I have a set of health care information in FHIR json format. Before storing such sensitive information i would like to anonymize the sensitive information such as name , phone number, address, physician contacts, etc.
I googled about anonimizing the fhir json and found there exist some security labels which classifies the information as Confidential , restricted , etc - https://www.hl7.org/fhir/valueset-security-labels.html
https://www.hl7.org/fhir/extension-auditevent-anonymized.html - This talks about the extension and the structure of anonymized data.
I am not sure how to replace these variables with security codes. Is there any api which does that or i have to find all the tags which are sensitive and loop the json and replace tags with security code.
Can someone give me an example of this can be achieved.
i have attached a sample json which i generated via synthea
{
"fullUrl": "urn:uuid:7a534e34-40a2-48ab-bc92-066d23251a8b",
"resource": {
"resourceType": "Patient",
"id": "7a534e34-40a2-48ab-bc92-066d23251a8b",
"text": {
"status": "generated",
"div": "<div xmlns=\"http://www.w3.org/1999/xhtml\">Generated by Synthea.Version identifier: v2.5.0-378-gee8c6470\n . Person seed: 6477291342685874262 Population seed: 1586936726889</div>"
},
"extension": [
{
"url": "http://hl7.org/fhir/StructureDefinition/patient-mothersMaidenName",
"valueString": "Hilma831 Luettgen772"
},
{
"url": "http://hl7.org/fhir/StructureDefinition/patient-birthPlace",
"valueAddress": {
"city": "Boston",
"state": "Massachusetts",
"country": "US"
}
},
{
"url": "http://synthetichealth.github.io/synthea/disability-adjusted-life-years",
"valueDecimal": 3.3178400761167306
},
{
"url": "http://synthetichealth.github.io/synthea/quality-adjusted-life-years",
"valueDecimal": 71.68215992388328
}
],
"identifier": [
{
"system": "https://github.com/synthetichealth/synthea",
"value": "7a534e34-40a2-48ab-bc92-066d23251a8b"
},
{
"type": {
"coding": [
{
"system": "http://terminology.hl7.org/CodeSystem/v2-0203",
"code": "MR",
"display": "Medical Record Number"
}
],
"text": "Medical Record Number"
},
"system": "http://hospital.smarthealthit.org",
"value": "7a534e34-40a2-48ab-bc92-066d23251a8b"
},
{
"type": {
"coding": [
{
"system": "http://terminology.hl7.org/CodeSystem/v2-0203",
"code": "SS",
"display": "Social Security Number"
}
],
"text": "Social Security Number"
},
"system": "http://hl7.org/fhir/sid/us-ssn",
"value": "999-80-4232"
},
{
"type": {
"coding": [
{
"system": "http://terminology.hl7.org/CodeSystem/v2-0203",
"code": "DL",
"display": "Driver's License"
}
],
"text": "Driver's License"
},
"system": "urn:oid:2.16.840.1.113883.4.3.25",
"value": "S99911013"
},
{
"type": {
"coding": [
{
"system": "http://terminology.hl7.org/CodeSystem/v2-0203",
"code": "PPN",
"display": "Passport Number"
}
],
"text": "Passport Number"
},
"system": "http://standardhealthrecord.org/fhir/StructureDefinition/passportNumber",
"value": "X66843459X"
}
],
"name": [
{
"use": "official",
"family": "Kreiger457",
"given": [
"Anthony633"
],
"prefix": [
"Mr."
]
}
],
"telecom": [
{
"system": "phone",
"value": "555-660-2614",
"use": "home"
}
],
"gender": "male",
"birthDate": "1939-12-25",
"deceasedDateTime": "2015-02-02T22:36:55+05:30",
"address": [
{
"extension": [
{
"url": "http://hl7.org/fhir/StructureDefinition/geolocation",
"extension": [
{
"url": "latitude",
"valueDecimal": 42.11942297974089
},
{
"url": "longitude",
"valueDecimal": -71.21834679934824
}
]
}
],
"line": [
"496 Tromp Mews Unit 96"
],
"city": "Walpole",
"state": "Massachusetts",
"postalCode": "02081",
"country": "US"
}
],
"maritalStatus": {
"coding": [
{
"system": "http://terminology.hl7.org/CodeSystem/v3-MaritalStatus",
"code": "M",
"display": "M"
}
],
"text": "M"
},
"multipleBirthBoolean": false,
"communication": [
{
"language": {
"coding": [
{
"system": "urn:ietf:bcp:47",
"code": "en-US",
"display": "English"
}
],
"text": "English"
}
}
]
},
"request": {
"method": "POST",
"url": "Patient"
}
}
There is no 'standard' way to anonymize because anonymization is a risk-reduction approach that requires knowledge of both the threats you're trying to defend against and how the data needs to be used. When combined with other data, any element could potentially be 'sensitive'. It could also be essential to the intended analysis.
The extension you're looking at is specific to audit event. The way you'd actually flag a Patient resource as having been anonymized is using the ANONYED security tag in resource.meta. I don't believe there's a standard extension for marking what elements should be anonymized. If there were such an element, the element would also need to indicated how the data should be anonymized and the alorithm would typically need to take into account multiple elements - or even multiple resources. E.g. Should dates be changed to a random value, or just have the day shifted? Should all dates in the set of related resources be shifted by the same amount or should they be adjusted independently. If you're changing the practitioners pointed to by different resources, should they be left as is, each one changed randomly, the same practitioner changed to the same random practitioner?
Related
Due to a (somehow) special requirement to store data alongside a FHIR resource, I am looking for the cleanest way to extend it to maintain such extra data, which is basically dynamic key-value lists.
On the one hand I have something like the following, with 0..* sets of key-value, that I have tried to store nesting three extensions (adjusting the context by level) being the last a single limitation to the accepted types (string/integer), although it does not allow me to repeat ids, anyway I am getting around it using prefixes XD.
"n-lists-of-key-value": [
{
"list1-key1": "list1-value1",
"list1-key2": "list1-value1",
"list1-key3": "list1-value1"
},
{
"list2-key1": "list2-value1",
"list2-key2": "list2-value2"
}
]
becomes:
···,
"extension": [ {
"id": "n-lists-of-key-value",
"url": "http://x.org/fhir/Extension/NListOfKeyValueExtension",
"extension": [ {
"id": "list_1",
"url": "http://x.org/fhir/Extension/ListKeyValueExtension,
"extension": [ {
"id": "list1-key1",
"url": "http://x.org/fhir/Extension/KeyValueExtension,
"valueString": "list1-value1"
}, {
"id": "list1-key2",
"url": "http://x.org/fhir/Extension/KeyValueExtension,
"valueString": "list1-value1"
}, {
"id": "list1-key3",
"url": "http://x.org/fhir/Extension/KeyValueExtension,
"valueString": "list1-value1"
} ]
}, {
"id": "list_2",
"url": "http://x.org/fhir/Extension/ListKeyValueExtension,
"extension": [ {
"id": "list2-key1",
"url": "http://x.org/fhir/Extension/KeyValueExtension,
"valueString": "list2-value1"
}, {
"id": "list2-key2",
"url": "http://x.org/fhir/Extension/KeyValueExtension,
"valueString": "list2-value2"
} ]
} ]
} ],
But on the other hand I have this, which is what I find most difficult to model, because trying to reuse the extension definitions I run into (conceptual?) difficulties with slicing and the use of the discriminators, failing to get past the error "Element matches more than one slice".
"my-3-fixed-named-sets": {
"fixed-named-set1": {
"list1-key1": "list1-value1",
"list1-key2": "list1-value2",
"list1-key3": "list1-value3"
},
"fixed-named-set2": {
"list2-key1": "list2-value1",
"list2-key2": "list2-value2"
},
"fixed-named-set3": {
"list2-key1": "list2-value1",
"list2-key2": "list2-value2"
}
}
becomes:
···,
"extension": [ {
"id": "my-3-fixed-named-sets",
"url": "http://x.org/fhir/Extension/NListOfKeyValueExtension",
"extension": [ {
"id": "fixed-named-set1",
"url": "http://x.org/fhir/Extension/ListKeyValueExtension,
"extension": [ {
"id": "list1-key1",
"url": "http://x.org/fhir/Extension/KeyValueExtension,
"valueString": "list1-value1"
}, {
"id": "list1-key2",
"url": "http://x.org/fhir/Extension/KeyValueExtension,
"valueString": "list1-value2"
}, {
"id": "list1-key3",
"url": "http://x.org/fhir/Extension/KeyValueExtension,
"valueString": "list1-value3"
} ]
}, {
"id": "fixed-named-set2",
"url": "http://x.org/fhir/Extension/ListKeyValueExtension,
"extension": [ {
"id": "list2-key1",
"url": "http://x.org/fhir/Extension/KeyValueExtension,
"valueString": "list2-value1"
}, {
"id": "list2-key2",
"url": "http://x.org/fhir/Extension/KeyValueExtension,
"valueString": "list2-value2"
} ],
}, {
"id": "fixed-named-set3",
"url": "http://x.org/fhir/Extension/ListKeyValueExtension,
"extension": [ {
"id": "list2-key1",
"url": "http://x.org/fhir/Extension/KeyValueExtension,
"valueString": "list2-value1"
}, {
"id": "list2-key3",
"url": "http://x.org/fhir/Extension/KeyValueExtension,
"valueString": "list2-value3"
} ]
} ]
} ],
I am aware that by indiscriminately including new extensions with different names/urls, I can end up ingesting all the data, but I suspect that there is a much cleaner way and that I am failing in , probably, the way to include an extra condition into the discriminator (https://hl7.org/fhir/profiling.html#slicing), so that I could point to the proper slide? or de fact of extending extensions or extending value[x] slices?
"discriminator": [
{
"type": "value",
"path": "url"
},
{ "???": "??Extension.value.slideName??
···
Then my specific question would be... can I slice the value[x] of the extension of a specific profile, and which in turn make them extensions? this lasy is to get having n key value lists
Any comment is very welcome :)
I have a payload for Claim.
I want to add an additional field LOB (Line of business) using extension, and POST it to FHIR Server, and then search the resource using that field like:
GET: http://fhir_server/Claim?lob=value
In the payload I have added the extension block:
{
"resourceType": "Claim",
"status": "active",
"extension" : [{
"url" : "<i entered some random url>",
"valueString" : "MAPD"
}],
"identifier": [{
"value": "TEST"
}],
"type": {
"coding": [{
"system": "http://terminology.hl7.org/CodeSystem/claim-type",
"code": "institutional"
}]
},
"use": "claim",
"patient": {
"reference": "f8d8477c-1ef4-4878-abed-51e514bfd91f",
"display": "John_Smith"
},
"billablePeriod": {
"start": "1957-04-12T21:08:35+05:30",
"end": "1957-04-12T21:23:35+05:30"
},
"created": "1957-04-12T21:23:35+05:30",
"provider": {
"reference": "urn:uuid:Organization",
"display": "PCP79032"
},
"priority": {
"coding": [{
"system": "http://terminology.hl7.org/CodeSystem/processpriority",
"code": "normal"
}]
},
"facility": {
"reference": "Location",
"display": "PCP79032"
},
"diagnosis": [{
"sequence": 1,
"diagnosisReference": {
"reference": "0316a0c4-0e46-e3fa-a7bf-9e4f3b9c9e92"
}
}, {
"sequence": 2,
"diagnosisReference": {
"reference": "932432b0-ac67-0f7a-1382-26c48050c62f"
}
}],
"insurance": [{
"sequence": 1,
"focal": true,
"coverage": {
"display": "Blue Cross Blue Shield"
}
}],
"item": [{
"sequence": 1,
"productOrService": {
"coding": [{
"system": "http://snomed.info/sct",
"code": "162673000",
"display": "General examination of patient (procedure)"
}],
"text": "General examination of patient (procedure)"
},
"encounter": [{
"reference": "67062d00-2531-3ebd-8558-1de2fd3e5aab"
}]
}, {
"sequence": 2,
"diagnosisSequence": [1],
"productOrService": {
"coding": [{
"system": "http://snomed.info/sct",
"code": "713458007",
"display": "Lack of access to transportation (finding)"
}],
"text": "Lack of access to transportation (finding)"
}
}, {
"sequence": 3,
"diagnosisSequence": [2],
"productOrService": {
"coding": [{
"system": "http://snomed.info/sct",
"code": "73595000",
"display": "Stress (finding)"
}],
"text": "Stress (finding)"
}
}],
"total": {
"value": 786.3299999999999,
"currency": "USD"
}
}
and created a SearchParameter using POST:
{
"resourceType" : "SearchParameter",
"id": "21761828-b929-416a-8813-d3646cf288f4",
"name": "lob",
"status": "active",
"url" : "https://fhir_server/StructureDefinition/f971fc4998167948838f8a8831ea914c",
"description": "Returns a Claim with extension.valueString matching the specified one in request.",
"code" : "lob",
"base" : [ "Claim" ],
"type" : "string",
"expression": "Claim.extension.where(url ='https://fhir_server/StructureDefinition/f971fc4998167948838f8a8831ea914c').value.string"
}
but while fetching the result (GET: https://fhirserver/Claim?lob=MAPD), I am getting the full dataset instead of the filtered result.
I also did a $reindex on the /Claim/<resource_id>/$reindex, but its not listing the field lob.
Am I missing something ?
Servers need to be configured to support new search parameters - you can't search by arbitrary elements, only by specific SearchParameters the server supports. That's true for both searching on extensions as well as core elements. You'll have to work with the owner of the server to support your desired search capability. Did you try posting on http://chat.fhir.org to find out if there was a non-extension mechanism to represent "line of business" on a Claim?
I have an array object like this:
{
"Items":[
{
"ShortName":"Product short name",
"Image":"https://www.andrew.cmu.edu/user/cfperron/cats/images/cat8.jpg",
"ManufacturerName":"MMM",
"CatalogName":"cats"
},
{
"ShortName":"Product2 short name",
"Image":"https://www.andrew.cmu.edu/user/cfperron/cats/images/cat7.jpg",
"ManufacturerName":"SSS",
"CatalogName":"Dogs"
}
]
}
how to get Images in ImageSet dynamically in adaptive cards? I was able to get rest of the values in FactSet. But stuck with images.
Ok I figured it out.
{
"type": "AdaptiveCard",
"body": [
{
"type": "TextBlock",
"size": "medium",
"weight": "bolder",
"text": "Search Results"
},
{
"type": "Container",
"items": [
{
"type": "ImageSet",
"imageSize": "medium",
"images": [
{
"type": "Image",
"url": "${Image}",
"size": "Medium"
}
]
},
{
"type": "FactSet",
"facts": [
{
"title": "Short Name",
"value": "${ShortName}"
},
{
"title": "Supplier Name",
"value": "${SupplierName} "
},
{
"title": "Price ",
"value": "${PriceAmount} "
}
]
}
],
"$data": "${$root['Items']}"
}
],
"$schema": "http://adaptivecards.io/schemas/adaptive-card.json",
"version": "1.2"
}
But next issue I am facing is if image url is empty, it throws an error "Adaptive Card Rendered error:
{
"message": "Cannot read property 'style' of null"
}"
How can we give a null check inside adaptive card for Image property?
How do I search for a person with BOTH given names I provide?
I have the following 2 patients who are "close". Everything (in the Human Name area) is the same except one of the GivenNames are the same.
Note "Apple" vs "Banana".
{
"resourceType": "Bundle",
"id": "269caf66-0ccc-43e7-b9a5-f16f84db0149",
"meta": {
"lastUpdated": "2019-11-20T19:30:26.858917+00:00"
},
"type": "searchset",
"link": [
{
"relation": "self",
"url": "https://localhost:44348/Patient?given=Jingerheimer"
}
],
"entry": [
{
"fullUrl": "https://localhost:44348/Patient/504f6bd3-e9b4-4846-8948-97bf09c70722",
"resource": {
"resourceType": "Patient",
"id": "504f6bd3-e9b4-4846-8948-97bf09c70722",
"meta": {
"versionId": "1",
"lastUpdated": "2019-11-20T19:26:11.005+00:00"
},
"identifier": [
{
"system": "ssn",
"value": "111-11-1111"
},
{
"system": "uuid",
"value": "da55d068e0784b359fa97498a11543c5"
}
],
"name": [
{
"family": "Smith",
"given": [
"John",
"Apple",
"Jingerheimer"
]
}
]
},
"search": {
"mode": "match"
}
},
{
"fullUrl": "https://localhost:44348/Patient/10054ce9-6141-4eca-bc5b-0978f8c8afcb",
"resource": {
"resourceType": "Patient",
"id": "10054ce9-6141-4eca-bc5b-0978f8c8afcb",
"meta": {
"versionId": "1",
"lastUpdated": "2019-11-20T19:26:48.962+00:00"
},
"identifier": [
{
"system": "ssn",
"value": "222-22-2222"
},
{
"system": "uuid",
"value": "52d09f9436d44591816fd229dd139523"
}
],
"name": [
{
"family": "Smith",
"given": [
"John",
"Banana",
"Jingerheimer"
]
}
]
},
"search": {
"mode": "match"
}
}
]
}
One has GivenNames that include "Apple". The other includes GivenNames that include "Banana".
This search works fine:
https://localhost:44348/Patient/?given=Jingerheimer
What I have tried is:
https://localhost:44348/Patient/?given=Jingerheimer&given=Apple
but that gives me no results.
Note, omitting "given=Jingerheimer" is not an option....that filters a bunch of others.
I'm trying to get
"Has BOTH of the given names I provide"
Your syntax is correct, so I think the server does not handle the search correctly. Can you check the self link for your second search to see if it reflects the search you performed? Does the result Bundle have an OperationOutcome detailing something went wrong? If all that seems okay, you'll need to check your server's code.
Let's say I have two JSON objects (call them 'websites' and 'links'). I need to end up with both objects in a single CSV (ideally in separate columns).
I'm working with something like this:
File.open("file.json", "w") do |f|
combined = [websites, links]
f.write(JSON.pretty_generate(combined))
end
And then I'm using the Ruby gem json2csv to convert this file to a CSV. But when I do, I get the following error:
error: undefined method 'keys' for #<Array:0x007fea8a8e33f8>
I can't figure out what's wrong. When I look in file.json, it appears to be structured like this: [{websites}, {links}]. From my limited knowledge of JSON, I think that's right, but I could easily be wrong.
Also, I know this won't get me separate columns in the CSV. If anyone has an answer for that part, major bonus points.
EDIT: JSON examples included below; error message changed after minor fix on my end.
websites:
{
"uri": "https://v1/websites",
"id": 28235674,
"background": null,
"createdDate": 1399585684000,
"lastActivityDate": 1430682494000,
"lastCommunicationDate": 1430682494000,
"lastNonCommunicationChronicleDate": 1430330886000,
"lastModifiedDate": 1449263116000,
"lastViewedDate": 1421429034000,
"preferredContactType": null,
"rss": "",
"emailAddresses": [
{
"email": "",
"type": "Work"
},
{
"email": "",
"type": "Work"
},
{
"email": "",
"type": "Work"
},
{
"email": "not found",
"type": "Work"
}
],
"phoneNumbers": [
],
"streetAddresses": [
],
"socialNetworks": [
{
"profileUrl": "http://twitter.com",
"name": "Twitter"
},
{
"profileUrl": "http://www.facebook.com",
"name": "Facebook"
},
{
"profileUrl": "http://plus.google.com",
"name": "GooglePlus"
},
{
"profileUrl": "http://www.linkedin.com",
"name": "LinkedIn"
},
{
"profileUrl": "http://twitter.com",
"name": "Twitter"
}
],
"contactUrls": [
],
"tags": [
"tag1",
"tag2"
],
"mostRecentActivity": "https://v1/history",
"mostRecentChronicle": "https://v1/history",
"mostRecentCommunication": "https://v1/history",
"mostRecentNonCommunicationChronicle": "https://v1/history",
"projectStates": "https://v1/websites",
"history": "https://v1/history",
"customFieldValues": [
],
"name": "",
"primaryDomain": "",
"domains": [
""
],
"associatedPeople": "https://v1/people",
"payments": "https://payments",
"links": "https://v1/links",
"type": "https://v1/websites"
}
links:
{
"uri": "https://v1/links/custom_fields",
"id": 15529329,
"value": "Name",
"backgroundColor": null,
"customField": "https://v1/links/custom_fields"
}
combined output:
[
{
"uri": "https://v1/websites",
"id": 28235674,
"background": null,
"createdDate": 1399585684000,
"lastActivityDate": 1430682494000,
"lastCommunicationDate": 1430682494000,
"lastNonCommunicationChronicleDate": 1430330886000,
"lastModifiedDate": 1449263116000,
"lastViewedDate": 1421429034000,
"preferredContactType": null,
"rss": "",
"emailAddresses": [
{
"email": "",
"type": "Work"
},
{
"email": "",
"type": "Work"
},
{
"email": "",
"type": "Work"
},
{
"email": "not found",
"type": "Work"
}
],
"phoneNumbers": [
],
"streetAddresses": [
],
"socialNetworks": [
{
"profileUrl": "http://twitter.com/",
"name": "Twitter"
},
{
"profileUrl": "http://www.facebook.com",
"name": "Facebook"
},
{
"profileUrl": "http://plus.google.com",
"name": "GooglePlus"
},
{
"profileUrl": "http://www.linkedin.com/",
"name": "LinkedIn"
},
{
"profileUrl": "http://twitter.com/",
"name": "Twitter"
}
],
"contactUrls": [
],
"tags": [
"tag1",
"tag2"
],
"mostRecentActivity": "https://v1/history/",
"mostRecentChronicle": "https://v1/history/",
"mostRecentCommunication": "https://v1/history/",
"mostRecentNonCommunicationChronicle": "https://v1/history/",
"projectStates": "https://v1/websites/",
"history": "https://v1/history",
"customFieldValues": [
],
"name": "",
"primaryDomain": "",
"domains": [
""
],
"associatedPeople": "https://v1/people",
"links": "https://v1/links",
"type": "https://v1/websites"
},
{
"uri": "https://links/custom_fields",
"id": 15529329,
"value": "Name",
"backgroundColor": null,
"customField": "https://links/custom_fields"
}
]
JSON.pretty_generate() is expecting a hash and you are passing an array of 2 hashes. Start with: combined.map { |c| f.write(JSON.pretty_generate(c)) } and then mapping them into your CSV should be as easy as following the CSV documentation.