INFECTIOUS DISEASE MODELLING AND HEALTH ECONOMICS Flashcards
WHAT IS INFECTIOUS DISEASE MODELLING?
A TOOL THAT CAN BE USED TO STUDY THE MECHANISMS BY WHICH DISEASES SPREAD, PREDICT FUTURE OUTBRAKS AND/OR TO EVALUATE STRATEGIES TO CONTROL/PREVENT DISEASE
WHAT IS AN (INFECTIOUS DISEASE) MODEL?
- MODEL IS A SET OF MATHEMATICAL AND STATISTICAL TECHNIQUES DESIGNED FOR THE ANALYTICAL EVALUATION OF DECISION MAKING UNDER UNCERTAINTY
- UNCERTAINTY IS REPRESENTED BY EVENT PROBABILITIES AND ASSOCIATED DISTRIBUTIONS
- MODELS USED WHEN THERE IS LIMITED OR INCOMPLETE CLINICL TRIAL DATA OR WHEN WE WANT TO EXPAND ON THE KNOWLEDGE FROM TRIALS
- MODELS SOMETIMES USED TO COMPARE ALTERNATIVE COURSES OF ACTION INSTEAD OF RELYING ON CLINICAL TRIALS (WHICH ARE EXPENSIVE AND TIME CONSUMING)
- MODELS COMBINE MEDICAL AND ECONOMICAL EVIDENCE, INFORMING BEST DECISION/OUTPUT WITH LIMITED RESOURCES
EXMPLES OF MODELS?
- DECISION TREE
- MARKOV MODEL
- OTHERS (PATIENT LEVEL SIMULATION, DISCRETE EVENT SIMULATION, DYNAMIC MODELS)
DYNAMIC MODELS ARE USUALLY USED FOR INFECTIOUS OR NON INFECTIOUS DISEASE?
INFECTIOUS D
DESCRIBE THE MODEL: DECISION TREE
- PRESENTS SERIES OF POSSIBLE CONSEQUENCES AND COST ASSOCIATED WITH EACH ONE
- MOVE FROM LEFT TO RIGHT
- USED FOR E.G. A SCREENING PROGRAM OR VACCINATION PROGRAM
- USUALLY FOR INTERVENTIONS WITH SHORT DURATION (UNDER A YEAR)
DECISION TREE MODEL IS USED FOR INTERVENTIONS WHICH LAST LESS THAN?
1 YEAR
MARKOV MODEL VS DECISION TREE MODEL; WHICH ONE IS USED FOR INTERVENTIONS WITH LONGER DURATIONS?
MARKOV M
WHAT ARE ‘EXTERNALITIES’ IN DISEASE MODELLING?
- ANY COST OR BENEFIT OF ONE PARTY THAT MIGHT IMPACT ON OTHERS
- IGNORING EXTERNALITIES IN MODELLING MAY LEAD TO INCORRECT RESULTS AND THUS WRONG DECISION
- CAN BE NEGATIVE (LIKE SMOKING) OR POSITIVE (PERSON VACCINATED, PERSON SCREENED ASYMPTOMATICALLY, TREATMENT FOR SYMPTOMATIC PEOPLE..)
WHY HERD IMMUNITY FOR COVID MIGHT BE IMPOSSIBLE?
- VACCINE ROLLOUT UNEVEN
- NEW VARIANTS CHANGE ERD IMMUNITY
- IMMUNITY MAY NOT LAST FOREVER
- VACCINES CHANGE HUMAN BEHAVIOUR
- VACCINES NOT COMPLETEY EFFECTIVE
HERD IMMUNITY?
THE DIRECT PROTECTION EXPERIENCED BY UNVACCINATED INDIVIDUALS RESULTING FROM PRESENCE OF IMMUNE INDIVIDUALS IN A POPULATION
STEPS: HOW ARE MODELS SET UP?
- MODELS ARE SIMPLIFICATIONS OF REALITY
- SET UP USING A SOFTWARE OF ONE’S CHOICE
- MODEL BUILDING IS AN ITERATIVE PROCESS: STEPS CAN BE REVISITED
STEPS: - IDENTIFY THE RESEARCH QUESTION
- IDENTIFY RELEVANT FACTS ABOUT THE INFECTION
- CHOOSE THE MODEL STRUCTURE
- IDENTIFY MODEL INPUT PARAMETERS
- SET-UP MODEL
- MODEL VALIDATION
- PREDICTION AND OPTIMIZATION
WHICH RELEVANT FACTS ABOUT THE INFECTION NEED TO BE IDENTIFIED FOR INFECTIOUS DISEASE MODELLING?
- WHAT IS THE LATENT/PREINFECTIOUS PERIOD (TIME FROM BEING INFECTED TO BECOMING INFECTIOUS)
- HOW LONG ARE PEOPLE INFECTED
- WHAT IS THE BASIC REPRODUCTION NUMBER, R0
- ARE ALL AGE GROUPS AFFECTED EQUALLY
WHAT HAS TO BE COONSIDERED WHEN CHOOSING THE MODEL STRUCTURE FOR INFECTIOUS DISEASE MODELLING?
- NATURAL HISTORY OF THE INFECTION (MODEL STRUCTURE SHOULD REFLECT THE NATURAL HISTORY OF A DISEASE, SHOWING ALL PERIODS!!!!!!!!!!!)
- ACCURACY AND TIME PERIOD OVER WHICH MODEL PREDICITIONS ARE REQUIRED (MAY NEED TO INCORPORATE KEY ASPECTS OF DEMOGRAPHY LIKE BIRTH DATE, DEATHS AND MIGRATION; IF LOOKING AT LONG TERM IMPACTS OF INFECTION) + (MAY NEED TO CONSIDER DIFFERENT STRAINS OR CHANGES TO THE POPULATION’S IMMUNITY TO THE DIFF STRAINS)
- RESEARCH QUESTION (THE MODEL STRUCTURE IS A FUNCTION OF THE RESEARCH QUESTION, SHOULD BE CLEARLY DEFINED)
WHAT DOES THE ‘SEIR’ MODEL OF INFECTIOUS DISEASE MODELLING STAND FOR?
SUSCEPTIBLE—>PREINFECTIOUS (EXPOSED)—>INFECTIOUS—> RECOVERED
WHAT ARE DETERMINISTIC MODELS?
MODELS THAT DESCRIBE WHAT HAPPENS ON AVERAGE IN A POPULATION (SAME RESULTS EACH TIME THEY ARE RAN)
WHAT ARE STOCHASTIC MODELS?
THEY ALLOW THE NUMBER OF INDIVIDUALS WHO MOVE THROUGH COMPARTMENTS TO VARY THROUGH CHANCE
- USEFUL FOR ASSUMPTIONS AND WHEN THERE IS A LOT OF UNCERTAINTY IN THE MODEL
- INCLUDE RANGS OF VALUES, AND DIFFER EACH TIME THEY RUN
TYPES OF MODELS?
- COMPARTMENTAL (PEOPLE CLASSIFIED INTO WELL KNOWN CLEARLY DEFINED CATEGORIES)
- INDIVIDUAL BASED OR MICROSIMULATION
- TRANSMISSION DYNAMIC OR DYNAMIC TRANSMISSION (ALLOWS CONTACT BETWEEN PEOPLE IN THE MODELS, MOSTLY FOR INFECTIOUS DISEASE)
- STATIC (NO MIXING BETWEEN THE INDIVIDUALS)
- NETWORK (OFTEN USED FOR STIs, VERY COMPLEX)
EAMPLES OF INFECTIOUS DISEASE MODEL INPUT PARAMETERS?
- RISK THAT SUSCEPTIBLE PEOPLE BECOME INFECTED BETWEEN A SPECIFIED TIME POINT
- PROPORTION OF INFECTED PEOPLE WHO BECOME INFECTIOUS
- PROPORTION OF INFECTIOUS INDIVIDUALS WHO RECOVER (I.E. BECOME IMMUNE, BUT THENMIGHT HAVE TO CONSIDER HOW LONG THE IMMUNITY LASTS)
- RELIABLE DATA
- UNCERTAINTY IN THE DATA
WHAT IS FORCE OF INFECTION (λ)?
THE RATE AT WHICH SUSCEPTIBLE INDIIDUALS BECOME INFECTED PER UNIT OF TIME/ THE PROBABILITY PER UNIT TIME THAT A SUSCEPTIBLE PERSON ACQUIRES AN INFECTION
BY MULTIPLYING THE FORCE OF INFECTION BY THE NUMBER OF SUSCEPTIBLE PEOPLE, WE GET?
RATE OF INFECTION
BASIC REPRODUCTION NUMBER (R0; R NAUGHT) INTERPRETATION?
- THE HIGHER THE NUMBER, THE MORE INFECTIOUS A CONDITION IS
R0=1, THE INCIDENCE OF THE DISEASEIN THE POPULATION REMAINS THE SAME, AND THERE IS NO CHANGE
R0<1, THE INCIDENCE OF THE DISEASE WILL DECREASE, AND THE DISEASE WILL EVENTUALLY BE ELIMINATED
R>1, THE INCIDENCE OF THE DISEASE WILL EXPONENTIALLY INCREASE
BASIC REPRODUCTION NUMBER IS…
AVERAGE NUMBER OF SECONDARY INFECTIOUS PERONS RESULTING FROM ONE INFECTIOUS PERSON FOLLOWING INTRODUCTION INTO A TOTALLY SUSCEPTIBLE POPULATION
NET REPRODUCTION NUMBER IS… (Rn)
AVERAGE NUMBER OF SECONDARY INFECTIOUS PERSONS RESULTING FROM ONE INFECTIOUS PERSON IN A GIVEN POPULATION (WHERE SUSCEPTIBILITY VARIES, NOT THE WHOLE POPULATION IS SUSCEPTIBLE/ NOT TOTALLY SUSCEPTIBLE)
EXAMPLES OF PROGRAMS/SOFTWARES THAT CAN BE USED FOR SETTING UP MODELS?
EXCEL, C++, R, PYTHON, BERKELEY MADONNA…
COST EFFECTIVENESS TRESHOLD USED BY NICE?
20,000-30,000 POUNDS PER QALY
WHY AND WHEN WAS NICE ESTABLISHED?
IN 1999 TO TOP THE POSTCODE LOTTERY IN HEALTHCARE
OPPORTUNITY COST?
THE BENEFIT YOU GIVE UP DUE TO NOT BEING ABLE TO SPEND MONEY ON MORE THAN ONE THING (LOST BENEFITS OF ALTERNATIVE COURSES OF ACTION)
COST EFFECTIVENESS PLANE?
X: EFFECT DIFFERENCE
Y: COST DIFFERENCE
SOUTH EAST QUADRANT: INTERVENTION LESS COSTLY BUT MORE EFFECTIVE
ALL QUANDRANT EXCEPT FOR THE NORTH WEST ONE REPRESENT ATTRACTIVE/POSSIBLE OPTIONS AND THE DECISION IS MADE BASED ON THE MAKER’S PREFERENCES; WILLINGNESS TO COMPROMISE PRICE OR EFFECTIVENESS IN TERMS OF HEALTH
MAIN AIM OF HEALTH ECON?
MAXIMISE HEALTH OF POPULATION WITH SCARCE RESOURCES
TYPES OF ECON EVALUATION?
COST EFFECTIVENESS ANALYSIS: EFFECTS MEASURED IN NATURAL UNITS (E.G. NUMBER OF CASES DETECTED), COST/LYG (LIFE YEARS GAINS), USED FOR E.F. DIFF TREATMENTS OF THE SAME DISEASE
COST UTILITY ANALYSIS: EFFECTS VALUED AS UTILITY IN QALYs, COST/QALY, CAN COMPARE DIFFERENT DISEASES
COST BENEFIT ANALYSIS: MONETARY VALUATION PLACED ON HEALTH OUTCOMES, COST:BENEFIT RATIO (BOTH IN POUNDS)
POSSIBLE PERSPECTIVES/VIEWPOINTS OF ECON ANALYSIS?
- INDIVIDUAL (WOULD INCLUDE E.G. TRAVEL COSTS, TIME OFF WORK)
- ORGANISTION (E.G. HOSPITAL)
- NHS (MOST COMMON)
- SOCIETAL
DATA SOURCES FOR ECON EVALUATION?
- NHS REFERENCE COSTS
- RESEARCH LITERATURE
- BRITISH NATIONAL FORMULARY
- EXPERT OPINION
(IF DATA FROM LITERATURE IS OLD, MAKE SURE TO INFLATE IT TO CURRENT TIMES + ALL DATA NEEDS TO BE FROM THE SAME YEAR + BETTER TO RELY ON EXPERTS THAN PROBABILITIES)
INCREMENTAL COST-EFFECTIVENESS RATIO?
The incremental cost-effectiveness ratio (ICER) is a statistic used in cost-effectiveness analysis to summarise the cost-effectiveness of a health care intervention. It is defined by the difference in cost between two possible interventions, divided by the difference in their effect
ICER=(COSTa-COSTb)/(EFFECTa-EFFECTb)
WHAT IS SENSITIVITY ANALYSIS?
A SET OF TECHNIQUES USED TO ANALYSE HOW SENSITIVE RESULTS ARE TO UNCERTAINTY
