ABM Summary 13 April 2018

Data Prep and Goals

These updated data from April 13 have fixed two prior issues: one was the persistence of a pregnant state through life…in other words pregnancy not being reset after birth resulting in increasing Crude Birth Rate. The second was the deterministic rather than stochastic assignment of PTB resulting in only those exceeding a threshold value on stress ever being eligible for preterm birth. Narjes has fixed each of these. So these results pick up from CBR and PTB and continue on to look at STI’s, stillbirth, DM, etc.

library(dplyr)
library(DBI)
library(ggplot2)

con <- dbConnect(RSQLite::SQLite(),  'DATA/abm1.sqlite')

dbListTables(con)

## [1] "abm_m1"       "abm_m2"       "abm_m3"       "abm_m4"      
## [5] "sqlite_stat1" "sqlite_stat4"

# abm_m4 is the current round of results

Birth Rates

‘GestationalAge’ is coded like this:

36 = Preterm
38 = Term
100 = Not pregnant

Crude birth rate by Race & SES

a3 <- tbl(con, 'abm_m4') %>%
  filter(Generation %in% c(1,2,3,4,5)) %>%
  mutate(Pregnant = ifelse(GestationalAge == 100, 0, 1)) %>%
  group_by(Age, Race, SocialClass) %>%
  summarize(CBR = mean(Pregnant)) %>%
  collect()

# CBR by race
g <- ggplot(a3, aes(x = Age, y = CBR, group = Race, color = Race))
g + geom_smooth() + scale_x_continuous(limits = c(13,50)) +
  scale_y_continuous(limits = c(0, .15)) +
  ggtitle('Crude birth rate by race')

# CBR by class
g <- ggplot(a3, aes(x = Age, y = CBR, group = SocialClass, color = SocialClass))
g + geom_smooth() + scale_x_continuous(limits = c(13,50)) +
  scale_y_continuous(limits = c(0, .15)) +
  ggtitle('Crude birth rate by SES')

Crude birth rate by Parity

a3b <- tbl(con, 'abm_m4') %>%
  filter(Generation %in% c(1,2,3,4,5),
         Parity %in% c(0,1,2,3,4)) %>%
  mutate(Pregnant = ifelse(GestationalAge == 100, 0, 1)) %>%
  group_by(Age, Race, Parity) %>%
  summarize(CBR = mean(Pregnant)) %>%
  collect()

# CBR by Parity (stratify on Race)
g <- ggplot(a3b, aes(x = Age, y = CBR, group = factor(Parity), color = factor(Parity)))
g + geom_smooth() + facet_grid(.~Race) +
  scale_x_continuous(limits = c(13,50)) + 
  scale_y_continuous(limits = c(0, .15)) +
  ggtitle('Crude birth rate by parity')

Pregnancy Outcomes

Preterm Birth by Age

a4 <- tbl(con, 'abm_m4') %>%
  filter(Generation %in% c(1,2,3,4,5),
         GestationalAge != 100) %>%
  mutate(PTB = ifelse(GestationalAge == 36, 1, 0)) %>%
  group_by(Race, SocialClass, Age) %>%
  summarise(PTBr = mean(PTB, na.rm = T)) %>%
  collect()

g <- ggplot(a4, aes(x = Age, y = PTBr, group = Race, color = Race))
g + geom_smooth() + ggtitle('%PTB by Age and Race')

g <- ggplot(a4, aes(x = Age, y = PTBr, group = SocialClass, color = SocialClass))
g + geom_smooth() + ggtitle('%PTB by Age and SES')

Does Maternal Stress predict PTB?

RSQLite::initExtension(con) # this extensions allows stdev, median queries in SQLite
a4.1 <- tbl(con, 'abm_m4') %>%
  filter(Generation %in% c(1,2,3,4,5),
         GestationalAge != 100) %>%
  mutate(PTB = ifelse(GestationalAge == 36, 1, 0)) %>%
  select(PTB, AverageStressScore, LQ, Race, SocialClass) %>%
  collect()
g <- ggplot(a4.1, aes(x = factor(PTB), y = AverageStressScore, fill = factor(PTB)))
g + geom_boxplot(width = 0.5) + 
  facet_grid(~Race) +
  theme_bw() + 
  ggtitle('Stress score by PTB & Race')

g + geom_boxplot(width = 0.5) + 
  facet_grid(~SocialClass) +
  theme_bw() + 
  ggtitle('Stress score by PTB & Social Class')

Does PTB predict child LQ?

Here I got confused…Is there an indicator for whether a child agent was herself born preterm? I think the GestationalAge variable represents the outcome of a mother-agents pregnancy rather than a child-agents own history.

Other pregnany outcomes

a5 <- tbl(con, 'abm_m4') %>%
  filter(Generation %in% c(1,2,3,4,5),
         PregnancyOutcome != 'NotPregnant') %>%
  group_by(Race, SocialClass) %>%
  mutate(PTB = ifelse(GestationalAge == 36, 1, 0),
         SB = ifelse(PregnancyOutcome == 'StillBirth', 1, 0),
         ShortIPI = ifelse(IPI == 'ShortIPI', 1, 0)) %>%
  summarize(PTBr = mean(PTB, na.rm = T),
            SBr = mean(SB, na.rm = T),
            IPIr = mean(ShortIPI, na.rm = T)) %>%
  collect()

library(kableExtra)
knitr::kable(a5, digits = c(2,3,2)) %>%
  kable_styling(bootstrap_options = c("striped", "hover"))

Race	SocialClass	PTBr	SBr	IPIr
Black	Low	0.21	0.01	0.143
Black	Middle	0.16	0.01	0.113
Black	Upper	0.10	0.01	0.104
White	Low	0.18	0.00	0.133
White	Middle	0.14	0.00	0.127
White	Upper	0.11	0.00	0.128

Non-pregnancy health status

STI Incidence by age

a6 <- tbl(con, 'abm_m4') %>%
  filter(Generation %in% c(1,2,3,4,5)) %>%
  group_by(Race, SocialClass, Age) %>%
  summarize(STIr = mean(STI, na.rm = T)) %>%
  collect()

g <- ggplot(a6, aes(x = Age, y = STIr, group = Race, color = Race))
g + geom_smooth() + facet_grid(.~SocialClass) +
  scale_x_continuous(limits = c(13,50)) +
  ggtitle('STI rate by race, class, age')