Questions and Answers: World Trade Center Responders Fatality Investigation Program

What is the World Trade Center Responders Fatality Investigation program?

The WTC Responders Fatality Investigation program was the data collection center for fatalities occurring among responders, workers and volunteers during and after the WTC disaster.

Why did you conduct this study?

Although numerous studies have identified health effects among those who were at the WTC site, reports of deaths occurring among those who were at the site were limited to news media reports and medical examiner/pathologist opinion. The National Institute for Occupational Safety and Health (NIOSH) requested that the New York State Department of Health conduct a study examining whether responders and workers were at a higher than normal risk for certain causes of death. The WTC Responders Fatality Investigation program was designed to be an early look at this question.

Who was included in this study?

Any WTC responder, worker or volunteer who died between September 12, 2001 and June 30, 2009. Responders included any person who responded to the WTC disaster or worked or volunteered at Ground Zero, the secure/exclusion zone, and/or the morgue or waste stream corridor including the Fresh Kills landfill in Staten Island between September 11, 2001 and June 30, 2002.

Why was there an end date for this study?

This research was funded by the National Institute for Occupational Safety and Health for a limited time period. It was meant only to be an early look at whether there were any groups that appeared at high risk for dying from specific causes. Other studies were designed to look at long-term consequences from the disaster.

How did you identify people who worked at the World Trade Center and subsequently died?

Because there was no central method to identify these responders, identifying the deaths occurred through a variety of methods. Daily reviews were conducted of on-line newspapers in New Jersey, Connecticut, New York, and eastern Pennsylvania, and of national obituary search engines using key words such as "ground zero," "WTC," "World Trade Center" and "9/11." Outreach and education was provided to over 200 groups about the project and the need to provide information on any WTC responder death. These included groups that were involved with the medical monitoring, treatment or research of WTC responders, worker advocacy organizations, labor unions, businesses that worked at the WTC site, and attorneys involved in lawsuits regarding WTC. Agreements were developed for data from FDNY, the WTC Health Registry, and the WTC Medical Monitoring and Treatment program to be shared with the WTC RFI program. A toll-free number and email account specific to this project were also established.

How many deaths were identified?

There were 836 deaths identified. The cause of death was confirmed for 814 of these. We were not able to find any information about the other 22 names identified that either confirmed the person was deceased or was present at the WTC sites.

How were deaths/cause of deaths confirmed?

An attempt was made to collect a death certificate for every WTC responder, rescue worker and volunteer, who died since September 12, 2001. Medical records, employers and family members also provided information.

Were you able to identify all deaths that occurred among the worker population?

Because we do not know who was present at the site, we do not believe that all deaths were identified. Although extensive outreach was conducted to find the responders, it appears that at most, only 47% of the deceased responders were identified using the methods described above. This estimate is based on using a statistical technique called capture-recapture analysis where a population size is estimated based on the number of reports received from different sources.

How did you decide if there were any causes of death that may be associated with exposure at the World Trade Center?

A statistical technique called "standardized mortality ratios" (SMRs) was conducted. This compares the number of known deaths to expected deaths. For this study, we compared the rates of deaths among the responder population to three different groups: the rates that occurred in the general United States population; the rates that occurred in the 16 counties surrounding and including NYC where 61% of the identified responder deaths occurred; and the rates that occurred in New York City. If the SMR is equal to 1.0, then this means the number of observed deaths equals that of the expected deaths; ratios higher than 1.0 indicate a higher number of deaths than expected.

We also used a technique called "proportionate cancer mortality ratios" (PCMRs) to look at the relative effect of each cancer type among all cancers. These are interpreted similarly to the Standardized Mortality Ratios.

The Standardized Mortality Ratios are all very low. Is there a reason for this?

There are many reasons suspected as to why the results are so low. The rate of death is calculated by dividing the number of known deaths by the number of people in the population. In this case, we believe that we have identified less than half of the known deaths which would decrease the rate. Likewise, because there was no record of who worked at the WTC site, we also do not know how many people were there. The rate will also be lower if the estimate of the number of people who worked at the site is too large. To look at whether this was the case, we did use a smaller number of total response workers, but the results did not change.

Another reason why the results may be so low is due to the "healthy worker effect". This occurs because the mortality rates in a working population are lower than the general population. This is primarily due to the fact that severely ill and disabled people are often not employed; and people who are employed may have access to better healthcare than those not employed, keeping them healthier. The choice of an inappropriate comparison population can contribute to the Healthy Worker Effect. Ideally, we would use a comparison population that consisted of workers similar to the people who worked at the WTC site, but we do not have death rates for a group like that. To partially control for this, mortality rates from three comparison populations were used. It was difficult to compare results across these populations due to the change in the number of deaths that occurred in each cause, and the low number of deaths for some of the comparisons may have contributed to rate instability.

What were the results of the study?

Because we were not able to identify most of the deaths, it is difficult to interpret the results. We do not know who we're missing and what they died from. The Health Worker Effect probably has a minimal effect on whether or not people develop cancer. This is why we looked at cancers using the PCMR technique. Using this technique can help us to better focus on identifying health outcomes that may need more study. But because so many deaths were missed, this technique is not useful or reliable in measuring the magnitude of the risk. Because the time period from the disaster to the end of this study was so short, only those cancers with high fatality rates and short latency periods (the period of time between the start of the disease and when symptoms become apparent) would be relevant. This study did identify two cancers where cancer rates in workers were higher than we expected – ovarian cancer and multiple myeloma. However, our analyses did not identify a common source of exposure among the workers who died from these cancers. Therefore, for the population of workers that we studied, we do not believe deaths from these cancers were related to exposures at the WTC.

What are the next steps?

The results of this study show that for future studies of this kind there needs to be identification of the majority of responders by all participating responder organizations. Since this did not occur at the WTC, we recommend that the currently established WTC medical monitoring programs and the WTC Health Registry match their populations to death certificates on a periodic basis to examine whether there is an increased risk among their cohorts. Because there is no appropriate external population to use for comparisons, comparisons within the cohort based on levels of exposure at the WTC site should be used wherever possible to control for the healthy worker effect. These studies should conduct specific analyses of those causes of death that were elevated in this project, even if they were not statistically significant to ensure that the data limitations were not masking a true risk.