Chapter 5.4 · Where illness keeps company

Multiple chronic conditions

Weighted by adults, a tract in the most deprived tenth is high on 5.7 of nine chronic conditions on average, against 0.5 in the least deprived tenth, and the conditions correlate at +0.49 across tracts, a figure that is +0.47 after an approximation of the model's census inputs and state effects is removed, which does not show that the overlap is real.

Author: Vishal Singh
Affiliation: NYU Stern
Published: 9 October 2026

Weighted by adults, a tract in the most deprived tenth of census tracts is in the highest 25% of all tracts for the prevalence of 5.7 (95% CI 5.5 to 5.8) of nine chronic conditions on average, against 0.5 (95% CI 0.4 to 0.6) in the least deprived tenth, a difference of 5.2 (95% CI 5.0 to 5.4) conditions. Across 82,781 tracts, the average correlation between the prevalences of two of the conditions is +0.49 (95% CI 0.48 to 0.51). This chapter asks which conditions are high in the same neighborhoods and how that overlap changes with deprivation. It is a chapter about places and says nothing about how many conditions any one person has.

The data are modeled estimates from CDC PLACES, which gives each tract one prevalence for each condition among adults and holds no record of any person.[1] In surveys that ask each adult about each condition, multiple chronic conditions means having two or more of them,[2] and a tract estimate cannot be turned into that count. A tract can be high on several conditions while no resident has more than one, and a tract can be high on none while some residents have several. A count of conditions in this chapter is a count of tract prevalences above a stated cutoff.

PLACES publishes 40 measures for tracts. The 9 conditions used here are the health outcome measures that cover all adults and have a tract estimate in every state: arthritis, cancer (non-skin) or melanoma, current asthma, coronary heart disease, COPD, depression, diagnosed diabetes, obesity and stroke. PLACES lists obesity, a measure of body weight, with the health outcomes. High blood pressure and high cholesterol (among adults ever screened) are left out because the 2024 release has no estimate of either for the 5,077 tracts of Florida, and a check on high blood pressure outside Florida appears at the end. A tract is high on a condition when its prevalence is in the highest 25% of tracts, with each tract counted once, and the count of a tract is the number of the nine conditions on which it is high. Means, correlations and shares of adults are population-weighted by the adults in each tract, and counts and shares of tracts count each tract once. Of the 83,522 tracts with estimates, 82,781 also have a rank on the Area Deprivation Index (ADI), a national percentile built from American Community Survey (ACS) data,[3] and the other 741 are left out. The approximation of the model's inputs in the fourth section needs ACS shares, and the ACS file lacks three of them for the District of Columbia, so that section leaves out the 203 tracts that lack a share, among them the 201 tracts of the District of Columbia. Every other statistic includes them.

The PLACES 2024 release uses survey responses from 2022, and from 2021 for blood pressure. Values are crude prevalence among adults, so the age structure of a tract moves them. The tract file carries no intervals, and the bands in this chapter come from resampling counties.

Stroke, diabetes, COPD and heart disease are high in the same tracts

The strongest correlations join stroke, diabetes, COPD, coronary heart disease and arthritis (Figure 1). Across tracts the prevalence of stroke correlates +0.90 (95% CI 0.89 to 0.92) with diabetes, +0.90 (95% CI 0.89 to 0.90) with COPD and +0.86 (95% CI 0.85 to 0.87) with heart disease, and the pairs among stroke, diabetes, COPD and heart disease run from +0.74 to +0.90. Arthritis correlates +0.90 (95% CI 0.89 to 0.91) with heart disease. At the other end, cancer and depression have the weakest correlations with the rest, and depression correlates only +0.06 (95% CI 0.01 to 0.11) with diabetes. The mean correlation of cancer with the other eight conditions is +0.22 (95% CI 0.19 to 0.25) and that of depression is +0.26 (95% CI 0.22 to 0.30), against +0.68 (95% CI 0.66 to 0.69) for COPD. Of the 36 pairs, 32 have an interval entirely above zero and 2 have one entirely below it, cancer with asthma (−0.13 (95% CI −0.20 to −0.07)) and cancer with obesity (−0.11 (95% CI −0.18 to −0.05)). One principal component accounts for 59.0% (95% CI 57.9% to 60.2%) of the variance of the nine standardized prevalences, so most of the difference between tracts in these conditions follows a single shared pattern. Every correlation here is between modeled estimates and is partly built in by the inputs the model shares across conditions, a point the section on the model's inputs takes up.

Figure 1

Stroke, diabetes, COPD and heart disease are high in the same tracts, with correlations from +0.74 to +0.90, while cancer correlates −0.13 with asthma

Population-weighted Pearson correlations between the tract prevalence of nine chronic conditions among adults, 82,781 census tracts, 2022

−1.000.00+1.00

A nine by nine grid of correlations between the tract prevalence of nine chronic conditions. Stroke, diabetes, COPD, coronary heart disease and arthritis correlate strongly and positively with one another. Asthma and obesity correlate moderately with those conditions and with each other. Cancer and depression have the weakest correlations with the other conditions, and cancer is slightly negative against asthma and obesity.
Measure	Cancer	Arthritis	Coronary heart disease	COPD	Stroke	Diabetes	Obesity	Asthma	Depression
Cancer	+1.00	+0.74, interval +0.72 to +0.76, n = 82,781	+0.63, interval +0.60 to +0.66, n = 82,781	+0.33, interval +0.30 to +0.36, n = 82,781	+0.25, interval +0.20 to +0.29, n = 82,781	+0.05, interval 0.00 to +0.11, n = 82,781	−0.11, interval −0.18 to −0.05, n = 82,781	−0.13, interval −0.20 to −0.07, n = 82,781	0.00, interval −0.07 to +0.06, n = 82,781
Arthritis	+0.74, interval +0.72 to +0.76, n = 82,781	+1.00	+0.90, interval +0.89 to +0.91, n = 82,781	+0.81, interval +0.80 to +0.82, n = 82,781	+0.74, interval +0.71 to +0.76, n = 82,781	+0.58, interval +0.52 to +0.64, n = 82,781	+0.48, interval +0.42 to +0.53, n = 82,781	+0.37, interval +0.32 to +0.42, n = 82,781	+0.25, interval +0.20 to +0.30, n = 82,781
Coronary heart disease	+0.63, interval +0.60 to +0.66, n = 82,781	+0.90, interval +0.89 to +0.91, n = 82,781	+1.00	+0.88, interval +0.87 to +0.89, n = 82,781	+0.86, interval +0.85 to +0.88, n = 82,781	+0.74, interval +0.71 to +0.76, n = 82,781	+0.48, interval +0.43 to +0.52, n = 82,781	+0.34, interval +0.28 to +0.38, n = 82,781	+0.17, interval +0.12 to +0.22, n = 82,781
COPD	+0.33, interval +0.30 to +0.36, n = 82,781	+0.81, interval +0.80 to +0.82, n = 82,781	+0.88, interval +0.87 to +0.89, n = 82,781	+1.00	+0.90, interval +0.89 to +0.90, n = 82,781	+0.77, interval +0.74 to +0.80, n = 82,781	+0.69, interval +0.67 to +0.72, n = 82,781	+0.61, interval +0.57 to +0.64, n = 82,781	+0.42, interval +0.37 to +0.46, n = 82,781
Stroke	+0.25, interval +0.20 to +0.29, n = 82,781	+0.74, interval +0.71 to +0.76, n = 82,781	+0.86, interval +0.85 to +0.88, n = 82,781	+0.90, interval +0.89 to +0.90, n = 82,781	+1.00	+0.90, interval +0.89 to +0.92, n = 82,781	+0.65, interval +0.62 to +0.68, n = 82,781	+0.55, interval +0.51 to +0.58, n = 82,781	+0.15, interval +0.11 to +0.19, n = 82,781
Diabetes	+0.05, interval 0.00 to +0.11, n = 82,781	+0.58, interval +0.52 to +0.64, n = 82,781	+0.74, interval +0.71 to +0.76, n = 82,781	+0.77, interval +0.74 to +0.80, n = 82,781	+0.90, interval +0.89 to +0.92, n = 82,781	+1.00	+0.69, interval +0.65 to +0.73, n = 82,781	+0.40, interval +0.36 to +0.46, n = 82,781	+0.06, interval +0.01 to +0.10, n = 82,781
Obesity	−0.11, interval −0.18 to −0.05, n = 82,781	+0.48, interval +0.42 to +0.53, n = 82,781	+0.48, interval +0.43 to +0.52, n = 82,781	+0.69, interval +0.67 to +0.72, n = 82,781	+0.65, interval +0.62 to +0.68, n = 82,781	+0.69, interval +0.65 to +0.73, n = 82,781	+1.00	+0.64, interval +0.59 to +0.69, n = 82,781	+0.41, interval +0.35 to +0.46, n = 82,781
Asthma	−0.13, interval −0.20 to −0.07, n = 82,781	+0.37, interval +0.32 to +0.42, n = 82,781	+0.34, interval +0.28 to +0.38, n = 82,781	+0.61, interval +0.57 to +0.64, n = 82,781	+0.55, interval +0.51 to +0.58, n = 82,781	+0.40, interval +0.36 to +0.46, n = 82,781	+0.64, interval +0.59 to +0.69, n = 82,781	+1.00	+0.62, interval +0.57 to +0.66, n = 82,781
Depression	0.00, interval −0.07 to +0.06, n = 82,781	+0.25, interval +0.20 to +0.30, n = 82,781	+0.17, interval +0.12 to +0.22, n = 82,781	+0.42, interval +0.37 to +0.46, n = 82,781	+0.15, interval +0.11 to +0.19, n = 82,781	+0.06, interval +0.01 to +0.10, n = 82,781	+0.41, interval +0.35 to +0.46, n = 82,781	+0.62, interval +0.57 to +0.66, n = 82,781	+1.00

Show data

Data for Stroke, diabetes, COPD and heart disease are high in the same tracts, with correlations from +0.74 to +0.90, while cancer correlates −0.13 with asthma
conf_high	conf_low	measure	n	other	r
n/a	n/a	Cancer	n/a	Cancer	1
0.755	0.718	Cancer	82,781	Arthritis	0.738
0.655	0.601	Cancer	82,781	Coronary heart disease	0.629
0.362	0.297	Cancer	82,781	COPD	0.332
0.287	0.203	Cancer	82,781	Stroke	0.246
0.106	−0.002	Cancer	82,781	Diabetes	0.052
−0.050	−0.180	Cancer	82,781	Obesity	−0.112
−0.069	−0.196	Cancer	82,781	Asthma	−0.127
0.057	−0.067	Cancer	82,781	Depression	−0.002
0.755	0.718	Arthritis	82,781	Cancer	0.738
n/a	n/a	Arthritis	n/a	Arthritis	1
0.914	0.888	Arthritis	82,781	Coronary heart disease	0.902
0.823	0.800	Arthritis	82,781	COPD	0.812
0.759	0.713	Arthritis	82,781	Stroke	0.736
0.635	0.520	Arthritis	82,781	Diabetes	0.577
0.531	0.421	Arthritis	82,781	Obesity	0.479
0.422	0.315	Arthritis	82,781	Asthma	0.374
0.305	0.197	Arthritis	82,781	Depression	0.253
0.655	0.601	Coronary heart disease	82,781	Cancer	0.629
0.914	0.888	Coronary heart disease	82,781	Arthritis	0.902
n/a	n/a	Coronary heart disease	n/a	Coronary heart disease	1
0.891	0.874	Coronary heart disease	82,781	COPD	0.883
0.875	0.851	Coronary heart disease	82,781	Stroke	0.863
0.763	0.710	Coronary heart disease	82,781	Diabetes	0.737
0.518	0.433	Coronary heart disease	82,781	Obesity	0.477
0.382	0.282	Coronary heart disease	82,781	Asthma	0.335
0.223	0.117	Coronary heart disease	82,781	Depression	0.172
0.362	0.297	COPD	82,781	Cancer	0.332
0.823	0.800	COPD	82,781	Arthritis	0.812
0.891	0.874	COPD	82,781	Coronary heart disease	0.883
n/a	n/a	COPD	n/a	COPD	1
0.904	0.887	COPD	82,781	Stroke	0.896
0.803	0.738	COPD	82,781	Diabetes	0.771
0.719	0.666	COPD	82,781	Obesity	0.694
0.641	0.569	COPD	82,781	Asthma	0.607
0.462	0.374	COPD	82,781	Depression	0.420
0.287	0.203	Stroke	82,781	Cancer	0.246
0.759	0.713	Stroke	82,781	Arthritis	0.736
0.875	0.851	Stroke	82,781	Coronary heart disease	0.863
0.904	0.887	Stroke	82,781	COPD	0.896
n/a	n/a	Stroke	n/a	Stroke	1
0.917	0.887	Stroke	82,781	Diabetes	0.902
0.675	0.623	Stroke	82,781	Obesity	0.649
0.580	0.507	Stroke	82,781	Asthma	0.545
0.193	0.109	Stroke	82,781	Depression	0.151
0.106	−0.002	Diabetes	82,781	Cancer	0.052
0.635	0.520	Diabetes	82,781	Arthritis	0.577
0.763	0.710	Diabetes	82,781	Coronary heart disease	0.737
0.803	0.738	Diabetes	82,781	COPD	0.771
0.917	0.887	Diabetes	82,781	Stroke	0.902
n/a	n/a	Diabetes	n/a	Diabetes	1
0.734	0.647	Diabetes	82,781	Obesity	0.691
0.456	0.356	Diabetes	82,781	Asthma	0.403
0.105	0.010	Diabetes	82,781	Depression	0.058
−0.050	−0.180	Obesity	82,781	Cancer	−0.112
0.531	0.421	Obesity	82,781	Arthritis	0.479
0.518	0.433	Obesity	82,781	Coronary heart disease	0.477
0.719	0.666	Obesity	82,781	COPD	0.694
0.675	0.623	Obesity	82,781	Stroke	0.649
0.734	0.647	Obesity	82,781	Diabetes	0.691
n/a	n/a	Obesity	n/a	Obesity	1
0.686	0.592	Obesity	82,781	Asthma	0.641
0.459	0.350	Obesity	82,781	Depression	0.407
−0.069	−0.196	Asthma	82,781	Cancer	−0.127
0.422	0.315	Asthma	82,781	Arthritis	0.374
0.382	0.282	Asthma	82,781	Coronary heart disease	0.335
0.641	0.569	Asthma	82,781	COPD	0.607
0.580	0.507	Asthma	82,781	Stroke	0.545
0.456	0.356	Asthma	82,781	Diabetes	0.403
0.686	0.592	Asthma	82,781	Obesity	0.641
n/a	n/a	Asthma	n/a	Asthma	1
0.660	0.572	Asthma	82,781	Depression	0.618
0.057	−0.067	Depression	82,781	Cancer	−0.002
0.305	0.197	Depression	82,781	Arthritis	0.253
0.223	0.117	Depression	82,781	Coronary heart disease	0.172
0.462	0.374	Depression	82,781	COPD	0.420
0.193	0.109	Depression	82,781	Stroke	0.151
0.105	0.010	Depression	82,781	Diabetes	0.058
0.459	0.350	Depression	82,781	Obesity	0.407
0.660	0.572	Depression	82,781	Asthma	0.618
n/a	n/a	Depression	n/a	Depression	1

Cells toward the teal end of the key are positive correlations and cells toward the brown end negative ones. The strongest cells join stroke, diabetes, COPD, heart disease and arthritis, and the weakest cells all involve cancer or depression, with cancer negative against asthma and obesity. Each cell is the correlation between two conditions across tracts, weighted by adults, with a 95% interval from a bootstrap over counties. Rows and columns share one order that places similar conditions together.Note: PLACES values are modeled estimates. A model fitted to survey responses is applied to each tract's census population, so every condition's estimate depends on the same age, sex, race, ethnicity and education inputs and on county and state effects, and correlations across tracts are partly built in. The tract file carries no intervals, and the intervals shown here come from a bootstrap over counties.Note: The diagonal is one by construction and has no interval. Values are crude prevalence among adults, not adjusted for age, so the age mix of a tract moves them.Note: The nine conditions are the PLACES health outcome measures that cover all adults and have an estimate in every state. High blood pressure and high cholesterol are left out because the 2024 tract release has no estimate of either for 5,077 Florida tracts; outside Florida the mean correlation of high blood pressure with the nine is +0.65.Source: CDC PLACES 2024 release (BRFSS 2022) and ADI 2023; modeled estimates, population-weighted (adults in each tract)

The correlations of cancer follow the age of the residents. PLACES gives crude prevalence, and across tracts the prevalence of cancer correlates +0.81 (95% CI 0.79 to 0.83) with the ACS share of residents aged 65 and over, against +0.64 (95% CI 0.61 to 0.66) for arthritis, +0.64 (95% CI 0.61 to 0.66) for heart disease, +0.23 (95% CI 0.19 to 0.26) for diabetes and −0.17 (95% CI −0.21 to −0.14) for depression. It correlates 0.00 (95% CI −0.07 to 0.07) with the ADI. Cancer correlates +0.74 (95% CI 0.72 to 0.75) with arthritis and +0.63 (95% CI 0.60 to 0.66) with heart disease, the two other conditions that line up with age, and close to zero with diabetes (+0.05 (95% CI 0.00 to 0.11)) and depression (0.00 (95% CI −0.07 to 0.06)). Two features of a tract therefore line up with the conditions. One is the age of its residents, which lines up with cancer, arthritis and heart disease. The other is deprivation, taken up in the sections below. Asthma correlates +0.64 (95% CI 0.59 to 0.69) with obesity. The modeled asthma estimates are set against air pollution in Chapter 5.3, Air pollution, and here asthma is one of nine conditions.

The same pattern can be counted. For the tracts that are high on stroke, the mean number of the other eight conditions that are also high is 4.9 (95% CI 4.8 to 5.0), and for tracts high on COPD it is 4.9 (95% CI 4.9 to 5.0) (Figure 2). For tracts high on depression it is 3.1 (95% CI 3.0 to 3.3) and for tracts high on cancer 2.5 (95% CI 2.4 to 2.6). If the conditions were unrelated across tracts, each mean would be 2.0. Weighting tracts by adults instead of counting each once moves each mean by 0.34 conditions or less.

Figure 2

Tracts high on stroke are also high on 4.9 of the other eight conditions on average, against 2.5 for tracts high on cancer and 2.0 if the conditions were independent

Mean number of the other eight conditions that are also high in a tract, for tracts in the highest 25% of tracts on each condition, 2022

Other conditions (of eight) that are also high in the same tract (mean number)

Show data

Data for Tracts high on stroke are also high on 4.9 of the other eight conditions on average, against 2.5 for tracts high on cancer and 2.0 if the conditions were independent
condition	conf_high	conf_low	estimate	n
Arthritis	4.65	4.48	4.57	20,620
Cancer	2.64	2.39	2.51	21,087
Asthma	3.91	3.55	3.73	20,589
Coronary heart disease	4.82	4.63	4.73	21,177
COPD	4.98	4.86	4.92	21,048
Depression	3.31	2.96	3.12	20,971
Diabetes	4.69	4.03	4.35	20,549
Obesity	4.15	3.86	3.99	20,885
Stroke	5.04	4.77	4.90	19,866

Every dot lies right of the benchmark rule, and cancer and depression lie furthest left. Each dot is the mean for the tracts that are high on that condition, with a 95% interval from a bootstrap over counties; the rule marks the number expected if the conditions were unrelated across tracts.Note: PLACES values are modeled estimates. A model fitted to survey responses is applied to each tract's census population, so every condition's estimate depends on the same age, sex, race, ethnicity and education inputs and on county and state effects, and correlations across tracts are partly built in. The tract file carries no intervals, and the intervals shown here come from a bootstrap over counties.Note: Weighting tracts by adults moves each mean by 0.34 conditions or less.Source: CDC PLACES 2024 release (BRFSS 2022) and ADI 2023; modeled estimates, unweighted (each tract counted once)

Tracts are high on all nine conditions, or on none, more often than unrelated conditions would give

Figure 3 counts, for each tract, the conditions on which it is high. 34.0% (95% CI 30.7% to 37.1%) of tracts are high on none of the nine, where unrelated conditions with the same shares of high tracts would leave 7.4%. At the other end 892 tracts, 1.1% (95% CI 0.9% to 1.3%) of the total, are high on all nine, where unrelated conditions would give 0.3 tracts. Tracts high on 5 or more of the nine make up 21.3% (95% CI 19.3% to 23.3%) of tracts against 4.9% for unrelated conditions. The benchmark is a reference point and not a hypothesis anyone held. It is the distribution that independent draws would give, with each condition's observed share of high tracts. Real tracts are not independent draws, in part because the model fits every condition to the same inputs. Weighted by adults, 38.1% of adults live in tracts that are high on none and 17.1% in tracts that are high on 5 or more, so large tracts are somewhat less likely to be high on many conditions.

Figure 3

34.0% of tracts are high on none of the nine conditions and 892 tracts are high on all nine, where independent conditions would give 0.3 tracts

Tracts by the number of the nine conditions whose prevalence is in the highest 25% of tracts, observed and expected if the conditions were independent, 82,781 tracts, 2022

Observed
If conditions were independent

Number of the nine conditions that are high in the tract

Show data

Data for 34.0% of tracts are high on none of the nine conditions and 892 tracts are high on all nine, where independent conditions would give 0.3 tracts
conf_high	conf_low	count	n	series	share
37.10	30.75	0	82,781	Observed	33.98
21.61	19.41	1	82,781	Observed	20.52
12.31	10.49	2	82,781	Observed	11.37
7.64	6.64	3	82,781	Observed	7.13
6.23	5.26	4	82,781	Observed	5.74
5.77	4.83	5	82,781	Observed	5.30
6.61	5.37	6	82,781	Observed	5.97
6.03	4.69	7	82,781	Observed	5.31
4.12	3.16	8	82,781	Observed	3.61
1.29	0.900	9	82,781	Observed	1.08
n/a	n/a	0	n/a	If conditions were independent	7.44
n/a	n/a	1	n/a	If conditions were independent	22.42
n/a	n/a	2	n/a	If conditions were independent	30.01
n/a	n/a	3	n/a	If conditions were independent	23.43
n/a	n/a	4	n/a	If conditions were independent	11.76
n/a	n/a	5	n/a	If conditions were independent	3.93
n/a	n/a	6	n/a	If conditions were independent	0.880
n/a	n/a	7	n/a	If conditions were independent	0.130
n/a	n/a	8	n/a	If conditions were independent	0.010
n/a	n/a	9	n/a	If conditions were independent	0

The observed bars are tallest at no high conditions and form a long right tail, while the benchmark bars peak at a low count and vanish toward the right. A tract is high on a condition when its prevalence is in the highest 25% of tracts. The observed bars are shares of tracts with a 95% interval from a bootstrap over counties; the benchmark applies each condition's observed share of high tracts as if the conditions were unrelated across tracts.Note: The count is a count of tract prevalences above a cutoff, not a count of conditions in any person. PLACES values are modeled estimates. A model fitted to survey responses is applied to each tract's census population, so every condition's estimate depends on the same age, sex, race, ethnicity and education inputs and on county and state effects, and correlations across tracts are partly built in. The tract file carries no intervals, and the intervals shown here come from a bootstrap over counties.Note: Tied values fall on one side of the cutoff, so the share of tracts that is high on one condition runs from 24.0% to 25.6%. The benchmark is a Poisson-binomial distribution with those shares. Shares of adults living in these tracts differ from shares of tracts (for example 38.1% of adults live in tracts that are high on none).Source: CDC PLACES 2024 release (BRFSS 2022) and ADI 2023; modeled estimates, unweighted (each tract counted once)

The cutoff is a choice, and the numbers depend on it. By construction a tract is high on 2.3 conditions on average at the chapter's cutoff, and ties at the cutoff move the share of tracts that is high on a single condition between 24.0% and 25.6%. Figure 4 repeats the count with other cutoffs. With the stricter cutoff of the highest 10% of tracts, 64.9% of tracts are high on none and 14 are high on all nine, against 0.0 expected for unrelated conditions. With the looser cutoff of the highest 33.3% the figures are 23.4%, 2,679 and 4.5. The number of tracts that are high on all nine exceeds the unrelated benchmark under every cutoff, and the share that is high on none depends strongly on the cutoff.

Figure 4

Tracts are high on all nine conditions more often than independent conditions would give under every cutoff, from 14 tracts against 0.0 at the strictest to 2,679 against 4.5 at the loosest

Count of high conditions per tract when a tract is high on a condition if it is in the highest 10%, 20%, 25% or 33.3% of tracts, 82,781 tracts, 2022

A tract is high on a condition if it is in the	Mean high conditions	Tracts high on none (%)	Tracts high on five or more (%)	Expected if independent (%)	Tracts high on all nine	Expected if independent
Highest 10% of tracts	0.90	64.9	6.2	0.1	14	0.00
Highest 20% of tracts	1.80	42.7	15.8	2.0	360	0.04
Highest 25% of tracts (used in this chapter)	2.26	34.0	21.3	4.9	892	0.32
Highest third of tracts	3.02	23.4	30.9	14.8	2,679	4.47

Show data

Data for Tracts are high on all nine conditions more often than independent conditions would give under every cutoff, from 14 tracts against 0.0 at the strictest to 2,679 against 4.5 at the loosest
all9	cutoff	exp9	ge5	ind_ge5	mean	none
14	Highest 10% of tracts	0	6.20	0.100	0.900	64.90
360	Highest 20% of tracts	0.040	15.80	2	1.80	42.70
892	Highest 25% of tracts (used in this chapter)	0.320	21.30	4.90	2.26	34
2679	Highest third of tracts	4.47	30.90	14.80	3.02	23.40

Each row moves the cutoff and repeats the count. The share of tracts that are high on none runs from 23.4% at the loosest cutoff to 64.9% at the strictest, and the observed number of tracts high on all nine exceeds the number expected for independent conditions in every row.Note: The count is a count of tract prevalences above a cutoff, not a count of conditions in any person. PLACES values are modeled estimates. A model fitted to survey responses is applied to each tract's census population, so every condition's estimate depends on the same age, sex, race, ethnicity and education inputs and on county and state effects, and correlations across tracts are partly built in. The tract file carries no intervals, and the intervals shown here come from a bootstrap over counties.Source: CDC PLACES 2024 release (BRFSS 2022) and ADI 2023; modeled estimates, unweighted (each tract counted once)

The count rises across the deprivation index, and the overlap does not strengthen

Figure 5 orders tracts on the ADI and shows the mean count in each tenth. It is 0.5 (95% CI 0.4 to 0.6) in the least deprived tenth and 5.7 (95% CI 5.5 to 5.8) in the most deprived, a difference of 5.2 (95% CI 5.0 to 5.4), and it rises at every step, with the two largest steps at the deprived end. The share of adults who live in tracts high on 5 or more conditions is 0.8% (95% CI 0.3% to 1.6%) in the least deprived tenth and 72.5% (95% CI 69.9% to 75.1%) in the most deprived. Part of this gradient is mechanical. The ADI is built from ACS data on income, education, employment and housing, and the PLACES model applies survey responses to each tract's census population using inputs from the same census sources, so conditions and index move together partly by design. The gradient of each condition taken one at a time belongs to Chapter 5.1, Health behaviors and the supply of care, and the count here adds only that the conditions which rise are high in the same tracts.

Figure 5

Weighted by adults, a tract in the most deprived tenth is high on 5.7 of nine conditions on average, against 0.5 in the least deprived tenth

Population-weighted mean number of the nine conditions that are high, by tenth of tracts on the ADI national percentile, 82,781 tracts, 2022

Mean number of high conditions (of nine)

Show data

Data for Weighted by adults, a tract in the most deprived tenth is high on 5.7 of nine conditions on average, against 0.5 in the least deprived tenth
conf_high	conf_low	decile	estimate	measure	n_units
0.550	0.380	1	0.450	count	8,274
0.700	0.550	2	0.620	count	8,118
1	0.780	3	0.880	count	8,511
1.10	0.920	4	1.01	count	8,261
1.39	1.19	5	1.29	count	8,189
1.83	1.60	6	1.71	count	8,318
2.46	2.22	7	2.34	count	8,305
3.34	3.07	8	3.21	count	8,238
4.49	4.21	9	4.35	count	8,354
5.81	5.52	10	5.66	count	8,213
1.57	0.310	1	0.800	share	8,274
1.62	0.850	2	1.19	share	8,118
3.33	2.08	3	2.67	share	8,511
4.93	3.49	4	4.19	share	8,261
7.79	5.88	5	6.79	share	8,189
12.89	10.27	6	11.53	share	8,318
20.66	17.07	7	18.80	share	8,305
33.81	29.22	8	31.48	share	8,238
51.73	46.65	9	49.11	share	8,354
75.09	69.88	10	72.53	share	8,213

The line rises at every step, and the last two steps are the largest. Each point is the mean for a tenth of tracts with a 95% interval from a bootstrap over counties. The second view plots the share of adults living in tracts that are high on 5 or more of the nine, which runs from 0.8% to 72.5%.Note: PLACES values are modeled estimates. A model fitted to survey responses is applied to each tract's census population, so every condition's estimate depends on the same age, sex, race, ethnicity and education inputs and on county and state effects, and correlations across tracts are partly built in. The tract file carries no intervals, and the intervals shown here come from a bootstrap over counties. The ADI is built from American Community Survey data that overlap the model's inputs, so part of each association with it is mechanical.Note: Tenths hold equal numbers of tracts except where tied ranks fall in one tenth. Values are crude prevalence among adults, not adjusted for age, so the age mix of a tract moves them.Source: CDC PLACES 2024 release (BRFSS 2022) and ADI 2023; modeled estimates, population-weighted (adults in each tract)

The conditions do not go together more closely where deprivation is higher. Inside the least deprived tenth the mean correlation among the nine is +0.40 (95% CI 0.37 to 0.44), and inside the most deprived tenth it is +0.34 (95% CI 0.32 to 0.36), a difference of +0.06 (95% CI 0.03 to 0.11) with the least deprived tenth higher. Inside the second to the ninth tenth it lies between +0.35 and +0.36. After the mean of each ADI tenth is subtracted, the average correlation across all tracts is +0.36 (95% CI 0.35 to 0.38), against +0.49 before. Holding the ADI tenth fixed therefore removes part of the overlap, and a correlation remains in every tenth. The count of high conditions rises with deprivation, while the overlap among the conditions inside a tenth is smaller in the most deprived tenth than in the least deprived.

An approximation of the model's census inputs leaves most of the overlap in place

Part of every correlation among the PLACES estimates in this chapter is built into the estimates. PLACES fits a model to survey responses, using the age, sex, race, ethnicity and education of each respondent, a county measure of poverty, and state and county effects, and applies it to the census population of each tract.[1] Conditions that are modeled from the same inputs move together wherever the inputs move, whatever the residents do. Within a county the model gives tracts different values only through the composition of their residents.

The data cannot take the model apart, but the ACS can approximate its inputs. For each condition, a weighted regression on 8 ACS tract shares, their squares and pairwise products, a county poverty rate and state effects reproduces 79% to 90% of the variance of tract prevalence, counting the part the state effects carry. The average correlation among the residuals is +0.47 (95% CI 0.46 to 0.48), against +0.49 (95% CI 0.48 to 0.51) for the conditions themselves on the same tracts and +0.49 on all tracts (Figure 6). With county effects in place of state effects and county poverty, the fit reproduces 62% to 85% of the variance within counties and the average correlation among the residuals is +0.52 (95% CI 0.51 to 0.52), higher than the raw value on the same tracts. Gradient-boosted trees fitted to the same inputs, scored on counties they did not see, reproduce 75% to 90% of the variance and leave a mean correlation of +0.45 (95% CI 0.44 to 0.47). Holding the state fixed gives +0.48 (95% CI 0.47 to 0.50) and holding the county fixed gives +0.43 (95% CI 0.42 to 0.43).

Figure 6

An approximation of the model's census inputs leaves the average correlation among the nine conditions at +0.47, against +0.49 in the raw data

Mean of the 36 pairwise population-weighted correlations among the nine conditions across 82,781 tracts, 2022, under seven comparisons

Compared within groups of tracts
After approximating the model's inputs

Show data

Data for An approximation of the model's census inputs leaves the average correlation among the nine conditions at +0.47, against +0.49 in the raw data
conf_high	conf_low	estimate	group	n	stage
0.511	0.477	0.495	Compared within groups of tracts	82,781	All tracts
0.498	0.471	0.484	Compared within groups of tracts	82,781	Within states
0.435	0.418	0.427	Compared within groups of tracts	82,781	Within counties
0.377	0.346	0.361	Compared within groups of tracts	82,781	Within tenths of the ADI
0.476	0.460	0.468	After approximating the model's inputs	82,578	Census shares, county poverty and state effects removed
0.521	0.511	0.516	After approximating the model's inputs	82,578	Census shares removed within counties
0.469	0.441	0.455	After approximating the model's inputs	82,578	Flexible fit (trees) on the same inputs and state

The top four dots compare tracts as they are and within states, counties and ADI tenths, and the bottom three remove an approximation of the model's inputs; all seven stay well right of zero. Each dot averages the 36 pairwise correlations after the stated step, with a 95% interval from a bootstrap over counties that does not refit the adjustment. The fits are an approximation of the PLACES model and not the model.Note: The approximation regresses each condition on 8 ACS tract shares (ages 65 and over and under 18, white, Black, Asian and Hispanic residents, and adults with a college degree and with no more than high school), with squares and pairwise products, and removes state or county effects. It uses a county poverty rate in place of the model's measure of county poverty. The ACS file has no sex shares, and the model also uses fitted county and state effects for each condition. The fit explains 79% to 90% of the variance of each condition, counting the state effects, and the tree fit 75% to 90% out of sample. With county effects the fit explains 62% to 85% of the variance within counties. The three adjusted rows leave out the 203 tracts that lack an ACS input, among them the 201 tracts of the District of Columbia, whose ACS file lacks three of the shares; on the tracts they use, the raw average correlation is +0.49.Note: PLACES values are modeled estimates. A model fitted to survey responses is applied to each tract's census population, so every condition's estimate depends on the same age, sex, race, ethnicity and education inputs and on county and state effects, and correlations across tracts are partly built in. The tract file carries no intervals, and the intervals shown here come from a bootstrap over counties.Source: CDC PLACES 2024 release (BRFSS 2022), ADI 2023 and ACS 2019-2023; modeled estimates, population-weighted (adults in each tract)

These numbers show that the inputs which can be measured here do not remove the overlap. They do not show that the overlap is real. The fits approximate the model and are not the model. The ACS file has no sex shares and no age groups beyond the shares under 18 and at 65 and over, the model works from individual respondents and not tract shares, and its county and state effects are fitted to each condition separately from the same survey. Whatever the conditions share and the fit misses, such as the age mix inside the older group, stays in the residuals and keeps them correlated. What the comparison supports is narrower. The shared part of the PLACES estimates is not a simple function of the census shares that were available, and the data cannot say how much of it comes from the model's other inputs and how much from the neighborhoods.

Where tracts are high on many conditions

Tracts that are high on many conditions are not spread evenly. Of the variance of the count, 22.4% lies between states and 29.8% between counties within states, a total of 52.2% between counties, and 47.8% lies between tracts within counties. The within-county part reflects the composition of residents, because the model gives tracts in one county different values only through that composition. Figure 7 maps, for each county, the share of adults who live in tracts that are high on 5 or more conditions. In 1,229 of the 3,128 counties with a value, 50% or more of adults do. By Census region the share is 25.8% of adults in the South, 20.2% in the Midwest, 8.5% in the Northeast and 7.0% in the West, against 17.1% nationally, and in West Virginia it is 81.7%. The counties with large shares cluster in the South and in parts of the Midwest.

Figure 7

In 1,229 counties, 50% or more of adults live in tracts that are high on 5 or more conditions, and 25.8% of adults in the South do, against 7.0% in the West

Share of adults living in tracts that are high on 5 or more of the nine conditions, by county, population-weighted, 2022

Loading county values

Adults in tracts high on five or more conditions (percent of adults)

Fewer adults in such tractsMore adults in such tracts

No data (3,143 counties)

Select a class to show only those places.

Find a county

Hover or tap a county for its value, or search by name. Zoom with the buttons on the map, then drag to move it.

Counties toward the high end of the key hold many adults in tracts that are high on several conditions at once. They cluster in the South and in parts of the Midwest. The map plots, for each county, the share of its adults who live in tracts that are high on five or more of the nine conditions; hover for the number of tracts and the mean count.Note: PLACES values are modeled estimates. A model fitted to survey responses is applied to each tract's census population, so every condition's estimate depends on the same age, sex, race, ethnicity and education inputs and on county and state effects, and correlations across tracts are partly built in. The tract file carries no intervals, and the intervals shown here come from a bootstrap over counties.Note: A county's share is a full count of its tracts and moves in large steps when the county has few tracts: 990 counties have fewer than 5 tracts. Gray counties have no tract in the sample. Puerto Rico is not covered by PLACES and is not drawn. Alaska is drawn, and Connecticut appears under its eight legacy counties, to which its tracts are assigned.Source: CDC PLACES 2024 release (BRFSS 2022), ADI 2023 and ACS 2019-2023; modeled estimates, population-weighted (adults in each tract)

Life expectancy is shorter in tracts that are high on more conditions

Death records give a check that does not depend on the PLACES model. USALEEP life expectancy at birth, built from deaths in 2010 to 2015, is 80.2 (95% CI 80.0 to 80.4) years in tracts that are high on none of the nine and 74.1 (95% CI 73.9 to 74.3) years in tracts that are high on eight, a gap of 6.1 (95% CI 5.8 to 6.3) years (Figure 8). In tracts that are high on all nine it is 75.2 (95% CI 74.9 to 75.4) years, 1.0 (95% CI 0.8 to 1.3) years above tracts that are high on eight, and the data here do not say why. Those tracts differ in composition: the share of residents aged 65 and over is 22.6% (95% CI 22.3% to 22.9%) in tracts high on all nine against 18.9% (95% CI 18.6% to 19.2%) at eight, and the Black alone share is 2.7% (95% CI 2.2% to 3.1%) against 14.9% (95% CI 12.9% to 16.9%). Across 77,934 tracts the count correlates −0.48 (95% CI −0.49 to −0.46) with life expectancy, and −0.25 (95% CI −0.27 to −0.22) once each tract is compared with the mean of its county. Within a county, tracts that are high on seven conditions differ from the county mean by −1.9 (95% CI −2.2 to −1.7) years, against +0.6 (95% CI 0.5 to 0.7) years for tracts that are high on none. Within a county the count varies only through the composition of tract residents, which also moves life expectancy, so this comparison does not isolate the conditions. Life expectancy at birth is built from death rates at each age, so the age mix of a tract does not move it the way it moves crude prevalence, which makes this a check on whether the count reflects more than age. The two sources cover different years, 2010 to 2015 and 2022, and a correlation between them does not show that the conditions shorten lives.

Figure 8

Life expectancy at birth is 6.1 years shorter in tracts that are high on eight conditions than in tracts that are high on none

Population-weighted mean tract life expectancy at birth by number of high conditions, 77,934 tracts, deaths in 2010 to 2015, institutional tracts dropped

Show data

Data for Life expectancy at birth is 6.1 years shorter in tracts that are high on eight conditions than in tracts that are high on none
conf_high	conf_low	count	estimate	measure	n_units
80.37	79.98	0	80.18	level	26,431
80	79.64	1	79.82	level	15,948
78.50	78.18	2	78.33	level	8,632
78.19	77.78	3	77.97	level	5,599
77.32	76.79	4	77.06	level	4,541
76.72	76.18	5	76.45	level	4,162
76.38	75.96	6	76.16	level	4,602
74.84	74.35	7	74.59	level	4,248
74.30	73.94	8	74.12	level	2,898
75.41	74.93	9	75.17	level	873
0.660	0.460	0	0.560	dev	26,417
0.800	0.610	1	0.700	dev	15,927
−0.230	−0.480	2	−0.350	dev	8,612
−0.180	−0.560	3	−0.360	dev	5,567
−0.770	−1.25	4	−1	dev	4,511
−1.02	−1.50	5	−1.26	dev	4,135
−0.980	−1.29	6	−1.14	dev	4,572
−1.70	−2.16	7	−1.94	dev	4,219
−1.60	−2	8	−1.80	dev	2,889
−0.290	−0.640	9	−0.470	dev	871

The line falls from 80.2 years where no condition is high to 74.1 years where eight are high, and ends higher at nine. Life expectancy comes from death records for 2010 to 2015 and is not a PLACES estimate; the count comes from the 2022 PLACES release. The second view subtracts each tract's county mean to compare tracts within counties.Note: Tract life expectancy covers 2010 to 2015 and the conditions cover 2022, so the two describe different years. USALEEP estimates were area-weighted from 2010 tracts onto 2020 tracts and suit comparisons of groups of tracts, not single tracts. The within-county view leaves out counties with a single USALEEP tract in the sample. Connecticut tracts carry planning-region codes in PLACES, which serve as its counties for the bootstrap.Note: Within a county the count varies only through the composition of tract residents, which also moves life expectancy. Life expectancy is not affected by the age mix of a tract the way crude prevalence is, which makes it a check on whether the count reflects more than age. The count itself comes from PLACES modeled estimates and carries the caveats in the notes to the earlier figures.Source: CDC PLACES 2024 release (BRFSS 2022), USALEEP life expectancy 2010-2015 (area-weighted onto 2020 tracts) and ACS 2019-2023; modeled estimates, population-weighted (ACS tract population)

What a randomized offer to move shows about two of the conditions

Causal evidenceRandomized housing-voucher experiment

Moving to Opportunity and measured obesity and diabetes

The tract patterns above do not say whether living in a high-poverty neighborhood changes health. One randomized study measured two of the nine conditions, with a definition that matches PLACES for one and not for the other. From 1994 to 1998 the Department of Housing and Urban Development randomly offered women with children in public housing in high-poverty tracts one of three things: a voucher usable only in low-poverty tracts, with moving counseling; an ordinary voucher; or no new offer. Of the women offered the low-poverty voucher, 48% used it to move, and every estimate below compares women by the offer they were randomized to (intention-to-treat) and not by whether they moved. A survey from 2008 to 2010 measured height, weight and glycated hemoglobin.[4]

Compared with the group offered no voucher, the group offered the low-poverty voucher differed in the prevalence of extreme obesity (a body mass index of 40 or more) by −3.4 (95% CI −6.4 to −0.4) points and in the prevalence of glycated hemoglobin of 6.5% or more by −4.3 (95% CI −7.8 to −0.8) points, intervals that exclude zero. Obesity as PLACES defines it, a body mass index of 30 or more, differed by −1.2 (95% CI −5.4 to 3.0) points, an interval that includes zero. The glycated hemoglobin cutoff counts diabetes that has not been diagnosed, while PLACES counts diagnosed diabetes. The group offered ordinary vouchers did not differ significantly from the group offered nothing.

The study measured two conditions and not a count of conditions, enrolled volunteers for a program, and does not show what a move would do to the number of conditions that are high in a tract.

What these data cannot show

A count of high conditions describes a tract and not a person. A correlation across places can differ widely in size from the correlation among the people who live in them,[5] so the overlap here cannot be read as people carrying several conditions at once, and no figure in this chapter says how many conditions an adult has. Drawing that conclusion from place data is ecological inference. The count of conditions per person comes from surveys that ask each adult, and Chapter 5.2, Health behaviors, person by person uses one for behaviors.

The PLACES values are modeled estimates, the tract file has no intervals, and the bands in this chapter describe the sampling of counties and not the model's own uncertainty. CDC advises against using the estimates to rank the overall health of an area or to track change at the local level, and this chapter does neither.[6] It does compare conditions with each other and with the ADI, and every such comparison carries the shared inputs described above. The approximations show that measured census shares do not remove the overlap, but they cannot separate what the model builds in from what the neighborhoods contribute. The staged tract file holds crude prevalence only, so age is part of every pattern, most visibly for cancer, arthritis and heart disease.

High is relative. A tract in the highest 25% of tracts for a condition is high compared with other tracts and not high by any clinical standard, and the cutoff moves every count, as the table shows. High blood pressure is left out, and outside Florida its mean correlation with the nine conditions is +0.65 (95% CI 0.64 to 0.67), against +0.51 among the nine on the same tracts, so it follows the same shared pattern. Counties are the clusters of the bootstrap, 3,129 of them, with Connecticut's planning regions standing in for its counties, while the map draws Connecticut as its eight legacy counties. The map has 15 gray counties with no tract in the sample, and 990 counties have fewer than 5 tracts, so their shares move in large steps. Puerto Rico is not covered by PLACES and is not drawn.

The conditions are measured once, in 2022, so the data say nothing about which condition comes first, whether tracts move into or out of the high group, or whether living in a tract changes any of the prevalences. The ADI describes 2019 to 2023 and the life expectancy check describes 2010 to 2015, so the sources do not describe one moment.

Data sources

CDC PLACESCDC PLACES 2024 release, census tracts, crude prevalence among adults of nine conditions (BRFSS 2022)Years: 2024 release (tracts); 2025 release (ZCTAs, 2023 data year) · Weights: adult population of the tract · License: Public domain (CDC, U.S. Department of Health and Human Services)
Area Deprivation IndexNeighborhood Atlas Area Deprivation Index, 2023, national percentileYears: 2023 release (version 4.0.1; built on ACS 2019-2023) · Weights: adult population of the tract · License: Free for non-commercial use; cite the Neighborhood Atlas and Kind and Buckingham
American Community SurveyACS 2019-2023 tract shares used to approximate the model's inputsYears: 2019-2023 (5-year) · Weights: adult population of the tract · License: Public domain (U.S. Census Bureau)
USALEEP life expectancyNCHS small-area life expectancy at birth, 2010-2015, 2020 tractsYears: 2010-2015 · Weights: ACS 2019-2023 tract population · License: Public domain (NCHS, CDC)
Published estimates from cited studiesEstimates quoted in the causal box (Ludwig and coauthors 2011)Years: as reported in each study · Weights: as reported in the study · License: Quoted from published research for commentary

All data sources

Methods

The chapter compares modeled prevalences of nine chronic conditions across census tracts, from the CDC PLACES 2024 release, with correlations, means and shares of adults weighted by the adult population of each tract. A tract is high on a condition when its prevalence is in the highest 25% of tracts, and the count of high conditions describes tracts and not people. The estimates come from a model that fits every condition to shared inputs, so part of the correlation among conditions is built in; the adjustments here approximate those inputs from census data and are not the model. Life expectancy covers deaths in 2010 to 2015, a different period from the PLACES estimates.

Which tracts and conditions are compared

The tract file of CDC PLACES (2024 release) gives one modeled prevalence per condition per tract. The chapter uses nine health outcome measures that cover all adults and have an estimate in every state: arthritis, cancer (non-skin) or melanoma, current asthma, coronary heart disease, COPD, depression, diagnosed diabetes, obesity and stroke.

The sample is the tracts with all nine conditions and an ADI rank. The approximations of the model's inputs use the tracts with every ACS input, which leaves out the tracts of the District of Columbia and a few others.

Figure explained here: Figure 1

What counts as high on a condition

A tract is high on a condition when its prevalence is in the highest 25% of tracts, with tied values at their mean rank, and the count is the number of conditions on which a tract is high; it describes tracts and not people.

The benchmark for independent conditions is a Poisson-binomial distribution with each condition's observed share of high tracts.

Figures explained here: Figure 2, Figure 3 and Figure 4

How tracts are weighted and where life expectancy comes from

Correlations, means and shares of adults are weighted by the tract's adult population; counts and shares of tracts count each tract once.

Life expectancy is the USALEEP estimate for 2010 to 2015 on 2020 tracts, weighted by ACS 2019-2023 tract population.

Figures explained here: Figure 5 and Figure 8

How the model's inputs are approximated

The adjustments for the model's inputs are weighted least-squares fits of each condition on eight ACS tract shares with squares and pairwise products, with state effects and a county poverty rate (or county effects), and a gradient-boosted-trees fit with folds by county; they are fitted once and not refitted inside the bootstrap, so their intervals condition on the fit.

They approximate the PLACES model, which uses individual age, sex, race, ethnicity and education, county poverty, and county and state effects.

Figure explained here: Figure 6

How uncertainty is measured

The variance split of the count is a weighted decomposition of every tract and has no interval.

Intervals are 95% percentile intervals from a bootstrap that draws whole counties with replacement (3,129 clusters; 2,000 replicates, seed 12345), with the same draws for every statistic, so a difference has an interval from shared draws.

Figure explained here: Figure 7

References

Greenlund, Kurt J.; Lu, Hua; Wang, Yan; Matthews, Kevin A.; LeClercq, Jennifer M.; Lee, Benjamin; Carlson, Susan A. (2022). PLACES: Local Data for Better Health. Preventing Chronic Disease 19: 210459. doi:10.5888/pcd19.210459
Ward, Brian W.; Schiller, Jeannine S. (2013). Prevalence of Multiple Chronic Conditions Among US Adults: Estimates From the National Health Interview Survey, 2010. Preventing Chronic Disease 10: 120203. doi:10.5888/pcd10.120203
Kind, Amy J. H.; Buckingham, William R. (2018). Making Neighborhood-Disadvantage Metrics Accessible: The Neighborhood Atlas. New England Journal of Medicine 378(26): 2456-2458. doi:10.1056/NEJMp1802313
Ludwig, Jens; Sanbonmatsu, Lisa; Gennetian, Lisa; Adam, Emma; Duncan, Greg J.; Katz, Lawrence F.; Kessler, Ronald C.; Kling, Jeffrey R.; Lindau, Stacy Tessler; Whitaker, Robert C.; McDade, Thomas W. (2011). Neighborhoods, Obesity, and Diabetes: A Randomized Social Experiment. New England Journal of Medicine 365(16): 1509-1519. doi:10.1056/NEJMsa1103216
Robinson, W. S. (1950). Ecological Correlations and the Behavior of Individuals. American Sociological Review 15(3): 351-357. doi:10.2307/2087176
Centers for Disease Control and Prevention (2024). PLACES Frequently Asked Questions: Using the Data. CDC PLACES website, page last reviewed August 29, 2024. www.cdc.gov/places/faqs/using-data/index.html

How to cite

@misc{singh2026multiplechronicconditions,
  author = {Singh, Vishal},
  title = {Multiple chronic conditions},
  year = {2026},
  url = {https://vishalsingh.org/zipcode/multiple-chronic-conditions},
  note = {In ZIP Code Destiny. Published 9 October 2026.}
}