Skip to Content
  • Letter
  • Open Access

1 January 2004

Onychauxic Dystrophic Toenails Requiring Debridement in Medicare Patients

Health Quantitative Sciences, School of Public Health, New York Medical College, Valhalla, NY 10595
Dear Dr. Frank,
I have read the article “Onychauxic Dystrophic Toenails Requiring Debridement in Medicare Patients: Prevalence and Anatomical Distribution” and the comments from the APMA Board of Trustees and Michael A. Turlik, DPM. As an epidemiologist, I am able to evaluate the study and the comments relating to study methodology and statistical approach. However, as one with no clinical training, I will refrain from addressing those comments dealing with the practice of podiatric medicine.
The article describes a cross-sectional study of the proportion of individuals requiring debridement of six or more nails. Samples were drawn from two groups of patients: nursing home patients and patients seen in private practice. In both of these groups, the proportion of patients requiring debridement of six or more nails appears substantially lower than the proportion of Medicare claims filed for debridement of six or more nails (95%) and the proportion of patients with five or fewer nails was substantially higher.
The study is a cross-sectional descriptive study. It is important to note that the study cannot, nor does it appear to even attempt to, test a hypothesis. Descriptive studies are usually conducted as preliminary steps (eg, gather preliminary data) toward developing more rigorous analytic designs that ultimately will test a hypothesis. Descriptive studies are usually referred to as “hypothesis-generating” studies. Indeed, no formal statistical tests of hypotheses were conducted.
One of the letter writers raised the point that the study sample does not reflect the target population: Medicare patients billed for nail debridement nationally. Technically, this criticism is correct. The denominator of the study estimate appears to be older individuals, whereas the denominator of the Medicare data is restricted to those filing claims. However, the practical implication is that the study sample (especially the nursing home sample) would bill Medicare for treatment if treatment were required.
A second criticism of the comparison focused on the geographic variability of data. In other words, can claims from a geographically restricted sample in New York be compared with national data? Again, technically, one would expect that local or regional experiences would differ—perhaps substantially—from national data. One could argue that a more reasonable comparison group might have been claims data from New York State. However, the size of the discrepancy between the national data and the sample is intriguing. While it is certainly possible that the difference may be due to sample selection factors, it may also be due to other factors, such as reporting or administrative policies peculiar to New York State. Whether any or all of these factors influence the discrepancy between the study estimates and the national data, a descriptive study is simply not rigorous enough to answer the question. Rather, a descriptive study need only draw attention to the discrepancy. Isolating the factors influencing the difference between national and local data is achieved through more rigorous studies, as recommended by the study’s authors.
Finally, there was criticism that the study did not discuss sample size or its estimation. (I am paraphrasing this because I am not sure of the exact intent of the letter writer’s comment). As I understand it, the writer is concerned that without a discussion of sample size or (and I am guessing) tests of statistical significance, the “reader is unable to discern whether the information presented is considered statistically valid.” As I have stated throughout, it does not appear to me that the study is designed to formally test a hypothesis; rather, its purpose appears to be to generate attention and interest in discrepant information.
I should note that it would have been useful for the prevalence estimates given in the study to be presented with 95% confidence intervals. With regard to the major findings (the number of individuals with six or more nails needing debridement), the sample size appears fairly large and would have yielded somewhat narrow confidence intervals. For example, for the two estimates given in the article, 13 of 100 (13%) nursing home patients and 140 of 317 (44.2%) private practice patients, the respective 95% confidence intervals would be 7.0% to 21.0% and 38.6% to 49.8%. These are exact binomial confidence intervals.
In sum, many of the criticisms expressed (eg, interrater reliability) would appear more appropriate for a study that is being offered as a rigorous analytic study designed to test a stated hypothesis. It does not appear that the study in question is being offered as such. The purpose of the current study appears to be merely to draw attention to the large discrepancy in estimates. The two critiques certainly identify several possible explanations for the discrepant results, as well as potential study pitfalls, which may be examined more rigorously in future studies. Indeed, I would suggest that this is precisely the purpose of this descriptive study and the subsequent critiques: to serve as a basis for focusing further research on the issue and refining the research methods necessary to draw valid conclusions.

Article Metrics

Citations

Article Access Statistics

Multiple requests from the same IP address are counted as one view.