Archive for the 'Epidemiology' Category

Vital Statistics: the Puerto Rico edition

When President Trump visited Puerto Rico after Hurricane Maria, he noted that only 16 people had died and compared that to the death toll after Hurricane Katrina:

“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 be very proud of what’s taken place in Puerto Rico.”

This was an improbably low number and immediately there were some posts suggesting that looking at vital statistics records could clarify this:

Everything that’s been reported about deaths in Puerto Rico is at odds with the official count
Eliza Barclay and Alexia Fernandez Campbell | Vox
October 11, 2017
According to this article, the method in use for the 16 certified deaths:

. . . “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”

Methodology suggested by John Mutter, Columbia University

. . . 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.

Rather than waiting for a year for Puerto Rican deaths to show up on the CDC website, investigators went to Puerto Rico to look at the vital registration system. Here’s a sampling.

Estimates of excess deaths in Puerto Rico following Hurricane Maria
Alexis Santo-Lozada and Jeffrey Howard | SocArXiv Papers
November 21, 2017

Nearly 1,000 More People Died in Puerto Rico after Hurricane Maria
Center for Investigative Journalism
December 7, 2017

Official Toll in Puerto Rico: 64. Actual Deaths May be 1,052
Frances Robles, Kenan Davis, Sheri Fink, and Sarah Almukhtar | New York Times
December 9, 2017
And, this one wins the prize for the best graphics. The first graph shows the excess deaths over previous years; the second is a table that shows causes of death – not just drownings and electrocutions, which are typical for hurricane events.

excess death graph

table showing causes of death

Missing Data

No not the ‘9’ code in data, but data that no longer exist.

The FiveThirtyEight blog has a nice post on data on drug use that no longer exists.

Data On Drug Use Is Disappearing Just When We Need It Most
Kathryn Casteel | FiveThirtyEight blog
June 29, 2017

The main source researchers use to determine patterns of drug use is the National Survey on Drug Use and Health and it doesn’t track very well with heroin mortality statistics. But, it is the best data researchers have:

Several other sources that researchers once relied on are no longer being updated or have become more difficult to access. The lack of data means researchers, policymakers and public health workers are facing the worst U.S. drug epidemic in a generation without essential information about the nature of the problem or its scale.

The rest of the post describes some data sources that no longer exist, which were useful indicators of heroin use:

Arrestee Drug Abuse Monitoring Program [ADAM].
Estimates of expenditures on heroin

Expenditures on illicit drug use are nicely summarized in this article:

Cocaine’s fall and marijuana’s rise: questions and insights based on new estimates of consumption and expenditures in US drug markets
Caulkins, J., B. Kilmer, P. Reuter, G. Midgette | Addiction
May 2015

And, not mentioned at all in the FiveThirtyEight post are patterns of drug usage from wastewater analysis – sewage epidemioloogy:

Real-Time Wasterwater Analysis Shows What Drugs are being Used Where
Douglas Main | Popular Science
June 11, 2014

Creating Residential Histories

This is a report on the NCI/SEERS web portal on a way to create residential histories of respondents/decadents for epidemiological research. The report (below) details how three commercial vendors were able to match the residential history of a small sample of federal government employees. Also available are the algorithms and software to reconcile conflicting addresses. Interested folks might want to browse other tools/papers in the NCI Geographical Information Systems and Science for Cancer Control webiste.

NCI/SEER Residential History Project
David Stinchcomb and Allison Roeser | Westat
May 2016

SAS residential history generation programs [3 programs]
[Summary] [Link to programs]

Twitter and Public Health

The NIH Office of Behavioral and Social Sciences blog highlighted research by Ann Marie White and Melanie Funchess on the use of “Twitter data to prevent violence and suicide through community-based helping networks.”

Their recent BSSR talk, “To tweet or not to tweet: Community-based participatory research approaches to advance wellness and violence prevention via social media,” highlighted their progress in developing community-based helping networks that take advantage of social media tools to improve public health in their neighborhoods.

Changing Causes of Death in Poor Countries

The World Bank has a new interactive chart showing how the leading causes of death are changing worldwide:

From The DataBlog:

Worldwide, the leading causes of death are changing, and they vary between rich and poor countries. In low-income countries, deaths from communicable diseases such as malaria and HIV/AIDS have fallen, while deaths from non-communicable diseases such as stroke and diabetes are on the rise.

How We Die in Michigan

Julie Mack of put together mortality statistics from the Michigan Department of Health and Human Services and found some interesting trends.

A century ago, in 1914, 13 percent of people died from heart disease and 6 percent from cancer. That’s an era when contagious disease and infection killed many people at a much younger age.

In 1964, a half-century ago, after the introduction of antibiotics, heart disease and cancer together accounted for 55 percent of Michigan deaths.

In recent years, heart disease has been declining as a cause of death, while cancer has been on the increase.

Remember: Causes of death are a zero-sum situation. Since everybody dies, if one cause goes down, another must increase.

See also: Michigan’s top 10 causes of death.

Lifetime Risk of HIV Diagnosis in U.S.

The CDC released a new Fact Sheet showing the lifetime risk of HIV diagnosis in the United States.

From the press release:

CDC researchers used diagnoses and death rates from 2009-2013 to project the lifetime risk of HIV diagnosis in the United States by sex, race and ethnicity, state, and HIV risk group, assuming diagnoses rates remain constant. Overall, the lifetime risk of HIV diagnosis in the U.S. is now 1 in 99, an improvement from a previous analysis using 2004-2005 data that reported overall risk at 1 in 78.

Playing with Mortality Visualizations

Nathan Yau of Flowing Data has been doing some interesting (and beautiful) visualizations of when and how people die. First was Years You Have Left to Live, Probably. Next was Causes of Death. And today he posted How You Will Die.

10 Years After Katrina

Here is a round-up of some of the reporting on New Orleans 10 years after Katrina:

From FiveThirtyEight:

The Pew Research Center: Remembering Katrina: Wide Racial Divide over Government’s Response

Wonkblog: What happened when Brad Pitt and his architects came to rebuild New Orleans


Emily Oster of FiveThirtyEight examines various claims about the benefits of breastfeeding:

If one takes the claims seriously, it is not difficult to conclude that breastfed babies are all thin, rich geniuses who love their mothers and are never sick a day in their lives while formula-fed babies become overweight, low-IQ adults who hate their parents and spend most of their lives in the hospital.

The truth is complicated.