Excess mortality during the COVID-19 pandemic: a geospatial and statistical analysis in Aden governorate, Yemen

  1. Emilie S Koum Besson
  2. Andy Norris
  3. Abdulla S Bin Ghouth
  4. Terri Freemantle
  5. Mervat Alhaffar
  6. Yolanda Vazquez
  7. Chris Reeve
  8. Patrick J Curran
  9. Francesco Checchi

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

The burden of COVID-19 in low-income and conflict-affected countries remains unclear, largely reflecting low testing rates. In parts of Yemen, reports indicated a peak in hospital admissions and burials during May–June 2020. To estimate excess mortality during the epidemic period, we quantified activity across all identifiable cemeteries within Aden governorate (population approximately 1 million) by analysing very high-resolution satellite imagery and compared estimates to Civil Registry office records.

Methods

After identifying active cemeteries through remote and ground information, we applied geospatial analysis techniques to manually identify new grave plots and measure changes in burial surface area over a period from July 2016 to September 2020. After imputing missing grave counts using surface area data, we used alternative approaches, including simple interpolation and a generalised additive mixed growth model, to predict both actual and counterfactual (no epidemic) burial rates by cemetery and across the governorate during the most likely period of COVID-19 excess mortality (from 1 April 2020) and thereby compute excess burials. We also analysed death notifications to the Civil Registry office over the same period.

Results

We collected 78 observations from 11 cemeteries. In all but one, a peak in daily burial rates was evident from April to July 2020. Interpolation and mixed model methods estimated ≈1500 excess burials up to 6 July, and 2120 up to 19 September, corresponding to a peak weekly increase of 230% from the counterfactual. Satellite imagery estimates were generally lower than Civil Registry data, which indicated a peak 1823 deaths in May alone. However, both sources suggested the epidemic had waned by September 2020.

Discussion

To our knowledge, this is the first instance of satellite imagery being used for population mortality estimation. Findings suggest a substantial, under-ascertained impact of COVID-19 in this urban Yemeni governorate and are broadly in line with previous mathematical modelling predictions, though our method cannot distinguish direct from indirect virus deaths. Satellite imagery burial analysis appears a promising novel approach for monitoring epidemics and other crisis impacts, particularly where ground data are difficult to collect.

Article activity feed

  1. Version published to 10.1136/bmjgh-2020-004564
  2. Version published to 10.1101/2020.10.27.20216366v2 on medRxiv
  3. ScreenIT

    SciScore for 10.1101/2020.10.27.20216366: (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

    No key resources detected.

    Results from OddPub: Thank you for sharing your code and data.

    Results from LimitationRecognizer: We detected the following sentences addressing limitations in the study:
    Study limitations: Problems with imagery availability and quality constrained the quantity of cemetery observations for analysis, reducing the accuracy of models and imposing limits on the period of analysis. Furthermore, the inherent inconsistencies (differences in viewing geometry, spatial and spectral resolution) of archive satellite imagery made it difficult to rapidly develop an automated, generalisable geospatial analysis method, and thus constrained us to a manual approach. Comparison with Civil Registry data and, crudely, with the projected crude death rate in Yemen (around 6 per 1000 person-years50) suggests that imagery analysis under-estimated the true death toll, particularly in more remote periods (note however that Aden’s young, working population might in fact have a lower death rate than nationally). Plausible explanations could include (i) incomplete identification of active cemeteries (we could not access a formal list of cemeteries or a land registry against which to validate our list of 11 sites); (ii) problems identifying fresh graves in cemeteries where ‘infilling’ took place, largely due to insufficient VHR imagery; (iii) use of small, informal burial plots; and (iv) burial of decedents in ancestral areas of origin outside the governorate. The latter two phenomena are plausible when considering the reported spike in burial costs during the analysis period13, and the ongoing economic crisis affecting Yemenis. Uncertainty in population denominators could ...

    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.

    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.

  4. Version published to 10.1101/2020.10.27.20216366v1 on medRxiv