Digital Motherhood Self-Tracking Apps for Breastfeeding Mothers—A Study on Usage and Effects on Maternal Well-Being

Smartphone apps for self-tracking breastfeeding emerged as a popular tool among new mothers. Yet, we know little about how mothers use these apps and, most importantly, how self-tracking breastfeeding relates to maternal well-being. After surveying a sample of German mothers engaging with breastfeeding trackers (n = 234; recruited via an online access panel), we identified three types of self-tracking usage: (1) straightforward basic trackers, (2) meticulous data collectors, and (3) advisory-oriented self-trackers. These usage types differ regarding the data they register, the algorithmic feedback they retrieve, and their conversational levels about parameters tracked. Our findings suggest that overall maternal well-being – in terms of confidence, stress, and self-worth – remains largely unaffected by different self-tracking usage. However, when considering only the mothers’ confidence concerning breastfeeding, breastfeeding self-efficacy is lower among those most engaged in tracking and higher among those least engaged with it. Implications of these findings are discussed in terms of whether breastfeeding trackers enhance or undermine mothers’ confidence in their breastfeeding abilities relative to the intensity of their self-tracking use. Thus, future research may include longitudinal designs to validate these findings and derive effective app-supported smartphone interventions for breastfeeding mothers.


Mothers' Use of Digital Media
Transitioning to motherhood entails deep-reaching physical, mental, and social changes, along with feelings of uncertainty, anxiety, and isolation (Plantin & Daneback, 2009). Meanwhile, healthcare services often do not meet women's increased need for information and reassurance (Cannon et al., 2018). An international survey, including primarily Western countries, on information seeking during pregnancy, found that approximately half (48.6%) of pregnant women surveyed were dissatisfied with prenatal care visits (Lagan et al., 2010). Thus, due to their ready accessibility, digital media became an integral part of motherhood in the global North . Moreover, recent studies in Asia (e.g., Jayaseelan et al., 2015;Wang et al., 2018), Africa (e.g., Flax et al., 2014;Trafford et al., 2020), and South America (e.g., Quintiliano-Scarpelli et al. 2021;Silva et al., 2019) suggest a similar pattern, indicating a cross-cultural trend toward digitally mediated motherhood on a global scale. As early as the mid-1990s, various websites, discussion forums, and so-called 'mommy blogs' provided (expectant) mothers with advice, social support, and an outlet to voice their experiences . While these sources are still in use today (Jaks et al., 2019), recent research has shown a notably high uptake of maternity apps for pregnancy-and parentingrelated information seeking (Kraschnewski et al., 2014). According to data from Listening to Mothers III, a large-scale survey conducted in the U.S., more than half (56%) of first-time mothers and nearly half (47%) of experienced mothers rated apps with pregnancy and childbirth information as "very valuable" (Declercq et al., 2013). Similar figures apply for pregnant women in an Irish study (59%), which also reported common use among educationally disadvantaged women (48%; O'Higgins et al., 2014). Baby care apps are especially valued for their novel functionalities, namely self-tracking functions (Hughson et al., 2018). These range from monitoring fertility and foetal growth to tracking the infant's development, sleeping patterns, and feeding habits -covering every stage of early motherhood . Given the complexity of nursing alone, breastfeeding apps with tracking components attracted particular attention from both medical research (e.g., Demirci & Bogen, 2017;Griffin et al., 2021;Wang et al., 2017;Wheaton et al., 2018) and social sciences (e.g., Dienelt et al., 2020;Lupton, 2017;Thornham, 2019). With functions such as customisation, reminders, visualisation of progress, and synchronisation to wearables and smart devices, selftracking apps are expected to ease mothers into mindful and efficient breastfeeding routines (Virani et al., 2019). Still, not all maternal app use turns out favourable. Some perceive mobile monitoring of breastfeeding as a time-consuming task that is incompatible with the day-to-day life of a new mother (Demirci & Bogen, 2017). Complicating matters further, prior research alerts about an excessive dependency on tracking devices that can suppress the development of own maternal capabilities (Thornham, 2019). This ambivalent study situation highlights the underlying mental health-related consequences of self-tracking use. Consequently, a thorough investigation into usage types will help us uncover the potential advantages and disadvantages associated with self-tracking breastfeeding.

Conceptualisation of Self-Tracking Usage
To identify differences in self-tracking usage among breastfeeding mothers, we draw on Lomborg et al.'s (2018) conceptualisation of self-tracking. Fundamentally, self-tracking is understood as a notion of flow: A user tracks personal data and logs it into the system, the system processes the data and transmits it back to the user, ultimately allowing her "to sift through everyday life and extract habitual and meaningful practices" (Thylstrup & Lomborg, 2017, p. 1). Considering this interplay between system and user, there are varying modes of engagement with the technology. These can be categorised into registration, algorithmic feedback, and conversation (see Karnowski & Reifegerste) and will be defined below in reference to breastfeeding trackers. Registration covers the basics of self-tracking, where information about the user is logged into the system. This step happens either manually (user-initiated) or automatically (systemmonitored, i.e., through sensors on the device). The user's aim might be to increase selfawareness through quantified parameters, generating consciousness about what Lupton (2014) calls "the hidden patterns in one's life that are otherwise undiscernible" (p. 13). Hence, already the simple act of taking notes -i.e., of time between feeds, length of feed, side of feed, or amount fed -in itself imparts knowledge about relevant details to observe when nursing (Dienelt et al., 2020).
Algorithmic feedback comprises two data-driven features of self-tracking regimens: allocution and consultation (see also Karnowski & Reifegerste, 2021). As a system-initiated mechanism, allocution initially keeps users "on track" by reminding them to register data inbetween tracking sessions. For instance, breastfeeding apps send push messages to offer general advice (e.g., "5 signs your baby is hungry") or to prompt users to breastfeed in time (e.g., "It's time to breastfeed your baby!"; Hughson et al., 2018). These are either default recommendations set by the system or personalised reminders configured by the user. Consultation takes feedback communication a step further. Through an analysis of accumulated data, the system allows users to discover patterns, trends, or progress in past tracking experiences. For this purpose, the collected data is often presented in visualisations, statistics, or explicit messages (Karnowski & Reifegerste, 2021). Hence, this tool is convenient to retrace breastfeeding habits. One option for this is to compare previous with current data (historical comparison); another option juxtaposes the users' data and a defined goal (normative comparison; Hermsen et al., 2016). In addition, maternal apps often incorporate reward systems or gamification elements (Lupton & Thomas, 2019).
Conversation, as a third mode of engagement with self-tracking technology, describes the user-initiated distribution and sharing of data with others. This distribution and sharing can occur directly within the app, via social networking sites, or face-to-face. In many instances, conversations serve the purpose of gaining support and validation. Therefore, mothers not only share and discuss their data with other app users but also incorporate it in medical appointments to demonstrate their maternal skills (Thornham, 2019).
To determine differences in individual self-tracking usage among breastfeeding mothers, we proceed from this conceptualisation and ask: RQ1: Which types of self-tracking usage can be identified among breastfeeding mothers regarding users' engagement with registration, algorithmic feedback, and conversation?

Psychological Dimension of Mobile Device Use
One often neglected aspect in examining users' engagement with self-tracking technology is the overall relationship between the user and her smartphone. Therefore, we will also consider the concept of smartphone self-extension (Park & Kaye, 2018). In principle, self-extension describes how material objects, i.e., smartphones or wearables, are perceived as an integral part of one's body or an extension of the self (Belk, 1988;Ross & Bayer, 2021). This embodiment of devices can shape how a person thinks or feels (Ross & Campbell, 2021). Ross & Bayer (2021) differentiate two dimensions of smartphone self-extension. Functional self-extension refers to the outsourcing and also the expansion of human (intellectual) capabilities (e.g., count number of steps). Identity self-extension concerns the appropriation of the smartphone to create one's sense of self -whether it is the customisation of the device to reflect the user's self, or even more blatant, the 'fusing' of the user with her device to the extent that smartphone use becomes existential to one's identity.
Complementing the identification of self-tracking usage types among breastfeeding mothers, we want to take a closer look at users' connectedness to their smartphones: RQ2: How do the identified types of self-tracking usage among breastfeeding mothers differ with regard to users' level of both functional and identity smartphone selfextension?

Self-Tracking and Maternal Well-Being
Maternal engagement in self-tracking can be seen as a liberating practise: It helps mothers understand their bodies' and their infants' cues and guides them through an otherwise stressful and emotionally challenging stage of life (Byrt & Dempsey, 2020). Correspondingly, self-tracking is associated with building maternal confidence (e.g., Gibson & Hanson, 2013). Mothers in Dienelt et al.'s (2020) qualitative survey expressed their appreciation for infant feeding trackers, especially in gaining some sense of confidence, mastery, and control. Affirmative algorithmic feedback conveys the certainty of "doing okay" (Thornham, 2019, p. 176). This sort of empowerment also improves breastfeeding efficacy (Dienelt et al., 2020. Accordingly, studies find enhanced intentions for breastfeeding exclusivity and duration following mobile health (mHealth) interventions (Ahmed et al., 2016;Litterbach et al., 2017). Nonetheless, there is no clear evidence that technically more advanced mHealth systems with interactive features lead to more promising breastfeeding rates than simple digital breastfeeding handouts (Griffin et al., 2021;Lewkowitz et al., 2020).
Breastfeeding, however, is not just a matter of willingness but largely depends on physiological factors (e.g., milk supply, mastitis) and, of course, on how well the baby cooperates (Awaliyah et al., 2019). Thus, self-tracking apps may cause frustration and annoyance if users cannot reach set goals (Costa Figueiredo et al., 2018). Contrary to the promised relief from worries (Johnson, 2014), scholars point to the risk of reinforced maternal stress (Demirci & Bogen, 2019). According to Sanders (2017), anxiety-provoking self-tracking experiences originate from the "sense of constant visibility" (p. 53). In the case of new mothers, this could be the fear of not meeting the socially imposed normative ideals of "good motherhood" (Thornham, 2019, p. 177). Here, critics raise concerns about datafication (van Dijck, 2014), an overreliance on quantified parameters. This trap of dataism may misguide women to perceive tracked data as more accurate and reliable than their own maternal subjectivity (Thornham, 2019). For instance, interviewed women repeatedly described situations in which they first grabbed the tracking device before latching on their crying baby for feeding (Dienelt et al., 2020;Thornham, 2019). In this case, self-tracking might create an illusion of control and inhibits actual maternal enjoyment (Sharon, 2017).
Considering the normative nature of self-tracking, users are constantly confronted with their maternal performance (Johnson, 2014). This constant confrontation inevitably affects feelings of self-worth, especially at a highly sensitive time like the breastfeeding period . Accordingly, Knittel et al. (2018) suggest that positive or negative feelings about data outcomes can boost or harm users' self-esteem. Here, it is worth noting that (breastfeeding) mothers' appreciation of the body and its functionality has fundamental associations with selfesteem (Hutchison & Cassidy, 2021;Rosenbaum et al., 2020). Furthermore, as Johnson (2014) posits, mothers could attain feelings of self-worth by sharing their maternal experience and images of their babies (as "the mother's work" (p. 337)) on digital media. However, empirical evidence on how maternal self-worth projects through conversations or self-tracking remains scant.
To investigate maternal well-being with respect to breastfeeding mothers' engagement with self-tracking technology, we hence ask: RQ3: How do the identified types of self-tracking usage among breastfeeding mothers differ with regard to the users' maternal well-being relating to confidence, stress, and self-worth?

Method
Building on related work using a qualitative approach (e.g., Dienelt et al., 2020;Lupton, 2017;Thornham, 2019), we set out to quantify the research problem. Since the identification of usage types (RQ1) requires precise descriptive answers to the questions "how many?", "how much?", and "how often?", a standardised questionnaire represents an adequate approach (Amaratunga et al., 2002;Mulisa, 2021). The logic of typological data then allows uncovering the similarities and differences of said usage types on smartphone self-extension (RQ2) and maternal wellbeing (RQ3). Hence, we conducted an online survey among German mothers (18 to 44 years) who have breastfed in the last 12 months and have used mobile self-tracking apps to monitor breastfeeding. Mothers for our sample were invited via an online access panel by Bilendi, which contacted 4241 women aged 18 to 44 with a child aged up to 24 months. Bilendi paid each participant €1.00 to complete the survey. The study was carried out in September 2021 on the survey platform SoSci survey. Data analysis was performed using R.

Measures 1
Modes of Self-Tracking. To assess participants' engagement with self-tracking systems, according to Lomborg et al. (2018), we followed Karnowski and Reifegerste's (2021) adaption and further customised it to apply to breastfeeding tracking. For registration, we asked participants to specify which parameters (respectively time, length, and amount of breastfeeding, pumping, and bottle feeding) they tracked how often and which device(s) they used to log information. We determined the forms of algorithmic feedback participants retrieved through the specifications general advice, visualisations, personalised advice, and rewards. Additionally, we asked if they received these as push messages. Concerning personalised advice (e.g., "It's time to breastfeed your baby"), participants were requested to state whether they self-adjust these or if they also make use of the apps' default recommendations. Acts of conversation were measured by the frequency in which users discussed their self-tracked data within the app, on social media, or face-to-face with family and friends and in medical appointments, e.g., with a midwife. We also assessed how participants first became aware of the app.
Smartphone Self-Extension. We measured both functional and identity smartphone selfextension using Ross and Bayer's 12-item smartphone Self-extension Scale (2021; scale from 1 = does not apply at all to 5 = fully applies, intermediate points labelled).
Maternal Well-Being. To determine maternal confidence, we put three measures to use: First, in terms of specific breastfeeding confidence, we used the short 14-item breastfeeding Self-Efficacy Scale (Dennis, 2003;e.g  breastfeeding because I feel proud and important while breastfeeding."; 24 items, scale from 1 = strongly disagree to 5 = strongly agree, intermediate points labelled) that allowed us to assess both the level of skill as well as willingness concerning mothers' perceptions in their breastfeeding abilities. In addition, we assessed overall maternal experience -outside of breastfeeding obligations -with the Being a Mother Scale (Matthey, 2010; e.g.: "I have felt confident about looking after my baby/toddler."; 13 items, scale from 1 = never to 5 = always, intermediate points labelled). According to previous studies, we captured maternal stress using the Perceived Stress Scale, which indicates the level of stress during the last month (Cohen et al., 1983, p. 394;Schneider et al., 2020; e.g.: "In the last month, how often have you felt nervous and 'stressed'?"; 10 items, scale from 1 = never to 5 = very often, endpoints labelled). Maternal self-worth was based on the Rosenberg Self-Esteem Scale (Rosenberg, 1965;von Collani & Herzberg, 2008; e.g.: "I certainly feel useless at times."; 10 items, scale from 1 = does not apply at all to 5 = fully applies, intermediate points labelled) and the Body Appreciation Scale-2 (Tylka & Wood-Barcalow, 2015; e.g.: "I respect my body."; 10 items, scale from 1 = never to 5 = always, intermediate points labelled), with which we gauged attitudes towards both one's inner self and outer appearance. An overview of descriptive statistics and reliability measures is presented in Table 1.
Demographic Characteristics. Finally, respondents' age, educational level, occupation, and family status were assessed. Moreover, we measured specific maternal characteristics like former breastfeeding experience and details on the breastfed child.

Participants
The recruitment initially resulted in 247 completed interviews, of which 13 had to be excluded, as they did not conform to the quality criteria (e.g., rushed through the questionnaire; Leiner, 2019). Hence, our final sample consists of 234 mothers with an average age of 31.1 years (SD = 4.5). 54.3% of respondents have a high, 38.5% middle, and 5.5% low educational level. 2 61.1% were not employed during their breastfeeding period, 21.8% worked part-time, and

Maternal Engagement with Breastfeeding Trackers: A Typology
Addressing our first research question on different types of self-tracking usage among breastfeeding mothers, we implemented a latent class analysis using R package poLCA (v. 1.4.1;Linzer & Lewis, 2011) 3 . To identify clusters of similar usage styles, the three main modes of engagement with self-tracking systems -registration, algorithmic feedback, and conversation -were included. Based on the one-to ten-class-solutions, the three-class-solution represents the best model fit as defined by the Bayesian Information Criterion (BIC, see Table 2), which leaves us with a satisfactory entropy of .93 (Celeux & Soromenho, 1996). Building on this model, we determined three types of self-tracking usage among breastfeeding mothers, which can be characterised as (1) straightforward basic trackers, (2) meticulous data collectors, and (3) advisory-oriented self-trackers (see Table 3).
Straightforward basic trackers make up the biggest group in our sample (43.0%). Mothers in this group are likely to stick to the basics of self-tracking usage on all modes. They appear to register relatively few parameters -mainly time and length of breastfeeding -and they do more sporadically than other users. Like the meticulous data collectors, it is improbable that they connect their tracking system to external devices. In line with a fairly basic approach to registration, their employment of algorithmic feedback is rather simplistic. Straightforward basic trackers likely access general advice and visualised results. Yet, in all likelihood, they largely avoid more personalised data handling features. Conversational-wise, it is unlikely that mothers in this group share their data within the app or on social media. Nonetheless, discussions about tracked data are more likely in interpersonal settings, at least now and then.   Meticulous data collectors constitute the second largest group (35.8%). They stand out for arguably the most elaborate self-tracking routine in terms of registration. Their eponymous characteristic 'meticulous' refers to how likely they are to collect data regularly and how many different types of data they probably keep track of. The only exception is the amount of milk drunk during breastfeeding and pumping length. Although not to the same extent as the advisory-oriented self-trackers, this usage type is probable to retrieve several forms of algorithmic feedback. While they are presumably not overly focused on reward systems, personalised advice is at least moderately used. Concerning conversation, this group differs from the straightforward basic trackers only in that they are more likely to engage in interpersonal talks slightly more often.
Advisory-oriented self-trackers are the smallest group (21.2%), characterised by a relatively high uptake of all modes of self-tracking. Although their registrative activities do not seem as thorough as those of the meticulous data collectors, they are very likely to log numerous types of data, at least occasionally. Despite the low rates, this type is more likely to incorporate wearables for breastfeeding purposes. Furthermore, they appear to hold strong demands for all kinds of algorithmic feedback, including personalised advice and rewards. Their conversational acts also reflect their engagement with 'advisory' aspects of self-tracking systems. Not only are they most likely to discuss their results in person, but unlike the other two groups, sharing tracked data within the app or on social media is more likely among this group.

Further Comparison of Self-Tracking Usage Types
Based on the LCA presented above, we assigned participants to the class to which they most likely belong to further analyse and compare these usage types (see Tables 2 and 3).
The initial motivation to download a self-tracking app for breastfeeding most often originated from the users themselves (73.5%), with no significant differences between the three usage types. Referrals from third parties to install a breastfeeding tracker are significantly more prevalent among advisory-oriented self-trackers, both from their personal network (37.5%) and medical professionals (18.8%; see Table 4). Fitting in with their different use of feedback features, we also see significant differences between the three groups in using push messages and setting up personalised advice. While only one-third of straightforward basic trackers (33.6%) enable push notifications, half of meticulous data collectors (51.4%) and two-thirds of advisory-oriented self-trackers (68.8%) do so. Taking a look at how the usage types make use of personalised advice, it becomes clear that straightforward basic trackers and advisoryoriented self-trackers primarily rely on self-adjusted prompts and reminders (71.4%; 86.4%), whereas default app recommendations are less of an option (28.6%; 13.6%). This ratio is less evident for the meticulous data collectors, who, next to self-adjusting (52%), also draw on preset advice from the app (48%).  Note. Means with different subscripts differ at the p = .05 level by Tukey's HSD. Scale from 1 to 5, the higher the value, the higher the level of smartphone self-extension.

Levels of Functional and Identity Smartphone Self-Extension
To answer RQ2, we compare the identified breastfeeding self-tracking types regarding their functional and identity smartphone self-extension (see Table 5). The three groups rank at a comparable, medium-high level concerning functional self-extension. Meanwhile -consistent with their greater affinity towards all modes and varieties of self-tracking -identity selfextension is significantly more pronounced among advisory-oriented self-trackers.

Self-Tracking Usage and Maternal Well-Being
We will now turn to the users' maternal well-being, answering our third research question. Again, we will juxtapose levels of maternal confidence, stress, and self-worth with the identified types of breastfeeding self-tracking (see Table 6). As part of maternal confidence, breastfeeding self-efficacy and motivation differ significantly depending on the type of self-tracking usage. Interestingly enough, self-efficacy ranks significantly higher among those who track the least (straightforward basic trackers) than those who track the most rigor (meticulous data collectors). Motivation for breastfeeding is significantly less apparent among meticulous data collectors and straightforward basic trackers than among advisory-oriented self-trackers. Nonetheless, outside of breastfeedingand thus outside of what is tracked -mothers in our sample experience motherhood hardly any differently. By the same token, no considerable differences can be observed for maternal stress and maternal self-worth: All mothers feel a moderate stress level. Indices of self-worth, that is, self-esteem and body appreciation, show a medium-high level throughout the sample. In sum, maternal well-being remains largely unaffected by different styles of self-tracking usage.

Sociodemographic and Maternal Characteristics of Self-Tracking Usage Types
To obtain a more thorough understanding of the three usage types, we explore mothers' sociodemographic and maternal background information, both of which might explain differences in how the different types employ breastfeeding trackers in their daily lives (see Tables 7 and 8).
Sociodemographically, the three usage types only differ in terms of their occupation while breastfeeding. Advisory-oriented self-trackers are more likely to hold full-time (29.2%) or parttime positions (27.1%) than straightforward basic trackers (10.8% full-time; 21.6% part-time) and meticulous data collectors (19.0% full-time; 17.9% part-time). Comparing the types in terms of their maternal characteristics reveals that advisory-oriented self-trackers have on average more than one child (M = 2.1) and correspondingly more breastfeeding experience (70.8%) than straightforward basic trackers (M = 1.6; 47.1%) and meticulous data collectors (M = 1.5; 39.3%). Still, none of the usage types have significantly more or less prior experience with breastfeeding trackers from previous breastfeeding periods. Around one-fifth of meticulous data collectors (20.2%) and advisory-oriented self-trackers (21.3%) had a premature delivery of their current breastfeeding child compared to significantly less straightforward basic trackers (4.0%). In addition, nearly half of the births among meticulous data collectors required a caesarean section (45.2%). This rate is considerably lower among advisory-oriented self-trackers (22.9%) and straightforward basic trackers (17.6%). Note. Means with different subscripts differ at the p = .05 level by Tukey's HSD. 1 Scale 1 to 5, the higher the value, the higher the level of breastfeeding self-efficacy, breastfeeding motivation, perceived stress, selfesteem, and self-worth; 2 Scale 1 to 4, the higher the value, the more positive mothers experience maternity.

Discussion
Following the proposed modes of engagement with self-tracking technology, we aimed to identify differences in how mothers use apps to monitor breastfeeding. Building on this typology and taking the dimension of smartphone self-extension into account, we set out to investigate how users' self-tracking usage ties in with their maternal well-being, particularly confidence, stress, and self-worth. Drawing on a survey with German mothers who use mobile breastfeeding trackers, we conducted a latent class analysis that factors in acts of registration, algorithmic feedback, and conversation. This procedure allowed us to identify three types of self-tracking usage: straightforward basic trackers, meticulous data collectors, and advisoryoriented self-trackers. Straightforward basic trackers show a laid-back approach to self-tracking. In terms of parameters tracked, they register the essentials, i.e., time and length of breastfeeding, but virtually nothing beyond that. Accordingly, as with the meticulous data collectors, usage of supplemental tracking devices is atypical. In line with that, their engagement with other modes of self-tracking can be described as somewhat restrained. While they use general advice and visualised statistics, more advanced features of algorithmic feedback like personalized advice, rewards, and push messages do not intrigue them as much. Compared to the two other usage types, straightforward basic trackers participate least in conversations about tracked data.
A far more extensive self-tracking program in terms of registration is evident among the meticulous data collectors. Mothers in this group rarely skip data entry: They keep track of the most parameters and do so in the most regular manner. Different kinds of algorithmic feedback seem to play a crucial role in how this usage type manages infant feeding, with general advice and visualisation being the most important. Furthermore, meticulous data collectors admittedly self-adjust personalised prompts, but contrary to the two other groups, they are more likely to rely on the app's default recommendations. Regarding conversation, they resemble straightforward basic trackers in that they shy away from online activities of data sharing. However, they are slightly more amenable to talk about tracked data face-to-face.
The modes with which the above-described groups engage only to a limited extent loom larger among the advisory-oriented self-trackers. Although not as broadly as the meticulous data collectors, this usage type still registers a significant share of parameters. These provide the basis for their advisory-oriented approach to self-tracking: they are most likely to retrieve all forms of algorithmic feedback, down to personalised advice and rewards. Moreover, they are more inclined to talk about their tracking experience across all channels. Their higher involvement in all modes of self-tracking is also reflected in that they at least partially connect the breastfeeding tracker with other devices.
Considering users' overall attachment to their smartphones, a pattern emerges that reflects the properties of the identified usage types. While there are no significant differences regarding functional self-extension, advisory-oriented self-trackers show significantly higher levels of identity smartphone self-extension compared to the other two usage types. Mirroring their almost ludic use of and high engagement with all three modes of self-tracking, this is a fitting result. These findings lead us to assume that users already firmly attached to their mobile devices might engage more with all offered features of self-tracking systems, while users showing a more distanced connection to their devices might stick to the basics.
While there are noticeable differences in how mothers engage with breastfeeding trackers, our results indicate no major disparities in how they feel mentally. Especially in terms of overall maternal experience, stress, and self-worth, all mothers reported relatively similar levels of well-being. When considering well-being solely regarding breastfeeding, varying levels of selfefficacy and motivation between the usage types become apparent. This finding is concerning insofar as it falls into the scope of what these mothers aim to manage through self-tracking. Notably, breastfeeding self-efficacy is significantly less pronounced among those who track the most rigorous (meticulous data collectors) compared to the straightforward basic trackers, who track least intensively. Therefore, a less data-driven nursing routine could imply that these mothers already feel more comfortable breastfeeding. Perhaps using a breastfeeding app serves the very purpose of confirmation rather than guidance and improvement. In contrast, more detailed self-tracking routines do not enhance mothers' perceptions of their ability to breastfeed. If anything, one could even argue that more elaborate tracking styles promote uncertainty as a side effect of the "constant visibility" (Sanders, 2017, p. 53) of one's maternal performance. Another explanation could be that mothers who extensively monitor breastfeeding already feel more doubtful and insecure about it, hence using self-tracking to obtain control. Indeed, this would also fit with the finding that significantly more women among the meticulous data collectors delivered their current breastfeeding child by caesarean section or prematurely, which also poses extra physiological challenges to infant feeding. However, on a more positive note, advisory-oriented self-trackers show the highest level of breastfeeding motivation. Arguably, this could stem from their feedback-focused and conversation-heavy usage style. In that sense, gamification elements, like rewards, or the exchange with others, keep them going. Maternal context information also indicates that mothers in this group have more breastfeeding (tracker) experience, possibly making them more versed and relaxed in this regard.

Limitations and Conclusions
This study comes with certain limitations that inform future research. First, our sample only consisted of mothers engaging with breastfeeding trackers, neglecting mothers who do not make use of such applications. Hence, future research should aim to incorporate non-trackers' perspectives on the value of these apps. A more heterogeneous sample would also enable a more comprehensive assessment and comparison of maternal well-being. Besides, the crosssectional design of our study did not allow evaluations of appropriation processes regarding breastfeeding trackers. Accordingly, future research on the use of self-tracking technologies could benefit from longitudinal studies to understand how users incorporate the system into their (breastfeeding) routine over time. This could conceivably involve a multimethod approach with a diary study in which mothers reflect on their mental state rather than just stating it off-the-cuff in the questionnaire. In addition, such a longitudinal approach could also give hints at causal mechanisms happening between self-tracking usage and maternal wellbeing. Finally, we must acknowledge the sampling via an online access as a limitation, as this procedure naturally harbours the risk of selection bias. In this respect, we should at least be aware of self-selection and noncoverage as constraints to the generalisability of our findings.
Still, the current study contributes to the literature on self-tracking technology and mothers' employment of mHealth applications. In particular, smartphone self-extension as a determinant on users' individual self-tracking experience proved to be a fertile theoretical consideration to enhance future studies on self-tracking. Beyond the mere identification of usage types, this study also provides important insights into the potentials and risks regarding mothers' perceptions in their breastfeeding abilities relative to the frequency and intensity of their selftracking use. To provide effective app-supported smartphone interventions, we need to know more about the current app market for breastfeeding trackers. Thus, it is essential to learn how satisfied women are with the currently available apps, whether the systems and features offered meet their individual needs, and whether they think that smartphone interventions can help