1Monitoring6.872/HST950Peter SzolovitsOutline• Problem of information overload• How to reason with/about time• Interpreting temporal dataProblem• ICU alarms sound roughly every 30 seconds, in a typical (full) ICU• Nurse takes ~minutes to resolve alarm• How to resolve?– Ignore (turn off) alarms– Prioritize– Automate– Make alarming algorithms more intelligentTime is Critical• Some systems have no explicit representation of time– E.g., Internist• <ABDOMEN TRAUMA RECENT>• <ABDOMEN TRAUMA REMOTE HX>• <CHEST PAIN SUBSTERNAL LASTING GTR THAN 20 MINUTE <S>>• <CHEST PAIN SUBSTERNAL LASTING LESS THAN 20 MINUTE <S>>• No representation of these pairs being “the same,” but at different times.• Note: Same problem with space; need orthogonality!Motivating Example: Distinguishing Four Possible RelationshipsBetween Transfusion and Jaundice• Post-transfusion antigen incompatibility hemolytic anemia• Post-Transfusion Hepatitis B: Acute Hepatitis• G-6-PD hemolytic anemia treated by transfusion• Post-Transfusion Hepatitis B: Chronic Activebloodtransfusionabdominalpainjaundice??Even Simple Models Help• PIP’s temporal model:– PAST, RECENT, NOW, SOON, FUTURE•Example: – in “chronic glomerulonephritis” model, “past acute GN”– (to my surprise), program hypothesized “future chronic GN” after diagnosing “now acute GN”AGNCGNpast present futurepast present future2What is Time?• (Macroscopically) unidirectional• Related to causality• May be modeled in various ways– Continuous quantity, as in differential equations– Discrete time points, as in discrete event simulations– Intervals, as in ordinary descriptions of durations, processes, etc.… or combinationsContinuous View• Differential equation view of world• States (state variables) evolve according to their laws002022xtvgtxvgtdtdxgdtxd++−=+−=−=Discrete Event View• Designated (countable) time points• Nothing “interesting” between events• Events may be defined by– Clock “ticks”– Interactions among objects in universe– Distinguished points in representation of state variables (e.g., highest point of cannon shell)What Can Be a Time Point?• Calendrical point—a specific date/time• Recognizable event—e.g., “when I had my tonsils out,” or “start of high school,” or “my ninth birthday”• Now—special, because it movesDiscrete Events areAssociated with State TransitionsE.g., Beck & Pauker’s model to help compute quality-adjusted years of survival:Constraint Propagationamong Time Points• Clearly, T(A,B)+T(B,C) = T(A,C)• But we only know lower/upper bounds on T• L(A,C) = L(A,B)+L(B,C)• U(A,C) = U(A,B)+U(B,C)• and thus, we can infer relationshipsA CBl, ul, ul, u3Interval “Overlaps” in TUPeseMost medical history temporal terminology is expressible in statements composed from TUP assertions.ANOREXIA><IRRITABILITY<ANOREXIA+´+++0-´-++0+´++´+IRRITABILITY>εεεIntervals and Points areAlternate Representations• “Overlaps” defined in terms of its endpoints:<Irritability Irritability><Anorexia Anorexia>+0,+∞+0,+∞+ε,+∞ +ε,+∞+ε,+∞+ε,+∞Initial Assertions• Completing all the relations not explicitly assertedANOREXIA>Externally asserted<ANOREXIA<IRRITABILITY3 DAYS2 DAYS7 DAYS5 DAYSLegendInferred3 DAYS4 DAYSConstraint<ANOREXIA<IRRITABILITY ANOREXIA>5 DAYS5 DAYS2 DAYS3 DAYS3 DAYS3 DAYSPropagated Constraint• Order n2edges in fully interconnected graph• Order n3computation• Work to localize propagation to semantically related events<ANOREXIA<IRRITABILITYANOREXIA>5 DAYS5 DAYS3 DAYS2 DAYS2 DAYS3 DAYSForms of Temporal Uncertainty• Lower/upper bounds on temporal distances• Central range + fringe• Continuous distributions4Interval View• Activities, processes take place over extended intervals of time• Observations are true over periods of time• Systems remain in steady (from some viewpoints) states over intervalsAllen’s Temporal IntervalsInference among Intervalsby CompositionX-YY-Ze.g., if A starts B and B overlaps C,what are the possible relationships between A and C?1. A before C2. A meets C3. A overlaps CTemporal Control Structure. -T. Russ• Processes maintaining truth of abstractions over specified interval• Update of past beliefs from corrections or new data.• Actions are permanent.Back to Monitoring• Detecting Trends• Language for Trend Description• Matching algorithms• Top-down vs. bottom-up vs. both• Learning trend detectors“Two-Point” Trend Detectors• Restricted to hospitals with the most complete information systems• Rind & Safran, 1992• Two point event detector– rise in creatinine > 0.5 mg/dl• Therapeutic context– Renally cleared or nephrotoxic medication– Possible care providers• Implementation– M procedures linked to E-mail5BIH Experience (cont’d)• Time series trial• 607 in 348 admissions during control periods• 497 events in intervention period• 369 alerts, sent to 584 different physicians, 9.25 recipients per alert• Improved response time• Improved outcomesRepresentationEasily implemented as an Arden Syntax MLMAre these two-point trend detectors sufficient?• If not, why not?• Noisy data.• Multi-phased processes. • Uncertainty over time. • Uncertainty over values.Issues in Trend Detection• Defining significant trends– Multiple variables– Multiple phases– Temporal and value uncertainty• Detecting trends from data• Generating alarms• Displaying, explaining results• Changing clinical contextPediatric GrowthMonitoringPediatric GrowthMonitoring• Data:- heights, weights - family history - bone ages - pubertal data, stages- hormone values• Disorders show characteristic patterns on growth chart.Boy with constitutional delay--HaimowitzCurve Fitting Approacha11 + e-b1 (t - c1)a31 + e-b3 (t - c3)a21 + e-b2 (t - c2)++Height(t) = ak= component k’s contribution to mature staturebk= a parameter proportional to the maximum growth velocity of the component (maximum rate of growth is (a * b) /4 centimeters per year)ck= the age in years at which the maximum growth velocity occursTriple-logistic curve [Thissen and Bock 1990]Describing Average Normal GrowthDescribing Average Normal Growth• Def. Z-score ≡ Number of standard deviations a patient's parameter is from the mean for that age.• From birth until age 2 - 3 years, height and weight vary together and establish baseline