The ZOE Covid Study App (which recently changed its estimates again, which doesn’t exactly instil confidence) released a study earlier in the month (or rather a press release with no link to an actual study) which claimed natural immunity following infection “only gave 65% protection against catching it again”. This compared to 71% protection from the AstraZeneca vaccine (rising to 90% for those who had tested positive for Covid before) and 87% protection from the Pfizer vaccine (rising to 95% for the previously test-positive). The researchers say the results came from during the U.K. Delta wave.
This is a surprisingly low estimate for the protection given by natural immunity. Other estimates have tended to be more like 80% against testing positive and 90% against symptomatic infection. A recent study from Israel (not yet peer-reviewed) found natural immunity was 13 times better than Pfizer vaccination at preventing PCR positives during the Delta surge and 27 times better at preventing symptomatic infection.
The ZOE result is similar, however, to a recent (very flawed) study based on the ONS infection survey, which claimed to find just 55% protection from natural infection. A similarly flawed study from Oxford University, also based on the ONS survey, found natural infection just 66% effective.
The main problem with the ZOE study is that it only looks at infections from May and June 2021. This was mostly a time of very low prevalence, though with the beginnings of the Delta surge occurring in the latter half.