30 Oct

What type of information does the CSPRG hold for the purposes of the IA and AA studies?

We have collected information on about 260,000 men and women from sources in the UK and abroad for the purposes of this study. We have used information from people who were investigated by colonoscopy because they had a positive test in a bowel cancer screening initiative, or because they had been referred by their doctor for investigation of their bowel symptoms. Below are details of where and for which periods of time we have collected this information, or links to where this information can be found.

UK hospitals

17 hospitals in the UK (see below) for colonoscopy procedures performed between 1972 and 2010. These hospitals were the source of most of the information we have used for the IA study. From these hospitals we have data on approximately 250,000 men and women. The hospital data collected for the IA study is de-identified, and a file that links our unique study number to the patient identifiers is held at the individual hospital where the data was collected from. No identifiers are held by the research group at CSPRG. The only personal information that we hold is full date of birth for the purpose of ensuring data integrity so that we can validate or query discrepancies in any follow-up data that we receive for these patients from other data sources. We do not hold the names or addresses for anyone whose information was used for this study and it will not be possible to identify individuals from the published results of this study. The clinical data collected consists of endoscopy reports, pathology reports, bowel cancers and deaths data. During analysis of our results we realised that, as the coding and analysis of data for this study took significantly longer than we anticipated, there is a gap in time between the period for which we collected data on bowel cancers and deaths from NHS Digital and the study period for which we collected data on patients procedures from the hospitals involved in the study. Therefore we need to obtain an update on the procedures that occurred during the more recent missing time period not covered by the original data collection.

The 17 UK hospitals where data was collected from and approximate date range of data collected:

  • Royal Sussex County Hospital, Brighton and Sussex University Hospitals NHS Trust, May 2001 to April 2008
  • St Mark’s Hospital, London North West Healthcare NHS Trust, January 1972 to July 2007
  • Glasgow Royal Infirmary, Greater Glasgow and Clyde NHS Trust, April 1996 to August 2009
  • Charing Cross Hospital and Hammersmith Hospital, Imperial College Healthcare NHS Trust, October 1997 to November 2007
  • St Mary’s Hospital, Imperial College Healthcare NHS Trust, January 1985 to July 2010
  • Cumberland Infirmary, North Cumbria Acute Hospitals Trust, October 1998 to September 2009
  • Queen Elizabeth Hospital, Queen Elizabeth Hospital NHS Trust, October 1999 to May 2006
  • Queen Mary’s Hospital, Queen Mary’s Sidcup NHS Trust, October 1988 to July 2009
  • Royal Liverpool University Hospital, Royal Liverpool and Broadgreen University Hospitals Trust, January 2000 to September 2009
  • Royal Surrey County Hospital, Royal Surrey County Hospital NHS Trust, September 1997 to May 2010
  • New Cross Hospital, Royal Wolverhampton Hospitals NHS Trust, January 1993 to November 2007
  • Royal Shrewsbury Hospital, Shrewsbury and Telford Hospitals NHS Trust, December 2001 to September 2009
  • Torbay District General Hospital, South Devon Healthcare NHS Foundation Trust, October 2000 to August 2007
  • St George’s Hospital, St George’s Healthcare NHS Trust, February 1992 to July 2009
  • University Hospital of North Tees, University Hospital of North Tees Trust, June 1986 to December 2006
  • Leicester General Hospital, University Hospitals of Leicester NHS Trust, April 1998 to March 2008
  • Yeovil District Hospital, Yeovil District Hospital Foundation Trust, February 1997 to May 2008

The AA study will re-use the data from the IA study ‘Hospital dataset’ and update the follow-up period to the end of December 2016, where possible, to give 10 years of follow-up data. To help us understand the cost implications, we will also ask for information on the type of bowel cancers that have occurred in this group of patients. This new information will be requested from Public Health England and the NHS National Services Scotland.

UKFSST

The UKFSST screening trial includes data on 2,051 people who were randomised, attended their screening appointment and had a colonoscopy at baseline. We used clinical data (including endoscopy, pathology, cancers and deaths) on these people.

NHS Bowel Cancer Screening Programme

The English Bowel Cancer Screening Programme pilot dataset includes 943 men and women who were aged 50 to 69 years old between 2000 and 2002 and who were referred for follow-up following colonoscopy for a positive Faecal Occult Blood Test (FOBt). We used the data from 3 pilot centres that collected data from follow-up colonoscopies. We obtained cancer and mortality data for these individuals with the help of NHS Digital. The only personal data we now hold is full date of birth.

Kaiser Permanente

The Kaiser Permanente Colorectal Cancer Prevention Program provided data on 2,203 men and women in the United States who attended bowel cancer screening from 1994 and were entered into a surveillance programme. This data is fully anonymised and we hold only clinical data and cancer and mortality data.