Members of Congress have requested a formal audit of the death toll in Puerto Rico, citing concerns about the “woefully underreported” number of people who had died in the aftermath of Hurricane Maria. [MORE]

From [VOX] Death tolls are the primary way we understand the impact of a disaster. And for nearly two weeks after Hurricane Maria hit Puerto Rico, as a humanitarian crisis was intensifying, the death toll was frozen at 16.

“Sixteen people certified,” Trump said on October 3 during his visit to the island, repeating a figure confirmed by the territory’s governor. "Everybody watching can really be very proud of what's taken place in Puerto Rico."

It was a moment that crystallized two conflicting narratives about the Puerto Rico disaster. The first one, from the federal government and Puerto Rico’s governor, is of a disaster that’s been managed well, with lives being saved and hospitals getting back up and running.

Lives surely have been saved in the response. But images and reports from the ground tell a story of people, cut off from basic supplies and health care, dying. They tell of hospitals running out of medication and fuel for their generators and struggling to keep up with the “avalanche of patients that came after the hurricane,” as one journalist put it.

The death toll from the hurricane is now up to 45, according to Gov. Ricardo Rosselló. But 90 percent of the 3.4 million American citizens on the island still don’t have power, and 35 percent still don’t have water to drink or bathe in. And given how deadly power outages can be, 45 deaths seems low, according to disaster experts.

At Vox, we decided to compare what the government has been saying with other reports of deaths from the ground. We searched Google News for reports of deaths in English and Spanish media from Puerto Rico since Hurricane Maria. We found reports of a total of 81 deaths linked directly or indirectly to the hurricane. Of those, 45 were the deaths certified by the government. The remaining 36 deaths were confirmed by local public officials or funeral directors, according to the reports. We also found another 450 reported deaths, most of causes still unknown, and reports of at least 69 people still missing.

The broader issue here relates to how storm deaths are counted. There are clear deaths from the storm, clear deaths indirectly from the storm, and then deaths that are harder to determine — for instance, a sick patient who died in a hospital experiencing frequent power outages. And then there’s the issue of how effective authorities are at finding and investigating the deaths to make sure they’re included in the count. The breakdown of these categories suggests that the government is being much more cautious in designating deaths as directly or indirectly hurricane-related, given the public information available.

At a Sunday news conference, Karixia Ortiz, press officer for the Department of Public Safety, said that “every death must be confirmed by the Institute of Forensic Science, which means either the bodies have to be brought to San Juan to do an autopsy or a medical examiner must be dispatched to the local municipality to verify the death,” according to an audio recording obtained by Huffington Post.

John Mutter, a disaster researcher at Columbia University who studied the death toll in the aftermath of Hurricane Katrina, says he’s skeptical of this methodology. “This is the way to go about it if you want to come up with smallest number possible,” he said, adding he suspects the death toll in Puerto Rico from Maria should already be in the hundreds based on what’s known about the conditions on the ground.

Our review of reports certainly suggests the real death toll is far higher than what the government has, thus far, estimated:

• In our search of local and US news reports, we found 36 deaths attributed to the hurricane in addition to the official 43. We cross-checked news accounts with the official death reports to make sure they didn't overlap.
•  NPR reported an additional 49 bodies with unidentified cause of death sent to a hospital morgue since the storm.
• The Los Angeles Times reported 50 more deaths than normal in one region in the three days after the hurricane.
• Puerto Rico's Center for Investigative Reporting reported 69 hospital morgues are at “capacity.” Exact figure is unknown.
• According to El Vocero newspaper, 350 bodies are being stored at the Institute of Forensic Sciences (equivalent to the state medical examiner's office), many of which are still awaiting autopsies. In the report, Héctor Pesquera, secretary of the Puerto Rico Department of Public Safety, did not say how many, if any, of the cadavers were there before the storm. (On Sunday, Pesquera denied claims that there was a backlog of unexamined bodies.)

“I don’t think there will be hundreds of deaths, but we will see,” Pesquera told reporters on Sunday. “We can’t speculate if there will be 100 or 200.”

Meanwhile, the situation on the ground remains life-threatening in some areas. And reporters and first responders are continuing to paint a much more aggressive picture about life and death on the island.

"It's horrible, it's horrible. It's a nightmare," a resident of the town of Atlaya told CBS on Tuesday. "There's barely any drinking water, not even in supermarkets; my fear is for my kids," another said.

Given the disparity in storylines, it’s worth taking a hard look at the numbers.

The best way to count storm deaths, according to a disaster researcher

There are no state or federal guidelines in the US for calculating storm death tolls for the medical examiners usually responsible for determining what constitutes a storm-related death. (And partly because many storm-related deaths aren’t recorded by the systems in place, the “official” Hurricane Katrina death toll is widely regarded as inaccurately low.)

Because we’ve had trouble reaching officials in Puerto Rico, it’s been difficult to decipher what exactly the process is for documenting and attributing deaths to disasters. But it is clear that the hurricane has disrupted it and that the government is now insisting that every body be inspected directly by the Institute of Forensic Sciences in San Juan before any death is attributed to the storm.

Mutter, the disaster researcher at Columbia, said that it’s very difficult after the fact to separate out deaths that would have happened anyway. “If they’re dismissing ones that would have happened anyway, that’s cheating,” he said.

The ideal way to calculate the death toll from the storm, he says, is to count all the deaths in the time since the event, and then compare that number to the average number of deaths in the same time period from previous years. Subtract the average number from the current number and that’s the death toll.

“When I first started hearing the deaths were only 16, and then 34, I thought there was something wrong,” he said. “Maria was bigger than the two previous storms, Harvey and Irma. And there’s no way to evacuate an island. All those people are still there. And then you look at damage and it’s profound. And now they’re saying only 45 people died, you’re saying come on, it couldn’t be.”

We found credible reports of additional hurricane-related deaths

In a review of local and national news reports that cite local officials and funeral directors around Puerto Rico, Vox identified at least 36 people who may have died in connection to the hurricane who are not accounted for in the official tally.

In some of these cases, local public officials named the victims and gave specific details. In other reports, details were scarce. It’s not clear why these deaths were not included in the official death toll. (A media representative for the governor’s office did not respond to an inquiry from Vox.)

• The mayor of Toa Baja told El Nuevo Día newspaper that at least nine people died after the storm. One man was dragged by a flood current, and eight others drowned. Residents confirmed these deaths to reporters and named some of the victims. The official death count only includes one drowning in Toa Baja.
• In Añasco, at least four people drowned as they tried to rescue people after the storm.
• In Canónavanas, the mayor reported that two elderly people died from panic attacks. The official death count only mentioned one person who died from respiratory problems in this town.
• The director of federal programs for the town of Barceloneta told Primera Hora that three sick people had recently died, including two with cancer, because they didn’t have access to the proper medical care after the hurricane.
• In Lajas, an elderly man died because he couldn't get oxygen at a local shelter.
• The mayor of Caguas told the Washington Post that a diabetic person died in a hurricane shelter because he didn't have access to medical care, and two other people killed themselves.
• Two funeral home directors in the town of Jayuya told a BuzzFeed reporter that the hurricane had directly or indirectly killed at least 18 people there in the past two weeks, including several people who died because they couldn’t get enough water, oxygen, or dialysis treatments. Only one of these deaths — a man who died in a landslide — was included in the government death count.
• The mayor of San Juan told CNN that two people died in a local hospital's intensive care unit because it ran out of diesel fuel. The official death count only mentions one death in San Juan from lack of access to medical care.

There are reports of another 450 who died of undetermined causes

News reports cite the deaths of more than 450 additional people since the hurricane, and 69 people have been reported missing. The additional deaths could be people who died as a direct result of the hurricane, or indirectly from the hurricane, or people who would have died even without the storm. For example, we found a report of one person who died because she didn't have enough oxygen tanks — this would count as an indirect death.

Here’s how we came up with figure of at least 450:

• An NPR reporter shadowed doctors at the Pavia Arecibo Hospital in Arecibo this week, where the lack of air conditioning was exacerbating patients' health problems. The hospital administrator told NPR there are 49 bodies at the hospital morgue from deaths after the storm. It's unclear how many were directly or indirectly linked to the storm.
• A reporter from the Los Angeles Times recently visited Lajas — a rural town on the southwestern side of the island — where elderly people can't get access to enough oxygen and insulin. The local funeral director said that at least 100 people had died in the area within three days of the storm’s passage, which is 50 percent higher than the area's normal death rate.
• Puerto Rico's medical examiners' office doesn't have enough staff to examine the cadavers or enough room to store them. Some 350 bodies were said to be at the Institute of Forensic Science, but it wasn’t clear how many were there before the storm.
• A reporter for the Center for Investigative Reporting in Puerto Rico spoke to doctors in half a dozen hospitals who said bodies are piling up at the morgues of the 69 hospitals in Puerto Rico. The majority of the hospital morgues that provided information are at full capacity. [MORE]