Plan of analysis and application to death from falls

1
Plan of analysis and application to death from
falls
2
Plan of analysis
3
Ideal objectives

• A method that allows us to
• Describe epidemiological situation
• Identify discrepancies
• time, place, person
• Define comparability
• Correcting factor
• Possible and impossible comparison
• Draw up recommendations to improve comparability

4
Described in a  technical note 

• Data to collect
• Indicator to process
• Identification of discrepancies and explanation
  through
• questionnaire
• mortality forum
• further investigation with country
• Recommendations for improving comparability
  proposal
• Partial comparative conclusion on epidemiological
  situations
• (Recomendation validation implemetation)

5
Example death from fall
6
Explore Eurostat data
7
Definition
8
Standardised mortality rate in 2003(deaths/100
000/year)
EU25 7.2
9
Death from fall2.5 to 23 deaths /100 000/year

• How to explain these differences
• Certification
• Codification
• Indicator production
• Epidemiological context
• gt need to explore
• time trends
• practices
• context

10
Explore time trends

• Absolute number
• Standardised mortality rate
• Crude mortality rate
• Age associated relative risk
• Sex associated relative risk

11
Time variation of absolute number of deaths at a
country level (Figure 0)

• Most important number of Falls related deaths
  concern 6 countries (IT, DE, FR, UK, PL, HU) with
  more than 2000 deaths/year
• Lack of data for IT since 2002
• Huge decrease
• for FX in 2000
• for DE in 98 an 2004, for 2004 lack of data
• for CH in 95
• for UK in 2001 after increase gt

12
Rank of falls as a Cause of death priority
(Figure 1)

• On last available data
• Among 26 group of CoD

13
Age standardised falls related mortality rate
evolution over time in Europe (Figure 2)

• Falls related SMR varies from 33.1 (HU 1994) to
  1.4/100000 resident/year (AL 04)
• To be checked
• .

14
Time variation at a country level (Figure 4)

• Stable countries

• Unstable countries

15
Comparison of absolute number and SMR ( age
standardised mortality rates) (Figure 3)

• Most countries have a parallel evolution of SMR
  and Abs Number over time.
• When punctual or regular discrepancies exist
  between SMR and Ab Numb., explanation should be
  obtained from countries
• variation of the population?
• variation of the distribution of the population?
• Variation of the distribution of the cases?
• SMR gt CMR under estimation of the risk of
  mortality by the CMR is due to of people with
  low risk is lower than the proportion of this age
  group in Europe
• To be able to compare these trends, it is
  important to check that both scales vary in the
  same proportion (lower limit / upper limit in the
  same ratio)

16
Risk of death from fall and sex
17
(No Transcript)
18
Sex ratio of mortality (SMRm/SMRf) Figure 5b

• Men have a moderately increased risk of death
  from falls when compared to women, in most of the
  countries (RR 1 3). This risk is not
  depending on age distribution (ratio of age
  standardised mortality rates). For all countries,
  the averaged RR vary between 1.8 and 2.3
  according to the years.
• In some countries, for a given year men have a
  very much increased risk of death from fall (RRgt
  4) AL in 99, BG in 01 and 04, CY in 2004, EE in
  2000 and 2002, IS in 2000, LT and RO all the
  years
• Women appear to be more at risk (RRlt1) in some
  cases IS (1995, 1997), MK (2000) and CZ (1995
• In a given country the RR is more or less stable
  except a huge increased risk a given year in some
  countries LU 1997

19
Mean CMR (/100 000/y) by age for UK
20
Age related risk of mortality from fall Figure
6Crude mortality by age group for Europe (6a)
21
Mean CMR (/100 000/y) by age for UK
22
RR age group/15-24
23
(No Transcript)
24
Average mortality rate ratio in relation with
age, for different years, by country
25
Average mortality rate ratio in relation with
age, for different years among all european
countries
26
Explore context
27
Explore context

• The use of Anamort questionnaire
• Changes in ICD version
• Application of updates
• Certification, Codification and indicator
  production differences
• The use of mortality Forum
• The litterature review

28
From the questionnaire
29
From the questionnaire

• False negative 23 countries
• Certification
• Lack of information
• Unspecified fracture/injury
• X59 (expo to unspec fact) or Y34 (unspec event,
  undet intent)
• I codes (cardio vascular system)
• Respiratory
• Neoplasm
•  Forgotten  fall (delay)
• Mental disorders and central nervous system
• Symtoms sign and and abnormal findings
• Coding
• Lack of query
• Pathological fracture
• Other External cause
• Femur fracture coded X59 if no mention of fall
• Epilepsy, alzeihmer
• Alcohol
• Death 1 month after death Y86

30
Explanations for huge decreases

• ICD9  accidental falls (E880-888) included E887,
  Fracture, cause unspecified
• ICD10 no correspondance with E887
• In practice if death with fracture with no
  other cause gt Code X59   Exposure to
  unspecified factor is given
• In 2000, X59 excess of 6 011 deaths coded E887 in
  ICD 9 that must be added

31
ICD 10 introduction
32
ICD 10 introduction
33
Conclusion on the effect of ICD 10 introduction

• Due to coding isssues, ICD 10 could have a
  mechanical effect of reduction of the mumber of
  death by fall in most countries
• Explore what happen in AT, FI, HU, NO
• Explore use of X59, Y34 during the same period
• Explore impact of fall by assault Y01 and suicide
  X80

34
From the questionnaire

• Risk factors for bias
• Age (yes 14 no 5) biases affect more elderly
  in both way according to countries
• Sex (yes 5 no 14) rarely falls among women are
  more often undertestimated (domestic violence and
  cases more superficially investigated)
• Region (yes 2 no 17) coder related
• Non resident will not change the result (26 vs 2)
  (UK include them SK not)
• Resident dying abroad will not change the results
  (UK exclude them)
• Other factors change and multiplicity of coders
  
• Improvement to be done
• Training of certifiers
• Standardisation of coding process (trained and
  stable coders, automatic coding system

35
Updates application and differences

• Cumulative updates
• Add a code R29.6 Tendency to fall, not classified
  elsewhere
• Add See at the beginning of this chapter for
  fourth-character subdivisions for falls
  (W00-W19)
• codes for place or activity
• Clarification on HIGHLY IMPROBABLE SEQUENCES
• accidents (V01-X59) reported as due to any other
  cause outside this chapter except
• (1) any accident (V01-X59) reported as due to
  epilepsy (G40-G41),
• (2) a fall (W00-W19) due to a disorder of bone
  density (M80-M85),
• (3) a fall (W00-W19) due to a (pathological)
  fracture caused by a disorder of bone density,
• Use
• 10 countries all updates, 10 some of them
• Year of application varies a lot
• Dificult if no automatic coding system
• Priority to continuity
• Impact
• Limited?
• Non application of the 3rd update will
  overestimate the number of falls

36
Mortality Forum
37
Mortality Forum

• Fall (W00-19) due to alcohol
• Could it be possible to assign to the FALL, but
  add a code from Y90-Y91
• Evidence of alcohol involvement determined by
  blood alcohol level
• Evidence of alcohol involvement determined by
  level of intoxication
• if "intoxicated" simply means "drunk,
• consider the sequence as highly improbable.
  gtFALL
• Else (more than drunk or undefined)
• code T510 (Toxic effect of alcohol) and X45
  (Accidental poisoning by and exposure to alcohol)
  for this case .
• How would you code the external cause for fall in
  tub when the nature of injury is not drowning?
• W01 W17 W19, W19.0, W18
• If no information on reason (malaise..) ?
• Multiple injuries to the deceased as a result of
  striking the ground following ejecting from a
  Sabre jet aircraft
• V97.0 Occupant of aircraft injured in other
  specified air transport accidents
• V95.8 -other aircraft accidents injuring occupant
  - as this captures the 'accident' to the aircraft
  which presumably would have resulted in death
  anyway had the aircrew not ejected!
• And NOT V97.1 Person injured while boarding or
  alighting from aircraft
• With ICD9, unknown causes of fracture are coded
  "E887 Fracture, cause unspecified".
• With ICD10 there is no code corresponding to E887
  and unknown causes of fracture are coded "X599
  Exposure to unspecified factor" together with all
  the others causes of traumatisms
• Personally, I find the solution suggested by
  Roberto attractive a code like the ICD-9 E887
  that identifies fractures without stated cause,
  but not in the "falls" block
• In cases with fracture of femur where external
  circumstances is not entered, we use unspecified
  fall (W19)

38
Mortality Forum

• killed when the skylift he was working from
  overturned machinery accident, or as a fall?
• V98 includes fall or jump from cable car, not on
  rails
• fall from a bicycle,
• V89.9 Person injured in unspecified vehicle
  accident  
• I a) Cerebromeningeal haemorrhage / b) Fall when
  in acute drunkenness, epileptic seizure II
  Chronic alcoholism with hepatic failure
• Epilepsy unless Fall is prior to epilepsy
• 1a Fracture Skull / 1b Myocardial Infarction
• X59.9
• 1a Fracture Skull / 1b epileptic attack
• To be honest, I do not quite understand why there
  should be an exception for epilepsy.
• person drowned after falling into a well
• person drowned after the car he was driving went
  off road and ended up in the sea
• fractures in connection with osteoporosis

39
Litterature review
40
Litterature review

• Reference manager
• 51 articles related to falls or fracture
• 36 discuss explanation of biases (others are
  purely descriptive), full text available
• Priorising falls or accidental falls give
  different priority and are supposed to be related
  to different intervention (see injury matrix)
• Underestimation of the magnitude of accidental
  falls
• Sequence osteoporosisgthip fracturegtfallgthospitalgtc
  omplicationgtdeath gt UCoD osteoporosis Or
  complication
• Lack of knowledge of appropriate rules certcod
• In order to avoid autopsies/inquest cert
• As much as fall occurred a long time before death
  cert
• As much as it concerns women, older persons (not
  true for intentional falls) cert
• No use of a place for specification of the cause
  of injury in the death certificate cert
• Lack of query cod
• Lack of autopsy cert
• No linkage between death certificate and medical
  record cod
• Various definitions/rules to determine intent
• Use of ICD10 instead of ICD9
• Overestimation accidental falls
• Hidden violence against elderly/childcert
• Underreporting of uncontrolled epilepsy

41
Litterature review 2

• Increase of accidental falls related deaths can
  be associated
• With development of the countries
• Increased risk of fall OR/AND increased CFR
  ( poor  care?)
• Race/ethnicity (Asian gtblack and white) Increase
  Age Sex (M) Urban setting
• low per capita income
• Low population density (p/km², in opposition with
  urban risk factor?)
• Early phase of industrial development?
• Number of co-morbidities
• Use of CNS and CV drugs
• Other
• Present results on falls by intent BUT
• falls could be examined as a whole
  (independently of intent) in order to develop
  common preventive passive measures (Bars on
  windows in mental facilities can prevent suicide,
  just as bars on windows in apartments prevent
  unintentional falls.)
• Concept of  Low Velocity Fall 
• Mortality ltltltmorbidity and cost
• Injured part after a fall can be predictor of
  death (hip, /-head /-thorax)
• Hip fractures should not be excluded from injury
  as it is done in some trauma registries as they
  participate to the workload of injury services

42
Comparison  accidental falls  (W0-W19) and
 all falls  (X80, Y01, Y30)
Spain 1600/2400
43
Comparison  accidental falls  (W0-W19) and
 all falls  (X80, Y01, Y30)
44
ANAMORT partner institutions

Institut de Veille Sanitaire (InVS), France




National Institute of Public Health, Denmark





Norwegian Institute of Public Health



Psytel,

France

Psytel



Kuratorium für Verkehrssicherheit

Austria

The statistical office of Estonia