SCOTTISH DIABETES CORE DATASET
EXPLANATORY NOTES
The following information explains the various columns within the Core Dataset, how they have been developed, how to use them, how they have changed from the CRAG Dataset (CRAG, September 2000), definitions, and future developments in clinical coding.
Content
The content of the CRAG Shared Care Dataset is largely maintained in this Core Dataset. Many of the field descriptions, parameters and field names are unchanged.
Layout
The Core Dataset retains the layout of the CRAG Dataset in that it has been set out in sections covering demographics, general data, diabetes control, cardiovascular disease, renal disease, feet, eye care, pregnancy, births and young diabetics.
New Read Codes
The development of this Core Dataset has required the NHS Information Authority in Birmingham, who provide support for Read Codes for the UK, to issue a substantial number of new Read Codes. Read Code Updates are produced every six months and the majority of the new codes were included in the March and September 2002 releases. The SCI Diabetes Collaboration Team is working to ensure that up-to-date releases of Read are implemented in diabetes systems supporting this dataset.
Fields without Read Codes
A small number of data items could not be Read coded. The majority of these are in the demographics section, e.g. previous surname, where their absence should have no effect.
Use of synonyms and Term Codes
Read codes have a 'preferred term' based on the wording thought to best describe the concept at the time the code was developed, and may have a number of 'synonymous terms' giving alternative ways of describing the same concept. For example: myocardial infarction (preferred) and heart attack (synonym). As far as possible, preferred terms (term code 00) have been used, but in some instances synonyms are used.
The preferred terminology changes over time such that a synonym may come to better represent prevailing medical language or knowledge. A good example of this is the change in classification of diabetes from Insulin-dependent and Non-insulin dependent diabetes mellitus to Type 1 and Type 2 diabetes mellitus. Read version 2 was developed when IDDM & NIDDM were the preferred terms. The Read classification regards Types 1 & 2 DM as synonyms for IDDM and NIDDM, although this has not necessarily been true in clinical usage. The current WHO classification of diabetes into Type 1 and Type 2 can be specifically coded within Read by tagging a two digit 'term code' onto the 5 digit Read codes for IDDM and NIDDM.
Laterality Codes
A column has been introduced to allow specification of laterality to certain data items where left or right is not specified in the Read code itself. These 'qualifiers' or 'attributes' can be tagged onto the Read code by the IT system.
Coding Comment Column
The coding comment column summarises the string of codes, values and text required to represent each data item. This is essentially for the benefit of system developers. In most circumstances clinical users would not be expected to need to refer to this column.
It is anticipated that IT system developers will probably want to automatically associate a date of entry with all Read Codes although this has not been specifically stated in the coding comments column. However, for many fields, the date of an event or investigation is required for clinical purposes and should be visible to the health care professional. This may not be the same as the date on which the data are entered on the system. In these instances the system must allow the health care professional to enter the appropriate date, and for these fields the coding comment column indicates the need to record a date.
Derived fields
In order to reduce duplicate data entry, a number of derived Fields have been introduced. For example, 'Year of first myocardial infarction' can be derived by the IT system selecting the earliest date from the 'Myocardial Infarction' field in which all MIs are recorded with a date.
Summary of major changes from CRAG dataset
A column is included detailing any changes made to the CRAG Dataset. The main changes are described below.
1. Fields removed. A small number of fields have been removed from the CRAG Dataset: Outcome of appointment; Planned review interval; Planned review date; Date of event; Referred to clinic nurse; Semi-quantitative 10-year CHD risk; and Person who interpreted eye examination.
2. Revised sections. The sections having undergone revision are: feet (amputation, sensation and ulceration fields); eyes (maculopathy, non-diabetic retinal lesions, laser therapy); pregnancy and births (children live & stillborn are now dealt with separately to abortions). The field 'Appointment with defined healthcare professional' has been divided into two: 'referral to defined healthcare professional' and 'seen by defined healthcare professional'.
3. New fields. A number of new fields have been added to the Core Dataset, particularly in the sections on eyes, young people with diabetes and dietary advice.
Definitions
Up-to-date definitions have been specified for data items where appropriate. As far as possible, definitions have been taken from published sources including the Definitions and Codes Manual NHSScotland (ISD, 2002); Report by the Working Group to Support Shared Care in Diabetes (CRAG, 2000); Management of Diabetes guidelines (SIGN, 2001); and Health Technology Assessment Report 1: Organisation of services for diabetic retinopathy screening (Health Technology Board for Scotland, 2002).
SNOMED-Clinical Terms
SNOMED-Clinical Terms (SNOMED-CT) is expected to replace Read Codes as the NHS preferred clinical terminology in the coming years, subject to successful evaluation and Scottish Executive approval. It will create a single unified terminology for use in acute and primary care, facilitating integration of computerised clinical information and is therefore an underpinning feature of the development of Electronic Patient Records and Electronic Health Records.
It is anticipated that some of the problems encountered during the production of this Read Coded dataset will be overcome once SNOMED-CT is introduced. This will be particularly beneficial in coding of the various specialist diabetic dataset modules proposed. The SCI-Diabetes Collaboration will consider migration to SNOMED-CT in due course.