Poster: A multilevel analysis of neighborhood, perceived park access, and hypertension in the Philadelphia area
See our Preventing Chronic Disease Paper on 2018 data
Is there a park or other outdoor space in your neighborhood that you’re comfortable visiting during the day?:
##
## 1 2
## 0.7753094 0.2246906##
## 1 2
## 0.7802654 0.2197346##
## 1 2 3
## 0.75269411 0.17712075 0.070185132012: Have you EVER been told by a doctor or other health professional that you have high blood pressure or hypertension? 1 = yes, 2 = no, 7 = pregnancy
2015 same
2018, same question added response 1 = yes, 2 = no, 6 = borderline high, 7 = pregnancy
No pregnancy in any years, 2018 has 35 people who are borderline. They’re changed to “no” in a new variable.
The rest of the analysis has Hypertension coded as Yes/No and Parks coded as Yes/No, with both “No’s” combined in 2018 data
##
## 1 2
## 1 2489 1015
## 2 5126 1196##
## 1 2
## 1 2483 982
## 2 5175 1176##
## 1 2 3
## 0 1701 318 116
## 1 1007 316 134##
## Pearson's Chi-squared test with Yates' continuity correction
##
## data: htn.parks.2012
## X-squared = 129.97, df = 1, p-value < 2.2e-16##
## Pearson's Chi-squared test with Yates' continuity correction
##
## data: htn.parks.2015
## X-squared = 125.57, df = 1, p-value < 2.2e-16##
## Pearson's Chi-squared test
##
## data: htn.parks.2018
## X-squared = 53.076, df = 2, p-value = 2.984e-12Now set up models for all 3 years separately.
predictors from Kjelstrom, Hass, & McIntyre: Age (4 categories), Sex (2 categories), Race (3 Categories: B W O), 150 pct poverty (indicator), Diabetes (indicator), Asthma (indicator)), Eversmoke (indicator), Mental health dx (indicator), obese (indicator), Education (3 Categories), exercise 3+ days (indicator)
Results from several models per year
Models: 1. “Empty” model (a.k.a. one-way ANOVA with random effects) 2. Level 2 model 3. Full model
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: HIGHBP ~ (1 | ZIPCODE)
## Data: d.2012
##
## AIC BIC logLik deviance df.resid
## 12092.8 12107.1 -6044.4 12088.8 9404
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.1783 -0.7372 -0.6330 1.1439 1.8868
##
## Random effects:
## Groups Name Variance Std.Dev.
## ZIPCODE (Intercept) 0.1719 0.4146
## Number of obs: 9406, groups: ZIPCODE, 206
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.68478 0.04109 -16.66 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: HIGHBP ~ PHILLY + (1 | ZIPCODE)
## Data: d.2012
##
## AIC BIC logLik deviance df.resid
## 12026.9 12048.3 -6010.4 12020.9 9403
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.146 -0.715 -0.630 1.128 1.779
##
## Random effects:
## Groups Name Variance Std.Dev.
## ZIPCODE (Intercept) 0.06645 0.2578
## Number of obs: 9406, groups: ZIPCODE, 206
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.85451 0.03851 -22.192 <2e-16 ***
## PHILLYYes 0.63544 0.06576 9.663 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## PHILLYYes -0.574## Generalized linear mixed model fit by ML
## Call: glmer(formula = HIGHBP ~ HAVEPARK + AGE4CAT + RACE3 + EDUCAT3 + EVRDIABA
## + EVRASTHA + EVERSMOK + MENTAL3A + NPOV150 + OBESECAT + XHALF2 + PHILLY + (1 |
## ZIPCODE), data = d.2012, family = binomial, nAGQ = 9)
##
## Estimates of Fixed Effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -3.01017 0.16185 -18.599 < 2e-16 ***
## HAVEPARKNo 0.17160 0.06237 2.751 0.005933 **
## AGE4CAT[35,50) 0.82886 0.11675 7.099 1.25e-12 ***
## AGE4CAT[50,65) 1.85031 0.11073 16.710 < 2e-16 ***
## AGE4CAT[65,Inf] 2.73869 0.11756 23.296 < 2e-16 ***
## RACE3Black 0.61292 0.07513 8.158 3.40e-16 ***
## RACE3Other 0.13443 0.12146 1.107 0.268374
## EDUCAT3HS or Coll 0.08038 0.10745 0.748 0.454411
## EDUCAT3Col Grad and up -0.13209 0.11658 -1.133 0.257188
## EVRDIABAYes 0.82462 0.07486 11.015 < 2e-16 ***
## EVRASTHAYes 0.30232 0.07258 4.165 3.11e-05 ***
## EVERSMOKYes 0.20662 0.05323 3.882 0.000104 ***
## MENTAL3AYes 0.28795 0.06916 4.164 3.13e-05 ***
## NPOV150Yes 0.17895 0.07505 2.385 0.017099 *
## OBESECATYes 0.83292 0.05720 14.561 < 2e-16 ***
## XHALF2Yes -0.20685 0.05282 -3.916 8.99e-05 ***
## PHILLYYes 0.23639 0.06485 3.645 0.000267 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Exponentiated Fixed Effects and Confidence Bounds:
## Estimate 2.5 % 97.5 %
## (Intercept) 0.04928337 0.03588699 0.06768053
## HAVEPARKNo 1.18719812 1.05059978 1.34155690
## AGE4CAT[35,50) 2.29071193 1.82217874 2.87971812
## AGE4CAT[50,65) 6.36178654 5.12063141 7.90377686
## AGE4CAT[65,Inf] 15.46664126 12.28372925 19.47429701
## RACE3Black 1.84582127 1.59308661 2.13865093
## RACE3Other 1.14388343 0.90156788 1.45132643
## EDUCAT3HS or Coll 1.08370264 0.87790192 1.33774784
## EDUCAT3Col Grad and up 0.87626056 0.69726956 1.10119904
## EVRDIABAYes 2.28100978 1.96972738 2.64148514
## EVRASTHAYes 1.35298963 1.17358497 1.55981970
## EVERSMOKYes 1.22951886 1.10770517 1.36472833
## MENTAL3AYes 1.33369002 1.16463062 1.52729032
## NPOV150Yes 1.19596328 1.03237762 1.38546995
## OBESECATYes 2.30001481 2.05608927 2.57287862
## XHALF2Yes 0.81314442 0.73318004 0.90183012
## PHILLYYes 1.26666833 1.11548472 1.43834213
##
## Estimates of Random Effects (Covariance Components):
## Groups Name Std.Dev.
## ZIPCODE (Intercept) 0.09948
##
## Number of obs: 8582, groups: ZIPCODE, 205
##
## logLik df AIC BIC
## -4469.79 18 8975.58 9102.62Models: 1. “Empty” model (a.k.a. one-way ANOVA with random effects) 2. Level 2 model 3. Full model - boundary issue due to near complete variance accounted for at Zip level
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: HIGHBP ~ (1 | ZIPCODE)
## Data: d.2015
##
## AIC BIC logLik deviance df.resid
## 12821.4 12835.8 -6408.7 12817.4 9986
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.1496 -0.7276 -0.6355 1.1830 1.8442
##
## Random effects:
## Groups Name Variance Std.Dev.
## ZIPCODE (Intercept) 0.1333 0.365
## Number of obs: 9988, groups: ZIPCODE, 258
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.66757 0.03638 -18.35 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: HIGHBP ~ PHILLY + (1 | ZIPCODE)
## Data: d.2015
##
## AIC BIC logLik deviance df.resid
## 12769.6 12791.3 -6381.8 12763.6 9985
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.1311 -0.7351 -0.6274 1.1508 1.7993
##
## Random effects:
## Groups Name Variance Std.Dev.
## ZIPCODE (Intercept) 0.06838 0.2615
## Number of obs: 9988, groups: ZIPCODE, 258
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.82792 0.03754 -22.05 <2e-16 ***
## PHILLYYes 0.52654 0.06413 8.21 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## PHILLYYes -0.585## boundary (singular) fit: see help('isSingular')## Generalized linear mixed model fit by ML
## Call: glmer(formula = HIGHBP ~ HAVEPARK + AGE4CAT + RACE3 + EDUCAT3 + EVRDIABA
## + EVRASTHA + EVERSMOK + MENTAL3A + NPOV150 + OBESECAT + XHALF2 + PHILLY + (1 |
## ZIPCODE), data = d.2015, family = binomial, control = glmerControl(optimizer =
## "bobyqa"), nAGQ = 8)
##
## Estimates of Fixed Effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -2.59817 0.16602 -15.650 < 2e-16 ***
## HAVEPARKNo 0.09683 0.06078 1.593 0.1111
## AGE4CAT[35,50) 0.78894 0.11966 6.593 4.30e-11 ***
## AGE4CAT[50,65) 1.81830 0.11511 15.796 < 2e-16 ***
## AGE4CAT[65,Inf] 2.71599 0.12028 22.580 < 2e-16 ***
## RACE3Black 0.72238 0.07029 10.276 < 2e-16 ***
## RACE3Other 0.13348 0.10897 1.225 0.2206
## EDUCAT3HS or Coll -0.29087 0.11459 -2.538 0.0111 *
## EDUCAT3Col Grad and up -0.58098 0.12102 -4.800 1.58e-06 ***
## EVRDIABAYes 1.09776 0.07463 14.710 < 2e-16 ***
## EVRASTHAYes 0.12584 0.07059 1.783 0.0746 .
## EVERSMOKYes 0.21803 0.05229 4.170 3.05e-05 ***
## MENTAL3AYes 0.35461 0.07065 5.019 5.19e-07 ***
## NPOV150Yes 0.07356 0.07224 1.018 0.3085
## OBESECATYes 0.77240 0.05502 14.038 < 2e-16 ***
## XHALF2Yes -0.23445 0.05108 -4.590 4.43e-06 ***
## PHILLYYes 0.14368 0.05938 2.420 0.0155 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Exponentiated Fixed Effects and Confidence Bounds:
## Estimate 2.5 % 97.5 %
## (Intercept) 0.07440926 0.05374174 0.1030249
## HAVEPARKNo 1.10167842 0.97794645 1.2410652
## AGE4CAT[35,50) 2.20106265 1.74093039 2.7828090
## AGE4CAT[50,65) 6.16137473 4.91692868 7.7207828
## AGE4CAT[65,Inf] 15.11963960 11.94422178 19.1392546
## RACE3Black 2.05932976 1.79428334 2.3635281
## RACE3Other 1.14279681 0.92303368 1.4148829
## EDUCAT3HS or Coll 0.74761201 0.59721852 0.9358781
## EDUCAT3Col Grad and up 0.55935261 0.44123458 0.7090907
## EVRDIABAYes 2.99745225 2.58957638 3.4695713
## EVRASTHAYes 1.13410504 0.98756776 1.3023858
## EVERSMOKYes 1.24363024 1.12248616 1.3778488
## MENTAL3AYes 1.42562419 1.24126816 1.6373612
## NPOV150Yes 1.07632940 0.93423734 1.2400328
## OBESECATYes 2.16496670 1.94364323 2.4114924
## XHALF2Yes 0.79100270 0.71564693 0.8742932
## PHILLYYes 1.15451805 1.02767754 1.2970138
##
## Estimates of Random Effects (Covariance Components):
## Groups Name Std.Dev.
## ZIPCODE (Intercept) 0
##
## Number of obs: 9326, groups: ZIPCODE, 253
##
## logLik df AIC BIC
## -4753.67 18 9543.34 9671.87Models: 1. “Empty” model (a.k.a. one-way ANOVA with random effects) 2. Level 2 model 3. Full model - same
## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: HIGHBP2_no6 ~ (1 | ZIPCODE)
## Data: d.2018
##
## AIC BIC logLik deviance df.resid
## 9938.7 9952.5 -4967.4 9934.7 7362
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.1764 -0.8317 -0.7359 1.1460 1.4968
##
## Random effects:
## Groups Name Variance Std.Dev.
## ZIPCODE (Intercept) 0.09844 0.3138
## Number of obs: 7364, groups: ZIPCODE, 243
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.38746 0.03585 -10.81 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1## Generalized linear mixed model fit by maximum likelihood (Laplace
## Approximation) [glmerMod]
## Family: binomial ( logit )
## Formula: HIGHBP2_no6 ~ PHILLY + (1 | ZIPCODE)
## Data: d.2018
##
## AIC BIC logLik deviance df.resid
## 9923.9 9944.6 -4958.9 9917.9 7361
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -1.1835 -0.8279 -0.7321 1.1439 1.4854
##
## Random effects:
## Groups Name Variance Std.Dev.
## ZIPCODE (Intercept) 0.07192 0.2682
## Number of obs: 7364, groups: ZIPCODE, 243
##
## Fixed effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -0.47806 0.04023 -11.883 < 2e-16 ***
## PHILLYYes 0.29637 0.06885 4.304 1.67e-05 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## PHILLYYes -0.568## boundary (singular) fit: see help('isSingular')## Generalized linear mixed model fit by ML
## Call: glmer(formula = HIGHBP2_no6 ~ HAVEPARK_new + RESPAGE_4CAT + RACE3 +
## EDUCAT3 + EVRDIAB_new + EVRASTHA + EVERSMOK + MENTAL3A + NPOV150 + OBESE_2CAT +
## XERCISE_2CAT + PHILLY + (1 | ZIPCODE), data = d.2018, family = binomial,
## control = glmerControl(optimizer = "bobyqa"), nAGQ = 8)
##
## Estimates of Fixed Effects:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -2.86598 0.31926 -8.977 < 2e-16 ***
## HAVEPARK_newNo 0.28652 0.09508 3.014 0.00258 **
## RESPAGE_4CAT2 0.65397 0.24511 2.668 0.00763 **
## RESPAGE_4CAT3 1.89919 0.22192 8.558 < 2e-16 ***
## RESPAGE_4CAT4 2.69483 0.22412 12.024 < 2e-16 ***
## RACE3Black 0.72867 0.11702 6.227 4.76e-10 ***
## RACE3Other 0.42510 0.16682 2.548 0.01083 *
## EDUCAT3HS or Coll -0.21367 0.21916 -0.975 0.32959
## EDUCAT3Col Grad and up -0.41923 0.22537 -1.860 0.06286 .
## EVRDIAB_newYes 0.93748 0.11317 8.284 < 2e-16 ***
## EVRASTHAYes 0.22040 0.11236 1.961 0.04982 *
## EVERSMOKYes 0.26286 0.08304 3.165 0.00155 **
## MENTAL3AYes 0.49820 0.11022 4.520 6.18e-06 ***
## NPOV150Yes 0.25423 0.10568 2.406 0.01615 *
## OBESE_2CATYes 0.70629 0.08933 7.906 2.65e-15 ***
## XERCISE_2CAT3+ days -0.19560 0.08295 -2.358 0.01837 *
## PHILLYYes -0.05305 0.09522 -0.557 0.57748
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Exponentiated Fixed Effects and Confidence Bounds:
## Estimate 2.5 % 97.5 %
## (Intercept) 0.05692747 0.03044854 0.1064333
## HAVEPARK_newNo 1.33177896 1.10535800 1.6045799
## RESPAGE_4CAT2 1.92316809 1.18954605 3.1092327
## RESPAGE_4CAT3 6.68049993 4.32426464 10.3206170
## RESPAGE_4CAT4 14.80297470 9.54057076 22.9680242
## RACE3Black 2.07232018 1.64759169 2.6065384
## RACE3Other 1.52973864 1.10310247 2.1213807
## EDUCAT3HS or Coll 0.80761595 0.52559941 1.2409518
## EDUCAT3Col Grad and up 0.65755510 0.42276112 1.0227495
## EVRDIAB_newYes 2.55354950 2.04554925 3.1877087
## EVRASTHAYes 1.24657184 1.00016902 1.5536787
## EVERSMOKYes 1.30064422 1.10528459 1.5305338
## MENTAL3AYes 1.64575976 1.32602259 2.0425935
## NPOV150Yes 1.28946325 1.04822013 1.5862274
## OBESE_2CATYes 2.02646691 1.70097719 2.4142406
## XERCISE_2CAT3+ days 0.82234092 0.69895128 0.9675132
## PHILLYYes 0.94833585 0.78687762 1.1429234
##
## Estimates of Random Effects (Covariance Components):
## Groups Name Std.Dev.
## ZIPCODE (Intercept) 0
##
## Number of obs: 3295, groups: ZIPCODE, 218
##
## logLik df AIC BIC
## -1797.06 18 3630.13 3739.93# From the
Park_effect$OR[3] <- 1.332
Park_effect$Lower95CI[3] <- 1.11
Park_effect$Upper95CI[3] <- 1.605