5. Calculation Of The Rates Of Diagnosis
With data sources identified and the data collected from the three participating Health Boards, work could commence on developing an estimated rate of diagnosis for dementia.
Firstly, data from all sources were combined together to form one file that held individual level data from both the local systems as well as secondary care data. The first instance of dementia was then identified for each individual, utilising CHI number, by selecting the earliest diagnosis date (or discharge date in the case of secondary care data).
The data was then aggregated to count the number of diagnoses per calendar year into 5 year age bands. The aggregations for the year 2014 were then used to produce an estimated rate of diagnosis for each age band. Given dementia is predominately diagnosed within the older population, age-specific rates were most appropriate at illustrating where in life new diagnoses are most common. These age-specific rates were calculated as follows:-
This calculation was applied to each specific NHS Board's diagnosis figures and population estimates, and the resulting age-specific rates per 1,000 population (rate x 1000) were as follows:-
Table 4. Age-Specific Rates for Newly Diagnosed Individuals with Dementia per 1,000 population.
|NHS Board||Under 60||60-64||65-69||70-74||75 -79||80-84||85-89||90+|
|NHS Ayrshire & Arran||0.1||0.9||2.1||6.4||14.3||28.7||43.9||49.9|
|NHS Greater Glasgow & Clyde||0.1||1.1||3.4||8.3||18.0||32.4||48.8||54.0|
A combined rate was derived to apply to the remaining 11 Health Boards in the absence of their local data. This was calculated by summing all three participating Health Boards diagnoses and dividing by their combined population. The combined rate was to act as a proxy-rate to account for differences in diagnosis numbers and population sizes between the remaining Health Boards.
The actual age-specific rates calculated for NHS Ayrshire & Arran, NHS Greater Glasgow & Clyde and NHS Lanarkshire were used for their own projections.
The rates of diagnosis were then applied to projected population estimates to produce projected numbers of newly diagnosed dementia cases for each year up until 2020.
The projected population estimates between 2014 and 2020 were taken from the 2012-base population projections for Scotland derived from National Records Scotland ( NRS) and can be accessed via their website  . Their web pages also detail important findings from the projections, such as areas where populations are projected to increase or decrease over time  .
Confidence intervals were also calculated using the Byar's Method to demonstrate the range of values wherein there is reasonable certainty the true incidence value lies.
Listed below are some points for consideration when interpreting the outputs that resulted from this methodology. As with most projects there will be limitations and it is important to recognise these when transforming outputs into information and action.
1. This report provides an estimate of the number of people newly diagnosed with dementia in 2014. However, it is possible for individuals to be living with dementia for quite some time without any formal diagnosis.
2. These rates have no state of permanence and are subject to change due to ever evolving interventions. People could start getting diagnosed earlier, therefore increasing the rate of diagnosis in younger age groups and decreasing the rate in older age groups. We recommend that this work is frequently re-run to take account of these factors
3. Whilst these results provide a reasonable estimate of the number of individuals newly diagnosed with dementia in a given year, it should be noted that the estimate is based on the service delivery model & processes within the three Health Boards that provided data and it should be considered that service delivery models & processes could vary across Scotland. Future reiterations of this project may consider obtaining more local data sources to use in calculating more site-specific rates.
4. Given the short time frames of the local data sources, it is difficult to fully capture true first incidence. For example, someone may have been formally diagnosed on a date prior to the time period of the data supplied by the Health Boards. Therefore, if this same individual was admitted to hospital in 2014 with dementia recorded as a co-morbidity, this methodology would then conclude 2014 was the first incidence for that individual.