Delete Records In Set After Classification?

I have a large (300K+ records) database with a sizeable amount of duplicate records. I want to delete the duplicates but this is not a matter of simply Remove Duplicates; I need to evaluate them before I do.

I am wondering what functions would:

1) select the specific record in a set of duplicates that makes a determination of a status
2) once the status has been determined for the set, delete all other records

Fields in my database:

ACCIDENT NUM (ID field, in text or General format)
DUP (for Duplicate, indicated by a character, for now its a "?")
OCC_KILLED (in Number format)
OCC_INJURED (in Number format)
SEVERITY (in text format)

Here are some scenarios:

ACC dup K I
12345 ? 0 0
12345 ? 1 2

Or:

ACC dup K I
123456 ? 0 1
123456 ? 1 0

Or:

ACC dup K I
1234567 ? 0 0
1234567 ? 0 2
1234567 ? 0 0

This is the formula for indicating if there are Duplicate records in the larger dataset:

=IF(OR(A2=A3,A2=A1),"?","")

I need to determine the Severity of the accident based on this:

If OCC_KILLED > 0 then SEVERITY = F (for Fatal)
IF OCC_INJURED > 0 and > OCC_KILLED then SEVERITY = I (for Injury)
IF OCC_KILLED and OCC_INJURED >= 0 then SEVERITY = F
IF OCC_KILLED and OCC_INJURED = 0 then SEVERITY = PDO (for Property Damage Only)

I have a code already in place for how to create the value for Severity but it DOES NOT account for duplicate records:

(in SEVERITY field):

=IF(A1<>0, "F",IF(B1<>0,"I","PDO")

sampleset.xls is a sample ot the database.