---
title: -log2markup- example
author: 
- |
  Niels Henrik Bruun \
  Unit of Clinical Biostatistics, \
  Aalborg University Hospital, Aalborg, Denmark


date:  2022-04-21

abstract: |
  **Background** Infant mortality rates and birth defect rates are very high 
  for low birth weight babies.
  A woman's behaviour during pregnancy can greatly alter the chances of carrying 
  the baby to term. \
  **Aim** To identify if smoking during pregnancy is associated with giving 
  birth to a low birth weight baby. \
  **Results** Smoking during pregnancy increases the risk for birthweights below 
  2500g both crude and adjusted for age and race.
  It is only white mothers, who have a higher risk.
...

# Data


## Outcomes

Two binary variables for birthweight below 1500g and 2500g.


## Exposure

Smoking during pregnancy is the exposure.


## Counfounders

The variables for mothers age and race are considered a confounder.


## Data management

~~~
use low bwt age race smoke using "http://www.stata-press.com/data/r12/hospid2.dta", clear
rename low bwlt2500
generate bwlt1500 = bwt < 1500 if !missing(bwt)
order bwlt1500, before(bwlt2500)
label variable bwlt1500 "birthweight < 1500g"
label variable bwlt2500 "birthweight < 2500g"
label define no_yes 0 "No" 1 "Yes"
label values bwlt* no_yes
label define smoke 0 "No" 1 "Yes"
label values smoke smoke
label data "From: http://www.stata-press.com/data/r12/hospid2.dta"
~~~~
# Results


## Data description



----------------------------------------------------------------------------------------------------------
Columns by: smoked during pregnancy                       No                    Yes                  Total

-------------------------------------    -------------------    -------------------    -------------------
n (%)                                             115 (60.8)              74 (39.2)            189 (100.0)

age of mother, median (iqi)                23.0 (20.0; 26.0)      22.0 (19.0; 26.3)      23.0 (19.0; 26.0)

race, n (%)                                                                                               

  white                                            44 (38.3)              52 (70.3)              96 (50.8)

  black                                            16 (13.9)              10 (13.5)              26 (13.8)

  other                                            55 (47.8)              12 (16.2)              67 (35.4)

----------------------------------------------------------------------------------------------------------
Table: **Table 1:** The variables


## Effect measurements



------------------------------------------------------------------------------------------------------------------------------------------------------------
                                                              Crude: RR      Lower 95% CI      Upper 95% CI      Adj: RR      Lower 95% CI      Upper 95% CI

---------------------    -------------------------------    -----------    --------------    --------------    ---------    --------------    --------------
birthweight < 2500g      smoked during pregnancy (Yes)             1.61              1.06              2.45         2.01              1.33              3.04

birthweight < 1500g      smoked during pregnancy (Yes)             1.04              0.18              6.08         1.60              0.27              9.41

------------------------------------------------------------------------------------------------------------------------------------------------------------
Table: **Table 2:** Crude and adjusted (age, race) effects. Smoking during pregnancy increases the risk for birthweights below 2500g both crude and adjusted for age and race.


## sub analyses

~~~
emc, at(14(2)28) yline(0) ytitle(log(RR)) name(fig1, replace): poisson bwlt2500 smoke age, vce(robust)
graph export output/fig1.png, width(8000) replace
~~~~
![**Figure 1: Effect modification by age.**](output/fig1.png)

~~~
poisson bwlt2500 i.smoke##i.race, vce(robust)
margins r.smoke, over(race)
marginsplot, yline(0) ytitle(log(RR)) title("") name(fig2, replace)
graph export output/fig2.png, width(8000) replace
~~~~
![**Figure 2: Effect modification by race.** 
It is only white mothers, who have a higher risk.](output/fig2.png)