although the majority of diabetic patients in United Kingdom are treated in primary care settings, data about this disease are mainly based on groups of patient selected from those in hospital care.it has been suggested that mortality statistics on diabetes are frequently biased by factors including referral bias in hospital populations.