ArticlePDF Available

Abstract and Figures

Maternal and infant health is a global healthcare problem affecting developing and developed countries alike. Pregnancy complications increase the risk of maternal and infant death, and are associated with adverse outcomes such as miscarriage, stillbirth, and preterm birth. Lifestyle modifications before and during pregnancy have been shown to reduce the risk of pregnancy complications. This paper evaluates the use of wearable sensors as an enabling technology to motivate healthier lifestyle behaviors during pregnancy. Wearable sensors have been used to track health and lifestyle in general populations. Pregnancy is associated with hormonal, physiological and psychological adaptations that need to be taken into account when quantifying lifestyle, and there is to date no solution available to accurately measure lifestyle behaviors during pregnancy. The main challenges lay in developing algorithms and systems capable of dealing with physiological adaptations, lifestyle adaptations, and enabling longitudinal monitoring throughout pregnancy. Addressing these challenges will open new opportunities towards an integrated and personalized approach to behavior change interventions based on sensor data. Putting such solutions in the hands of consumers may unlock a crowdsourcing approach to quantification of known risk factors, discovery of new markers for pregnancy complications, and to answering some of the fundamental questions that remain around adverse outcomes such as preterm birth
Content may be subject to copyright.
IEEE
Proof
INVITED
PAPER
1Wearable Sensors for Healthier
2Pregnancies
3This paper evaluates the use of wearable sensors as an enabling technology to
4motivate healthier lifestyle behaviors during pregnancy in developing and
5developed countries.
6By Julien Penders,Marco Altini,Chris Van Hoof, and Eric Dy
7ABSTRACT |Maternal and infant health is a global healthcare
8problem affecting developing and developed countries alike.
9Pregnancy complications increase the risk of maternal and
10 infant death, and are associated with adverse outcomes such as
11 miscarriage, stillbirth, and preterm birth. Lifestyle modifica-
12 tions before and during pregnancy have been shown to reduce
13 the risk of pregnancy complications. This paper evaluates the
14 use of wearable sensors as an enabling technology to motivate
15 healthier lifestyle behaviors during pregnancy. Wearable
16 sensors have been used to track health and lifestyle in general
17 populations. Pregnancy is associated with hormonal, physio-
18 logical and psychological adaptations that need to be taken
19 into account when quantifying lifestyle, and there is to date no
20 solution available to accurately measure lifestyle behaviors
21 during pregnancy. The main challenges lay in developing
22 algorithms and systems capable of dealing with physiological
23 adaptations, lifestyle adaptations, and enabling longitudinal
24 monitoring throughout pregnancy. Addressing these chal-
25 lenges will open new opportunities towards an integrated
26 and personalized approach to behavior change interventions
27 based on sensor data. Putting such solutions in the hands of
28 consumers may unlock a crowdsourcing approach to quantifi-
29 cation of known risk factors, discovery of new markers for
30 pregnancy complications, and to answering some of the
31 fundamental questions that remain around adverse outcomes
32 such as preterm birth.
33 KEYWORDS |Author, please supply index terms/keywords for
34 your paper. To download the IEEE Taxonomy go to http://www.
35 ieee.org/documents/taxonomy_v101.pdf.AQ1
36I. INTRODUCTION
37Pregnancy complications represent a serious risk to both
38maternal and infant health and are associated with adverse
39outcomes such as miscarriage, preterm birth, stillbirth,
40and low birth weight. Worldwide 289,000 women died in
412013 during and following pregnancy and childbirth. Most
42of these deaths occurred in low-resource settings and could
43be prevented. In the United States alone, 650 women die
44each year as a result of pregnancy or delivery complica-
45tions [1]. Ten percent of pregnancies are considered to be
46high risk and that number is increasing with the increase
47in maternal age and incidence of chronic disease. Overall,
4894.1% of deliveries reported some type of pregnancy or
49delivery-related complications in the United States in
502008, accounting for $17.4 billion in hospital stays alone
51[2]. Preterm birth is by far the most prevalent adverse
52outcome in pregnancy, with 500,000 babies born preterm
53every year in the United States alone, accounting for a total
54(direct and indirect) cost of $26 billion per year.
55Ahealthylifestyleisessentialforahealthypregnancy.
56The Center for Disease Control and Prevention recom-
57mends all women of reproductive age to adopt healthy
58lifestyles, in particular maintain a healthy diet and weight,
59be physically active, quit all substance use for good, and
60prevent injuries [1]. Lifestyle modifications before and
61during pregnancy have the potential to reduce the risk of
62maternal and fetal complications and chronic diseases [3].
63Furthermore, pregnancy has been identified as a window
64of opportunity for long-term lifestyle changes [4].
65Recent advances in mobile health, or m-health,
66technologies, and in particular wearable sensor technolo-
67gies, open new possibilities in monitoring health condi-
68tions remotely and tracking health parameters in daily life
69environments.AQ2 Wearable technologies empower people
70with new tools to engage in healthier lifestyles and habits,
71and reduce their risk of chronic diseases through behavior
72change. Mobile health technologies have already been
Manuscript received August 26, 2014; revised November 3, 2014;
accepted December 12, 2014.
J. Penders and E. Dy are with Bloom Technologies, San Francisco, CA, USA.
M. Altini is with Bloom Technologies, Diepenbeek, Belgium.
C. Van Hoof is with imec, Leuven, Belgium.
Digital Object Identifier: 10.1109/JPROC.2014.2387017
0018-9219 !2015 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
|Proceedings of the IEEE 1
IEEE
Proof
73 shown to reduce the impact of chronic diseases through
74 behavior change [5], and to effectively improve women’s
75 health [6].
76 This paper evaluates wearable sensors as an enabling
77 technology to motivate healthier lifestyle behaviors during
78 pregnancy and reduce the risk of lifestyle-related preg-
79 nancy complications. Wearable sensor technologies,
80 behavior change and pregnancy monitoring have been
81 largely covered in literature. This paper is a first attempt to
82 marry these three spaces, providing a technology perspec-
83 tive on promoting healthy pregnancies within the context
84 of behavior change.
85 The rest of the paper is organized as follows. In the first
86 section we introduce the relation between lifestyle
87 behaviors, risk factors and pregnancy complications. The
88 main maternal adaptations during pregnancy are summa-
89 rized in the second section, whereas the third section
90 summarizes how lifestyle behaviors relate to pregnancy
91 outcomes. The fourth section reviews wearable sensor
92 technology solutions that have been developed to date,
93 specifically for the field of pregnancy monitoring. Next,
94 the main challenges in designing wearable sensor systems
95 and algorithms for lifestyle behavior monitoring during
96 pregnancy are discussed. The paper closes by sketching the
97 main opportunities at the intersection of wearable
98 technology, behavior change and obstetrics, with the
99 promise to reduce pregnancy complications and improve
100 outcomes.
101II. LIFESTYLE-RELATED PREGNANCY
102COMPLICATIONS
103Like much of our health, pregnancy outcomes are affected
104by a variety of risk factors such as genetic factors,
105environmental factors, clinical history, and lifestyle factors
106(see Fig. 1). Although the first three are mainly out of our
107control, lifestyle factors can be influenced by our lifestyle
108behaviors. Lifestyle-related pregnancy complications im-
109pact not only short and long term maternal health, but also
110have lasting impact on the health of the unborn baby. The
111most common maternal health conditions include: ane-
112mia, urinary tract infections, mental health conditions
113(depression), hypertension, gestational diabetes mellitus,
114and hyperemesis gravidarum (extreme morning sickness).
115Hypertension is the second leading cause of maternal
116death in the United States. It can be chronic, gestational or
117pregnancy induced, and in all cases it significantly
118increases the risk of preeclampsia and preterm birth.
119Preeclampsia has an incidence of 5%–8% [7]. Gestational
120diabetes affects 2%–10% of all pregnancies [8] and this
121number could be doubled accounting for all nonreported
122cases. The most common fetal and baby’s health problems
123include preterm birth, stillbirth, birth defects, fetal growth
124problem and low birth weight. In addition, pregnancy
125complications may lead to delivery complications such as
126Caesarian deliveries, extended delivery time, or the need
127for special delivery interventions.
Fig. 1. The relationship between lifestyle behaviors, risk factors, and pregnancy complications. Lifestyle risk factors, clinical history, genetic
risk factors and environment risk factors all influence pregnancy complications. Lifestyle risk factors can be impacted by behavior change to
reduce the risk of pregnancy complications.
Penders et al.:Wearable Sensors for Healthier Pregnancies
2Proceedings of the IEEE |
IEEE
Proof
128 This paper defines lifestyle behaviors as one’s daily
129 habits or routines that can be adapted, interrupted and
130 taken up again upon one’s decision. Clear relations have
131 been reported between pregnancy complications and
132 outcomes, and the following lifestyle behaviors: physical
133 activity [9], [10], sleep [11], stress [12], diet and weight
134 management [13], [14], smoking [15], and drinking [16].
135 This paper focuses on lifestyle behaviors that can be
136 tracked using wearable sensors, and thus for which
137 technology offers new opportunities for designing behavior
138 change interventions based on objective monitoring, in
139 particular during pregnancy.Theseinclude:physical
140 activity, sleep, stress, diet, and weight management.
141 III. MATERNAL ADAPTATIONS DURING
142 PREGNANCY
143 Pregnancy is associated with hormonal, psychological and
144 physiological adaptations that assist fetal growth and
145 prepare the woman’s body for labor. Understanding these
146 adaptations is important to develop accurate lifestyle
147 behavior models and distinguish normal changes from
148 abnormal ones. For the sake of this paper, we focus on
149 physiological adaptations that affect measurable lifestyle
150 markers.
151 A. Body Weight Adaptations
152 Maternal weight increases as the woman’s body is
153 preparing for delivery and the fetus is growing. While the
154 relationship between maternal body weight and birth
155 weight has long been established, only recently has it been
156 investigated with respect to pregnancy outcome. Recom-
157 mendations and guidelines for healthy weight gain have
158 been introduced and are constantly updated reflecting
159 changes in the general population [1].
160 B. Autonomic and Cardiovascular Adaptations
161 The autonomic nervous system (ANS) has been shown
162 to play a role in adapting the body during pregnancy
163 [18]–[20]. Gruge et al. measured a decreased sympathetic
164 activity until mid-pregnancy, followed by a return to
165 prepregnancy sympathetic activity in the third trimester.
166 Kuo et al. reported a biphasic shift from the autonomic
167 nervous activity toward a lower sympathetic and higher
168 vagal modulation in the first trimester, and then toward a
169 higher sympathetic and lower vagal modulation in late
170 pregnancy. This biphasic change is reflected on heart rate
171 variability (HRV) measurement: the high frequency (HF)
172 component of HRV increases during first trimester, then
173 progressively reduces during the second and third
174 trimester.
175 Important cardiovascular changes altering the physiol-
176 ogy of pregnancy have also been observed. Heart rate (HR)
177 increases during pregnancy, starting between two and five
178 weeks of pregnancy and continuing well into the third
179 trimester [21]. On average, the resting HR raises
1808 beats/min by the eighth week, and reaches an increase of
18116 beats/min by the end of pregnancy [22]. The
182mechanism of the increased HR is not yet clearly
183identified, and may be attributed to sympathetic reflex
184adjustments to maintain arterial blood pressure despite a
185reduction of systemic and peripheral vascular resistance
186[23], [24].
187A notable increase in blood volume takes place
188between 10 and 20 weeks of pregnancy. Resting cardiac
189output (Q) is increased as early as the fifth week of
190pregnancy as a result of the increased heart rate, stroke
191volume and blood volume. Resting Q increases by 1 L/min
192at 8 weeks of gestation, which represents over 50% of the
193overall change in pregnancy. During the third trimester,
194resting Q increases only minimally, primarily because of
195the increase in HR as term approaches [25].
196Systolic blood pressure is not increased in normal
197pregnancy due to decreased peripheral vascular resistance.
198Diastolic and therefore median blood pressure may
199decrease up to 15 mmHg in mid-pregnancy [25].
200C. Respiratory and Metabolic Adaptations
201Energy consumption increases during pregnancy and
202especially in later stages, in great part due to an increase in
203Resting Metabolic Rate (RMR) [26]. The higher RMR is
204significantly correlated with higher body weight during
205pregnancy, and remains unchanged if adjusted by body
206weight. Thus, changes in resting metabolism seem to be
207mostly due to the increase in body weight due to the
208growth of the fetus. The correlation between total energy
209expenditure (TEE) and RMR during pregnancy is high,
210indicating that the contribution of physical activity energy
211expenditure (PAEE) to TEE during pregnancy is small and
212relatively constant from one person to another.
213Changes in submaximal VO2during pregnancy depend
214on the type of exercise performed. During maternal rest
215or submaximal weight-bearing exercise (e.g., walking,
216stepping, treadmill exercise), the absolute maternal VO2
217
(L/min) is significantly increased compared with the
218nonpregnant state. The magnitude of change is approx-
219imately proportional to maternal weight gain, similarly
220to what was found during measurements of RMR [27].
221However, when pregnant women perform submaximal
222weight-supported exercises (e.g., level cycling), where
223the energy cost of locomotion is not altered by maternal
224morphological changes, the findings are contradictory.
225Some studies reported significantly increased absolute
226VO2[27], [28] while many others [29], [30] reported
227unchanged or only slightly increased absolute VO2
228
compared with the nonpregnant state.
229D. Musculoskeletal Adaptations
230Anatomical changes occurring as the woman’s body is
231preparing for delivery and due to the growing fetus can
232affect the mother’s musculoskeletal system. The increased
233weight in pregnancy may significantly increase the forces
Penders et al.:Wearable Sensors for Healthier Pregnancies
|Proceedings of the IEEE 3
IEEE
Proof
234 across joints such as the hips and knees and cause changes
235 in posture, which in turn may affect balance, predisposing
236 pregnant women to increased risk of falling. Despite a lack
237 of scientific evidence on increased musculoskeletal inju-
238 ries during pregnancy, the medical community tends to
239 prescribe low intensity weight bearing exercise during
240 pregnancy.
241 IV. LIFESTYLE BEHAVIORS AND
242 PREGNANCY COMPLICATIONS
243 Maternal adaptations during pregnancy lead to changes in
244 lifestyle behaviors that may impact pregnancy complica-
245 tions. Understanding these changes is important in order
246 to develop accurate models of lifestyle behaviors during
247 pregnancy. Here we focus on lifestyle behaviors that can
248 be tracked using state of the art wearable technology,
249 namely: physical activity, sleep, stress, diet, and weight
250 management.
251 A. Physical Activity
252 Walking is the most popular type of activity amongst
253 pregnant women, followed by activities such as swimming
254 laps, weight lifting, gardening, and aerobics [32]. Behav-
255 ioral changes in activity patterns have been reported
256 during the pregnant and postpartum states [26], [33].
257 Pregnant women spent significantly more time in activities
258 of light intensities ðMETG¼2:5Þ, and significantly less
259 time in activities of moderate intensities ðMET >2:5Þ,
260 when compared to postpartum states.
261 Physical exercise may reduce the risk of preeclampsia,
262 prevent gestational diabetes and reduce postpartum
263 musculoskeletal issues [9], [10]. Physical exercise may
264 also have a positive impact on breastfeeding and postpar-
265 tum weight loss as well as on offspring health and
266 development. While there are only a limited number of
267 studies published on the relationship between maternal
268 physical activity and the occurrence of preeclampsia, they
269 all point to a reduced risk by at least 35% and as much as
270 69% as a result of recreational exercise and higher
271 occurrence of walking and stair climbing [10]. Physical
272 inactivity is reported to be a key risk factor for obesity and
273 type-2 diabetes [34]. While limited studies have addressed
274 the link between maternal mental health and physical
275 exercise, studies report a positive effect on self-esteem and
276 general wellbeing as well as reduced insomnia [36], and a
277 randomized control trial has suggested that postpartum
278 exercise may be effective in reducing the symptoms of
279 depression [37]. The impact of moderate physical exercise
280 on reducing other side effects of pregnancy such as
281 insomnia, anxiety, nausea, leg cramps and lower back pain
282 have also been reported [9].
283 B. Sleep
284 The majority of women experience altered and
285 disturbed sleep during pregnancy [11]. General discomfort,
286increased nighttime urinary frequency, fetal movement,
287fatigue are factors that are responsible for increased sleep
288disturbance. Following delivery hormonal levels quickly
289drop to their pre-pregnancy levels and it takes on average
29012 weeks till uninterrupted sleep, largely driven by the
291maturing of the circadian rhythm of the child.
292These adverse sleep changes due to pregnancy may
293lead to sleep disorders [11], the most common being
294insomnia, sleep disorder breathing (SDB) and restless leg
295syndrome. SDB has been signaled as a risk factor for small-
296for-gestational-age births and for pregnancy-induced
297hypertension. There are currently no guidelines for
298treatment of pregnancy-associated sleep apnea but pub-
299lished guidelines for the treatment of SDB in the general
300population can serve as a reference (e.g., guidelines from
301the American Academy of Sleep Medicine [35]). After
302pregnancy, and typically after six months, the majority of
303the underlying causes have disappeared and disorders do
304not persist.
305C. Stress
306Throughout the course of pregnancy, mood fluctua-
307tions are common and symptoms generally improve during
308the second trimester and worsen during the third
309trimester. Reproductive hormones changes, psychosocial
310factors (weight gain and body image), maternal stress and
311worries, sleep disorders and perceived lack of control all
312contribute to stress vulnerability.
313Research studies have shown that depression and
314anxiety during pregnancy, stressful life events, poor social
315support and a history of depression are the major risk
316factors for postpartum depression. As a result, 10%–20%
317of women develop major depressive disorder during and
318after pregnancy [38]. Studies have also revealed that
319anxiety and stress during pregnancy have impact on fetal
320health [39]and on developmental outcome in infancy [12].
321D. Diet and Weight Management
322Currently 20–40% of women gain more than the
323recommended weight during pregnancy. Epidemiological
324studies have shown that the average pregnancy related
325weight retention is between 0.5–3.0 kg [40].
326Maternal obesity and excessive gestational weight gain
327are established risk factors for most pregnancy complica-
328tions, in particular preeclampsia, gestational diabetes
329mellitus, perinatal mortality, and stillbirth [41]. Further-
330more, excess weight gain during pregnancy combined with
331lifestyle changes during the postpartum period is an
332important contributor to obesity among women [13]and
333infant [42]. The management of weight gain during
334pregnancy and the weight loss post partum are most
335effectively achieved through a combination of exercise and
336calorie restriction. Aside from the beneficial effect on
337limiting weight gain, dietary intervention and cholesterol
338lowering diets have shown promise for reducing gesta-
339tional hypertension, preeclampsia and preterm birth,
Penders et al.:Wearable Sensors for Healthier Pregnancies
4Proceedings of the IEEE |
IEEE
Proof
340 although larger studies are needed. Postpartum, there
341 were no reported adverse impacts of the suggested calorie
342 restriction on lactation. Evidence is increasing that
343 adequate diet also reduces gestational diabetes [14].
344 V. TRACKING LIFESTYLE BEHAVIORS
345 DURING PREGNANCY
346 Recent advances in mobile and wearable sensor technol-
347 ogies enable seamless tracking of lifestyle parameters, and
348 open new opportunities for motivating behavior change.
349 Numerous wearable sensor systems and solutions have
350 been reported for tracking health parameters and
351 behaviors. Monitoring lifestyle behaviors during pregnan-
352 cy however presents additional challenges due to the
353 specific maternal adaptations associated with pregnancy.
354 Furthermore, the natural changes in lifestyle behaviors
355 associated to pregnancy adaptations shall be taken into
356 account when designing behavior interventions for preg-
357 nant women. In this section we review wearable technol-
358 ogies and algorithms available to date for tracking lifestyle
359 behaviors specifically during pregnancy. We refer the
360 interested reader to other publications for detailed reviews
361 of wearable sensor technologies for health [45]–[47]and
362 lifestyle [48], [49]monitoring in general populations.
363 Besides wearable sensors, unobtrusive sensors integrated
364 in the environment may also be used to monitor health
365 parameters. Here we focus exclusively on wearable
366 sensors, and refer the interested reader to other publica-
367 tions for relevant work using unobtrusive sensors, for
368 example the work by Liu et al. [55].
369 A. Physical Activity
370 Physical activity is the lifestyle factor that has received
371 the most attention to date. Questionnaires, pedometers
372 and accelerometers have been used in clinical research to
373 study the effect of different aspects of physical activity
374 (type, energy expenditure (EE), frequency and duration)
375 on pregnancy outcomes. Accelerometer-based approaches
376 have reported moderate to low accuracies, reaching 77.5%
377 and 56.9% accuracy in measuring steps for pregnant
378 women depending on the type of pedometers used [33].
379 Studies comparing the performance of questionnaires and
380 pedometers show agreement between the two approaches
381 but fail in providing any conclusions with regards to their
382 absolute accuracy in measuring physical activity [43]. The
383 use of more advanced wearable sensors combining physical
384 and physiological signals has lead to inconclusive results in
385 pregnancy, with underestimation and overestimation
386 reported in different studies [44], [50].
387 Most systems introduced to date used basic technolo-
388 gies such as step counters or EE monitors relying on single
389 regression models to estimate EE from movement. More
390 research is available on the general population, where
391 single regression models and step counters have been
392 replaced by machine learning based methods. Most
393notably, activity-specific EE models able to detect both
394activity type and the associated EE have consistently
395outperformed previous methods [51]. Our group has
396shown that combining physical and physiological infor-
397mation improves the accuracy of physical activity and
398energy expenditure measurement in non-pregnant popu-
399lation [52]. This accuracy improvement is expected to
400extend to pregnant population as well, although no study
401up to date investigated the accuracy of recently developed
402activity-specific models during pregnancy, with respect to
403reference activity type and indirect calorimetry.
404The large number of consumer solutions available for
405activity tracking reflects the mature stage of wearable
406technologies and algorithms for physical activity assess-
407ment. Recently, specific versions of activity trackers
408targeted to the women market have been released (see
409Fig. 2). Smartphone apps, possibly coupled with a
410wearable activity or heart rate tracker, allow people to
411track their activity and EE, using state of the art
412algorithms. The accuracy and validity of these consumer
413solutions is yet to be established. Special care needs to be
414considered especially in pregnant populations, where level
415of activities, as well as posture, change significantly
416compared to the general population. There is today no
417commercial solution available for accurately measuring
418physical activity in pregnant women.
419B. Sleep
420Generic sleep monitoring technologies have been used
421in research studies in pregnancy with varying perfor-
422mances. PolySomnoGraphy (PSG) provides the most
423detailed and complete view on sleep and has been used
424in a few sleep pregnancy studies [53]. PSG is the gold
425standard for sleep monitoring but its applicability to real-
426life settings is limited. Actigraphy provides a much less
427intrusive alternative and has the major advantage that it
428can be used in the home environment [54]. However
429actigraphy only provides information about sleep time,
430which limits its usability for understanding underlying
431physiological mechanisms linking sleep to pregnancy
Fig. 2. Activity trackers specifically targeted to women. Left: Shine by
Misfit Wearables. Right: Fitbit Flex.
Penders et al.:Wearable Sensors for Healthier Pregnancies
|Proceedings of the IEEE 5
IEEE
Proof
432 disorders. The majority of sleep studies on pregnant
433 women however still rely on questionnaires to assess the
434 overall sleep quality during pregnancy as well as specific
435 disorders [56]–[59]. The Berlin questionnaire is one of the
436 most used questionnaire to diagnose Obstructive Sleep
437 Apneas (OSA), but was shown to have poor performance in
438 predicting obstructive sleep apneas in pregnancy [60].
439 Our group, amongst others, has shown that wearable
440 sensors can be used for sleep monitoring with the potential
441 to move sleep monitoring from the expensive hospital
442 environment to home [61], [62]. Recent algorithms
443 developed in the general population using wearable
444 sensors showed that sleep stages can be detected by means
445 of HRV, thus without the need of invasive PSG devices
446 [63]. Such approaches would need to be investigated and
447 validated during pregnancy, especially in view of the
448 autonomic and cardiovascular adaptations associated to
449 pregnancy.
450 A multitude of smartphone apps allow people to log
451 their sleep time. A few of them also provide automatic
452 tracking based on sensors (microphone and accelerome-
453 ter) embedded in the smartphone. Most commercially
454 available activity trackers provide an estimation of sleep
455 duration. Recently new products have been introduced
456 that rely on physiological signals (EEG, ECG, GSR) to
457 provide additional information regarding sleep architec-
458 ture. The accuracy of these consumer solutions has yet to
459 be established and validated [64]. There is as of today no
460 pregnancy specific sleep monitoring solution.
461 C. Stress
462 Stress assessment in general population is usually done
463 using questionnaires such as the Perceived Stress Scale
464 [65] or salivary cortisol sampling. Pregnancy is associated
465 with specific stressors, and there are very few instruments
466 available that measure pregnancy related stress. The
467 pregnancy experience scale was designed to address
468 pregnancy-specific contributors to maternal psychological
469 state [66], [67]. It captures both positive and negative
470 events and provides an indication of the balanced stress
471 state in pregnancy, both of which have been shown to have
472an impact on pregnancy health [67]. Other scales have
473been designed to address the needs of specific geographies
474and cultures [68], [69]. A few studies have reported the
475successful use of HRV as a measure of stress during
476pregnancy, as an early indicator of gestational depression
477[70]and pregnancy induced hypertension [71], [72].
478Several groups, including our own, have reported
479physiological-based stress monitoring systems, mainly
480confined to laboratory or pre-defined settings [73]–[75].
481These systems capture physiological responses activated by
482the ANS to provide an estimation of arousal and stress (see
483Fig. 3). HR, HRV, galvanic skin response and respiration
484rate have been the most used parameters. Recent work
485moving towards arousal and stress assessment in real-life
486conditions showed promising results in small samples of
487the general population. These methods use machine
488learning techniques for context recognition, combined
489with physiological markers of stress such as HR or HRV
490[76] to contextualize and better interpret the physiological
491response to daily life stressors. Such methods would need
492to be validated during pregnancy, as autonomic and
493cardiovascular adaptations associated to pregnancy may
494impact the relation between autonomic nervous system
495markers and stress.
496Most consumer solutions for stress monitoring are
497smartphone apps that let people log their stress using
498digital versions of analog stress scales. A few recent apps
499allow users to track HRV, based on which they derive a
500stress level estimation. These apps presumably offer an
501objective measure of stress, but suffer from the specificity
502issues associated to HRV. To date, there is no pregnancy
503specific stress monitoring solutions.
504D. Diet and Weight Management
505While diet is half of the energy balance equation,
506technological solutions are still lacking, both in epidemi-
507ological research and the consumer market where most
508solutions rely on questionnaires. The food, beverage, and
509medication intake questionnaire (FBMIQ) is one of the
510most widely used to track diet in clinical research studies
511involving pregnant women.
Fig. 3. The imec’s ECG necklace (left) and wrist-based sensor (right) measure heart rate variability and Galvanic Skin Response, respectively,
from which an estimation of stress level is derived.
Penders et al.:Wearable Sensors for Healthier Pregnancies
6Proceedings of the IEEE |
IEEE
Proof
512 Leveraging the emergence of wearable technologies, a
513 few research groups have reported approaches to auto-
514 matically track food and calorie intake in small samples of
515 the general population. Different techniques were pro-
516 posed, using microphones [77] or wrist-based inertial
517 sensors [78]. These are early concepts however, and
518 further research is required to make these solutions robust
519 and reliable for accurate tracking of calorie and energy
520 intake.
521 A multitude of apps are available for pregnant women
522 to manually track their diet and nutrition. There is
523 however no automated method commercially available for
524 diet and food intake tracking.
525 VI. CHALLENGES
526 Wearable technologies for lifestyle monitoring have seen a
527 growing deployment in the general population. Yet very
528 few pregnancy specific wearable health solutions have
529 been introduced so far. In this section, we discuss the
530 challenges in applying wearable sensor technologies to
531 pregnancy monitoring, and highlight the important
532 aspects to take into consideration when applying tech-
533 nologies initially developed for the general population, to
534 pregnancy.
535 A. Physiological Adaptations
536 Technologies developed to measure lifestyle changes
537 during pregnancy need to account for the physiological
538 changes associated to the different stages of pregnancy.
539 General methods developed for physical activity and EE
540 estimation are inaccurate during pregnancy because of the
541 associated cardiovascular, respiratory and metabolic
542 changes. First, the extra energy requirements needed for
543 supporting the rise in rest metabolic rate associated to
544 maternal and fetal tissue growth should be accounted for
545 when determining RMR and EE during various activities.
546 Second, the change in the slope of the HR-VO2 regression
547 curves indicates that the amount of energy burned at a
548 given HR changes as pregnancy progresses [27], [28], [79].
549 Overlooking this physiological variation would lead to an
550 overestimation of EE at rest, and an underestimation
551 during physical activity. New algorithms must be devel-
552 oped to automatically recalibrate the individual relation
553 between physiological signals such as HR and EE or
554 oxygen consumption as pregnancy progresses. Similarly,
555 physical activity monitoring algorithms need to be
556 revisited to account for specific changes in gait and
557 balance patterns affected by pregnancy specific musculo-
558 skeletal adaptations.
559 The autonomic nervous system readjusts during
560 pregnancy to meet the additional needs of the changing
561 body and growing fetus. Pregnant women have been
562 shown to have lower HF but higher LF/HF and LF
563 compared with their non-pregnant counterpart [80].
564 Furthermore the relative contributions to the HF, LF and
565LF/HF vary with pregnancy stages [20]. These fluctuations
566in HRV complicate its usability as a measurement of
567autonomic system dysfunction and stress [19], [20].
568Further clinical research is required to gain a better
569understanding of the changes in HRV that are due to
570pregnancy in order to isolate them from stress-induced
571HRV changes. Accordingly, algorithms developed for
572stress monitoring based on physiological signals need to
573be revisited in the context of pregnancy to account for the
574normal autonomic fluctuations associated with the differ-
575ent stages of pregnancy.
576Algorithms implemented in sleep monitoring systems
577should account for the cardiovascular and respiration
578adaptations associated to pregnancy, especially when
579tracking sleep patterns longitudinally. Solutions that
580integrate built-in detection of sleep disorders should
581account for the normal physiological changes associated
582to pregnancy to avoid misdiagnosis of sleep disorder
583conditions.
584B. Lifestyle Adaptations
585Physiological, psychological and hormonal changes
586lead to lifestyle adaptations throughout pregnancy. Phys-
587ical activity levels evolve towards less weight-bearing and
588lower intensity activities during pregnancy. Walking is the
589most popular physical activity in pregnant women,
590followed by activities such as swimming laps, weight
591lifting, gardening, and aerobics. When developing algo-
592rithms for physical activity monitoring during pregnancy,
593the focus should go towards increasing the accuracy of
594physical activity measures for these activity classes that are
595the most representative for pregnant women, if necessary
596at the expense of lowering the accuracy in other part of the
597physical activity compendium.
598Furthermore, detailed data about exercise duration,
599frequency, type and intensity, especially for sporadic
600activities (i.e., not only intentional physical exercise),
601become crucial to define activity programs and interven-
602tions that will promote healthier pregnancies. For
603example, there is preliminary evidence indicating that
604intermittent exercise may increase oxidative stress, which
605could presumably increase the risk of preeclampsia [81].
606This complicates the assessment of physical activity, since
607it becomes essential to differentiate regular from irregular
608exercise. Better algorithms are needed to monitor inter-
609mittent exercise and transitions between different physical
610activities, and provide a more accurate estimation of
611physical activity and EE that can be used to drive behavior
612change interventions in pregnant women.
613Similarly, information about the sleep architecture and
614fragmentation is important when evaluating sleep quality
615and possible sleep disorders [11]. Systems and algorithms
616that combine actigraphy with physiological signals are
617needed to provide the necessary detailed information
618regarding architecture and disruption of sleep [43].
Penders et al.:Wearable Sensors for Healthier Pregnancies
|Proceedings of the IEEE 7
IEEE
Proof
619 Pregnancy is associated with the apparition of new
620 stressors such as fear of pregnancy problems or child
621 malformations. In addition, it has been shown that both
622 hassles and uplifts need to be considered in assessing the
623 overall stress status of pregnant women [67]. Accordingly,
624 algorithms for stress monitoring need to account for
625 pregnancy-specific stressors and capture both positive and
626 negative stress events, and combine them into a meaning-
627 ful and accurate integrated measure of stress.
628 C. Longitudinal Monitoring
629 Continuous monitoring is required throughout all
630 phases of pregnancy to monitor lifestyle markers, and
631 identify changes in lifestyle behaviors and early risk factors
632 associated to pregnancy complications. Achieving longitu-
633 dinal monitoring requires the development of truly
634 wearable, miniaturized and non-intrusive technology
635 solutions that lower the barrier to usability. Significant
636 progress has been achieved in the last few years to reduce
637 the size while increasing battery lifetime of wearable
638 systems [82]. Recent technology achievements in flexible
639 and stretchable electronics [83], advanced chip integration
640 and algorithm development are paving the way towards
641 low-cost, multi-sensory conformable systems that provide
642 accurate and reliable information in every day situations
643 and longitudinally. This is a general challenge in the
644 successful deployment of wearable sensor technologies for
645 consumer health, which discussion goes beyond the scope
646 of this paper.
647 VII. OPPORTUNITIES
648 In this section we discuss the opportunities around
649 applying wearable sensor technologies and machine
650 learning techniques to pregnancy monitoring. While
651 wearable technologies for lifestyle monitoring are getting
652 more widespread in the general population, very few
653 pregnancy specific wearable health solutions have been
654 introduced so far. We chose three areas where technolog-
655 ical innovations can have a significant impact during
656 pregnancy: motivating behavioral change, an integrated
657 approach to tracking and crowdsourcing clinical research.
658 A. Technology Enabled Behavioral Change
659 Maintaining a healthy lifestyle takes on a different
660 meaning during pregnancy than during other phases of
661 life. The definition of appropriate lifestyle targets should
662 be carefully assessed when designing behavior change
663 interventions for pregnant women. The American College
664 of Obstetrician and Gynecologists (ACOG) defines guide-
665 lines for physical activity and recommends 30 min of
666 activity per day [31]. The Institute Of Medicine (IOM) sets
667 healthy target ranges for weight gain in pregnancy [17].
668 These guidelines must be built into behavior change
669 interventions targeted specifically at pregnant women
670 because applying lifestyle targets or goals from the general
671population may put pregnant women at risk for pregnancy
672complications.
673Mobile and wearable sensor technologies have already
674been adopted to trigger behavioral change [5]. Deploying
675systems not only able to provide targeted messaging, but
676also to quantify lifestyle behavior, can open new opportu-
677nities for personalized interventions and optimized feed-
678back loops, eventually driving higher compliance. The
679design of reliable monitoring technologies and accurate
680algorithms is crucial to providing feedback and objective
681monitoring that can serve as the basis for quantitative and
682sensor based behavior change interventions.
683B. Integrated Approach
684Quantifying and tracking lifestyle behaviors using
685wearable technology requires an integrated approach.
686First, data streams from smartphone and wearable sensors
687should be integrated to improve the accuracy of data
688analytics aiming at quantifying lifestyle behaviors (see
689Fig. 4). For instance, physiological data from wearable
690sensors shall be integrated with location and movement
691data from smartphones to provide a more accurate and
692context aware estimation of physical activity and stress.
693Next, lifestyle behaviors should be considered together
694(holistically), because they are interrelated. The impact of
695activity, stress, sleep and diet on health and outcome
696during pregnancy should be looked at and analyzed
697considering the complex relations between variables,
698instead of isolating single ones.
699Additional challenges arise when trying to measure and
700impact multiple lifestyle behaviors at once. Lifestyle
701markers are not unique to lifestyle behaviors. HR and
702HRV for instance are influenced by both physical activity
703and stress. Advanced data analytics are needed that can
704help understand how each lifestyle behavior is contribut-
705ing to specific physiological variables and avoid confound-
706ing factors. Recent developments in machine learning and
707pattern recognition are leading to the development of
708context-aware algorithms that are able to interpret
709physiological data and user behavior differently, depend-
710ing on the specific situation relative to the user. Such
711algorithms should work together to provide accurate
712assessment of lifestyle behaviors based on multiple
713lifestyle markers and other sources of information.
714Capturing lifestyle behavioral markers holistically by
715means of technological innovations will allow researchers
716to investigate how different lifestyles impact the develop-
717ment of pregnant complications. This approach can
718ultimately help in defining more precise and personalized
719lifestyle-related risk factors and therefore help individuals
720in better manage health during pregnancy (see Fig. 5).
721C. Crowdsourcing Clinical Research
722The growing deployment of smartphone apps and
723wearable devices generates an exponentially increasing
724amount of wellness and health data. This consumer-generated
Penders et al.:Wearable Sensors for Healthier Pregnancies
8Proceedings of the IEEE |
IEEE
Proof
725 data represent a mine of new information that can be
726 leveraged for new clinical research, a new practice recently
727 qualified as crowdsourcing clinical research [84]. The rapid
728 proliferation of online communities as well as mobile
729 phones apps targeting the pregnant population is a clear sign
730of the penetration of crowdsourced data in the pregnancy
731field. From a technology perspective, the possibility to
732gather lifestyle behavior data on pregnant populations much
733larger than what can be done by means of standard clinical
734trials opens numerous opportunities.
Fig. 4. Integrated approach to lifestylebehavioralmarkers assessment.Raw data from multiple sourcesare used by context-aware data analytics
to estimate lifestyle behavior markers representative of a person’s lifestyle.
Fig. 5. Integrated approach to personalized lifestyle-related risk factor assessment during pregnancy. Lifestyle behavior markers allow
monitoring different lifestylebehaviors, such asphysical activity,sleep, stress and diet.Studying how thesedifferent behaviors affect pregnancy
complications canenable refinementand discovery of personalized lifestyle-related risk factors, ultimatelyenabling betterrecommendations for
pregnant women.
Penders et al.:Wearable Sensors for Healthier Pregnancies
|Proceedings of the IEEE 9
IEEE
Proof
735 Crowdsourcing studies, either organized by researcher
736 and clinical staff, or simply participant-driven, can
737 generate large-scale datasets tackling important limitations
738 of current observational studies, where findings are often
739 limited by sample size. Large-scale datasets can allow for
740 more stratification of pregnant women and potentially
741 help in better understanding the link between different
742 lifestyle behaviors and pregnancy outcomes, related to
743 women specific characteristics and environment. The
744 rapid pace at which technology is advancing, and the
745 reduced delays compared to standard clinical studies can
746 significantly reduce the time to unveil new medical
747 findings or clarify current inconsistencies in results
748 reported from small studies in different subpopulations.
749 One of the main challenges of crowdsourcing clinical
750 research is scientific rigor. Especially when data are
751 consumer-generated or derived by participant-driven
752 studies, often lacking reference points or traditional study
753 design, data reliability can be questionable. Proper
754 methods should be developed to account for such
755 unsupervised data collection settings. One approach
756 when dealing with the inherent lack of reliability and
757 repeatability of today’s wearable sensor data is to define
758 quality of information metrics to systemize and automate
759 the evaluation of the quality of consumer-generated data.
760 Early approaches have been introduced that either look at
761 application agnostic metrics [84] or application specific
762 metrics [85]. Application agnostic metrics provides more
763 suitable information in case the data need to be used across
764 applications. In the case of pregnancy monitoring,
765 application specific metrics may be preferred as they
766 provide intelligible information that can be related to the
767 parameters of interest. Furthermore, algorithms intended
768 to take consumer-generated data as input should be
769 designed with the necessary robustness to deal with
770 missing or low-quality data. Other challenges such as
771 security, privacy and data ownership will need to be
772 addressed before crowdsourcing can be deployed as a
773 mainstream technique for clinical research. These chal-
774 lenges are generic to any consumer driven approaches to
775 new health marker discovery. Pregnancy here may present
776 some advantages because it spans over a well defined
777 period of time during which the woman is closely followed
778 by her doctor, therefore allowing for semi-supervised
779 reporting of reference points and outcomes. Despite all
780 these challenges, crowdsourcing health research promises
781 to accelerate clinical discovery in complex and still largely
782 misunderstood physiological and medical fields such as
783 pregnancy.
784 VIII. SUMMARY
785 Lifestyle-related pregnancy complications are increasingly
786 impacting maternal and infant health. Adopting healthy
787 lifestyle behaviors during pregnancy reduces pregnancy
788 complications, thus leading to improved outcomes and
789long lasting reduced risk of chronic diseases for both the
790mother and the baby. Accurately monitoring lifestyle
791behaviors is essential to understand a woman’s behavior,
792define efficient strategies to induce behavior change, and
793evaluate the efficiency of behavior change interventions.
794Recent advances in wearable sensor technologies offer new
795opportunities to accurately measure lifestyle behaviors and
796drive behavior change accordingly.
797This paper focused on four lifestyle behaviors (physical
798activity, sleep, stress, and diet and weight management)
799that have been shown to influence pregnancy complica-
800tions such as hypertension, preeclampsia, gestational
801diabetes mellitus, depression, and to impact the risk of
802preterm birth and other pregnancy outcomes. Innovations
803in wearable sensor technologies have been shown to
804enable the measurement of lifestyle behaviors in the
805general population. Specific solutions tailored to the needs
806and the peculiarities of pregnant woman are yet to be
807developed.
808Important challenges need to be taken into account
809when applying wearable sensor technologies to accurately
810measure behavior in pregnant women. First, pregnancy is
811associated with physiological adaptations that impact the
812cardiovascular, metabolic, autonomic and musculoskeletal
813systems. New algorithms for physical activity, stress and
814sleep monitoring in pregnant women must be developed to
815account for normal fluctuations in physiological signals
816associated with the different stages of pregnancy. Physi-
817ological changes in turn lead to lifestyle adaptations
818throughout pregnancy. Algorithms need to be tailored to
819the changing lifestyle habits of pregnant women, and
820provide the necessary resolution to enable accurate
821measurements. Finally, the need for continuous and
822longitudinal monitoring requires the development of truly
823wearable, miniaturized and nonintrusive technology solu-
824tions that lower the barrier to usability and increase
825compliance.
826Addressing these challenges will lead to new opportu-
827nities. Monitoring lifestyle behavior with pregnancy
828specific wearable technologies paves the way for quanti-
829tative and sensor-based strategies to behavior change
830interventions. It leads to personalized interventions and
831optimized feedback loops that have the potential to drive
832higher compliance and efficiency. Wearable sensors offer a
833holistic view on lifestyle behaviors and open the door to a
834two-level integrated approach to behavior modeling,
835integrating multiple data sources and multiple behaviors.
836This can provide new insights on how different lifestyles
837impact the development of pregnant complications. This
838approach can ultimately help in defining more precise and
839personalized lifestyle-related risk factors.
840In the hand of consumers, wearable sensors unlock
841exponentially growing datasets that contain a new source
842of knowledge largely unexploited to date. These crowd-
843sourced datasets tackle important issues of current
844observational studies such as limitations in sample size
Penders et al.:Wearable Sensors for Healthier Pregnancies
10 Proceedings of the IEEE |
IEEE
Proof
845 and population stratification. Important challenges in
846 trustworthiness, privacy, security, and ownership of the
847 data need to be addressed. However, rapid technology
848 advancements and the exponential leverage associated to
849 crowdsourcing data collection offer a unique opportunity
850to accelerate clinical research, clarify current inconsisten-
851cies in research, understand and validate the relation
852between lifestyle behaviors and pregnancy outcomes, and
853eventually discover new markers for pregnancy complica-
854tions and outcomes. h
REFERENCES
[1] Centre for Disease Control and Prevention.
[Online]. Available: http://www.cdc.gov/
reproductivehealth/MaternalInfantHealth/
[2] A. Elixhauser and L. M. Wier, ‘‘Complicating
Conditions of Pregnancy and Childbirth,
2008,’’ HCUP Stat. Brief, no. 113, May 2011.
[3] R. Luoto, M. F. Mottola, and
L. Hilakivi-Clarke, ‘‘Pregnancy and lifestyle:
Short- and long-term effects on mother’s and
her children’s health,’’ J. Pregnancy, vol. 2013,
2013, Art. ID 537526.
[4] L. A. Hagen Haakstad, N. Voldner, and K. Bo,
‘‘Stages of change model for participations in
physical activity during pregnancy,’’ J.
Pregnancy, vol. 2013, 2013, Art. ID 193170.
[5]
AQ3 C. Free and G. Philipps, ‘‘The effectiveness
of mobile-health technology-based health
behaviour change or disease management
interventions for health care consumers: A
systematic review,’’ PloS Medicine, vol. 10,
no. 1, p. 21001362, 2013.
[6] E. Derbyshire and D. Dancey, ‘‘Smartphone
medical applications for women’s health:
What is the evidence-base and feedback?’’
Int. J. Telemed. Applicat., vol. 2013, 2013,
Art. ID. 782074.
[7] J. Villar, L. Say, A. M. Gulmezoglu,
M. Meraldi, M. D. Lindheimer, A. P. Betran,
and G. Piaggio, ‘‘Eclampsia and pre-eclampsia:
A health problem for 2000 years,’’ in
Pre-eclampsia, H. Critchly, A. MacLean,
L. Poston, and J. Walker, Eds. London, U.K.:
RCOG Press, 2003, pp. 189–207.
[8] B. E. Metzger and D. R. Coustan,
‘‘Proceedings of the fourth international
work-shop-conference on gestational diabetes
mellitus,’’ Diabetes Care, vol. 21, no. 2,
pp. B1–B167, 1998.
[9] B. Lewis, M. Avery, E. Jennings, N. Sherwoo d,
B. Martinson, and A. L. Crain, ‘‘The effect of
exercise during pregnancy on maternal
outcomes: Practical implications for
practice,’’ Amer. J. Lifestyle Med., vol. 2, no. 5,
pp. 441–455, 2008.
[10] J. M. Pivarnik et al., ‘‘Impact of physical
activity during pregnancy and postpartum on
chronic disease risk,’’ J. Amer. College Sports
Med., vol. 38, no. 5, pp. 989–1006, 2006.
[11] W. Pien and R. J. Schwab, ‘‘Sleep disorders
during pregnancy,’’ Sleep, vol. 27, no. 7,
pp. 1407–1417, 2004.
[12] A. C. Huizink, P. G. Robles de Medina,
E. J. H. Mulder, G. H. A. Visser, and
J. K. Buitelaar, ‘‘Stress during pregnancy is
associated with developmental outcome in
infancy,’’ J. Child Psychol. Psych., vol. 44,
pp. 810–818, 2003.
[13] S. Thangaratinam et al., ‘‘Interventions to
reduce or prevent obesity in pregnant women:
A systematic review,’’ Health Technol Assess.,
vol. 16, no. 31, pp. 1–191, 2012.
[14] J. C. Y. Louie, T. P. Markovic, N. Perera,
D. Foote, P. Petocs, G. P. Ross, and
J. C. Brand-Miller, ‘‘A randomized controlled
trial investigating the effects of a low -glycemic
index diet on pregnancy outcomes in
gestational diabetes mellitus,’’ Diabetes Care,
vol. 34, no. 11, pp. 2341–2346, 2011.
[15] M. Sexton and J. Hebel, ‘‘A clinical trial of
change in maternal smoking and its effect
on birth weight,’’ JAMA, vol. 251, no. 7,
pp. 911–915, 1984.
[16] G. Carson et al., ‘‘Alcohol use and pregnancy
consensus clinical guidelines,’’ JOGC, vol. 32,
no. 8, pp. S1–S32, 2012, supplement 3.
[17] Institute of Medicine, ‘‘Weight gain during
pregnancy: Reexamining the guidelines,’’
Rep. Brief, 2009.
[18] Q. Fu and B. D. Levine, ‘‘Autonomic
circulatory control during pregnancy in
humans,’’ Sem. Reproductive Med., vol. 27,
no. 4, p. 330, 2009.
[19] S. H. Ghuge, V. V. Patil, R. G. Latti, and
K. D. Thorat, ‘‘A comparative study of
cardiovasculare sympathetic activity in three
trimesters of pregnancy,’’ Pravara Med. Rev.,
vol. 3, no. 1, 2011.
[20] C.-D. Kuo, G.-Y. Chen, M.-J. Yang, H.-M. Lo,
and Y.-S. Tsai, ‘‘Biphasic changes in
autonomic nervous activity during pregnancy,’’
Brit. J. Anaesth., vol. 84, pp. 323–329, 2000.
[21] S. Hunter and S. C. Robson, ‘‘Adaptation of
the maternal heart in pregnancy,’’ Brit. Heart
J., vol. 68, no. 6, p. 540, 1992.
[22] J. F. Clapp, III, ‘‘Maternal heart rate in
pregnancy,’’ Amer. J. Obst. Gynecol., vol. 152,
no. 6, pp. 659–660, 1985.
[23] K. Ueland, M. Novy, E. Peterson, and
J. Metcalfe, ‘‘Maternal cardio-vascular dy-
namics. iv. the influence of gestational age on
the maternal cardiovascular response to pos-
ture and exercise,’’ Amer. J. Obst. Gynecol.,
vol. 104, no. 6, p. 856, 1969.
[24] A. Poppas et al., ‘‘Serial assessment of the
cardiovascular system in normal pregnancy
role of arterial compliance and pulsatile
arterial load,’’ Circulation, vol. 95, no. 10,
pp. 2407–2415, 1997.
[25] S. Fujitani and M. R. Baldisseri,
‘‘Hemodynamic assessment in a pregnant and
peripartum patient,’’ Crit. Care Med., vol. 33,
no. 10, pp. S354–S361, 2005.
[26] K. Melzer, Y. Schutz, M. Boulvain, and
B. Kayser, ‘‘Pregnancy-related changes in
activity energy expenditure and resting
metabolic rate in Switzerland,’’ Eur. J. Clin.
Nutr., vol. 63, no. 10, pp. 1185–1191, 2009.
[27] M. W. Carpenter, S. P. Sady, M. A. Sady,
B. Haydon, D. R. Coustan, and
P. D. Thompson, ‘‘Effect of maternal
weight gain during pregnancy on exercise
performance,’’ J. Appl. Physiol., vol. 68, no. 3,
pp. 1173–1176, 1990.
[28] M. Pernoll, J. Metcalfe, T. L. Schlenker,
J. E. Welch, and J. A. Matsumoto, ‘‘Oxygen
consumption at rest and during exercise in
pregnancy,’’ Respir. Physiol., vol. 25, no. 3,
pp. 285–293, 1975.
[29] A. P. Heenan, L. A. Wolfe, and G. Al Davies,
‘‘Maximal exercise testing in late gestation:
Maternal responses,’’ Obst. Gynecol., vol. 97,
no. 1, pp. 127–134, 2001.
[30] S. Sady, M. Carpenter, M. Sady, B. Haydon,
B. Hoegsberg, E. Cullinane, P. Thompson, and
D. Coustan, ‘‘Prediction of vo2max during
cycle exercise in pregnant women,’’ J. Appl.
Physiol., vol. 65, no. 2, pp. 657–661, 1988.
[31] R. Artal, R. Wiswell, and B. Drinkwater,
‘‘Exercise in pregnancy,’’ in Proc. Amer. Congr.
Obst. Gynecol., Baltimore, MD, USA, 1991.
[32] K. R. Evenson and F. Wen, ‘‘Measuring
physical activity among pregnant women
using a structure one-week recall
questionnaire: Evidence for validity and
reliability,’’ Int. J. Behav. Nutr. Phys. Activ.,
vol. 7, p. 21, 2010.
[33] J. Liu, S. N. Blair, Y. Teng, A. R. Ness,
D. A. Lawlor, and C. Riddoch, ‘‘Physical
activity during pregnancy in a prospective
cohort of british women: Results from the
avon longitudinal study of parents and
children,’’ Eur. J. Epidemiol., vol. 26, no. 3,
pp. 237–247, 2011.
[34] J. C. Dempsey, T. K. Soren sen, M. A. Williams,
I. M. Lee, R. S. Miller, E. E. Dashow, and
D. A. Luthy, ‘‘Prospective study of gestational
diabetes mellitus risk in relation to maternal
recreational physical activity before and
during pregnancy,’’ Amer. J. Epidemiol.,
vol. 159, no. 7, pp. 663–670, 2004.
[35] . [Online]. Available: http://www.aasmnet.
org/
[36] D. D. Costa, N. Rippen, M. Dritsa, and
A. Ring, ‘‘Self-reported leisure-time physical
activity during pregnancy and relationship to
psychological well-being,’’ J. Psychosom. Obst.
Gynecol., vol. 24, no. 2, pp. 111–119, 2003.
[37] K. Armstrong and H. Edwards, ‘‘The effects of
exercise and social support on mothers
reporting depressive symptoms: A pilot ran-
domized controlled trial,’’ Int. J. Mental Health
Nurs., vol. 12, pp. 130–138, 2003.
[38] L. J. Miller, ‘‘Postpartum depression,’’ JAMA,
vol. 287, pp. 762–765, 2002.
[39] A. J. Ferreira, ‘‘Emotional factors in prenatal
environment,’’ J. Nervous Ment Dis., vol. 141,
pp. 108–118, 1965.
[40] S. A. Gore, D. M. Brown, and D. S. West,
‘‘The role of postpartum weight retention
in obesity among women: A review of
the evidence,’’ Ann. Behav. Med., vol. 26,
pp. 149–159, 2010.
[41] K. Raatikainen, N. Heiskanen, and
S. Heinonen, ‘‘Transition from Overweight to
Obesity Worsens Pregnancy Outcome in a
BMI-dependent Manner,’’ Obesity, vol. 14,
no. 1, pp. 165–171, 2006.
[42] A. Fraser, K. Tilling, C. Macdonald-Wallis,
N. Sattar, M. J. Brion, L. Benfield, and
D. A. Lawlor, ‘‘Association of maternal weight
gain in pregnancy with offspring obesity and
metabolic and vascular traits in childhood,’’
Circulation, vol. 121, no. 23, pp. 2557–2564,
2010.
[43] L. A. Haakstad, I. Gundersen, and K. Bo,
‘‘Self-reporting compared to motion monitor
in the measurement of physical activity during
pregnancy,’’ Acta Obstet Gynecol. Scand.,
vol. 89, no. 6, pp. 749–756, 2010.
[44] S. Berntsen, S. N. Stafne, and S. Morkved,
‘‘Physical activity monitor for recording
energy expenditure in pregnancy,’’ Acta
Obstet. Gynecol. Scand., vol. 90, no. 8,
pp. 903–907, 2011.
[45] A. Pantelopoulos and N. Bourbakis, ‘‘A survey
on wearable biosensor systems for health
monitoring,’’ in Proc. IEEE 30th Ann. Int. Conf.
Eng. Med. Biol. Soc., 2008, pp. 4887–4890.
Penders et al.:Wearable Sensors for Healthier Pregnancies
|Proceedings of the IEEE 11
IEEE
Proof
[46] A. Pantelopoulos and N. G. Bourbakis,
‘‘A survey on wearable sensor-based systems
for health monitoring and prognosis,’’ IEEE
Trans. Syst., Man, Cybern., Part C: Applicat.
Rev., vol. 40, no. 1, pp. 1–12, 2010.
[47] Y. Hao and R. Foster, ‘‘Wireless body sensor
networks for health-monitoring applications,’’
Physiol. Measur., vol. 29, no. 11, p. R27, 2008.
[48] S. Brownsell, D. Bradley, S. Blackburn,
F. Cardinaux, and M. S. Hawley, ‘‘A systematic
review of lifestyle monitoring technologies, ’’ J.
Telemed. Telecare, vol. 17, no. 4, pp. 185–189,
2011.
[49] A. G. Bonomi and K. R. Westerterp,
‘‘Advances in physical activity monitoring and
lifestyle interventions in obesity: A review,’’
Int. J. Obesity, vol. 36, no. 2, pp. 167–177,
2011.
[50] K. M. Smith, L. M. Lanningham-Foster,
G. J. Welk, and C. G. Campbell, ‘‘Validity of
the SenseWear Armband to predict energy
expenditure in pregnant women,’’ Med. Sci.
Sports Exercise, vol. 44, no. 10, pp. 2001–
2008, 2012.
[51] A. G. Bonomi, A. H. Goris, B. Yin, and
K. R. Westerterp, ‘‘Detection of type,
duration, and intensity of physical activity
using an accelerometer,’’ Med. Sci. Sports
Exercise, vol. 41, no. 9, pp. 1770–1777, 2009.
[52] M. Altini, J. Penders, R. Vullers, and O. Amft,
‘‘Personalizing energy expenditure estimation
using physiological signals normalization
during activities of daily living,’’ Physiol.
Measurement, vol. 35, no. 9, p. 1797, 2014.
[53] K. A. Lee, M. E. Zaffke, and G. McEnany,
‘‘Parity and sleep patterns during and after
pregnancy,’’ Obstetrics & Gynecology, vol. 95,
no. 1, pp. 14–18, 2000.
[54] K. A. Lee and C. L. Gay, ‘‘Sleep in late
pregnancy predicts length of labor and type of
delivery,’’ Amer. J. Obst. Gynecol., vol. 191,
no. 6, pp. 2041–2046, 2004.
[55] Z. Liu, L. Jiang, W. Chen, K. I. Kitamura, and
T. Nemoto, ‘‘Characterization of maternal
body movement during sleep before and
after parturition,’’ in Proc. IEEE 2011
Annu. Int. Conf. Eng. Med. Biol. Soc., 2011,
pp. 1503–1506.
[56] M. A. Williams, R. S. Miller, C. Qiu,
S. M. Cripe, B. Gelaye, and D. Enquobahrie,
‘‘Associations of early pregnancy sleep
duration with trimester-specific blood
pressures and hypertensive disorders in
pregnancy,’’ Sleep, vol. 33, no. 10, p. 1363,
2010.
[57] C. Hedman, T. Pohjasvaara, U. Tolonen,
A. Suhonen-Malm, and V. V. Myllyla
¨, ‘‘Effects
of pregnancy on mothers’ sleep,’’ Sleep Med.,
vol. 3, no. 1, pp. 37–42, 2002.
[58] F. L. Facco, J. Kramer, K. H. Ho, P. C. Zee,
and W. A. Grobman, ‘‘Sleep disturbances in
pregnancy,’’ Obst. Gynecol., vol. 115, no. 1,
pp. 77–83, 2010.
[59] K. Shiga, K. Murata, and H. Kodama, ‘‘Effects
of sleep disturbances during pregnancy on
cardiac autonomic modulation in the resting
state,’’ Int. J. Gynecol. Obst., vol. 119, no. 2,
pp. 149–153, 2012.
[60] F. K. Sahin, G. Koken, E. Cosar, F. Saylan,
F. Fidan, M. Yilmazer, and M. Unlu,
‘‘Obsructive sleep apnea in pregnancy
and fetal outcome,’’ Amer. J. Obstet. Gynecol.,
vol. 202, no. 6, p. 552.e1-7, 2010.
[61] N. de Vicq, F. Robert, J. Penders,
B. Gyselinckx, and T. Torfs, ‘‘Wireless Body
Area Network for Sleep Staging,’’ in Proc. Int.
Conf. Biologic. Circuits Syst., 2007.
[62] S. Lee, L. Yan, T. Roh, S. Hong, and H.-J. Yoo,
‘‘The Smart Patches and Wearable Band
(W-Band) for comfortable sleep monitoring
system,’’ in Proc. IEEE 2011 Annu. Int. Conf.
Eng. Med. Biol. Soc., 2011, pp. 6915–6918.
[63] X. Long, P. Fonseca, R. Haakma, R. M. Aarts,
and J. Foussier, ‘‘Spectral boundary adaptation
on heart rate variability for sleep and wake
classification,’’ Int. J. Artif. Intell. Tools, vol. 23,
no. 3, 2014.
[64] J. M. Kelly, R. E. Strecker, and M. T. Bianchi,
‘‘Recent developments in home
sleep-monitoring devices,’’ ISRN Neurol.,
vol. 2012, 2012, Art. ID 768794, 10 pages.
[65] S. Cohen, T. Kamarck, and R. Mermelstein,
‘‘A global measure of perceived stress,’’ J.
Health Social Behav., pp. 385–396, 1983.
[66] J. A. DiPietro, M. M Ghera, K. Costigan, and
M. Hawkins, ‘‘Measuring the ups and downs
of pregnancy stress,’’ J. Psychosom. Obstet.
Gynecol., vol. 25, no. 3–4, pp. 189–201, 2004.
[67] M. Seligman, T. Steen, N. Park, and
C. Peterson, ‘‘Positive psychology progress:
Empirical validation of interventions,’’ Amer.
Psychol., vol. 60, pp. 410–421, 2005.
[68] C. H. Chen, H. M. Chen, and T. H. Huang,
‘‘Stressors associated with pregnancy as
perceived by pregnant women during three
trimesters,’’ The Kaoshiung J. Med. Sci., vol. 5,
no. 9, pp. 505–509, 1989.
[69] A. Kazi, Z. Fatmi, J. Hatcher, U. Niaz, and
A. Aziz, ‘‘Development of a stress scale for
pregnant women in the South Asian context:
The A-Z Stress Scale,’’ East Mediterr. Health J.,
vol. 15, no. 2, pp. 353–361, 2009.
[70] A. K. Shea, M. V. Kamath, A. Fleming,
D. L. Streiner, K. Redmond, and M. Steiner,
‘‘The effect of depression on heart rate
variability in pregnancy,’’ Clin. Autonom. Res.,
vol. 18, no. 4, p. 203, 2008.
[71] G. Pal, P. Shyma, S. Habeebullah, P. Pal,
N. Nanda, and P. Shyjus, ‘‘Vagal withdrawal
and sympathetic overactivity contribute to the
genesis of early-onset pregnancy-induced
hypertension,’’ Int. J. Hypertension, 2011.
[72] G. Pal, P. Shyma, S. Habeebullah, P. Shyjus,
and P. Pal, ‘‘Spectral analysis of heart rate
variability for early prediction of
pregnancy-induced hypertension,’’ Clin.
Exper. Hyperten., vol. 31, no. 4, pp. 330–341,
2009.
[73] J. A. Healey and R. W. Picard, ‘‘Detecting
stress during real-world driving tasks using
physiological sensors,’’ IEEE Trans. Intell.
Transport. Syst., vol. 6, no. 2, pp. 156–166,
2005.
[74] E. Jovanov, A. O’Donnell Lords, D. Raskovic,
P. G. Cox, R. Adhami, and F. Andrasik,
‘‘Stress monitoring using a distributed
wireless intelligent sensor system,’’ IEEE Eng.
Med. Biol. Mag., vol. 22, no. 3, pp. 49–55,
2003.
[75] J. Wijsman, B. Grundlehner, H. Liu,
H. Hermens, and J. Penders, ‘‘Towards mental
stress detection using wearable physiological
sensors,’’ in Proc. IEEE 2011 Eng. Med. Biol.
Soc. Annu. Int. Conf., 2011, pp. 1798–1801.
[76] M. Kusserow, O. Amft, and G. Troster,
‘‘Modeling arousal phases in daily living
using wearable sensors,’’ IEEE Trans. Affect.
Comput., vol. 4, no. 1, pp. 93–105, 2013.
[77] O. Amft and G. Tro
¨ster, ‘‘On-body sensing
solutions for automatic dietary monitoring,’’
IEEE Perv. Comput., vol. 8, no. 2, pp. 62–70,
2009.
[78] Y. Dong, A. Hoover, J. Scisco, and E. Muth,
‘‘A new method for measuring meal intake
in humans via automated wrist motion
tracking,’’ Appl. Psychophysiol. Biofeedback,
vol. 37, no. 3, pp. 205–215, 2009.
[79] M. Katz, D. Kroll, Y. Shapiro, N. Cristal, and
I. Meizner, ‘‘Energy expenditure in normal
labor,’’ Israel J. Med. Sci., vol. 26, no. 5,
pp. 254–257, 1990.
[80] C. C. Yang, T.-C. Chao, T. B. Kuo, C.-S. Yin,
and H. I. Chen, ‘‘Preeclamptic pregnancy is
associated with increased sympathetic and
decreased parasympathetic control of HR,’’
Amer. J. Physiol.-Heart Circ. Physiol., vol. 47,
no. 4, p. H1269, 2000.
[81] A. F. Saftlas, N. Logsden-Sackett, W. Wang,
R. Woolson, and M. B. Bracken, ‘‘Work,
leisure-time physical activity, and risk of
preeclampsia and gestational hypertension,’’
Amer. J. Epidemiol., vol. 160, no. 8,
pp. 758–765, 2004.
[82] R. F. Yazicioglu, T. Torfs, P. Merken,
J. Penders, V. Leonov, R. Puers, and
C. Van Hoof, ‘‘Ultra-low-power biopotential
interfaces and their applications in wearable
and implantable systems,’’ Microelectron. J.,
vol. 40, no. 9, pp. 1313–1321, 2009.
[83] J. A. Rogers, T. Someya, and Y. Huang,
‘‘Materials and mechanics for stretchable
electronics,’’ Science, vol. 327, no. 5973,
pp. 1603–1607, 2010.
[84] C. Bisdikian, L. M. Kaplan, M. B. Srivastava,
D. J. Thornley, D. Verma, and R. I. Young,
‘‘Building principlesfor a quality of information
specification for sensor information,’’ in Proc.
12th Int. Conf. Inf. Fusion, 2009, pp. 1370–1377.
[85] I. Armenti, P. Asare, J. Su, and J. Lach, ‘‘A
methodology for developing quality of
information metrics for body sensor design,’
in Proc. Conf. Wireless Health, 2012, p. 2.
ABOUT THE AUTHORS
855 Julien Penders, photogr aph and bio graphy not available at the time of
856 publication.
857 Marco Altini, photograph and biogr aphy not ava ilable at th e time of
858 publication.
859Chris Van Hoof, phot ograph and biography not available at the time of
860pu blicatio n.
861Eric Dy,photographandbiographynotavailableatthetimeof
862pu blicatio n. AQ4
Penders et al.:Wearable Sensors for Healthier Pregnancies
12 Proceedings of the IEEE |
... The role of mHealth during pregnancy is being increasingly investigated [19][20][21][22][23]. Using apps to support pregnant women enhances the traditional pregnancy care model by providing additional educational information and empowering women to look after their own health [21,[23][24][25][26]. Sensors and monitoring technologies that automatically track specific health indicators have been integrated into mHealth solutions to support pregnant women's self-care behaviors [22,[27][28][29]. ...
... However, none of these have been specifically designed for pregnancy care. While research has explored women's and clinicians' views of mHealth in pregnancy, there is still a knowledge gap regarding the preferences of mHealth monitoring among pregnant women and their clinicians as well as the suitability of mHealth monitoring for different conditions [24,29,30,33]. Recent studies have highlighted the importance of patient-centered design and behavior decision research in the development of mHealth solutions for pregnancy [19,23,28]. ...
... Potential Interests mHealth technologies for health and lifestyle monitoring have been used in the general population. There is a growing interest in introducing mHealth solutions to support the pregnancy journey, which is the period of a woman's life that involves significant physiological changes and potential risks [29,44]. In this study, we examined the interests and perceptions of women and clinicians regarding the use of mHealth for health and well-being monitoring during pregnancy. ...
Article
Background Mobile health (mHealth) technologies, such as wearable sensors, smart health devices, and mobile apps, that are capable of supporting pregnancy care are emerging. Although mHealth could be used to facilitate the tracking of health changes during pregnancy, challenges remain in data collection compliance and technology engagement among pregnant women. Understanding the interests, preferences, and requirements of pregnant women and those of clinicians is needed when designing and introducing mHealth solutions for supporting pregnant women’s monitoring of health and risk factors throughout their pregnancy journey. Objective This study aims to understand clinicians’ and pregnant women’s perceptions on the potential use of mHealth, including factors that may influence their engagement with mHealth technologies and the implications for technology design and implementation. Methods A qualitative study using semistructured interviews was conducted with 4 pregnant women, 4 postnatal women, and 13 clinicians working in perinatal care. Results Clinicians perceived the potential benefit of mHealth in supporting different levels of health and well-being monitoring, risk assessment, and care provision in pregnancy care. Most pregnant and postnatal female participants were open to the use of wearables and health monitoring devices and were more likely to use these technologies if they knew that clinicians were monitoring their data. Although it was acknowledged that some pregnancy-related medical conditions are suitable for an mHealth model of remote monitoring, the clinical and technical challenges in the introduction of mHealth for pregnancy care were also identified. Incorporating appropriate health and well-being measures, intelligently detecting any abnormalities, and providing tailored information for pregnant women were the critical aspects, whereas usability and data privacy were among the main concerns of the participants. Moreover, this study highlighted the challenges of engaging pregnant women in longitudinal mHealth monitoring, the additional work required for clinicians to monitor the data, and the need for an evidence-based technical solution. Conclusions Clinical, technical, and practical factors associated with the use of mHealth to monitor health and well-being in pregnant women need to be considered during the design and feasibility evaluation stages. Technical solutions and appropriate strategies for motivating pregnant women are critical to supporting their long-term data collection compliance and engagement with mHealth technology during pregnancy.
... Various studies have shown that pregnant women use apps and wearable technologies and find them to be useful sources of HI (Hearn et al., 2013;Rodger et al., 2013;Kraschnewski et al., 2014;Thomas and Lupton, 2016), which offer exciting new opportunities (Lupton and Pedersen, 2016;Limbu et al., 2021). Various technologies enable women to communicate with each other, monitor the progress of their pregnancy, contribute to physical activity and alleviate anxiety through interactive, "professional" and private consultations (Tripp et al., 2014;Penders et al., 2015). A number of wearable technologies and apps can measure physical activity and other changes in lifestyle behaviours, as well as continuously selfmonitor physiologic parameters (e.g. ...
... A number of wearable technologies and apps can measure physical activity and other changes in lifestyle behaviours, as well as continuously selfmonitor physiologic parameters (e.g. blood pressure, blood sugar level, heartbeat, oxygen uptake, weight gain, sleep, physical activity, step counts, diet and fetus) (Penders et al., 2015;Lau et al., 2018;Cheung and Wu, 2019;Daly et al., 2019;Runkle et al., 2019) and provide useful HI in monitoring changes in pregnancy health status (Johnson, 2014;Lee, 2019). These applications are claimed to be empowering technologies that allow women to take control of their experiences and are part of a broader "'device-ification' of mothering" (Johnson, 2014). ...
Article
Purpose This study aims to increase the understanding of the pregnancy-related information behavior (IB) of pregnant women in Estonia. Design/methodology/approach The research involved a quantitative research methodology consisting of a semi-structured questionnaire. Data was collected from pregnant Estonian women through a self-administered Web-based questionnaire using a convenience sampling during the period from January to February 2019. A total of 300 pregnant women answered the questionnaire. The data were analysed using statistical analysis and the results of the study were compared with the results of previous studies. Findings The three topics on which information was most frequently sought were: fetal development, use of medicines during pregnancy and symptoms of pregnancy. The main sources of information were the internet and the midwife. The most reliable and valuable source of information was a midwife. Health-related information was sought mainly because it helped women make decisions related to pregnancy and childbirth. A number of factors facilitate the information seeking process. In addition, widespread access to the internet and technological skills facilitated IB. The following factors hindered the search for information: the controversy and/or ambiguity of information published on the internet and the time spent searching for information. Most women used wearable technologies during pregnancy. Research limitations/implications This study has several limitations. First, the weakness of online surveys is the potential lack of representativeness, as it excludes from the survey those who do not have access to or ability to use the internet for various reasons (Evans and Mathur, 2005; Limbu et al., 2021). Second, as most recruitment for the study took place online, there was a risk that those who did not use the internet could be excluded from the survey. Third, as the questionnaire was also shared in the Facebook news feed by the Women’s Clinic and Maternity Hospital of the East Tallinn Central Hospital, it may be that the respondents recruited through it more often used the support provided by medical professionals. Fourth, due to the volume limits of the study, it is not possible to present all the results of the study on the basis of socio-demographic characteristics and stage of pregnancy. Therefore, the findings cannot be generalized to the broader population and future studies should explore a larger and more representative populations. Practical implications This study will give some useful information to help to improve the services offered for pregnant women in Estonia. Social implications The findings of this study may inform how to better support this target group. Originality/value There is a lack of research in Estonia that focuses on the IB of pregnant women and this research fills this gap.
... depression or stress) and provide information to bring about changes and support therapeutic interventions [17]. Prevention, protection and rehabilitation interventions, such as lifestyle changes through longitudinal monitoring during pregnancy, should be implemented in order to improve emotional management and reduce negative consequences [18]. ...
Article
Full-text available
Background Technology-based approaches during pregnancy can facilitate the self-reporting of emotional health issues and improve well-being. There is evidence to suggest that stress during pregnancy can affect the foetus and result in restricted growth and preterm birth. Although a number of mobile health (mHealth) approaches are designed to monitor pregnancy and provide information about a specific aspect, no proposal specifically addresses the interventions in parents at risk of having small-for-gestational-age (SGA) or premature babies. Very few studies, however, follow any design and usability guidelines which aim to ensure end-user satisfaction when using these systems. Results We have developed an interactive, adaptable mHealth system to support a psycho-educational intervention programme for parents with SGA foetuses. The relevant results include a metamodel to support the task of modelling current or new intervention programmes, an mHealth system model with runtime adaptation to changes in the programme, the design of a usable app (called VivEmbarazo) and an architectural design and prototype implementation. The developed mHealth system has also enabled us to conduct a proof of concept based on the use of the mHealth systems and this includes data analysis and assesses usability and acceptance. Conclusions The proof of concept confirms that parents are satisfied and that they are enthusiastic about the mHealth-supported intervention programme. It helps to technically validate the results obtained in the other stages relating to the development of the solution. The data analysis resulting from the proof of concept confirms that the stress experienced by parents who followed the mHealth-supported intervention programme was significantly lower than among those who did not follow it. This implies an improvement in the emotional health not only of the parents but also of their child. In fact, the babies of couples who followed the mHealth-supported programme weigh more than the babies of couples under traditional care. In terms of user acceptance and usability, the analysis confirms that mothers place greater value on the app design, usefulness and ease of use and are generally more satisfied than their partners. Although these results are promising in comparison with more traditional and other more recent technology-based approaches.
... The use of advanced sensor technology in the field of maternal health helps to detect the complications of pregnancy in the early stage and to motivate and provide better health facilities to improve these complications. Currently, these wireless or wearable sensors are used to detect risk factors such as blood pressure, heart rate, breathing rate, etc., comprising the patient's lifestyle and behavior in the subclinical period of an unfavorable pregnancy result [54]. ...
Article
Full-text available
Currently, information and communication technology (ICT) allows health institutions to reach disadvantaged groups in rural areas using sensing and artificial intelligence (AI) technologies. Applications of these technologies are even more essential for maternal and infant health, since maternal and infant health is vital for a healthy society. Over the last few years, researchers have delved into sensing and artificially intelligent healthcare systems for maternal and infant health. Sensors are exploited to gauge health parameters, and machine learning techniques are investigated to predict the health conditions of patients to assist medical practitioners. Since these healthcare systems deal with large amounts of data, significant development is also noted in the computing platforms. The relevant literature reports the potential impact of ICT-enabled systems for improving maternal and infant health. This article reviews wearable sensors and AI algorithms based on existing systems designed to predict the risk factors during and after pregnancy for both mothers and infants. This review covers sensors and AI algorithms used in these systems and analyzes each approach with its features, outcomes, and novel aspects in chronological order. It also includes discussion on datasets used and extends challenges as well as future work directions for researchers.
... While physical activity is one of the cornerstones in diabetes management (American Diabetes Association, 2016;Harrison et al., 2016), automatic collection of physical activity data has gained minimal attention in GDM apps (Skar et al., 2018). Chan and Chen (2019) reported that interventions for increasing physical activity amongst pregnant women were more effective with wearable devices than without, and automatic self-tracking of lifestyle (e.g., nutrition, physical activity, and symptoms) has been argued to help in countering pregnancy-related health risks (Penders et al., 2015;Runkle et al., 2019). In T2D, data collection on physical activity has typically been based on steps recorded by a mobile phone, smartwatch, or a pedometer attached to the belt (Årsand et al., 2015, 2010). ...
Article
Full-text available
Gestational diabetes mellitus (GDM) has considerable and increasing health effects as it raises both the mother’s and offspring’s risk for short- and long-term health problems. GDM can usually be treated with a healthier lifestyle, such as appropriate dietary modifications and engaging insufficient physical activity. While telemedicine interventions requiring weekly or more frequent feedback from health care professionals have shown the potential to improve glycemic control amongst women with GDM, apps without extensive input from health care professionals are limited and have not shown to be effective. We aimed to improve the efficacy of GDM self-management apps by exploring desirable features in a review. We derived six desirable features from the multidisciplinary literature and we evaluated the state of implementation of these features in existing GDM apps. The results showed that features for increasing competence to manage GDM and for providing social support were largely lacking.
... As gravidity is a special body condition both medically and physically [17,18], we did not use the data of normal people to train the supervised machine learning algorithm for the activity recognition of gravidas. Instead, we collected a novel dataset of 10 physical activities from 61 gravidas who were at various stages of gravidity. ...
Article
Full-text available
We propose a physical activity recognition and monitoring framework based on wearable sensors during maternity. A physical activity can either create or prevent health issues during a given stage of pregnancy depending on its intensity. Thus, it becomes very important to provide continuous feedback by recognizing a physical activity and its intensity. However, such continuous monitoring is very challenging during the whole period of maternity. In addition, maintaining a record of each physical activity, and the time for which it was performed, is also a non-trivial task. We aim at such problems by first recognizing a physical activity via the data of wearable sensors that are put on various parts of body. We avoid the use of smartphones for such task due to the inconvenience caused by wearing it for activities such as “eating”. In our proposed framework, a module worn on body consists of three sensors: a 3-axis accelerometer, 3-axis gyroscope, and temperature sensor. The time-series data from these sensors are sent to a Raspberry-PI via Bluetooth Low Energy (BLE). Various statistical measures (features) of this data are then calculated and represented in features vectors. These feature vectors are then used to train a supervised machine learning algorithm called classifier for the recognition of physical activity from the sensors data. Based on such recognition, the proposed framework sends a message to the care-taker in case of unfavorable situation. We evaluated a number of well-known classifiers on various features developed from overlapped and non-overlapped window size of time-series data. Our novel dataset consists of 10 physical activities performed by 61 subjects at various stages of maternity. On the current dataset, we achieve the highest recognition rate of 89% which is encouraging for a monitoring and feedback system.
... High-risk complications are estimated to occur in 10 percent of pregnancies, and the evidence reveals the growing rate of high-risk pregnancies [1]. One of the most common complications that can occur during pregnancy is gestational diabetes mellitus (GDM) [2,3], accounting for more than 80% of diabetes cases during pregnancy [2]. ...
Article
Full-text available
This study attempted to review the evidence for or against the effectiveness of mobile health (m-health) interventions on health outcomes improvement and/or gestational diabetes mellitus (GDM) management. PubMed, Web of Science, Scopus, and Embase databases were searched from 2000 to 10 July 2018 to find studies investigating the effect of m-health on GDM management. After removing duplications, a total of 27 articles met our defined inclusion criteria. m-health interventions were implemented by smartphone, without referring to its type, in 26% (7/27) of selected studies, short message service (SMS) in 14.9% (4/27), mobile-based applications in 33.3% (9/27), telemedicine-based on smartphones in 18.5% (5/27), and SMS reminder system in 7.1% (2/27). Most of the included studies (n=23) supported the effectiveness of m-health interventions on GDM management and 14.3% (n=4) reported no association between m-health interventions and pregnancy outcomes. Based on our findings, m-health interventions could enhance GDM patients' pregnancy outcomes. A majority of the included studies suggested positive outcomes. M-health can be one of the most prominent technologies for the management of GDM.
Article
As the growth of the technology rises day to day but still we cannot able to overlook any wearable device which is the friendliest one for pregnant women. Our proposed system is a wearable device which monitors the health condition of expectant mothers’ and transmits data to the respective physician especially in rural areas. With these kind of real time wearable systems, doctors are able to provide higher quality medical services and more personalized healthcare to these women. This health monitoring system would allow a pregnant woman to interact with a physician with almost full functional capability. Preventive measure taken by continuoushealth monitoringof patient from early stages and guidance to avoid prenatal risks is the prime objective of this system. It is only applicable after 16 weeks of pregnancy. The health monitoring device constantly measures the body temperature and heartbeat of the womb and whenever there are fluctuations from the normal value it sends the information to gynecologist at remote place through GSM.The usage of these advanced technologies for pregnant women’s care facilitates optimal care to them and thereby pregnancy period mortality can be reduced substantially.
Chapter
The desire for efficient, timely, and cost-effective medical diagnostics has impelled biosensor technology and research progression. There are significant challenges in the development process of biosensors to meet the continually increasing demands. The fabrication and development process should produce high performance and yield without sacrificing cost-effectiveness. This need necessitates continuous and lean development methodologies to fabric, characterize, and miniaturize biosensors and biosensing systems to render more effective outcomes. Clinical biosensors combine detection and transduction units to detect chemical or biological substances or responses and transform them into electrical, optical, thermal, piezoelectric, or electrochemical signals. Most biosensors aim to detect a biological signal from a specific analyte to monitor the biological functions and environment. Another lesser-known class of biosensors, wearable biosensors for clinical applications, is used to measure biopotential, pathophysiological, or biological signals noninvasively. This chapter focuses on the lesser-known class of biosensors deployed using wearable form factors, emphasizing noninvasiveness, cost-effectiveness, and portability. This chapter also summarizes the technological challenges that have deterred the development and execution of biosensors., with a brief description of the opportunities and future outlook for biosensors in clinical and healthcare applications.KeywordsBiosensorsBioelectronicsHealthcareWearablesNanobiotechnologyVital signs
Chapter
mHealth (mobile Health) systems are turning out very useful in their application to the Life Sciences as they can assist users in several ways by acquiring, storing, visualizing and processing information. These systems consist of hardware and software especially designed to provide required functionalities and properties in order to satisfy stakeholders’ needs. This is also of special importance in health as end-users can find in the use of ICTs (Information and Communication Technologies) a key help to address and improve treatment and recovery processes of physical and mental health. This paper presents the design of a mHealth system, which is intended to support a psycho-educational programme devised and supervised by health professionals for pregnant women (and their partners) with SGA (Small for Gestational Age) foetuses. We pay special attention to adaptation and usability properties to tailor the psycho-educational programme and facilitate its use by patients through structured and updated health information and tasks. It tries to avoid obstacles for technology acceptance. As a result, a prototype of the mHealth system has been developed and used by pregnant women and partners in a proof of concept; pregnant women were motivated and informed positive feedback.
Article
Full-text available
In this paper we propose a generic approach to reduce inter-individual variability of different physiological signals (HR, GSR and respiration) by automatically estimating normalization parameters (e.g. baseline and range). The proposed normalization procedure does not require a dedicated personal calibration during system setup. On the other hand, normalization parameters are estimated at system runtime from sedentary and low intensity Activities of Daily Living (ADLs), such as lying and walking. When combined with activity-specific EE models, our normalization procedure improved EE estimation by 15 to 33% in a study group of 18 participants, compared to state of the art activity-specific EE models combining accelerometer and non-normalized physiological signals.
Article
Full-text available
Background. Smartphone medical applications have a major role to play in women's health with their roles being very broad, ranging from improving health behaviours to undertaking personalised tests. Objective(s). Using Medline, Web of Knowledge, and the PRISMA guidelines 15 randomized controlled trials (RCTs) were identified, with mobile interventions being tested on 1603 females, in relation to key aspects of health. Using a similar systematic approach an iPhone database search identified 47 applications (apps) developed to improve women's health. Findings. Ten RCTs used text messaging or app interventions to support weight loss, with significant improvements being observed in eight studies. For other aspects of women's health RCTs are needed to determine possible health benefits. iPhone store data analysis identified that a substantial number of women's health apps did not have star ratings or feedback comments (68 and 49 per cent, resp.), raising concerns about their validity. Conclusion. Peer-review systems, supporting statements of evidence, or certification standards would be beneficial in maintaining the quality and credibility of future health-focused apps. Patient groups should also ideally be involved in the development and testing of mobile medical apps.
Article
It is generally agreed that pregnant women snore more than those who are not pregnant. Snoring in pregnancy has been associated with hypertension and is a risk factor for both preeclampsia and intrauterine growth restriction. Obstructive sleep apnea (OSA) may also be a cause of fetal compromise, since other disorders associated with maternal hypoxia can lead to adverse fetal outcomes. In this prospective observational study, both polysomnography (PSG) and the nonstress test (NST) were performed in pregnant women whose questionnaire responses suggested symptoms of OSA. Participating in the study were 35 women in whom the Berlin Scale questionnaire demonstrated symptoms of OSA. OSA was diagnosed by PSG in 4 of these women (11.4%). Two women had mild and 2 had moderate OSA. The women with OSA were older than the others and had higher body mass indices and larger neck circumferences. Two of the 4 women with OSA had gestational diabetes and one had cardiovascular disease. Three of the 31 women who did not have OSA had chronic hypertension, and 2 each had gestational diabetes, cardiovascular disease, and pregnancy-induced hypertension (preeclampsia). In 3 of the 4 cases of OSA, fetal heart decelerations accompanied maternal blood desaturation. The newborn infants had lower Apgar scores and birth weights than infants whose mothers did not have OSA. Three of the 4 infants whose mothers had OSA were admitted to a newborn healthcare unit. The investigators believe that, because early diagnosis and treatment of OSA may lessen adverse outcomes, all pregnant women should be checked for symptoms of OSA and those who are suspected should be offered PSG.
Article
Research over the past 20 years has focused on the safety of physical activity during pregnancy. Guidelines for health care providers and pregnant/postpartum women have been developed front the results of these studies. The overwhelming results of most studies have shown few negative effects on the pregnancy of a healthy gravida, but rather, be beneficial to the maternal-fetal unit. Recently, researchers have begun to consider the role of maternal physical activity in a more traditional chronic disease prevention model, for both mother and offspring. To address the key issues related to the role of physical activity during pregnancy and postpartum on chronic disease risk, the American College of Sports Medicine convened a Scientific Roundtable at Michigan State University in East Lansing. MI. Topics included preeclampsia, gestational diabetes, breastfeeding and weight loss, musculoskeletal disorders, mental health, and offspring health and development.
Article
These clinical trial results are the first, to our knowledge, from a prospective, randomized, and controlled experiment demonstrating that a reduction of smoking during pregnancy improves the birth weight of the infant. Nine hundred thirty-five pregnant smokers were randomly assigned to treatment and control groups; the former received smoking intervention. At the eighth month of pregnancy, differences between the two groups in salivary thiocyanate level and reported smoking were statistically significant. For single, live births, the treatment group infants had a mean birth weight that was 92 g heavier and were 0.6 cm greater in length than the control group infants. The decrement in weight related to smoking cannot be fully explained by gestational age. The findings suggest that some fetal growth retardation can be overcome by the provision of antismoking assistance to pregnant women.(JAMA 1984;251:911-915)
Conference Paper
Body sensors networks (BSNs) are emerging technologies that are enabling long-term, continuous, remote monitoring of physiologic and biokinematic information for various medical applications. Because of the varying computational, storage, and communication capabilities of different components in the BSN, system designers must make design choices that trade off information quality with resource consumption and system battery lifetime. Given these trade-offs, there is the possibility that the information presented to the health practitioner at the end point may deviate from what was originally sensed. In some cases, these deviations may cause a practitioner to make a different decision from what would have been made given the original data. Engineers working on such systems typically resort to traditional measures of data quality like RMSE; however, these metrics have been shown in many cases to not correlate well with the notions of information quality for the particular application. Objective metrics of information distortion and its effects on decision making are therefore necessary to help BSN designers make more informed trade-offs between design constraints and information quality and to help practitioners understand the kind of information being produced by BSNs, on which they have to base decisions. In this paper, we present a general methodology for developing such metrics for various BSN applications, illustrate how this methodology can be applied to a real application through a case study, and discuss issues with developing such metrics.
Article
In this work, we introduce methods for studying psychological arousal in naturalistic daily living. We present an activity-aware arousal phase modeling approach that incorporates the additional heart rate (AHR) algorithm to estimate arousal onsets (activations) in the presence of physical activity (PA). In particular, our method filters spurious PA-induced activations from AHR activations, e.g., caused by changes in body posture, using activity primitive patterns and their distributions. Furthermore, our approach includes algorithms for estimating arousal duration and intensity, which are key to arousal assessment. We analyzed the modeling procedure in a participant study with 180 h of unconstrained daily life recordings using a multimodal wearable system comprising two acceleration sensors, a heart rate monitor, and a belt computer. We show how participants' sensor-based arousal phase estimations can be evaluated in relation to daily activity and self-report information. For example, participant-specific arousal was frequently estimated during conversations and yielded highest intensities during office work. We believe that our activity-aware arousal modeling can be used to investigate personal arousal characteristics and introduce novel options for studying human behavior in daily living.