It's a distortion of reality. The origin is probably https://www.rmv.se/verksamheter/medicinska-aldersbedomningar/metoder/ where they describe the method used. The result of an examination is given in different levels of confidence, "The assessment of the completed survey indicates that the surveyor is 18 years of age or older", "The assessment of the survey conducted may indicate that the surveyor is 18 years or older" and "The assessment of a survey conducted possibly indicate that it is under 18 years of age". Even at best the method have a 10% failure rate. When a second opinion is made the result differs in 85% of the cases.SvD
The figure 80% is given in https://www.metro.se/artikel/nya-siffror-80-procent-av-unders%C3%B6kta-ensamkommande-%C3%B6ver-18-%C3%A5r-men-metoden-f%C3%A5r-kritik The raw numbers are that 6880 medical age assessments were submitted. 2481 statements were given. The result is that 2002 were assessed as being 18 or older. In 25 of those it was judged as "possibly" and 432 "may be".
Summary: In about 80% of the suspected cases the suspicion turned out to be possibly correct. Medical age assessment is of course not used in every case, only where someone claims to below 18, but is suspected of being above 18. So the high number may say more about how good they are at judging age from just observing them (psychological and psychological).