University:
Boston University
Course:
SAR HP 151 | Introduction to Health Professions
Academic year:

2020
Views:

284

Pages:

19
Author:

Hope Palmer

Disease clusters: What’s real, what’s not, and how to tell the difference Outline • 1. What is a disease cluster? – Unofficial and official definitions • 2. What are some typical ways evaluate reports of clusters? – CDC Protocol • 3. What are some statistical methods? – Woburn and Cape Cod studies • 4. How do you tell what’s “real” Location and Disease • location = clue to cause • history (e.g., Snow’s Cholera studies) – Migrant studies (e.g., breast cancer in Japanese women who move to the US) • contemporary examples • • • • Lyme disease, infectious diseases (H1N1 ‘09) heart disease and “Mediterranean diet” cancer in Woburn, Tom’s River NJ, Birth defects, scleroderma, polycythemia rubra vera Terms of the debate Epidemiology 1995 Daniel Wartenberg Some of the issues • typical kinds of cluster studies – “too much cancer in my neighborhood,” “lots of kids with autism near the plastics factory” • epidemiologic aspects: – small-scale, low statistical power – non-communicable diseases (usually) – long latency (sometimes, e.g., cancer) • suspected environmental cause Cluster Epidemiology : the statistical analysis of spatial or spacetime distributions of disease, with the goal of “detecting clusters.” • local data compared to standard population • p-values; significance testing • significant clustering means a cluster is detected, or follow-up necessary • “pre-epidemiology” “What is a cluster?” • CDC (2008): • “a cluster of public health concern is defined by an unusual aggregation, real or perceived, of health events in time and space that is reported to a health agency” • Dictionary: • “a group of the same or similar elements occurring closely together; a bunch (like grapes)” [American Heritage Dictionary] • Street: • “if it looks like a duck, walks like a duck, and quacks like a duck, it’s a duck.” Cluster Modifiers apparent perceived putative actual suspected true possible reported circumstantial significant Source: Jacquez et al. Cluster Two basic concepts of “cluster” • Cluster1 (cluster-as-excess) : an observable excess of cases in space-time. • Cluster2 (cluster-as-causal-excess) : an observable excess of cases in space-time due to a cause of interest. Rothman “Sobering Start…” • Cluster studies: • • • • • small numbers inhomogeneous case definition biased ascertainment latency poor information on exposures, etc. • Conclusion “no new statistical methodologies are needed to refine our study of disease clusters…” Cluster Epidemiology vs. Spatial Epidemiology • Cluster Epidemiology – – – – – – statistical analysis p-values significance testing seeks to detect “clusters” issue of boundaries, etc. faith in statistical testing • “Normal” Epidemiology – data description – measuring disease incidence vs. location – identifying related causes – analogy: occupational epidemiology CDC Protocol (MMWR, July 26, 1990/39(RR-11);1-16) • 1. Initial contact and response – Gather information from concerned citizen • 2. Assessment – Preliminary calculations, case verification, literature review, feedback to advisory committee • 3. Major feasibility study – Assess potential study design, logistics, cost CDC Protocol (continued) • 4. Etiologic Investigation – Develop protocol and conduct study – Expected to “contribute to epidemiologic and public health knowledge” • Examples include Woburn childhood leukemia follow-up, Tom’s River childhood cancer study, Camp Lejeune studies, many others over past 20 years New England example • “A suspected cancer cluster is more likely to be a true cluster, rather than a coincidence, if it involves: – A large number of cases of a similar type . . . – A rare type of cancer . . . – An increased number of cases of a certain type in an age group that is not usually affected by that type of cancer.” New England example (cont.) • “statistical significance will determine if additional steps are required . . . The concerned individual or community is provided with an official letter. . . • “Because a variety of factors often work together to create the appearance of a cluster where nothing abnormal is occurring, most reports of suspected cancer clusters are not shown to be true clusters.” Case Examples • Childhood leukemia and contaminated drinking water in Woburn, MA • Breast cancer on Cape Cod and in Massachusetts Woburn Citizens’ Concerns • Years of toxic waste dumping from leather tanneries, chemical plants • Strong odors in East Woburn • Water tasted bad periodically • Apparent cluster of childhood leukemia in East Woburn neighborhood • Local and State officials unresponsive n o i t c A n e z i t i C • Created grassroots organization called FACE (For A Cleaner Environment) • Called for health and environmental studies • Demanded clean-up of contaminated properties • Several families filed lawsuit (subject of book and movie, “A Civil Action”) Woburn Childhood Leukemia Cases 1969-1979