Please note the position of...

Please note the position of QEXL/Q-UEL consortium of universities and small companies in regard to underlying theoretical principles of clinical decision support.

Robson, B. and Boray, S. “Implementation of a web based universal exchange and inference language for medicine. Sparse data, probabilities and inference in data mining of clinical data repositories”. Computers in Biology and Medicine (in press, preprint downloadable http://www.computersinbiologyandmedicine.com/article/S0010-4825(15)00257-7/abstract)

Robson, B., Caruso, T, and Balis, U. G. J. (2014) “Suggestions for a Web Based Universal Exchange and Inference Language for Medicine. Continuity of Patient Care with PCAST Disaggregation. Computers in Biology and Medicine, 56: 51–66. “We describe here the applications of our recently proposed Q-UEL language to continuity of patient care between physicians, specialists and institutions as mediated via the Internet, giving examples derived from HL7 CDA and VistA of particular interest to workflow. Particular attention is given to the Universal Exchange Language for healthcare as requested by the US President׳s Council of Advisors on Science and Technology (PCAST) released in December 2010, especially in regard to disaggregation of the patient record on the Internet…”

Robson, B. (2014) “ Hyperbolic Dirac Nets for Medical Decision Support. Theory, Methods, and Comparison with Bayes Nets” Computers in Biology and Medicine, in 2014 Aug;51:183-97. “We recently introduced the concept of a Hyperbolic Dirac Net (HDN) for medical inference on the grounds that, while the traditional Bayes Net (BN) is popular in medicine, it is not suited to that domain: there are many interdependencies such that any "node" can be ultimately conditional upon itself. A traditional BN is a directed acyclic graph by definition, while the HDN is a bidirectional general graph closer to a diffuse "field" of influence. Cycles require bidirectionality; the HDN uses a particular type of imaginary number from Dirac׳s quantum mechanics to encode it. Comparison with the BN is made alongside a set of recipes for converting a given BN to an HDN, also adding cycles that do not usually require reiterative methods. This conversion is called the P-method. Conversion to cycles can sometimes be difficult, but more troubling was that the original BN had probabilities needing adjustment to satisfy realism alongside the important property called "coherence". The more general and simpler K-method, not dependent on the BN, is usually (but not necessarily) derived by data mining, and is therefore also introduced. As discussed, BN developments may converge to an HDN-like concept, so it is reasonable to consider the HDN as a BN extension.”

Deckelman, S., and Robson, B. (2014). “Split-Complex Numbers and Dirac Bra-Kets”, Vol. 14:3, 135-149, Communications in Information and Systems (CIS).

Robson, B. (2014) “POPPER, a Simple Programming Language for Probabilistic Semantic Inference in Medicine.” Computers in Biology and Medicine, 56: 107-23 “Our previous reports described the use of the Hyperbolic Dirac Net (HDN) as a method for probabilistic inference from medical data, and a proposed probabilistic medical Semantic Web (SW) language Q-UEL to provide that data. Rather like a traditional Bayes Net, that HDN provided estimates of joint and conditional probabilities, and was static, with no need for evolution due to "reasoning". Use of the SW will require, however, (a) at least the semantic triple with more elaborate relations than conditional ones, as seen in use of most verbs and prepositions, and (b) rules for logical, grammatical, and definitional manipulation that can generate changes in the inference net. Here is described the simple POPPER language for medical inference. It can be automatically written by Q-UEL, or by hand. Based on studies with our medical students, it is believed that a tool like this may help in medical education and that a physician unfamiliar with SW science can understand it. It is here used to explore the considerable challenges of assigning probabilities, and not least what the meaning and utility of inference net evolution would be for a physician.”

Robson, B., Caruso, T, and Balis, U. G. J. (2013)”Suggestions for a Web Based Universal Exchange and Inference Language for Medicine”, Computers in Biology and Medicine,1;43(12):2297-310.

like0

Comments

Two final synopses refelecting our Q-UEL-based thinking on CDS.

Barry Robson's picture

May I take the liberty of providing two more synopses of two recent efforts that should round off a quick overview of our thinking? That should complete our input for a while! These provide more overview of our research concerning what it takes to have a web-based learning and automated reasoning system for medical decision support.

Robson, B. and Boray, S. (2015), Implementation of a web based universal exchange and inference language for medicine. Sparse data, probabilities and inference in data mining of clinical data repositories.  Computers in Biology and Medicine (in press). “We extend Q-UEL, our universal exchange language for interoperability and inference in healthcare and biomedicine, to the more traditional fields of public health surveys. These are the type associated with screening, epidemiological and cross-sectional studies, and cohort studies in some cases similar to clinical trials. There is the challenge that there is some degree of split between frequentist notions of probability as (a) classical measures based only on the idea of counting and proportion and on classical biostatistics as used in the above conservative disciplines, and (b) more subjectivist notions of uncertainty, belief, reliability, or confidence often used in automated inference and decision support systems. Samples in the above kind of public health survey are typically small compared with our earlier “Big Data” mining efforts. An issue addressed here is how much impact on decisions sparse data should have. We describe a new QUEL compatible toolkit including data analytics application DiracMiner that also delivers more standard biostatistical results,  DiracBuilder that uses its output to build Hyperbolic Dirac Nets (HDN) for  decision support, and HDNcoherer that ensures that probabilities are mutually consistent. Use is exemplified by participating in a real word health-screening project, and also by deployment in a industrial platform called the BioIngine, a cognitive computing platform for health management.”

Robson, B. (2015), MARPLE, a Hyperbolic Dirac Net for Probabilistic Semantics in Medical Decision Support.Theory, methods, and comparison with IBM's Watson. (in preparation). “MARPLE stands for Medical Automated Reasoning Programing Language Environment. It is a suite of Q-UEL software applications, i.e. using Q-UEL tags for communications and reasoning operations, for more advanced reasoning. The aim continues to be that of Q-UEL, i.e. development of a probabilistic Semantic Web or “Thinking Web”, which learns and reasons, although Q-UEL also turns out to be a powerful local tool.  The language used is a more advanced derivative of POPPER (Robson, 2015), combined with the use of a new generation of Q-UEL XTRACT autosurf and spawn tags ( Robson, Caruso and Baylis; 2103). Unlike IBM’s Watson, Q-UEL Dirac bra-operator-ket statements < subject expression | relationship expression | object expression> are fundamentally probabilistic, wherever possible, from the outset, in the manner suitable for  medical metrics. They also follow the hyperbolic-complex or dual probability concept of the Hyperbolic Dirac Net  (Robson, 2013) , allowing use of probabilistic knowledge networks that are general graphs, and reasoning as graph evolution by metastatements in the POPPER-like manner. Also, automated inference is distributed across the Internet and on local end-user devices that do not need to be high performance machines."

like0

Proposal for target CDS test case.

Barry Robson's picture

This is a comment on the quality test case. One can make predictions and test quality in e.g. ROC curve represntation, but ideally, an advanced CDS, as opposed to a simple data mining and/or analytic portal, should at least be able to pass a medical qualification examination, say USMLE, but (to be more realistic for use in practice) with more than 5 answer options. For typical exam questions of that newer preferred type; see for example S. M. Case and D. B. Swanson, Constructing Written Test Questions for the Basic and Clinical Sciences (Third edition), National Board of Medical Examiners (2015), from which the following is taken. A. Ankylosing spondylitis B. Intervertebral disc infection C. Multiple myeloma D. Myofascial pain E. Osteoporosis F. Spinal stenosis G. Spondylolysis H. Tuberculosis of the spine For each patient with back pain, select the most likely diagnosis. A 26-year-old man has insidious onset of low back pain and early morning stiffness.The pain alternates from side to side and occasionally radiates into the buttocks and back of the thighs, but not below the knees.The patient has acute anterior uveitis, diffuse low back and sacroiliac tenderness, and restricted range of motion at the hips.His erythrocyte sedimentation rate is 40mm/h; latex fixation test is negative; and mild hypoproliferative anemia is present. By way of example, prototype Q-EL CDS systems extract e.g. < patient age(years):=26 | has | pain:=(when(day):=morning, where(anatomy):=(diffuse:=‘low back’:, radiates:=(buttocks, thighs), tenderness:= sacroiliac) and infection:= uveitis and anemia:=(mild, hypoproliferative) ‘erythrocyte sedimentation rate(mm/h)’:= 40 > and so far hit the strongest clue in the following comorbiity association, consistent with answer A. < ‘acute anterior uveitis’ Pfwd:=0.3 | if | ‘ankylosing spondylitis’ Pbwd:=0.2 >

like0