Posts Tagged ‘epidemiology’

Applied quantile regression in environmental health

This week, Sheryl Magzamen and colleagues published a report of early life lead exposure and educational outcomes (exam scores) in Wisconsin school children that illustrates how quantile regression can reveal relationships masked by conditional means approaches. The greater effect of lead exposure on children in the
lower tail of exam scores is an example of how “In many cases there may not be one, unique slope that effectively characterizes the changes across the probability distribution.”

CDC releases National Asthma Control Program state profiles

CDC has set up a new page collecting short (two-page) burden of disease profiles from the 36 National Asthma Control Program grantee states across the US.

These summaries highlight key statistical data, such as prevalence and health care use in adults and children, as well as data on patient education and medication use from the Asthma Call-Back Survey.

Great to have these standardized briefs collected in one place. The obvious next step: Make the data underlying these PDFs readily available.

CDC – Asthma – National Asthma Control Program State Profiles

Parents misperceive asthma control in kids

The rise of asthma control and impairment as the main indicators of management has renewed interest in a longstanding challenge: Variability in the perception and experience of asthma symptoms. Parents and children have been shown to differ in their assessments of the existence of asthma, let alone the presence or severity of specific symptoms. And the meaning of symptoms, and the ties to medication taking, are other matters entirely.

A new report from a large interview study suggests that worldwide, few children and adolescents achieve control of their asthma and experience frequent symptoms. A significant portion (11 percent) reported mild asthma attacks at least weekly, while 35 percent required oral corticosteroids or hospitalization at least annually.

The team interviewed 1,284 parents of children with asthma in six countries (Canada, Greece, Hungary, the Netherlands, South Africa and the UK) and 943 of the children themselves. The results highlight the impact of frequent morbidity on daily life: Asthma restricted the child’s activities in 39 percent of families and caused 70 percent to change their lifestyle. The article was published in the European Respiratory Journal.

One reason for the significant morbidity may be parental misperception of asthma control. Parents in the study tended to underestimate the severity of their child’s asthma while overestimating the level of control. While 73 percent of parents described their child’s asthma as mild or intermittent, 40 percent of children/adolescents had C-ACT scores ≤19, indicating inadequate control. In addition, even fewer (14.7%) achieved complete control as defined by the more stringent Global Initiative for Asthma (GINA) guidelines.

Parent misperception of control in childhood/adolescent asthma: the Room to Breathe survey

W.D. Carroll, J Wildhaber and PLP Brand

Disease labels in national surveys – the case of COPD

The new CDC Framework for COPD Prevention, much better thought of as a well developed agenda for applied public health, estimates that half of the people with COPD in the US have not been diagnosed.

The report – developed by a group of experts during a workshop in 2010 – proposes first among its four goals that the US improve the collection, analysis, dissemination, and reporting of COPD-related public health data. In particular, it highlights the need to develop and initiate new data collection within existing surveys, an obviously efficient and valuable objective. It recommends:

Refining the definition of COPD in existing surveillance systems by adding the terms ‘COPD,’ and ‘chronic obstructive pulmonary disease’ to the currently used ’emphysema’ and ‘chronic bronchitis’ terms.

This sentence brought me to a stop, frustrated that this change still needs to be made.

Our public health agencies must soon recognize some basic limitations of our surveillance systems: If they attempt to assess prevalence by asking people to report a physician diagnosis of some disease, they have no choice but to coevolve with diagnostic and popular nomenclature.

In this case, COPD has now been the dominant label for several years. Changes in awareness and labeling have been driven by many forces, including an improved understanding of the pathophysiology and natural history of the disease. But also by:

  • the rise of national organizations with names like the COPD Foundation, which was founded way back in 2004,
  • the marketing and education efforts of our own federal agencies (see the NHLBI page for example, or, well, the title of the report, which one downloads from the COPD page at CDC),
  • an increase in direct to consumer marketing of pharmaceuticals for the disease (such as this page for a branded formulation of tiotropium).

If we expect accurate and reliable prevalence estimates, we need to require that national surveys such as BRFSS, NHANES, and NHIS match the terminology in circulation. It is no longer acceptable to field surveys characterized by diverging epidemiological measures and popular labels.

A quick look at search frequency for COPD and emphysema (via Google Trends) shows that, in the US, search for COPD is more than twice as frequent compared to that for emphysema, and has been since about 2007. The dramatic, transient increases in search frequency for emphysema, which followed highly-publicized celebrity deaths attributed to emphysema (Johnny Carson in January of 2004) and mis-attributed to emphysema (Amy Winehouse in 2008), underscore the dynamic popular landscape of disease labels but also show the resilience of the longer-term trend.

Moreover, there are not large regional variations here; search for COPD (in blue) dominates that for emphysema (in red) across all regions of the country.

Inertia in the methods of our national surveys undermines the utility and value of their resulting estimates and, while it may make for more stable measures, assures that they are steadily providing information about an increasingly inappropriate category.

Acute health effects from chlorine gas exposure

In January 2005, a train derailment in South Carolina released 42-60 tons of chlorine gas in the middle of a small town. I was part of the CDC team that went to South Carolina to help the Dept of Health and Environmental Control respond to the disaster and investigate the health effects.

One of our papers, which looks in detail at the clinical presentation, hospital course, and pathology among those individuals who died or were hospitalized as a result of their exposure, was just published in the January 2009 issue of the American Journal of Emergency Medicine [abstract] (free download at their site). 

I’ve written more about the derailment and our investigation here.