Breastatistics

Introduction
FSP posted last year about how her love for data had driven her, during gestational diabetes, to collect extra blood glucose data from herself and graph it. Inspired by her example, today I present Dr Jekyll's Milk Pumping Data for November.

Methods
I was away from Small for ~9.5 hrs, 5 days a week. During that time, Small was exclusively fed expressed breastmilk. I initially pumped three times a day, around 9 am, noon, and 3 pm, reducing to twice a day towards the end of the month. Pumping duration of each epoch was typically 30-40 minutes, corresponding with (a) the time required to drain most available milk; (b) the time required to obtain enough milk to make the hassle worthwhile; and (c) the maximum time I was willing to spend, or rather longer. The pump used was a secondhand Medela Freestyle.

On weekends, Small was fed ad libitum direct from the source. Due to the effects of indoor lighting and winter in the northern hemisphere, he was kept on a 14:10 light:dark cycle.

Results
Asymmetry of production
Despite no obvious size disparity, Righty consistently delivered more product than Lefty. Across all observations (n = 50), Righty matched or outpaced Lefty (Fig. 1).

Fig. 1: Righty delivers the goods better than Lefty. R-squared of
best-fit line, 0.59. Dashed line represents unity.

Effect of time on output
We had hypothesized that production would peak early in the week, due to the extra stimulation on the weekend. Surprisingly, there was no significant within-week effect (Fig. 2), although the relatively low N at this point may have limited statistical power.

Fig. 2: We'll see what effect restarting Monday lab meetings
will have. Error bars represent SEM.

It is well-established that milk production decreases in the afternoon relative to the morning. Our data were consistent with this finding (Fig. 3), although the fucking ANOVA is only 0.07.

Fig. 3: I went ahead and ran the T-tests anyhow. Morning > afternoon,
p = 0.045. But the midday versus afternoon is only p = 0.09.
Disregarding the ANOVA made me feel naughty.

Pumping is less effective than slurping
As has been seen previously, a pump is not as efficient at milk extraction as a baby. However, thanks to diligent monitoring, the deficit held steady at an average of 1 oz/day (Fig. 4). The total deficit over the period under study was -16.6 oz, which was compensated for by previously frozen supplies. It is worth noting that an additional 12.2 oz was pumped outside of work (data not shown), reducing the month's net deficit to -4.4 oz.

Fig. 4: Small is an eating machine.

Discussion

Pumping wouldn't suck so much if it were consistently possible to work during sessions. Frequently, though, working reduces output. So we'll see how long I keep this up.

Also, carting one's own bodily fluids around in lab is really weird. But the prospect of writing this post was oddly motivational.

Happy Thanksgiving, Americans.