Skip to main content
Back to Steel City Stories
The Year Pittsburgh Became America’s Deadliest City — Steel City Stories
HistoryChapter 27 of 4118 min read

The Year Pittsburgh Became America’s Deadliest City

Steel City Stories

Preparing the first passage.

Steel City Stories · Chapter 27

The receipt is a barracks count and a dead soldier’s name, not a municipal slogan. By October 4, 1918, influenza had reached the temporary army barracks on the Oakland campuses of the University of Pittsburgh and the school that is now Carnegie Mellon. A Pittsburgh Post-Gazette reconstruction published in 2020, working from the city’s newspapers, names Charles N. Patterson, twenty-nine, of Aspinwall, as one soldier already dead, with another dying and a third seriously ill. Military authorities took over part of Magee Hospital. A photograph preserved in the University of Pittsburgh’s archives, and captioned in that same reconstruction, shows one of seven temporary barracks on the Pitt campus, each described as able to hold a thousand men. Whether every bunk was full, the caption records the scale of the wartime city Pittsburgh was running while a respiratory epidemic arrived. Oakland was not a quiet civic campus. It was a military housing project set down in a dense, smoky, streetcar city, and the virus used the housing.

The University of Michigan’s American Influenza Epidemic digital encyclopedia, which builds a day-by-day file from Pittsburgh papers, starts the local political story one day earlier in spirit and the same week in fact. On October 3, Pennsylvania’s acting commissioner of health, Dr. Benjamin Franklin Royer, ordered saloons, theaters, movie houses, and other amusements closed across the state and banned public meetings, funerals, parades, and visits to the sick. He left schools and churches to local authorities. Pittsburgh officials did not wait for the paper to arrive. Mayor Edward V. Babcock, health director Dr. W. H. Davis, and school superintendent William M. Davidson met and decided to carry out the order they expected by morning. Davis reminded physicians that influenza had been made a reportable disease in Pennsylvania on September 21. Babcock’s objection is on the record in the encyclopedia’s quotation from the next day’s meeting. He told reporters that, in view of local conditions and the absence of any epidemic of influenza, the state department’s orders appeared too drastic. The epidemic then made the sentence obsolete. Within days the city had hundreds of cases.

What closed, and what the city insisted on keeping open

On Friday, October 4, Davis ordered hospitals to stop unnecessary visits and to report empty beds. Churches were asked to shorten services, to turn away the visibly ill, and, soon after, to end regular services while remaining available for private prayer. Jury trials stopped. Elevator loads were capped at six. Night school, which lacked medical inspectors, was halted, and the encyclopedia puts the affected night enrollment at about fifteen thousand. Day schools stayed open. Davis reinforced them with thirty-two additional doctors and argued that a supervised classroom was safer than the street. That argument is the chapter’s policy spine. Pittsburgh did not ignore the epidemic. It staged its closures. Amusements and crowds fell under the state ban immediately. The institution that gathered more than a hundred thousand children every morning was treated as a medical asset rather than as a crowd.

The encyclopedia’s October file shows how long that theory lasted. By the third week, Superintendent Davidson thought attendance was about seventy-five percent of normal, and school hygiene director Dr. H. B. Burns spoke of as many as a thousand pupils sick. On October 21, Councilman Enoch Rauh forced a real count of public, parochial, and private schools. The next day Davidson and Davis told the Board of Public Education that attendance was down thirty percent, on the order of twenty-seven thousand absences. Of those, 6,070 pupils, about eight percent of the student body, had influenza. The rest were absent from other illness or from parental fear. Fifty-three students and two teachers were already dead of influenza or pneumonia. Burns still told the board that the five hours a child spent in school were the best place for them, and that Pittsburgh’s medical inspection was among the best in the country. The board voted not to close.

The vote did not survive the weekend in the form the board intended. On October 23 Davis met again with Davidson, Burns, and his bureau heads. Absenteeism was approaching forty percent. The reason for staying open, medical supervision of the children who were actually there, had collapsed if the children were not there. The encyclopedia says that on the next day, October 24, Davis ordered public, private, and parochial schools closed and told parents to keep children at home and off streetcars, streets, and playgrounds. A 2020 Post-Gazette narrative dates the school-closing order to October 23 and quotes officials telling children to keep off the streets. A Journal of the American Medical Association review of city interventions, the Markel study often cited for the national comparison, also uses October 24 for Pittsburgh’s school closure, against a gathering ban it dates to October 4. The one-day disagreement is small and worth keeping. The policy fact is not small. Schools remained open for about three weeks after the state had closed the saloons and the theaters. When they finally shut, the encyclopedia’s own authors argue, Pittsburgh’s social-distancing measures overlapped for only about two weeks. Layering that is late is not the same intervention as layering that is simultaneous.

The two death rates that are not one number

Pittsburgh’s claim to have been the deadliest large American city in the epidemic is a sourced claim, and it is not a single integer. The Michigan encyclopedia says the average excess death rate for Eastern cities was 555 per 100,000, and that Pittsburgh’s excess death rate was 807 per 100,000, worse than Philadelphia at 748 and Boston at 710. The Markel study in JAMA, using a comparative excess-death measure, reports Pittsburgh’s cumulative excess mortality as 807 per 100,000 and ranks the city forty-third out of forty-three cities in its study set, the worst burden in that table. Those two scholarly projects are pointing at the same order of result. They are not pointing at the figure 861, and they are not pointing at a raw count of 4,600.

The 861 figure has a different citation and a different window. The Post-Gazette’s 2020 reconstruction attributes to the Proceedings of the National Academy of Sciences an estimate of 861 flu-related deaths per 100,000 people in Pittsburgh between September 14, 1918, and May 31, 1919, and it uses that estimate to call the city’s Spanish-flu death rate the nation’s highest. The same article later calculates Pittsburgh’s flu deaths as an estimated 4,500 to 6,000 people, in a population it places between 534,000 and 588,000, which it converts to a range of 842 to 1,020 per 100,000. A separate Post-Gazette note on a small community’s records says most studies put the city’s deaths between 4,500 and 6,600, roughly one percent of a population it gives as 588,000. Pitt Med’s historical essay says an official tally of forty-six large cities put Pittsburgh among the worst, with a 1918 death rate of 25.4 per 1,000, more than double the year before, and that only Baltimore and Nashville had a higher per-capita rate in that particular comparison. In the same essay the phrase “nearly 4,600 Pittsburgh fatalities” appears beside a state figure of more than 35,000 dead in Pennsylvania.

None of these may be collapsed into “the” toll. Excess deaths per 100,000 are not a count of death certificates labeled influenza. A flu-related rate for a season running into May 1919 is not the same object as a 1918 crude death rate of 25.4 per 1,000, which includes every cause. “Nearly 4,600” is one essay’s compression. “4,500 to 6,000” and “4,500 to 6,600” are ranges that admit the certificates and the estimates do not match. This chapter can say, because the pages say it, that Pittsburgh sits at or next to the top of the major-city comparisons researchers actually published, that 807 per 100,000 is the excess-death figure in the Michigan encyclopedia and in the JAMA city study, and that 861 per 100,000 is a different flu-related estimate, for a stated window, reported by the Post-Gazette from the National Academy of Sciences proceedings. It will not pretend those are one statistic written down in a single health-department ledger and then copied badly. The deadliest-city sentence survives the distinction. The false precision does not.

Twelve ambulances, and the city they could not cover

The ambulance fact is usable because a newspaper reconstruction opened for this chapter attaches it to a contemporary paper. The Post-Gazette, citing the Pittsburgh Daily Post, reports that six of the city’s twelve ambulances broke down. They served a population the article places between 534,000 and 580,000. Calls came in hour by hour from every corner of the city, and nobody had time to repair the vehicles that were being driven to pieces. Twelve ambulances is not a metaphor for unpreparedness. It is a fleet size. Six broken is a maintenance failure inside a surge. A city that could barracks thousands of soldiers in Oakland could not keep a dozen motor ambulances on the street. That disproportion is the public-health system the epidemic actually met.

Hospitals overflowed into buildings that were not hospitals. The Post-Gazette lists emergency beds at Washington Park in the Hill District, room discussed for about three hundred; at Kingsley House, at Fulton or Fullerton and Bedford, a site now associated with the arena land, where the bed count is given as somewhere between one hundred and one hundred fifty; at the county courthouse annex on Ross Street, about three hundred; at the Marine Hospital in Arsenal Park at Penn and Fortieth; at the Concordia club in Oakland, up to one hundred cots; and in smaller conversions from Hays to Turtle Creek, where Westinghouse employees helped put up an emergency hospital in about forty hours. The Michigan encyclopedia adds the voluntary layer. The Red Cross sent seventy graduate nurses, forty practical nurses, and 172 aides and helpers into emergency hospitals, caring for close to three thousand patients, while also managing more than twelve thousand home cases. More than six hundred Mothers of Democracy organized brigades. Bishop Regis Canevin, on October 26, put Catholic resources at the health department’s disposal. The Irene Kaufmann Settlement took nursing in the Hill. The Free Burial Association buried poor Jews without charge. These are not ornaments. They are the evidence that the formal system’s twelve ambulances were only the motorized edge of a much larger, improvised care economy, and that the improvised economy was still losing.

Case counts moved faster than the ordinances. The Post-Gazette says the city had 4,291 cases by October 15. The encyclopedia says that on October 16 more than eight hundred new cases appeared and the reported total reached 4,445, and that Davis believed official reports were catching about half the truth. In one eighteen-hour stretch, from 4 p.m. on October 17 to 10 a.m. on October 18, the Post-Gazette records 746 new cases. Half of reality, if Davis was right, means every one of those headlines was a floor. A reportable disease is only as real as the reports. Pittsburgh had made influenza reportable on September 21 and still, a month later, had a health director saying the file was incomplete. The deadliest-city rankings sit on top of that incomplete file. They are the best comparisons the scholars can build. They are not a claim that every Pittsburgh death was written down the afternoon it happened.

A mayor, a commissioner, and an armistice parade

The fight between Babcock and Royer is the political chapter inside the medical one, and the encyclopedia documents it without needing anyone’s unspoken motives. Saloon interests lobbied Harrisburg to lift the ban. Royer called an early lifting criminal and then invited the mayor to talk. On November 1 he suggested Pittsburgh might reopen amusements as soon as November 4. He then moved the date to November 9. Babcock announced that the city would not act as the state’s police agent. The encyclopedia quotes him: the state was supreme, but the city would refuse to enforce, and Royer could try to quarantine Pittsburgh if government needs did not intervene. Royer called the proclamation an invitation to lawlessness. He told his Pittsburgh counsel, Nelson McVicar, to prosecute defiant owners. Most theaters and saloons hesitated. The ones that opened drew warrants, and McVicar later dropped cases to reward the ones that stayed shut. Royer lifted the state ban at noon on November 9, as he had reset it. Davis said schools could reopen on Monday, November 18, after medical inspection. The Post-Gazette’s looser formulation, that schools reopened gradually between November 11 and November 18, fits inside that date rather than contradicting the decision.

November 11 is the complication Davis could not inspect away. The war ended. Pittsburgh held a mass parade anyway. The Post-Gazette traces it from the Mon Wharf down Grant, Fifth, Liberty, and Sixth, and across the Sixth Street Bridge into the North Side, with the mayor at the front. Crowd counts in the papers were vague, recorded as thousands. People fainted in the crush. Speeches were abandoned because the noise was too loud. The encyclopedia and Pitt Med both treat the armistice gatherings, and an official civic celebration, as a plausible reason for a fresh spike of cases. A city that had spent five weeks arguing about whether a saloon could open put thousands of bodies on the same downtown streets because the war was over. The epidemic did not respect the headline. On December 5 the Pittsburgh Press could report that “only” twenty-four people had died in a day of the virus or its complications, while the Pittsburgh Post said thirty-four. Even the decline was an argument about a number.

Why Pittsburgh’s rate was so high is the question the encyclopedia refuses to answer with one villain. It notes that Pittsburgh’s closures began early relative to its own curve, earlier in the local epidemic than Boston’s, and that Babcock’s brief defiance is unlikely, by itself, to explain an excess death rate worse than Philadelphia’s. The differences it is willing to discuss are the staggered school closure and the air. Philadelphia and Boston closed schools earlier, so their measures stacked. Pittsburgh waited three weeks to add schools to a gathering ban, and the overlap was short. Modern studies cited there find that early, sustained, layered distancing changed the peak. The encyclopedia also notes Pittsburgh’s smoke. A population living under coal-fired mill pollution had higher pneumonia mortality in ordinary years, especially in the fall and winter inversions, and later research has associated heavy air pollution with pneumonia deaths. That is a hypothesis grounded in the city’s industrial atmosphere, not a laboratory proof that soot, rather than the virus, killed the 1918 dead. It belongs in a Pittsburgh chapter because the air was not a backdrop. It was a chronic respiratory burden sitting under an acute one. The mills did not adjourn for Royer’s order. War production was the reason the barracks were in Oakland. The streetcars kept moving, windows ordered open, aisles sprinkled with disinfectant, passengers still packed. A public-health system can close a theater and still be defeated by a shift change.

A city that stayed at work

The ordinances did not stop the reason Pittsburgh was crowded. War production continued. The barracks in Oakland existed because the Army was housing men next to the universities, and the Post-Gazette’s caption material puts seven of those buildings on the Pitt campus. Streetcars remained the way wage earners reached the mills. State rules, as the newspaper reconstructs them, told trolley companies to keep windows open and to disinfect the aisles, including the clumsy practice of sprinkling passengers’ legs along with the floor. A respiratory virus does not negotiate with a shift whistle. Men who could not be spared from a mill could not be protected by a theater closure. That is the limit of Royer’s order, and it is a limit he did not invent. A health commissioner can shut a saloon. He cannot, in October 1918, shut the steel district and the cantonment at the same time without a different kind of government than Pennsylvania had. The encyclopedia’s puzzle, why a city that acted early still died at a higher excess rate than Philadelphia, has this industrial residue in it. Early action against leisure is not early action against the workplace.

The volunteer system is the evidence of that limit, and the numbers are large enough to respect. Seventy graduate nurses, forty practical nurses, and 172 aides from the Red Cross, nearly three thousand emergency-hospital patients, and more than twelve thousand people nursed at home are the encyclopedia’s figures for the organized charity response. Six hundred members of the Mothers of Democracy, parish relief in all twenty-seven wards under Bishop Canevin’s offer, the Irene Kaufmann Settlement covering the Hill, and free burial for poor Jews are the same story at neighborhood scale. Pittsburgh was not passive. It was inventive and still short. Inventiveness that depends on nuns, clubwomen, and a settlement house is a compliment to those people and an indictment of the standing capacity. The twelve ambulances, six of them broken, are the motorized version of the same shortage. A city does not discover in a plague that it owns only a dozen ambulances. It discovers that the dozen was the plan.

Underreporting sits under every ranking. Davis’s statement that official reports caught about half the cases, recorded in the encyclopedia for the week when the counted total passed four thousand, means the 4,291 and 4,445 figures in the newspapers are floors. A death rate built from certificates is more stable than a case rate built from exhausted doctors, and still not perfect. People died at home. People were recorded as pneumonia. People were buried before a clerk agreed on a cause. The scholarly choice of excess deaths, 807 per 100,000 in both the Michigan file and the JAMA study, is an attempt to escape that clerk. Excess mortality asks how many more people died than the city’s ordinary seasons would predict. It will catch influenza that was filed under another name, and it will also catch deaths that were not influenza if the baseline is wrong. That is why this chapter refuses to translate 807 into “4,600 people” by arithmetic performed in the writing. The essays that say nearly 4,600, or 4,500 to 6,600, are making their own translation from their own sources. They are cited as translations. The rate and the count stay in separate sentences because combining them would manufacture a precision the health department of 1918 did not have.

What “deadliest” is allowed to mean

The year Pittsburgh became, in the comparative tables, America’s deadliest large city was not a year in which the city did nothing. Royer’s order was obeyed before the local case count looked like an epidemic, which is when Babcock called it too drastic. Churches, courts, elevators, night schools, and saloons were all touched. Volunteers built a second medical city in parks, settlement houses, and club rooms. The failure was specific. Day schools stayed open until the supervision argument died of absenteeism. The ambulance fleet was twelve vehicles, and half of them broke. The mayor and the state commissioner spent the first days of November fighting over who would police a reopening. The armistice put a crowd back on Fifth Avenue while new cases were still being counted in the hundreds. And the air the mills made did not become a variable in the 1918 ordinances, even though later writers looking at pneumonia in this city cannot leave it out.

The numbers that support the title are the comparative rates, used with their names attached. An excess death rate of 807 per 100,000, in the Michigan encyclopedia and the JAMA city study, put Pittsburgh at the bottom of a forty-three-city list, which is another way of saying the worst. A flu-related estimate of 861 per 100,000, for September 14, 1918, through May 31, 1919, is the Post-Gazette’s citation to the National Academy of Sciences proceedings. A body count of nearly 4,600, or of 4,500 to 6,600, is an estimate range in later Pittsburgh journalism and in Pitt Med, not a figure this chapter pretends to have totaled from certificates. Deadliest, in that disciplined sense, means the published comparisons place this city at the extreme. It does not mean a single clerk wrote 4,600 and 861 in the same cell and closed the book. The system that met the virus was a wartime industrial city with a small motor fleet, a late school order, a magnificent and insufficient volunteer nursing effort, and lungs already living in smoke. That is a public-health history. It is not a mystery about a germ that happened to dislike Pittsburgh.

Patterson’s name, the Oakland barracks, the late school order, the broken ambulances, and the two incompatible rates are enough to hold the title without a false ledger. Pittsburgh in the autumn of 1918 was a war plant and a cantonment that tried to govern a virus with closures aimed first at crowds at leisure. It gathered an extraordinary volunteer nursing effort and still watched its comparative death rates go to the bottom of the scholarly tables. Deadliest, attached to 807 per 100,000 excess deaths and to the separate 861 flu-related estimate, is a statement about those tables. It is not a statement that the city failed to notice the sick. Babcock noticed them and still called the first order too drastic, then later refused to police the state’s calendar. Davis noticed them and still kept the day schools open until absenteeism destroyed the reason for keeping them open. The virus did not need a more dramatic government than that. It needed a city whose lungs, streetcars, barracks, and workplaces were already a single system. They were.

Evidence Ledger

What remains disputed

The school-closure date is either October 23 or October 24, depending on whether a reader follows the Post-Gazette’s 2020 reconstruction or the Michigan encyclopedia and the JAMA study. The practical difference is one day. The three-week delay after the amusement ban is not disputed. The death figures are a family of measurements, not a corrupted copy of one measurement. Excess deaths at 807 per 100,000, flu-related deaths at 861 per 100,000 over a stated season, a 1918 crude rate of 25.4 per 1,000, and body-count estimates from nearly 4,600 to as high as 6,600 answer different questions. Population denominators in the same articles range from about 534,000 to 588,000. Ambulance counts are taken from the Post-Gazette’s reading of the Pittsburgh Daily Post, not from a motor-pool ledger reopened here. How much of the excess mortality is attributable to smoke, to the late school order, to wartime crowding, or to underreporting is explicitly uncertain in the best secondary account. Davis’s belief that official case counts caught about half the illnesses is a health director’s judgment, not a corrected census. The deadliest-city ranking is only as strong as the comparison set. In the JAMA set of forty-three cities, Pittsburgh is last, which those authors treat as worst. A different set, such as Pitt Med’s remark that Baltimore and Nashville had higher 1918 per-capita death rates of all causes, can reorder the podium without touching the 807 figure. Both statements can be true because they are not the same race.

This is an original public expansion of the short website chapter. It does not replace the Google Drive FULL MASTER.

Archival photographs relevant to this chapter. Sourced from Wikimedia Commons with attribution — not modern stock art.

Penn Central Station, Pittsburgh (5961021539)

Penn Central Station, Pittsburgh (5961021539)

Wikimedia Commons · Penn Central Station, Pittsburgh (5961021539).jpg · Hugh Llewelyn · CC BY-SA 2.0

Pittsburgh - Downtown from Mt. Washington (1963)

Pittsburgh - Downtown from Mt. Washington (1963)

Wikimedia Commons · Pittsburgh - Downtown from Mt. Washington (1963).jpg · Roger Wollstadt from Sarasota, Florida · CC BY-SA 2.0

LarimerSchool

LarimerSchool

Wikimedia Commons · LarimerSchool.jpg · Lee Paxton · CC BY-SA 4.0

Related Stories