Among hospitalised adults registered with COVID-19 there was a higher proportion of patients with various underlying conditions than in the population as a whole. The clinical picture changes with age. We found a tendency towards a higher proportion registered with non-communicable diseases such as cardiovascular diseases, cancer, type 2 diabetes and COPD than in the general population in many age groups, among both women and men. There was little difference in the prevalence of registered conditions between persons who had tested positive for COVID-19 and the general population.
Currently there is little research evidence as to whether persons with specific underlying diseases and conditions are at a higher risk of a serious course of COVID-19 (7). In most previous studies, the analyses are based on population groups where there might be a selection to be tested for SARS-CoV-2 and to be hospitalised with COVID-19 (8, 9). This might influence the results (7, 10).
Advanced age alone, or the presence of multiple diseases, is associated with a severe course of COVID-19 (7). However, it is difficult to determine whether the increased risk of developing severe disease after COVID-19 is due to the presence of multiple diseases or whether it is a reflection of the fact that older persons with these diseases are often hospitalised as a precautionary measure.
It is difficult to distinguish between these factors in our project as well. Being hospitalised for COVID-19 may in itself be an expression of the disease severity. We did, however, observe that even though the length of the hospitalisation period may reflect the severity of the COVID-19 infection, some patients also died during or shortly after a short stay in hospital. It will thus be difficult to rate the illness severity among hospitalised patients exclusively based on the length of their stay or use of intensive care.
Analyses of individual diseases have pointed out that cardiovascular diseases, cancer, diabetes and COPD are overrepresented among persons who developed severe disease after COVID-19, although the findings are not entirely unambiguous across the studies (4). For example, in a study by Petrilli et al., chronic pulmonary diseases are not overrepresented among hospitalised patients (11). For other diseases, such as diabetes, the findings are more consistent across studies and well as with our own findings. A Danish non-peer reviewed study by Riley et al. shows that the proportion with diabetes is higher among adults with a serious course of COVID-19 (10). We have little knowledge of how the diseases are distributed by age and sex among hospitalised patients when compared to the general population. Our data therefore have value despite the small and non-random nature of the data material, which needs to be interpreted with caution.
The proportion of hospitalised persons with multiple diseases such as cardiovascular diseases, cancer, type 2 diabetes and COPD increases with age and can explain part of the difference between age groups. It is interesting to note that among the oldest – 80 years and older – we find that compared to the general population, a higher proportion of those who tested positive to COVID-19 are registered with dementia, while the proportion with dementia among patients hospitalised with COVID-19 is not higher than in the general population. This may partly be due to the Norwegian testing procedures, whereby people who are at higher risk, including elderly nursing home residents with dementia, may have been tested more frequently, but perhaps not hospitalised to the same extent.
The strength of this registry study lies in its inclusion of the entire Norwegian population aged 20 years and older. This means that we have shown how the proportion of the population with the large disease groups of cardiovascular diseases, cancer, type 2 diabetes and COPD are distributed by sex and age among persons who have tested positive or been hospitalised for COVID-19. This is important when assessing the potential effects of pandemics that can affect persons in these large disease groups. We included separate analyses for cancer patients who were undergoing active treatment, since we can assume that they will be more vulnerable than other cancer patients. The proportion of patients undergoing treatment for cancer was also higher among those hospitalised for COVID-19 than in the general population.
A registry study will include non-validated data. Here, this does not apply to the definition of the patient group, since one of the strengths of our study is the high degree of correspondence between the Norwegian Surveillance System for Communicable Diseases and the coding in the Norwegian Patient Registry for hospitalised patients. As regards all the listed diseases, on the other hand, non-validated data will be included. Validation of data from the Norwegian Patient Registry may, for example, show that 'suspicion of a disease' has been coded as 'the disease'. Furthermore, we have insufficient knowledge about the quality of the Norwegian Registry for Primary Health Care, since it was established relatively recently. Overestimation of disease prevalence is generally likely in registry studies without validation, and this is a weakness. Unfortunately, the project had no data on multimorbidity for all the diseases. This means that an individual can be affected by more than one disease, in the same way as the general population with which the hospitalised patients are compared.
Follow-up times vary between the different registries. The Norwegian Patient Registry contains information on completed hospital care episodes, and consequently we have no data on those who were hospitalised in late April/early May. Access to data on currently hospitalised patients could have further strengthened the study. This group is likely to have encompassed a proportion with long hospitalisation periods and a more serious disease course may assume, however, that this weakness of the study will have little effect on the results, since this group is small. The regulations stipulate that the code for COVID-19 infection should not be used as a main diagnosis. In the hospital data, it is thus impossible to distinguish between patients who have been admitted for COVID-19 and patients who have been diagnosed with COVID-19 during hospitalisation for other conditions.
Because of the Norwegian testing strategies in the spring of 2020, the likelihood of being tested was higher for a person with an underlying disease than for an otherwise healthy person with the same symptoms. If many people have had COVID-19 without being tested, this could weaken the findings in our study. The majority of these would then most likely be in the group of 'otherwise healthy' persons, and we would have expected an overrepresentation of chronically ill patients among those who had tested positive. This seems however not to be the case.
With access to data about those who tested negative for COVID-19 we could have estimated the degree of selection caused by the testing strategies. Many of those who were first confirmed to have COVID-19 were fit and healthy ski tourists, and this may also have served to reduce the proportion of chronically ill persons among those who tested positive. A person with an underlying disease will also tend to be hospitalised with milder symptoms than an otherwise healthy person. We know that during the influenza pandemic caused by the A(H1 N1) virus in 2009–2010, the likelihood of hospitalisation was higher among persons with type 2 diabetes than among the general population, while the disease severity and complications was not any higher (12). The weaknesses of observation studies that use non-random material, as in parts of this study, and how this affects the results have been described in a number of recently published articles (8, 9, 13).
The objective of this study was to describe how different diseases were distributed among persons with confirmed COVID-19 and among patients hospitalised for COVID-19 when compared to equivalent figures for the general population. The main purpose has been to contribute to monitoring and emergency preparedness. Having such an overview is important in order to follow the development in Norway further. Linking registry data with personal data that include all persons tested for COVID-19 and where the group that tested positive to COVID-19 is not selected, could have helped answer questions about the risk of being infected with or hospitalised for COVID-19, and thus provided more evidence about the underlying causes of the development of severe disease.