Audrey Gaskins

Nutrition and Epidemiology, Harvard T.H. Chan School of Public Health
  • Harvard T.H. Chan School of Public Health
  • United States of America

Influencer Of

Recent Comments

Jun 27, 2018
Replying to RUTH BIRK

We appreciate the importance that
Dr. Gaskins sees in our study as the first that investigates the correlation
between adherence to pre-defined dietary patterns and semen qualities.

We agree with some of the limitations mentioned by Dr. Gaskins and it is clear that further research will
be needed before our findings can lead to clinical recommendations.

However, we want to clarify one technical point. Dr. Gaskins was wandering about the fact that in some of the dietary scores the standard deviation for individuals in the second and third
quantiles was 0, suggesting some kind of an error in our calculations. We want
to explain the source of this phenomenon. First, note that the dietary scale of
the AHEI-2010 ranges between 0 and 110 based on 11 components of 10 points
each. Similarly, the aMEd scale consists of 9 components, 1 point each (see
Table 1 in Harmon et. al. Am J Clin Nutr. 2015 Mar; 101(3): 587–597). As a
result, the population scores are discrete (and clearly crude) and not continuous.
Second, the distribution of the scores in the population is normal, in the
sense that most of the scores are in the middle range. Third, we made the
reasonable decision not to split individuals with the same score into different
quantiles but to adjust the quantile boundaries accordingly (this decision is
reflected in the manuscript in the fact that the number of individuals in each
quartile is not identical, see Table 1). The combination of these three factors
can lead, and indeed led, to a situation where all individuals in a central
quantile have the same score, resulting in a standard deviation of zero. This
situation can also explain the fact that the trends that we see are
statistically significant when comparing characteristics between quartiles Q1
and Q4, and are not evident across the middle quantiles Q2 and Q3.

Authors of the article 

Dietary patterns are positively associated with semen quality

Ruth, I must have misunderstood how you calculated the AHEI index in your population. Usually, it is standard practice to assign 0 to 10 points for each component where a score of 10 indicates that the recommendations were fully met and a score of 0 represents the least healthy dietary behavior. Intermediate intakes are scored proportionately between 0 and 10. So someone could score a 5 for a particular component if they were half way to reaching the recommended intake. Given that people are not assigned a dichotomous 10 or  0 based on meeting or not meeting the recommendation, the distribution of AHEI in a population generally tends to have a wider range and variability of scores compared to metrics such as the aMED in which people only can score from 0 to 9 in integers.

Audrey Gaskins