-
Notifications
You must be signed in to change notification settings - Fork 3
/
Copy pathpcsupply.R
816 lines (677 loc) · 32.5 KB
/
pcsupply.R
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
# Primary care and health outcomes in the United States
# (c) 2018, Sanjay Basu, basus@stanford.edu
#install.packages("devtools")
#install.packages("tidyverse")
#install.packages("readxl")
#install.packages("labelled")
#install.packages("plm")
#install.packages("reshape")
#install.packages("stargazer")
#install.packages("httr")
#install.packages("mctest)
#install.packages("olsrr")
#install.packages("lme4")
#install.packages("vioplot")
rm(list=ls())
library(devtools)
library(tidyverse)
library(readxl)
library(labelled)
library(httr)
setwd("~/Data/ahrf")
# Download raw AHRF files ------------------------------------------------------
# See https://github.com/jjchern/ahrf/blob/master/data-raw/prep_county.R
url = "https://datawarehouse.hrsa.gov/DataDownload/AHRF/AHRF_2016-2017.ZIP"
fil_zip = tempfile(fileext = ".zip")
if(!file.exists("data-raw/county/ahrf2017.asc")) {
download.file(url, fil_zip)
dir.create("data-raw/county")
unzip(fil_zip, exdir = "data-raw/county", junkpaths = TRUE)
}
list.files("data-raw/county")
raw_src = "data-raw/county/ahrf2017.asc" # Raw data
dic_src = "data-raw/county/ahrf2016-17.sas" # SAS dictionary file
doc_src = "data-raw/county/AHRF 2016-2017 Technical Documentation.xlsx"
# Find out the line for the first field: F00001 ---------------------------
read_excel(doc_src) %>%
pull(X__1) %>%
grepl("F00001", .) %>%
which() -> bgn_line
bgn_line
# Prepare the layout file -------------------------------------------------
read_excel(doc_src,
col_names = c("field", "col_col", "year_of_data", "var_label",
"characteristics", "source", "date_on"),
skip = bgn_line) %>%
filter(grepl("^F[0-9]", field)) %>%
separate(col_col, c("col_start", "col_end")) %>%
mutate_at(c("col_start", "col_end"), as.integer) -> ahrf_county_layout
ahrf_county_layout
# Prepare the county AHRF file --------------------------------------------
read_fwf(file = raw_src,
col_positions = fwf_positions(start = ahrf_county_layout$col_start,
end = ahrf_county_layout$col_end,
col_names = ahrf_county_layout$field)) -> ahrf_county
ahrf_county
# Delete raw data as it’s too large ---------------------------------------
unlink(raw_src)
# Download raw IHME LE files ---------------------------------------
# See https://github.com/BuzzFeedNews/2017-05-us-health-care/blob/master/index.Rmd
url = "http://ghdx.healthdata.org/sites/default/files/record-attached-files/IHME_USA_COUNTY_LE_MORTALITY_RISK_1980_2014_NATIONAL_XLSX.zip"
fil_zip = tempfile(fileext = ".zip")
{
download.file(url, fil_zip)
dir.create("data-raw/county")
unzip(fil_zip, exdir = "data-raw/county", junkpaths = TRUE)
}
list.files("data-raw/county")
raw_src = "data-raw/county/IHME_USA_COUNTY_LE_MORTALITY_RISK_1980_2014_NATIONAL_Y2017M05D08.xlsx" # Raw data
life_expect_us <- read_excel(raw_src, sheet = 1, skip = 3, col_names = FALSE) %>%
select(1:10)
names(life_expect_us) <- c("place","fips","le_1980","le_1985","le_1990","le_1995","le_2000","le_2005","le_2010","le_2014")
# Clean the FIPS codes, adding zeros where necessary ---------------------------------------
life_expect_us <- life_expect_us %>%
mutate(fips = ifelse(nchar(fips)==4|nchar(fips)==1,paste0("0",fips),fips))
# Extract data for counties only ---------------------------------------
life_expect_counties <- life_expect_us %>%
filter(grepl(",",place)) %>%
separate(place, into=c("place","state"), sep = ", ")
# Select columns with life expectancy data and convert from text string with confidence intervals to numbers ---------------------------------------
counties_clean <- life_expect_counties %>%
select(4:11) %>%
mutate_all(funs(as.numeric(substring(.,1,5))))
# Create data frame with names and abbreviations for states and District of Columbia ---------------------------------------
states <- data_frame(state.name,state.abb)
names(states) <- c("state","abb")
dc <- data_frame("District of Columbia", "DC")
names(dc) <- c("state","abb")
states <- bind_rows(states,dc)
# Join that to the counties' names, states, and FIPS codes ---------------------------------------
states_names <- life_expect_counties %>%
select(1:3) %>%
inner_join(states) %>%
mutate(place = paste0(place,", ",abb))
# Recombine AHRF with processed life expectancy data ---------------------------------------
life_expect_counties <- bind_cols(states_names,counties_clean)
# Linearly impute 2015 LE ----
life_expect_counties$le_2015 = life_expect_counties$le_2014 + 1/4*(life_expect_counties$le_2014-life_expect_counties$le_2010)
# keep relevant years ----
life_expect_counties <- life_expect_counties %>%
select(fips, le_2005, le_2010, le_2015)
# Save LE ----------------------------------------------------------------
save(life_expect_counties,file="life_expect_counties")
# Download raw CHRD files ---------------------------------------
chrd <- read_csv("http://www.countyhealthrankings.org/sites/default/files/CHR_TRENDS_CSV_2018.csv")
chrd <- chrd %>%
separate(yearspan, into=c("time","yearend"), sep = "-") %>%
unite(fips,"statecode","countycode",sep="") %>%
select(c(1,3,4,9)) %>%
mutate(time = as.integer(time)) %>%
# filter(time == 2000 | time==2005 | time==2010) %>%
unite(var_year,"measurename","time") %>%
spread(var_year,rawvalue) %>%
select(c("fips","Adult obesity_2005","Adult obesity_2010","Adult obesity_2013")) %>%
mutate(fipscode = as.character(fips),
obese_2005 = `Adult obesity_2005`*100,
obese_2010 = as.numeric(`Adult obesity_2010`)*100,
obese_2015 = obese_2010+5/3*(as.numeric(`Adult obesity_2013`)*100-as.numeric(`Adult obesity_2010`)*100)
) %>%
select(c("fipscode","obese_2005","obese_2010","obese_2015"))
# add in tobacco smoking data----
chrd_tob_15 <- read_csv("http://www.countyhealthrankings.org/sites/default/files/2017CHR_CSV_Analytic_Data.csv", skip = 1)
chrd_tob_15 <- chrd_tob_15 %>%
rename(tob_2015 = "measure_9_value") %>%
mutate(tob_2015 = as.numeric(tob_2015*100),
fipscode = as.character(fipscode)) %>%
select(c(fipscode,tob_2015))
chrd <- left_join(chrd,chrd_tob_15, by=c("fipscode"="fipscode"))
#url <- "http://www.countyhealthrankings.org/sites/default/files/2012%20County%20Health%20Rankings%20National%20Data_v2.xls"
#GET(url, write_disk(tf <- tempfile(fileext = ".xls")))
#chrd_tob_10 <- read_excel(tf, "Ranked Measure Data",skip = 1)
chrd_tob_10 = read.csv("~/Downloads/2012 County Health Rankings National Data_v2.csv")
chrd_tob_10 <- chrd_tob_10 %>%
rename(tob_10 = "X..Smokers") %>%
mutate(tob_2010 = as.numeric(tob_10),
tob_2005 = as.numeric(tob_10),
fipscode = as.character(FIPS))%>%
select(c(fipscode,tob_2005,tob_2010))
chrd <- left_join(chrd,chrd_tob_10, by=c("fipscode"="fipscode"))
chrd[chrd<0]=0
# Save CHRD ----------------------------------------------------------------
save(chrd,file="chrd")
# Download urban/rural data ----
#https://www.cdc.gov/nchs/data/oae/NCHSUrbruralFileDocumentation.pdf
library(readr)
# url = ("ftp://ftp.cdc.gov/pub/Health_Statistics/NCHS/Datasets/OAE/urbanrural/NCHSURCodes2013.txt")
# ctyurb <- read_table2(url, col.names = FALSE)
ctyurb <- read_table2("data-raw/county/NCHSURCodes2013.txt",col_names = FALSE)
ctyurb <- ctyurb %>%
rename(fips = X1,
urb_2015 = X9,
urb_2005 = X10) %>%
select(fips,urb_2005,urb_2015)
ctyurb$urb_2005[is.na(ctyurb$urb_2005)]=ctyurb$urb_2015[is.na(ctyurb$urb_2005)]
ctyurb$urb_2010 = round((as.integer(ctyurb$urb_2005)+as.integer(ctyurb$urb_2015))/2)
ctyurb <- ctyurb %>%
mutate(urb_2005 = as.numeric(urb_2005),
urb_2015 = as.numeric(urb_2015),
urb_2010 = as.numeric(urb_2010))
ctyurb$urb_2005[ctyurb$urb_2005>6]=NA
ctyurb$urb_2010[ctyurb$urb_2010>6]=NA
ctyurb$urb_2015[ctyurb$urb_2015>6]=NA
ctyurb$urb_2005[ctyurb$urb_2005<5]=0
ctyurb$urb_2005[ctyurb$urb_2005>=5]=1
ctyurb$urb_2010[ctyurb$urb_2010<5]=0
ctyurb$urb_2010[ctyurb$urb_2010>=5]=1
ctyurb$urb_2015[ctyurb$urb_2015<5]=0
ctyurb$urb_2015[ctyurb$urb_2015>=5]=1
save(ctyurb,file="ctyurb")
# Organize into panel data ---------------------------------------
rm(list=ls())
library(tidyverse)
setwd("~/Data/ahrf")
load("~/Data/ahrf/ahrf_county")
load("~/Data/ahrf/life_expect_counties")
load("~/Data/ahrf/chrd")
load("~/Data/ahrf/ctyurb")
# County-level non-federal [non-military] MD providers = total general practice (GP and fam med) + general internal medicine ---------------------------------------
# Other providers [DO, NP] available only from 2010
ahrf_county %>%
select(county = F04437,
fips = F00002,
gim_2005 = `F11209-05`, # Gnrl Int Med, PC, Office Based
fp_2005 = `F08860-05`, # MD's, Tot Gen Pract, PC,Off Bsd
ped_2005 = `F11706-05`, # Peds, Tot PC, Off Bsd
pop_2005 = `F11984-05`,
gim_2010 = `F11209-10`,
fp_2010 = `F08860-10`,
ped_2010 = `F11706-10`,
pop_2010 = `F04530-10`,
gim_2015 = `F11209-15`,
fp_2015 = `F08860-15`,
ped_2015 = `F11706-15`,
pop_2015 = `F11984-15`,
urb_2013 = `F00020-13`, # Rural-Urban Continuum Code , See https://www.ers.usda.gov/data-products/rural-urban-continuum-codes/documentation/
inc_2005 = `F13226-05`, # Median Household Income , See https://www.bea.gov/newsreleases/regional/lapi/lapi_newsrelease.htm
inc_2010 = `F13226-10`,
inc_2015 = `F13226-15`,
ed_2006 = `F14480-06`, # % Persons 25+ Yrs w/<HS Diploma , See https://www.census.gov/programs-surveys/acs/data.html
ed_2011 = `F14480-11`,
medct_2010 = `F15299-10`, # Stan,Risk-Adj Per Cap Medcr Cst, See https://www.cms.gov/Research-Statistics-Data-and-Systems/Statistics-Trends-and-Reports/Medicare-Geographic-Variation/GV_PUF.html
medct_2015 = `F15299-15`,
eld_2005 = `F14083-05`, # Population Estimate 65+ , Census County Char File
eld_2010 = `F14840-10`,
eld_2015 = `F14083-15`,
fem_2005 = `F13907-05`, # Pop tot fem, Census
fem_2015 = `F13907-15`,
blkm_2005 = `F13910-05`, # Pop Black male, Census
blkm_2015 = `F13910-15`,
blkf_2005 = `F13911-05`, # Pop Black fem, Census
blkf_2015 = `F13911-15`,
hism_2005 = `F13920-05`, # Pop Hisp male, Census
hism_2015 = `F13920-15`,
hisf_2005 = `F13921-05`,
hisf_2015 = `F13921-15`,
unemp_2005 = `F06795-05`, # Unemployment Rate, 16+ , Bureau of Labor Stats
unemp_2010 = `F06795-10`,
unemp_2015 = `F06795-15`,
poll_2006 = `F15338-06`, # Days w/8-hr Avg Ozone ovr NAAQS, CDC EPH Tracking Network
poll_2010 = `F15338-10`,
poll_2012 = `F15338-12`,
pov_2005 = `F13321-05`, # Percent Persons in Poverty , Census SAIPE
pov_2010 = `F13321-10`,
pov_2015 = `F13321-15`,
tot_2005 = `F11215-05`, # M.D.'s, Total Ptn Care Non-Fed , AMA Phys Master File
tot_2010 = `F11215-10`,
tot_2015 = `F11215-15`,
hobed_2005 = `F08921-05`, # Hospital Beds , AHA Survey Database
hobed_2010 = `F08921-10`,
hobed_2014 = `F08921-14`,
unins_2010 = `F14751-10`, # % <65 without Health Insurance , Census SAHIE
unins_2015 = `F15474-15`,
mcare_2011 = `F15549-11`, # Mdcre Enrllmnt, Agd & Dsbld Tot, CMS Enroll Dashboard
mcare_2015 = `F15549-15`,
homeval_2006 = `F14613-06`, # median home value
homeval_2011 = `F14613-11`
) %>%
mutate(pop_2005 = as.integer(pop_2005),
pop_2010 = as.integer(pop_2010),
pop_2015 = as.integer(pop_2015),
fp_2005 = as.integer(fp_2005),
fp_2010 = as.integer(fp_2010),
fp_2015 = as.integer(fp_2015),
gim_2005 = as.integer(gim_2005),
gim_2010 = as.integer(gim_2010),
gim_2015 = as.integer(gim_2015),
ped_2005 = as.integer(ped_2005),
ped_2010 = as.integer(ped_2010),
ped_2015 = as.integer(ped_2015),
pc_2005 = (as.integer(gim_2005)+as.integer(fp_2005)+as.integer(ped_2005))/pop_2005*100000, # PC providers per 100k pop
pc_2010 = (as.integer(gim_2010)+as.integer(fp_2010)+as.integer(ped_2010))/pop_2010*100000,
pc_2015 = (as.integer(gim_2015)+as.integer(fp_2015)+as.integer(ped_2015))/pop_2015*100000,
inc_2005 = as.integer(inc_2005)*1.38, # adjust for inflation to 2015 USD, see https://data.bls.gov/cgi-bin/cpicalc.pl?cost1=1&year1=200001&year2=201501
inc_2010 = as.integer(inc_2010)*1.23, # adjust for inflation to 2015 USD, see https://data.bls.gov/cgi-bin/cpicalc.pl?cost1=1.00&year1=200501&year2=201501
inc_2015 = as.integer(inc_2015),
ed_2005 = (as.integer(ed_2006)-0.2*(as.integer(ed_2011)-as.integer(ed_2006)))/10, # linear interp
ed_2010 = (as.integer(ed_2011)-0.2*(as.integer(ed_2011)-as.integer(ed_2006)))/10,
ed_2015 = (as.integer(ed_2011)+0.8*(as.integer(ed_2011)-as.integer(ed_2006)))/10,
medct_2005 = (as.integer(medct_2010)-(as.integer(medct_2015)-as.integer(medct_2010)))/100, # linear interp
medct_2010 = as.integer(medct_2010)/100,
medct_2015 = as.integer(medct_2015)/100,
eld_2005 = as.integer(eld_2005)/as.integer(pop_2005)*100, # as %age
eld_2010 = as.integer(eld_2010)/as.integer(pop_2010)*100,
eld_2015 = as.integer(eld_2015)/as.integer(pop_2015)*100,
fem_2010 = (as.integer(fem_2005)+as.integer(fem_2015))/2/as.integer(pop_2010)*100, # linear interp
fem_2005 = as.integer(fem_2005)/as.integer(pop_2005)*100, # as %age
fem_2015 = as.integer(fem_2015)/as.integer(pop_2015)*100,
blk_2005 = (as.integer(blkm_2005)+as.integer(blkf_2005))/as.integer(pop_2005)*100,
blk_2015 = (as.integer(blkm_2015)+as.integer(blkf_2015))/as.integer(pop_2015)*100,
blk_2010 = (blk_2005+blk_2015)/2, # linear interp
his_2005 = (as.integer(hism_2005)+as.integer(hisf_2005))/as.integer(pop_2005)*100,
his_2015 = (as.integer(hism_2015)+as.integer(hisf_2015))/as.integer(pop_2015)*100,
his_2010 = (his_2005+his_2015)/2, # linear interp
unemp_2005 = as.integer(unemp_2005)/10,
unemp_2010 = as.integer(unemp_2010)/10,
unemp_2015 = as.integer(unemp_2015)/10,
poll_2005 = as.integer(poll_2006)-1/4*(as.integer(poll_2010)-as.integer(poll_2006)),
poll_2010 = as.integer(poll_2010),
poll_2015 = as.integer(poll_2012)+3/2*(as.integer(poll_2012)-as.integer(poll_2010)),
pov_2005 = as.integer(pov_2005)/10,
pov_2010 = as.integer(pov_2010)/10,
pov_2015 = as.integer(pov_2015)/10,
spec_2005 = ((as.integer(tot_2005))/pop_2005*100000)-pc_2005,
spec_2010 = ((as.integer(tot_2010))/pop_2010*100000)-pc_2010,
spec_2015 = ((as.integer(tot_2015))/pop_2015*100000)-pc_2015,
hobed_2015 = (as.integer(hobed_2014)+1/4*(as.integer(hobed_2014)-as.integer(hobed_2010)))/pop_2015*100000,
hobed_2005 = as.integer(hobed_2005)/pop_2005*100000,
hobed_2010 = as.integer(hobed_2010)/pop_2010*100000,
unins_2005 = (as.integer(unins_2010)-(as.integer(unins_2015)-as.integer(unins_2010)))/10,
unins_2010 = as.integer(unins_2010)/10,
unins_2015 = as.integer(unins_2015)/10,
mcare_2005 = (as.integer(mcare_2011)-6/4*(as.integer(mcare_2015)-as.integer(mcare_2011)))/pop_2005*100,
mcare_2010 = (as.integer(mcare_2011)-1/4*(as.integer(mcare_2015)-as.integer(mcare_2011)))/pop_2010*100,
mcare_2015 = as.integer(mcare_2015)/pop_2015*100,
homeval_2005 = as.integer(homeval_2006)*1.18-0.2*(as.integer(homeval_2011)*1.06-as.integer(homeval_2006)*1.18), # linear interp w/ CPI adjustment to 2015 $US
homeval_2010 = as.integer(homeval_2011)*1.06-0.2*(as.integer(homeval_2011)*1.06-as.integer(homeval_2006)*1.18),
homeval_2015 = as.integer(homeval_2011)*1.06+0.8*(as.integer(homeval_2011)*1.06-as.integer(homeval_2006)*1.18)
) -> ahrf_county
ahrf_county[ahrf_county<0]=0
#lapply(ahrf_county, summary)
ahrf_county <- ahrf_county %>%
select(county,
fips,
fp_2005,
fp_2010,
fp_2015,
gim_2005,
gim_2010,
gim_2015,
ped_2005,
ped_2010,
ped_2015,
pc_2005,
pc_2010,
pc_2015,
inc_2005,
inc_2010,
inc_2015,
ed_2005,
ed_2010,
ed_2015,
medct_2005,
medct_2010,
medct_2015,
eld_2005,
eld_2010,
eld_2015,
fem_2005,
fem_2010,
fem_2015,
blk_2005,
blk_2010,
blk_2015,
his_2005,
his_2010,
his_2015,
unemp_2005,
unemp_2010,
unemp_2015,
poll_2005,
poll_2010,
poll_2015,
pov_2005,
pov_2010,
pov_2015,
spec_2005,
spec_2010,
spec_2015,
hobed_2005,
hobed_2010,
hobed_2015,
unins_2005,
unins_2010,
unins_2015,
mcare_2005,
mcare_2010,
mcare_2015,
homeval_2005,
homeval_2010,
homeval_2015,
tot_2005,
tot_2010,
tot_2015,
pop_2005,
pop_2010,
pop_2015)
# Join data to the AHRF subset ----
counties_data <- left_join(ahrf_county,life_expect_counties, by=c("fips"="fips"))
counties_data
# Join data to the urban/rural subset ----
counties_data <- left_join(counties_data,ctyurb, by=c("fips"="fips"))
counties_data
# Join data to CHRD files ----
counties_data <- left_join(counties_data,chrd, by=c("fips"="fipscode"))
counties_data
#lapply(counties_data,summary)
paneldata = counties_data
paneldata = paneldata %>%
select(-one_of("fips"))
# Reshape wide to long; note that guam and puerto rico don't have cvd available and make up most of the NA's, so need to exclude them when counting NA's for 50 states+DC ----
paneldata = data.frame(paneldata)
panel = reshape(paneldata, varying =dput(names(paneldata[,2:dim(paneldata)[2]])),
direction="long",idvar="county",sep="_")
plot(panel$pc,panel$le)
plot(log(panel$pc),(panel$le), xlab = "Log (# of primary care physicians/10k pop)", ylab = "Age-adjusted life expectancy at birth (years)")
save(panel,file="pcpanel")
save(paneldata,file="pcpaneldata")
# descriptive stats ----
rm(list=ls())
library(tidyverse)
library(stargazer)
setwd("~/Data/ahrf")
load("~/Data/ahrf/pcpanel")
load("~/Data/ahrf/pcpaneldata")
panel$tot = panel$pc + panel$spec
panel05 = panel[,2:dim(panel)[2]]
panel05 = panel05[panel05$time==2005,]
panel10 = panel[,2:dim(panel)[2]]
panel10 = panel10[panel10$time==2010,]
panel15 = panel[,2:dim(panel)[2]]
panel15 = panel15[panel15$time==2015,]
paneldiff = panel15-panel05
tableone05 = matrix(0,nrow=dim(panel)[2]-1,ncol=3)
tableone10 = matrix(0,nrow=dim(panel)[2]-1,ncol=3)
tableone15 = matrix(0,nrow=dim(panel)[2]-1,ncol=3)
tableoned = matrix(0,nrow=dim(panel)[2]-1,ncol=3)
for (i in 1:(dim(panel)[2]-1)){
tableone05[i,1:3] = c(mean(na.omit(panel05[,i])),quantile(panel05[,i],na.rm=T,c(.025,.975))[1],quantile(panel05[,i],na.rm=T,c(.025,.975))[2])
tableone10[i,1:3] = c(mean(na.omit(panel10[,i])),quantile(panel10[,i],na.rm=T,c(.025,.975))[1],quantile(panel10[,i],na.rm=T,c(.025,.975))[2])
tableone15[i,1:3] = c(mean(na.omit(panel15[,i])),quantile(panel15[,i],na.rm=T,c(.025,.975))[1],quantile(panel15[,i],na.rm=T,c(.025,.975))[2])
tableoned[i,1:3] = c(mean(na.omit(paneldiff[,i])),quantile(paneldiff[,i],na.rm=T,c(.025,.975))[1],quantile(paneldiff[,i],na.rm=T,c(.025,.975))[2])
}
colnames(panel05)
table05=cbind(colnames(panel05),tableone05)
table10=cbind(colnames(panel10),tableone10)
table15=cbind(colnames(panel15),tableone15)
tabled = cbind(colnames(paneldiff),tableoned)
stargazer(table05,type="text")
stargazer(table10,type="text")
stargazer(table15,type="text")
stargazer(tabled,type="text")
sum(na.omit(panel$pc[panel$time==2005]==0))
sum(na.omit(panel$pc[panel$time==2015]==0))
sum(na.omit(panel$pc[panel$time==2005]>100))
sum(na.omit(panel$pc[panel$time==2015]>100))
pc_gainloss = (panel$pc[panel$time==2015]-panel$pc[panel$time==2005])
summary(pc_gainloss)
summary(pc_gainloss[panel$urb==0])
summary(pc_gainloss[panel$urb==1])
summary(pc_gainloss[panel$pov<16])
summary(pc_gainloss[panel$pov>=16])
summary(pc_gainloss[panel$blk<9])
summary(pc_gainloss[panel$blk>=9])
summary(pc_gainloss[panel$his<8])
summary(pc_gainloss[panel$his>=8])
quantile(pc_gainloss,c(.025,.975),na.rm=T)
quantile(pc_gainloss[panel$urb==0],c(.025,.975),na.rm=T)
quantile(pc_gainloss[panel$urb==1],c(.025,.975),na.rm=T)
quantile(pc_gainloss[panel$pov<16],c(.025,.975),na.rm=T)
quantile(pc_gainloss[panel$pov>=16],c(.025,.975),na.rm=T)
quantile(pc_gainloss[panel$blk<9],c(.025,.975),na.rm=T)
quantile(pc_gainloss[panel$blk>=9],c(.025,.975),na.rm=T)
quantile(pc_gainloss[panel$his<8],c(.025,.975),na.rm=T)
quantile(pc_gainloss[panel$his>=8],c(.025,.975),na.rm=T)
pcn_gainloss = (panel$fp[panel$time==2015]-panel$fp[panel$time==2005])+
(panel$gim[panel$time==2015]-panel$gim[panel$time==2005])+
(panel$ped[panel$time==2015]-panel$ped[panel$time==2005])
summary(pcn_gainloss)
summary(pcn_gainloss[panel$urb==0])
summary(pcn_gainloss[panel$urb==1])
summary(pcn_gainloss[panel$pov<16])
summary(pcn_gainloss[panel$pov>=16])
summary(pcn_gainloss[panel$blk<9])
summary(pcn_gainloss[panel$blk>=9])
summary(pcn_gainloss[panel$his<8])
summary(pcn_gainloss[panel$his>=8])
quantile(pcn_gainloss,c(.025,.975),na.rm=T)
quantile(pcn_gainloss[panel$urb==0],c(.025,.975),na.rm=T)
quantile(pcn_gainloss[panel$urb==1],c(.025,.975),na.rm=T)
quantile(pcn_gainloss[panel$pov<16],c(.025,.975),na.rm=T)
quantile(pcn_gainloss[panel$pov>=16],c(.025,.975),na.rm=T)
quantile(pcn_gainloss[panel$blk<9],c(.025,.975),na.rm=T)
quantile(pcn_gainloss[panel$blk>=9],c(.025,.975),na.rm=T)
quantile(pcn_gainloss[panel$his<8],c(.025,.975),na.rm=T)
quantile(pcn_gainloss[panel$his>=8],c(.025,.975),na.rm=T)
spec_gainloss = (panel$spec[panel$time==2015]-panel$spec[panel$time==2005])
summary(spec_gainloss)
summary(spec_gainloss[panel$urb==0])
summary(spec_gainloss[panel$urb==1])
summary(spec_gainloss[panel$pov<16])
summary(spec_gainloss[panel$pov>=16])
summary(spec_gainloss[panel$blk<9])
summary(spec_gainloss[panel$blk>=9])
summary(spec_gainloss[panel$his<8])
summary(spec_gainloss[panel$his>=8])
quantile(spec_gainloss,c(.025,.975),na.rm=T)
quantile(spec_gainloss[panel$urb==0],c(.025,.975),na.rm=T)
quantile(spec_gainloss[panel$urb==1],c(.025,.975),na.rm=T)
quantile(spec_gainloss[panel$pov<16],c(.025,.975),na.rm=T)
quantile(spec_gainloss[panel$pov>=16],c(.025,.975),na.rm=T)
quantile(spec_gainloss[panel$blk<9],c(.025,.975),na.rm=T)
quantile(spec_gainloss[panel$blk>=9],c(.025,.975),na.rm=T)
quantile(spec_gainloss[panel$his<8],c(.025,.975),na.rm=T)
quantile(spec_gainloss[panel$his>=8],c(.025,.975),na.rm=T)
specn_gainloss = ((panel$tot[panel$time==2015]-panel$pc[panel$time==2015])-
(panel$tot[panel$time==2005]-panel$pc[panel$time==2005]))*(panel$pop[panel$time==2005])/100000
summary(specn_gainloss)
summary(specn_gainloss[panel$urb==0])
summary(specn_gainloss[panel$urb==1])
summary(specn_gainloss[panel$pov<16])
summary(specn_gainloss[panel$pov>=16])
summary(specn_gainloss[panel$blk<9])
summary(specn_gainloss[panel$blk>=9])
summary(specn_gainloss[panel$his<8])
summary(specn_gainloss[panel$his>=8])
quantile(specn_gainloss,c(.025,.975),na.rm=T)
quantile(specn_gainloss[panel$urb==0],c(.025,.975),na.rm=T)
quantile(specn_gainloss[panel$urb==1],c(.025,.975),na.rm=T)
quantile(specn_gainloss[panel$pov<16],c(.025,.975),na.rm=T)
quantile(specn_gainloss[panel$pov>=16],c(.025,.975),na.rm=T)
quantile(specn_gainloss[panel$blk<9],c(.025,.975),na.rm=T)
quantile(specn_gainloss[panel$blk>=9],c(.025,.975),na.rm=T)
quantile(specn_gainloss[panel$his<8],c(.025,.975),na.rm=T)
quantile(specn_gainloss[panel$his>=8],c(.025,.975),na.rm=T)
cor.test(pc_gainloss,spec_gainloss)
cor.test(pcn_gainloss,specn_gainloss)
# center and scale ----
rm(list=ls())
library(tidyverse)
library(plm)
library(lme4)
library(stargazer)
setwd("~/Data/ahrf")
load("~/Data/ahrf/pcpanel")
load("~/Data/ahrf/pcpaneldata")
panel$tot = panel$pc + panel$spec
zpanel = panel
zpanel$fp = log(zpanel$fp+1)
zpanel$gim = log(zpanel$gim+1)
zpanel$pc = log(zpanel$pc+1)
zpanel$inc = log(zpanel$inc)
zpanel$ed = log(zpanel$ed+1)
zpanel$medct = log(zpanel$medct)
zpanel$eld = log(zpanel$eld)
zpanel$fem = log(zpanel$fem)
zpanel$blk = log(zpanel$blk+1)
zpanel$his = log(zpanel$his+1)
zpanel$unemp = log(zpanel$unemp)
zpanel$poll = log(zpanel$poll+1)
zpanel$pov = log(zpanel$pov)
zpanel$spec = log(zpanel$spec+1)
zpanel$hobed = log(zpanel$hobed+1)
zpanel$unins = log(zpanel$unins)
zpanel$mcare = log(zpanel$mcare+1)
zpanel$obese = log(zpanel$obese)
zpanel$tot = log(zpanel$tot+1)
zpanel$tob = log(zpanel$tob)
zpanel$homeval = log(zpanel$homeval+1)
exp(mean(na.omit(log(panel$tot+1))))-exp(sd(na.omit(log(panel$tot+1))))
exp(mean(na.omit(log(panel$tot+1))))+exp(sd(na.omit(log(panel$tot+1))))
exp(mean(na.omit(log(panel$pc+1))))-exp(sd(na.omit(log(panel$pc+1))))
exp(mean(na.omit(log(panel$pc+1))))+exp(sd(na.omit(log(panel$pc+1))))
exp(mean(na.omit(log(panel$spec+1))))-exp(sd(na.omit(log(panel$spec+1))))
exp(mean(na.omit(log(panel$spec+1))))+exp(sd(na.omit(log(panel$spec+1))))
exp(mean(na.omit(log(panel$pov))))-exp(sd(na.omit(log(panel$pov))))
exp(mean(na.omit(log(panel$pov))))+exp(sd(na.omit(log(panel$pov))))
exp(mean(na.omit(log(panel$ed+1))))-exp(sd(na.omit(log(panel$ed+1))))
exp(mean(na.omit(log(panel$ed+1))))+exp(sd(na.omit(log(panel$ed+1))))
exp(mean(na.omit(log(panel$fem))))-exp(sd(na.omit(log(panel$fem))))
exp(mean(na.omit(log(panel$fem))))+exp(sd(na.omit(log(panel$fem))))
exp(mean(na.omit(log(panel$blk+1))))-exp(sd(na.omit(log(panel$blk+1))))
exp(mean(na.omit(log(panel$blk+1))))+exp(sd(na.omit(log(panel$blk+1))))
exp(mean(na.omit(log(panel$his+1))))-exp(sd(na.omit(log(panel$his+1))))
exp(mean(na.omit(log(panel$his+1))))+exp(sd(na.omit(log(panel$his+1))))
exp(mean(na.omit(log(panel$unemp))))-exp(sd(na.omit(log(panel$unemp))))
exp(mean(na.omit(log(panel$unemp))))+exp(sd(na.omit(log(panel$unemp))))
exp(mean(na.omit(log(panel$hobed+1))))-exp(sd(na.omit(log(panel$hobed+1))))
exp(mean(na.omit(log(panel$hobed+1))))+exp(sd(na.omit(log(panel$hobed+1))))
exp(mean(na.omit(log(panel$mcare+1))))-exp(sd(na.omit(log(panel$mcare+1))))
exp(mean(na.omit(log(panel$mcare+1))))+exp(sd(na.omit(log(panel$mcare+1))))
exp(mean(na.omit(log(panel$medct))))-exp(sd(na.omit(log(panel$medct))))
exp(mean(na.omit(log(panel$medct))))+exp(sd(na.omit(log(panel$medct))))
exp(mean(na.omit(log(panel$tob))))-exp(sd(na.omit(log(panel$tob))))
exp(mean(na.omit(log(panel$tob))))+exp(sd(na.omit(log(panel$tob))))
exp(mean(na.omit(log(panel$obese))))-exp(sd(na.omit(log(panel$obese))))
exp(mean(na.omit(log(panel$obese))))+exp(sd(na.omit(log(panel$obese))))
exp(mean(na.omit(log(panel$poll+1))))-exp(sd(na.omit(log(panel$poll+1))))
exp(mean(na.omit(log(panel$poll+1))))+exp(sd(na.omit(log(panel$poll+1))))
exp(mean(na.omit(log(panel$homeval+1))))-exp(sd(na.omit(log(panel$homeval+1))))
exp(mean(na.omit(log(panel$homeval+1))))+exp(sd(na.omit(log(panel$homeval+1))))
zpanel$fp = (zpanel$fp-mean(na.omit(zpanel$fp)))/(2*sd(na.omit(zpanel$fp)))
zpanel$gim = (zpanel$gim-mean(na.omit(zpanel$gim)))/(2*sd(na.omit(zpanel$gim)))
zpanel$pc = (zpanel$pc-mean(na.omit(zpanel$pc)))/(2*sd(na.omit(zpanel$pc)))
zpanel$spec = (zpanel$spec-mean(na.omit(zpanel$spec)))/(2*sd(na.omit(zpanel$spec)))
zpanel$tot = (zpanel$tot-mean(na.omit(zpanel$tot)))/(2*sd(na.omit(zpanel$tot)))
zpanel$ed = (zpanel$ed-mean(na.omit(zpanel$ed)))/(2*sd(na.omit(zpanel$ed)))
zpanel$medct = (zpanel$medct-mean(na.omit(zpanel$medct)))/(2*sd(na.omit(zpanel$medct)))
zpanel$eld = (zpanel$eld-mean(na.omit(zpanel$eld)))/(2*sd(na.omit(zpanel$eld)))
zpanel$fem = (zpanel$fem-mean(na.omit(zpanel$fem)))/(2*sd(na.omit(zpanel$fem)))
zpanel$blk = (zpanel$blk-mean(na.omit(zpanel$blk)))/(2*sd(na.omit(zpanel$blk)))
zpanel$his = (zpanel$his-mean(na.omit(zpanel$his)))/(2*sd(na.omit(zpanel$his)))
zpanel$unemp = (zpanel$unemp-mean(na.omit(zpanel$unemp)))/(2*sd(na.omit(zpanel$unemp)))
zpanel$poll = (zpanel$poll-mean(na.omit(zpanel$poll)))/(2*sd(na.omit(zpanel$poll)))
zpanel$pov = (zpanel$pov-mean(na.omit(zpanel$pov)))/(2*sd(na.omit(zpanel$pov)))
zpanel$hobed = (zpanel$hobed-mean(na.omit(zpanel$hobed)))/(2*sd(na.omit(zpanel$hobed)))
zpanel$unins = (zpanel$unins-mean(na.omit(zpanel$unins)))/(2*sd(na.omit(zpanel$unins)))
zpanel$mcare = (zpanel$mcare-mean(na.omit(zpanel$mcare)))/(2*sd(na.omit(zpanel$mcare)))
zpanel$obese = (zpanel$obese-mean(na.omit(zpanel$obese)))/(2*sd(na.omit(zpanel$obese)))
zpanel$tob = (zpanel$tob-mean(na.omit(zpanel$tob)))/(2*sd(na.omit(zpanel$tob)))
zpanel$inc = (zpanel$inc-mean(na.omit(zpanel$inc)))/(2*sd(na.omit(zpanel$inc)))
zpanel$homeval = (zpanel$homeval-mean(na.omit(zpanel$homeval)))/(2*sd(na.omit(zpanel$homeval)))
zpanel$le = zpanel$le*365.25
# address collinearity with VIFs----
# paneltest = (panel[complete.cases(panel),])
# paneltest = data.frame(paneltest[,3:25])
# x = as.matrix(paneltest[,1:15,17:23])
# y = as.matrix(paneltest[,16])
# library(mctest)
# omcdiag(x,y)
# imcdiag(x,y)
#
# library(olsrr)
# testmodel = lm(le~.,
# data = panel)
# ols_vif_tol(testmodel)
#
# testmodel = lm(le~pc+
# urb+ed+medct+fem+blk+his+unemp+poll+pov+hobed+mcare+obese+tob+spec+homeval,
# data = panel)
# ols_vif_tol(testmodel)
# Regressions ----
reg_min = (lmer(le~pc+ (1+pc| county)+ (1|time) ,
data = zpanel))
stargazer(reg_min,type="text",style="asq",single.row=T,
ci = T,digits=1,
ci.level = 0.95)
reg_tot = (lmer(le~tot+urb+pov+ed+fem+blk+his+unemp+hobed+mcare+medct+tob+obese+poll+homeval+ (1+tot| county)+ (1|time) ,
data = zpanel))
stargazer(reg_tot,type="text",style="asq",single.row=T,
ci = T,digits=1,
ci.level = 0.95)
lotot=exp(mean(na.omit(log(panel$tot+1))))-exp(sd(na.omit(log(panel$tot+1))))
hitot=exp(mean(na.omit(log(panel$tot+1))))+exp(sd(na.omit(log(panel$tot+1))))
lotot
hitot
summary(reg_tot)$coefficients[2]/(hitot-lotot)
reg_pc = (lmer(le~pc+urb+pov+ed+fem+blk+his+unemp+hobed+mcare+medct+tob+obese+poll+homeval+ (1+pc| county)+ (1|time) ,
data = zpanel))
stargazer(reg_pc,type="text",style="asq",single.row=T,
ci = T,digits=1,
ci.level = 0.95)
lopc=exp(mean(na.omit(log(panel$pc+1))))-exp(sd(na.omit(log(panel$pc+1))))
hipc=exp(mean(na.omit(log(panel$pc+1))))+exp(sd(na.omit(log(panel$pc+1))))
lopc
hipc
summary(reg_pc)$coefficients[2]/(hipc-lopc)
reg_spec = (lmer(le~spec+urb+pov+ed+fem+blk+his+unemp+hobed+mcare+medct+tob+obese+poll+homeval+ (1+spec| county)+ (1|time) ,
data = zpanel))
stargazer(reg_spec,type="text",style="asq",single.row=T,
ci = T,digits=1,
ci.level = 0.95)
lospec=exp(mean(na.omit(log(panel$spec+1))))-exp(sd(na.omit(log(panel$spec+1))))
hispec=exp(mean(na.omit(log(panel$spec+1))))+exp(sd(na.omit(log(panel$spec+1))))
lospec
hispec
summary(reg_spec)$coefficients[2]/(hispec-lospec)
reg_base = (lmer(le~pc+spec+urb+pov+ed+fem+blk+his+unemp+hobed+mcare+medct+tob+obese+poll+homeval+(1+pc| county)+ (1|time) ,
data = zpanel))
stargazer(reg_base,type="text",style="asq",single.row=T,
ci = T,digits=1,
ci.level = 0.95)
summary(reg_base)$coefficients[2]/(hipc-lopc)
summary(reg_base)$coefficients[3]/(hispec-lospec)
confint(reg_base,method="Wald")[7]/(hipc-lopc)
confint(reg_base,method="Wald")[29]/(hipc-lopc)
confint(reg_base,method="Wald")[8]/(hispec-lospec)
confint(reg_base,method="Wald")[30]/(hispec-lospec)
# Sensitivity analyses ----------
reg_melu = lmer(le~pc+pc*urb+spec+urb+pov+ed+fem+blk+his+unemp+hobed+mcare+medct+tob+obese+poll+homeval + (1+pc| county)+ (1|time) ,
data = zpanel)
stargazer(reg_melu,type="text",style="asq",single.row=T,
ci = T,digits=1,
ci.level = 0.95)
reg_melp = lmer(le~pc+pc*pov+spec+urb+pov+ed+fem+blk+his+unemp+hobed+mcare+medct+tob+obese+poll+homeval + (1+pc| county)+ (1|time) ,
data = zpanel)
stargazer(reg_melp,type="text",style="asq",single.row=T,
ci = T,digits=1,
ci.level = 0.95)
reg_melb = lmer(le~pc+pc*blk+spec+urb+pov+ed+fem+blk+his+unemp+hobed+mcare+medct+tob+obese+poll+homeval + (1+pc| county)+ (1|time) ,
data = zpanel)
stargazer(reg_melb,type="text",style="asq",single.row=T,
ci = T,digits=1,
ci.level = 0.95)
reg_melh = lmer(le~pc+pc*his+spec+urb+pov+ed+fem+blk+his+unemp+hobed+mcare+medct+tob+obese+poll+homeval + (1+pc| county)+ (1|time) ,
data = zpanel)
stargazer(reg_melh,type="text",style="asq",single.row=T,
ci = T,digits=1,
ci.level = 0.95)