Send letter to journal:
Re: Administrative data and occupational health research

Dear Editor,

We write to you with respect to our paper published in the Journal about the risk of occupational injury in foreign workers in Spain. Using the first available data in Spain on the subject, for the year 2003, the study found that foreign workers ran risks much higher than those of Spanish workers for both non-fatal and fatal occupational injury. The same data have recently become available for 2004, and we would like to briefly discuss the results, methodology used to obtain them, and some conclusions we believe to be important to the study of health and social issues such as these using secondary data.

Both anonymous data sets were obtained from the Ministry of Labour and Social Issues (MTAS), which maintains a registry of fatal and non- fatal occupational injuries in insured workers. Nationality was classified as Missing/Unknown (000), Spanish (724), or Foreign (any other value) according to codes employed by MTAS.

The results obtained are very clearly contradictory, showing an increase in the number of both non-fatal and fatal injuries in Spanish workers, and an enormous reduction in all injuries in foreign workers from 2003 to 2004. The reduction is such that for non-fatal injuries in 2004, foreign status appears to have a slight protective influence: relative risks (Spanish workers as reference) were 0.95 (Confidence Interval at 95%: 0.94-0.96) for non-fatal and 1.2 (0.9-1.4) for fatal. Furthermore, there is also a sizeable reduction in the number of cases where the nationality was missing, going from 19,689 such cases in 2003 to only 2,402 in 2004.

The data for both years are similar in many ways, such as sample size and proportion of men and women, and were treated in the same way both years. However, the data have one difference that is very relevant here. In 2003, the variable nationality included codes that were “invented”; that is, not part of the original coding scheme. As those codes neither indicated missing data (000) nor Spanish nationality (724), we included them as foreign workers. In 2004 there are no such erroneous codes; all those included correspond to a country listed in the coding scheme.

Obviously, such results cast doubt on the results obtained in 2003. We fully acknowledge that the decision to include the “invented” codes as foreigner workers is one that can be questioned. But we also see here a valuable opportunity to question and debate the reliability of currently available administrative data; as we can now doubt the results obtained from 2003, we could just as easily doubt the information from 2004. To echo thoughts from the original study, the data we have are limited both in scope and in quality. Though we are halfway through 2006, data from 2004 have only recently become available, and we will have to wait until 2005 data are released to truly evaluate the patterns of occupational injury studied here.

Such doubts are extremely important in a field heavily dependent on rigorously collected administrative data. These data are often the only feasible way to examine broadly certain aspects of population health and track changes over time. As public health researchers we have a responsibility to make use of and report on the data available to us. We also have a responsibility to acknowledge data limitations and advocate for improvement. With migration as an ever-increasing reality, in Spain and elsewhere, it is critical to stay abreast of the occupational health situation of those new arrivals living and working among us. To do so, we must have good data.

As we wait for further data, we will continue our study of immigrant occupational health through the use of other study designs, and hope that this issue will remain a pressing one in the public health field. Our results put into sharp relief situations critical to the understanding of worker health in Spain and in other European countries. Equally, they highlight questions that remain to be answered, not the least of which is the gaping hole in the information we have on foreign workers.

Sincerely,

Emily Q. Ahonen, MPH

Fernando G. Benavides, MD, PhD

Occupational Health Research Unit. Pompeu Fabra University (Barcelona)

References