Removing Multiple Admissions by ID

[Anindit Chhibber]

I have a data set that has admission dates by person ID (MRN). I would like to remove multiple admissions within 60 days for an individual keeping the first admission date.

* Example generated by -dataex-. To install: ssc install dataex
clear
input long MRN int AdmissionDate
  2972 20095
  3897 20452
 10173 20378
 26278 20251
 32706 20298
 33589 20228
 44859 20168
 47134 20142
 48637 20196
 48637 20450
 60699 20143
 95513 20335
113290 20447
119693 20093
121764 20356
138628 20226
138628 20234
141747 20169
144469 20111
144469 20150
144469 20235
144469 20248
145367 20089
145367 20293
145367 20298
149005 20162
149005 20235
151019 20350
152751 20402
152751 20408
154419 20334
156158 20201
161737 20408
162545 20156
165449 20383
165449 20389
165449 20432
174615 20221
175158 20231
179846 20209
182568 20257
184044 20313
186791 20431
188730 20200
198986 20248
200303 20415
200303 20416
212159 20317
213272 20277
213272 20370
242826 20122
257949 20326
259887 20167
259887 20185
263665 20185
264747 20415
284315 20263
284745 20341
315564 20117
315564 20141
315564 20250
326132 20282
326132 20296
332528 20165
381723 20386
391391 20291
391391 20421
392928 20404
419861 20177
429514 20272
434175 20228
434175 20246
462739 20109
462739 20120
462739 20151
462739 20161
462739 20170
465989 20424
465989 20425
465989 20450
469957 20231
479592 20144
479592 20380
479592 20439
481135 20222
481135 20228
507921 20323
519272 20261
520221 20434
520221 20440
538678 20178
539718 20139
539718 20195
544395 20389
557579 20116
577163 20249
577163 20263
582676 20204
589358 20324
589358 20357
end
format %td AdmissionDate

[William Lisowski]

Suppose you had a person with exactly three admission dates: May 1 and June 15 (45 days apart) and July 1 (16 days after June 15, 61 days from May 1).

What date(s) would you keep and what date(s) would you remove?

[Anindit Chhibber]

I am sorry i didn't make this clear. I would like to keep May 1 and July 1 but not June 15. Suppose there is a fourth admission date for the same person July 5th, would like to remove that too because its within 60 days of July 1. Hope this makes sense. Thanks!

[William Lisowski]

This code, run on your data, seems to do what you want.

generate start60 = AdmissionDate
format %td start60
bysort MRN (AdmissionDate): replace start60 = start60[_n-1] if _n>1 & AdmissionDate<start60[_n-1]+60
generate todrop = AdmissionDate != start60
list, clean