People with high blood sugar, known as diabetes, are at risk of many other chronic diseases and dying earlier. It is called pre-diabetes when people have higher than normal sugar levels but not yet achieve diabetes levels, It is believed that the risk of other chronic diseases is accumulating from the pre-diabetes status. In recent decades, researchers also found that diabetes may increase the risk of many types of cancer, for example pancreatic cancer. But for now, there is no cancer screening programme for people with diabetes or pre-diabetes. Therefore, we will examine whether there are more cancer cases from pre-diabetes. It will help find out the best time window to start help people with pre-diabetes or diabetes screen cancer. In addition, we will also examine if cancer diagnosis changes by age, sex and ethnicity to help specific groups.
Using data from Clinical Practice Research Datalink (CPRD) with linkage to Hospital Episodes Statistics (HES) and Office of National Statistics (ONS) Death Registration, we will investigate the trends in cancer incidence and mortality in people with pre-diabetes and diabetes to identify possible cancer screening window. Participants with a first ever HbA1c measure between 6.0%-6.4% or a diagnosis of pre-diabetes and no prescriptions of glucose-lowering drugs and diagnosis codes of any types of diabetes will be extracted as pre-diabetes groups. They will be followed-up from the date of confirmed pre-diabetes to outcomes of interest occurred, the diagnosis of diabetes, death, transfer-out the practice, last data collection date, whichever came first. Participants with a first ever diagnosis code of diabetes will be extracted as the diabetes group, and they will be followed-up from the diagnosis of diabetes to outcomes of interest occurred, death, the latest available of HES linkage, whichever came first. Age-Period-Cohort (APC) analysis discerns three types of time varying phenomena: Age effects, period effects and cohort effects. In our study, APC analysis incorporated with Lexis Diagram Observations will be applied to quantitatively assess the effect of age, calendar year and durations of (pre-)diabetes on the cancer. Trends in cancer incidence and mortality, by all-sites and site-specific, will be modelled with Poisson Regression, rates in people with pre-diabetes and diabetes will be presented by age, durations and calendar year. These will help to identify the time window to implement cancer screening among people diagnosed with diabetes, site-specific cancers as outcomes will provide information on which types of cancer should be screened. Given the high risk of diabetes and diabetes-related complications in South Asians and differences in cancer outcomes by gender, these analyses will be stratified by White and South Asians, and by male and female, to demonstrate the potential inequities among ethnicities and sex.
Health Outcomes to be Measured:
1. All-sites cancer incidence
2. All-sites cancer mortality
3. Site-specific cancer incidence: 13 cancers including pancreatic, liver, gallbladder, female breast, endometrial, thyroid, colorectal, gastric, bladder, kidney, oesophagus and ovarian cancer and multiple myeloma
4. Cancer-specific mortality: 13 cancers including pancreatic, liver, gallbladder, female breast, endometrial, thyroid, colorectal, gastric, bladder, kidney, oesophagus and ovarian cancer and multiple myeloma
HES Admitted;ONS;Patient IMD