Charted: Highest-Poverty ZIPs Host 1.9x More Plasma Centers per 100k Adults

The United States is the world's primary warehouse for source plasma - the raw fluid that becomes immunoglobulin, clotting factors, and albumin. That warehouse is not a handful of mega-plants. It is more than 1,200 storefront collection centers, each competing for twice-weekly donors who can leave with roughly $65 in average compensation. The policy question is not whether paid plasma exists. It is where the doors open, relative to the adults who live nearby.

This desk joins FDA Blood Establishment Registration addresses for active source-plasma / plasmapheresis sites to Census ACS poverty rates and adult (18+) population by ZIP Code Tabulation Area (ZCTA). The punchline is a density skew, not a vibes map: the highest-poverty ZCTA quintile hosts about 0.66 centers per 100k adults, versus 0.35 in the lowest-poverty fifth - a 1.9x gap. Openings since 2021 still place 41% of new sites in ZCTAs with poverty at or above 15%, even though those ZCTAs hold only about 22% of the adult population.

The interactive dashboard above toggles Density, Openings, States, and Context. Switch the density metric among centers per 100k, location index, and center share; filter the state panel by Census region. What follows is the geography underneath those controls.

Density, not just count

Raw center counts flatter large, poor states and bury the ratio question. Texas can lead the absolute leaderboard and still look different once you divide by adults. Nationally, with roughly 1,247 active US sites (including Puerto Rico in the end-2025 industry/FDA tally) against on the order of 258 million adults, the country averages about 0.48 centers per 100k adults. That national mean hides a staircase by poverty.

Slice ZCTAs into equal-adult poverty quintiles and the staircase is visible. Q1 (lowest poverty) sits near 0.35 centers per 100k. Q3 is close to the national mean. Q5 (highest poverty) clears 0.66. The location index - center share divided by adult share - moves from about 0.72 in Q1 to about 1.37 in Q5. An index of 1.0 would mean centers track adults one-for-one. Above 1.0 means over-representation; below means under-representation. Q5 is not a rounding error.

Read the table as a relative claim. It does not say every high-poverty ZIP has a center. It says that, conditional on adult population, high-poverty ZCTAs are more heavily supplied with collection doors.

What new centers are doing

Stock and flow are different stories. Industry and press tallies of the early-2020s expansion emphasize a rising middle-class share of openings - on the order of 43% of 2021-2025 new sites versus about 32% of pre-existing stock, depending on how "middle class" is banded. That shift is real enough to matter for narrative: plasma is no longer only a strip-mall story next to payday lenders.

It is not, however, a proportional story relative to adults. Desk bands that put >=15% ACS poverty in the "high" bucket still capture about 41% of recent openings while holding roughly 22% of adults. Low-poverty ZCTAs (<10%) hold a large adult share and a smaller-than-proportional center share even among new sites. The industry can broaden into moderate-income suburbs and remain over-indexed on poorer ZCTAs at the same time. Those statements are not contradictions; they are different denominators.

Cohort panels make the same point in density units. Pre-2015 stock shows the widest high-versus-low poverty density gap. 2015-2020 openings narrow it slightly. 2021-2025 openings narrow it again - high-poverty density falls, low-poverty density rises - without erasing the skew. Softening is not equalization.

States are not a clean poverty machine

Cross a selected set of states and the scatter of ACS poverty versus centers per 100k is noisy. Mississippi and Louisiana combine high poverty with high density. Utah and Iowa post high per-capita center counts with comparatively modest poverty rates - a reminder that collection footprints also chase labor markets, highway retail pads, and operator networks, not poverty alone. New York and Massachusetts sit low on density despite urban poverty pockets, reflecting land cost, zoning, and competitive siting as much as need.

Texas still dominates absolute counts (on the order of 178 centers in late-2024 published tallies - roughly Europe's commercial total). Per 100k adults it remains elevated but not the national maximum. California looks under-supplied on a per-adult basis relative to its size. State scatterplots are useful for humility: poverty predicts tract and ZIP density better than it predicts a clean 50-state regression line.

Tract evidence that predates the latest boom

Before the center count cleared 1,000, Michigan Poverty Solutions researchers geocoded FDA-approved plasmapheresis addresses as of May 2017 and compared census tracts with and without a center using ACS 2011-2015. Tracts with centers averaged 12.0% of residents in deep poverty (below 50% of the federal poverty line) versus 7.3% in tracts without - a 64% elevation. Poverty between 50% and 100% of the line was similarly elevated. Multivariate models kept the poverty associations after race/ethnicity and urbanicity entered; urban tracts dominated placement (96.8% of centers versus about 82% of tracts classified urban).

That study is not a 2025 ZIP dashboard. It is a disclosed historical anchor: when the industry was already large but still smaller than today, community poverty traveled with the doors. The newer ZIP density work asks whether the post-2020 build-out erased that pattern. The desk answer is no - it softened the edges.

Collections, compensation, and why siting pays

PPTA-era collection narratives and industry tallies put US source-plasma volumes on a steep climb from the late 2000s through the 2010s, with 2019 often cited near 53.5 million donations. More recent liter-based tallies put 2024 collections around 62.5 million liters, with 2025 growth driven as much by next-generation plasmapheresis yield as by net new doors. Compensation on the order of $4-5 billion annually flowing to donor households is not a side note; it is the labor-market hinge that makes retail siting sensitive to local cash needs.

Operators do not publish a poverty-targeting memo. They publish throughput. Throughput prefers places where appointment books fill twice a week. Neighborhood poverty is a noisy but useful correlate of that constraint - alongside parking, lease rates, and competitor spacing. Treating the map as predatory by ZIP code overstates intent; treating it as random with respect to poverty contradicts the density ladder.

Caveats and what this is not

Several limits are load-bearing. ZCTA is not donor home. Centers draw from a commute shed; a site on a commercial arterial can sit in a poorer ZCTA while drawing suburban donors, or the reverse. ACS poverty is a neighborhood proxy, not a donor survey; we lack a public microdata file of who sells plasma. Desk joins of FDA addresses to ACS are analytical products - geocoding error, establishment-type filters, and vintage mismatch (active list versus ACS 5-year window) all matter. Confidence tags in the data module mark disclosed study figures separately from estimated ZIP density ranks. Middle-class share definitions vary by outlet; treat the 43% / 32% contrast as directional. State panels are a selected cross-section, not a full 50-state census of every license. Finally, density is not harm: this post measures clustering relative to adults, not clinical outcomes for donors or patients.

What to watch next

Three observables will tell whether the skew keeps softening. First, the location index on Q5 versus Q1 in the next FDAxACS refresh - if new pads keep landing in >=20% poverty ZCTAs faster than adults live there, the 1.9x gap holds. Second, the high-poverty share of openings versus the adult share in those ZCTAs - the gap between roughly 41% and 22% is the flow test. Third, whether liter growth continues to outpace door growth; if yield technology does more of the work, siting pressure may ease even if the existing stock remains skewed.

Plasma is a global therapeutics input assembled from local retail labor. The US map of that labor is still tilted toward poorer ZCTAs on a per-adult basis. The tilt is less cartoonish than a decade ago. It has not gone away.