This rather startling and important fact about polio is demonstrated very thoroughly by Roman Bystrianyk and Suzanne Humphries in their study of many vaccines called Dissolving Illusions (2013). I give a portion of the Polio change material without omission and including footnotes below. It is taken from a Kindle edition without pagination, starting about 40% into the book:

Specific polio diagnosis was not pursued with laboratory testing before 1958. The diagnostic criteria for polio were very loose prior to the field trials for the vaccine in 1954. Before the vaccine was deployed, health-care professionals were vigilantly programmed to be on the lookout for polio. After the trials, they were vigilantly noting who developed polio—vaccinated or unvaccinated— and made every effort to diagnose a non-polio illness in a vaccinated person. Dr. Bernard Greenberg, head of the department of biostatistics of the University of North Carolina School of Public Health and chairman of the Committee on Evaluation and Standards of the American Public Health Association, stated in 1960:

Prior to 1954 any physician who reported paralytic poliomyelitis was doing his patient a service by way of subsidizing the cost of hospitalization and was being community-minded in reporting a communicable disease. The criterion of diagnosis at that time in most health departments followed the World Health Organization definition : “Spinal paralytic poliomyelitis: signs and symptoms of nonparalytic poliomyelitis with the addition of partial or complete paralysis of one or more muscle groups, detected on two examinations at least 24 hours apart.” Note that “two examinations at least 24 hours apart” was all that was required.… Laboratory confirmation and presence of residual [longer than 24 hours] was not required. {366}

The practice among doctors before 1954 was to diagnose all patients who experienced even short-term paralysis (24 hours) with “polio.” In 1955, the year the Salk vaccine was released, the diagnostic criteria became much more stringent. If there was no residual paralysis 60 days after onset, the disease was not considered to be paralytic polio. This change made a huge difference in the documented prevalence of paralytic polio because most people who experience paralysis recover prior to 60 days. Dr. Greenberg said:

The change in 1955 meant that we were reporting a new disease, namely, paralytic poliomyelitis with a longer-lasting paralysis. Furthermore diagnostic procedures have continued to be refined. Coxsackie virus and aseptic meningitis have been distinguished from paralytic poliomyelitis. Prior to 1954 large numbers of these cases were mislabeled as paralytic poliomyelitis . Thus, simply by changes in diagnostic criteria , the number of paralytic cases was predetermined to decrease in 1955-1957, whether or not any vaccine was used. {367}

{Humphries, Suzanne; Bystrianyk, Roman (2013-12-12). Dissolving Illusions (Kindle Locations 3147-3167). Kindle Edition. }