"While not a diagnostic tool, this study playfully probes whether the dimensions of human bones, like rivers and stock markets, follow a peculiar mathematical pattern, revealing a complex and nuanced relationship with Benford’s Law."
This intriguing investigation by orthopedic surgeons in Hangzhou, China, delves into an unexpected mathematical curiosity: whether the numerical data describing human bones adheres to Benford’s Law, a principle that governs the distribution of leading digits in many naturally occurring datasets. The researchers meticulously analyzed X-ray measurements from 132 bones across three adult patients, comparing raw lengths, perimeters, and projected areas against the predictable patterns of this law. While the initial findings suggest that bone dimensions do not perfectly align with this mathematical quirk, the study opens a window into the potential applications of such analyses in validating scientific data and understanding the inherent variability within biological measurements, all while emphasizing its purely descriptive nature and distinct separation from clinical diagnoses.
A Mathematical Enigma in the Skeletal System
In a study that blurs the lines between radiology and number theory, two orthopedic surgeons from Hangzhou, China, have embarked on an unusual exploration of human anatomy. Yunlong Zhi and Cheng Ji, affiliated with the Affiliated Hangzhou First People’s Hospital, School of Medicine, Westlake University, posed a thought-provoking question: Do the numerical descriptions of human bones behave in a manner analogous to the numbers that characterize rivers, stock market fluctuations, or international trade volumes? Their findings, published in the International Journal of General Medicine, suggest a complicated answer, one that the researchers are careful to frame as a mathematical curiosity rather than a definitive medical insight.
The core of their investigation revolves around Benford’s Law, a fascinating observation about the frequency distribution of leading digits in many real-world numerical datasets. This law posits that in such datasets, the digit ‘1’ appears as the leading digit approximately 30.1% of the time, ‘2’ about 17.6%, and so on, with the frequency of leading digits decreasing as the digit’s value increases, culminating in ‘9’ appearing only about 4.6% of the time. This phenomenon has been observed across a diverse range of data, from financial transactions and scientific measurements to word frequency counts in various languages. The underlying principle is that datasets spanning several orders of magnitude tend to exhibit this pattern. However, datasets confined to a narrower numerical range often deviate from this predictable distribution.
The surgeons’ decision to investigate human bones stemmed from this very principle. They hypothesized that if bone dimensions, when measured and quantified, exhibited a similar leading-digit distribution, it could have implications for understanding data integrity in anatomical studies. However, they were acutely aware of the inherent limitations. Human long bones, by their anatomical nature, are constrained within a limited range of sizes, typically falling within one or two orders of magnitude. This constraint, they anticipated, might influence how well bone measurements conform to Benford’s Law. Therefore, their study was designed as a descriptive inquiry, aiming to characterize the numerical patterns rather than to definitively assert that bones "obey" the law.
The Rigorous Methodology Behind the Curiosity
To conduct their research, Zhi and Ji meticulously screened patients who had undergone X-rays of at least half their long bones following trauma evaluations between January 2019 and December 2023. One patient was excluded from the study due to hallux valgus, commonly known as a bunion deformity, to ensure the analyzed bones were as representative of typical anatomy as possible. This left a cohort of three adult patients: a 31-year-old man, a 66-year-old man, and a 55-year-old woman. Crucially, none of these individuals had experienced fractures, tumors, or other significant skeletal deformities that could skew the measurements.
From these three patients, the researchers systematically measured 44 long bones each, yielding a total of 132 observations. The measurements included raw bone length, perimeter, and projected area. Recognizing that geometrical relationships involve higher powers of linear dimensions, they also calculated squared and cubed values of the bone lengths. These derived values were intended to explore how two-dimensional and three-dimensional quantities might relate to the linear measurements and potentially influence their adherence to Benford’s Law.
The analysis revealed a complex interplay between the bone measurements and the mathematical prediction. Raw bone lengths showed a statistically significant deviation from the expected Benford distribution, with a p-value of 0.001. This deviation was particularly pronounced in the frequency of numbers starting with the digit ‘2’. In the bone length data, ‘2’ appeared as the leading digit 29.5% of the time, a marked contrast to the 17.6% predicted by Benford’s Law. Similarly, perimeters also deviated significantly (p = 0.001), but in the opposite direction, with only 3% of perimeter values starting with ‘2’.
In contrast, projected areas and squared lengths exhibited a much closer alignment with Benford’s Law, showing no statistically significant deviations. The p-values for these measurements were 0.260 and 0.292, respectively, indicating that their leading-digit distributions were not substantially different from what the law would predict. The cubed lengths came even closer to conformity, with a p-value of 0.910.
To further quantify the degree of deviation, the researchers employed the Mean Absolute Deviation (MAD), a straightforward measure of how far observed proportions stray from expected ones. The MAD values followed a consistent pattern: 0.042 for length, 0.037 for perimeter, 0.024 for area, 0.026 for squared length, and a minimal 0.014 for cubed length. This reinforced the observation that as the dimensionality of the measurement increased (from linear to squared and cubed lengths), the data showed a greater tendency to align with Benford’s Law.
However, the authors themselves were quick to identify a potential confounding factor in their most compelling results. They noted that cubing a number inherently expands its numerical range. This expansion, they explained, can artificially push a dataset toward a Benford-like pattern, suggesting that the improved adherence observed with cubed lengths might be an arithmetic artifact rather than a reflection of a true biological phenomenon.
The Critical Limitations and the Path Forward
The study’s limitations section is notably candid, highlighting the need for careful interpretation of the findings. The most significant limitation is the exceptionally small sample size. Analyzing bones from just three patients, while yielding 132 measurements, does not represent a diverse population. Furthermore, the bones from a single individual are not independent observations; the proportions and measurements of one person’s skeleton inherently influence dozens of data points. This interdependence restricts the generalizability of the findings to a broader population.
Another crucial caveat lies in the methodology of X-ray imaging. Standard front-to-back X-rays provide two-dimensional projections, effectively flattening a three-dimensional object into a shadow. Consequently, they cannot accurately measure true three-dimensional surface area, volume, or mass. What the researchers termed "projected area" was, in essence, this flattened shadow. The authors explicitly cautioned that these projections should not be construed as genuine three-dimensional measurements.
The researchers also urged against over-interpreting the statistical significance of the p-values. A p-value above 0.05, while indicating a failure to reject the null hypothesis (i.e., that the data conforms to Benford’s Law), does not equate to definitive proof of conformity. Moreover, the chi-square test, while useful for detecting deviations, cannot quantify the degree of adherence. The authors meticulously summarized their findings as "descriptive evidence of digit-distribution characteristics in long-bone morphometric data," emphasizing their observational rather than conclusive nature.
The research was supported by the Medical and Health Research Project of Zhejiang Province, and the authors declared no conflicts of interest. The study underwent a single, anonymous peer review process, receiving comments from two reviewers.
Benford’s Law: A Tool for Scientific Integrity in Medicine
The rationale behind this seemingly academic pursuit is rooted in a practical application of first-digit analysis within the scientific community. Benford-style checks have become a valuable tool for investigators examining the integrity of research data, particularly for detecting fabricated or manipulated numbers. Invented datasets, often created without adherence to natural numerical distributions, tend to fail these statistical tests. This approach has been explored in various biomedical contexts, including the analysis of histology measurements and the validation of COVID-19 case reporting data from European countries.
The authors of the bone study propose that establishing reference patterns for skeletal measurements could, in the future, contribute to the assessment of data quality and the validation of morphometric databases. However, they are equally clear about the strict boundaries of their findings. These results must not be used as diagnostic criteria for skeletal diseases or as evidence to deem any individual bone measurement as abnormal.
For the general reader, the most salient takeaway is a modest yet important one: a mathematical pattern that governs many natural datasets does not neatly apply to the dimensions of human bones. The probable reason is the limited range of variation in bone sizes, which is insufficient for the pattern to emerge distinctly. The question of whether this relationship might change with larger, more diverse samples and the use of true three-dimensional scanning techniques remains open. The authors suggest that future studies should address this by conducting analyses stratified by crucial demographic and pathological factors, including sex, age, height, ethnicity, and the presence of skeletal diseases. This nuanced approach promises to further illuminate the intricate connection between the quantitative world of mathematics and the complex biological reality of the human skeleton.
The study, funded by the Medical and Health Research Project of Zhejiang Province, underwent a single anonymous peer review and the authors declared no conflicts of interest.
