Redlining Scoping Review Individual Study Effect Sizes and Outcome List

What the evidence says

The strongest causal evidence in this set comes from a randomized housing-voucher experiment, not from historical redlining maps directly: children whose families used an experimental voucher to move to a lower-poverty neighborhood before age 13 had annual incomes $3,477 higher in their mid-twenties (31% above a control-group mean of $11,270), a 2.5 percentage point (16%) increase in college attendance over a 16.5% control-group base rate, lower neighborhood poverty in their own adulthood, and a lower likelihood of single parenthood among females. The mechanism the source identifies is exposure time: gains "decline with the child's age at move," and children who were 13–18 at random assignment showed no measured earnings gain (point estimate -$967, not statistically significant), while pooled adult outcomes across the whole experimental group showed little or no effect (-4.7% on individual earnings, 95% CI upper bound +12%). This is evidence about neighborhood poverty exposure during childhood, not about HOLC grading specifically.

For historical HOLC grading itself, the birth-outcome evidence is narrow but directionally consistent at small magnitude: grade C (yellow-lined) neighborhoods increased odds of preterm birth (OR 1.02, 95% CI 1.00–1.05) and small-for-gestational-age birth (OR 1.03, 95% CI 1.00–1.05) relative to grade B, statewide in California, with a regional subgroup in the San Francisco–Oakland metro area showing a larger preterm birth effect (OR 1.10, 95% CI 1.01–1.19) and an elevated low birth weight odds (OR 1.09, 95% CI 1.00–1.20).

The largest cluster of findings covers cardiovascular, metabolic, and renal outcomes tied to HOLC grade, from a single review source. Residing in a HOLC grade D (redlined) neighborhood increased risk of major adverse cardiovascular events by 13.9% (HR 1.139, 95% CI 1.083–1.198) and shifted those events an average of 45 days earlier; bystander CPR was less common in grade D (35.8%) than grade A (41.8%); and cardiometabolic and cardiovascular disease burden increased stepwise from grade A to grade D, including hypertension, diabetes, obesity, CKD, coronary artery disease, and stroke. Diabetes-related mortality, childhood BMI, and community-level stroke prevalence were all higher in redlined areas, and CKD patients in grade B, C, or D neighborhoods had higher risk of new cardiovascular complications than those in grade A. Alongside disease burden, the same source reports access effects: residents of redlined neighborhoods had lower rates of routine healthcare visits, higher rates of being uninsured, more reliance on Medicaid or dual Medicare-Medicaid coverage, less pre-dialysis nephrology consultation, more catheter-based (versus fistula) dialysis access, and worse kidney function at dialysis onset. Not every measured factor moved in the expected direction: walkability alone had no measured effect on childhood BMI.

A separate causal analysis of air pollution, using a spatial causal-inference model across 69 cities, found that grade B (Still Desirable) neighborhoods had higher NO2 (0.48 ppb, 95% CI 0.35–0.62) and higher PM2.5 (0.07 µg/m³, 95% CI 0.03–0.11) than grade A, with the harmful effect strongest in Los Angeles and Atlanta.

Above these individual studies, three scoping reviews describe the outcome landscape without supplying pooled effect sizes. One reports that studies "nearly universally" find an association between redlining and health-relevant outcomes but states that "heterogeneity in study design precludes direct comparison of results." Another lists redlining as linked to adverse maternal and neonatal outcomes, breast cancer mortality and subtype, HIV outcomes, stroke prevalence, cancer survival, and healthcare utilization, and explicitly states it "did not conduct a quantitative synthesis or meta-analysis," which "limits the ability to assess pooled effect sizes or compare strength across pathways." A third reports that 57% of studies found associations between discriminatory housing practices and mental health problems among Black American adults, and that redlining "produce[s] housing contexts and conditions involving environmental stressors" for Black residents, again without a quantified effect size. A fourth source notes only that a "Redlining Index" measure was used in two studies, on breast cancer survival and preterm birth, without reporting what those studies found.

Where the evidence splits

Within the same California birth-outcome study, the direction of effect reverses across adjacent grade comparisons: grade C versus grade B increased odds of preterm birth (OR 1.02) and SGA (OR 1.03), but grade D versus grade C decreased odds of SGA (OR 0.94, 95% CI 0.92–0.96) — the opposite of what a monotonic worse-grade-worse-outcome pattern would predict. The source does not explain this reversal.

The cardiovascular review reports several paired findings where White subgroups show better outcomes in redlined areas than the general grade-D pattern would predict, while Black subgroups show the expected worse outcomes: Black veterans from historically redlined neighborhoods had higher heart-failure risk than Black veterans in HOLC A–C areas, while White veterans from HOLC grade D neighborhoods had lower heart-failure risk than White veterans in HOLC A–C areas. Likewise, Black residents of HOLC grade C neighborhoods had higher risk of 30-day readmission or death after heart-failure hospitalization, while White residents of historically redlined areas had lower risk of 30-day readmission (with no difference in death or composite outcomes). Here the difference in population — race — is stated directly by the source as the stratifying variable, though the source does not explain the underlying mechanism for why the direction reverses by race.

The same review also reports that atrial fibrillation incidence and percutaneous coronary intervention rates were lower in HOLC grade D than grade A, despite grade D otherwise carrying higher rates of nearly every other cardiovascular and metabolic condition in this set. The source itself calls this "paradoxical" and offers survivor bias or unmeasured confounding only as a possible, unconfirmed explanation — it does not establish which applies.

Thresholds and dose-response

Age at move is a stated threshold in the voucher-mobility evidence: children who moved to a lower-poverty neighborhood before age 13 showed the earnings and college-attendance gains described above; children who were 13–18 at random assignment showed no significant earnings effect (point estimate -$967). Beyond this cutoff, the same source reports a continuous dose-response relationship: gains "decline with the child's age at move," consistent with "every extra year of exposure to a low-poverty environment during childhood" being beneficial.

For HOLC grading, one source states a stepwise dose-response across the full grade range: cardiometabolic risk factors and outcomes increased "progressing from HOLC grade A to grade D," including hypertension, diabetes, obesity, CKD, coronary artery disease, and stroke.

Open empirical questions

• What explains the reversed direction for small-for-gestational-age odds between HOLC grade D and grade C (OR 0.94) when the C-versus-B comparison in the same study runs the opposite way? The source does not offer a mechanism.
• What explains the race-stratified reversals in the cardiovascular review — White veterans and White residents of redlined areas showing better outcomes (lower heart-failure risk, lower readmission risk) than their counterparts in less-redlined grades, and lower atrial fibrillation incidence and PCI rates in grade D generally? The source raises survivor bias and unmeasured confounding only as possibilities, not established explanations.
• None of the three scoping reviews in this set (Swope, Kisa, Smith) conducted quantitative synthesis or meta-analysis, so no pooled effect size exists for redlining's relationship to breast cancer, HIV outcomes, cancer survival, or healthcare utilization broadly, despite associations being reported in individual studies.
• The direction and magnitude of gentrification's effect on housing consistency and housing costs is not specified by the source that mentions it.