In March 1947 a traveler who had come by bus from Mexico City died of smallpox in a New York hospital, and nobody recognized the disease until weeks later. Once the health department confirmed it in early April, the city set out to vaccinate nearly everyone. The figure usually repeated is about 6.35 million people in a matter of weeks, although the count and the time window vary by source. This piece covers how the department organized that campaign, where the vaccine came from, what the record leaves unclear, and why mass-prophylaxis planners still cite it.

The index case: a bus from Mexico City and a missed diagnosis

Most accounts name the first patient as Eugene Le Bar, an American businessman who had been living in Mexico, where smallpox was still circulating in the 1940s. He traveled to New York by bus with his wife and arrived at the start of March 1947. He was already ill on the trip, and in the days after arrival he developed a fever and a rash. He was seen at Bellevue Hospital and then moved to Willard Parker Hospital, the city’s hospital for communicable diseases, where he died on March 10 in most accounts.

His illness was not diagnosed as smallpox while he was alive. The retellings I trust say the clinicians who saw him considered other explanations for the rash, and some accounts mention a drug reaction as the working diagnosis. I am not going to describe his clinical course in more detail than that, because the bedside particulars in popular histories vary and I cannot trace them to the medical record. The broad outline is consistent, though. An adult with fever and a rash moved through two of the largest hospitals in the city, and the disease was not recognized until it had produced further cases.

The missed diagnosis is easy to understand and should not be treated as a scandal. By 1947 smallpox had become uncommon in the United States, and most of the variola circulating in the country in the preceding decades had been the milder form, variola minor. Many American physicians practicing then had never seen a case of the severe form, variola major. A clinician who has never seen a disease is badly placed to recognize it, especially in a hospital whose staff carried a reasonable assumption that smallpox belonged to another place and time. That assumption is the part of the story that still applies to practicing clinicians today.

Four weeks unrecognized: the hospital cases that exposed it

The outbreak became visible through secondary cases. In the weeks after Le Bar’s death, other people developed fever and rash, and the accounts tie these cases to the hospitals where he had been treated. When the illness in those patients was confirmed as smallpox in early April, investigators traced the chain back to the traveler who had died in March. The record that settles the order of events is the health department’s own published account and the department’s records. Anyone teaching from it should use those sources rather than anniversary journalism, which tends to tighten the timeline for narrative effect.

The gap of roughly four weeks between the index death and the confirmation explains everything the department did next. Smallpox has an incubation period that usually runs from about one to two weeks, so a month of unrecognized exposure allowed for at least one generation of cases and possibly the start of another. Visitors, patients, staff and family members had moved through the affected wards and then out into the city. Some of the contacts could be identified from hospital records, and others could not. The department therefore faced an unknown number of exposed people scattered across a city of millions, and it could not be certain that the cases already in hand were the only ones.

A second feature of the timing mattered. A contact who is vaccinated soon after exposure has a good chance of avoiding the disease or getting a milder case. CDC’s current guidance says vaccination within about three days of exposure can prevent smallpox or greatly reduce its severity, and that vaccination a few days later than that may still offer some benefit. By the time the department knew what it was dealing with, that window had long closed for the earliest contacts. Its realistic aim was to get ahead of the next generation of cases, and it had little time to do so.

The delay sat at recognition, not at the response

The 1947 response is remembered for its speed, but the outbreak grew during the four weeks before anyone named the disease, and almost no speed in dispensing could have recovered that time for the early contacts. In mass-prophylaxis planning, the clock starts at recognition, which depends on a clinician thinking of the diagnosis and a laboratory being able to confirm it. Exercises that begin with the confirmed case already announced skip the part of the timeline that went wrong in New York.

The decision to vaccinate a city of nearly eight million

The health commissioner was Dr. Israel Weinstein, and the mayor was William O’Dwyer. In early April the department announced the outbreak and urged residents to be vaccinated, and the effort expanded quickly into a campaign aimed at the whole city. The census population of New York was a little under 7.5 million in 1940 and a little under 7.9 million in 1950, so the 1947 population fell in that range. Many residents had been vaccinated as children, since school vaccination requirements were common, but immunity from smallpox vaccination wanes over the years, and adult protection could not be assumed.

The reasoning behind going citywide rather than vaccinating only known contacts follows from the delay described above. With contacts that could not all be identified and a month of possible spread, a contacts-only strategy would have depended on finding people the department had no way to find. The alternative used later in the global eradication program, usually called surveillance and containment or ring vaccination, relies on fast case detection and thorough contact tracing, both of which had already failed in New York by the time the disease was confirmed. Vaccinating the population was the strategy that still worked without knowing where the virus had gone.

The political side of the decision deserves attention because it is often left out of public health accounts. The campaign needed the mayor’s support, money and people from city departments outside public health, and cooperation from hospitals, physicians and the press. Accounts describe O’Dwyer being vaccinated publicly and the city’s newspapers and radio stations carrying the department’s message. I have not found a single document setting out who decided what and when, so I am describing the decision in general terms. For the department’s own reasoning, Weinstein’s published report on the outbreak in the American Journal of Public Health in 1947 is the place to start, and the New York City Municipal Archives hold departmental records.

Sites and staffing: clinics, hospitals, precincts and volunteers

The department had a base to build on. New York in 1947 had a network of district health centers and child health stations, the municipal hospitals, and a large private medical community. The accounts describe free vaccination offered at health department facilities and hospitals and then pushed out into buildings that were not normally clinical, with police station houses the example most often mentioned. Photographs from the period show long lines of people waiting outside. Clinics extended their hours, and the accounts describe vaccination continuing into the evening and on weekends at the peak.

Staffing came from wherever the city could find it. Department physicians and nurses formed the core, and they were joined by hospital staff, private physicians who volunteered, and lay volunteers, including Red Cross workers, who handled the lines, the paperwork and the flow of people so that clinicians could keep vaccinating. Accounts also describe private physicians vaccinating patients in their offices. I have not seen a reliable headcount of volunteers, and the figures that circulate in retellings do not come with a source, so I am not repeating one.

The vaccination procedure itself made this scale possible. Smallpox vaccine in that era was applied to the skin, usually with a needle using the multiple pressure or scratch technique, rather than drawn up and injected. A trained vaccinator could therefore move through a line quickly, and the procedure could be taught to additional clinical staff in a short time. That is very different from a mass event that requires injection, a dose calculation, or an individual medical screening interview, and the difference goes straight to throughput. Planners who quote the 1947 numbers as a benchmark for a modern point of dispensing need to account for it.

The public behavior that made the lines long also made them work. The accounts describe people turning out in very large numbers, largely voluntarily, in response to the department’s messaging. That turnout is part of what the city got right, but the department could not count on it in advance. A planner today should not assume that a population in a different era, with different levels of trust in public health, will queue the same way.

Vaccine supply: cold calf lymph and every source the city could reach

The vaccine used in 1947 was glycerinated calf lymph, produced by growing vaccinia virus on the skin of calves. It lost potency at room temperature, so it had to be kept cold from production to use, which made refrigeration and transport a constraint at every site. The freeze-dried vaccine that later made smallpox vaccination practical in the tropics during the eradication program was not yet in routine use in New York. Each clinic, police precinct and hospital line therefore depended on a cold supply chain run with the refrigeration of the 1940s.

The accounts describe the city drawing on several sources at once. The health department had its own long-standing capacity to produce vaccine, and commercial manufacturers increased their production for the city. Federal agencies and the military are also described as contributing stock. I am not giving dose counts by source because the figures I have seen do not agree and I cannot trace them to a supply ledger. The department’s published report and the municipal records are where those numbers would be, if they survive in a usable form.

The supply story matters for current planning because it shows a city taking every source it could reach in the absence of any national stockpile. The Strategic National Stockpile did not exist in 1947. The city’s position depended on its own laboratory, its relationships with nearby manufacturers, and its ability to persuade federal and military holders to release stock. A modern jurisdiction gets countermeasures through a federal and state request process that did not exist then. That process brings its own delays and steps, from the request through state receipt and staging to local distribution, and a local plan has to account for every one of them.

The numbers: about 12 cases, 2 deaths and a vaccination count to cite as a range

The outbreak itself was small. Most accounts give 12 cases and 2 deaths, with Le Bar among the dead. I have seen secondary sources with slightly different totals, and I cannot trace those variants to a primary record, so I give 12 cases and 2 deaths as the figures in most accounts and note that variants circulate. The small size of the outbreak is part of why the campaign is remembered as a success. It is also why the campaign’s effect is hard to measure, because nobody can observe how many cases would have followed without it.

The vaccination count needs more care. The figure most often repeated is about 6.35 million people, and it traces to the health department’s own reporting of the campaign, including Weinstein’s published account. Other retellings, mostly later journalism and anniversary pieces, give a lower round figure of about 5 million. The time window also varies. Some accounts say about two weeks, while others describe the campaign as running for most of a month, and I suspect much of the disagreement comes from whether a writer counts the peak weeks or the whole effort. I have not been able to trace the 5 million figure to a separate primary tally, so I treat it as a secondary-source variant rather than an independent count.

The figures should be cited as a range with sources attached: about 5 million to about 6.35 million vaccinations, over a period that accounts describe as anywhere from roughly two weeks to about a month, with the higher figure coming from the health department’s reporting. A training slide that reads “6.35 million in two weeks” combines the upper vaccination figure with the shortest time window, and that pairing produces a throughput rate that no single source clearly supports.

Do not build a throughput benchmark on the tidiest pairing of numbers

Planners sometimes divide the highest vaccination count by the shortest time window and use the result as a target for modern dispensing sites. The 1947 figures do not support that arithmetic, because the count and the window come from different accounts, the count may include repeat vaccinations, and the procedure was a skin scratch rather than an injection or an oral medication with screening. Use the campaign to show that a city can reach most of its population quickly, and take actual throughput targets from your own exercises and from current state and federal guidance.

What the record says less about: reactions, duplicates and cost

Smallpox vaccine made with live vaccinia virus carries real risks, including rare serious complications such as encephalitis, and some of those complications can be fatal. A campaign on the scale of New York’s would be expected to produce some serious reactions. Later discussions of the 1947 campaign, particularly during the debates over the civilian smallpox vaccination program of 2002 and 2003, raised the question of how many people it harmed. I have seen figures quoted for deaths from vaccine complications in the 1947 campaign, but I cannot trace them to a reliable primary source, so I am not giving one. The honest statement is that the campaign carried a cost in adverse events and that the size of that cost is not well documented in the sources I can verify.

The vaccination total has a definitional problem as well. In a voluntary campaign run at many sites, with private physicians vaccinating in their offices as well, some people were almost certainly vaccinated more than once, and some vaccinations probably went uncounted. I have not seen a statement of how the department handled duplicates, so the totals are better read as counts of vaccinations reported than as verified counts of distinct people protected. The vaccinations also did not all take. Vaccine takes were normally checked by looking at the site several days later, and I have not seen take rates for the campaign as a whole.

I have not found a reliable figure for the cost of the campaign, including staff overtime, vaccine, and the time lost by millions of people standing in line. The absence of these details does not make the campaign a failure, but it does mean that the version taught in planning courses usually covers only the outbreak and the count. A planner who uses New York 1947 to argue for a mass campaign should also acknowledge that the campaign accepted some vaccine harm and that its overall effect has never been formally measured.

Why planners still cite it: points of dispensing and the limits of the comparison

The campaign is still cited because it is one of the few well-documented cases of an American city delivering a countermeasure to most of its population within weeks, using sites and staff that were mostly not built for the job. Modern mass-prophylaxis doctrine rests on the same idea. CDC’s Cities Readiness Initiative, launched in 2004, set the goal of getting medical countermeasures to an entire metropolitan population within a fixed time after a decision to do so, and local plans meet that goal through points of dispensing. Those are open sites for the general public and closed sites where an employer, hospital or institution dispenses to its own people. The way the 1947 city leaned on police precincts, hospitals and private physicians’ offices is an early version of that mixed model.

The comparison has limits that should be stated plainly. The countermeasure most current plans assume is an oral antibiotic for an anthrax release, and that scenario runs against a much shorter clock than the one New York faced, with dispensing that involves screening and medication choices. Smallpox vaccination today would also be different, because the vaccines in the national stockpile, the screening for people at higher risk of complications, and the legal and consent framework have all changed since 1947. Current smallpox response doctrine favors finding cases and vaccinating rings of contacts first, with broader vaccination considered according to the situation. Verify current federal smallpox response guidance with CDC and your state health department rather than assuming 1947 practice applies.

The lessons that do carry over concern organization. New York’s campaign depended on a health department with standing relationships with the mayor, the police, hospitals, private physicians, manufacturers and the press, all in place before the outbreak, and on public trust that turned an announcement into lines around the block. A modern jurisdiction can reproduce those relationships before an event, and that preparation is where most of the work in a points of dispensing plan actually lies.

From the communications side of the work I have done, the part of the 1947 story I find most transferable is the message discipline. Accounts describe the city speaking with essentially one voice through the department, the mayor and the press. Modern plans tend to have the dispensing logistics worked out in detail and the public information function sketched in a paragraph, even though turnout depends on that function.

Closed points of dispensing are where 1947 most directly applies

The city’s use of hospitals, private offices and non-clinical public buildings to reach people where they already were is the closest historical parallel to the closed POD agreements in a modern mass-prophylaxis plan. Fire, EMS, police and dispatch agencies are usually expected to protect their own workforce and families first so that they can keep working during a public campaign. If your agency has not signed a closed POD agreement with the local health department, or has signed one but never staffed it in an exercise, your own responders may end up standing in the public line.

What to do at your agency

  • The fire chief, EMS director or police chief asks the local health department’s emergency preparedness coordinator whether the agency is listed as a closed point of dispensing, and the task is finished when the signed agreement, or a written statement that none exists, is filed with the agency’s continuity plan.
  • The agency’s designated POD lead pulls the current roster of employees and household members the agency would cover under that agreement and compares it to the head count the health department has on file, and the task is finished when the two numbers match or the health department has the corrected figure in writing.
  • The training officer adds the 1947 New York campaign to the next scheduled infectious disease or mass-prophylaxis briefing, and the task is finished when the session notes give the vaccination count as a range of about 5 million to about 6.35 million attributed to its sources, give 12 cases and 2 deaths as the figures in most accounts, and state that the recognition delay was about four weeks.
  • The EMS medical director confirms that the agency’s infection control protocol tells crews what to do when they encounter fever with an unusual rash in a patient with recent international travel, and the task is finished when the protocol names who the crew notifies and that notification path has been checked against the health department’s current after-hours number.
  • The PSAP supervisor confirms that the dispatch center has the health department’s current after-hours reportable disease contact entered in its resource directory, and the task is finished when a test call has reached a person at that number and the date is logged.
  • The emergency manager reviews the public information section of the existing mass-prophylaxis or pandemic annex and writes one paragraph naming who approves joint messages among the health department, the elected executive and public safety agencies, and the task is finished when the health officer has initialed the revised page.

Takeaways

  • A traveler who arrived in New York by bus from Mexico City in early March 1947 died of smallpox on March 10 in most accounts, and the disease was not recognized until secondary cases were confirmed in early April.
  • The four-week gap between the index death and confirmation is why the health department could not rely on tracing contacts and chose to vaccinate the city as a whole.
  • Most accounts give 12 cases and 2 deaths for the outbreak, while variant totals circulate in secondary sources that I cannot trace to a primary record.
  • The vaccination count should be cited as a range from about 5 million in later retellings to about 6.35 million in the health department’s own reporting, over a window that sources describe as anywhere from two weeks to about a month.
  • The campaign drew on health centers, hospitals, police precincts, private physicians and lay volunteers, and it used a skin-scratch procedure that allowed far faster throughput than injection or screened oral dispensing.
  • Vaccine supply came from the city’s own production, commercial manufacturers and federal and military stock, at a time when no national stockpile existed and the vaccine had to be kept cold.
  • The record says little that can be verified about vaccine complications, duplicate vaccinations and cost, so the campaign should be taught with those gaps stated rather than filled.
  • Planners cite 1947 as an early model for points of dispensing, but current smallpox doctrine, vaccines and dispensing scenarios differ enough that throughput targets should come from current guidance and local exercises.
Questions or a different view?

Reach me through the contact page. I read every message.