2. Spatial distribution of older people and disability in Scotland
2.1 Aim and approach
The first stage of the project consisted of investigating the spatial distribution of older people, including older people with disabilities, using small-scale geographical units ( i.e. data zones). The aim is to evaluate and compare how these groups of people locate across Scotland using the Scottish Government 8-fold urban-rural classification defined at the level of data zones.  The eight types of areas are shown in Figure 2: Large Urban Areas (over 125,000 people); Other Urban Areas (10,000-125,000 people); Accessible Small Towns (3,000-10,000 people, within a 30 minute drive time to an urban area); Remote Small Towns (3,000-10,000 people, with a drive time between 30 and 60 minutes to aa urban area); Very Remote Small Towns (3,000-10,000 people, with a drive time of over 60 minutes to an urban area); Accessible Rural Areas (less than 3,000 people, within a drive time of 30 minutes to an urban area); Remote Rural Areas (less than 3,000 people, with a drive time of between 30 and 60 minutes to an urban area); Very Remote Rural (less than 3,000 people, with a drive time of over 60 minutes to an urban area).
Data zones are small geographical units derived from Census output areas, and are designed to contain between 500 and 1,000 household residents and be representative of communities, 'fit inside' larger boundaries ( e.g. those of local authorities) and can be combined to other larger areas. Statistics based on data zones are therefore more 'fine-grained' than statistics presented for larger administrative areas ( e.g. local authorities) and allow for a more detailed spatial analysis of the distribution of older people.
Whilst it is common knowledge that older people tend to be more prevalent in sparsely populated areas, there is limited understanding of how (if) this pattern differs across different types of rural areas and small towns, compared to urban areas. By combining data zone level data with the 8-fold urban-rural classification, we can provide a more detailed understanding of such patterns.
Figure 2: The 8-fold urban-rural classification, 2011/2012  .
The first task was to assess which data could be used to measure 'older people' and 'disability'. Based upon researchers' past experience in identifying and using small scale population data for other projects ( e.g. Copus and Hopkins, 2015; Slee et al., 2014), Table A.1. in Appendix A provides a summary of the data sources used in the mapping and analysis of the spatial distribution of older people and people with disabilities. Two key sources of data were identified and used:
- Scotland's Census 2011. Data tables at 2011 data zone level were available from the Data Warehouse facility  , as was an index of Census output tables.
- Scottish Neighbourhood Statistics ( SNS) data download facility  provided a range of indicators related to population characteristics, services, communities and economic measures which are available at different geographical scales.
Defining older people and disability
The 2011 Census provided detailed information on population age, which was aggregated to produce totals of people aged 65 or over and 85 or over in each data zone. In addition, the household population who were both aged 65 or over, and were living alone/in one person households (a group who may be particularly isolated) was also extracted. The definition of disability within the Equality Act (Section 6(1) of the Equality Act 2010) provided guidance in the selection of variables. The total populations whose day-to-day activities were limited either 'a little' or 'a lot' by a long-term health problem or disability, and the numbers who reported  that they had a) a physical disability and b) a learning disability were derived and extracted at data zone level from Census tables. Additionally, the number of people who were both aged 65 or over, and had a limiting long-term health problem or disability, was identified and extracted.
In addition to the Census data, data on benefit claimants, hospital admissions and cancer registrations available via SNS were also used to measure the incidence of disability. As people with coronary heart disease and cancer are disabled  , data available from SNS on the number of admissions to Scottish hospitals with a main diagnosis of these diseases was used (these data are based on data zone residents, rather than the location of hospitals)  . Furthermore, information on the number of new cancer registrations among data zone residents over the 2005-9 period was used. In addition, information on the number of claimants of Attendance Allowance, a benefit eligible to disabled people aged 65 or over for care purposes  , was available via SNS for all people (aged 65 or over) and for men and women. Information for hospital admissions was based on the year 2012 and Attendance Allowance data used relates to the fourth quarter of 2012; data on new cancer registrations is based on the 2005-9 period as noted above.
The spatial analysis considered the measures described above, which are shown in Table 1. For space reasons we present and discuss only the key variables of interest in the main body of the report (highlighted in bold in Table 1), while the remaining variables are presented and discussed in Appendix B.
Table 1: Measures of age and disability used in the spatial analysis.
*The mapping and spatial analysis of these variables is presented in Appendix B.
Measuring the spatial distribution of older people and disability in Scotland
To evaluate the spatial distribution of older people, and older people with disabilities, across Scotland we compared the number of people with, or frequency of, protected characteristics in a given data zone with the number of people with, or frequency of, protected characteristics for the whole of Scotland. This was done by constructing location quotients for each of the variables in Table 1. Location quotients ( LQ) provide an easy and intuitive way of quantifying the degree to which a certain attribute ( e.g. age above 65 years) is concentrated in a given region compared to the national average ( i.e. the reference). Consider the following example. Across Scotland, 16.8% of the population was aged 65 or over in 2011. If the proportion of people aged 65 or over was equally distributed across urban and rural areas in Scotland ( i.e. zero concentration), the LQ would be equal to 1 (or 100%) for all rural and urban areas in Scotland. However, this is not likely to be the case and indeed we know that the ratio of local-to-national shares of older people tends to be greater than 1 (or 100%) for rural areas and smaller than 1 (or 100%) for urban areas. This in turn indicates that rural areas have higher shares of older people than expected ( i.e. the national reference or average). The percentages can be interpreted in a straightforward way: a figure of 200% shows that the number of people with a protected characteristic is double that expected ( i.e. according to national average), and a figure of 50% means that the population with a protected characteristic is exactly half that expected ( i.e. according to national average). Table 2 shows the incidence of the measures considered in the spatial analysis for overall Scotland.
Table 2: Incidence of protected characteristics (age and disability) for overall Scotland.
|Protected characteristic group||Specific characteristic/indicator||Denominator group||% / ratio|
|Age||Population aged 65 or over||Total population (2011)||16.81|
|Age||Population aged 85 or over||Total population (2011)||2.00|
|Age||Household population aged 65 or over and living alone||Total household population (2011)||6.00|
|Age||Male population aged 65 or over||Total population (2011)||7.25|
|Age||Female population aged 65 or over||Total population (2011)||9.57|
|Disability||Population with limiting long-term health problem or disability||Total population (2011)||19.65|
|Disability||Population with physical disability||Total population (2011)||6.71|
|Disability||Population with learning disability||Total population (2011)||0.50|
|Disability||Hospital admissions with a diagnosis of coronary heart disease||Total population (2012)||0.0049|
|Disability||Hospital admissions with a diagnosis of cancer||Total population (2012)||0.0267|
|Disability||Number of new cancer registrations||Total population (2009)||0.0273|
|Age and disability||Population aged 65 or over and with a limiting long-term health problem or disability||Total population (2011)||8.94|
|Age and disability||Claimants of Attendance Allowance||Total population (2012)||2.97|
|Age and disability||Male claimants of Attendance Allowance||Total population (2012)||0.97|
|Age and disability||Female claimants of Attendance Allowance||Total population (2012)||2.00|
Statistics are percentages given to two decimal places, or the ratio of the number of hospital admissions in question/new cancer registrations to the total population (given to four decimal places). These figures were used to calculate the expected values in regions and data zones.
The location quotients (expressed in %) were calculated for the variables in Table 1 and each individual data zone to produce: (i) maps showing the values obtained for each individual data zone; (ii) summary tables showing the average values for each of the eight types of areas defined as per the SG 8-fold urban-rural classification. During the analysis, for some indicators, some of the more extreme data zone-level values are highlighted: these are identified as possible locations of interest and are not, by themselves, representative of Scotland-wide patterns.
Section 2.2 presents and discusses the main findings from the spatial analysis of older people ( section 2.2.1.), older people with disabilities ( section 2.2.2), and overall main conclusions ( section 2.2.3.).
Overall, the information presented in tables 3-4 and figures 3-4 provides confirming evidence on some key well-know facts about population ageing, youth out-migration to cities, and age structures in Scotland: that there is a disparity in the share of older people in the total population between urban and non-urban areas in Scotland, and that this disparity becomes significantly greater for remote parts of Scotland ( i.e. 'remoteness effect'). The information presented also shows that in some cases accessible rural areas behave much in the same way as urban areas (Table 4). Despite these broad-brush patterns, however, there appears to be considerable variation in the spatial concentration of older people across data zones within each of the eight types of areas (figures 3-4).
Population aged 65 or over
Although urban areas contained around two thirds of Scotland's population aged 65 or over in 2011, the share of older residents who were living in large urban areas was below the national average, especially in large urban areas. On the other hand, the share of older residents was higher than expected ( i.e. national average) for all types ( i.e. accessible, remote, very remote) of small town and rural area, but was particularly higher for remote areas. There is hence evidence of a 'remoteness effect': residents aged 65 or over tended to be considerably more concentrated in remote and very remote areas. Remote small towns and very remote rural areas each had an older population that was almost 30% higher than that expected. This overall pattern is supported by data zone-based mapping in Figure 3, which suggests a tendency for more remote areas of Scotland (parts of Dumfries and Galloway, Argyll, some more isolated islands and the far north) to have well above-expected numbers and shares of older residents. By contrast, dark blue coloured areas (indicating that the population aged 65 or over was half that expected or less) were concentrated in the large cities, and areas surrounding Aberdeen, Edinburgh and Glasgow appear to have contained relatively low numbers of older residents. Very remote rural data zones where the population aged 65 or over was more than double that expected are at Dornoch (S01010760), Arran (S01011174 and S01011176) and parts of Argyll and Bute (S01007351, S01007354, S01007330).
Table 3: Population aged 65 or over.
|Region||Total population||Population aged 65 or over||Population aged 65 or over (% of expected)||% of Scotland population||% of Scotland population aged 65 or over|
|Very Remote Rural||162,017||35,209||129.25||3.06||3.95|
|Very Remote Small Towns||67,549||12,988||114.36||1.28||1.46|
|Remote Small Towns||134,493||29,102||128.70||2.54||3.27|
|Accessible Small Towns||472,352||85,329||107.44||8.92||9.58|
|Other Urban Areas||1,640,430||274,272||99.44||30.98||30.81|
|Large Urban Areas||2,061,792||313,056||90.31||38.94||35.16|
Figure 3: Population aged 65 or over, Scotland.
Population aged 85 or over
The pattern found for a second cohort of older people - the population aged 85 or over in 2011 - is similar to that found in the analysis of the population aged 65 or over. The share of the population aged 85 or over in remote and very remote rural areas and remote and very remote small towns is far higher than that expected ( i.e. national average), while it tends to be lower than expected for urban areas and accessible rural areas. This is especially true for remote and very remote small towns, with a population aged 85 or over 49% and 21% respectively higher than expected. Figure 4 clearly shows more spatial variation in data zone-level values than for the 65 or over population described above. Parts of remote small towns where this older population is particularly over-represented include data zone S01007010 (at Huntly in Aberdeenshire), where the population aged 85 or over is more five times that expected. Two remote small town data zones in the North Berwick area (S01008269, S01008270) have more than four times the expected number of residents aged 85 or over. It is possible that the locations of care homes and sheltered housing may influence some of these particularly extreme values.
Table 4: Population aged 85 or over.
|Region||Total population||Population aged 85 or over||Population aged 85 or over (% of expected)||% of Scotland population||% of Scotland population aged 85 or over|
|Very Remote Rural||162,017||3,973||122.62||3.06||3.75|
|Very Remote Small Towns||67,549||1,635||121.03||1.28||1.54|
|Remote Small Towns||134,493||4,002||148.79||2.54||3.78|
|Accessible Small Towns||472,352||9,801||103.75||8.92||9.25|
|Other Urban Areas||1,640,430||31,943||97.37||30.98||30.16|
|Large Urban Areas||2,061,792||39,754||96.41||38.94||37.54|
Figure 4: Population aged 85 or over, Scotland.
Population aged 65 or over and with a limiting long-term health problem or disability
Table 5 shows that people aged 65 or over who were affected by a long-term health condition or disability were over-represented in remote and very remote areas, but particularly in remote small towns. By contrast, the share of older people with a long-term health condition or disability was only slightly above expected for accessible small towns and under-represented in accessible rural areas and, although to less extent, in large urban areas. Figure 5 shows a similar spatial pattern to that of the population who were aged 65 or over (Figure 3), with a contrast between a high number of data zones with above-expected numbers of older people with long-term conditions in remote areas, and below-expected populations in more accessible urban and small town areas and urban areas. Similarly to the previous measures for the spatial concentration of older people, table 5 and figure 5 also reveal a 'remoteness effect' in the spatial distribution of older people suffering from limiting long-term health problems or disabilities.
Table 5: Population aged 65 or over and with a limiting long-term health problem or disability.
|Region||Total population||Population aged 65 or over with limiting condition/disability||Population aged 65 or over with limiting condition/disability (% of expected)||Population with limiting condition/disability (% of 65 or over population)||% of Scotland population||% of Scotland population aged 65 or over with limiting condition/disability|
|Very Remote Rural||162,017||16,916||116.76||48.04||3.06||3.57|
|Very Remote Small Towns||67,549||6,696||110.86||51.56||1.28||1.41|
|Remote Small Towns||134,493||14,931||124.15||51.31||2.54||3.15|
|Accessible Small Towns||472,352||44,278||104.83||51.89||8.92||9.35|
|Other Urban Areas||1,640,430||148,142||100.99||54.01||30.98||31.29|
|Large Urban Areas||2,061,792||175,744||95.33||56.14||38.94||37.12|
Figure 5: Population aged 65 or over with a limiting long-term health problem or disability, Scotland.
Claimants of Attendance Allowance
Attendance Allowance is a benefit which is eligible to individuals aged 65 or over who are disabled and require care  . The proportion of claimants of attendance allowance was well above the national average in very remote and remote small towns and very remote rural areas. By contrast, they were below the national average in accessible and remote rural areas, especially in accessible rural areas (only 78% of national average). The proportion of claimants of attendance of allowance was similar to the national average in other urban areas, accessible small towns, and large urban areas. This pattern is illustrated in Figure 6 at the level of data zones.
There are some interesting differences between this indicator and the previous indicator ( i.e. older people affected by limiting long-term health conditions or disabilities), particularly for accessible rural areas and large urban areas. While the proportion of older people with a limiting long-term health condition or disability is equally under-represented in both accessible rural areas and large urban areas (see Table 5), the proportion of older people claiming attendance allowance is well under-represented in accessible rural areas (only 78% of national average) but slightly over-represented in large urban areas (3% above national average). It is not clear what underlies these differences; reasons can include one or a combination of the following factors: a) differences in income levels, and hence the capacity to cover for disability-related costs, between older people in large urban areas and accessible rural areas, b) differences in the degree of severity of disability, and hence disability-related costs, faced by older people in large urban areas and accessible rural areas, and c) differences in benefit fraud between older people in large urban areas and accessible rural areas. Evidence on residence-based hourly rates of pay for 2014 indicates that median hourly rates are highest for accessible rural areas (£12.42) and smallest for remote rural areas (£11.28), with an intermediate value (£11.55) for the rest of Scotland ( i.e. urban areas and small towns)  . In addition to this, data from the Scottish Household Survey ( SHS) indicates that the percentage of households with lower net annual income levels is largest in large urban areas and smallest in accessible rural areas, while the percentage of households with higher net annual income levels is highest in accessible rural areas and among the lowest in large urban areas  . Although these figures do not prove that the disparity in the expected shares of attendance allowance claimants between large urban areas and accessible rural areas results from differences in income levels, they suggest this may a possible, even if partial, explanation. Future research may wish to address this issue more specifically, however such endeavour is not part of the scope of this study.
Table 6: Claimants of Attendance Allowance.
|Region||Total population||Attendance Allowance claimants*||Attendance Allowance claimants (% of expected)||Attendance Allowance claimants (% of 65 or over population)||% of Scotland population||% of Scotland's Attendance Allowance claimants|
|Very Remote Rural||160,772||5,290||110.73||14.38||3.03||3.35|
|Very Remote Small Towns||68,037||2,530||125.14||18.74||1.28||1.60|
|Remote Small Towns||124,252||4,535||122.83||15.80||2.34||2.87|
|Accessible Small Towns||456,662||13,885||102.33||15.94||8.59||8.79|
|Other Urban Areas||1,615,398||48,670||101.39||17.14||30.40||30.80|
|Large Urban Areas||2,062,877||63,155||103.03||19.71||38.82||39.96|
* Note that the regional numbers of claimants is summed from data zone-level figures for the fourth quarter of 2012 which were rounded from actual values.
Figure 6: Claimants of Attendance Allowance, Scotland.
Figures 7 and 8 provide a summary of the level of over- or under- representation of older people and older people with limiting long-term health conditions or disabilities for each of the areas defined in the SG 8-fold urban-rural classification. On these figures, for each of the eight areas, dots to the right of the vertical grey line show that the indicated protected characteristic was found more frequently than expected in that area.
Figure 7 indicates that remote and very remote rural areas and small towns have a higher than expected concentration of older people, a pattern which is present across all indicators related to age. It is also apparent that older people become more over-represented in progressively more remote rural areas, and that older men are more over-represented in all types of rural areas in comparison to older women. This may be related to the historically higher employment in land based industries ( e.g. agriculture) in rural areas, which tend to employ far higher numbers of men than women.
The indicators related to age and disability (Figure 8) show that the spatial concentration of the population who are both older and affected by a disability or a limiting long-term health condition is larger than expected in remote and very remote small towns, and in very remote rural areas. The spatial concentration of older and disabled population is below expected ( i.e. national average) in accessible rural areas and large urban areas, but the proportion of older and disabled population claiming attendance allowance is well under-represented in accessible rural areas whilst slightly over-represente din large urban areas. The reasons for this disparity were not addressed in this study, but a simple comparison of median earnings between accessible rural areas and non-rural areas revealed that earnings are higher in accessible rural areas, which could help explain (at least partially) the difference.
Although the above than expected figures of older people and older people with disabilities or limiting long-term health conditions in remote and very remote small towns and rural areas represent considerably smaller absolute numbers of people when compared with large urban areas, the negative effect on the well-being and resilience of communities is likely to be stronger in in the former. These areas are remote and hence face challenges in terms of accessibility to important medical and care facilities.
Figure 7: Summary of indicators related to age.
Figure 8: Summary of indicators related to age and disability.
Table 7 shows pairwise correlations between data-zone level populations of older people and protected characteristics related to disabilities. The value of the coefficients of pairwise correlation is high, with values between 0.53-0.71, for the association between older people and limiting long-term conditions or physical disabilities: that is, the areas with high numbers of older people typically also tend to have high populations affected by limiting long-term conditions or physical disabilities. However, the value of the coefficients of pairwise correlation is much smaller for the association between older people and learning disabilities, suggesting a weak co-occurrence of these attribute across data zones. The association between older age and higher rates of disability is expected (Scottish Government, 2014:110); this is because physical and limiting long-term health problems tend develop and appear as people become older. The same relation, however, does not apply to the occurrence of learning disabilities, which explains the largely smaller pairwise correlations.
Table 7: Correlations between data zone-level protected characteristic frequencies.
|Limiting long-term health problem or disability||Physical disability||Learning disability|
|Aged 65 or over||0.63||0.59||0.19|
|Aged 85 or over||0.56||0.53||0.18|
|Aged 65 or over and living alone||0.71||0.68||0.26|
All figures show Spearman's rank correlation coefficient to two decimal places. All correlations are significant at the 99% confidence interval.
Across all indicators, there is a tendency for the frequencies of older people and older people with disabilities to be higher in remote/very remote small towns and rural areas compared to the national average. By contrast, the frequencies of older people and older people with disabilities tend to be smaller than expected or similar to the national average, in urban areas and smaller than expected in accessible rural areas. It is important to acknowledge that these rural areas and small towns have total populations which are very low in comparison with urban areas. Therefore, small absolute numbers of people with protected characteristics, or relatively small counts of protected characteristics, will affect the figures for these regions. However, the higher than expected frequencies of people with protected characteristics in remote areas of Scotland should not be ignored. In particular, the high concentrations of potentially vulnerable older and disabled people in remote and very remote small towns require careful consideration, despite the larger overall numbers of older and disabled people in urban areas. Some remote small towns in the south-west of Scotland were identified by Atterton (2012) as being particularly vulnerable based on an index of demographic and economic indicators, and work in 2015 found that the socio-economic performance of remote small towns was, on average, poorer than that of accessible small towns and rural areas (Copus and Hopkins, 2015). Accessible rural areas, where protected characteristics related to age and disability have been found to be under-represented, also have stronger socio-economic performance compared with remote rural areas and remote/accessible small towns (Copus and Hopkins, 2015): the demographic profile and relative wealth of these areas may be associated with the under-representation of old age and disability.
Email: Graeme Beale, email@example.com