Novel At-Home Mother’s Milk Conductivity Sensing Technology as an Identification System of Delay in Milk Secretory Activation Progress and Early Breastfeeding Problems: Feasibility Assessment

Background Prolonged exclusive breastfeeding is a public health priority and a personal desire by mothers; however, rates are low with milk supply challenges as a predominant cause. Early breastfeeding management at home is key. Milk electrolytes, mainly sodium ions, are accepted as biomarkers of secretory activation processes throughout the first weeks after birth and predictors for prolonged breastfeeding success, although they are not incorporated into routine care practice. Objective The aim of this study was to test the feasibility of a novel handheld smartphone-operated milk conductivity sensing system that was designed to compute a novel parameter, milk maturation percent (MM%), calculated from milk sample conductivity for tracking individual secretory activation progress in a real-world home setting. Methods System performance was initially evaluated in data collected from laboratory-based milk analysis, followed by a retrospective analysis of observational real-world data gathered with the system, on the spot in an at-home setting, implemented by lactation support providers or directly by mothers (N=592). Data collected included milk sample sensing data, baby age, and self-reported breastfeeding status and breastfeeding-related conditions. The data were retroactively classified in a day after birth–dependent manner. Results were compared between groups classified according to breastfeeding exclusivity and breastfeeding problems associated with ineffective breastfeeding and low milk supply. Results Laboratory analysis in a set of breast milk samples demonstrated a strong correlation between the system’s results and sodium ion levels. In the real-world data set, a total of 1511 milk sensing records were obtained on the spot with over 592 real-world mothers. Data gathered with the system revealed a typical time-dependent increase in the milk maturation parameter (MM%), characterized by an initial steep increase, followed by a moderate increase, and reaching a plateau during the first weeks postpartum. Additionally, MM% levels captured by the system were found to be sensitive to breastfeeding status classifications of exclusive breastfeeding and breastfeeding problems, manifested by differences in group means in the several-day range after birth, predominantly during the first weeks postpartum. Differences could also be demonstrated for the per-case time after birth–dependent progress in individual mothers. Conclusions This feasibility study demonstrates that the use of smart milk conductivity sensing technology can provide a robust, objective measure of individual breastfeeding efficiency, facilitating remote data collection within a home setting. This system holds considerable potential to augment both self-monitoring and remote breastfeeding management capabilities, as well as to refine clinical classifications. To further validate the clinical relevance and potential of this home milk monitoring tool, future controlled clinical studies are necessary, which will provide insights into its impact on user and care provider satisfaction and its potential to meet breastfeeding success goals.

Device precision performance was further tested in a series of frozen breast milk specimen.Milk samples were selected from MyMilk stored sample set, derived from samples voluntarily sent by mothers for various informational lab tests.Mothers provided voluntary additional data and provided a waiver and a consent for further anonymous use of remaining samples for internal R&D by the company.Milk samples were kept frozen (-20c) in 1 ml aliquots.Before analysis, either by laboratory grade instruments or the evaluated apparatus, samples were brought to stable room temp.Eleven samples were used for repeated measurements, 4 different breast milk samples and 7 spiked milk samples with elevated concentration of KCl to cover the wide range of concentrations that reflect milk samples from various time points and states (sample average conductivity range 2193-7614 µS/cm).Breast milk Samples were tested by the devices in 5-9 different days to evaluate precision.Coefficient of variation was calculated for each set Novel at-Home secretory activation milk conductivity sensing technology Supplementary Materials 2 of testing (Device 1: non spiked samples, n=4, 5 days repeats, %CV=3.5%;Spiked samples, n=7, 9 days repeats, %CV=6.8%;Device 2: spiked samples, n=7, 5 days repeats, %CV=10%).
Device performance and stability for long periods were tested by evaluating system repeatability within a period of several months.The device was monitored for 11 months by repeated testing of 5 levels of KCl solutions (3-50mM, 471-6654 µS/cm)), and for 8 months period with repeated scanning of 5 frozen breastmilk samples (2598-6483µS/cm).CV% was calculated from tests on 6-8 separate dates distributed along the assessment period (KCl solutions, n=5 samples, 8 repeats, CV=4.7%;Milk samples, n=5 samples, 6 repeats, CV%=3.8).
We further compared the results obtained by device apparatus to results obtained for a serial of large number of breast milk samples by lab-grade conductivity meter (LAQUA Twin conductivity meter EC-33, HORIBA, (Supplementary Figure S1B, full range 1475-5260 uS/cm, n=79, R^2=0.95,Recovery=95.8%(CV% 6.3)).
To understand the relation between the apparatus measurement and sample sodium concentrations, to this large set of milk samples, we plotted the measured relative conductivity against the sodium concentration measured for each sample.Milk sodium was measured in defatted breast milk by laboratory grade ion selective electrode module (The ISE module of the Roche/Hitachi Cobas 6000 c501 system), and in whole milk sample by portable Na + ion selective electrode analyzer (LAQUA Twin, NA-11 HORIBA), apparatus that has been reported to be validated for accurate analysis of human milk sodium levels [37].Results were found to be well correlated supporting Na as a major contributing factor to Milk conductivity (Supplementary

Milk biochemical correlation analysis
For testing the correlation between milk conductivity and various milk components and conditions, we performed a retrospective analysis of a dataset extracted from MyMilk's laboratory database.Records were generated empirically by in-house laboratory-based breast milk testing.MyMilk laboratories provides non-diagnostic breastmilk testing services, and routine Breast milk samples are received for testing for nutritional composition analysis, breast pain origin assessment and R&D purposes.Dataset include voluntarily reported information about birth date, habits, breastfeeding status and additional maternal and baby indicators.The Novel at-Home secretory activation milk conductivity sensing technology Supplementary Materials 4 company is the owner of the legally registered user database (database #7000655996, Israel Privacy Protection Authority database registry), and all users agreed to privacy policy and terms of use allowing for storing and using the data for R&D purposes.Data was retrospectively extracted depending on the analyte required for analysis.
For comparing the dynamics of milk conductivity data to milk sodium (Na + ), a known secretory activation marker, we first plotted the milk parameters relative to days from birth, revealing similar sharp decline in the early postpartum period (Supplementary Figure S2A,B), similar to the reported dynamics of milk Na + in the lactogenesis phase [15].According to previously reported link between milk Na + level and breast inflammation [29], we analyzed a second laboratory-generated dataset generated from samples reported and tagged with inflammation associated breast pain, and observed elevated milk Na + and milk conductivity in the inflammation associated breast pain dataset, compared with the normal dataset (Supplementary Figure S2 A,B).High correlation was found between milk sodium levels and milk conductivity levels in this dataset as well (Pearson's r=0.918,P<.001 Supplementary Figure S2 C).Analysis of cases with inflammation-associated breast pain beyond day 5 postpartum (N= 50 normal, N=19 Pain), show significantly higher milk conductivity (2487±83 µS/cm vs 3774±151 µS/cm, ANOVA F(67,1)=61.74 P<.001) and higher Na + in pain classified group (8 vs 13 mmol/L, , ANOVA F(67,1)=6.84P=.01) Supplementary Figure S2D).

Setting MM% equation
MM% is computed based on the equation MM%= 1-[(X-Xmin)/(Xmax-Xmin)]*100, where X is the raw device output, and Xmin and Xmax are the pre-set.To defined the pre-set limits, min value (Xmin) and max value (Xmax), an empirical dataset of 625 scans were used (Baby age at scan: day 0 -10 days, heterogenic breastfeeding status (exclusive/some/significant formula, with/wo problem) were used to reflect the full range; n=625; Full range 1095-9814µS/cm, Mean±STD 3761±1177µS/cm).Dataset showed A Gaussian bell-shape probability histogram with acceptable asymmetry probability distribution of a random variable about its mean (Skewness 1.32) and no heavy tailed (Kurtosis 2.88).Dataset 2.5 th and 97.5 th percentiles were set as Xmin and X max, respectively, for MM% equation.
Apparatus usage for milk sensing.User guide included Instructions for milk collection, and device use directly by the LC / mother.Users were instructed to collect 0.3-0.5mlmilk sample, from each breast separately, either before a breastfeeding session or >1 hour from the last feed, and to scan as soon as possible to sample expression.Adequate sample volume is visible by milk appearance in a control slit in the milk cell.The value appearing on the device LCD screen after 5-10 seconds is recorded into an App (Mothers App: MyLee, iOS, ID1533231342, MyMilk laboratories LTD ; Lactation support provider App: MyMilk scan by MyMilk Laboratories (invitation only), built with Appsheet) (Supplementary Figure S5).Users were instructed to reject a read due to technical error If the number was not stabilized within several seconds, and to repeat the scan.Users were instructed to wash the device between scans and at the end of use, by rinsing the milk chamber tip under running tap water and air drying.The use of the device was on the sole discretion of the user, and was not monitored for compliance.
Common technical problems included inaccurate device performance due to disconnected or leaky sample chamber, battery corrosion or low voltage, dirty chamber or electrodes, or misentry of result to App, and users were encouraged to periodically monitor QA by provided KCl solutions.Common incorrect usage by users were mainly linked to water remains in the chamber, partial fill up due to small milk volume mainly in first usages or non-complete transition of the sample into the sample cell.Users were guided on good practices and regarding common technical issues, and were supported remotely with any technical problems.