Page tree
Skip to end of metadata
Go to start of metadata

This wiki is maintained as a guide for analysts who are seeking to improve patient outcomes through better use of data for clinical care, audit and research at the Edinburgh Cancer Centre and across Scotland. If you would like to contribute please get in touch with Peter Hall p.s.hall@ed.ac.uk. 

Contents

Please Note

This site is currently under construction!



Edinburgh Cancer Centre Information Programme

The Edinburgh Cancer Centre provides comprehensive cancer care for South East Scotland, serving patients within the 4 regional Health Board – NHS Lothian, NHS Fife, NHS Dumfries and Galloway and NHS Borders.  In 2014 the Edinburgh Cancer Centre began a Cancer Information Modernisation Programme. The programme focuses on obtaining improved access to cancer patient data to support service planning, delivery of better quality care and improved outcomes for patients.  The ultimate aim being to deliver a world class resource for use in collaboration with national analytical partners, researchers and Scottish Government’s Innovative Healthcare Delivery Programme.

Serving a population of 1.4 million, the Edinburgh Cancer Centre has for 40 years maintained a clinical process and outcomes database that integrates with regional electronic health systems and describes the complete cancer journey. This ability to provide a comprehensive description of the clinical pathway and cancer journey is essential for robust HTA. By linking this data with modern prescribing, treatment and diagnostic databases, wider health outcomes, healthcare activity and costs can now be measured in addition to the various necessary clinical outcomes measures. The Edinburgh Cancer Information Modernisation programme is based on an update of this legacy database, using it as the foundation for a fully integrated system. 

The programme is facilitated by close working with the Cancer Research UK Edinburgh Cancer Research Centre which has international expertise in Bioinformatics, Genetics and Molecular Biology applied to cancer. 


What is Cancer Informatics?

Information technology plays a key role in fighting cancer through a growing field called cancer informatics.

Cancer informatics is where information science, computer science and health care intersect. It’s about acquiring, storing and using information about cancer most thoroughly and efficiently.

Its tools include computers, clinical guidelines, and information and communication systems. It uses information about patient physical, psychological and social characteristics ('phenotype') and their molecular characteristics ('genotype'). These characteristics are studied in relation to how patients interact with the health system and how patients are affected ('outcomes').

A key feature is the synthesizing of data about cancer in ways such as analyzing captured data, gathering evidence from clinical trials, applying information into clinical practice, evaluating outcomes of changes in clinical practice, and generating new hypotheses for investigation.

The Edinburgh Cancer Informatics Programme is striving to use methods in cancer informatics to implement a “rapid-learning system for cancer care.” The “rapid-learning’ comes from the healthcare system being able to keep itself constantly informed on how it is affecting the patients it cares for.

Cancer informatics is a broad field, getting deeper and more complex as medical research and information technology progress. The field is all about collaboration, not just among physicians and researchers but between doctors and health IT professionals and patients and the public. 

Return to Top

Patient perspectives

Making better use of NHS data provides huge opportunities to benefit patients and the public, whilst respecting patient confidentiality. There is a strong and united patient voice in the UK guiding the use of data for this purpose.  #usemydata provides some useful insight. The safe use of data in Scotland is governed by the Caldicott Guardians, operating through the Public Benefits and Privacy Panel.



Analysing and presenting data

Defining a cohort

New diagnoses of cancer are recorded in the Scottish Cancer Registry (SMR06). Cancer statistics are usually reported over the range of ICD-10 codes that define invasive cancer C00-C96. It is common to exclude C44 (non-melanoma skin cancer) because it is non-life threatening, it is poorly or inconsistently recorded and there are large numbers of cases. The registry website contains more information http://www.isdscotland.org/Health-Topics/Cancer/FAQ . 

Patient and population characteristics (phenotype and genotype)

Defining the characteristics of patients with breast cancer using the Scottish Morbidity Record

  • Demographics: Age, gender, Deprivation (Scottish Index of Multiple Deprivation; SIMD), 
  • Geography: Rurality, Health Board of Residence, Health Board of Diagnosis, Health Board of Treatment 
  • Comorbidity: Charlson score (continuous), Charlson variables (17; binary)
  • Stage: AJCC Stage, TNM, Tsize, Nnumber, grade
  • Molecular subgroup: ER, HER2
  • Treatment: Surgery, Chemo, RT, Hormone therapy, Biological/other therapy

Deprivation

A project conducted by ISD and funded by Macmillan in 2017 provides an excellent description of methods suitable for the characterisation of patient populations by SIMD. It presents an overview of methods for studying the association between survival and deprivation. See the project page published by Macmillan for full details.

By default for cancer analysis, we are using the ISD preferred method for ranking and categorising by Scottish quantiles.

Staging

T and N stage can be clinical (assessed by clinical or radiological evaluation) or pathologically (by microscopic assessment following surgical excision). Clinical staging follows universal definitions across all diagnosed cancer following the manual produced by the American Joint Committee on Cancer

Breast Cancer Staging

Outcomes

Toxicity

Toxicity data collected in clinical trials conforms to the Common Terminology for Adverse Events (CTCAE) definitions. In routine care, toxicities are graded along the same scale of 0 to 5. Clinical teams attempt to align their toxicity reporting as closely as possible with CTCAE, but there may be some drift from these precise definitions. Within South East Scotland, patients undergoing Systemic Anti-Cancer Therapy are assessed for toxicity prior to each cycle of treatment and the graded toxicities are recorded in coded form in ChemoCare.

Disease-specific survival

Defined as: The percentage of people in a study or treatment group who have not died from a specific disease in a defined period of time. The time period usually begins at the time of diagnosis or at the start of treatment and ends at the time of death. Patients who died from causes other than the disease being studied are not counted in this measurement.

Assignment of cause of death is complex and guided by ICD-10 standards: http://apps.who.int/classifications/icd10/browse/Content/statichtml/ICD10Volume2_en_2010.pdf

Consideration needs to be given to competing risks in cohort or research studies (candidate methods include Fine and Gray or Cox subdistribution models). For epidemiological interpretation, consider relative survival or net survival.





Datasets

Regional Datasets

The South East Scotland (SES) Cancer Database

The South East Scotland Oncology Database has been populated and maintained by the Department of Oncology at ECC since it was established in 1974.  The database contains a broad spectrum of data items across various stages of the cancer patient journey including diagnosis, treatment, follow-up and recurrence.  It is the richest single source of cancer data in NHS Lothian and a unique facility within South East Scotland. In 2019 the database underwent modernisation, becoming the South East Scotland Cancer Database.

A data dictionary and coding manual is available for the version of the database used until 2017

The up to date data dictionary is available on this wiki: The South East Scotland (SES) Cancer Database

TrakCare

TrakCare (InterSystems) is the nationally adopted electronic health record. Structured as a Patient Administration System, it provides epidose-structured data describing inpatient, outpatient and daycase attendances as well as radiology procedures. Within NHS Lothian it hosts the module used for cancer data entry by multi-disciplinary teams. Hospital coding (ICD-10 diagnostic codes and OPCS procedure codes) takes place within TrakCare. Radiology and pathology reports are also contained within TrakCare as unstructured or semi-structured text.

ChemoCare (SACT)

Chemotherapy prescribing system has been in use since 2006. ChemoCare has been adopted as the standard SACT prescribing system in all Scottish Health Boards. An ongoing programme of work is harmonising the data definition across Scotland. A single ChemoCare instance is used across the SCAN Health Boards.

ARIA (Radiotherapy)

Radiotherapy system is the point of care software used for radiotherapy planning and delivery. Until recently data from ARIA was exported to Public Health England as part of the NATCANSAT database. This has now stopped and a regional solution for bespoke and routine data extraction is in development within the Edinburgh Cancer Centre, which provides radiotherapy services for the whole of SCAN. The Edinburgh Cancer Centre received a data dump from NatCanSAT that contains an RTDS compatible dataset from April 2012 - October 2016.

Bespoke audit databases

There are a large number of legacy and actively maintained databases designed with specific purposes in mind. An example is that maintained by the Secondary Breast Cancer Clinical Nurse Specialists service within NHS Lothian. This records the demographic and clinical details of all patients seen with secondary (stage IV) breast cancer. Similar databases are maintained by other specialist services, such as the Cancer of Unknown Primary service at the Edinburgh Cancer Centre. The majority of the data bases are simple Excel spreadsheets.

EDGE (Clincial Trials)

Is the UK national clinical trials patient database. This contains patient detains with regards to screening, approaches, participation, randomisation and withdrawal for all clinical trials operated within the Scottish Cancer Research Networks.

Edge Website.

Patient Level Information and Costing System (PLICS)

NHS Lothian is the first Health Board in Scotland to implement its own PLICs system. PLICS uses a hybrid costing method, adhering to standardised methods to assign costs to the individual components of care episodes. Example components are nursing time, medicines cost and surgical theatre time.

Clinical Genetics

Held in isolation within NHS Lothian, the Clinical Genetics Service holds information on germline genetic testing (such as BRCA gene mutations) for local patients as well as providing testing on a national basis for specific diseases.

National Datasets

The Scottish Cancer Registry (SMR06) 

SOCRATES, the main database for The Scottish Cancer Registry and has been collecting information since 1958. 

For more info: http://www.isdscotland.org/Health-Topics/Cancer/Scottish-Cancer-Registry/

ICDO Codes

International classification of diseases for oncology (ICD-O) codes are recorded in SMR06 with codes for both morphology and site.

Tumour type indicates the morphology (histology) of the tumour and comprises of the first four digits of the ICD-O morphology code and a fifth digit which denotes the behaviour of the tumour.

CodeDescription
0Benign
1Uncertain whether benign or malignant
2Carcinoma in situ
3Malignant - primary site
6Malignant - metastatic site or secondary site
9

Malignant - uncertain whether primary or metastatic site

Diagnosis on the same day

It is possible for patients to be diagnosed with multiple primary tumours on the same day; this must be accounted for when looking at patient history as these tumours can potentially be flagged as tumour history.

Tertiary referral centres and regional analysis

SMR06 identified Health Board of diagnosis separately to Health Board of residence. In an example cohort derived for breast cancer, roughly 36% of patients in SMR06 diagnosed within NHS Lothian each year are non-Lothian residents. 


The Scottish Morbidity Record

See the SMR Data Manual on the ISD website.

SMR01 The General/Acute and Inpatient Day Case dataset collects episode level data on hospital inpatient and day case discharges from hospitals in Scotland. 

Data dictionary: http://www.ndc.scot.nhs.uk/Data-Dictionary/SMR-Datasets//SMR01-General-Acute-Inpatient-and-Day-Case/

SMR99 (NRS Mortality): National Records of Scotland (NRS) classifies the underlying cause of death and up to 7 contributing causes (in terms of the International Statistical Classification of Diseases and Related Health Problems) on the basis of the information collected on the death certificate together with any additional information provided by official sources, such as the doctor who certified the death and pathologists. 

SMR00: The Outpatients dataset collects episode level data from patients on new appointments at outpatient clinics in all specialties (not A&E).  

A&E Data: Emergency Departments and departments that provide Accident and Emergency (A&E) services submit data to ISD through the A&E datamart. 

Social care (SOURCE): ***In development***


Prescribing (PIS): The Prescribing Information System (PIS) holds details of prescription items dispensed in the community, prescription details, dispenser details remuneration of drug items and patient CHI number.  

SCI Store: SCI Store is an information repository for a National Laboratory Results data mart to hold clinical biochemistry data (e.g. Patient demographics, Laboratory investigation reports, Radiology reports, etc.)

NHS 24:

Scottish Ambulance Service:

GP Out of Hours:

Patient Level Information Costs (PLICs): produced to support the strategic planning of health and social care integration authorities. Data pulled in from Costbook, SMR01, SMR02, SMR04, SMR50, A&E, PIS, NRS Deaths. Cancer-specific costings are not yet included within this project.

Quality Performance Indicators (QPIs) /Cancer Audit

Quality Performance Indicators (QPIs) /Cancer Audit: The Scottish Cancer Taskforce established the National Cancer Quality Steering Group (NCQSG) in 2008 which includes responsibility for the development of small sets (approximately 10-15 indicators) of tumour specific national quality performance indicators (QPIs). The QPIs have been developed collaboratively with the three Regional Cancer Networks (NOSCAN, SCAN, WOSCAN), Information Services Division (ISD), and Healthcare Improvement Scotland.

QPI data dictionary

Breast screening:  


ICDO Codes

International classification of diseases for oncology (ICD-O) codes are recorded in SMR06 with codes for both morphology and site.

Tumour type indicates the morphology (histology) of the tumour and comprises of the first four digits of the ICD-O morphology code and a fifth digit which denotes the behaviour of the tumour.

Code

Description

0

Benign

1

Uncertain whether benign or malignant

2

Carcinoma in situ

3

Malignant - primary site

6

Malignant - metastatic site or secondary site

9

Malignant - uncertain whether primary or metastatic site

However when analysing the fifth digit of the morphology code for the C50* tumours the only result was 3 – malignant – primary site.

Diagnosis on the same day

It is possible for patients to be diagnosed with multiple primary tumours on the same day; this must be accounted for when looking at patient history as these tumours can potentially be flagged as tumour history.

Non Lothian Residents

Roughly 36% of patients in SMR06 each year are non-Lothian residents. These patients have been removed from the analysis as the data for their full history lies with the health board of residence which at the moment there is no access to.

To see a breakdown by treatment type please see the summary table.

Example projects

Methods for causal inference

SATURNE

SATURNE aims to address whether chemotherapy is effective in patients with characteristics out-with the eligibility criteria of historical clinical trials. It also aims to find out whether online decision tools (Adjuvant! Online and NHS predict) are valid and reliable tools for aiding chemotherapy decisions in Scottish patients. SATURNE also presents an opportunity to test the feasibility and validity of regression discontinuity design (RD), a new data analysis method from econometrics that may allow us to use existing healthcare records to estimate the effectiveness of chemotherapy in situations where randomised trials are not possible. More info.

Routine use of Patient Reported Outcomes (PROs)

The Scottish Cancer Innovation Challenge is looking at New approaches to record and integrate cancer PROMs (Patient Reported Outcome Measures) and PREMs (Patient Reported Experience Measures). CMOP looked at prostate cancer and melanoma. It compared case ascertainment of castrate resistant prostate cancer using the routinely capture NHS data against manual scrutiny of records by the research team.  More info

Medicines outcomes capture

CMOP

The CMOP programme is establishing effectiveness of cancer medicines and exploring collection of patient reported outcome measures (PROMs).

Return to Top

External References

Write a comment…