Long Term Monitoring of Health Inequalities: Headline Indicators – October 2014

Annex 2: Data sources and quality

Data quality

Except where the source data is held by Scottish Government (i.e. the mental wellbeing indicator), aggregate data is provided by National Records of Scotland for the all-cause mortality and alcohol mortality indicators, and by ISD Scotland for all other indicators in this report. Scottish Government statisticians carry out quality assurance checks on the aggregate data, comparing it with past trends and against other published data, such as national level data published by NRS or ISD. For the mental wellbeing indicator, Scottish Government statisticians quality assure the aggregate data in the same way but take the additional step of double checking the programming methods used to derive the figures within the responsible team.

ISD Scotland and NRS are responsible for the quality assurance of their own datasets. Detailed information on the quality control of the relevant ISD datasets is available online^[12]. National Records of Scotland have published detailed information on the quality of data on deaths^[13]. Analysts at both ISD and NRS are provided with income-employment decile-datazone lookups and population estimates before a request for aggregate data is submitted.

Revisions and timeliness of report

Our general approach to revisions and release schedules is described at the following web address: http://www.scotland.gov.uk/Topics/Statistics/Browse/Health/TrendHealthOutcome/Results

For this year's publication, the mental wellbeing indicator could not be updated as the data would not be available until December 2014. Users will be consulted about the most appropriate time and method for the next update to that indicator. The healthy life expectancy indicator is updated every second year and will next be included in the 2015 report. The following revisions apply to this year's report:

• Revisions to income-employment methodology

Income-employment deprivation deciles have been recalculated for all years. This takes account of revised population estimates between years 2002 and 2010, by datazone, published by National Records of Scotland in May 2014. Other changes to the application of SIMD in the income-employment decile methodology were implemented, in discussion with the technical group and Scottish Government statisticians. These are:

For years 1997 up to 2004, deciles have been derived from the income and employment domains of SIMD 2004. This is a revision from previous reports which defined income-employment deciles on SIMD 2006 for the earliest years in the time series. This change has a relatively small impact on inequalities compared to previous reports. The benefit is that deciles are now based on SIMD data closer to the year of each point in the time series.

Datazone rankings for the SIMD 2009 income domain were corrected in 2010, but the corrections were not reflected in this report until this year's update. For 2007 and 2008, the revised version of the SIMD 2009 income domain has been used in the derivation of income-employment deciles in this report. The impact of the correction for the full list of datazone ranks is shown at the following web address: http://www.scotland.gov.uk/Topics/Statistics/SIMD/SIMD09V2DZImpact

For 2009, the first annual update to income and employment data in SIMD 2009 (denoted SIMD 2009 + 1) has been used to derive income-employment deciles. This is a change from previous reports which used the unrevised, original SIMD 2009 ranks. This change also means deciles are based on underlying SIMD data closer to the relevant point in the time series.

For 2010, the second annual update (denoted SIMD 2009 + 2) has been used, in line with previous reports. For 2011 onwards, SIMD 2012 has been used, also in line with previous reports.

Information on use of SIMD to analyse change over time is provided on the Scottish Government website at the following address: http://www.scotland.gov.uk/Topics/Statistics/SIMD/ChangeOverTimePaper

• Rebased population estimates for Scotland

Population estimates for Scotland between the years 2002 and 2010 were updated by National Records of Scotland (NRS) in December 2013, and then in May 2014 at datazone level. The revised estimates are based on the 2011 Census. Further information is available on the NRS website: http://www.nrscotland.gov.uk/statistics-and-data/statistics/statistics-by-theme/population/population-estimates/mid-year-population-estimates/mid-2002-to-mid-2010-revision

These revisions impact the calculation of age-standardised rates and also the income-employment methodology described above.

• New European Standard Population

Because the 2013 ESP is more heavily weighted towards older ages than the 1976 ESP, there is a significant increase in most age-standardised rates when using the 2013 ESP. While age-standardised (mortality and incidence) rates included in this report are all shown for specific age groups, there is still a significant impact of moving to the new ESP. Although there is a step change in rates, patterns over time are unchanged. Detailed information on the impact of moving to the 2013 ESP on mortality rates in Scotland has been published by NRS: http://www.nrscotland.gov.uk/files//statistics/age-standardised-death-rates-esp/age-standard-death-rates-background.pdf

The focus of this report is on inequalities, i.e. the absolute and relative differences in rates between areas on a scale of deprivation. The scale of impact of the new ESP is smaller when measuring relative inequalities. Analysis of the net impact of revisions to income-employment methodology, rebased population estimates and the new ESP on all indicators was carried out in the quality assurance stage of this report's production. The effect is slightly greater in later years, for two reasons. Firstly, population estimates being rebased on the 2011 Census has a greater impact in those years. Secondly, the correction to the application of SIMD 2009 for years 2007 and 2008, and the move to using the first annual update of SIMD 2009 for the year 2009, introduces small changes for those years.

For all indicators, patterns over time followed a similar pattern to those previously reported. The following chart illustrates the net impact of these methodological changes on relative inequality trends for the premature mortality indicator.

• Probable suicide coding

There was a change to the way suicides were coded following an update of death classification by the World Health Organisation (WHO). This resulted in some deaths previously coded under 'mental and behavioural disorders due to psychoactive substance use' now being classed as 'self-poisoning, and the change was reflected in the 2013 version of this report with a 'break in series' in trends from 2011. This year, probable suicide rates have been presented entirely on the basis of the old coding, in order to aid interpretation of long-term trends.

• Low birthweight: RII

Quality assurance of data for previous years has identified that relative inequality in the low birthweight indicator was incorrectly displayed for the year 2008 in previous versions of this report. RII had been erroneously displayed as 0.87 rather than 0.77 due to a data input error.

Pre-release access

In accordance with the Pre-release Access to Official Statistics (Scotland) Order 2008, pre-release access to these statistics was provided to Scottish Government policy and communications colleagues 5 working days before release for the purposes of briefing ministers. NHS Health Scotland were also provided pre-release access from 5 working days before release for the purposes of making a statement or issuing a press release at the time of, or shortly after, publication.

Indicators

• Premature Mortality (from all causes, aged under 75 years)

Source: National Records of Scotland.
Definition: European age-standardised rates of deaths from any cause amongst those aged under 75 years.

• Low birthweight

Source: NHS Information Services Division (ISD); SMR02 maternity dataset.
Definition: The figures are presented as a percentage of all live singleton births (not including home births or births in non-NHS hospitals). Low birthweight is defined as <2,500g - the standard World Health Organisation definition. Figures for the most recent year are provisional.

• Healthy birthweight

Source: NHS Information Services Division (ISD); SMR02 maternity dataset.
Definition: A baby is considered to be of healthy birthweight (a weight appropriate for its gestational age) when it lies between the 5th and 95th centile for weight at its gestational age. Gestational age is a way of expressing the age or development of a baby. It is typically based on an antenatal ultrasound scan. However, it may also be estimated from the number of weeks since the mother's last normal menstrual period. Centiles for birthweight charts for gestational age are derived from Scottish data on births between the years 1998 and 2003^[14].

• Coronary Heart Disease - first ever hospital admission for heart attack aged under 75 years

Source: NHS Information Services Division (ISD); SMR1/01 records (all inpatient and daycase discharges).
Definition: European age-standardised rates of first ever hospital admission for acute myocardial infarction (heart attack) amongst those aged under 75 years. The following World Health Organisation International Classification of Disease coding was used: ICD10 'I21-I22'; ICD9 '410'.

• Coronary Heart Disease - deaths aged 45-74 years

Source: NHS Information Services Division (ISD); using deaths data from National Records of Scotland.
Definition: European age-standardised rates death from coronary heart disease amongst those aged 45-74 years. The following World Health Organisation International Classification of Disease coding was used: ICD10 'I20-I25'; ICD9 '410-414'. Because of the dynamic nature of the linked database, previous years' data are sometimes updated in subsequent publications.

• Cancer - incidence rate aged under 75 years

Source: NHS Information Services Division (ISD); Scottish Cancer Registry.
Definition: European age-standardised rates of new cases of cancer amongst those aged under 75 years.

All Cancers- cancer defined as all malignant neoplasms excluding non-melanoma skin cancer. The following World Health Organisation International Classification of Disease coding was used: ICD10 'C00-C96' excluding 'C44' (the Scottish Cancer Registry does not use code 'C97').
Prostate cancer (males only)- ICD-10 C61
Breast cancer (females only)- ICD-10 C50
Cancer of the trachea, bronchus and lung- ICD-10 C33-C34
Colorectal cancer- ICD-10 C18-C20

• Cancer - deaths aged 45-74 years

Source: NHS Information Services Division (ISD); Scottish Cancer Registry.
Definition: European age-standardised rates of deaths from cancer amongst those aged under 45-74 years.
All cancers- cancer defined as all malignant neoplasms excluding non-melanoma skin cancer. The following World Health Organisation International Classification of Disease coding was used: ICD10 (2000 onwards) 'C00-C97' excluding 'C44'.
Prostate cancer (males only) - ICD-10 C61
Breast cancer (females only) - ICD-10 C50
Cancer of the trachea, bronchus and lung- ICD-10 C33-C34
Colorectal cancer- ICD-10 C18-C20

• Alcohol - first ever hospital admission aged under 75 years

Source: NHS Information Services Division (ISD).
Definition: European age-standardised rates of first ever hospital admission for alcohol related conditions amongst those aged under 75 years. These rates include hospitals discharges where alcohol-related problems are recorded as either primary or secondary reasons for admission to hospital and will cover first ever admission since 1981 (a patient may have had admissions prior to 1981 which would not be recorded in this database). These figures exclude private hospitals, mental illness hospitals, psychiatric units and maternity hospitals and include Scottish residents only. Caution is necessary when interpreting these figures. The recording of alcohol misuse may vary from hospital to hospital. Where alcohol misuse is suspected but unconfirmed it may not be recorded by the hospital. The following revised World Health Organisation International Classification of Disease coding was used: ICD10: F10, K70, X45, X65, Y15, Y90, Y91, E244, E512, G312, G621, G721, I426, K292, K860, O354, P043, Q860, T510, T511, T519, Y573, R780, Z502, Z714, Z721.

• Alcohol - deaths aged 45-74 years

Source: National Records of Scotland.
Definition: European age-standardised rates of death from alcohol related conditions amongst those aged 45-74 years. The definition of alcohol related deaths includes deaths where there was any mention of alcohol related conditions on the death certificate, rather than just as the main cause of death. The following World Health Organisation International Classification of Disease coding was used: ICD10 F10, G31.2, G62.1, I42.6, K29.2, K70, K73, K74.0, K74.1, K74.2, K74.6, K86.0, X45, X65, Y15; ICD9 291, 303, 305.0, 425.5, 571.0, 571.1, 571.2, 571.3, 571.4, 571.5, 571.8, 571.9, E860.

• All-cause mortality aged 15-44 years

Source: National Records of Scotland.
Definition: European age-standardised rates of deaths from any cause amongst those aged 15-44 years. Specific breakdowns for deaths from assault, drug related deaths and suicide are also provided, as the major causes of death for which there are large inequalities amongst young people. There may be some double counting in these breakdowns. The following World Health Organisation International Classification of Disease coding was used: Assault ICD10 'X85-Y09', 'Y87.1' ICD9 'E960-969'; Drug-related ICD10 'F11-16', 'F19', 'X40-44', 'X60-64', 'X85', 'Y10-Y14'; Suicide (intentional self-harm + undetermined intent) ICD10 'X60-84', 'Y87.0' ICD9 'E950-959', 'E980-989'.