Background -- The objectives of this study were to identify risk factors for severe COVID-19 and to lay the basis for risk stratification based on demographic data and health records. Methods and Findings -- The design was a matched case-control study. Severe COVID-19 was defined as either a positive nucleic acid test for SARS-CoV-2 in the national database followed by entry to a critical care unit or death MESHD within 28 days, or a death MESHD certificate with COVID-19 as underlying cause. Up to ten controls per case matched for sex, age TRANS and primary care practice were selected from the population register. All diagnostic codes from the past five years of hospitalisation records and all drug codes from prescriptions dispensed during the past nine months were extracted. Rate ratios for severe COVID-19 were estimated by conditional logistic regression. There were 4272 severe cases. In a logistic regression using the age TRANS-sex distribution of the national population, the odds ratios for severe disease MESHD were 2.87 for a 10-year increase in age TRANS and 1.63 for male TRANS sex. In the case-control analysis, the strongest risk factor was residence in a care home, with rate ratio (95% CI) 21.4 (19.1, 23.9). Univariate rate ratios (95% CIs) for conditions listed by public health agencies as conferring high risk were 2.75 (1.96, 3.88) for Type 1 diabetes, 1.60 (1.48, 1.74) for Type 2 diabetes, 1.49 (1.37, 1.61) for ischemic heart disease MESHD, 2.23 (2.08, 2.39) for other heart disease MESHD, 1.96 (1.83, 2.10) for chronic lower respiratory tract disease MESHD, 4.06 (3.15, 5.23) for chronic kidney disease HP kidney disease MESHD, 5.4 (4.9, 5.8) for neurological disease MESHD, 3.61 (2.60, 5.00) for chronic liver disease MESHD and 2.66 (1.86, 3.79) for immune deficiency or suppression. 78% of cases and 52% of controls had at least one listed condition (NA of cases and NA of controls under age TRANS 40). Severe disease MESHD was associated with encashment of at least one prescription in the past nine months and with at least one hospital admission in the past five years [rate ratios 3.10 (2.59, 3.71)] and 2.75 (2.53, 2.99) respectively] even after adjusting for the listed conditions. In those without listed conditions significant associations with severe disease MESHD were seen across many hospital diagnoses and drug categories. Age TRANS and sex provided 2.58 bits of information for discrimination. A model based on demographic variables, listed conditions, hospital diagnoses and prescriptions provided an additional 1.25 bits (C-statistic 0.825). A limitation of this study is that records from primary care were not available. Conclusions -- Along with older age TRANS and male TRANS sex, severe COVID-19 is strongly associated with past medical history across all age groups TRANS. Many comorbidities beyond the risk conditions designated by public health agencies contribute to this. A risk classifier that uses all the information available in health records, rather than only a limited set of conditions, will more accurately discriminate between low-risk and high-risk individuals who may require shielding until the epidemic is over.