1. Epidemiology and Global Health

Projected long-term effects of colorectal cancer screening disruptions following the COVID-19 pandemic

  1. Pedro Nascimento de Lima  Is a corresponding author
  2. Rosita van den Puttelaar
  3. Anne I Hahn
  4. Matthias Harlass
  5. Nicholson Collier
  6. Jonathan Ozik
  7. Ann G Zauber
  8. Iris Lansdorp-Vogelaar
  9. Carolyn M Rutter
  1. RAND Corporation, United States
  2. Erasmus MC, Netherlands
  3. Memorial Sloan Kettering Cancer Center, United States
  4. Argonne National Laboratory, United States
  5. Fred Hutchinson Cancer Center, United States
https://doi.org/10.7554/eLife.85264
Share this article
Cite this article
  1. Pedro Nascimento de Lima
  2. Rosita van den Puttelaar
  3. Anne I Hahn
  4. Matthias Harlass
  5. Nicholson Collier
  6. Jonathan Ozik
  7. Ann G Zauber
  8. Iris Lansdorp-Vogelaar
  9. Carolyn M Rutter
(2023)
Projected long-term effects of colorectal cancer screening disruptions following the COVID-19 pandemic
eLife 12:e85264.
https://doi.org/10.7554/eLife.85264
  2. Download BibTeX
  3. Download .RIS
2 figures, 4 tables and 4 additional files

Figures

Figure 1 with 1 supplement
Download asset Open asset
Screening benefits lost due to disruptions by cohort and scenario.

Notes: Each dot represents the estimated life-years lost per 1000 individuals or life-days lost from one model under the high sensitivity scenario. Results are ordered from highest to lowest reduction in benefit induced by the pandemic. Scenarios that result in less than 2 life-days lost per person are omitted from this figure and presented in a Supplementary figure. This figure does not present a counterfactual no-screening scenario for the 50-year-olds.

Figure 1—figure supplement 1
Download asset Open asset
Estimated loss of life in minor disruption scenarios.

Notes: Each dot represents the estimated life-years lost per 1000 individuals or life-days lost from one model under the high sensitivity scenario. Results are ordered from highest to lowest reduction in benefit induced by the pandemic. This figure supplement only shows scenarios that resulted in in less than 2 life-days lost per person. This figure does not present a counterfactual no-screening scenario for the 50-year-olds.

Figure 2 with 1 supplement
Download asset Open asset
Life-years gained (LYG) in high- vs. low-sensitivity scenarios.

Notes: Each dot represents one scenario considered in this study. The horizontal axis displays the number of LYG estimated in that scenario under a high colonoscopy sensitivity scenario. The vertical axis shows the results for the same cohort under a low colonoscopy sensitivity scenario. If sensitivity did not affect the estimate, then all points would be on top of a 45-degree line. Different colors represent CRCSPIN and MISCAN models.

Figure 2—figure supplement 1
Download asset Open asset
Life-years lost in high- vs. low-sensitivity scenarios.

Notes: Each dot represents one scenario considered in this study. The horizontal axis displays the number of life-years lost due to disruptions (LYL) estimated in that scenario under a high colonoscopy sensitivity scenario. The vertical axis shows the results for the same cohort under a low colonoscopy sensitivity scenario. If sensitivity did not affect the estimate, then all points would be on top of a 45-degree line. Different colors represent CRCSPIN and MISCAN models.

Tables

Table 1
Study cohorts and scenarios.
CohortNo-disruption counterfactualCRC screening disruption scenario
DescriptionLabel
Unscreened
50-year-olds (U50)		Decennial COL from age 50 to 70Short-term delays of [d] months*U50 | C[d]m
Long-term delay (COL at age 65 and 75)U50 | C@65
Annual FIT from age 50 to 75Short-term delays*U50 | F[d]m
Unscreened
60-year-olds (U60)		Decennial COL from age 60 to 70Short-term delays*U60 | C[d]m
Long-term delay (COL at age 65 and 75)U60 | C@65
Annual FIT from age 60 to 75Short-term delays*U60 | F[d]m
COL-adherent
60-year-olds (C60)		Decennial COL from age 50 to 70Short-term delays*C60 | C[d]m
Switch to annual FIT and short-term delaysC60 | F[d]m
Long-term delay (COL at age 65 and 75)C60 | C@65
Discontinue screeningC60 | U
FIT-adherent
60-year-olds (F60)		Annual FIT from age 50 to 75Short-term delays*F60 | F[d]m
Discontinue screeningF60 | U
FIT-semi-adherent
60-year-olds (f60)		Biannual FIT from age 50 to 56, annual FIT from age 60 to 75Short-term delays*f60 | F[d]m
Discontinue screeningf60 | U
Unscreened
70-year-olds (U70)		COL at age 70Short-term delays*U70 | C[d]m
Long-term delay (COL at age 75)U70 | C@75
Annual FIT from age 70 to 75Short-term delaysU70 | F[d]m
COL-adherent
70-year-olds (C70)		Decennial COL from age 50 to 70Short-term delays*C70 | C[d]m
Switch to annual FIT and short-term delaysC70 | F[d]m
Long-term delay
Perform COL at age 75		C70 | C@75
Discontinue screeningC70 | U
FIT-adherent
70-year-olds (F70)		Annual FIT from age 50 to 75Short-term delays*F70 | F[d]m
Discontinue screeningF70 | U

  1. Notes: This table presents the scenarios considered in this study. Each scenario corresponds to a combination of a population cohort, indicated by their age during the first COVID-19 lockdowns (March 2020), a pre-pandemic, and a post-pandemic screening regimen. The scenarios aim to represent possible combinations of screening regimens followed in the United States. The first letter in the scenario code represents screening before the pandemic and the second letter represents screening after the pandemic.

  2. *

    Delays of 3, 9, and 18 months. Letter d stands for the number of months of delays.

Table 2
Per lesion test sensitivity and specificity.
Sensitivity*Specificity†
TestAdenoma1–5 mmAdenoma6–9 mmAdenoma ≥10 mmPreclinical cancer
Colonoscopy, high sensitivity‡0.750.850.950.950.86
Colonoscopy, low sensitivity§0.550.700.900.950.86
FIT
MISCAN0.000.1140.1590.62565/0.8860.97
CRC-SPIN0.050.150.22*0.740.97

  1. Notes: This table presents the assumed test characteristics. We simulated two colonoscopy sensitivity scenarios seeking to represent a range of colonoscopy sensitivity of gastroenterologists in the United States.

  2. *

    Sensitivity is for lesions within reach of the scope. We assume the same test characteristics for follow-up and surveillance colonoscopy as for screening colonoscopy.

  3. †

    For FIT, the lack of specificity reflects detection of bleeding from other causes. We assume other-cause bleeding is independent of adenoma status. For colonoscopy, the lack of specificity reflects detection of non-adenomatous lesions, but specificity is handled in post-processing in cost-effectiveness analyses. Since this study does not consider burden outcomes, specificity is not considered in this paper. Specificity values were obtained from Lin et al., 2021.

  4. ‡

    Baseline scenarios used in Zauber et al., 2008.

  5. §

    In line with low-sensitivity scenarios compatible with Rutter et al., 2022.

  6. CRC-SPIN uses per-person test sensitivity for stool-based tests that are based on the size of the most advanced lesion. To account for the likelihood that a person with multiple adenomas is more likely than a person with only one to have a positive stool test, MISCAN uses lesion-based sensitivities instead of person-based sensitivities. Lesion-based sensitivities were derived by calibrating the person-based sensitivities to the number of people having one or more small/medium/large adenomas or cancers detected by stool-based testing with diagnostic colonoscopy, divided by those having one or more small/medium/large adenomas or cancers detected by colonoscopy screening.

Table 3
CRC surveillance intervals.
Finding at second-most recent colonoscopy*†Finding at first-most recent colonoscopy*†Interval‡ to next colonoscopy, years
No prior colonoscopyNormal colonoscopySee note below§
1–2 adenomas <10 mm7
3–4 adenomas <10 mm3
10 adenomas <10 mm or any adenoma ≥10 mm3
>10 adenomas1
Normal colonoscopyNormal colonoscopy10
1–2 adenomas <10 mm7
3–4 adenomas <10 mm3
5–10 adenomas <10 mm or any adenoma ≥10 mm3
>10 adenomas1
1–2 adenomas <10 mmNormal colonoscopy10
1–2 adenomas <10 mm7
3–4 adenomas <10 mm3
5–10 adenomas <10 mm or any adenoma ≥10 mm3
>10 adenomas1
3–4 adenomas <10 mmNormal colonoscopy10
1–2 adenomas <10 mm7
3–4 adenomas <10 mm3
5–10 adenomas <10 mm or any adenoma ≥10 mm3
>10 adenomas1
5–10 adenomas <10 mmNormal colonoscopy5
or1–2 adenomas <10 mm5
any adenoma ≥10 mm3–4 adenomas <10 mm3
5–10 adenomas <10 mm or any adenoma ≥10 mm3
>10 adenomas1
>10 adenomas of any sizeNormal colonoscopy5
1–2 adenomas <10 mm5
3–4 adenomas <10 mm3
5–10 adenomas <10 mm or any adenoma ≥10 mm3
>10 adenomas1
  1. *

    A normal colonoscopy is one in which no adenomas, SSPs (not currently simulated), or CRC is detected.

  2. †

    This table omits the case where CRC is detected at a screening, diagnostic, or surveillance colonoscopy because the CISNET CRC models do not simulate detailed events following CRC diagnosis.

  3. ‡

    The Multi-Society Task Force provides a range for some intervals (e.g. the interval for 3–4 adenomas <10 mm is 3–5 years). In such cases, we selected the shortest intervals provided.

  4. §

    A person whose first screening or diagnostic colonoscopy is normal does not enter surveillance but instead resumes screening with the original modality 10 years after the normal colonoscopy. The exception to the 10-year waiting period is when the first colonoscopy is a screening colonoscopy with an x-year interval, where x>10. In that case, the next colonoscopy is in x years.

Table 4
Projected life-years (LY) per 1000 individuals.
ScenarioModelNo screeningScreening without disruptionsScreening with disruptionsLoss due to disruptions
LY[a]LY[b]LYG[b-a]LY[c]LYG[c-a]LY[b-c]% LYG loss[(b-c)/b]
U50 | C3mCRCSPIN31,59531,89329931,89229721
MISCAN31,22231,49427331,49127031
U50 | C9mCRCSPIN31,59531,89329931,89029531
MISCAN31,22231,49427331,49026852
U50 | C18mCRCSPIN31,59531,89329931,88629172
MISCAN31,22231,49427331,48826662
U50 | C@65CRCSPIN31,59531,89329931,76617212743
MISCAN31,22231,49427331,39016910438
U50 | F3mCRCSPIN31,59531,86627131,86527010
MISCAN31,22231,48326131,48125911
U50 | F9mCRCSPIN31,59531,86627131,86226831
MISCAN31,22231,48326131,48025821
U50 | F18mCRCSPIN31,59531,86627131,85826473
MISCAN31,22231,48326131,47825652
U60 | C3mCRCSPIN23,33623,55722123,560224-3-1
MISCAN23,11423,30919523,30419053
U60 | C9mCRCSPIN23,33623,55722123,55521921
MISCAN23,11423,30919523,30118784
U60 | C18mCRCSPIN23,33623,55722123,547211104
MISCAN23,11423,30919523,296182147
U60 | C@65CRCSPIN23,33623,55722123,5121764520
MISCAN23,11423,30919523,2671534221
U60 | F3mCRCSPIN23,33623,52819123,529193-2-1
MISCAN23,11423,29117723,28817431
U60 | F9mCRCSPIN23,33623,52819123,52518921
MISCAN23,11423,29117723,28517163
U60 | F18mCRCSPIN23,33623,52819123,51918384
MISCAN23,11423,29117723,280166116
C60 | C3mCRCSPIN23,24323,54129823,54129900
MISCAN23,07723,32024323,31824121
C60 | C9mCRCSPIN23,24223,54129823,54129900
MISCAN23,07723,31924323,31724021
C60 | C18mCRCSPIN23,24223,54129823,54029800
MISCAN23,07723,32024323,31723931
C60 | F3mCRCSPIN23,24223,54129823,53228993
MISCAN23,07723,32024323,310233104
C60 | F9mCRCSPIN23,24323,54129823,53228993
MISCAN23,07723,31924323,31023394
C60 | F18mCRCSPIN23,24323,54129823,53228993
MISCAN23,07723,32024323,309232114
C60 | C@65CRCSPIN23,24323,54129823,53829531
MISCAN23,07723,32024323,308231115
C60 | UCRCSPIN23,24323,54129823,43519210636
MISCAN23,07723,32024323,19511812551
F60 | F3mCRCSPIN23,30723,57927223,57626931
MISCAN23,14423,37723423,37723300
F60 | F9mCRCSPIN23,30623,57827223,57526931
MISCAN23,14323,37623423,37623400
F60 | F18mCRCSPIN23,30723,57827223,57326752
MISCAN23,14323,37723423,37623310
F60 | UCRCSPIN23,30723,57927223,46716011141
MISCAN23,14323,37623423,2831419340
f60 | F3mCRCSPIN23,31423,56224923,56024721
MISCAN23,13623,35521923,35321721
f60 | F9mCRCSPIN23,31423,56324923,55924541
MISCAN23,13623,35521923,35221631
f60 | F18mCRCSPIN23,31323,56224923,55624263
MISCAN23,13623,35521923,35021452
f60 | UCRCSPIN23,31323,56224923,41910514358
MISCAN23,13423,35321923,2036815069
U70 | C3mCRCSPIN15,97316,10212816,09912632
MISCAN15,74815,86611715,85710898
U70 | C9mCRCSPIN15,97316,10212816,09412086
MISCAN15,74815,86611715,8521031412
U70 | C18mCRCSPIN15,97316,10212816,0861131612
MISCAN15,74815,86611715,845972118
U70 | C@75CRCSPIN15,97316,10212816,052795039
MISCAN15,74815,86611715,815665143
U70 | F3mCRCSPIN15,97316,0699516,0679322
MISCAN15,74815,8409215,8338578
U70 | F9mCRCSPIN15,97316,0699516,0639056
MISCAN15,74815,8409215,830811011
U70 | F18mCRCSPIN15,97316,0699516,058841112
MISCAN15,74815,8409215,824761617
C70 | C3mCRCSPIN15,68315,96828515,96828500
MISCAN15,59015,82423415,82323410
C70 | C9mCRCSPIN15,68315,96828515,96828500
MISCAN15,59015,82423415,82323310
C70 | C18mCRCSPIN15,68415,96928515,96828500
MISCAN15,58915,82423415,82223321
C70 | C@75CRCSPIN15,68315,96828515,96828500
MISCAN15,59015,82423415,81922952
C70 | F3mCRCSPIN15,68315,96828515,96428152
MISCAN15,59015,82423415,81822863
C70 | F9mCRCSPIN15,68315,96828515,96428141
MISCAN15,59015,82423415,81722873
C70 | F18mCRCSPIN15,68315,96828515,96428142
MISCAN15,58915,82423415,81722773
C70 | UCRCSPIN15,68315,96828515,9302473813
MISCAN15,58115,81523415,7261468938
F70 | F3mCRCSPIN15,76416,02425916,02325900
MISCAN15,67615,90222615,90222600
F70 | F9mCRCSPIN15,76516,02425916,02325910
MISCAN15,67715,90322615,90322510
F70 | F18mCRCSPIN15,76616,02525916,02425810
MISCAN15,67715,90322615,90322600
F70 | UCRCSPIN15,76416,02425915,9902263313
MISCAN15,67715,90322615,8731963013

  1. Notes: Outcomes calculated over the lifetime of a cohort of 1000 average-risk, CRC-free individuals with age at pandemic defined in the scenario description. The scenario column describes colorectal cancer screening disruption scenarios, as presented in Table 1. Life-years (LY) and Life-years gained (LYG) are computed over the remaining lifespan of individuals starting at the beginning of 2020. All values refer to cohort-level estimates – that is, the expected LY of an average-risk person. This table presents results assuming high colonoscopy sensitivity. Supplementary file 1 presents additional results for the low colonoscopy sensitivity scenario, and Supplementary files 2 and 3 present CRC cases and deaths outcomes.

Additional files

MDAR checklist
https://cdn.elifesciences.org/articles/85264/elife-85264-mdarchecklist1-v1.pdf
Supplementary file 1

Projected lifetime life-years (LY) per 1,000 individuals.

https://cdn.elifesciences.org/articles/85264/elife-85264-supp1-v1.xlsx
Supplementary file 2

Projected lifetime CRC cases per 1,000 individuals.

https://cdn.elifesciences.org/articles/85264/elife-85264-supp2-v1.xlsx
Supplementary file 3

Projected lifetime CRC deaths per 1,000 individuals.

https://cdn.elifesciences.org/articles/85264/elife-85264-supp3-v1.xlsx

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Pedro Nascimento de Lima
  2. Rosita van den Puttelaar
  3. Anne I Hahn
  4. Matthias Harlass
  5. Nicholson Collier
  6. Jonathan Ozik
  7. Ann G Zauber
  8. Iris Lansdorp-Vogelaar
  9. Carolyn M Rutter
(2023)
Projected long-term effects of colorectal cancer screening disruptions following the COVID-19 pandemic
eLife 12:e85264.
https://doi.org/10.7554/eLife.85264