Working Group Public Health Statistics

Transcription

1 Directorate F: Social Statistics and Information Society Unit F-5: Health and food safety statistics Doc. ESTAT/F5/11/HEA/04 Working Group Public Health Statistics Luxembourg, June 2011 Item 5 of the agenda The healthy Life Years (HLY) indicator

2 1. Introduction The aim of this document is to give an overview of the situation of the Healthy Life Years (HLY) indicator as defined and used by the Commission. Focus will be given on the data Having the code of practice with its quality criteria as the framework, this document will describe the different problems especially in relation to relevance, coherence and comparability. To tackle those problems and to improve the future indicators, a number of actions on short and long term are proposed and placed in a broader context including the actions of SANCO and EHIS. A final chapter lists the proposals and questions to the Working Group delegates. 2. State of the art The HLY indicator (also called disability-free life expectancy or DFLE) measures the number of remaining years that a person of a certain age can be expected to live without disability; in other words, this is a health expectancy indicator. HLY is currently calculated using mortality statistics and data on self-perceived disability. Mortality data comes from Eurostat s demographic database, while self-perceived disability data is based on the EU statistics on Income and Living Conditions (EU-SILC). The relevant EU-SILC question, the GALI question, is: "For at least the past six months, to what extent have you been limited because of a health problem in activities people usually do? Would you say you have been: strongly limited? / limited? / not limited at all?". The HLY indicator is calculated separately for males and females and at two ages: birth and the age of 65. The calculation of the HLY started in 1995 for 9 Member States. At that time the ECHP (European Community Household Panel) (and not EU-SILC), implemented from 1994 to 2001, was used as the source for self-perceived disability. For the years extrapolation techniques were applied. From 2004, SILC was progressively introduced. From 2005 onwards series are complete for the EU25 MS and from 2007 onwards for the EU27 MS. The tables on Eurostat's web site are updated with the values for 2009 (not completely: IT, UK and EU27 are missing). Changes for the HLY at birth in 2009 compared to 2008 are rather small: for all but one country they are below the value of 5 % (positive or negative). However changes for the HLY at 65 are bigger than 10 % (positive and negative) for 6 countries. Good result to mention is that for all HLY indicators there are fewer changes beyond the threshold than for In terms of values the 2009 HLY at birth for women and men range from around 52 to 70, the 2009 HLY at 65 for women range from 2.8 to 14.6 and the 2009 HLY at 65 for men range from around 3.4 to The table in Annex I show the figures for : values and annual changes in percentages as well as line charts reflecting the nature of the trends. (The Excel file is included in the background documents: HLY xls). 1

3 4. Problem statement The value of the HLY as a policy indicator has to be evaluated towards 5 important quality criteria: i.e. relevance, coherence, comparability, clarity and timeliness. (see the ESS code of practice: Relevance Relevance is defined by the need and the use by the different policy users. Since the Lisbon strategy the HLY is established as an important long time series indicator. Several Commission Services have integrated this indicator in policies such as Sustainable Development Strategy, the Open Method of Coordination on Social Inclusion and Social Protection and Health Inequalities. Recently a new EU 2020 flagship initiative on Active and Healthy Ageing (AHA) was set up. This AHA defines as the global headline target an increase of HLY by 2 years in the EU on average. The policy relevance of the HLY is consequently not questionable. Currently computed separately for males and females and at two ages: birth and the age of 65, some new requirements have been asked for the future: HLY at regional level (NUTS 2); HLY at other ages (DG SANCO suggested 'at 55' in the context of health inequalities and possibly also for active healthy ageing initiative) (this is quite simple and we plan to do some simulations in the coming months); HLY by socio-economic status Coherence The link between the Life Expectancy (LE) and the Healthy Life Years indicators is of special interest and some users would expect a close link between the two. But the ranking between countries (for men and women separately) for the two indicators and the trend of the two indicators show a completely different picture. While life expectancy has increased in all countries from 2005 to 2009, the computed indicator for HLY shows a decreasing trend for quite a number of countries even if the trend is diverging between gender/countries. For this reason and on request of DG REGIO, an alternative calculation method was tested based on the SILC self-perceived health question (replacing the GALI question). The corresponding results are much more consistent with the Life Expectancy indicator. For example, female HLY at birth are the highest in Bulgaria in 2007 of all MS. Using health perception, BG comes 25th, which is much closer to its ranking based on life expectancy alone 29th. HLY put Germany and Finland 23rd and 24th in the ranking. HLY based on health perception places them 13th and 7th, considerably closer to their LE ranking (11th and 5th)." However, this new defined indicator does not correspond to the original policy need defining a disability-free expectancy Comparability These characteristics have to be analysed according to 2 dimensions: over time within a country and between countries. For the long time series starting from 1995 the comparability within a country is problematic as different sources and techniques were used. From 2004 onwards the SILC question has been 2

4 introduced but the national version of the GALI question for some countries has changed (see the tables and the annual changes marked in red). For the comparison between countries we mainly have to address the situation that some countries have not yet implemented the standard GALI question. The cultural factor explaining possible differences between countries is a "normal" feature in health survey data and is not further analysed here. Concerning the GALI question the situation can be synthesised as follows 1 : (provisional information) For : AT, BE, FR, IE, LU (2008 to check), MT (2008 to check), RO (2008 to check), SI had an identical GALI question over the period and all have steady trends in HLY in the country reports showing the stability of the GALI; CY, CZ, DK, EE, IT, LV, PL, SK, ES had changes in the GALI over the period but these explained the discontinuities in trends in HLY and in these 9 countries the GALI question is now correct; HU, NL and PT had changes in the GALI over the period which explained discontinuities in trends but the GALI is still not correct; DE, FI, EL, LT, UK there were changes to the GALI explaining discontinuities but we are not sure whether the GALI is now correct as we do not have the EU-SILC 2008 version or it has not been translated; in Bulgaria the GALI does not seem to have changed but there is a big discontinuity in the trend though we do know that the question is not correct; in Sweden there have been changes but the time does not correspond with the discontinuity and we do not know whether the 2008 question is correct. For 2009: The 3 questions of the minimum European health module (so called MEHM) (including the GALI) were OK in 13 countries: BE, CY, CZ, DE, ES, FI, FR, GR, HU, IT LV, SI, and SK. Moreover the GALI was OK in DK; The 3 questions (including the GALI) were OK in 2008 in 3 countries but we don t have the 2009 questions yet: IE, NL and SE. A priori they are OK in 2009; Therefore in total the GALI is now possibly OK in 17 MS out of 27; The GALI question is still not OK for AT, and UK and self-perceived morbidity (Q2) is still not OK for UK; The 2009 questions are available for 7 additional countries but we need to get their back translation in English. They are BG, DK (only for Q1), EE, LT, MT, PL, and PT; The 2008 and 2009 questions are both missing for 3 countries prevent our check. They are LU, NO, and RO; To summarize we need the MEHM questions for 2008 and 2009 in LU, NO and RO and we need back translation for 2009 for BG, DK (only for self-perceived health (Q1)), EE, LT, MT, PL, and PT. 1 EHEMU information and SILC assessment reports 3

5 Repetitive efforts were undertaken to harmonise the national versions of the GALI question in the SILC surveys but a full harmonisation and stability is not yet attained. In that context several assessments took place: At the WG meeting of October 2009 a document was presented with an overview of the SILC questions compared to the EHIS standards (attached in the background documents and also available on Circa in WG meetings: document 2009 Item 7.4 SILC - assessment of the health questions.doc). At the WG "Statistics on living conditions" of May 2011 a document on the improvement of comparability of health variables was presented (the document is included in the background documents). Also SANCO joint action partner (EHEMU) tackles the comparability and translation problem of the SILC question Clarity A first issue concerns the metadata: explanations on breaks in series need to be presented mostly linked to changes of the GALI question in time - as well as the deviations of the national GALI questions towards the standard. A second issue is how to understand or explain the values and the trends especially between countries and in relation to other indicators such as the life expectancy. Examples: Swinging trends (up and down): see tables and charts mentioned before Important breaks: see tables and charts mentioned before Different rankings: for example DK has ranking 21 for HLY at birth for women but for HLY at 65 the rank is 4. On the contrary SE is on rank 1 for all HLY (except the HLY at birth for women where the rank is 2) (see also Annex I) Difference in ordering for the life expectancy (or HLY in percentage of the total life expectancy) (or possibly "opposing" trends in HLY and LE) A last issue is the terminology especially in an international and scientific context. The HLY is for experts a wrongly chose name as it is in fact a disability free life expectancy (DFLE) and as such one type of a healthy life years indicator. However as it is now so well established and long time in use a change in wording would only confuse the users. 5. Future actions 5.1. Proposal for the short to medium term The Eurostat proposal for the short to medium term is to stick to the initial computational method based on mortality statistics and data on self-perceived disability GALI SILC question. But priority is to be given to improving the harmonization of the SILC GALI question, to develop an explanatory analysis for a joint analysis with the life expectancy indicator in close collaboration with the EHLEIS Joint Action and to improve the data dissemination on the Eurostat website. Depending on the output of the WG discussion, additional computations (regional level (NUTS 2); at other ages (50 or 55); by socio-economic status) would be carried out. 4

6 SILC: harmonisation of the GALI question We propose to reinforce the coordination of the national authorities involved in the EHIS and SILC to implement the national version of the standard question which the EHIS projects (Wave I and II) have implemented. The WG public health delegates are asked to check and if needed to reinforce this coordination and updating process. Joint Action EHLEIS DG SANCO and the INSERM of France have recently signed a Joint Action agreement for 36 months. This JA comprises around 20 co-beneficiaries (public health authorities, statistical offices, universities, etc from BE, CZ, DE, DK, EL, FR, IT, NL, SE UK.) Objectives are: computation and dissemination, monitoring and analysing trends and gaps in life and health expectancies, develop methods for computing comparable Health expectancies by socioeconomic status, contribute towards identifying the main determinants of healthy life and research to develop an alternative summary measure of population health (SMPH). The JA will also focus on a broader international context and compare HLY with US and Japan. Furthermore one working package includes also calibration of the GALI with EHIS data and validation of GALI with information from HES (health examination surveys). It is interesting to note that some of the project partners are also strongly involved in EHIS. It is of utmost importance to coordinate Eurostat's activities with those of SANCO and the JA partners, not only in consistency and effectiveness of results, also in sharing work and expertise and becoming as such more cost-efficient. Dissemination: tables, metadata and "pedagogical" explanations We propose to focus on the series from 2005 onwards (SILC based data for all countries) and for the analysis of the trend to start in 2007 (in the understanding that problems caused by the imperfect GALI questions are mostly tackled). Accordingly we propose then to split the web site table in 2. All observations for which the SILC question changed or is not harmonised shall be flagged and the metadata page will link to a tabular overview country/year. On Eurostat's "Statistics Explained" page on HLY we can illustrate how to understand some of the more surprising findings. ( We also propose in collaboration with SANCO and their joint action partner to link to country reports, scientific articles via the Wikipedia and EHEMU-HLY web sites 5.2. Proposal for the longer term Linked and in line with the planned analysis in the JA we propose on a longer term to: calculate HLY based on the EHIS question and analyse the differences with the SILC based HLY; analyse other EHIS data (morbidity, functional limitations, self care, ) in relationship to the GALI and the 2 other MEHM questions. 5

7 6. Decisions for the Working Group Eurostat requests the WG members: 1. to confirm the actual definition and calculation method and to endorse as a first priority the harmonisation of the GALI question in the national SILC surveys; time line ; 2. to support the dissemination plans and contribute to the necessary metadata in explaining difficulties of national series; time line: continuously; 3. to endorse and follow-up the collaboration with the JA EHLEIS for an improved explanatory dissemination and for the comparative analysis with EHIS data; time line /2015; 4. to give opinion on the new requirements for a HLY indicator computed at NUTS 2 level, for other age groups and by socio-economic status; 5. to endorse on the longer term the analysis and development of a revised HLY indicator (SMPH summary measurement of population health); time line

8 Annex I: HLY Data HLY at birth - Women EU27 : : : : EU25 : : : : BE BG : : : CZ : DK DE : EE IE EL ES FR IT : CY : LV : LT : LU HU : MT : NL : AT PL : PT RO : : : SI : SK : FI SE UK : : IC : NO Percentage annual change / in bold/red > ABS(5%) EU % EU % - BE -0.5% 0.2% 1.4% 1.5% 6.5% BG 0.2% -11.2% CZ -1.3% 0.2% 5.7% -0.2% - DK -0.5% -9.9% 0.4% -1.6% -0.9% DE 0.5% -1.5% 0.4% 5.3% - EE 3.1% 4.8% 1.7% 2.8% -2.1% IE 0.3% -0.5% 0.4% 1.4% -0.3% EL -7.4% -1.9% -1.1% 1.0% 3.1% ES -2.2% 0.6% -0.6% 0.3% 1.0% FR -1.6% 0.0% 0.2% -0.4% 0.3% IT % -3.4% -3.6% -5.9% CY 1.1% 3.8% -0.8% 9.2% - LV 3.1% 0.7% 3.0% -1.8% - LT 2.7% 2.8% 2.8% 3.4% - LU 2.3% -0.6% 4.5% -0.5% 3.2% HU 0.0% 0.7% 1.1% 5.7% - MT -1.8% 1.8% 2.0% -1.3% - NL 0.0% -6.1% 0.8% 0.2% - AT 1.8% -2.6% 0.5% 2.0% -1.0% PL -0.8% 2.1% -2.0% -6.1% - PT -2.3% -0.2% -0.6% 1.6% 9.0% RO -1.9% 0.5% SI 1.0% -2.2% 2.1% 1.9% - SK 0.0% -6.4% 2.7% -3.5% - FI -1.7% 2.4% 10.0% 0.6% -0.9% SE 1.2% 3.2% -0.7% 6.3% 3.6% UK - 0.3% 1.5% 0.2% - IC -0.7% -3.1% 9.9% 1.2% - NO -1.5% 4.2% 4.2% -0.4% -1.7% 7

9 HLY at Birth Men E EU27 : : : : EU25 : : : : BE BG : : : CZ : DK DE : EE IE EL ES FR IT : CY : LV : LT : LU HU : MT : NL : AT PL : PT RO : : : SI : SK : FI SE UK : : IC : NO Percentage annual change / in bold/red > ABS(5%) EU % EU % - BE 0.8% -0.2% 0.8% 1.8% 5.7% BG 0.0% -7.6% CZ -0.5% -0.2% 6.1% -0.2% - DK -0.8% -7.6% -0.4% -1.0% 0.1% DE 1.6% -5.1% 0.5% 6.4% - EE 4.0% 6.5% 0.2% 2.9% -3.6% IE 0.8% 0.8% -0.8% 0.5% 0.6% EL -8.0% -0.8% -0.6% 0.9% 3.1% ES -1.9% 0.9% -0.8% 0.8% 1.1% FR 0.2% -1.0% 0.5% 1.1% 1.3% IT % -2.9% -1.5% -3.9% CY 0.9% 2.4% -2.0% 8.1% - LV 2.1% 1.2% 0.8% -0.2% - LT 4.4% 2.2% 1.9% 2.3% - LU 0.5% 4.2% 2.0% -1.9% 5.2% HU 2.0% -0.7% 1.5% 4.2% - MT 0.6% -0.3% 1.2% -0.6% - NL -1.6% -5.0% 1.1% 0.0% - AT 2.1% -0.7% 0.0% 1.0% -0.5% PL -0.5% 1.7% -1.4% -4.6% - PT -1.7% 1.2% -2.2% 2.1% 6.0% RO -0.8% -0.7% SI 2.0% 1.4% 1.7% 2.3% - SK 0.6% -6.5% 2.0% -1.1% - FI -0.9% 3.4% 7.2% 2.3% -2.6% SE 1.9% 2.5% 0.6% 4.5% 3.5% UK - 0.2% -0.2% 2.8% - IC -3.1% -2.6% 6.6% 2.1% - NO -1.6% 5.1% 1.4% 0.3% 0.0% 8

10 HLY at 65 Women E EU27 : : : : EU25 : : : : BE BG : : : CZ : DK DE : EE IE EL ES FR IT : CY : LV : LT : LU HU : MT : NL : AT PL : PT RO : : : SI : SK : FI SE UK : : IC : NO Percentage annual change / in bold/red > ABS(10%) EU % EU % - BE -1.0% -1.0% 5.1% 3.2% 13.1% BG -2.2% -34.0% CZ 2.4% -1.2% 16.9% 2.9% - DK -2.4% -14.0% 1.4% 0.0% 4.4% DE -1.5% -12.0% 2.7% 23.7% - EE 26.2% 2.4% 5.1% 14.7% -24.4% IE 1.9% -1.0% -1.0% 6.1% 2.1% EL -18.5% -13.8% -7.8% 3.0% 5.3% ES -2.3% -13.1% 5.3% 3.3% -4.2% FR -6.1% -1.0% 4.2% 1.1% -5.1% IT % -15.3% -4.5% -27.0% CY 10.4% 5.5% 1.4% 50.0% - LV 16.3% 14.0% 2.4% -22.2% - LT 6.3% 16.7% 3.8% 20.9% - LU -1.7% 8.4% 17.6% -1.1% 1.1% HU -11.1% 8.6% 5.5% 10.0% - MT -1.8% 3.6% 13.4% -12.6% - NL 7.3% -20.7% 8.0% 2.8% - AT 9.6% -5.2% 2.7% 13.6% -5.7% PL -1.3% 7.1% -13.6% -19.8% - PT 0.0% 1.9% -10.2% 15.7% 37.8% RO -10.3% 1.3% SI 6.5% -6.1% 4.2% 11.8% - SK 7.7% -36.6% 7.9% -28.3% - FI 0.0% 1.1% 18.9% 13.8% -5.8% SE 5.8% 0.0% -0.7% 27.5% -0.9% UK - 2.6% 3.6% 0.0% - IC -2.9% -15.2% 28.9% 5.8% - NO -5.4% 13.0% 10.1% 0.8% -5.6% 9

11 HLY at 65 Men E EU27 : : : : EU25 : : : : BE BG : : : CZ : DK DE : EE IE EL ES FR IT : CY : LV : LT : LU HU : MT : NL : AT PL : PT RO : : : SI : SK : FI SE UK : : IC : NO Percentage annual change / in bold/red > ABS(10%) EU % EU % - BE 1.9% 2.0% 6.3% 4.4% 7.1% BG -3.4% -24.3% CZ 8.1% -7.5% 19.4% 3.1% - DK -6.7% -8.4% 4.0% -3.8% -0.8% DE 3.2% -19.5% 0.0% 18.5% - EE 41.0% 11.4% -12.5% 17.6% -26.1% IE 8.5% -1.1% 3.3% 1.1% 7.1% EL -19.1% -9.2% -3.0% 6.3% 1.1% ES -6.1% -4.9% 4.0% 3.1% -1.0% FR 1.1% -6.5% 8.1% 4.9% -1.2% IT % -9.2% -3.3% -19.6% CY 6.5% 4.5% -5.3% 40.3% - LV -2.1% -5.9% 15.9% -12.0% - LT 3.5% 9.6% -10.3% 13.7% - LU 0.0% 20.0% 4.7% -7.5% 8.1% HU 3.6% 3.8% 6.0% 0.0% - MT 5.8% 1.0% 4.0% -5.7% - NL -3.1% -13.4% 2.8% 4.8% - AT 11.0% 0.0% 4.3% 4.5% -6.9% PL -1.4% 6.2% -9.7% -13.3% - PT 0.0% -2.9% 0.0% 9.7% 24.0% RO -6.5% 1.3% SI 1.1% 2.2% 8.4% 12.2% - SK 17.2% -29.3% 2.5% -16.7% - FI 1.3% 2.6% 27.9% -1.6% -11.4% SE 5.4% 0.8% -0.8% 22.9% 4.0% UK - 3.9% -1.0% 1.0% - IC -8.0% -12.1% 15.4% 11.5% - NO -4.3% 13.7% 0.0% 1.6% 2.5% 10

12 HLY: country ranking according descending value of 2009 (For IT and UK: 2008 values) HLY birth women HLY birth men HLY 65 women HLY 65 men 2009 rank 2009 rank 2009 rank 2009 rank standard deviation of rankings AT AT AT AT BE BE BE BE BG BG BG BG CY CY CY CY CZ CZ CZ CZ DE DE DE DE DK DK DK DK EE EE EE EE EL EL EL EL ES ES ES ES FI FI FI FI FR FR FR FR HU HU HU HU IC IC IC IC IE IE IE IE IT IT IT IT LT LT LT LT LU LU LU LU LV LV LV LV MT MT MT MT NL NL NL NL NO NO NO NO PL PL PL PL PT PT PT PT RO RO RO RO SE SE SE SE SI SI SI SI SK SK SK SK UK UK UK UK