Income Inequality and CVD Risk Factors: Descriptives

Project

We aim to examine the relation of neighborhood-level income inequality with psychosocial and biomedical risk factors of CVD within a racially and socioeconomically diverse cohort of adults in Baltimore, Maryland.

Psychosocial risk factors include: depressive symptoms, anger, anxiety, and perceived stress.

Biomedical risk factors include: hypertension (or blood pressure), hypercholesterolemia (as measured by cholesterol ratio or cholesterol levels), diabetes (or hemoglobin A1C), and obesity (body mass index).

Covariates include: age, sex, race, and median neighborhood income. (I also plan on adding the below medical clusters, medications and substance use)

The interaction of neighborhood-level income inequality and individual income will also be examined to investigate whether individual-level income level (or poverty status) moderates the relation between neighborhood-level income inequality and CVD risk factors.

This Document

These are descriptives of the HANDLS Wave 1 MRV dataset with exclusions already taken out, and with listwise complete.

Exclusions are:

Neurological (Stroke, TIA, Brain Cancer, Dementia, Alzheimers, Parkinsons Disease, Multiple Sclerosis, Epilepsy)
Heart Failure
Schizophrenia
Dialysis (one person)

Please let me know if anything looks funky.

Updates

2014-05-05: Added Transformation, Discussion, and Frequency Table sections.
2014-05-01: Using imputed values of CES. Added frequency table for glucose and sqrt-transformed glucose histogram and qqplot.
2014-05-01: Displaying descriptives for only listwise complete cases. Added histograms and QQPlots for log-transformed continuous variables.
2014-04-30: Using imputed values where available and added qqplots

Transformation

None: CES, SBP, DBP, BMI, LDL
log10-transform: Glucose, HDL, Total Cholesterol (Run analyses with and without transformations)
Possibly dichotomize: Anxiety, PSS, Anger

Discussions

Alan: In general, tests for normality are too stringent for most purposes with linear regression, which is "robust" when the errors of prediction are somewhat non-normally distributed.

The Q-Q plots provide a visual reference with no test so it's a judgment call. Mostly, if the sample quantiles follow the theoretical quantiles for most of the range (say more than two-thirds or so) then it's probably ok.

Glucose definitely needs a transformation

Education is a little weird, but probably reflects an underlying categorical pile up at years when degrees are awarded (HS, AA, Bachelor's, etc).

For as much as it's skewed, depression didn't look too bad, but probably needs transforming.

Not sure what's going on for anxiety, but probably there are lots of folks who aren't anxious. Anxiety may be a candidate for dichotomization. To some extent, similar comments apply to perceived stress and anger.

The rest, maybe except for HDL and cholesterol, seem pretty reasonable.

Shari: Thanks Alan! This is very helpful. I was kind of on the fence re. HDL and cholesterol as well. I generally prefer raw units for ease of interpretation, but wondered about running those models with and without transformation. (The analyses with transformation could be acknowledged in a footnote if they're the same)

I'd be inclined to leave CES-D in raw units. Re. anxiety, PSS, and anger, can you please remind me if you're using the "negative affect" composite in these analyses? (Answer: NO) In general, I don't tend to like dichotomizing these continuous outcomes.

Descriptives

**Total Sample Descriptives (listwise n= 1891 )**
Variable	N	Mean or Percent	SD	Min	Max	Skew	Kurtosis
Baseline Age (years)	1891	47.5	9.2	30.0	64.0	-0.1	-1.0
Sex (% Male)	1891	43.4
Race (% Black)	1891	58.5
Poverty Status	1891	41.8
Education (years)	1891	12.6	3.0	1.0	21.0	0.6	0.8
Depressive Symptoms (Imputed)	1891	14.3	11.1	0.0	59.0	1.0	0.5
Anxiety	1891	2.9	3.2	0.0	10.0	0.8	-0.7
Perceived Stress	1891	5.4	3.3	0.0	16.0	0.3	-0.5
Anger	1891	9.9	6.3	0.0	28.0	0.5	-0.5
Hypertension (based on self-report, medication or clinic BP)	1891	41.5
SBP (Imputed)	1891	120.0	17.4	79.0	190.0	0.7	0.6
DBP (Imputed)	1891	72.7	10.7	39.0	110.0	0.2	-0.1
Diabetes (based on self-report, medication or labs)	1891	14.5
Glucose (Imputed)	1891	103.6	40.0	33.0	554.0	4.4	25.3
Body Mass Index (kg/m2)	1891	29.8	7.7	14.7	61.9	0.9	0.8
Obesity (BMI >= 30)	1891	41.9
LDL (Imputed)	1891	109.6	34.9	14.0	265.6	0.5	0.7
LDL >= 160	1891	7.9
HDL (Imputed)	1891	53.2	17.3	11.0	175.0	1.6	5.5
HDL < 40	1891	19.6
Total Cholesterol (Imputed)	1891	187.3	42.4	83.0	651.0	1.3	8.6
Total Cholesterol >= 240	1891	9.6
CVD (%)	1891	3.8
Metabolic Diseases (%)	1891	25.3
Psychotropic Medication Use (%)	1891	14.4
Anti-hypertensive Medication Use (%)	1891	26.4
Diabetes Medication Use (%)	1891	8.7
Hypercholesterolemia Medication Use (%)	1891	11.4
Current Smoking (%)	1891	19.4
Current Alcohol Use (%)	1891	54.9
Current Illicit Drug Use (%)	1891	16.3
Gini Index (x 100)	1891	44.2	6.7	35.2	60.5	0.9	0.1
Census Tract Median Household Income (in tens of thousands)	1891	3.6	1.2	1.2	8.8	1.2	3.6
Census Tract Median Family Income (in tens of thousands)	1891	4.0	1.2	1.3	8.2	0.8	1.0

Histograms

plot of chunk unnamed-chunk-2

Frequency Tables

**Glucose (imputed) frequency**
v	Freq	CumSum	CumPercent
33	1	1	0.0
42	1	2	0.1
56	1	3	0.2
60	1	4	0.2
61	1	5	0.3
62	1	6	0.3
63	2	8	0.4
65	2	10	0.5
66	1	11	0.6
67	2	13	0.7
68	1	14	0.7
69	2	16	0.8
70	6	22	1.2
71	4	26	1.4
72	2	28	1.5
73	8	36	1.9
74	6	42	2.2
75	10	52	2.8
76	16	68	3.6
77	18	86	4.5
78	26	112	5.9
79	25	137	7.2
80	27	164	8.7
81	20	184	9.7
82	43	227	12.0
83	43	270	14.3
84	50	320	16.9
85	57	377	19.9
86	55	432	22.9
87	60	492	26.0
88	55	547	28.9
89	63	610	32.3
90	55	665	35.2
91	76	741	39.2
91.5762176757642	1	742	39.2
92	77	819	43.3
92.2076951834963	1	820	43.4
92.8391726912284	1	821	43.4
93	62	883	46.7
94	78	961	50.8
94.1021277066925	1	962	50.9
94.8980882326809	1	963	50.9
95	75	1038	54.9
95.1475846483279	1	1039	54.9
95.7790621560599	1	1040	55.0
96	63	1103	58.3
96.2868381385851	1	1104	58.4
96.410539663792	1	1105	58.4
97	51	1156	61.1
97.6734946792562	1	1157	61.2
97.890992753085	1	1158	61.2
98	61	1219	64.5
98.1812706617813	1	1220	64.5
98.5952500956846	1	1221	64.6
99	57	1278	67.6
99.2267276034166	1	1279	67.6
99.3184307868054	1	1280	67.7
99.5679272024524	1	1281	67.7
99.7854252762812	2	1283	67.8
100	40	1323	70.0
100.871663710966	1	1324	70.0
100.995365236173	1	1325	70.1
101	30	1355	71.7
101.048380291745	1	1356	71.7
101.212863310002	1	1357	71.8
101.679857799477	2	1359	71.9
101.844340817734	1	1360	71.9
102	42	1402	74.1
102.31133530721	1	1403	74.2
102.475818325466	1	1404	74.2
102.601613215906	1	1405	74.3
102.942812814942	2	1407	74.4
103	34	1441	76.2
103.988269756577	2	1443	76.3
104	26	1469	77.7
104.278547665273	1	1470	77.7
104.910025173005	1	1471	77.8
105	22	1493	79.0
105.127523246834	1	1494	79.0
105.46872284587	1	1495	79.1
105.759000754566	1	1496	79.1
105.882702279773	1	1497	79.2
106	10	1507	79.7
106.047185298029	1	1508	79.8
106.100200353602	2	1510	79.8
106.172980188469	2	1512	80.0
106.731677861334	1	1513	80.0
107	22	1535	81.2
107.02195577003	1	1536	81.2
107.145657295237	2	1538	81.3
107.653433277762	1	1539	81.4
107.994632876798	1	1540	81.4
108	14	1554	82.2
108.698890219398	1	1555	82.2
108.916388293227	1	1556	82.3
109	11	1567	82.9
109.33036772713	2	1569	83.0
109.494850745386	1	1570	83.0
109.547865800959	1	1571	83.1
109.712348819215	1	1572	83.1
109.889065399994	1	1573	83.2
109.961845234862	1	1574	83.2
110	11	1585	83.8
110.303044833898	1	1586	83.9
110.93452234163	1	1587	83.9
111	9	1596	84.4
111.152020415459	1	1597	84.5
111.442298324155	1	1598	84.5
111.565999849362	1	1599	84.6
111.783497923191	1	1600	84.6
111.856277758058	1	1601	84.7
112	11	1612	85.2
112.020760776314	1	1613	85.3
112.238258850143	1	1614	85.3
112.993437883082	1	1615	85.4
113	19	1634	86.4
113.119232773522	1	1635	86.5
113.968208355083	1	1636	86.5
114	7	1643	86.9
115	12	1655	87.5
115.862640878279	1	1656	87.6
116	5	1661	87.8
117	8	1669	88.3
118	10	1679	88.8
119	7	1686	89.2
120	14	1700	89.9
121	7	1707	90.3
122	4	1711	90.5
123	2	1713	90.6
124	5	1718	90.8
125	3	1721	91.0
126	4	1725	91.2
127	2	1727	91.3
128	3	1730	91.5
129	4	1734	91.7
130	2	1736	91.8
131	1	1737	91.9
132	3	1740	92.0
133	4	1744	92.2
134	1	1745	92.3
135	2	1747	92.4
136	2	1749	92.5
137	4	1753	92.7
139	2	1755	92.8
140	2	1757	92.9
141	2	1759	93.0
142	1	1760	93.1
143	3	1763	93.2
144	2	1765	93.3
145	1	1766	93.4
146	3	1769	93.5
147	2	1771	93.7
150	7	1778	94.0
151	2	1780	94.1
153	1	1781	94.2
154	1	1782	94.2
159	1	1783	94.3
160	1	1784	94.3
161	2	1786	94.5
162	1	1787	94.5
163	1	1788	94.5
164	2	1790	94.7
165	2	1792	94.8
166	2	1794	94.9
167	1	1795	94.9
168	2	1797	95.0
169	1	1798	95.1
171	3	1801	95.2
175	1	1802	95.3
176	1	1803	95.3
177	2	1805	95.5
178	1	1806	95.5
179	1	1807	95.6
180	2	1809	95.7
182	1	1810	95.7
183	2	1812	95.8
184	1	1813	95.9
185	1	1814	95.9
189	2	1816	96.0
192	1	1817	96.1
196	1	1818	96.1
197	1	1819	96.2
198	1	1820	96.2
199	1	1821	96.3
201	1	1822	96.3
202	1	1823	96.4
204	1	1824	96.5
207	1	1825	96.5
209	1	1826	96.6
211	1	1827	96.6
212	1	1828	96.7
213	1	1829	96.7
221	1	1830	96.8
225	2	1832	96.9
226	1	1833	96.9
228	2	1835	97.0
229	2	1837	97.1
230	1	1838	97.2
233	1	1839	97.2
234	1	1840	97.3
235	1	1841	97.4
236	1	1842	97.4
241	1	1843	97.5
244	1	1844	97.5
247	1	1845	97.6
251	1	1846	97.6
252	2	1848	97.7
255	1	1849	97.8
257	1	1850	97.8
258	1	1851	97.9
259	1	1852	97.9
260	2	1854	98.0
261	1	1855	98.1
264	1	1856	98.2
267	1	1857	98.2
268	1	1858	98.2
269	1	1859	98.3
270	1	1860	98.4
271	1	1861	98.4
275	2	1863	98.5
283	1	1864	98.6
284	1	1865	98.6
285	1	1866	98.7
287	1	1867	98.7
290	1	1868	98.8
292	1	1869	98.8
293	2	1871	98.9
294	1	1872	99.0
298	1	1873	99.0
299	1	1874	99.1
306	1	1875	99.2
311	1	1876	99.2
312	1	1877	99.3
316	1	1878	99.3
323	1	1879	99.4
331	1	1880	99.4
337	1	1881	99.5
346	1	1882	99.5
351	2	1884	99.6
363	2	1886	99.7
366	1	1887	99.8
384	1	1888	99.8
393	1	1889	99.9
430	1	1890	100.0
554	1	1891	100.0

**Education frequency**
v	Freq	CumSum	CumPercent
1	2	2	0.1
2	3	5	0.3
3	1	6	0.3
4	2	8	0.4
5	2	10	0.5
6	11	21	1.1
7	26	47	2.5
8	60	107	5.7
9	100	207	10.9
10	170	377	19.9
11	201	578	30.6
12	650	1228	64.9
13	137	1365	72.2
14	177	1542	81.5
15	46	1588	84.0
16	35	1623	85.8
17	45	1668	88.2
18	127	1795	94.9
19	46	1841	97.4
20	31	1872	99.0
21	19	1891	100.0

**CES (imputed) frequency**
v	Freq	CumSum	CumPercent
0	69	69	3.6
1	59	128	6.8
2	68	196	10.4
3	98	294	15.6
4	85	379	20.0
5	82	461	24.4
6	97	558	29.5
7	82	640	33.8
8	74	714	37.8
9	81	795	42.0
10	71	866	45.8
11	56	922	48.8
11.5568172773325	1	923	48.8
11.7769244595959	1	924	48.9
12	61	985	52.1
12.0520584374252	1	986	52.1
12.4372460063862	1	987	52.2
12.5669826633045	1	988	52.2
13	62	1050	55.5
13.4528919107474	1	1051	55.6
13.6675185746217	1	1052	55.6
13.6729990930109	1	1053	55.7
13.7280258885767	1	1054	55.7
13.8325989613192	1	1055	55.8
14	58	1113	58.9
14.2782938442353	1	1114	58.9
14.5731108878512	2	1116	59.0
14.9582984568122	1	1117	59.1
15	53	1170	61.9
15.1233788435097	1	1171	61.9
15.5775231585628	1	1172	62.0
15.9627107275238	1	1173	62.0
16	43	1216	64.3
16.0727643186555	1	1217	64.4
17	61	1278	67.6
17.0884102230168	1	1279	67.6
17.3635442008461	1	1280	67.7
17.5286245875436	2	1282	67.8
17.6937049742412	1	1283	67.8
17.8037585653729	1	1284	67.9
17.9138121565046	1	1285	68.0
18	62	1347	71.2
18.1889461343339	1	1348	71.3
19	34	1382	73.1
19.2541383158718	1	1383	73.1
19.8594330670962	1	1384	73.2
20	37	1421	75.2
21	31	1452	76.8
22	30	1482	78.4
23	30	1512	80.0
24	24	1536	81.2
25	38	1574	83.2
26	35	1609	85.1
27	27	1636	86.5
28	25	1661	87.8
29	27	1688	89.3
30	17	1705	90.2
31	16	1721	91.0
32	12	1733	91.6
33	16	1749	92.5
34	15	1764	93.3
35	15	1779	94.1
36	17	1796	95.0
37	11	1807	95.6
38	15	1822	96.3
39	13	1835	97.0
40	7	1842	97.4
41	8	1850	97.8
42	5	1855	98.1
43	2	1857	98.2
44	13	1870	98.9
45	1	1871	98.9
46	4	1875	99.2
47	2	1877	99.3
48	1	1878	99.3
49	2	1880	99.4
50	3	1883	99.6
51	1	1884	99.6
52	2	1886	99.7
53	1	1887	99.8
54	1	1888	99.8
55	1	1889	99.9
57	1	1890	100.0
59	1	1891	100.0

**Anxiety frequency**
v	Freq	CumSum	CumPercent
0	699	699	37.0
1	230	929	49.1
2	163	1092	57.8
3	118	1210	64.0
4	120	1330	70.3
5	117	1447	76.5
6	105	1552	82.1
7	93	1645	87.0
8	80	1725	91.2
9	95	1820	96.2
10	71	1891	100.0

**Perceived Stress frequency**
v	Freq	CumSum	CumPercent
0	141	141	7.5
1	113	254	13.4
2	156	410	21.7
3	181	591	31.2
4	221	812	42.9
5	167	979	51.8
6	198	1177	62.2
7	152	1329	70.3
8	232	1561	82.5
9	125	1686	89.2
10	97	1783	94.3
11	42	1825	96.5
12	28	1853	98.0
13	14	1867	98.7
14	17	1884	99.6
15	4	1888	99.8
16	3	1891	100.0

**Anger frequency**
v	Freq	CumSum	CumPercent
0	70	70	3.7
1	87	157	8.3
2	70	227	12.0
3	78	305	16.1
4	125	430	22.7
5	98	528	27.9
6	119	647	34.2
7	108	755	39.9
8	135	890	47.1
9	87	977	51.7
10	120	1097	58.0
11	91	1188	62.8
12	85	1273	67.3
13	88	1361	72.0
14	68	1429	75.6
15	79	1508	79.8
16	61	1569	83.0
17	61	1630	86.2
18	36	1666	88.1
19	48	1714	90.6
20	43	1757	92.9
21	33	1790	94.7
22	36	1826	96.6
23	18	1844	97.5
24	19	1863	98.5
25	11	1874	99.1
26	7	1881	99.5
27	7	1888	99.8
28	3	1891	100.0