They are consistent with both, And if the standard were preponderance of evidence and the only evidence on the table where the forensic report, it's two to one odds in favor of homicide by strangulation.
One thing I don't know regarding a medical examiner's job is whether they are supposed to take in additional evidence, and if an ME is required to give one single conclusion. If they are required to give one single conclusion, then it still makes sense that the New York medical examiner would conclude the thing that is more probable in light of circumstances, even if the forensic evidence alone indicates something else is more probable. Because a 25% frequency isn't low enough that one would bet on the 50% frequency event being what occurred in light of the fact that the homicide would have had to happen In a secure and protected facility with an isolated prisoner.
It's almost feels like a Bayesian versus frequentist analysis question.
But in any case, if the medical examiner is required to give a single answer, and their reputation suffers if that answer turns out to be wrong, I would assume that they would factor in mitigating circumstances to allow for the lower probability scenario to outweigh the higher probability scenario in isolation. And the forensics expert who is not responsible for analyzing future cases and may not be constrained to give a single answer is free to answer based on raw probabilities without factoring in circumstances.