Dataset Information

Prevalence of COVID-19-related risk factors and risk of severe influenza outcomes in cancer survivors: A matched cohort study using linked English electronic health records data.

ABSTRACT:

Background

People with active cancer are recognised as at risk of COVID-19 complications, but it is unclear whether the much larger population of cancer survivors is at elevated risk. We aimed to address this by comparing cancer survivors and cancer-free controls for (i) prevalence of comorbidities considered risk factors for COVID-19; and (ii) risk of severe influenza, as a marker of susceptibility to severe outcomes from epidemic respiratory viruses.

Methods

We included survivors (≥1 year) of the 20 most common cancers, and age, sex and general practice-matched cancer-free controls, derived from English primary care data linked to cancer registrations, hospital admissions and death registrations. Comorbidity prevalences were calculated 1 and 5 years from cancer diagnosis. Risk of hospitalisation or death due to influenza was compared using Cox models adjusted for baseline demographics and comorbidities.

Findings

108,215 cancer survivors and 523,541 cancer-free controls were included. Cancer survivors had more diabetes, asthma, other respiratory, cardiac, neurological, renal, and liver diseases, and less obesity, compared with controls, but there was variation by cancer site. There were 205 influenza hospitalisations/deaths, with cancer survivors at higher risk than controls (adjusted HR 2.78, 95% CI 2.04-3.80). Haematological cancer survivors had large elevated risks persisting for >10 years (HR overall 15.17, 7.84-29.35; HR >10 years from cancer diagnosis 10.06, 2.47-40.93). Survivors of other cancers had evidence of raised risk up to 5 years from cancer diagnosis only (HR >5 years 2.22, 1.31-3.74).

Interpretation

Risks of severe COVID-19 outcomes are likely to be elevated in cancer survivors. This should be taken into account in policies targeted at clinical risk groups, and vaccination for both influenza, and, when available, COVID-19, should be encouraged in cancer survivors.

SUBMITTER: Carreira H

PROVIDER: S-EPMC7788436 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:BackgroundA history of multiple myeloma, prostate cancer, and breast cancer has been associated with adverse bone health, but associations across a broader range of cancers are unclear. We aimed to compare the risk of any bone fracture and major osteoporotic fractures in survivors of a wide range of cancers versus cancer-free individuals.MethodsIn this population-based matched cohort study, we used electronic health records from the UK Clinical Practice Research Datalink linked to hospital data. We included adults (aged ≥18 years) eligible for linkage, and we restricted the study start to Jan 2, 1998, onwards and applied administrative censoring on Jan 31, 2020. The cancer survivor group included survivors of the 20 most common cancers. Each individual with cancer was matched (age, sex, and general practice) to up to five controls (1:5) who were cancer-free. The primary outcomes were any bone fracture and any major osteoporotic fracture (pelvic, hip, wrist, spine, or proximal humeral fractures) occurring more than 1 year after index date (ie, the diagnosis date of the matched individual with cancer). We used Cox regression models, adjusted for shared risk factors, to estimate associations between cancer survivorship and bone fractures.Findings578 160 adults with cancer diagnosed in 1998-2020 were matched to 3 226 404 cancer-free individuals. Crude incidence rates of fractures in cancer survivors ranged between 8·39 cases (95% CI 7·45-9·46) per 1000 person-years for thyroid cancer and 21·62 cases (20·18-23·18) per 1000 person-years for multiple myeloma. Compared with cancer-free individuals, the risk of any bone fracture was increased in 15 of 20 cancers, and of major osteoporotic fractures in 17 of 20 cancers. Effect sizes varied: adjusted hazard ratios (HRs) were largest for multiple myeloma (1·94, 95% CI 1·77-2·13) and prostate cancer (1·43, 1·39-1·47); HRs in the range 1·20-1·50 were seen for stomach, liver, pancreas, lung, breast, kidney, and CNS cancers; smaller associations (HR <1·20) were observed for malignant melanoma, non-Hodgkin lymphoma, leukaemia, and oesophageal, colorectal, and cervical cancers. Increased risks of major osteoporotic fracture were noted most substantially in multiple myeloma (2·25, 1·96-2·58) and CNS (2·12, 1·56-2·87), liver (1·62, 1·01-2·61), prostate (1·60, 1·53-1·67), and lung cancers (1·60, 1·44-1·77). Effect sizes tended to reduce over time since diagnosis but remained elevated for more than 5 years in several cancers, such as multiple myeloma and stomach, lung, breast, prostate, and CNS cancers.InterpretationSurvivors of most types of cancer were at increased risk of bone fracture for several years after cancer, with variation by cancer type. These findings can help to inform mitigation and prevention strategies.FundingWellcome Trust.

Project description:BackgroundCharacterising the size and distribution of the population at risk of severe COVID-19 is vital for effective policy and planning. Older age, and underlying health conditions, are associated with higher risk of death from COVID-19. This study aimed to describe the population at risk of severe COVID-19 due to underlying health conditions across the United Kingdom.MethodsWe used anonymised electronic health records from the Clinical Practice Research Datalink GOLD to estimate the point prevalence on 5 March 2019 of the at-risk population following national guidance. Prevalence for any risk condition and for each individual condition is given overall and stratified by age and region with binomial exact confidence intervals. We repeated the analysis on 5 March 2014 for full regional representation and to describe prevalence of underlying health conditions in pregnancy. We additionally described the population of cancer survivors, and assessed the value of linked secondary care records for ascertaining COVID-19 at-risk status.ResultsOn 5 March 2019, 24.4% of the UK population were at risk due to a record of at least one underlying health condition, including 8.3% of school-aged children, 19.6% of working-aged adults, and 66.2% of individuals aged 70 years or more. 7.1% of the population had multimorbidity. The size of the at-risk population was stable over time comparing 2014 to 2019, despite increases in chronic liver disease and diabetes and decreases in chronic kidney disease and current asthma. Separately, 1.6% of the population had a new diagnosis of cancer in the past 5 y.ConclusionsThe population at risk of severe COVID-19 (defined as either aged ≥70 years, or younger with an underlying health condition) comprises 18.5 million individuals in the UK, including a considerable proportion of school-aged and working-aged individuals. Our national estimates broadly support the use of Global Burden of Disease modelled estimates in other countries. We provide age- and region- stratified prevalence for each condition to support effective modelling of public health interventions and planning of vaccine resource allocation. The high prevalence of health conditions among older age groups suggests that age-targeted vaccination strategies may efficiently target individuals at higher risk of severe COVID-19.

Project description:IntroductionDisease registers and electronic health records are valuable resources for disease surveillance and research but can be limited by variation in data quality over time. Quality may be limited in terms of the accuracy of clinical information, of the internal linkage that supports person-based analysis of most administrative datasets, or by errors in linkage between multiple datasets.ObjectivesBy linking the National Down Syndrome Cytogenetic Register (NDSCR) to Hospital Episode Statistics for England (HES), we aimed to assess the quality of each and establish a consistent approach for analysis of trends in prevalence of Down's syndrome among live births in England.MethodsProbabilistic record linkage of NDSCR to HES for the period 1998-2013 was supported by linkage of babies to mothers within HES. Comparison of prevalence estimates in England were made using NDSCR only, HES data only, and linked data. Capture-recapture analysis and quantitative bias analysis were used to account for potential errors, including false positive diagnostic codes, unrecorded diagnoses, and linkage error.ResultsAnalyses of single-source data indicated increasing live birth prevalence of Down's Syndrome, particularly in the analysis of HES. Linked data indicated a contrastingly stable prevalence of 12.3 (plausible range: 11.6-12.7) cases per 10 000 live births.ConclusionCase ascertainment in NDSCR improved slightly over time, creating a picture of slowly increasing prevalence. The emerging epidemic suggested by HES primarily reflects improving linkage within HES (assignment of unique patient identifiers to hospital episodes). Administrative data are valuable but trends should be interpreted with caution, and with assessment of data quality over time. Data linkage with quantitative bias analysis can provide more robust estimation and, in this case, stronger evidence that prevalence is not increasing. Routine linkage of administrative and register data can enhance the value of each.

Project description:BackgroundEmergency admissions are a major source of healthcare spending. We aimed to derive, validate, and compare conventional and machine learning models for prediction of the first emergency admission. Machine learning methods are capable of capturing complex interactions that are likely to be present when predicting less specific outcomes, such as this one.Methods and findingsWe used longitudinal data from linked electronic health records of 4.6 million patients aged 18-100 years from 389 practices across England between 1985 to 2015. The population was divided into a derivation cohort (80%, 3.75 million patients from 300 general practices) and a validation cohort (20%, 0.88 million patients from 89 general practices) from geographically distinct regions with different risk levels. We first replicated a previously reported Cox proportional hazards (CPH) model for prediction of the risk of the first emergency admission up to 24 months after baseline. This reference model was then compared with 2 machine learning models, random forest (RF) and gradient boosting classifier (GBC). The initial set of predictors for all models included 43 variables, including patient demographics, lifestyle factors, laboratory tests, currently prescribed medications, selected morbidities, and previous emergency admissions. We then added 13 more variables (marital status, prior general practice visits, and 11 additional morbidities), and also enriched all variables by incorporating temporal information whenever possible (e.g., time since first diagnosis). We also varied the prediction windows to 12, 36, 48, and 60 months after baseline and compared model performances. For internal validation, we used 5-fold cross-validation. When the initial set of variables was used, GBC outperformed RF and CPH, with an area under the receiver operating characteristic curve (AUC) of 0.779 (95% CI 0.777, 0.781), compared to 0.752 (95% CI 0.751, 0.753) and 0.740 (95% CI 0.739, 0.741), respectively. In external validation, we observed an AUC of 0.796, 0.736, and 0.736 for GBC, RF, and CPH, respectively. The addition of temporal information improved AUC across all models. In internal validation, the AUC rose to 0.848 (95% CI 0.847, 0.849), 0.825 (95% CI 0.824, 0.826), and 0.805 (95% CI 0.804, 0.806) for GBC, RF, and CPH, respectively, while the AUC in external validation rose to 0.826, 0.810, and 0.788, respectively. This enhancement also resulted in robust predictions for longer time horizons, with AUC values remaining at similar levels across all models. Overall, compared to the baseline reference CPH model, the final GBC model showed a 10.8% higher AUC (0.848 compared to 0.740) for prediction of risk of emergency admission within 24 months. GBC also showed the best calibration throughout the risk spectrum. Despite the wide range of variables included in models, our study was still limited by the number of variables included; inclusion of more variables could have further improved model performances.ConclusionsThe use of machine learning and addition of temporal information led to substantially improved discrimination and calibration for predicting the risk of emergency admission. Model performance remained stable across a range of prediction time windows and when externally validated. These findings support the potential of incorporating machine learning models into electronic health records to inform care and service planning.