English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

In 1963, the late Bobby Bragan, then manager of the Milwaukee Braves, poked fun at the statistics commonly used by baseball analysts, observing that “according to the percentage people” a person who was standing “with one foot in the oven and one foot in an ice bucket” was, on average, “perfectly comfortable.” Nearly 5 decades later, the problem highlighted by Bragan’s colorful assessment of the limitations of summary statistics has grown exponentially, especially in health care. Bombarded daily with an estimated 34 gigabytes of information in nonwork activities alone, American health care consumers are exposed to an overwhelming amount of statistical information about any health care condition that they choose to investigate. For example, a Google search on the term “H1N1 vaccine” returns more than 3 million hits in 0.28 seconds. The resulting potential for data overload should prompt us to pay close attention to how we package and summarize these data. It is perhaps not surprising that a great deal of research in numerous medical specialties focuses on consumers’ perceptions of data about their health risks, often finding that consumers either underestimate their risk or fail to take action based on known risks. For example, in a case-control study of genetic counseling for BRCA1/2 testing in a sample of women with a family history of breast or ovarian cancer, Armstrong et al. (2005) found that 43 of 200 (21.5%) cases (i.e., patients who had undergone genetic counseling) and 21 of 181 (11.6%) controls perceived their lifetime risk of developing breast cancer as high. Risk was perceived as low for 121 of 200 (60.5%) cases versus 145 of 181 (80.1%) controls. Similarly, Moore et al. (2010) studied a sample of AfricanAmerican women seeking care in urban health centers, finding that overweight and normal-weight women reported the same perceived risk of weight-related illnesses; and Holmes et al. (2007) studied auditory symptoms and hearing protection use in adults aged 18 to 27 years, finding that few used hearing protection consistently, although 6% reported hearing loss and 20% reported symptoms including ear pain and tinnitus after exposure to loud noises. Slightly rephrased, perhaps the issue at the heart of all of these studies is to what extent health care consumers believe that summary statistics, such as averages, actually apply to them. The challenge for managed care, as both a user and purveyor of data and information, is to ensure that data-weary consumers are provided with accurate but comprehensible summaries of which behaviors and treatments are most likely to benefit instead of harm. This task, already challenging because of the complexity of health care information, has been made even more difficult by the increasing use and promulgation of retrospective observational research designs. That is, when health care outcomes are analyzed retrospectively, the only limits to the methodological options available are the imaginations and budgets of the investigators. The resulting myriad of measures and outcomes in the published research literature can cause both conflicting findings and bewilderment about what study results really mean. Retrospective analyses of administrative claims data pose several particular challenges in summarizing risks and benefits. A few key examples are discussed in this editorial.

It Depends on What Mean Means: Averaging Versus Event Patterning in Analyses of Administrative Claims Data