the state’s original plan and unfortunate plan to send Narcan to so many wrong places. We’ll update readers when we hear back.