Bias in Sleep-Disordered Breathing Research: A Critical Review
This paper critically examines potential sources of bias in Peppard et al. (2013), a community-based time series study on the increased prevalence of sleep-disordered breathing in adults. The analysis identifies several key methodological concerns: selection bias arising from the use of a Wisconsin-based convenience sample that was extrapolated to the broader U.S. population; longitudinal cohort aging that confounds obesity and health outcome trends over time; and the uncritical use of BMI as a proxy for weight despite its well-documented limitations. The paper concludes that while no study is entirely free of bias, the authors only partially addressed these issues, raising questions about the generalizability of their findings.
- Introduction: The Ubiquity of Bias in Research: All research contains unavoidable, often unconscious bias
- Overview of Peppard et al. (2013): Study examines rising sleep-disordered breathing in Wisconsin adults
- Selection Bias and Sample Representativeness: Wisconsin sample not representative of U.S. population
- Cohort Aging as a Confounding Factor: Aging participants inflate observed health decline over time
- BMI as a Flawed Proxy and Remaining Limitations: BMI used uncritically despite well-known measurement flaws
- Conclusion: Biases partially addressed but validity remains uncertain
✍️ How to write this paper — guide, tools & examples ▾
What makes this paper effective
- The paper opens with a strong theoretical framing — acknowledging that all research contains bias — before applying that framework to a specific study, giving the critique intellectual grounding.
- Each identified bias is explained in concrete, accessible terms, making the critique clear and easy to follow without sacrificing analytical depth.
- The conclusion is balanced: the paper does not dismiss the study outright but fairly assesses whether the remaining biases undermine validity.
Key academic technique demonstrated
The paper demonstrates critical appraisal of a published study — a core skill in health and social sciences. The writer systematically identifies methodological weaknesses (selection bias, cohort effects, proxy measurement issues) while acknowledging the authors' own awareness of these limitations. This balanced critique avoids overclaiming while still offering substantive evaluation.
Structure breakdown
The paper is organized into a general introduction on bias, a brief study summary, three focused critique sections (sampling, aging cohort, BMI), and a concluding validity assessment. This moves logically from broad to specific and back to broad, a pattern well-suited to critical review writing at the undergraduate level.
Introduction: The Ubiquity of Bias in Research
Bias is ever-present in research. While we are generally aware of the basic forms that bias takes, it often arises unconsciously, which makes it more difficult to detect. All studies contain bias — these biases begin with our own enculturation. The lenses through which we examine every facet of the world are shaped by our experiences, and thus the bias of our own culture, of how we studied science, and of our preferred methodologies underpins every decision made in a research project. No study can be said to be entirely free from bias. It is therefore more important to recognize where bias exists and how it influences study design, the interpretation of findings, and ultimately our vision of objective truth.
Overview of Peppard et al. (2013)
Peppard et al. (2013) studied the increased prevalence of sleep-disordered breathing in adults. The authors examined time series data from an ongoing community-based study in Wisconsin to determine whether there has been an increase in sleep-disordered breathing in adults, and to draw conclusions about causality. The authors compared current modelled rates against historical rates, all using data drawn from that survey.
Sleep-disordered breathing includes conditions such as sleep apnea and hypopnea, both of which are known to have negative health consequences. The authors proceeded on the premise that factors like hypertension and related weight issues are contributing factors to such disorders, and that an increase in obesity would therefore also lead to an increase in sleep-disordered breathing in adults over the time period of the study.
Selection Bias and Sample Representativeness
There are several potential sources of bias in this study. The researchers are based in Wisconsin, and the study relied on a Wisconsin-based dataset. The availability of longitudinal data going back to the 1980s was clearly appealing — its existence essentially facilitated the study design. However, the study population, which began with over 1,500 participants, is not necessarily reflective of the broader population. Ethnically, the sample is unlikely to align with the United States as a whole, and may not even be representative of Wisconsin's general population.
Although participants were chosen randomly, the demographics of the sample are not likely to mirror those of the United States. This constitutes a form of selection bias: the availability of this particular dataset is what led to its use, rather than any determination that it was the most appropriate sample. The authors did attempt to extrapolate findings from this group to the entire country, a step that is methodologically questionable given these sampling limitations.
Cohort Aging as a Confounding Factor
Another significant source of bias is that the study participants grew older over the course of the study. When the time series study began, participants were between 30 and 70 years of age. They were assessed at a baseline and then reassessed every four years thereafter, meaning the same individuals were included at every interval. The problem with this design is that health outcomes tend to decline naturally with age, so rates of obesity and related health conditions would be expected to increase in any aging cohort — independent of broader population trends.
This aging effect skews the results. If the study had instead used a fresh random sample of, say, 40-year-olds at each interval, it would have more effectively tracked changes in the population over time by eliminating the confounding influence of longitudinal cohort aging. As designed, it is difficult to disentangle true population-level increases in sleep-disordered breathing from the expected health deterioration that accompanies an aging sample.
Conclusion
The authors were aware of these issues and tried to work around them, but even so they were not able to completely remove such biases from their findings. The central question is whether they were able to reduce these biases sufficiently that their findings retain validity. There is no compelling reason to conclude that the study's findings are invalid — the authors clearly made reasonable efforts. However, there are biases in this paper that are not fully discussed and not fully accounted for in the study design. A more explicit acknowledgment of these limitations, particularly the sampling constraints and the confounding effect of cohort aging, would have strengthened the study's credibility and its claims to generalizability.
References
Peppard, P., Young, T., Barnet, J., Palta, M., Hagen, E., & Hla, K. (2013). Increased prevalence of sleep-disordered breathing in adults. American Journal of Epidemiology. Retrieved April 4, 2016, from https://aje.oxfordjournals.org/content/early/2013/04/13/aje.kws342.full
Create your account
Always verify citation format against your institution’s current style guide requirements.