Forum Discussion
Tracking for patients
this is a sample from my project.
as you can see we have the id no. and the service that the patient used with the cost, the date and the provider for the service.
what I'm expecting: to find the abuser/misuser of the service by the no. of visits to the provider, dates are near to each other or going for the same diagnosis with high no. of frequency anything weird i can find will help
MarinaSaadalla You send a nice file with over 80 thousand records. I cannot possibly oversee what kind of data you have in there. But perhaps a few tips to get started. You may want to split the P.O.ID (column Q) so that you have the PO number and the year. And then perhaps also split the Specific Assessment (column U) in Code and Description.
Then you could do a first summary (count and cost) by P.O. number (or whatever column identifies an individual), Assessment code and Date, using a pivot table. But the number of POs and Assessment codes is quite big, so perhaps you need start by counting the number of visit per PO per month. That could eliminate most of them. Then you focus on those POs that have many occurrences, as the others are probably "not weird". But that's for you to judge.
Riny_van_Eekelen thank for your fast response ... i made the sample more simple may it becomes more clear for you to help me.
thanks in advance
MarinaSaadalla Thanks! I assume that the Card numbers represent the patients. I created a pivot table that counts the number of visits per card number, per assessment, per day. Showing only those that had more than 1 visit in total. That limits the list to 10 patients, one of which had 3 visits within two days. Up to you to judge the weirdness.
Now, in your real file with 68000 records, you could start in the same way, but vary the filtering limits to limit the output.
Riny_van_Eekelen I found highlighted card numbers in the data sheet, how did you do it ?