Re: [Gnumed-devel] Lab import (data) matching

[James Busser]

On 3-Feb-08, at 2:18 PM, Karsten Hilbert wrote:

> - the importer auto-matches those results which coincide in either:
>   - firstname, lastname, dob, and gender
>   - or lab request ID and - when available - any of the above
>   - or lab request ID only if none of the above is available
>     (IOW pseudonymized results)
>   with *existing* patients

Building on what you suggested, how about, for a first  
implementation, defining two "levels" of matching:

1) auto-matching and
2) suggested matches

Auto-matching:

	lab_request_id AND is_pending AND (date of birth OR surname OR  
firstname)
	lab_request_id AND is_pending AND (null or blank person identifiers  
i.e anonymous patient)
        date of birth AND surname AND firstname AND (if sex is provided) sex
        ... the last one above gives good coverage for unexpected copies
        ... the last one should ignore mismatched lab_request_ids in case
                the lab is including the id from a request originated elsewhere
                but is being correctly copied

Suggested matching:

        lab_request_id
                AND (date of birth OR surname OR firstname)
                AND NOT is_pending

        lab_request_id
                AND (null or blank person identifiers i.e anonymous patient)
                AND NOT is_pending

	... either of the above could be a correction on a test previously  
marked final

        *first name variants use case
        date of birth
                AND first two characters of surname (see reference)

	... is the above the limit of what should be attempted for a first  
iteration?
	... work would remain to be done deciding e.g. to split the workflow  
between single-candidate suggested matches versus multiple-candidate  
matches
	... shall we reserve for later iterations more criteria including  
the following
        ... shall I wiki a link to this thread?

        *alias use case
        date of birth
                AND sex
                AND (phone number or postal code)

        *birthdate variants use case
        two among (year of birth, month of birth, date of birth)
                AND first two characters of surname
		AND first character of firstname matches one initial in GNUmed  
firstnames

        *other?

Reference:
--------------
Of interest, a registry study that tested match keys based on various  
combinations (name substrings, date of birth, gender, and health  
center) to link encounter and laboratory files found that an optimal  
match key was the first two letters of the last name with the date of  
birth, which had a sensitivity of 92.7% and a positive predictive  
value of 99.5% (even across ethnicity). This would suffice for auto- 
matching for research, but not for clinical care, with one obvious  
vulnerability being a multiple-birth within a family. This implies  
that we should perhaps consider multiple-birth as an attribute of a  
GNUmed person identity
        http://www.citeulike.org/group/98/article/408895

other links:
--------------

Scottish Care Information Store Patient Matching Reference Guide
(nicely lists potential match actions which could be insert /  
update / add alias  / patient exception and no action)
(also gives a figure of what a widget might look like)
	http://www.sci.scot.nhs.uk/products/store/SHOW/User_and_Training/ 
Patient%20Matching%20Reference%20Guide.pdf

Open Enterprise Master Person Index
        http://wiki.openempi.org/index.php/Requirements

Building a comprehensive clinical information system from components.  
The approach at Intermountain Health Care.
        http://www.ncbi.nlm.nih.gov/pubmed/12695790

A bit of blog discussion:
	http://www.neotool.com/blog/2007/01/26/patient-data-matching-with- 
hl7-messages/