In September 2015, pediatrician Mona Hanna-Attisha made national headlines when she reported a near doubling of children’s lead levels in Flint, Michigan after the city’s water source was switched from Lake Huron to the Flint River. Her research findings triggered a public health crisis that remains unresolved to this day.

In April 2016, Matt O’Brien of the Associated Press, reporting on an analysis of U.S. Environmental Protection Agency (EPA) data, found that the drinking water system of Providence, Rhode Island was one of the largest in the country to repeatedly exceed federal lead standards. Except for some recent rumblings by Kate Nagle in GoLocalProv and Sam Bell on ABC6’s In the Arena, the story seems to have come and gone.

In a brochure distributed to its customers, Providence Water has offered several tips to reduce lead in tap water, such as running the water rather than drinking what first comes out of the faucet, and using only cold water for cooking and preparing baby formula. The agency’s most recent annual report for 2015 touts that it cleaned and lined 12.3 miles of water mains and flushed an additional 100 miles. Still, its annual water quality reports continue to show lead concentrations that exceed the “action levels” established by EPA’s 1991 Lead and Copper Rule.

So I decided to ask Providence Water for its database of lead water measurements and see for myself where the city stands. Parts of Providence, I find, have indeed had as much lead in their tap water as Flint, especially the area around Ward 2 on Providence’s East Side. The lead statistics reported by Providence Water obscure the seriousness of the problem by combining Providence with neighboring Cranston, Johnston and North Providence. While there is some evidence of an improvement during the past 2–3 years, Providence still appears to be far from solving a serious public health problem.

Lead in Drinking Water

Infants, young children and pregnant mothers are particularly susceptible to lead in tap water. The most important source of lead exposure to infants is baby formula reconstituted with contaminated tap water. Canadian researchers recently found that each additional 1 part per billion (ppb) of lead in drinking water raised the blood lead levels of 1-to-5-year-old children by 35 percent. One researcher has reported increased fetal deaths during the Washington DC drinking water lead crisis of 2000–2004.

To comply with the Lead and Copper Rule, the EPA requires that community water systems perform regular measurements on samples of tap water at multiple homes considered to be at high risk of contamination. The samples are supposed to be first draw, that is, straight out of the tap after the water has sat in the pipes for at least 6 hours. In its evaluation of the measurements, the EPA focuses not on the average lead level, but on the 90th percentile, that is, the cutoff value that exceeds the bottom 9 out of 10 measurements but falls below the top 1 out of 10 measurements. While there is no scientifically accepted safe level of lead in water, the EPA has established an action level of 15 ppb. If the 90th percentile lead level exceeds 15 ppb, the water system has to take a series of corrective steps, including corrosion control treatment, source water monitoring, public education, and lead service line replacement.

The 90th percentile calculations mandated by the EPA are certainly not the most natural way to analyze data on lead levels in drinking water. So many researchers prefer simply to report the proportion of lead measurements that exceed the 15-ppb threshold. If that proportion is greater than 10 percent, then the EPA action level has been exceeded.

Flint Water

During 2015, concerned residents of Flint began collecting their own water samples. The Flint Water Study’s initial analysis of 252 first-draw samples during Aug.-Sep. 2015 showed a 90th percentile lead level of 25 ppb, well above the EPA action level of 15. Here is a map showing the variation in lead levels by ward within the City of Flint. The blue numbers, which show the proportion of first-draw samples over 15 ppb, exceeded 10 percent in every ward but Ward 1. The highest blood lead levels in Flint children, Dr. Hanna-Attisha reported, were in Wards 2, 5 and 6.

In an update issued in March 2016, the Flint Water Study reported some improvement in first-draw lead levels. The 90th percentile level had dropped from a revised value of 28.5 ppb in 2015 to 22.8 ppb in 2016, but remained well above the EPA action level of 15 ppb.

Providence Water: The Official Data

Let’s start with the official figures from the annual water quality reports issued by Providence Water since 2010.

90th Percentile Lead Concentrations Reported by Providence Water Report Year

90th Percentile (ppb)*

2010 ** 30 2011 21 2012 25 2013 30 2014 16 2015 15 2016 16 * The EPA action level is 15 ppb. ** The 2010 Water Quality Report refers to sampling performed during 2009.

Up to 2013, the table suggests, Providence’s numbers were just as high as Flint’s. But during 2014–2016, things had improved to the point where Providence was just at or slightly over the EPA’s action level.

There were, however, three critical problems with the official data.

Not Just Providence

The first problem was that the official numbers included three cities served by Providence Water other than Providence itself — namely, Cranston, North Providence and Johnston. In fact, of the 2,479 first-draw samples with a street address that were recorded in the database from 2012 through May 2017, only 884 (or 32 percent) were located in Providence proper. The inclusion of Cranston, North Providence and Johnston in the official numbers made the lead-in-water problem certainly look a lot less menacing. For first-draw samples for all years combined, the 90th percentile level for Providence and the three other cities was 14.5 ppb. But when I broke it down, the corresponding 90th percentile levels were 21.5 ppb for Providence and 13 ppb for the other cities.

What About the Other Measurements?

The second problem was that the database included numerous measurements besides the first-draw samples. For addresses within the city of Providence, nearly one-quarter of the measurements were not coded simply as “1STDRAW.” Some of them included 5-minute, 10-minute and even 20-minute flushes. While one would expect these additional measurements to show lower lead levels, they frequently did not.

The graph below shows the trends within the city of Providence in the 90th percentile lead levels for first-draw samples and all samples. In 2016, the 90th percentile for first-draw samples was 20 ppb, well above the figure reported by Providence Water for that year. But the 90th percentile for all samples was 36 ppb. In fact, 28 percent of the samples in Providence in 2016 were over the cutoff of 15 ppb, a statistic comparable to the values seen in Flint’s Wards 2, 5 and 6.

Ward 2 on the East Side: The Most Contaminated of Them All

The map displayed at the top of this article shows the breakdown by ward within the city of Providence. In order to get statistically stable results, I have included all measurements — whether first-draw or flush — from 2012 to May 2017. The idea was to produce a map that could be compared to the one drawn above by the Flint Water Study. The most appropriate comparison is between the red percentages in the Flint map and the black percentages in the Providence ward map below. While Providence Wards 1, 13 and 14 have had lead profiles comparable to the highest risk wards in Flint, Providence Ward 2 on the East Side has been the most contaminated of them all.

What About the Water Mains?

In Flint, the water source was temporarily switched from Detroit-supplied Lake Huron water to the Flint River in anticipation of a new line from Lake Huron. In Providence, by contrast, the lead supposedly comes not from the ultimate water source, but from water mains, local feeds from the water mains to individual houses, and lead-laden fixtures within the houses. Of these three contributors, the water mains may be the most critical. There are 169 first-draw measurements in the Providence Water database with no recorded address. For these measurements, the 90th percentile lead level was 44 ppb. Are they the water mains?

Providence Water: Has There Really Been Any Improvement?

Despite numerous attempts to parse the raw data this way and that way, I could not reproduce the official numbers reported by Providence Water. One peculiarity that I encountered in attempting to replicate the official results is that the database is loaded with identical repeated measurements recorded at the same address on the same date. Should these measurements be considered separate observations in calculating the 90th percentile? Or should they be collapsed into a single value?

This leads me to the third — and most fundamental — problem with the official data. You can’t be sure there’s really an improvement in lead levels unless you go back and sample the same homes year after year. If instead you sample the most contaminated addresses in 2010–2013 and then sample the least contaminated sites in 2014–2017, you’re going to get some impressive graphs that really don’t mean anything.

I was able to locate 26 homes in the database that had been sampled in each year from 2014 through 2017. For this small sample, the respective 90th percentile levels did show a downward trend: from 21.4 ppb in 2014, to 19.6 ppb in 2015, to 17.3 ppb in 2016, to 11.5 ppb in the first five months of 2017. The problem, of course, is that nobody is going to conclude that Providence has finally begun to tackle its lead problem on the basis of 26 homes in a city of thousands. And there just aren’t enough data in 2017 to draw any firm conclusions yet.

Where Does Providence Stand?

Back in 2012, the RI Department of Health and the EPA entered into a consent agreement with the Providence Water Supply Board that gave the Board extra time to clean up its lead problem. As part of the agreement, the Board had to convene a national panel of experts. According to the Board’s most recent audited financial statement for fiscal 2016, further extensions were granted in succeeding years. A page on its website entitled “Where Does Providence Stand?” states, “Providence Water is currently negotiating a Consent Agreement with the RI Department of Health.”

Maybe we should abandon our annual ritual of giving Providence Water more time to clean up its lead problem. Maybe it’s time for a private group of citizens, such as Flint’s Concerned Pastors for Social Action, to seek redress in the courts. Maybe it’s time for the Mayor of the City of Providence to get his back up off the wall and assert some leadership.