Two information issues in injury surveillance: 1 ...

Two information issues in injury surveillance: 1 ...

Injury indicators: purposes, progress, prognosis James Harrison AIHW National Injury Surveillance Unit Research Centre for Injury Studies Flinders University - Adelaide - South Australia Overview Injury indicators Purposes: Why have indicators? Principles: Concepts, definitions & standards Progress: Where are we up to? Problems: Constraints Prognosis: Next steps AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University Purposes: Why have indicators?

Information to support injury prevention & control Support planning and policy-making Guide targeting and priority-setting by Measuring and monitoring injury, its consequences, injury risk-factors and exposure to them (Contributing to) evaluation of interventions AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University Purposes: Origin of indicators Steamengine Pressuretime indicator Control valves Operator

AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University Purposes: Origin of indicators Injury prevention program Injury indicator Start or alter prevention program Injury prevention practitioners AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University What to indicate? Things worth knowing Steam-engine Risk of explosion Efficiency Effect of altered settings

Injury prevention Burden of injury Risk exposure Effect of intervention especially if not directly observable AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University How to indicate it? A. Identify and specify another thing, which: 1. Can be measured 2. Varies with the thing worth knowing Steam-engine Pressure-time relationship Injury prevention Injury incidence Prevalence of risk exposure Reach of an intervention B. Make a device / information system able to provide measured values of A. Steam-engine indicator Injury indicator AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University

What is a good indicator? One with attributes suitable for its purpose Steam-engine indicator Varies predictably with timepressure Not much influenced by other things Sensitive enough to show important changes Timely enough to allow good control of engine Injury prevention indicator Varies predictably with injury occurrence Not much influenced by other things Sensitive enough to show important changes Timely enough to allow effective response to findings Well-defined Comparable (esp. over time) AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University What is a good indicator? I Information objective

Assessment as topic for indicator Trends in total burden of injury Broad scope is appealing Lack information for reliable measurement, especially over time Trends in incidence of fatal Narrower scope: may not be a reliable and severe injury guide to (eg) trends in burden of total Measurement is feasible AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University Principles: concepts, definitions, standards Indicators require data Necessary but not sufficient Must be COMPARABLE Foundations for comparability

Clear meaning for key concepts: eg. injury, severity, outcome Classifications Operational definitions eg. Recommended Framework for reporting injury mortality data Derived measures eg. STIPDA Consensus Recommendations for using hospital discharge data for Injury Surveillance Reporting standards eg. ICD-10, ICD-10-AM, ICECI

eg. ICISS severity, remoteness, Socio-Economic Status Institutional arrangements eg. WHO-FIC, Injury ICE, STIPDA AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University Injury outcome indicators Consensus criteria (ICE on Injury Statistics) Define in terms of anatomical/physiological damage Serious injury Case ascertainment independent of extraneous factors Representative of target population Available data Well-documented (definitions, methods, etc) (Cryer, Langley, et al 2005) AIHW National Injury Surveillance Unit

Research Centre for Injury Studies, Flinders University Progress: Where are we up to? Australia as an example Focus on trends in severe & fatal injury c. 10 years at national level Deaths and hospitalised injury Onion metaphor: layers of work (iterative) Getting, understanding, cleaning & documenting data Scope: defining and operationalising definition of injury Methods to avoid multiple counting Methods to assess & control changing sample fraction AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University Full specification can be complex Incident fatal and severe injury due to Transport Fatal cases Deaths occurring in Australia during the year from 1 July 2002 and registered

within six months of the end of that year where any Multiple Cause of Death code is in the range S00-T75 or T79 and Underlying Cause of Death is in the range V01-V98. If Underlying Cause is not an External Cause, then include if (any) (first-mentioned external cause) Multiple Cause of Death code is in the range V01-V98 and the person was a resident of Australia. Non-fatal highthreat to life cases Episodes of admitted patient (acute) care in any acute-care hospital that (commenced) (ended) during the year from 1 July 2002 where (principal) (any) diagnosis code is in the range S00-T75 or T79 and (any) (the corresponding) external cause code is in the range V01-V75 or V79 and the person was not admitted following transfer from an acute care hospital and the (all injuries multiplicative) (worst injury) ICISS severity score is < 0.941 and the person is recorded as residing in Australia and the episode ended with the patient alive. Population Estimated resident population at 31 December 2000. Rates Numerator values are sums of fatal cases and the non-fatal instances of types of injury that pose high threat to life. Rates are adjusted for age by the direct method using the Australian population in 2001 as the reference (and stratified or adjusted for remoteness of place of residence and Indigenous status). AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University

but different methods can give very different values How many deaths following injury due to a fall are recognised as such? Of deaths registered in Australia in 2002: Underlying cause of death coded to Accidental Falls (W00-W79) 629 Underlying cause of death coded to Unspecified cause (X59) with a fracture as an additional cause 888 Mostly falls by people aged 75 and older at death Cause unspecified mainly because ABS does not seek one at this age Underlying cause of death coded to Natural cause (i.e. not injury) with additional cause codes for injury condition(s) and for fall as external cause of injury 1,518 Underlying cause is coded to circulatory disease in 49% Small differences in wording on death certificates affect whether coded as injury deaths. AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University even if differences are subtle Injury all ages (NHPA indicator 1.2) Australia 1993-94 to 2002-03

1,400 1,200 cases/100,000 pop 1,000 all separations 800 Excl transfers from acute hospitals 600 Excl transfers to acute hospitals 400 200 2002-03 2001-02 2000-01 1999-00 1998-99

1997-98 1996-97 1995-96 1994-95 1993-94 0 Year of separation Age-adjusted rates AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University Has the age-adjusted rate risen? Falls ages 75+ (NHPA indicator 1.2) Australia 1993-94 to 2002-03 4,000 3,500 cases/100,000 pop 3,000

2,500 all separations 2,000 Excl transfers from acute hospitals Excl transfers to acute hospitals 1,500 1,000 500 2002-03 2001-02 2000-01 1999-00 1998-99 1997-98 1996-97 1995-96

1994-95 1993-94 0 Year of separation Age-adjusted rates AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University Aspects of burden of hospitalised injury Hospitalisation due to Accidental Falls ages 65 and older, Australia 1995-6 to 2002-3 2,500 60,000 2,000 1,500 40,000 30,000 1,000 20,000 500

10,000 Cases Crude rate Adj. Rate 2002-03 2001-02 2000-01 1999-00 1998-99 1997-98 0 1996-97 0 1995-96 Cases 50,000

Cases/100,000 population 70,000 AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University Aspects of burden of hospitalised injury Hospitalisation due to Accidental Falls ages 65 and older, Australia 1995-6 to 2002-3 1,200,000 1,000,000 No injury Dx; Fall ext. cause Other (AdDx injury) 600,000 Rehabilitation (AdDx injury) Injury (PrDx) 400,000 200,000 2002-03 2001-02 2000-01 1999-00

1998-99 1997-98 1996-97 0 1995-96 Bed-days 800,000 AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University Problems: Constraints Data quality (incl timeliness) Difficulty and costliness of validation studies

special studies; WA linked data Data (mostly) not person-linked Variation in % of cases incident in population admitted to hospital? (Especially for low severity injuries?) >episode (ie record) / incident case of admitted injury: tricky to allow for this source of multiple counting of cases hospital data and deaths data: do in-hospital injury deaths appear as injury deaths in national deaths data collection? Access (administrative issues, privacy) Complexity (esp for clever data cleaning) Classification Australian clinical modification of ICD-10 (ICD-10-AM) revisions (each 2y) allow useful enhancement of Injury & External Cause codes but (so far) the revised codes are not used for routine mortality data

AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University Signal & noise: example 1 Nature Injury (principal diagnosis) All 1993-94 to 2002-03 75 ye ars and olde r 6000.0 10.0 4000.0 3000.0 ratio (series 1:2) rate (/100,000 pop) 5000.0 A ustralia separations A ustralia cases (est.) sub-area A cases (est.) 2000.0 ratio [1:2]

1000.0 0.0 Age-adjusted rates 2002-03 2001-02 2000-01 1999-00 1998-99 1997-98 1996-97 1995-96 1994-95 1993-94 1.0 AIHW National Injury Surveillance Unit

Research Centre for Injury Studies, Flinders University Signal & noise: example 2 M e chanism Injury (principal diagnosis) M otorcyclist 199394 to 2002-03 Australia 0-4 ye ars and olde r (case s (e st.)) 40.0 35.0 rate (/100,000 pop) 30.0 25.0 20.0 15.0 NSW, Vic, A CT & NT Qld, SA, WA & Tas 10.0 5.0 Age-adjusted rates 2002-03 2001-02 2000-01

1999-00 1998-99 1997-98 1996-97 1995-96 1994-95 1993-94 0.0 AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University Prognosis: Next steps What?

Broader scope: of burden (eg rehab, late care) Broader scope: of severity (eg ED) Broader scope: exposures as well as outcomes Prediction of non-fatal outcomes Better denominators Shorter lag (+/- model-based prediction) How? Standards (more, better, more widely used) Linkage Incremental system improvements AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University Summary Purposes: Why have indicators?

Principles: concepts, definitions & standards Coming together quite good trends data for serious injury Problems: The technical foundation; essential for comparability Progress: To estimate trends (and differences) in important phenomena, which are not directly observable Constraints are real, tricky but generally manageable Prognosis: Exciting times (nb linkage)

Grail: ICISS-like measure(s) for important non-fatal outcomes AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University After dinner? AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University Understanding the issue Demographics of injured people Types of injury Severity and outcomes Risk factors and exposure Effectiveness and use of interventions

AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University Targeting and priority-setting (potential) topics for indicators Frequency Risk Burden Potential for improvement AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University Approaches Opportunistic Using existing data and other resources Making the most of them for injury surveillance

Understanding, creative use, incremental enhancements Purposive / strategic Data sources and other components of injury surveillance developed specially to serve this purpose AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University Approaches Opportunistic Purposive Case data sources ABS mortality NHMD VEMD NCIS QISU-ED Trauma registers

ASCIR Classifications ICD (vanilla) ICD-10-AM ICECI ICD, NHDD ICE on Injury Statistics Concepts, definitions AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University Status Fatal and serious injury Hospital separations and deaths Cross-sectional and trends Better understanding of data -> better use

Good useable system Will benefit from further validation, etc Basis for Reporting of indicators Numerous analyses of specific topics Use with cost-models for good injury costings Use in GBD models AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University Status Occurrence

Incident fatal & severe injury Characterised in terms of ICD-10-AM, LoS, threat to life, demographic characteristics etc. Period: deaths - many years, hospitalised c. 10 years 1-2 year lag (2-3 y latency for system changes) Risks and burden Population-based rates Age/sex/remoteness/SEIFA/CoB/etc (+/- indigenous status) Other denominators Potentially diverse; patchy availability Other units Cost, DALYs AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University Status Change in occurrence (trends) Challenges:

Hospital data (largely) not person-linked: multiple counting differences in admission/recording: variable sample variations in data quality Deaths Injury death vs UCoD External Cause changes in coding some types of death Manageable with care AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University Future? Technical developments Maintenance Operational links AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University

Future? Technical developments 1. Standards For comparable & meaningful data 2. Outcomes Measure & describe outcomes of injury (in addition to death) 3. Linkage Illness service utilisation data, case data/event data, case data/population data 4. Denominators / exposure data Better population data (nb Indigenous); other denominators 5. Knowledge, attitudes, behaviour of Australians concerning injury, injury prevention, interventions AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University 1. Standards Needed for comparable & meaningful data Including: Operational definition(s) of injury

Classifications ICD-10, ICD-10-AM, ICECI Data standards / minimum data sets For deaths, hospital data, other sources Australia: Technical Review & Revision of injury indicators International: ICE on Injury Statistics; ICECI NDS-IS, NCIS data set Survey questions Opportunities:

Consideration/contribution to/adoption of international standards Further development & updating of Australian standards AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University 2. Outcomes Measure & describe outcomes of injury Overview: Fairly good methods to measure threat to life of injury AIS & derivatives, GCS etc, ICISS Much of burden of injury is due to consequences other than death Methods to measure threat to health of injury are immature Progress is likely to require large and expensive prospective studies

Opportunities: Collaborative project(s), national or international AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University 3. Linkage Potential for more information from existing data May include: Person-based linkage of illness service utilisation data, etc Linking case data to event data (eg hospital / crash) Linking case data to population data (eg enhanced census) Linking special injury register data to any of these Opportunities:

Collaborative projects using WA Linked Health Data Advocacy for similar capabilities elsewhere National developments AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University 4. Denominators/exposure data Potential for more information from existing case data Types of denominators: Better population data Injury risk factor / exposure data Eg. more specific sports-participation data, alcohol use data, travel/vehicle use data Injury protective factor / intervention

nb reliable estimates of Indigenous population Eg participation by older persons in specified exercise activities Opportunities: Diverse AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University 5. Knowledge, attitudes, behaviour Knowledge, attitudes, behaviour of Australians concerning injury, injury prevention, interventions Themes: Knowledge

Attitudes What is the direction (positive or negative) and strength of attitudes towards injury prevention generally and to specific interventions. Behaviour What is the extent, distribution and reliability of knowledge about injury occurrence, consequences and potential for prevention. What are the population patterns of certain behaviours that increase or decrease injury risk Opportunities: CATI surveys AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University Future? Maintenance

Maintain current data sources & quality Ensure good understanding of data eg loss of items from ABS mortality data eg signs of possible deterioration of injury coding in some hospital records Validation studies (eg ARC project) Continue to tap value from existing sources Novel analytic methods (eg ICISS) More efficient analysis & access (cf US WISQARS) Minor changes (eg ICD-10-AM biennial revisions) AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University Future? Operational links

Better understanding of uses and users Governments, researchers, industry, community Look for lessons in related activities Other public health surveillance ane information activities Injury surveillance elsewhere Close liaison with policy-makers Consult, discuss, seek feedback Individually and though national forums Health and other sectors

Seek two-way connection with policy in which surveillance information helps shape policy supports its implementation and monitoring AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University Review Injury surveillance systems & information Purposes Information to support injury prevention & control Approaches Opportunistic and purposive Status Good system for severe and fatal injury Future

Standards; Outcomes; Linkage; Denominators/exposure data; Knowledge, attitudes, behaviour Maintenance Operational links and relationships AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University Why better now? Ten year series of data Quality fairly good and has generally improved Better understanding Completeness Similarity between jurisdictions

Of data strengths and limitations Development of concepts and methods Eg severity measures, injury definitions AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University Trends in injury incidence: issues Records are for inpatient episodes, not persons injured Proportion of incident injury cases resulting in admission might change over time. Identifiability of separation records as relating to an injury might change over time AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University Burden of hospitalised injury Hospitalisation due to Accidental Falls ages 65 and older, Australia 1995-6 to 2002-3 2,500

60,000 2,000 1,500 40,000 30,000 1,000 20,000 500 10,000 Cases Crude rate Adj. Rate 2002-03 2001-02 2000-01 1999-00 1998-99

1997-98 0 1996-97 0 1995-96 Cases 50,000 Cases/100,000 population 70,000 AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University SCI: trends and priorities Persisting Spinal Cord Injury due to motor vehicle accidents, ages 15y and older, Australia 1995-6 to 2003-4 10 6 Motorcyclists Other MVTA

4 M/cyclists, 15-34y 2 2003-04 2002-03 2001-02 2000-01 1999-00 1998-99 1997-98 1996-97 0 1995-96 Cases/100,000 population 8

Year of injury AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University Improvements Understanding the data New uses of the data Estimating incident cases Comparability over time & between places Coding validation (ARC project) Completeness (external cause codes, activity, etc) ICD-based severity measures Classification

ICD-10-AM 3rd ICD-10-AM 5th AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University Classification ICECI Development was prompted by dissatisfaction with ICD External Causes classification: Lagged behind theory (eg energy transfer concept) Technical defects (eg not multi-axial) Developed by international collaborative group Challenges:

Enable refinement Keep current Translation (concepts as well as language) Minimise burden of maintenance AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University AIHW National Injury Surveillance Unit Research Centre for Injury Studies, Flinders University

Recently Viewed Presentations

  • Chapter 7: Food and Agriculture

    Chapter 7: Food and Agriculture

    Which of the following is not a function of food demand? A. Net Income B. Population size C. Share of income spent on food D. Income available through GDP 7. T F Although Asia's meat production is rising rapidly, the...
  • Classifications ULCERATIVE, VESICULAR, AND BULLOUS LESIONS THE PATIENT

    Classifications ULCERATIVE, VESICULAR, AND BULLOUS LESIONS THE PATIENT

    Gingival cyst of infants. Odontogenic keratocyst. Dentigerous cyst. Eruption cyst. Lateral periodontal cyst. Gingival cyst of adult. Botryoid odontogenic cyst. Calcifying epithelial odentogenic cyst. Radicular cyst,apical and lateral. Residual cyst. Paradental cyst and mandibular infected buccal cyst. Inflammatory collateral cyst
  • 슬라이드 1 - DBGuide.net

    슬라이드 1 - DBGuide.net

    -- Avaya 신속하고 일관성 있는 빌드, 배포 프로세스 구축 = 개발 주기 단축 신속한 소프트웨어 개발 주기 = 향상된 팀 생산성 형상관리, 빌드 팀은 효율적이지 못한 수작업, 반복적인 작업에서 벗어나 보다 생산적인 활동을 수행 운영 유지 활동 노력이 경감됨 ...
  • 36.1  Describe the structural components of language.  Language:

    36.1 Describe the structural components of language. Language:

    Inborn Universal Grammar (Noam Chomsky) - suggested that language acquisition is inborn, which helps explain why we pick up language so quickly as children. Language . Acquisition Device: humans are born with a build-in predisposition to learn grammar rules.
  • Betula nigra BET - you - lah NY - grah

    Betula nigra BET - you - lah NY - grah

    Betula nigraBET - you - lah NY - grah. COMMON NAME: River Birch. Form: medium sized tree, 50 to 70 feet tall and wide, oval or pyramidal. FOLIAGE: lustrous medium or dark green, alternate, simple leaves, doubly serrate margin, diamond...
  • Impedance Biosensor - Sabancı University myWeb Service

    Impedance Biosensor - Sabancı University myWeb Service

    Many biosensors are affinity-based,meaning they use an immobilized capture probe that bindsthe molecule being sensed - the target or analyte -selectively, thus transferring the challenge of detecting atarget in solution into detecting a change at a localizedsurface. This change can...
  • Western Heritage - Weebly

    Western Heritage - Weebly

    Sumerian Writing . Sumerian cuneiform is the earliest known writing system. Its origins can be traced back to about 8,000 BC and it developed from the pictographs and other symbols used to represent trade goods and livestock on clay tablets....
  • Year 1: Big question: Why is Easter the most important ...

    Year 1: Big question: Why is Easter the most important ...

    To know the story of Daniel and the Lion's Den, and the stories of Ruth, Abraham and Sarah. To know the miracle of Jesus Calming the Storm. To know the encounter of Thomas doubting. To know the parable of the...