SCOTTISH DIABETES SURVEY 2002
RESULTS
Overview
Each NHS Board was asked to submit data to the Monitoring Group by 30 September 2002. Data was submitted by 14 boards. No data was submitted from Orkney and so 2001 survey data has been used throughout. It is clear that during 2002 most areas were in a period of development - establishing registers, implementing clinical systems and collating and validating data. The net result is that the 2002 Survey is a record of consolidation rather than significant advance. Nevertheless, it is encouraging that the data which were submitted for the 2002 Survey are more robust than the data available in 2001.
All figures shown in this report represent percentage information for those patients registered and included in the survey. The percentage information shown therefore overestimates the actual screening performance in any area. The amount of this overestimate is determined by the prevalence of registered diabetes compared to the true prevalence of diabetes in any area. Thus, those areas which have a high prevalence of registered diabetes may seem to have similar screening performance to those of other areas when in fact they are covering their population in a far more robust way. In the 2003 survey the Monitoring Group intend to use the estimated prevalence of diabetes in each area, corrected for population demographics, and use these figures to produce more meaningful comparisons across Scotland.
Prevalence of diabetes and prevalence of registered diabetes
The 2002 Scottish Diabetes Survey identified 103,835 people with diabetes in Scotland.
The prevalence of diabetes and the number of people reported on local diabetes systems and reported to the Scottish Diabetes Survey are not yet synonymous. The gap between these two figures will not begin to close until diabetes clinical management systems are implemented and in use throughout Scotland. Even then, there are always likely to remain some people with diabetes who have not been diagnosed and therefore are not known to the health service. Clinical management systems can only ever be a record of diagnosed diabetes; it is a matter of research to calculate how many people remain undiagnosed. As noted in last year's report, estimates of the overall prevalence of clinically diagnosed diabetes in Europe based on published studies suggested a prevalence of approximately 3% in 1997. Projections indicated an increase to around 3.6% by 2000 and to over 4% by 2010. (11)
Acknowledging that the true prevalence may be around 3.6%, using figures from Tayside, which is recognised in Scotland as having the most accurate diabetes register, it is estimated that about 3% of the Scottish population have been diagnosed as having diabetes; that is, over 150,000 people. Table 1 looks at the picture across Scotland if this estimate (3%) is correct. The expected figures for individual areas has been calculated by applying national age-specific rates for diabetes to NHS Board populations. Differences in the expected rate between NHS Boards reflects differences in age structure only. This expected rate does not take account of sex, ethnicity or deprivation which will also have an influence on the prevalence. Table 1 shows that two areas have already exceeded this expected prevalence. This is perhaps evidence that our estimate of 3% is too low. It is recognised that more sophisticated estimates of prevalence are required and the Monitoring Group recommends that work be commissioned to develop more accurate ways of deriving figures.
The box below summarises the ways in which registers may over- or under-estimate the number of patients.
Potential reasons for incorrect ascertainment |
Possible reasons for higher than expected figures:- | Possible reasons for lower than expected figures:- |
Duplicate entries/patients with multiple records (perhaps due to name change or moving house) Inclusion of patients who have moved out of the area Inclusion of patients who have died
| |
Table 1: Expected prevalence of diabetes in Scotland 2002
| Population (a) | % of Scottish population | Expected pop. with diabetes Age std rate/100 pop (b) | Expected pop. with diabetes Age std
(Number) | Registered people with diabetes
(2002 SDS) | Estimated pop. with diabetes registered by September 2002 |
| Scotland | 5,114,600 | 100.0% | 3.00 | 153,438 | 103,835 | 67.7% |
1 | Argyll & Clyde | 423,500 | 8.3% | 3.06 | 12,940 | 9,522 | 73.6% |
2 | Ayrshire & Arran | 373,400 | 7.3% | 3.16 | 11,802 | 9,026 | 76.5% |
3 | Borders | 106,900 | 2.1% | 3.44 | 3,678 | 2,929 | 79.6% |
4 | Dumfries & Galloway | 145,800 | 2.9% | 3.47 | 5,057 | 5,150 | 101.8% |
5 | Fife | 350,400 | 6.9% | 3.06 | 10,724 | 9,920 | 92.5% |
6 | Forth Valley | 278,000 | 5.5% | 2.99 | 8,318 | 6,845 | 82.3% |
7 | Grampian | 523,400 | 10.2% | 2.92 | 15,309 | 5,726 | 37.4% |
8 | Greater Glasgow | 904,400 | 17.7% | 2.87 | 25,984 | 4,191 | 16.1% |
9 | Highland | 208,600 | 4.1% | 3.16 | 6,590 | 2,156 | 32.7% |
10 | Lanarkshire | 562,000 | 11.0% | 2.85 | 16,027 | 16,358 | 102.1% |
11 | Lothian | 783,600 | 15.3% | 2.84 | 22,268 | 18,917 | 85.0% |
12 | Orkney | 19,480 | 0.4% | 3.24 | 631 | 377 | 59.7% |
13 | Shetland | 22,440 | 0.4% | 2.83 | 634 | 608 | 95.9% |
14 | Tayside | 385,500 | 7.5% | 3.25 | 12,546 | 11,277 | 89.9% |
15 | Western Isles | 27,180 | 0.5% | 3.42 | 930 | 833 | 89.6% |
aEstimated population figures at 30 June 2000
b expected figure has been calculated by applying national age-specific rates to NHS board population. Differences in expected rate between NHS boards reflects differences in age structure only. The rate does not take account of sex, ethnicity or deprivation which will also have an influence on the prevalence.
Figure 2: Diabetes register: percentage of total population
In the 2002 Scottish Diabetes Survey 2.03% of the Scottish population have been identified as having diabetes. As gaps in ascertainment have been identified by NHS Boards it is unsurprising that this recorded percentage is less than the predicted prevalence of diabetes in Scotland. In particular, certain areas have only been able to identify cases in subsections of their populations. A number of Boards are currently providing data predominantly for patients seen in secondary care. The very low percentage of cases in Greater Glasgow reflects the fact that these data are based on three Local Health Care Co-operative areas; figures for Grampian exclude significant areas of the region; and Highland data includes only patients seen in secondary care. The Scottish Diabetes Survey is intended to be a survey of NHS Board populations, as each board is responsible for the diabetes care of all patients in their areas. Results are therefore presented based for populations. However, it is anticipated that over time ascertainment will improve and the survey will more accurately reflect population prevalence.
The proportion of the population identified as having diabetes in the Survey ranged from 0.46% to 3.53% in different health board areas. By standardising for age the effects of different age structures in the populations of different health board areas can be taken into account (figure 3). This still shows considerable variation in proportion of the population recorded as having diabetes throughout the country. Most of this variation is likely to be due to differences in case ascertainment and the maturity of the register in each area. These marked variations between areas should reduce over time as registers develop.
Figure 3: Diabetes register: age-standardised prevalence by NHS Board
Note: Glasgow - prevalence has been calculated using the population of the 3 LHCC's
Differences in the completeness of local registers inevitably impacts on the registered prevalence for Scotland as a whole. As the largest board in Scotland, the impact of Greater Glasgow is particularly marked. The figures submitted by Greater Glasgow for the 2002 Survey only include data from three LHCC's in South Glasgow, which together provide care to a little over a fifth of the Glasgow total (192,374 out of 904,400). These data have been collected in the course of implementation of the Glasgow Diabetes Project, and therefore the data currently available reflects the current stage of the project. The project is now being extended to the North of Glasgow and data will be collected from practices in the north for the first time. It is anticipated that the submission in 2003 will include data for the south and a large part of the north of Glasgow. The registered prevalence in South Glasgow 2.18%. However, adding in the population for the rest of Glasgow reduces the registered prevalence to only 0.46%. If the figures for South Glasgow were reflected across the whole board area the registered prevalence for Scotland would be 2.36%.