revision

report/SAR-2023-017-BH-v01.Rmd

@@ -135,7 +135,7 @@ deaths %>%
 ```
 
 Under the "multiple observations per individual" approach, most outcome events are dropped for the complete case dataset.
-This happens because that dataset uses the exposure at discharge, but there are no Zip codes recorded for that measurement time.
+This happens because that dataset uses the exposure at all follow-up times, but most individuals that have expired did not have Zip codes recorded for that follow-up session.
 This way no DCI scores were available for most individuals, resulting in a sample of size 2 (Table 2).
 By applying the binary comparison between the two imputation approaches we found that both LOCF and LOCF+NOCB data frames are equal.
 A single outcome event was added to those datasets after the imputation is applied to the underlying SES data.

report/SAR-2023-017-BH-v01.md

@@ -82,7 +82,7 @@ The data procedures, design and analysis methods used in this report are fully d
 
 <!-- ## Statistical analyses -->
 
-This analysis was performed using statistical software `R` version 4.3.0.
+This analysis was performed using statistical software `R` version 4.3.2.
 
 # Results
 
@@ -126,7 +126,7 @@ Table: **Table 2** Number of death events available to models in each dataset
 |mult_locf+nocb | 694|
 
 Under the "multiple observations per individual" approach, most outcome events are dropped for the complete case dataset.
-This happens because that dataset uses the exposure at discharge, but there are no Zip codes recorded for that measurement time.
+This happens because that dataset uses the exposure at all follow-up times, but most individuals that have expired did not have Zip codes recorded for that follow-up session.
 This way no DCI scores were available for most individuals, resulting in a sample of size 2 (Table 2).
 By applying the binary comparison between the two imputation approaches we found that both LOCF and LOCF+NOCB data frames are equal.
 A single outcome event was added to those datasets after the imputation is applied to the underlying SES data.