We read with interest the article by Pendse, et al.
[1] regarding role of transcutaneous bilirubin measurement in preterm
neonates. The study is yet another attempt to validate new era of
sensor-based technology that is quick and non-invasive. However, we have
few concerns:
1. The authors used a statement: "Bland-Altman
analysis was used to ‘visualize’ the agreement between TSB and TCB." It
appears that the authors have confused Association with Agreement.
Correlation/regression or analysis of variance (t-test) are measures of
association and not of agreement. An excellent association does not
necessarily mean good agreement. Bland and Altman developed a simple
statistical tool to measure agreement between two methods way back in
1986 [2]. However, the tool is often used inappropriately [3].
2. Authors also highlighted that 90% of the data
points fall within 95% confidence interval. This is a statistical fact,
and we cannot claim anything about agreement from this. As it is a kind
of estimation problem, a larger sample size is advisable. Bland himself
suggested minimum of 100 observations with justification [4]. A
complicated regression analysis or a tool based on Structural Equation
Modeling (SEM) [5] may be used to analyze method comparison studies but
Bland-Altman method is popular due to its simplicity and ease of
interpretation. The authors could have tested the reliability along with
validity as they have three transcutaneous bilirubin observations.
We suggest the authors to continue the study with
large sample (at least 100) and present the findings again with
Bland-Altman analysis as primary (and only) analysis and interpretation
based on the new findings.
References
1. Pendse A, Jasani B, Nanavati R, Kabra N.
Comparison of transcutaneous bilirubin measurement with total serum
bilirubin levels in preterm neonates receiving phototherapy. Indian
Pediatr. 2017;54:641-3.
2. Bland JM, Altman DG. Statistical methods for
assessing agreement between two methods of clinical measurement. Lancet.
1986;1:307-10.
3. Phatak AG, Nimbalkar SM. Method comparison
(Agreement) studies: Myths and rationale. J Clin Diagn Res.
2017;11:JI01-J3
4. Bland JM. How can I decide the sample size for a
study of agreement between two methods of measurement? Available from:
https://www-users.york.ac.uk/~mb55/meas/sizemeth.htm. Accessed
March 5, 2018.
5. Alanen E. Everything all right in method comparison studies? Stat
Methods Med Res. 2012;21:297-309.