SITUA liON ANALYSIS FOR POLICY REPORT SAP/TRlWPROI001102
PACIFIC ISLAND MORTALITY EVALUATION
Strengthening capability for collection, evaluation, analysis, project and reporting of mortality data from Pacific Island countries
Regional Office for the Western Pacific Manila, Philippines
June 2002
2. METHODS ... 2
2.1 Survey of existing mortality data ... 2
2.2 Evaluation of Validity of mortality data ... 2
3. RESULTS 3.1 Problems with currently available Pacific Island mortality data ... 3
3.2 Best estimates of current Pacific Island mortality ... 4
3.3 Priority countries requiring assistance ... 4
3.4 Summary table ... 5
4. CONCLUSiONS ... 7
ANNEXES: ANNEX 1 - MELANESIA SUMMARY, GRAPHS AND TABLES ... 9
ANNEX 2 - MICRONESIA SUMMARY, GRAPHS AND TABLES ... 21
ANNEX 3 - POLYNESIA SUMMARY, GRAPHS AND TABLES ... 39
ANNEX 4 - FORM FOR SURVEY OF PACIFIC ISLAND MORTALITY INFORMATION AND NOTE ON RESPONSE ... 67
ANNEX 5 - INTERNET SEARCH ... 73
ANNEX 6 - CANBERRA VISIT ... 77
ANNEX 7 - SPC VISIT.. ... 79
REFERENCES ... 81
1. INTRODUCTION
The Pacific Island countries are at different stages of the demographic transition, with some populations still experiencing relatively high mortality and fertility, while others manifest lower mortality and declining fertility. Many of these countries are also passing through the epidemiological transition, which is associated with complex changes in causes of mortality by age and sex over time and between countries.
The availability of accurate and current mortality data and trends is of vital importance for comparison purposes between countries to assist in the generation of the political will necessary to channel public and private resources into health-related activities. The consequences of inaccurate (low) mortality data are an under-estimate of the gravity and urgency of the health situation, and can result in under-investment in health development. Many international and aid agencies use infant and childhood mortality rates and life expectancy as part of the process of decision making for allocation of funds. Trends in mortality can also provide an overall assessment of the results of investment in social and economic development and health services.
In the mortality data from Pacific Island countries presently available, considerable contradictions are evident, especially with regard to implausible increases in life expectancy in several countries during the 1990's, and variation in data supposedly for the same period reported by different agencies.
The objectives of this project are to:
(1) identify and obtain recent empirically derived mortality data for Pacific Island countries, examine for validity, and report accurate mortality by sex;
(2) assist countries with assessment and analysis of their mortality data, including identification of important gaps; and
(3) assist countries with filling important gaps in mortality data and provide opportunities for group and individual training.
Phase 1 (2001) of the project involved: a detailed examination of material already obtained, searches of web sites and databases; telephone, email and fax contact with key informants in international agencies, universities and Pacific Island countries; visits to the Australian National University (ANU) in Canberra and the Secretariat for the Pacific Community (SPC) in Noumea;
and a survey of mortality information held by Ministries/Departments of Health and Statistics in Pacific Island countries through the World Health Organization. Mortality data were evaluated from information on: source and methods, plausibility, and accuracy and consistency.
It is envisaged that Phase 2 (2002) would involve country visits to priority countries identified by the Phase 1 assessment. The purpose of the visits would be to obtain further mortality data and provide countries with assistance in evaluation and analysis of their own data, as well as identification of important gaps in mortality information collection and analysis. Phase 3
would involve: projects to rectify gaps in mortality information, detailed regional mortality analysis, and training activities including fellowships.
2. METHODS
2.1 Survey of existing mortality data
A systematic survey was undertaken of international and regional agencies, and universities and research institutes, known to be involved in the production and dissemination of mortality data in the Pacific Island region. The primary source of mortality data was traced through references in secondary material, and by interviews with authors of reports and key informants as far as possible. Relevant web sites were searched as well as electronic health and demographic databases. During Phase 1 information was gathered from institutions and a survey of countries with visits to ANU (Canberra) and SPC (Noumea).
A data collection instrument was sent to the Health and Statistics Ministries/Departments of the Pacific Island countries through WHO offices and representatives requesting published and un-published mortality data from 1990. Follow-up of this survey instrument by telephone and email was undertaken.
2.2 Evaluation of validity of mortality data
Mortality data was evaluated with respect to its: source and methods; plausibility; and accuracy and consistency.
Sources and methods: The source of mortality data was identified, and the methods used for computation and projections were ascertained as far as possible. These were evaluated with respect to known strengths and weakness of sources and methods. Some countries are known to have relatively good vital registration by the civil system or death recording by health departments.
However, completeness of death recording is suspect in many Pacific Island countries. Use of model life tables for imputation of adult mortality from child mortality, or as part of indirect methods, were examined in relation to the appropriateness of the model life tables used. Projections were examined in relation to methodology and assumptions, when these were available.
Plausibility: Infant and childhood mortality and life expectancy figures were evaluated with respect to plausibility compared with other similar countries, the cause-structure of mortality, and reported changes in relation to previous estimates of greater validity.
Accuracy and consistency: If mortality estimates derived by independent methods requiring different assumptions provide approximately the same estimates, then confidence in such figures is enhanced. For example, concurrence of death recording corrected for under-enumeration and indirect methods from demographic analysis of censuses.
- 3 -
3. RESULTS
3.1 Problems with currently available Pacific Island mortality data
The problems encountered with assessment of Pacific Island mortality data are set out below.
3.1.1 Death or vital registration data
There is considerable use of under-enumerated death recording or vital registration data, especially within countries, without any attempt to estimate and correct for under-enumeration using sample surveys or the Brass Growth Balance method. Both infant and adult mortality data are underestimated using these calculations.
(1) In some small island countries, especially with populations below 50 000, vital registration or death recording data are used to produce annual mortality estimates. The small number of deaths produces significant fluctuations from year to year with some estimates implausibly high and others implausibly low. Averages over three to five years, or in some cases seven to ten years, are required.
(2) Possibly inappropriate life tables are used in some countries to adjust and smooth empirical death recording or vital registration data. Depending on the locations of the countries in the demographic and health transitions, some model life tables may be inappropriate.
(3) In some small island countries with extensive connections with more developed metropolitan countries, out-migration of seriously ill persons for treatment may lead to under- enumeration of deaths of these people (especially the French associates of New Zealand and the United States of America). Furthermore, sailors or expatriate workers may die overseas and not be registered although they were counted in the census. This usually causes greater under- enumeration of male deaths than female deaths, and thus sex differences in adult mortality, which are implausibly small in some countries.
3.1.2 Imputation and indirect demographic methods
(1) In some countries, adult mortality is imputed from infant and under-five mortality using possibly inappropriate life tables. Such methods are very approximate, especially considering that the infant and childhood mortality from death registration or indirect methods may not be accurate.
(2) Some estimates of mortality from indirect demographic methods do not produce plausible results (compared with previous estimates and similar countries), and this may be due to uncorrected inaccuracies in the census or survey data, or problems in the analysis.
(3) Possibly inappropriate life tables are used for estimates, which rely on the combination of indirect measures of childhood and adult mortality.
3.1.3 Projections
(1) Projection methodology for life expectancy and infant mortality estimates is rarely mentioned or described. However, there is indirect evidence that projections are frequently employed to produce 'current' estimates.
(2) There is indirect evidence that projections of life expectancy increases using a fixed fraction of a year of life expectancy per annum (such as one-half a year per annum) have been used in some countries. These projections are implausible for countries in the midst of the epidemiological (health) transition with an epidemic of non-communicable disease.
3.1.4 General comments
(1) There is very poor documentation of sources and methodology of mortality estimates in almost all published reports.
(2) There are significant differences between data published by different agencies for similar periods for many countries. It is usually not clear how such different estimates were derived.
(3) It may be the case that for some countries published mortality estimates are sourced from similar neighbouring countries or regional averages, or imputed from macro- economic (e.g. GOP/capita) or social data.
(4) The age-specific mortality upon which life expectancies are based is frequently not provided.
3.2 Best estimates of current Pacific Island mortality
Table 1 presents he best recent mortality estimates in the opinion of the writer. The data given are not necessarily the most recent. For some countries, the best mortality estimates come from late 1980s or early 1990s data. Projections have not been made. Experience indicates that previous projections of Pacific Island mortality data have frequently been over-optimistic and not confirmed by subsequent empirical information. The data in the table are from published sources or data provided by countries or agencies.
3.3 Priority countries requiring assistance
A number of countries require assistance with analysis methods, development of vital registration systems, development and/or administration of census surveys (see Table 1).
Countries singled out for assistance which could all benefit from detailed assessment of available vital registration data are: Fiji, Federated States of Micronesia, Marshall Islands, Nauru, Cook Islands, Samoa, Tonga and Tuvalu.
- 5 -
3.4 Table 1: Summary table best recent mortality estimates. Pacific Islands
Country Best recent mortality estimates Method Comment
(Popn '000) Year Life Year Infant
expectancy at mortality rate birth (yrs) (/1000) both
Male Female sexes
MELANESIA
Fiji 1996 65 69 1996 20 Demographic analysis Assistance with
(810 in 1997) of 1996 census analysis of current
vital registration data required
New Caledonia 1999 70 76 1999 6.3 Accurate vital No assistance
J197 in 1996L registration needed
Papua New 1991 52 51 1995 76 Life expectancy based Results of analysis
Guinea on 1991 census and of 2000 census
(3 608 in 1990) IMR on inter-censual awaited No
survey (1995) possibility of using death or vital registration data
Solomon 1999 61 62 1999 66 Demographic analysis No possibility of
Islands of 1999 census using death or vital
(404 in 1999) reaistration data
Vanuatu 1989 62 64 1989 45 Demographic analysis Analysis of 1999
(193 in 1999) of censuses census yielded life
There may be expectancies of possibilities of using 66(M) 69 (F) and 27 death or vital (IMR) which are not registration data entirelv olausible MICRONESIA
Federated 1990 64 67 1994 46 Life tables imputed Possibilities of using
States of - from childhood death or vital
Micronesia 1994 mortality derived from registration data
(118 in 1997) indirect methods Assistance reauired
Guam 1995 72 76 1995 9 Accurate vital No assistance
(146 in 1997) reQistration needed
Kiribati 1995 59 65 1995 62 Demographic analysis Unpublished data
(81 in 1999) of 1995 census from 1999 census
similar to 1995 census estimates Some possibilities of using death recordinQ data MELANESIA
Marshall 1994 60 63 1994 63 Life tables imputed More recent
Islands from childhood estimates
(51 in 1999) mortality implausible
Assistance required for analysis of death recording data
Nauru 1991 54 61 1991- 13 Hospital death Assistance needed
(10 in 1992) - 1993 registration data with periodic
1993 analysis of death
registration data
Country Best recent mortality estimates Method Comment
(Popn '000) Year Life Year Infant
expectancy at mortality rate birth (yrs) (/1000) both
Male Female sexes
Northern 1994 67 73 1992- 10 Vital registration data May benefil from
Mariana
-
1996 Probably accurate assessment of vitalIslands 1996 registration data
(59 in 1995)
Palau 1995 64 70 1995 20 Vital registration data May benefit from
(19 in 1998) Probably accurate assessment of vital
registration data POLYNESIA
American 1995 68 76 1991- 13 More information More information
Samoa 1995 required required
(63 in 1999)
Cook Islands 1995 68 72 1991- 16 Vital registration data Requires assistance
(17in 1999)
-
1996 of with analysis of1997 vital registration
data
French 1996 69 74 1996 10 Accurate vital No assistance
Polynesia registration data required
(220 in 1996)
Niue 1991 70 1991- 18 Probably accurate May benefil from
(2 in 1997)
-
1996 vital registration data assessment of vital1997 registration data
Samoa 1998 65 72 1998 25 Vital statistics sample Requires assistance
(175 survey with vital statistics
in 1999) sample survey and
development of death registration
Tokelau 1996 68 70 1991- 32 Vital registration data May benefit from
(1.5 1995 Probably accurate assessment of vital
in 1996) registration data
Tonga 1996 70 72 1996 19 Demographic analysis Assistance required
(97 in 1999) of 1996 census to develop vital
registration MELANESIA
Tuvalu 1991 64 70 1990- 27 Life expectancy data Assistance required
(11 in 1999) 1999 from demographic in assessment and
analysis of 1991 analysis of vital
census registration data
May be too high
Wallis and 1990 67 71 1990- 15 Probably accurate May benefil from
Futuna
-
1995 vital registration assessment of vital(14 in 1996) 1995 registration data
- 7 -
4. CONCLUSIONS
The results of this initial survey represent published available information on Pacific Island mortality to the end of 2001. The 'best estimates' in the summary table above contain some information of questionable validity, but are the most accurate and plausible available in the opinion of the writer.
Apart from quoting infant mortality data directly from some of the few countries who provided detailed information to this project, no extensive assessment or analysis of vital registration data and its validity has been made. However, initial examination of these data suggested in some cases significant under-registration, and considerable annual fluctuations in countries with small populations. In other cases, death recording data suggested lower life expectancies than published sources.
Detailed assessment and analysis of mortality data from countries that would appear to be able to benefit from assistance would need to be undertaken in collaboration with the appropriate government departments in such countries. Publication of mortality estimates based on adjusted death or vital registration data that differed from present estimates would need to incorporate details of the methodology, and be conducted in collaboration with Pacific Island countries and concerned internat:..:...1 and regional agencies.
From this Phase I report it is possible to identify a selection of Pacific Island countries that would benefit from a more detailed assessment, particularly of un-analysed and un-published mortality information. This could lead to publication and dissemination of more reliable mortality and life expectancy information for these countries.
This report has also identified a number of problems in mortality assessment and analysis which could be addressed by workshops for relevant staff in Pacific Island countries. Some workshops could be sub-regional since the types of problems differ across the spectrum of Island countries.
There could also be opportunities created for fellowship attachments or post-graduate study for staff of Island countries who wish to gain further training and expertise in mortality assessment and analysis.
ANNEX 1 MELANESIA SUMMARY, GRAPHS AND TABLES
Fiji
Population in 1997: 809900
The most important observation conceming Fiji mortality data is the dramatic reported increase in life expectancy from around 61 years in males and 65 years in females in 1986 to over 70 years in males and over 75 years in females in the mid 1990s, only to be revised downwards to 65 years for males and 69 years for females by 2000. The data for the mid 1980s and the late 1990s, which are quite similar were derived from the demographic analysis of the 1986 and 1996 censuses using indirect techniques, supplemented by Department of Health data. These estimates are most likely correct. It is unclear how the earlier estimates of life expectancy of over 70 years were derived, but some may have come from analysis of under-enumerated vital registration data and other estimates may have been produced by over-optimistic projections from life expectancy from the 1986 census.
There is some possibility that selective out-migration since the late 1980s may have affected community mortality profiles, however, this would most likely have been persons with relatively low mortality, and would have affected the Indo-Fijian component only. It is unlikely that changes in health service provision would affect mortality in such a way as the recorded data indicate. The most likely reason for the recorded changes in Fiji life expectancy over the last 15 years is that the changes are largely artefactual.
Infant mortality of 15-20/1000 may be at the lower end of plausibility, although there are often significant problems with under-enumeration of early neo-natal deaths (mis-classified as stillbirths) in data derived from death recording.
If mortality has been mostly stationary in Fiji over the period in question this may be a consequence of a combination of: a continued rise in non-communicable disease mortality which could balance reduction in mortality from other causes; lack of improvement in health and social services and aspects of social and economic development related to health status and mortality;
and/or an out-migration of segments of the population with low mortality which disguised mortality improvement in the remaining population. These suggestions would require more detailed investigation. However, data on change in life expectancy between the 1986 and 1996 censuses for Fijians and Indo-Fijians do not suggest a major difference between the races, which would not suggest differential out-migration as a cause.
Whereas the 1996 census gave life expectancy estimates of 65 years for males and 69 years for females, most recent data from the Fiji Ministry of Information (2000) gives life expectancies of 61 and 65 years respectively for males and females. It is unclear how the 2000 estimates were derived.
- 10 -
Annex 1 (Fiji)
It is recognised that mortality data from the Department of Health is improving in quality and completeness. The most important task in Fiji is to develop Health Department capacity to devise methods for analysis of Department of Heath mortality data, which correct for under-enumeration and provide consistent and realistic estimates of Fiji mortality on a continuing basis.
Fiji Male Life Expectancy Rjl Female Life Expectancy
80
I
_ 80
E
75 ~ 75•••
.; 70 ~
.
, •.
.; 70• •
65
•
r; 65• •
~ 60
•
• :ii 60~ 55 ~ 55
a. ll' 50
I 50
: 45
~ 45
:s
40...J 40
1970 1980 1990 2000 2010 1970 1980 1990 2000 2010
Period (year) Period (year)
Annex 1 (Fiji) Table 1
RECENT ESTIMATES
Reference Ref.
Sources/Methods
E(X) IMR No. Comments
Date Male Female Date 11000
-
- -
nla 18 wwwunicef.om - 2001 113 Increase in life2000 61.4 65.2 2000 16.3 Fiji Ministry of
81 expectancy to 72 and
Information, 2000 75 was projected, but
nla nla nla 1997 20 UNICEF,1999 151 no difference in life
1996 64.5 68.7 1996 20 www.sRc.org.nc -
114 expectancy or infant
2001 mortality rate was
nla nla nla 1997 16.8 Fiji Ministry of Health, detected between the
1997 82 1986 and 1996
nla nla nla 1996 17.1 Fiji Ministry of Health,
82 censuses using
1997 indirect methods.
1996 64.5 68.7 1996 20 Rakaseta V, 1999 15 Indirect methods
1996 census The estimate of IMR of 1996 64.5 68.7 1996 22 Fiji Bureau of
10
711000 is very unlikely.
Statistics, 1998
PrGbably Y, a year The 2000 estimates South Pacific per year added to from the Fiji Ministry of
1995 72 75 1994 16
Commission, 1998 23
1986 census life Information show expectancy figures. lower life expectancy
Secretariat Pacific estimates than
1995 72 75 1994 16
Community, 1998 20 1986 census previous data;
1990 methods not given
- 69.5 73.7 nla nla United Nations, 2000 30
1995 Population Division
1996 64.8 69.0 1986 22 Rakaseta V, 1999 15 Fiji Bureau of Statistics
1996 1996
69.7 75.8 7 www.SRc.org.nc -
- -
2001 1141998 1998
Fiji Co-ordinating
1993 M + F
=
71 1993 16.8 Committee on 84 Children, 19951992 nla
I
nla 1992 16.6 Fi'i Government, 1993 45 1990 M + F=
71 1990 25 World Bank, 1995 47Fiji Co-ordinating
1990 M + F
=
69 1990 22 Committee on 84Children, 1995 1986 61.0 65.
1986 Fiji Department for
2
-
Women and Culture 461986 61.4 65.
1986 20.9 Fiji Bureau of
43 1986 census
2 Statistics, 1986
1985 M + F
=
70 1985 29 World Bank, 1995 47- 13 -
New Caledonia Population in 1996: 196 836
Annex 1
New Caledonia mortality and life expectancy data are based on vital registration of deaths, which is considered to be accurate. There are some deaths in New Caledonian residents who seek treatment abroad (Australia, France), which may not be included in the vital statistics data.
The increase in life expectancy and decline in infant mortality over the last decade might be partly influenced by in-migration of Europeans with relatively low mortality, but evidence available suggests that mortality decline is also affecting the Provinces outside of Noumea, which are overwhelmingly Melanesian.
The decline in mortality is probably partly a consequence of improved social and economic conditions and improved health services for the local population (of all ethnicities) consequent upon the accords agreed upon with France a decade ago.
There are no implications for changing the mortality recording system.
New Caledonia Male Life Expectancy
_ 80---'-1
I!! 75..
.; 70 .",..~
t' 65
i 60
1l 55
=- 50
5 ~~ j I
1970 1980 1990 2000 2010 Period (year)
New Caledonia Infant Mortality Rate
~ :~1--- -- I
~ 70
l 60
~ 50 I
~ o 40 30 I
=. 20
~ 1~~j-- __ --~~-.~~--~
1970 1980 1990 2000 2010 Period (year)
New Caledonia Female Life Expectancy
~
80
75 .,; 70~ 65
Ii 60
1l 55
i
50~ 45
... 40 +-~-~-_--l
1970 1980 1990 2000 2010 Period (year)
Annex 1 (New Caledonia) Table 2
RECENT ESTIMATES E(X)
Date Male Female Date 1999 69.8 75.8 1999
1998 70.3 76.7 1998
1997 68.8 76.4 1997
1995 68 76 1995
1995 68 76 1995
1990
- 69.7 74.7 nla
1995
1993 67.9 75.7 1993 1992 68.2 75.1 1992 1991 66.7 74.3 1991 1990 67.9 73.1 1990 1989 66.7 71.9 1989
1987 1987
-
67.5 73.8 -1989 1989
IMR /1000
6.3
6.9
5 8 8 nla
7.6 8.6 8.2 10.7 11.2
11.2
Reference Ref.
Sources/Methods
No. Comments
Age specific mortality
Nouvelle Caledonie 120 based on generally
accurate civil vital
Direction des Affaires registration. Problems
Sanitaires et Sociales
93 are: sparsity of deaths
de Nouvelle- in one year of data,
Caledonie, 1998 and suspiciously low
Secretariat Pacific
21 IMR Plausible
Community. 1999 improvement in life
Secretariat Pacific
20 expectancy over 10
Community, 1998 years.
South Pacific
Commission, 1998 23 SPC has different
estimates for same
United Nations, 2000 30 period -1996-1998.
Nouvelle Caledonie, 1999 119
Nouvelle Caledonie, 1999 119
Nouvelle Caledonie, 1999 119
Nouvelle Caledonie, 1999 119
Nouvelle Caledonie, 1999 119
Rallu JL, 1991 18
- 15 -
Papua New Guinea Population in 1990: 3607954
Annex 1
There is no reliable vital registration of deaths in PNG either through statistics offices or the health department. All estimates of mortality are based on indirect demographic methods through censuses or surveys. The last empirical life expectancy estimates were from the 1990 census and suggested life expectancies of 52 years for males and 51 years for females. Estimates have been made for the mid-1990s of 55 and 54 years for males and females respectively, but these are probably projections.
Infant and child mortality is estimated from inter-censual surveys as well as censuses. The infant mortality appears to be around 70-80/1000 with no trends in the last decade.
The 2000 census is still being edited and entered, and the mortality analysis of this census is not due to commence until mid-2002, and results may not be available until 2003.
There is no alternative in PNG at present except to use indirect methods to estimate community mortality. While efforts to improve registration of deaths should continue the data are not expected to be sufficiently reliable for decades. Mortality estimates are infrequent because censuses occur only at 10-year intervals, although infant and childhood mortality estimates are also available from inter-censual surveys. A possibly to improve available mortality data would be to add techniques for calculation of adult mortality to inter-censual surveys, and/or to use sample vital registration areas.
~ 80 75
! 70
~ 65 c 60
~ 55
e-
50~
..
45Papua New Guinea Male Life Expectancy
---1
....
...J 40 +---_-_-_---l
1970 1980 1990 2000 2010 Period (year)
Papua New Guinea Infant Mortality Rate
~ 80 ~-~.-.---
~ :;; 60
~ 40
"E ~ 20
~ o +----_ _ -~-__l
1970 1980 1990 2000 2010 Period (year)
Papua New Guinea Female Life Expectancy
~ 8 0 , - - - , :I 75
~ ,.. 70 :; 65
¥ 60
~ 55 •
:5
50 . . .1970 1980 1990 2000 2010 P. rlod (year)
Annex 1 (Papua New Guinea) Table 3
RECENT ESTIMATES
E(X) IMR
Date Male Female Date /1000
1991 ,1986
-
54.6 53.5-
771996 1996
- - -
nfa 79nfa nfa nfa 1995 76
1990
-
55.2 56.7 nfa nfa1995
nfa 55 54 nfa 77
1991 52 51 1991 82
1991 52 51 1991 82
1991 52.2 51.4 1991 82 1990 52.2 51.4 1986
-96 77
Reference Ref.
Sources/Methods Comments No.
All data from censuses www.spc.org.nc-
114 and surveys.
2001
www.unicef.org -2001 113 PNG National
Statistical Office, 1996 68 United Nations, 2000 30
UNDP, 1999 29
South Pacific Commission, 1998 23 Secretariat Pacific Community, 1998 20
Hayes G, 1996 12
United Nations, 2000 30
- 17 -
Annex 1 Solomon Islands
Population in 1999: 404 000
Vital registration and health department death recording are variable and incomplete in Solomon Islands. All mortality data are derived from indirect demographic methods applied to censuses.
Life expectancy data given for the 1990s are implausibly high compared to the 1980s.
There have been difficulties in calculating mortality levels from the last census (1999), but the most recent estimates of life expectancy of 61 years for males and 62 years for females are plausible;
as is the IMR of 66/1000.
Solomon Islands Male Life Expectancy
80,---~
"F 75
~ 70
;: 65
li u 60
~ 55
Q,
~ 50
~ 45
•
• • •
...J 40 +--__ ~--~--~___"
0 90
0 80
q
- 8.
70 60l!' 50
~ t: 40
0 30 E 20
-
I:..
10-
.5 01970 1980 1990 2000 2010 Period (year)
Solomon Islands Infant Mortality Rate
... • •
1970 1980 1990 2000 2010 Period (year)
80
e
75: 70
;: 65 » u i 60
1l
Q, 55=
50~ 45
...J
Solomon Islands Female Life Expectancy
•
• • •
1970 1980 1990 2000 2010 Period (year)
Annex 1 (Solomon Islands) Table 4
RECENT ESTIMATES
E(X) IMR
Date Male Female Date 11000 1999 60.6 61.6 1999 66
n/a n/a nla 1999 38
1990
-
68.4 72.7 nla nfa1995 1990
-
63.5 67.3 nla nfa1995
nJa 64 66 nJa 38
1986 60 61 1986 38
1986 60 61 1986 38
1980 1980
- 59.9 61.4
-
38.31985 1985
1980 1980
-
59.9 61.4-
381984 1984
1980 1980
- 59.9 61.4 - 38
1984 1984
-
- - nla 37Reference Ref.
Sources/Methods Comments No.
SPC 166 1999 census Mosl recent data
UNICEF, 1999 158 plausible
United Nations, 2000 30 Reported data show
an increase of 1 year per year from mid
United Nations, 2000 30 1980s to mid 1990s
which is implausible.
UNDP, 1999 29
South Pacific
23
Large discrepancy in
Commission, 1998 IMR between
Secretariat Pacific UNDP/UNICEF and
Community, 1998 20 those reported by
SPC.
Bartleman et al., 1988. 1 Census
Methods for UN projections need Secretariat Pacific
21 Indirect estimates investigation Community, 1999
www.sgc.org.nc - 2001 114
wW'A'.un[GefoIQ - 2001 113
- 19 -
Vanuatu
Population in 1999: 193219
Annex 1
Vanuatu mortality and life expectancy data are based on demographic analyses of censuses, however, the quality of some of these analyses may be questionable. This particularly applies to the available analysis of the 1999 census data. In this analysis the infant mortality rate was estimated at 26 and the life expectancy for males was 67 years and for females was 70 years.
The Vanuatu mortality data from the 1999 census was re-analysed by SPC and results suggested life expectancies of 66 years and 69 years for males and females, with an IMR of 27/1000.
However, these estimates may not be plausible and there may be problems with the census data itself.
The increase in life expectancy and decline in infant mortality over the last two decades might be partly influenced by improved environmental conditions and health services, but there are questions concerning validity of analyses.
Vanuatu Male Life Expectancy
80
'F 75
!
70,.. 65 u i 60
~ 55
~ 50
~ 45
•
• •• •
:::; 40 +---~_-~----"
c> 100
& 80
~ ,.. 60
'"
~ 40 1: ~ 20 .l!!
1970 1980 1990 2000 Period (year)
Vanuatu Infant Mortality Rate
•
• • •
•
B o+---_-_-_~
2010
1970 1980 1990 2000 2010 Period (year)
Vanuatu Female Life Expectancy
80 ----
~ 75
! 70
~ 65
:i 60
i
.. 50 55~ 45
•
• •
• •
~ 40 +--_-_-_~
1970 1980 1990 2000 2010 Period (year)
Annex 1 (Vanuatu Table 5
RECENT ESTIMATES
Reference Ref.
Sources/Methods Comments
E(X) IMR No.
Date Male Female Date /1000
1999 66 69 1999 27 SPC 166 1999 census Reported data show
1999 67 70 1999 26 Vanuatu National
79 1999 census an increase in life
Statistics Office, 2000 expectancy in the
nfa nfa nfa 1999 39 UNICEF,1999 154 early 1990s of almost
nfa 64 67 nfa 45 UNDP, 1999 29 1 year per year which
nfa nfa nfa 1989 55-60 Vanuatu Census Dala
75 1989 census is implausible.
Office, 1991
www.sgc.org.nc- Decrease of IMR of 45
1989 61.5 64.2 1989 45
2001 114 to 26 deaths per 1000
South Pacific live births since 1989
1989 62 64 1989 45
Commission, 1998 23 may not be plausible.
1989 61.5 64.2 1989 45 Vanuatu & UN 76
1979 Methods for UN
61.5 64.2 1985 45 Secretariat Pacific
21 projections need
-
1989 Community, 1999 investigation
1979 Nfa 56 1979 94 Vanuatu Census Data
75 1979 census Office, 1991
1979 54 56 1979 94 Vanuatu & UN 76
- 21 -
MICRONESIA SUMMARY, GRAPHS AND TABLES Federated States of Micronesia
Population Estimate in 1997: 118000
ANNEX 2
FSM mortality and life expectancy data are derived from both censuses and vital registration. It is acknowledged that vital registration is under-reported, but considered to be slowly improving. There is a disclaimer in the 1999 FSM statistical yearbook indicating that the data is to be used for comparative purposes, and does not actually represent the number of deaths.
Direct measures of mortality could be calculated using deaths registered with the Department of Health, however ,these events are under-registered and inconsistently covered and at best the data provides a pattern of mortality. Small numbers of persons and deaths and inaccuracies in reporting make direct estimations difficult.
It appears that the most recent life expectancy estimates have been imputed from elltimates of childhood mortality.
For the census year 1994, IMR was calculated at 46/1000 using indirect methods from the census and 25/1000 using vital registration data. The level of infant mortality does not appear to have improved in the past decades.
Federated States of Micronesia Male Life Expectancy
80
"F 75
~ 70
-z;: 65
:i 60
i 55
'"
~ 50
:S
45•
•• •
1970 1980 1990 2000 2010 Period (year)
Federated States of Micronesia Female Life Expectancy
80
'E" 75
! 70
~ 65 ... ,
Ii 60 •
1! 55
~50
~ 45
...J 40 +--_~ _ _ _ _ ---<
1970 1980 1990 2000 2010 Period (year)
Federated State. of Micronesia Infant Morta IIty Rate
90
§ 80 . : 70
i. 60
i:' 50
~ 40
~ 30 - 20
i
10• ••
...
o+---~-~-_---<
1970 19S0 1990 2000 2010 PeTiod (year)
Annex 2 (Federated States of Micronesia) Table 6
RECENT ESTIMATES
Reference Ref,
Sources/Methods Comments
E(X) IMR No,
Date Male Female Date 11000
nla Nla nla 1999 28 UNICEF,1999 153 There is a
nla nla nla 1997 26.8 FSM Government,
42 Registration data considerable
1999 discrepancy in IMR
nla nla nla 1996 19.5 FSM Government,
42 Registration data between SPC/UNDP
1999 and UNICEF, and IMR
FSM Government, from the FSM
nla nla nla 1995 20.9
1999 42 Registration data government is
FSM Government, variable.
nla nla nla 1994 24.9
1999 42 Registration data
-
-
- nla 20 www.unicef.org - 2001 1131994 nla 65.2 1994 47 FSM Government,
41 1994 census 1996
South Pacific Adult mortality and
1994 64 67 1994 46 29 life table based on
Commission, 1998 1 .
CEBCS technique
nla 64 67 nla 46 UNDP, 1999 29
Nla nla nla 1993 21.7 FSM Government,
42 Registration data 1999
nla nla nla 1992 21 FSM Maternal and Child Health, 1991/92 89 1992 62.2 65.9 nla 52 FSM GovernmenU
UNICEF, 1996 90
1990 1990
64.6 66.8 45 www.sQc.org.nc -
- -
2001 1141992 1992
1990 1990
- 64.6 66.8
-
45 Demmke e/ al., 1997 3 1994 census1992 1992
1990 1990
Secretariat Pacific
- 64.6 66.8
-
45 Community, 1999 211992 1992
1979 1979 Vital registration
- 56.5 59.3 - 45.2 Taylor R, 1989 26 Adjusted death
1981 1981 recording
- 23-
Guam
Population in 1997: 146,328
Aproximately 129,000 Civilians and 16,000 Military Personnel (including families)
Annex 2
Infant mortality and life expectancy is generally determined from registered death data. In the period 1990-1995, the IMR was 9/1000.
Life expectancy for 1995 was estimated to be 73 years for males and 77 years for females.
While registration of births and deaths for Guam is likely to be complete, the influence of immigration and emigration is not documented. There are implications of migration of military personnel, including families, for life expectancy estimates. This may lead to assumptions of higher life expectancy for the indigenous Chamorro population, which comprises 43 percent of the population of Guam, than may be the case.
Guam
Male Life Expectancy
8 0 , - - - ,
~ 75 : 70
,.,
;: 65 u :; 60
~ 55
Q.
; 50
~ 45
• ••
40~--~--______ ~
1970 1980 1990 2000 Period (year)
Guam Infant Mortality Rate
2010
~ 80 ---.----.~
1D
Q. 60
~ ~ 40 o E 20
C •
J! 0 ~~--~----~
.E 1970 1980 1990 2000 2010 Period (year)
Guam
Female Life Expectancy
80 . , - - - -
~ 75
:I 70
,.,
~ 65
; 60
i
55:I 50
"
:5 45
• ••
40~ ____ ~--__ --~
1970 1980 1990 2000 2010 Period (year)
Annex 2 (Guam) Table 7
RECENT ESTIMATES E(X)
Date Male Female Date 1995 72.5 76.6 1995
1995 73 77 1995
1990
- 72,2 76,0 nla
1995
1990 1990
-
71.5 75,6 -1994 1995
1990
1990 69,8 74.4 -
1995
1988 1990
-
69,8 74,43-
1992 1995
IMR 11000
9 9 Nla
9
9
9
Reference Ref.
SourceslMethods Comments No.
Secretarial Pacific
21 Accurate vital
Community, 1999 registration data.
Sou Ih Pacific
23 Unnecessary to use
Commission, 1998 model life tables which
yield higher life
United Nations, 2000 30 expectancies,
Vital registration and census,
West Model life Demmke et ai, 1997 4
Table matching registered deaths for 1990 -1994, www,sRc,org.nc-
2001 114
Vital registration.
West Model life Demmke et ai, 1997 4 Table matching
registered deaths for 1990 -1994,