That's the rate when an infection establishes itself in blood. In a known outbreak (Royal Brompton Hospital), I recall that 9 patients out of 50 (18%) developed candidemia, but none of them died.

However, a study from India reports:

Among 1,400 ICU-acquired candidemia cases (overall incidence of 6.51 cases/1,000 ICU admission), 65.2 % were adult. Though the study confirmed the already known risk factors for candidemia, the acquisition occurred early after admission to ICU (median 8 days; interquartile range 4-15 days), even infecting patients with lower APACHE II score at admission (median 17.0; mean ± SD 17.2 ± 5.9; interquartile range 14-20). The important finding of the study was the vast spectrum of agents (31 Candida species) causing candidemia and a high rate of isolation of Candida tropicalis (41.6 %). Azole and multidrug resistance were seen in 11.8 and 1.9 % of isolates. Public sector hospitals reported a significantly higher presence of the relatively resistant C. auris (8.2 vs. 3.9 %; p = 0.008) and C. rugosa (5.6 vs. 1.5 %; p = 0.001). The 30-day crude and attributable mortality rates of candidemia patients were 44.7 and 19.6 %, respectively. Logistic regression analysis revealed significant independent predictors of mortality including admission to public sector hospital, APACHE II score at admission, underlying renal failure, central venous catheterization and steroid therapy.

Notes:

• C. auris is observed together with other pathogens like C. rugosa and C. tropicalis, leading to the question of which of them is the killer, or whether co-infection is the killer

• the mortality rate is given as a percent of people who developed candidemia (had Candida infections in their bloodstream), not as a percent of the total; upon hasty reading this can fool a person, and has fooled people before

• the mortality rate is split into two variables: crude mortality (the patient is dead, the patient tests positive) and attributable mortality (the patient is dead, we can see how Candida killed the patient)

• they found correlations which increased the risk; since they list "admission to a public sector hospital" among the factors, low standards of care in the Indian public healthcare system, or lack of experience in dealing with Candida (including wrong treatment protocols) may explain part of excess mortality

The study from Venezuela (a country afflicted with civil conflict and poverty) reports:

A total of 13 critically ill pediatric and 5 adult patients, with a median age of 26 days, were included. All were previously exposed to antibiotics and multiple invasive medical procedures. Clinical management included prompt catheter removal and antifungal therapy. Thirteen patients (72%) survived up to 30 days after onset of candidemia. AFLP fingerprinting of all C. auris isolates suggested a clonal outbreak. The isolates were considered resistant to azoles, but susceptible to anidulafungin and 50% of isolates exhibited amphotericin B MIC values of >1 μg/ml.

Unfortunately, the summary does not give adequate clues to understand the underlying conditions of the people (as I notice - mostly children). It mentions they were "critically ill", but doesn't mention if they were critically ill with Candida or before getting Candida. With such statistics as given in the summary - it is truly hard to evaluate the danger posed by a pathogen, because it's hard to isolate it from other factors. However, 30-day mortality was 28%, which is miserable.

My main conclusion seems to be: the first line of defense is having a hospital system that is ready to detect and deal with Candida. If this exists, many deaths can be avoided.

Of course, vaccination would be far preferable to combating it at a later stage.