Health care use attributable to COVID-19: A propensity matched national electronic health records cohort study of 249,390 people in Wales, UK

  1. J Kennedy
  2. M Parker
  3. M Seaborne
  4. M Mhereeg
  5. A Walker
  6. V Walker
  7. S Denaxas
  8. N Kennedy
  9. S.V Katikireddi
  10. S Brophy

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

Background

To determine the extent and nature of changes in infected patients healthcare utilization, we studied healthcare contact in the 1-4 weeks and 5-24 weeks following a COVID-19 diagnosis compared to propensity matched controls.

Methods

Survival analysis was used for time to death and first clinical outcomes including clinical terminology concepts for post-viral illness, fatigue, embolism, respiratory conditions, mental and developmental conditions, fit note, or hospital attendance. Increased instantaneous risk for the occurrence of an outcome for positive individuals was quantified using hazard ratios (HR) from Cox Regression and absolute risk was quantified using relative risk (RR) from life table analysis.

Results

Compared to matched individuals testing negative, surviving positive community-tested patients had a higher risk of post-viral illness (HR: 4.57, 95%CI: 1.77-11.80, p=0.002), fatigue (HR: 1.47, 95%CI: 1.24-1.75, p<0.001) and embolism (HR: 1.51, 95%CI: 1.13-2.02, p=0.005) at 5-24 weeks post-diagnosis. In the four weeks after COVID-19 higher rates of sick notes were being issued for community-tested (HR: 3.04, 95%CI: 0.88 to 10.50, p<0.079); the risk was reduced after four weeks, compared to controls. Overall healthcare attendance for anxiety, depression was less likely in those with COVID-19 in the first four weeks (HR: 0.83, 95%CI: 0.73-1.06, p=0.007). After four weeks, anxiety, depression is less likely to occur for the positive community-tested individuals (HR: 0.87, 95%CI: 0.77-1.00, p=0.048), but more likely for positive hospital-tested individuals (HR: 1.16, 95%CI: 1.00-1.45, p=0.053). Although statistical associations between positive infection and post-infection healthcare use are clear, the absolute use of healthcare is very.

Conclusions

Community COVID-19 disease is associated with increased risks of post-viral illness, fatigue, embolism, depression, anxiety and respiratory conditions. Despite these elevated risks, the absolute healthcare burden is low. Either very small proportions of people experience adverse outcomes following COVID-19 or they are not presenting to healthcare.

Trial registration

Data held in SAIL databank are anonymised and therefore, no ethical approval is required. All data in SAIL has the permission from the relevant Caldicott Guardian or Data Protection Officer and SAIL-related projects are required to obtain Information Governance Review Panel (IGRP) approval. The IGRP approval number for this study is 1259.

Article activity feed

  1. ScreenIT

    SciScore for 10.1101/2022.04.21.22274152: (What is this?)

    Please note, not all rigor criteria are appropriate for all manuscripts.

    Table 1: Rigor

    NIH rigor criteria are not applicable to paper type.

    Table 2: Resources

    Software and Algorithms
    SentencesResources
    Final data preparations specific to survival analysis were performed in RStudio 2021.09.0 such as setting reference groups for the Cox proportional hazard models (12).
    RStudio
    suggested: (RStudio, RRID:SCR_000432)
    Risk ratio and confidence intervals calculations were performed in Microsoft Excel (Version 2201.) and hazard ratio plots (figures 5 - 8) were also manually constructed in Microsoft Excel.
    Microsoft Excel
    suggested: (Microsoft Excel, RRID:SCR_016137)

    Results from OddPub: We did not detect open data. We also did not detect open code. Researchers are encouraged to share open data when possible (see Nature blog).

    Results from LimitationRecognizer: We detected the following sentences addressing limitations in the study:
    Strengths and Limitations: This study has several strengths, it utilises an entire country’s (Wales) primary and secondary care data and so gives a national perspective of outcomes of COVID-19, making the work generalisable as it is a total population cohort. The use of propensity matching has the advantage of adjusting for many variables and so adjusting for propensity to contact COVID and account for a rich set of covariates which are associated with infection risk, so giving robust findings of association outcomes with surviving COVID. In addition, we were able to match and so control for differences between those tested in the community compared to when they attend a hospital. However, the matching did reduce the sample size of COVID patients from 249,390 to 98,600 which means a loss of 60% of COVID cases who did not have a match, this would result in a loss of precision for detecting rare events. Some limitations of this study include, the study only investigates the first occurrence and does not reflect total burden or duration of an existing problem for example, this study showed higher levels of fatigue in those with COVID, but it did not show how long this fatigue lasts for as the analysis gives a time to first mention of a fatigue diagnosis. This study examines engagement with healthcare and so can reflect use and burden to the health care system due to COVID specifically. However, it cannot capture the unmet need of people who have morbidity associated with COVID b...

    Results from TrialIdentifier: No clinical trial numbers were referenced.

    Results from Barzooka: We did not find any issues relating to the usage of bar graphs.

    Results from JetFighter: We did not find any issues relating to colormaps.

    Results from rtransparent:
    • Thank you for including a conflict of interest statement. Authors are encouraged to include this statement when submitting to a journal.
    • Thank you for including a funding statement. Authors are encouraged to include this statement when submitting to a journal.
    • No protocol registration statement was detected.

    Results from scite Reference Check: We found no unreliable references.

    About SciScore

    SciScore is an automated tool that is designed to assist expert reviewers by finding and presenting formulaic information scattered throughout a paper in a standard, easy to digest format. SciScore checks for the presence and correctness of RRIDs (research resource identifiers), and for rigor criteria such as sex and investigator blinding. For details on the theoretical underpinning of rigor criteria and the tools shown here, including references cited, please follow this link.

  2. Version published to 10.1101/2022.04.21.22274152v1 on medRxiv