Are Death Statistics in Turkey

Accurate? A Study on Ill-Defined

Causes of Death in Izmir

Dr.Raziye Ozdemir Soncag

Assoc.Prof.Zeliha Ocek

Prof.Gonul Dinc Horasan

Why are cause of death statistics important?

Cause of death statistics are primary outputs as evidence for;

Health policy and planning purposes

Monitoring specific health programmes

Identifying and prioritizing health research activities

2

Quality of globally available information on

causes of death

Source: World Health Organisation, School of population Health University of Queensland, Improving the quality and use of birth,

death and cause-of death information: guidance for a standards-based review of country practices, 2010. 3

Death registration in Turkey

Turkish Statistical Institute (TURKSTAT) is responsible for death

statistics

The main problems are;

Cause of death statistics are collected from only provincial and district

centers

Incomplete registration

Incomplete notification

Late notification

Missing information in death certificates

Inaccurate data

4

Aim

To determine the proportion of ill-defined underlying causes of

deaths among all deaths reported to TURKSTAT in Izmir in 2010.

5

Methods

Death

Certificate

Ill defined ICD-10 codes were determined;

Symptoms, signs, and ill-defined conditions (R00–R9);

Injuries where the intent is not determined (Y10–Y34,Y872);

Cardiovascular disease categories lacking diagnostic meaning, such as cardiac arrest and

heart failure (I47.2, I49.0, I46, I50, I51.4, I51.5, I51.6, I51.9, I70.9)

Cancer deaths coded to categories for secondary or unspecified sites ( C76, C80, C97)

(Mathers et al, 2005)

All underlying causes of death in 2010 were

coded using ICD-10

N=18933

Provincial health

directorate TURKSTAT

The last cause of death

6

Results

The percentage of unknown and ill-defined causes of death

in total deaths

Causes of death was not reported;

2,3%

Ill-defined ; 27,4%

Defined ; 70,3%

7

Distrubution of unknown and ill-defined codes

by catagories

Cardiovascular diseases; 59,6%

Symptoms; 25,0%

Causes of death was not reported; 7,7%

Injuries ; 4,0% Cancers; 3,8%

8

The percentage of defined and ill-defined codes by

age groups

0

10

20

30

40

50

60

70

80

90

100

66,4

43,9

62

42,4

52,9 46,2

57,7 57,5 61,7

70,6 73,7 74,5 75,4 76,1 74,6 68,2

62,5 70,3

33,6

56,1

38

57,6

47,1 53,8

42,3 42,5 38,3

29,4 26,3 25,5 24,6 23,9 25,4 31,8

37,5 29,7

Defined Ill-defined

9

Distrubution of ill-defined codes by some variables

0

10

20

30

40

50

60

70

80

90

100

71,6 68,7 70,3 70,4 69,4 73,7

49

72,6

40,3

19,2 22,2

67,3 74,3 77,2 74,4

23,7

72,8

28,4 31,3 29,7 29,6 30,6 26,3

51

27,4

59,7

80,8 77,8

32,7 25,7 22,8 25,6

76,3

27,2

Defined İll-defined

Pe

rce

nta

ge

X2=18,8

P=0,000

X2= 240,9

P=0,000 X2= 845,3

P= 0,000

X2= 1137,4

P= 0,000

X2= 984,1

P= 0,000

X2= 0,0

P= 0,989

Sex

The place of

recidence

The place of

death

Manner of death

Institute

Autopsy

10

Conclusion

The proportion of ill defined causes of death in TURSTAT statistics were very

high.

This problem was common among the deaths with autopsy or the reported

deaths by forensic institutions.

The reporting should be completed after the autopsy in these instutions.

The restriction of the ill defined codes during electronic data entry may decrease

the amount of ill defined death causes in mortality statistics.

The system should be appropriately designed to record events correctly.

These results strongly support the need to include training in death certificate

completion as part of physician education.

11

Thank you for your attention