Table 3

Hypothetical data from 10 ambulance records and 20 emergency department records

Ambulance dataEmergency department data
CaseYearDayHospBirth yearBirthdaySexCaseYearDayHospBirth yearBirthdaySex
From past experience or calculations using the larger list, we might assume that: (1) About 90% of injured patients brought by ambulance will generate an emergency department record, so the prior probability of a match between records on the two lists might be about 0.045 (equation 6), and the prior odds (equation 1) about 0.047. (2) The chances that records truly from the same event will have the same admission year, admission date, hospital, birth year, birth date, and sex (the M probabilities) might be 0.99, 0.95, 0.99, 0.95, 0.99, and 0.95 respectively. (3) The chances that variables would match for randomly selected records (the U probabilities) might be 0.50 for admission year, 0.0027 (1/365) for admission date, 0.4 for hospital X or Y, 0.2 for hospital Z, 0.01 for birth year, 0.0027 for birth date, 0.60 for males, and 0.40 for females.
A0101Jan01X1950Jan21ME0101Jan01X1950Jan21M
A0201Jan01X1950May01FE0201Jan10Z1987Jul17M
A0301Jan10Y1975Dec27E0301Feb23X1992Oct19M
A0401Aug13X1977Apr29FE0401Apr22Y1979May09M
A0501Sep12Y1980Feb16FE0501May02X1929Nov12F
A0601Dec31Z1919Sep16ME0601May23Y1964Dec01M
A0702Feb02X1924Mar26FE0701Jun01X1950May01F
A0802Jun10Y1951Mar29ME0801Aug14X1977Apr29F
A0902Aug06Y1953Apr17E0901Sep12Y1980Feb16F
A1002Sep21Z1956Jun03FE1001Oct21Y1985Mar12M
E1102Jan01Z1919Sep16M
E1202Jan10Y1975Dec27F
E1302Feb02X1924Mar26F
E1402May16X1924Oct12M
E1502Jun10Y1951Mar29M
E1602Jul04Z1982Jun12M
E1702Aug05Y1953Apr17M
E1802Aug06Y2002Apr17F
E1902Sep21Z1956Jun03F
E2002Nov22X1917May29M