Yesterday the first paper related to the National ALS Registry was released (here or here). The ALS Registry is a very difficult project. Because ALS is not a nationally notifiable condition, it is difficult to accurately identify every patient. This study aims to get an idea of the population of ALS patients by collecting data through to two different means. Both datasets are worthwhile. However, both are also imperfect. My concerns are not related to the long-term value of the Registry. My concern is that this study's conclusions are based on the idea that the combination of the two imperfect studies captures the entire population. There is absolutely nothing in the data that suggests this is the case. Any results based on this assumption are flawed at best, and disingenuous and reckless at worst. It is counterproductive to disseminate statistics that we know are wrong. The results are already being quoted by ALS organizations and the media. I cannot understand how the CDC or the ALS organizations can justify pretending that these results are representative of the US ALS population.
The database portion of the project began in 2006. It includes administrative data from four major sources. These sources combined cover 90 million patients. There are well over 300 million people in the US. Therefore, we know these databases will not cover every patient.