18:38:54 RRSAgent has joined #hcls 18:38:54 logging to http://www.w3.org/2007/11/08-hcls-irc 18:39:22 hello rrsagent, I was saying there are about 31 people here, and they are introducing themselves. 18:41:10 Susie has joined #hcls 18:41:54 zakim, who's here? 18:41:54 sorry, jar, I don't know what conference this is 18:41:56 On IRC I see Susie, RRSAgent, FabienG, jar, Zakim, chimezie, KerstinF, ericP 18:42:21 alanr has joined #hcls 18:42:28 zakim, list 18:42:28 I see T&S_XMLSEC(TPAC)8:00AM, XML_PMWG()10:00AM, SW_HCLS(TPAC)1:00PM active 18:42:31 also scheduled at this time are AB_(TPAC07)8:30AM, TAG_Weekly()1:00PM, WAI_UAWG()1:00PM, WS_Policy(Editors)1:00PM 18:42:31 zakim, who is on the call? 18:42:31 sorry, alanr, I don't know what conference this is 18:42:32 On IRC I see alanr, Susie, RRSAgent, FabienG, jar, Zakim, chimezie, KerstinF, ericP 18:42:46 Zakim, this is HCLS 18:42:46 ok, ericP; that matches SW_HCLS(TPAC)1:00PM 18:42:58 zakim, who is on the call? 18:42:58 On the phone I see ??P0, ??P1, +46.4.67.7.aabb 18:42:59 Zakim, who is here? 18:43:00 On the phone I see ??P0, ??P1, +46.4.67.7.aabb 18:43:01 On IRC I see alanr, Susie, RRSAgent, FabienG, jar, Zakim, chimezie, KerstinF, ericP 18:43:17 ha 18:43:56 restrictions on new logo are described on w3 web site 18:44:06 intros done. eric n is introducing agenda 18:44:29 +Kerstin_Forsberg 18:44:40 -Kerstin_Forsberg 18:44:49 eric n: what is hcls? 18:45:00 (i'm reporting on what's on his slide, not asking) 18:45:34 AdrianP has joined #hcls 18:45:36 i am self-appointed scribe, apparently 18:46:11 banderss has joined #HCLS 18:46:23 The agenda for the meeting is at: http://esw.w3.org/topic/HCLS/F2F/2007/11/Agenda 18:47:15 EricN introduces HCLS - http://www.w3.org/2001/sw/hcls/ 18:48:00 excellent, thanks for the uris 18:48:07 or are they iris? 18:48:56 Katsu has joined #hcls 18:50:11 There will be a HCLS Workshop at WWW 2008 in Beijing 18:50:27 Karen has joined #hcls 18:51:29 Invite RRSAgent 18:51:44 next issue of scientific american has semantic web theme with hcls article 18:51:54 karen, rrsagent is already on. 18:53:18 rrsagent, where am i? 18:53:18 See http://www.w3.org/2007/11/08-hcls-irc#T18-53-18 18:53:24 karen, I'm scribing 18:53:34 thanks! 18:53:50 no writing anything down now because rrsagent has the url to erin n's powerpoint 18:55:00 - +46.4.67.7.aabb 18:55:08 i think scribing here on irc makes most sense. maybe we can trade off, should i continue now or do you want to continue? 18:55:32 i can continue here 18:55:36 Please continue and let me know when you're tired 18:55:45 excellent, go ahead, let me know when you get tired. 18:55:48 alrighty 18:55:49 + +46.4.67.7.aacc 18:55:57 maybe you can paste in your previous notes or give a uri 18:56:23 -> http://esw.w3.org/topic/HCLS/ClinicalObservationsInteroperability/Minutes20071108 Initial notes on ESW Wiki 18:57:08 TanyaH: Partners HC is a strong supporter of this W3C IG 18:57:09 Next speaker: Tonya Hongsermeir 18:57:37 ... Most EHR lack knowledge (uner-nourished). 18:57:50 do we have a uri to karen's slides? 18:58:29 Don't know; I'll ask EriP when he returns 18:58:47 -> http://esw.w3.org/topic/HCLS/ClinicalObservationsInteroperability/November2007F2FAgenda.html?action=AttachFile&do=get&target=Tonya_F2F_Slides.ppt Tonya's slides 18:58:57 oops tonya 19:01:22 - +46.4.67.7.aacc 19:01:27 Partners legacy challenges; roadmap to leverage SW technologies; eager for tools and approaches 19:02:15 eneumann has joined #hcls 19:02:16 TonyaH: There is nothing today for solving dense problems for knowledge aquisition and maintainance 19:02:56 ... we want to get rid / reduce of the vertical silos 19:03:47 ... Content Management Servers (Documentum) are used for document management 19:04:20 alanr has joined #hcls 19:04:22 + +46.4.67.7.aadd 19:04:42 who just joined? 19:05:13 TanyaH: alot of ontology content is being handled / modeled as rules. The intersection and the tools that support the intersection are still not yet mature 19:06:26 Next Speaker: VipulK 19:07:13 VipulK: we are looking at the intersections between the biological / clinical spectrum and research / practice 19:08:41 VipulK: if there is one thing common across these it is the sharing of clinical observations 19:09:34 -> http://esw.w3.org/topic/HCLS/ClinicalObservationsInteroperability/November2007F2FAgenda.html?action=AttachFile&do=get&target=Vipul_F2F_Slides.ppt Vipul's slides 19:11:10 ericP has joined #hcls 19:14:11 VipulK goes over blood pressure models in both DCM and SDTM 19:14:34 ... demonstrates how these models can be expressed in OWL 19:17:20 VipulK: patient data can be merged by making coherent connections between observations on the same patient 19:24:09 alanr has joined #hcls 19:24:36 Question: was it difficult mapping terms and identifying criteria for definitions? 19:24:48 ... was OWL sufficiently expressive? 19:25:02 VipulK: Initially OWL was expressive enough 19:25:47 ... eventually the pragmatism must be considered in determining which languages to use 19:26:33 ScottM: Subsumption and equivalence may not be sufficient 19:27:57 Question: how were the mappings generated / created 19:28:09 VipulK: manually (across 2 models) 19:28:35 ... one issue: we want real world data from healthcare providers ... 19:28:53 JoanneL: which tools were used? 19:29:04 VipulK: we mostly used spreadsheets and protege 19:29:12 Question (from audience): how are you validating 19:29:34 VipulK: 1) validation by 'exports' (scribe not sure) 19:29:53 ... 2) ultimate test is to run it in a system 19:30:01 .... 3) use inference to test consistency 19:30:28 Next speaker: Rachel Richesson 19:31:03 s/'exports'/experts 19:31:38 -> http://esw.w3.org/topic/HCLS/ClinicalObservationsInteroperability/November2007F2FAgenda.html?action=AttachFile&do=get&target=Rachel_F2F_Slides.ppt RachelR's slides 19:32:16 Rachel describes projects her group does 19:33:07 ...NIH projects for example in diabetes 19:33:38 EMR: Electronic Medical Record 19:34:01 EHR: Electronic Healthcare Record 19:34:35 Rachel: The goal is to identify / screen patients via data in the EMR/EHR 19:34:50 Rachel goes over a slide of the process and the data involved 19:38:14 clinicaltrials.gov is source of sample protocol data 19:43:35 ...sites challenges around terminology of standards; reviews chart 19:44:12 CHI= Consolidated Healthcare [?] 19:44:41 Q: Does it have to include MEDRA (sp?) 19:45:13 Rachel: MEDRA would be in the findings range 19:45:37 ...the chart information will change; but is a frame today 19:46:01 ...CDSIC = Clinical Data Standards Interchange Consortium 19:46:27 ...people and activities overlap; however some different terminology in use 19:46:39 ...eg CDISC uses NCI 19:47:58 Tonya: We're mapping some things to SNOMED, but have to create our versions sometimes 19:48:56 Kerstin has joined #hcls 19:51:21 Rachel: we have use cases but need standards 19:51:30 ...one level for mapping 19:51:39 ...another is a bottom up approach to look at the data 19:51:50 ...good context to know there is a variety of representation issues to deal with 19:52:02 ...not there yet on delivery end or clinical end 19:53:10 alanr has joined #hcls 19:53:13 ...HL7 is one to watch for healthcare 19:53:31 ping 19:53:41 ...BRIDG trying to bring together HL7 and CDISC 19:54:12 ...Detailed Clinical Models; lots of ways to represent this 19:54:35 ...what we need to remember is what the models are and how we use them consistently 19:54:44 [visual cartoon] 19:55:15 what is the biggest conduit for getting patient data into a bytes and chars? 19:55:28 Fabien? 19:55:37 I have a message for you 19:55:44 data entry personel? 19:57:10 i wonder if we can play with these machines/interfaces 19:57:34 Katsu2 has joined #hcls 19:57:39 when someone says "taxonomy-direccted input", i don't have a feel for it 'till i see the UI 19:57:50 and the pain of the person inputing the data 19:59:03 For those virtually present (Bo and Kerstin), feel free to raise a question by typing 'q+' 19:59:36 Participants discuss legal issues with releasing data 20:00:02 ...challenges of anonymizing data, not releasing patient information 20:00:44 http://browse.opengalen.org/ 20:00:48 in firefox 20:00:55 example of taxonomic choice 20:01:38 Zakim, please dial ericP-mobile 20:01:38 ok, ericP; the call is being made 20:01:40 +EricP 20:02:11 -EricP 20:06:26 breaking for coffee now, rrsagent. 20:07:21 -??P0 20:08:11 -??P1 20:09:21 +??P0 20:09:40 Zakim, ??P0 is f2f 20:09:40 +f2f; got it 20:09:49 Zakim, please dial ericP-mobile 20:09:49 ok, ericP; the call is being made 20:09:51 +EricP 20:10:03 -EricP 20:19:47 +Kerstin_Forsberg 20:24:44 +??P16 20:25:53 -??P16 20:30:44 - +46.4.67.7.aadd 20:31:42 scribeNick: eneumann 20:31:50 + +46.4.67.7.aaee 20:32:37 Marcus Collins from Novartis is attending... 20:32:50 Tom Oniki presentation... 20:33:01 alanr has joined #hcls 20:33:17 Zakim, who is on the phone? 20:33:17 On the phone I see f2f, Kerstin_Forsberg, +46.4.67.7.aaee 20:33:35 Tom: working with Stan Huff and Joey Coyle from Intermountain 20:33:48 ...not-for profit serving Utah 20:34:13 ... need clinical models, and computerized decision support, automated data analysis 20:34:41 ... this requires standard data structure at a fine-granular level 20:35:12 ... one use case in Clin Decision Support is CT recruitment 20:35:57 ...Clinical Element Model (CEM): 12 yrs of effort; for data entry, decision logic, core services... 20:36:09 CEML is an XML model 20:36:34 ... Type, Key, and Value Choice 20:37:19 ... Data and Item (also a CEM) 20:38:51 ... use HL7 dv.3 ata types 20:39:06 s/dv.3 ata/v.3 data 20:39:25 .. can make collection of Observation items 20:40:01 .. qualifier give more info on value choices (recursive items) 20:42:13 ...do qualify the data or the subject? 20:42:33 Joey: qualifiers of modifiers is more complicated, incl. negation 20:43:01 alanr: the qualifier item may vary depending on how you use the data 20:43:05 alanr, how about { [ a :BunchOfMeasurements ; :bodyPos [ a :BodyPos ; :bodyPosition :sitting ] ; systolicBP "120"^^:mmHg ] } ? 20:43:42 Tom: data is usually 'controlled terminology codes' 20:43:56 how do you know that bodypos is of the same body as the bp is taken from 20:44:08 *you know* 20:44:10 ... eager to learn more about SW 20:44:13 but we want this explicit 20:44:30 s/:BunchOfMeasurements/:BunchOfBodyMeasurements/ 20:44:39 could be a rule - that would be fine 20:45:09 just makes the semantics of :bodyPos more obvious from its domain 20:45:16 yes 20:45:18 ... need codes + structure: numbness and location?? 20:45:50 ...dryweight vs weight with type 'dry' 20:46:55 ... try and define a single data storage model 20:49:03 eric: can all the groups decide one one clinical model standard in reasonable time, or should we be willing to work with multiple standards and mappings between them? 20:49:17 writing it three ways: [ a :BunchOfBodyMeasurements ; :weight [ a :DryWeight ; :massValue "70"^^kg ] ; :weight "70"^^:dryKg ] ; :dryWeight "70"^^si:kg ] 20:51:13 Tom: pre-coordinated single-name way of defining things (snomed style?), vs combinations of codes and values 20:52:22 ...UI and how the data is actually stored; hiding the various levels from the user 20:53:40 q+ 20:54:37 ack ericP 20:54:45 .. coded domains: measurements verses coded assrertions: when to use one or the other 20:55:57 ... things we seek: explicit model for CEM 20:56:09 ... single way to store and access data elements 20:56:27 ... interlingua for sharing modelled data between institutions 20:56:36 ... extensibility 20:57:10 ...apps the address data generally, i.e., new kinds of data 20:57:57 ... Status: partnering with GE healthcare, with func requirement sheet 20:58:17 ... would like to discuss creating the needed models 20:59:09 ... no data stored yet against these models 21:00:33 Ronen: how does the DCM relate to OpenEHR? 21:00:47 Tom: there is probably a lot of potential mappings 21:01:05 ... HL7 RIM templates are another example 21:01:16 i'm curious to learn how vipul's models differed consistently from HL7 21:01:46 Chimezie presentation 21:02:31 Chimezie: focus mainly on demo, but some background: SemanticDB is a CV registry for doing outcome research 21:02:46 ... to provide knowledge for decision support systems 21:03:32 ... more emphasis on longitudinal (life long) content, not just clinical records; how to avoid changing DB schemas each time? 21:03:51 ... provide a more flexible system built on XML and SW technologies 21:04:47 ... problems: fragmeted data; inflexible to extend; idiosyncratic terms; overlapping content so queries are difficult 21:05:18 Karen has joined #hcls 21:05:35 ... director gave CC a mandate: use better models for data handling; what is the next set of technologies? 21:07:08 alanr, sound like Exhibit + SPARQL to you? 21:07:24 yup 21:07:27 ... solutions: extending un-anticipated domains is important; should not rely on dev for UI- automated; expressive languages and drawing inferences; highly distributed; scalable to large data volume; use standards! 21:08:00 .. why CC joined W3C 21:08:28 ... XML to RDF using GRDDL and OWL 21:09:08 ...drop down driven terms for queries; use of xforms; 21:09:13 CC = Cleveland Clinit [distinguished from Creative Commons for benefit of later readers] 21:09:23 +??P11 21:09:35 thanks eric 21:10:15 ... quality of data is important for getting more use out of it 21:10:27 ...CyCorp reasoner used 21:10:59 -Kerstin_Forsberg 21:11:51 ... queries: NL strings to parsed phrases to SPARQL 21:12:25 ... extendinto new areas of research; use of local terminology; high-fidelity integration 21:13:34 ... Cyc Ontology- high level; several levels of logic including micro-theories 21:14:09 ... patient record core onto, at center of other terminologies: heart rhythm, meds, surgery, events... 21:14:20 q+ /win go 23 21:15:10 ... separation of domain goals from DB optimization issues (joins, normal forms) 21:15:30 ...demo now based at texas 21:18:08 - +46.4.67.7.aaee 21:18:30 +[IBM] 21:19:15 ... Cyc takes clauses and builds a SPARQL query from it; also optimizes it 21:19:41 but they are not using a very fast database/sparql implementation. Based on Mysq 21:19:43 mysql 21:20:54 would be interesting to test this in virtuoso 21:24:32 Chintan's presentation: 21:25:21 Chintan: At Biomedical semantics dept at Columbia working with IBM 21:25:51 ... ontological reasoning of Patient reasoning for CTs 21:26:14 ... recruitment is ain bottleneck (50% oftime) 21:26:21 ...low participation rates 21:26:37 ... lack of time by clinicians and patients 21:26:47 ... idea is to automate this better 21:27:04 ... data not well encoded digitally 21:27:45 ...CT criteria is not well expression -- semantic gap 21:28:12 ... electronic records to eligibility criteria mapping 21:29:08 ... our approach is to use knowledge in terminologies and ontologies: MED and SNOMED 21:29:21 MED = Medical Entities Dictionary 21:29:50 ... classificaiton of therapies and diseases 21:30:15 ... reason on data+terminology 21:31:20 ... TB reasoning as infectious agent and cause for meningitis 21:31:41 ...challenges: reasoning needsto scale large for terms 21:32:14 ... trials data is also large and queries tend to be very expressive 21:32:47 ...SHER is a scalable highly expressive reasoner uses of novel summarization approach 21:33:22 q+ to ask what the distinction is between local knowledge and domain knowledge 21:33:25 ...mappings of MED to SNOMED; clinical data repositoty (CDR) ETL'ed to RDF ABox 21:34:51 q+ also sparql owl inference enabled? 21:35:05 ... query formulator to input inclusion/exclusion criteria 21:36:19 ...local + domain knowledge mapping (89% MED concepts map to SNOMED) 21:37:57 q+ alanr to ask if their sparql owl is inference enabled 21:38:04 ... Incomplete data handling 21:38:56 ...increased response rate using sound but incomplete logic 21:39:33 ... now trying to integrate with Columbia's CTM systems 21:40:51 alanr: how much loss with incomplete reasoner? 21:41:12 Kavitha: we can combine both and get both benefits? 21:42:04 Josh: How does this comapre to rule systems? 21:42:25 s/Josh/Jyotish 21:43:09 brainstorm session now... 21:43:53 Topic: Brainstorming 21:44:05 Vipul: Goal: what can we do to move forward? 21:44:55 ... is this an attractive use case? 21:45:10 chimezie: doesn't intersect with our data, but it addresses VMR content 21:45:39 rachel: have a foot in HC and research worlds 21:46:06 ... CDISC folks not here, but appeard helpful to Brahn Kissler 21:46:56 june: need exists 21:47:17 ... we created a forum for families of patients of alz 21:47:29 ... we post results of about 7 studies 21:47:50 ... federally mandiated barrier between HC and research 21:48:02 ... considered an ethical breach 21:48:18 ... not sure this use case falls in the real world 21:48:25 -[IBM] 21:48:26 ... who would do it and how 21:48:50 vipul: ok if a machine referrs folks to studies? 21:48:59 june: grey area 21:49:23 george allen: other way to do it ... 21:49:41 alanr: to use cases, right? 21:50:05 ... 2nd where clinicial refers patient is ruled out 21:50:44 aaron: my org is currently working on cohort selection and cohort identification 21:51:26 june: is there a way for research selection criteria to feed back to acquired data 21:52:03 alanr: issue is whether there are too many to eval without feedback criteria 21:52:12 steve: good use case 21:52:21 ... isn't last demo the same? 21:52:32 vipul: clearly aspect of matching is there 21:52:59 ... but we are more ambitious 'cause we want to map to HL7 21:53:17 alanr: the use case description does not mention this 21:53:43 aaa: interesting to me 21:53:56 ... inclusion/exclusion rules are very broad 21:54:15 ... think there was an ex from rachel's presentation 21:54:45 rachel: e.g. ability to lie flat in this instrument 21:55:12 bbb: SOP: completely opaque 21:55:22 ... manhandle data out of EPRs into flat files 21:55:29 ... then search through flat files 21:55:37 ... could be an excel spreadsheet 21:56:23 s/bbb/rona/ 21:56:40 tom (IMH): qualified yes 21:57:07 ... transformation form UI model to storage model and other model archetypes more interesting to us right now 21:58:19 ccc: make sure you have examples for everything in the use case 21:58:40 s/ccc/joey/ 21:58:47 george: yes 21:58:59 ... have a masters informatics student now doing this 21:59:11 ... the research gives us the variables of interest 21:59:38 ... other vars tell us whether we want to approach this patient 21:59:53 alanr: you mean extra questions to ship to the interviewer? 22:00:07 george: yes -- clear up grey areas 22:00:19 susie: not in dev at Lilly 22:00:38 ... but, happy to share this UC with folks in dev 22:00:52 ... this seems like a good UC 22:01:27 don: yes, but we have a product that uses a totall diffn't technology (hidden markov models, ...) 22:01:55 ... need to id clinical DB and funder that will want to use that DB 22:02:52 ... have specifics of mapping before you go to far 22:03:06 vipul: need to make sure it's consumer-driven 22:03:16 dan: terrific 22:03:31 ... good thing: mapping to a specfic EMR system 22:03:51 ... the person who will do most of that mapping is the clinical research corrdinator 22:04:09 vipul: ok, so need intuitive interface 22:04:29 Karen: what do we need to show regulatory bodies 22:04:50 vipul: started with tbat, but reallized that reg agency was not involved 22:05:05 marcus: will bring to research arm 22:05:33 marcus: good use case 22:05:41 s/research/development 22:06:12 ... similar use ceses in social networks like telecom 22:06:26 vipul: not sure we'll have resorces to follow these up 22:06:50 eee [NIST]: haven't heard metrics for success 22:07:05 ... how much cross-over is possible 22:07:52 joanne: metric for success: integrator reliability 22:08:07 Adrian: Different data mining techniques used including e.g. logic based approaches such as inductive logic programming (ILP) 22:08:34 max: can see applicability of semweb here 22:08:43 dbooth: interested in feedback 22:11:07 alanr: blocking factor is whether we can get some patient data 22:11:20 ... but yes to this UC 22:11:32 Topic: data issue 22:11:42 joanne: yes to UC 22:12:02 ACTION: Vipul to get some relevent patient data 22:12:12 alanr: rachel analyzed criteria 22:12:41 ... might be easiest to release if you could pick out the criteria reducing the amoutn you have to anonymized 22:13:22 ... e.g. if it has no need of family history, we might be able to use deceased data 22:13:35 joanne: what's the inclusion criteria 22:13:45 rachel: we sampled five protocols 22:13:54 joanne: might want to make something up 22:14:18 eneumann: have synthetic data on a web page 22:14:57 http://www.cse.buffalo.edu/~szhong/papers/kanony.pdf 22:15:03 ... is there confidence in K-anonymization? 22:15:34 rona: doesn't work on small populations 22:15:41 see http://privacy.cs.cmu.edu/people/sweeney/kanonymity.html 22:16:48 vipul: how many do we need? 22:16:55 ericP why do we need more than three? 22:17:01 ericP: why do we need more than three? 22:17:12 vipul need to cover all of the selection criteria 22:17:20 vipul: need to cover all of the selection criteria 22:17:54 eneumann: recommend narrowing the first step 22:18:04 ... see what surprises come up 22:18:27 Topic: resources 22:18:40 vipul: need to know time commitments 22:18:53 ... george, your masters student 22:19:42 ACTION: will try to get a student? 22:19:57 ACTION: george will try to get a masters student? 22:20:08 joey: not prepared to commit at this point 22:20:55 rona: will try build up over next 5-6 months 22:21:04 ... would hope to get 1 22:21:33 s/rona/ronen 22:21:33 s/rona/ronen/ 22:21:53 scott: if we get involved, would be on a mapping issue 22:22:13 ... ... using associated text resources 22:22:33 ... 1 student if we get involved 22:22:45 eneumann: does your group interact with T&O? 22:22:53 scott: yes 22:24:04 ... lots of folks in AMS involved in ontology mapping but i don't command their resoruces 22:24:17 ... can ask them if they would join us 22:25:02 ... lots of resources, but it's a young technology 22:25:16 ronen: time frame? 22:25:51 eneumann: have this done, say, in a year? 22:26:21 vipul: Banff project took a year 22:27:39 joanne: would like to capture the process 22:28:53 aaron: can't give time commitment 22:29:12 Michal Z: train 20 students, can probably persueade 2 22:29:24 Rachel: maybe 1day / week 22:31:32 chimezie: can't promise -- lack of time and divergent tragectories 22:32:10 susie: can participate 22:32:38 ... if it becomes a well-scoped project, and Lilly dev folks are interested, i can probably fund a student 22:33:14