The expert who blew the whistle on the Flint water crisis says the only way to protect the nation's school children against lead in drinking water is regular testing of virtually every fountain or sink they might use during the day.
But an NBC News survey of the country's 20 biggest cities shows that very few school districts have met that standard.
Some, like Los Angeles, have not done district-wide testing in years; others, like Houston, have tested only a handful of schools. The uneven approach could leave hundreds of thousands of kids at risk of exposure to a toxin that can stunt intellectual development and cause physical and behavioral problems.
"As long as we have lead in the plumbing, this poses a health risks to students," said Marc Edwards, the Virginia Tech professor who uncovered elevated lead levels in Washington D.C. in 2003 and in Flint, Michigan, last year.
"Schools need to be testing to protect the students."
Chicago - which has a lead-poisoning rate at least twice the national average - had no record of testing its 600-plus schools until launching a pilot program with just 28 buildings last week, and it has not gotten the results.
Chicago schools spokeswoman Emily Bittner said buildings were not tested in the past because there's no state or federal requirement and because Chicago's municipal water supply meets all standards. The furor over Flint's lead levels was the impetus for the new effort, she said.
"It's part of a national conversation," Bittner said. "It's to reassure people."
It's true that schools that rely on a municipal water supply and don't run their own system are not required by federal regulators to test what comes out of their taps for lead; efforts to mandate such testing have failed in Congress.
But Edwards points out that just because the water that leaves a city treatment plant is lead-free doesn't mean it won't contain lead when it pours from a faucet in a kindergarten classroom, after traveling through plumbing systems.
Even though most schools aren't connected to pure lead service lines, any plumbing installed before a 2014 lead-free law could contain some of the heavy metal - in the solder or fixtures, for instance - which could leach into the water as the result of corrosion or aging, he said.
The link between lead and aging infrastructure means that one-time testing isn't enough, Edwards said. A school that was deemed safe a decade ago could yield elevated lead today as pipes corrode and disgorge loose fragments into the water, he said.
"These problems are getting worse with time, not better," he said.
There is no safe level of lead exposure, experts say, and young children are most vulnerable to its effects. The Environmental Protection Agency says action should be taken any time a level higher than 15 parts per billion is found in a public water system or 20 ppb from a school fixture.
Across the country, cities big and small that have tested schools for lead in the wake of the Flint scandal have gotten troubling results. In Newark, New Jersey, a lab found elevated levels in almost half its schools in March, and fountains were shut down. Last week, in Howland, Ohio, four out of six schools tested above the limit.
Here's how the 20 most populous cities in the nation have handled testing:
In the country's biggest school system, with more than 1 million students, testing is ongoing since 2002. Of about 89,000 samples, 1.13 percent exceeded EPA guidelines on first-draw samples, and that dropped to less than a tenth of a percent on second draw, after the faucet was running to eliminate stagnant water. Remediation includes flushing and equipment replacement. The city removed all lead service lines to schools between 2008 and 2010.
The Los Angeles Unified School District, which has 640,000 students, tested 850 sites in 2008. At least 11 percent of drinking fountains had lead above the action level, but officials say only 2 percent were still too high after flushing protocols were used - and those were shut off. Because of budget issues, it wasn't until a bond approval last year that LA launched a cleanup plan, including pipe or fountain replacement, for the other 9 percent; it won't be complete for another 18 months. The school district plans to start random testing next year, saying it would be too expensive to do them all again.
The city, which has 396,000 students enrolled in public school, began its first documented testing last week, taking samples from 28 buildings in what it referred to as a "pilot project." Once those results are analyzed in mid-May, Chicago schools will "develop a comprehensive lead testing approach," according to City Hall, which indicated in a press release that it's banking on proposals for new federal testing programs to take up the slack.
The city has 283 public schools, with 215,000 students, but has tested just five of them. None of the sites, which were chosen at random, had lead levels above the action level, officials said. At one school, the amount of lead in fountain water was higher than in other parts of the building so the equipment will be replaced.
More than 20,000 outlets in 308 buildings were tested for lead between 1999 and 2009. Any that tested elevated were subject to remediation - including plumbing replacement - and repeatedly retested until they were normal. No testing has been done since the initial $5 million program, but the city, which has 134,000 public-school students, plans to do some random testing this year.
The city leaves testing up to 33 school districts. Deer Valley Unified School District, which has 33,927 students and 38 schools, said it tests only the one that has its own water supply and the results were normal. The Pendergast district (10,000 students on a dozen campuses), Paradise Valley (32,000 children in 42 schools) and Alhambra (14,200 students in 15 schools) reported they do no testing.
The San Antonio Water System said none of its residential testing last year showed elevated lead levels but that it's up to schools districts to check educational buildings. Northside Independent School District (104,000 students), San Antonio Independent School District (53,000) and South San Antonio Independent School District (10,000) students, reported they do no testing. The North East Independent School District, which had 68,000 students, said it has not tested but is planning a random survey "to alleviate people's fears."
The San Diego Unified School District, which encompasses 130,000 students at 226 facilities, does not do random or routine testing because the city has eliminated lead service lines and the schools have lead-free plumbing, a spokesperson said. About 20 years ago, there was a district-wide sampling, and no results were over the limit, though the records were not archived.
The school district, which has 160,000 students on 227 campuses, does not routinely test for lead in the water but will test when it believes there is an issue, said spokesman Andre Riley. Since Jan. 2015, they have tested at least four schools, all of which had results well below the action level, Riley said.
Since 2006, every school is tested every year at two sites, usually one faucet and one water fountain. The 2016 results have not come in but last year, of the 40 schools tested, none had results over 15 ppb. Starting next year, all schools will be tested twice.
The Austin Independent School District, which has 83,000 students at 130 schools, said it tested 47 schools between April 12 and April 27 and "all were within the acceptable level so there is no cause for concern."
The city has not tested since 2000. Between 1988 and 2000, it tested every drinking water outlet in 150 schools, and 40 schools had results above the action level. "All non-compliant fixtures and sources were eliminated," Duval County Public Schools said. "Schools constructed after the year 2000 have met all current specifications for lead free piping, fittings and fixtures."
The city has 11 school districts. The largest, Indianapolis Public Schools, with 30,000 students, said it tests "periodically," but did not provide specifics. Two schools have been tested "in recent months" and both were well below the action level, a spokeswoman said, adding that lead piping was removed in the district. The Metropolitan School District of Lawrence Township, which has 15,000 students, tested 20 buildings between February and March and said it got negative results on all.
San Francisco Public Schools, which enrolls 57,000 students at 156 schools, said its last district-wide test for lead was "a few years ago," but did not provide an exact time. Five schools are currently out of compliance and have been placed on a bottled-water program until renovations and retesting are complete.
All schools in the city, which has 51,000 students, were tested in 2008. Sixteen of 122 schools had results above action level. New water coolers were installed and the water was retested.
The school district, which has 86,000 students in 130 schools, says "no need for testing for lead has been identified to date." One school, Arlington Heights High School, has been part of the city's testing program since 2009; the 2015 result from that sample was 6.2 ppb, below the action level.
Charlotte-Mecklenburg Schools, which has 146,000 students, said testing is regularly done at only one school, which has its own water system; the last results there, in August 2015, found no elevated lead. For the other 167 schools in the city, "there may be occasions we test for lead levels due to parent/school concerns but on these few occasions the results were negative," the district said in a statement.
Detroit Public Schools, home to about 52,000 students, took first-draw and after-flushing samples from three sites at each of its 62 elementary schools this spring; 19 schools, nearly a third, had results over the action level. The highest was 1,500 ppb, or 10 times the cutoff. All those schools are on bottled water while retesting and remediation is done, and testing of the city's 38 middle and high schools is now underway.
The city has 60,000 public school students on 94 campuses. It tested 10 of them in 2009 and none were above the EPA action level. They tested 15 this year and none required action, although one sample came back at 10 ppb and the district has asked the school to investigate.
Since 2004, each of the city's 98 public schools - where 54,000 students are enrolled -- have been tested at least every three years, and measured against a cutoff that is more stringent than the EPA's. In the most recent results, 99.5 percent of samples from the first-draw of water were lower than 15 ppb and 99.7 of samples taken after 30-seconds of flushing were below the EPA threshold. Edwards said that Seattle's program "is a model, among the best in the country."