Background: Quality antenatal care (ANC) is recognised as an opportunity for screening and early identification of pregnancy-related complications. In rural Ghana, challenges with access to diagnostic services demotivate women from ANC attendance and referral compliance, leading to absent or late identification and management of high-risk women. In 2016, an integrated diagnostic and clinical decision support system tagged ‘Bliss4Midwives’ (B4M), was piloted in Northern Ghana. The device facilitated non-invasive screening of pre-eclampsia, gestational diabetes and anaemia at the point-of-care. This study aimed to explore the experiences of pregnant women with B4M, and its influence on service utilisation (“pull effect”) and woman-provider relationships (“woman engagement”). Methods: Through an embedded study design, qualitative methods including individual semi-structured interviews and non-participant observation were employed. Interviews were conducted with 20 pregnant women and 10 health workers, supplemented by ANC observations in intervention facilities. Secondary data on ANC registrations over a one-year period were extracted from health facility records to support findings on the perceived influence of B4M on service utilisation. Results: Women’s first impressions of the device were mostly emotive (excitement, fear), but sometimes neutral. Although it is inconclusive whether B4M increased ANC registration, pregnant women generally valued the availability of diagnostic services at the point-of-care. Additionally, by fostering some level of engagement, the intervention made women feel listened to and cared for. Process outcomes of the B4M encounter also showed that it was perceived as improving the skills and knowledge of the health worker, which facilitated trust in diagnostic recommendations and was therefore believed to motivate referral compliance. Conclusions: This study suggests that mHealth diagnostic and decision support devices enhance woman engagement and trust in health workers skills. There is need for further inquiry into how these interventions influence maternal health service utilization and women’s expectations of pregnancy care. Keywords: mHealth, Antenatal, Woman-provider interaction, Prenatal screening, Ghana, Clinical decision support * Correspondence: email@example.com Athena Institute, Faculty of Science, Vrije Universiteit, Amsterdam, The Netherlands Department of Public Health, Institute of Tropical Medicine, Maternal and Reproductive Health Unit, Antwerp, Belgium Full list of author information is available at the end of the article © The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Abejirinde et al. BMC Pregnancy and Childbirth (2018) 18:209 Page 2 of 11 Background integrated diagnostic and Clinical Decision Support Sys- Worldwide, significant progress has been made to re- tem (CDSS) that enables non-invasive point-of-care duce maternal mortality and morbidity. However, many screening for pre-eclampsia, gestational diabetes and an- low and middle-income countries (LMICs) remain sig- aemia. B4M components include a non-invasive device for nificantly burdened by preventable maternal deaths and measuring haemoglobin via infrared sensors mounted on pregnancy-related morbidities [1, 2]. Substantive efforts a finger clip; a self-inflating blood pressure cuff; and an to improve maternal health globally is therefore a key automated reader for urinary protein and glucose through agenda under the Sustainable Development Goals [3, 4]. dipsticks. Data from all diagnostic devices are automatic- Furthermore, there is an observed shift from higher pro- ally or manually linked to an android tablet equipped with portions of maternal deaths due to direct causes such as decision support algorithms. Additionally, a traffic light haemorrhage and sepsis, to increasing mortality related system (red, orange and yellow colours) visually indicates to indirect causes such as heart disease and anaemia, i.e. risk category or referral urgency for pre-eclampsia, gesta- the obstetric transition [5, 6]. tional diabetes and anaemia, while prompting the health Quality antenatal care (ANC) is widely recognised as an worker on counselling and treatment. Before commencing important opportunity for screening and early identifica- the pilot project, users (midwives and community health tion of pregnancy-related complications such as pre- nurses) were trained on the technical and operational eclampsia, anaemia, and gestational diabetes [7, 8]. In functions of the device. Training also included modules cases of late or failure to diagnose, these conditions may on the principles of quality ANC and management of cause severe morbidity and mortality. It is estimated that pregnancy complications. haemorrhage, often in combination with anaemia, as well The project hypothesised that through this as preeclampsia, account for about 27 and 14% of all ma- mHealth-health worker interface, the interpersonal ternal deaths worldwide [9, 10]. Specifically, in Ghana, the process of care- a component of quality ANC- would im- Maternal Mortality Ratio (MMR) in 2015 was reported at prove. It was also expected that coupled with treatment or 319 per 100,000 live births with a large range of uncer- referral advice, the traffic light signalling would increase tainty (from 216 to 458) . Haemorrhage (39%), hyper- women’s engagement with the midwife, and women’s tensive disorders (35%) and unsafe abortions (7%) were compliance. Additionally, by being responsive to the needs the main direct causes, while indirect causes (26%) include of pregnant women for screening tests, the project antici- severe anaemia, diabetes and malaria . pated that use of B4M would increase demand and utilisa- In rural and remote areas of Ghana, comprehensive tion of ANC services. Summarily, the pilot project offered ANC is often hampered by staff shortages, lack of diag- the opportunity to assess the influence of mHealth on nostic equipment and supplies or weak referral linkages ANC demand or service utilization (“pull effect”), and on [13, 14]. Pregnant women are therefore often referred to the interpersonal process of care (“woman engagement”). distant health facilities, laboratories or private facilities This paper therefore explores the experiences of for routine diagnostics. The resultant loss of time and women exposed to the B4M device, to answer the re- money required to visit these facilities, demotivate search questions: i) How did women experience the use women from ANC attendance and referral compliance, of Bliss4Midwives during their routine antenatal care and may result in late identification and management of consultations? ii) What influence did Bliss4Midwives pregnancy-related complications. have on woman-provider relationships and on ANC ser- Mobile Health (mHealth) solutions, defined as “the vice utilization? provision of health services and information via mobile Due to inactive use of B4M devices in the Northern technologies”  have been reported as beneficial for im- region at the time of data collection, only findings in the proving ANC services in LMICs [16–18]. There is a grow- Upper East region are reported. ing body of literature stating that mHealth solutions can potentially reduce gaps and inefficiencies in health service Methods delivery including point-of-care screening, integration of Design records and streamlining care processes [19–22]. How- Qualitative methods including individual semi-structured ever, there is sparse evidence on the influence of such in- interviews and non-participant observation were terventions on service utilisation and woman engagement. employed. Data collection was embedded into a broader Using a mHealth intervention in rural Ghana as a case realist evaluation of midwives’ adoption and utilization of study, this paper aims to address this gap. the B4M device. In 2016, a one-year proof-of-concept study tagged Bliss4Midwives (B4M) was launched in two regions of Setting Northern Ghana- Upper East and Northern regions, with The Upper East region (UER) is one of the 10 adminis- a goal to improve ANC services. The B4M device is an trative regions in Ghana and is further sub-divided into Abejirinde et al. BMC Pregnancy and Childbirth (2018) 18:209 Page 3 of 11 districts and municipalities. The B4M project was B4M-ANC care process (see Additional file 2). This was piloted in Bawku municipal (one health facility- identi- done to understand the experience of use and nature of fied as Facility A) and Binduri district (three health facil- the mHealth-mediated consultation from different ac- ities- identified as B, C and D) in the UER. Both Bawku tors. Health workers were asked to share their percep- municipal and Binduri district are predominantly rural tions of the influence of B4M on women’s behaviour, and about half of their population is illiterate [23, 24]. compliance to referral or clinical recommendations, and Three B4M devices were placed in health facilities on a on women’s attitudes to the use of mHealth for ANC. fixed (permanently stationed in facilities A and D) or ro- As part of the realist evaluation in which this study was tational basis (one device shared by facilities B and C) embedded, workers were asked to respond to a and used to screen women attending ANC. colour-guided Likert five-point scale (ranging from ‘strongly disagree’ to ‘strongly agree’) in response to spe- Data collection cific questions about B4M use. The questionnaire sec- Exit interviews following ANC screening were con- tion of the interview explored perceptions of how ducted using a semi-structured interview guide. Ques- women responded to the device. Specifically: “I think tions explored women’s initial reactions to the device, more pregnant women are coming for ANC now that we its perceived benefits, their views on its (potential) ef- use B4M in the Health Facility” and “I think that preg- fect on ANC uptake behaviour (i.e. pull effect), qual- nant women follow my advice more now that I use B4M ity of service delivery and their desire for continued for consultation.” Responses to these questions were use of the device. therefore included in the analysis for this paper. We Interview guides for pregnant women were developed aimed to interview all health workers who had been in English language and translated to the local language- trained in B4M use, had used the device at least once Kusaal (see Additional file 1). These were discussed with post-training and were available at the time of data col- local program managers before piloting and subsequent lection. Due to staff rotations, not all midwives who modification (ambiguous words identified and refined, were initially trained were available to be interviewed exclusion of questions perceived to be culturally in- and not all health workers in each facility were trained appropriate). The first author and a trained female Re- on the use of B4M. A summary of the number and cat- search Assistant (RA) who is fluent in both English and egory of interviews is presented in Table 1. Kusaal conducted 20 interviews with pregnant women All interviews were audio recorded and later tran- in June 2017. Sixteen exit interviews were conducted in scribed and translated where necessary to English lan- the four health facilities immediately after ANC visits guage. Two independent individuals transcribed a with B4M use. Using disaggregated data from the project random number of interviews done in Kusaal to assure database and with the help of health facility staff, add- the quality of transcription and translation. Duration of itional women who had been exposed to the device dur- interviews with pregnant women was between 11 and ing past ANC visits were traced at community level and 34 min (median 21.36 min). Because interviews with invited to the health facility for retrospective interviews health workers were part of a broader evaluation object- (n = 4). Respondents were selected by convenience sam- ive, these lasted longer- from 35 to 91 min (median pling based on their attending ANC at the health facility 52.29 min). during data collection, or based on previous ANC In order to triangulate and validate findings from in- screening with the device and availability for interview. terviews, one researcher using a semi-structured check- Women were interviewed irrespective of gestation and list and observation guide conducted non-participant type of ANC visit (i.e. first or follow-up). Interviews were conducted in locations close to health facilities, but not Table 1 Categories of respondents per health facility in the immediate vicinity of ANC consultations. Based Health Workers Pregnant Women on the level of education and preference of the respond- Exit interviews Retrospective interviews ent, interviews with pregnant women were conducted in Health Facility English or Kusaal. Depending on the nature and unique Facility A 3 6 0 circumstances related to timing and respondents’ con- venience, the RA was sometimes the main interviewer Facility B 2 7 0 while the principal researcher observed or functioned as Facility C 3 0 1 the main interviewer, with the RA translating. Facility D 2 3 3 In addition, 10 semi-structured interviews in English TOTAL 10 16 4 were conducted with health workers (midwives and Facility A is a district hospital in Bawku Municipality and is the first level community health nurses) who operated the B4M device referral point for facilities B, C and D which are health centers in in intervention facilities and thereby were engaged in the Binduri district Abejirinde et al. BMC Pregnancy and Childbirth (2018) 18:209 Page 4 of 11 observations of ANC consultations in three intervention patterns. Qualitative data analysis was supported using sites (facilities A, B and D). These were documented NVivo qualitative data analysis Software; QSR Inter- using handwritten notes. Non-participant observation is national Pty Ltd. Version 11, 2014. stated here to mean that the researcher is not an active participant in the facility, but interacts occasionally with Ethical considerations the people in a non-intrusive way through questions and Approval for this study was granted by the Navrongo active listening, if the opportunity presents itself . Health Research Centre Institutional Review Board (Ap- Secondary data on first ANC visits (i.e. registrations) proval ID: NHRCIRB18) and the EMGO+ Scientific over a one-year period- from December 2015 to Decem- Committee of the Amsterdam Public Health Institute ber 2016 (i.e. 6 months before and 6 months after the (Reference Number: WC2017–026). Verbal and or writ- pilot commenced in June 2016) were extracted from ten consent was secured from all respondents. Consent health facility records to supplement findings on the per- was secured prior to all interviews, and respondents ceived pull effect of B4M. signed or appended their thumbprints to an informed consent form. Respondents for exit interviews received Data analysis transport reimbursement (equivalent value of USD$1 - All transcripts were read and a preliminary codebook $2). Health workers were not reimbursed since inter- was developed guided by the main themes explored in views were conducted at their work place. the interviews. Two researchers developed codes induct- ively. One researcher coded all transcripts while another Results coded a random number of six transcripts to test Between June 2016 and April 2017, 950 ANC screenings consistency of codes and support data analysis. Codes (including follow-up visits) were conducted with the were clustered into themes based on similar or recurring B4M device in the UER; 284 screenings from health Table 2 Socio-demographic characteristics of respondents Characteristic Pregnant women (n = 20) Health workers (n = 10) Age <18 10 18–24 8 1 25–29 3 2 30–34 1 5 35–39 2 1 40+ 1 1 Missing data 4 0 Number of Births 0–112 2–34 4–52 Missing data 2 ANC Observations Conducted Yes 14 No 6 Cadre Midwife 8 Enrolled Nurse 1 Community Health nurse 1 Years of Experience 0–4 4 5–9 5 >10 1 An auxiliary cadre similar to health assistants Abejirinde et al. BMC Pregnancy and Childbirth (2018) 18:209 Page 5 of 11 facility A, 69 in health facility B, 129 in C and 468 in D. are friendly and not shouting).However,these ratingscould Socio-demographic characteristics of interviewees are not be directly attributed to the B4M device, although a presented in Table 2. few respondents specifically stated that their satisfaction of the visit was because of the otherwise absent diagnostic ser- Women’s first impressions of B4M vice provided by B4M. Analysis showed that pregnant Pregnant women’s first impressions of the Bliss4Midwives women appreciated the device for detecting their health device ranged from fear and happiness to curiosity or a neu- problems and saving time and money that would have been tral disposition (Table 3). The non-invasive Haemoglobin otherwise expended on diagnostic referrals. Some women clip required women to insert an index finger into the also believed the device improved the knowledge of health peg-like clip and was a main reason for fear. This was be- workers and made them pay closer attention to specific as- cause some beneficiaries did not know what to anticipate on pects of the woman’shealth(Table 4). insertion. Two women specifically reported a feeling of dis- Prior to the intervention, diagnostic screening was not comfort during the screening. frequently conducted at the point of care and now con- Happiness was related to technological novelty or appre- stituted an additional task in the ANC workflow, neces- ciation for easier and additional diagnostic services. As sitating more time. For this reason, a drawback on the noted during ANC observations, some attendees were neu- perceived benefits of B4M use was that the ANC process tral- rarely questioning or initiating conversations with the took longer time, which made women impatient. Health health worker, and were passive actors in the consultation workers interpreted this impatience to mean that women process. Notably, the few women who demonstrated curi- were not accepting the device. Reports on time estimates osity were formally educated (completed basic primary edu- and perceived delays with B4M use ranged from 10 to cation). Health workers also corroborated reports of 30 min extra as compared to the standard workflow. happiness and fear. Technical difficulties (software freezes, slow response The ability of providers to clearly explain the device func- time), and procedural issues (low user dexterity with op- tion and help women understand its benefits facilitated a erating the device) were two main factors that contrib- positive disposition to B4M. Attempts by health workers to uted to delays. While there were reports of rare incorporate use of B4M while handling a high volume of occasions where one or two respondents voiced work in understaffed settings, sometimes resulted in situa- their displeasure about time delays, midwives more tions where not all women attending ANC were screened commonly observed signs of displeasure through with B4M. An unintended consequence of this adaptation body language. One midwife mentioned an extreme strategy was that while some women felt special for being se- case where the turnaround time was too long, and a lected, others felt otherwise excluded or came up with plaus- woman got restless and left the facility. ible explanations of why they were not screened with the Interestingly, while ANC attendees in the district hos- device. pital (facility A) were particularly grateful to B4M for saving time, their peers in the health centres did not Perceived benefits of B4M make this association. Contrarily, up to two women ANC respondents felt satisfied about the services offered at from health centres B, C and D associated the use of the facilities. When asked to rate the quality of their just B4M with a time delay. Observations revealed that the concluded ANC visit on a scale of 1–10, most respondents reported time saving benefits of B4M from respondents gave scores between 6 and 10 points with specific credit in facility A was connected to women otherwise having given to counselling and staff attitude (e.g. health workers to spend up to 2 hours at the hospital laboratory to Table 3 Initial reactions of ANC attendees to B4M device Reaction Supporting Quote Fear “It was because I had never seen it and that was my very first time of seeing it, so I was afraid” -Pregnant woman 1, Facility D Happiness “I was excited because I was hopeful that if I have any sickness in my system, the machine will let me know.” -Pregnant woman 3, Facility A “I would say for us here, when you hear that you have a device that is going to help you do something, they get excited…. some (women) get excited, but others…. it is the explanation that you have given them of what (the box) can do so that they can accept it.” -HCW 3, Facility A Neutral “I wasn’t happy and wasn’t afraid as well, I just knew it was part of the care they were providing” -Pregnant woman 6, Facility B “I thought using the box is only when it becomes necessary. So whether or not she uses it, it is the same… They are the professionals and know the best for me. As for me, I cannot tell.” -Pregnant woman 1, Facility A Curiosity “That, because after she said she was going to use it for this particular thing, when she was doing it, I was eager and ready to see whether I belong to any of the diseases she has mentioned or I’m free, so I was just waiting to see what the machine will tell me” -Pregnant woman 2, Facility B Abejirinde et al. BMC Pregnancy and Childbirth (2018) 18:209 Page 6 of 11 Table 4 Pregnant women’s Perceptions of B4M Factor Supporting Quote Service Provision “(I like) every part of the machine. I like it because the machine will tell you the right BP…. and then it will tell you the types of disease or the problems you are finding in your system.” -Pregnant woman 2, Facility B Improved Knowledge “(I think the midwife has more knowledge). Because it is through those things that she used to get the information about my condition. If not, (the results I got from) outside, she could have recorded it and give me the card to go away, but because of the machine she has took her time to also take me through this procedure to know whether it is really true that I have this disease and …she is sure that what they brought outside is comparing to her machine.” - Pregnant woman 2, Facility B Efficiency “It has helped to make her efficient. It is because the things we could have gone to the lab to do that requires so much time, this one is faster and easier.” -Pregnant woman 3, Facility A Time delays “I: So in your opinion which particular place do you spend a lot of time? R: At the machine I: So where do you think when worked on, will help reduce the time spent? At the records, the machine, the palpation area, the dispensary? R: No, just at the machine I: ….so the machine is what causes the delay R: Yes” -Pregnant woman, Facility D process routine tests. Except where other laboratory about the device, its functions and the procedures for tests had to be conducted, B4M therefore became a wel- the screening. On the other hand, the structured format come time-efficient alternative in the larger facility. of B4M compelled prolonged woman-provider inter- ANC workflow is usually structured in stations such action leading to improved counselling that was other- that women have to move from one point to another for wise previously rushed or absent. specific tests or different aspects of the consultation Observations and follow-up questions to workers (palpation, updating records, weight, height and blood revealed that most women were indifferent about the pressure check, collection of routine drugs). In this con- details of care and did not necessarily want to be in- text, workers felt the integrated function of B4M made volved. Women were reportedly more keen to receive ANC routine easier and more comfortable. However, assurances about their health status without spending one minor but crucial deviation from the perception that much time at the facility. In two of the three facilities B4M facilitated focused antenatal care was noted in ob- where B4M use was observed, the physical positioning servation visits showing that workflow was still broken of thedevicerelativetothe womanand B4Muser when women needed to provide their urine samples. prohibited (in the case of opposite seating) or was This usually necessitated going to an outhouse, which suboptimal for (e.g. screen not within field of vision) prolonged consultation time and was sometimes visual engagement by the woman. frowned upon by other workers within the facility who With a few exceptions, midwives would usually explain saw this as an interruption of the general workflow. the summary of the B4M encounter to the woman after In health facilities where workers helped women (re)- screening and use the tests results as an opportunity to negotiate their ANC experience and workflow, women educate them on diet, birth preparedness and pregnancy better understood the value and benefits of the box and care prior to scheduling the follow-up appointment. Al- this improved acceptability while diminishing complaints though this partly explained why the B4M encounter about time delays. As one health worker put it “they de- took longer, pregnant women also felt listened to and velop some love in it and every time they come, even if it more included in the care process. is wasting time, they will still wait”- an indication that women may have also adjusted their expectations and “Because I saw that everything, everything was ok. But factored in the extra time needed for use of the device. that place (i.e. another health facility) I can’t see. But here, everything is plain and I will know that… I can Women’s engagement with B4M see it. I see what are the problems, what they are Prior to using B4M for the first time, the project re- advising me on and other things.” -Pregnant woman 6, quired that health workers informed women about the Facility D. device using a structured information sheet. Following this, the woman was required to confirm consent by appending a thumbprint or signature. This mandatory Trust and compliance with B4M step served as an opening for interactive communica- Prior to project implementation, it was found that apart tion, but was not always carried out. Observations from challenges in the health system with referral link- showed that not all women received a full explanation ages, women would often not comply with referral Abejirinde et al. BMC Pregnancy and Childbirth (2018) 18:209 Page 7 of 11 recommendations. Reasons for this included doubt in women said interaction with the box was proof that ANC the severity of their problem, concerns about cost and was properly completed. time implications, and fear of being unattended or get- ting lost in an unfamiliar large health facility. “You know when I come and they don’t use it to One of the B4M intervention hypotheses was that the attend to me and I just get back home like that, it traffic-light module of the system (which should be visible doesn’t actually feel like you came” -Pregnant woman to the woman during ANC consultations) would facilitate 3, Facility B. women’s engagement and motivate referral compliance. Interviews explored the confidence of women in the diag- nostic decisions from B4M and the possibility that women Pull effect were more assured about the validity and necessity of re- An anticipated effect of B4M use in health facilities was ferral instructions if it came from a digital third party. that its innovativeness and the provision of diagnostic Analysis showed that most women were confident with options for women would encourage more first time the screening results and a couple of respondents noted (registrants) and follow-up ANC visits (attendants). In- that they would not have believed a referral recommenda- terviews revealed that all, except two health workers, felt tion if it had not come from the device. that the use of B4M in their health facility created a de- mand for ANC services. In some cases, it was believed “Using the box makes me actually believe that it is that B4M attracted women from other health centres to indeed true, because with the box she is telling me theirs. In response to the Likert-scale question “I think what she has seen” -Pregnant woman 7, Facility B. that more pregnant women are coming for ANC now that we have this device in our health facility”, two However, there were a few reports of women who health workers neither agreed nor disagreed, three while acknowledging benefits of the device, did not ne- slightly agreed and five were strongly in agreement. A cessarily trust its decisions or recommendations. In ne- minority, however, felt that more women were now reg- gotiating this digital trust, educated respondents were istering for ANC, but that this was unrelated to B4M. found most likely to critically reflect on results and rec- Health workers attributed this perception of a positive ommendations from B4M. pull effect to women spreading news about the device All health workers agreed to varying extents (eight within their communities, especially following a testi- strongly agreed and two slightly agreed) to the ques- mony of improved wellbeing after adhering to the B4M tion “I think pregnant women follow my advice more advice. now that I used the box for consultation”.Reasons in- cluded the perception that women’s respect for “When they saw it the first time and those (I screened) workers increased- an extension of additional skills and told them their problem, they were telling their acquired through B4M. Many health workers felt that colleagues- ‘Oh I went to the hospital, they are now women’s compliance to referral and counselling rec- having a machine’. So everybody wants to come and ommendations happened because the machine, which see the machine.” -HCW 2, Facility B. was seen as a more knowledgeable or accurate medium, was ‘demanding’ it. However, no pregnant woman reported discussing the Some health workers were also of the opinion that the availability of the device with any of their peers or learn- device mediated woman-provider relationship by enhan- ing about its presence in the facilities from a third party. cing trust. Although we did not have enough data points to deter- mine the extent to which number of ANC visits was as- “Even here the moment you work with them and they sociated with the presence of B4M in health facilities, we have trust with you, that is all. They believe and they compared descriptive data on ANC registrations (1st trust whatever you tell them…. And with the box, it visits) in UER intervention facilities 6 months prior to has even enhanced our work. They believe and they B4M (December 2015–May 2016) and 6 months follow- trust that if we tell them something, it is true because ing commencement of the pilot (July 2016–December the box has actually said it.” -HCW 2, Facility D. 2016). Interestingly, results did not show a dramatic in- crease in total visit numbers. In some instances, we no- Interestingly, some women personified the device, report- ticed a decline over time (Table 5). ing that the box “advised them” while others described When women were asked to comment on their dis- what “the machine said”. Just as the receipt of routine drugs position to return for ANC even in the absence of the (iron and folate) at the end of a visit was considered proof box, all noted that it would not influence their attend- that ANC had been effectively carried out, two pregnant ance of follow-up visits, although they may be curious or Abejirinde et al. BMC Pregnancy and Childbirth (2018) 18:209 Page 8 of 11 Table 5 Number of ANC registrants 6-months pre- and 6- trained professional. Given the intervention context- pre- months post- B4M initiation dominantly rural with high illiteracy levels, it was not sur- Facility A Facility B Facility C Facility D prising that only a couple of B4M beneficiaries who are also educated, were engaged actively in the care process Dec 2015- May 2016 792 199 79 171 . Future interventions of a similar nature should go July 2016- Dec 2016 647 205 88 135 beyond technical training; additional training in commu- Difference − 145 + 6 + 9 −36 nication and negotiation skills is necessary to facilitate ac- Data from June 2016 was not included in this analysis because the ceptability and to actively involve women in the care intervention commenced in the middle of the month Facility A is a district hospital in Bawku Municipality and is the first level process. referral point for facilities B, C and D which are health centers in This study found that although pregnant women could Binduri district not fully engage with the B4M-mediated process, they slightly disappointed if it was absent. This may have were still satisfied with the care received. The design of been due to the lack of alternatives as pointed out by B4M compelled health workers to follow structured one respondent: guidelines and procedures, such that they had to directly communicate with women. In so doing, the device “Of course I will (keep visiting this health facility). served as a platform for improved woman-provider Where else will I go? I was even coming here when the interaction. This made women feel heard and listened box was not here” - Client 1, Facility B. to, although an indirect consequence of this was longer consultation time. Missed opportunities for improved Although most women felt their future ANC behav- engagement occurred when workers did not take the iour would not be motivated by the presence of the de- entry point offered by the intervention to explain its vice, their expectation for future use was affirmative. function to end-beneficiaries. By virtue of providing diagnostic services, which would otherwise take a longer Discussion time or incur additional time and financial commitments This paper aimed to explore the experiences of pregnant of pregnant women, B4M served to overcome gaps in women screened with B4M device and its influence on unnecessary diagnostic referral while influencing the ANC service utilization and woman-provider relation- likelihood that women would comply with referral rec- ships. There has been significant focus on the feasibility ommendations. A note of warning, however, is that if and acceptability of mHealth from the lens of the tech- women become only compliant to clinical recommenda- nology user. However, less attention has been paid to tions when delivered by technology, there is risk of a di- the views and perceptions of the end beneficiary- per- lemma should woman-mHealth trust rank above spectives which are equally important in the outcome woman-provider trust. and long-term use of mHealth devices . While technology has shown promise in improving ac- The reported enthusiasm of women to B4M can be cess and efficiency of some health care processes, it is linked to the novelty effect of mHealth and are similar generally agreed that it is not yet a substitute for to reactions of health worker users in other studies [21, real-time face-to-face interaction and interpersonal rela- 27]. While there are mHealth innovations where the user tionship . There are mixed reports on how mHealth is the direct beneficiary of the functions offered [16, 28], may facilitate or hinder woman-provider interaction the nature and function of B4M necessitated a triad rela- [30–32]. A study on the perceptions of community tionship between the device, the health worker user and health workers in Brazil to the use of a data collection the direct beneficiary i.e. pregnant women. Findings that and transmission tool, reported social barriers to utiliz- some women were initially fearful of the device shed ing the cell phone intervention despite its reported ad- light on the importance of incorporating both technol- vantages . Such barriers included the negative ogy users and beneficiaries in the mHealth-assimilation impact of the tool on social interaction (i.e. interpersonal process. Specifically, reactions of fear in the B4M study connection) with pregnant women. Interrupted eye con- were directly related to device components, which can tact and passive involvement especially with illiterate be overcome with careful explanation and guidance from and older women impeded trust building and effective health workers. The potential risk of skipping this engagement. On the other hand, assessment of the mo- process of careful negotiation is the spread of myths and bile santé Nouna intervention in rural Burkina Faso misconceptions, which could hamper eventual integra- showed expectations of improved relationships between tion of mHealth into routine care. the health facility and the community with mHealth as The neutral disposition of some pregnant women to an intermediary . B4M may have been related to low literacy levels and feel- The meanings of expertise that women attached to the ings of inadequacy about questioning the competence of a B4M device while personifying it were notable. This Abejirinde et al. BMC Pregnancy and Childbirth (2018) 18:209 Page 9 of 11 made B4M seem like an active expert-participant in the impressive rates (up to 90%) of ANC attendance and consultation. A review of studies found that when bene- utilization of skilled delivery . While this study could ficiaries of care attribute a status of specialised or not prove increased ANC registrations due to B4M use, “higher up” knowledge to CDSS, it enhances their trust health workers had a strong perception of a pull effect. in results and recommendations . Woman-provider The counter-intuitive results from B4M facilities with trust has been shown to play a role in the continuity of continuous use versus rotating use, where registrations maternal care and in women’s preferences for facility de- marginally decreased in the former group compared to livery [34, 35] and ANC visits have been earmarked as the latter, further caution against hasty conclusions re- an entryway to establish such trust . Although some garding a pull effect and hint at other factors at play to health workers believed that the B4M device mediated explain these differences. However, some studies in simi- the woman-provider relationship by enhancing trust, we lar settings have reported improvements in health ser- could not ascertain the pattern of interactions regarding vice utilization, specifically ANC attendance, due to the trust effect. Were ANC attendees trusting health mHealth [28, 38, 39]. In addition to a small data set, the workers more because they used the device, or did it absence of agreement between workers’ perceptions and only enhance existing trust? Especially with B4M where routine data from the B4M sites may be due to incom- screening results required voluntary compliance from plete records, pre-existing trends of high ANC coverage women to seek further care, their trust in test results in study regions, or minimal options at competing health and the midwife’s recommendations was crucial. Screen- facilities from which clients could be pulled. In light of ing, which is only one of many steps in the chain of care updated guidelines recommending eight ANC contacts is therefore only relevant when followed by prompt and during pregnancy , and the continued expansion of proper management of the conditions detected. Unfortu- mHealth use in LMICs, it is beneficial to identify the nately, broader health system deficiencies such as lack of contexts and mechanisms by which mHealth positively resources and fragmented referral systems limit the cap- influences demand and service utilization. ability for a full chain of service delivery, which is be- yond the scope of what mHealth can do. Study limitations Interestingly, perceptions of time efficiency differed The B4M intervention was a short-term proof of concept between types of health facility. The mixed reports about and as such its use was confined to a limited number of the time efficiency of B4M use in health centres (facil- health facilities in study regions. Retrospective interview- ities B, C, D) and the hospital (facility A) appeared to be ing of some pregnant women as well as B4M users who linked to the workload associated with the latter. B4M had not used the device for up to 2 months before the allowed women bypass burdensome bottlenecks of the interview, introduced a risk of recall bias. We attempted care process in the bigger health facility. Studies have to counter this by using visual cues of the device and shown that service delivery advantages of mHealth vary prompting respondents with specific events associated depending on characteristics of the facility in which it is with device use. In addition we used multiple data sources placed and aspect of care it is used for. Reports from a for triangulation. By conducting women’s exit interviews multi-country study show greater time efficiency when in locations removed from the immediate service delivery CDSS was used in the delivery room than for ANC . environment, other biases (e.g. social desirability and con- This may explain why B4M was considered more time firmation) were minimized. A researcher who was not fa- efficient at the district hospital than at lower levels miliar with the local language conducted observations. where women had more direct contact with all points of This limited the ability to fully capture verbal elements of the ANC workflow without too many interruptions in what was observed. However, when possible, information the care process. Overall, our findings show that B4M gaps were filled by follow-up questions to health workers. saves time compared to the standard where women from Research bias due to the researchers’ presence during the health centres have to delay screening and travel a dis- initial sets of observations was inevitable. To counter the tance for tests. In this standard scenario, ANC consult- observer effect, multiple observation visits were con- ation becomes interrupted till a later date, after test ducted over a number of days, with the expectation that results have been acquired. These are time inefficient those being observed will resume their natural behaviour processes that may not be immediately appreciated by over time . Not all respondents declared their age and pregnant women and overall influence the timeliness of health record booklets were not readily accessible, leading treatment and continuity of care. to information gaps on socio-demographic characteristics. In the last decade, the effect of community focused in- formation, education and communication campaigns Conclusions about the value of ANC seems to have paid off. Even in Pregnant women’s experiences and perceptions of rural areas of Northern Ghana, studies have reported mHealth-supported service delivery can influence how Abejirinde et al. BMC Pregnancy and Childbirth (2018) 18:209 Page 10 of 11 the implementation process translates to desired care and explained to non-literate respondents after which they appended their thumbprints. outcomes. The B4M proof-of-concept showed that mHealth influences the trust ANC attendees have in Competing interests health workers and in referral recommendations, which RD and RAA work for organizations involved in the implementation of the Bliss4Midwives proof-of-concept and were actively involved in the project. may positively impact compliance of women to treat- JvR is a member of the editorial board (section editor) of BMC Pregnancy ment. Additionally, by fostering some level of engage- and Childbirth. Other authors declare no competing interests. ment, the intervention made women feel listened to and cared for. There is, however, a need for further inquiry Publisher’sNote into how mHealth shapes women’s expectations of ma- Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. ternal care and the perceived pull effect. Future studies should also focus on how beneficiaries, in addition to Author details mHealth users, directly or indirectly influence mHealth Athena Institute, Faculty of Science, Vrije Universiteit, Amsterdam, The Netherlands. Department of Public Health, Institute of Tropical Medicine, adoption, utilization and health outcomes. Maternal and Reproductive Health Unit, Antwerp, Belgium. ISGlobal, Hospital Clínic- Universitat de Barcelona, Barcelona, Spain. Simavi, Amsterdam, The Netherlands. Presbyterian Health Services-North, BolgatangaUpper East Additional files RegionGhana. Department of Obstetrics, Leiden University Medical Centre, Leiden, The Netherlands. Additional file 1: Question Guide: Client exit interviews. (DOCX 145 kb) Additional file 2: Question Guide: Health worker interviews. (DOCX 141 kb) Received: 24 January 2018 Accepted: 25 May 2018 Abbreviations References ANC: Antenatal care; B4M: Bliss4Midwives; CDSS: Clinical decision support 1. WHO. Trends in Maternal Mortality: 1990 To 2015: estimates by WHO, system; LMICs: Low and middle-income countries; mHealth: Mobile health; UNICEF, UNFPA. Geneva: World Bank Group and the United Nations RA: Research assistant; UER: Upper East Region Population Division; 2015. 2. Filippi V, Chou D, Ronsmans C, Graham W, Say L. Levels and Causes of Acknowledgements Maternal Mortality and Morbidity. In: Black RE, Laxminarayan R, Temmerman The authors would like to thank the Bliss4Midwives implementing M, Walker N, editors. Reproductive, Maternal, Newborn, and Child Health. organizations (Stichting Cordaid; Relitech B.V.; Ned. Org. voor Toegepast Third Edit ed. Washington (DC): The International Bank for Reconstruction Natuurwetenschappelijk Ond. TNO; Stichting Simavi; Stichting Enviu and Development / The World Bank; 2016. Nederland; Association of Church-based Development NGOs, Ghana; and 3. Requejo JH, Bhutta ZA. The post-2015 agenda: staying the course in Presbyterian Health Services-North, Ghana) for their support in carrying out maternal and child survival. Arch Dis Child. 2015;100:76–81. this study and in facilitating evaluation activities. We would also like to thank 4. Scott H, Isabella D. Accountability for improving maternal and newborn the local research assistants, translators and transcribers, study participants, health. Best Pract Res Clin Obstet Gynaecol. 2016;36:45–56. and the district administrators and health facility personnel involved. The first 5. Graham W, Susannah W, Peter B, Filippi V, Giorgia G, Sandra V, et al. author appreciates the additional financial support of Professor Vincent De Diversity and divergence: the dynamic burden of poor maternal health. Brouwere towards the stakeholders’ dissemination meeting organized at the Lancet. 2016;388:2164–75. https://doi.org/10.1016/S0140-6736(16)31533-1. end of the study. 6. Souza JP, Tunçalp Ö, Vogel JP, Bohren M, Widmer M, Oladapo OT, et al. Obstetric transition: the pathway towards ending preventable maternal Funding deaths. BJOG. 2014;121(Supplement 1):1–4. The Bliss4Midwives proof-of-concept was funded by the Life Sciences and 7. World Health Organisation. WHO Recommendations on Antenatal Care for Health for Development (LS&H4D) grant: Number- LSH14GH16, from the a Positive Pregnancy Experience. Geneva: WHO Press; 2016. http://apps. Dutch Ministry of Foreign Affairs (RVO). The LS&H4D grant, as well as the who.int/iris/bitstream/10665/250796/1/9789241549912-eng.pdf . Erasmus Mundus Joint Doctorate Fellowship Specific Grant Agreement 8. RequejoJH, BryceJ,Barros AJD,Berman P,BhuttaZ, Chopra M, et al. 2015–1595, which IOA is a beneficiary of, financed data collection activities. Countdown to 2015 and beyond: fulfilling the health agenda for Funding agencies had no role in study design, analysis, or preparation of the women and children. Lancet. 2015;385:466–76. https://doi.org/10.1016/ manuscript. S0140-6736(14)60925-9. 9. Say L, Chou D, Gemmill A, Tunçalp Ö, Moller A-B, Daniels J, et al. Global Availability of data and materials causes of maternal death: a WHO systematic analysis. Lancet Glob Health. Interview guides for health workers and pregnant women specific to this 2014;2:e323–33. https://doi.org/10.1016/S2214-109X(14)70227-X. study are included as additional files. The questionnaire and project datasets 10. Menéndez C, Romagosa C, Ismail MR, Carrilho C, Saute F, Osman N, et are available from the corresponding author on reasonable request. al. An autopsy study of maternal mortality in Mozambique: the contribution of infectious diseases. PLoS Med. 2008;5:e44. https://doi. Authors’ contributions org/10.1371/journal.pmed.0050044. IOA led study design, data collection, analysis, interpretation and writing; RD 11. WHO, UNICEF, UNFPA, World Bank Group, United Nations population contributed to design, analysis, interpretation and writing; RAA contributed division. Trends in Maternal Mortality: 1990 to 2015: estimates by WHO, to data collection and writing; AB, MZ, JvR and VDB supervised the study UNICEF, UNFPA, World Bank Group and the United Nations Population and contributed to design, interpretation and writing. All authors have read Division. Geneva, Switzerland: WHO Press; 2015. http://www.afro.who.int/ and approved the final manuscript. sites/default/files/2017-05/trends-in-maternal-mortality-1990-to-2015.pdf 12. Ghana Health Service. Family Health Division Annual Report for 2015. 2015. Ethics approval and consent to participate http://www.ghanahealthservice.org/downloads/2015_FAMILY_HEALTH_ Approval for this study was granted by the Navrongo Health Research DIVISION_ANNUAL_REPORT.pdf. Centre Institutional Review Board (Approval ID: NHRCIRB18) and the EMGO+ 13. Ayanore MA, Pavlova M, Groot W. Focused maternity care in Ghana: results Scientific Committee of the Amsterdam Public Health Institute (Reference of a cluster analysis. BMC Health Serv Res. 2016;16:395. https://doi.org/10. Number: WC2017–026). Consent was secured prior to all interviews using an 1186/s12913-016-1654-5. informed consent form available in English and Kusaal languages. Literate 14. Amoakoh-Coleman M, Agyepong IA, Kayode GA, Grobbee DE, Klipstein- respondents read and signed the consent form, while the form was read Grobusch K, Ansah EK. Public health facility resource availability and Abejirinde et al. BMC Pregnancy and Childbirth (2018) 18:209 Page 11 of 11 provider adherence to first antenatal guidelines in a low resource setting in western Kenya. Int J Med Inform. 2015;84:207–19. https://doi.org/10.1016/j. Accra, Ghana. BMC Health Serv Res. 2016;16:505. https://doi.org/10.1186/ ijmedinf.2014.12.005. s12913-016-1747-1. 33. Duclos V, Yé M, Moubassira K, Sanou H, Sawadogo NH, Bibeau G, et al. 15. Hagan D, Uggowitzer S. Information and communication Technologies for Situating mobile health: a qualitative study of mHealth expectations in the Women’s and Children’s health- a planning workbook. Geneva: WHO Press; rural health district of Nouna, Burkina Faso. Heal Res Policy Syst. 2017; 2014. http://www.who.int/pmnch/knowledge/publications/ict_mhealth.pdf. 15(Suppl 1):47. https://doi.org/10.1186/s12961-017-0211-y. 34. Jack SM, DiCenso A, Lohfeld L. A theory of maternal engagement with 16. Oliveira-Ciabati L, Vieira CS, Franzon ACA, Alves D, Zaratini FS, Braga GC, et public health nurses and family visitors. J Adv Nurs. 2005;49:182–90. https:// al. PRENACEL – a mHealth messaging system to complement antenatal doi.org/10.1111/j.1365-2648.2004.03278.x. care: a cluster randomized trial. Reprod Health. 2017;14:146. https://doi.org/ 35. Sheppard VB, Zambrana RE, O’Malley AS. Providing health care to low- 10.1186/s12978-017-0407-1. income women: a matter of trust. Fam Pract. 2004;21:484–91. https://doi. 17. Feroz A, Perveen S, Aftab W. Role of mHealth applications for improving org/10.1093/fampra/cmh503 antenatal and postnatal care in low and middle income countries: a 36. Mensah N, Sukums F, Awine T, Meid A, Williams J, Akweongo P, et al. systematic review. BMC Health Serv Res. 2017;17:704. https://doi.org/10. Impact of an electronic clinical decision support system on workflow in 1186/s12913-017-2664-7. antenatal care: the QUALMAT eCDSS in rural health care facilities in 18. Amoakoh-Coleman M, Borgstein AB-J, Sondaal SFV, Grobbee DE, Miltenburg Ghana and Tanzania. Glob Health Action. 2015;8 https://doi.org/10.3402/ AS, Verwijs M, et al. Effectiveness of mHealth interventions targeting health gha.v8.25756. care workers to improve pregnancy outcomes in low- and middle-income 37. Gudu W, Addo B. Factors associated with utilization of skilled service countries: a systematic review. J Med Internet Res. 2016;18:e226. https://doi. delivery among women in rural northern Ghana: a cross sectional org/10.2196/jmir.5533. study. BMC Pregnancy Childbirth. 2017;17:159. https://doi.org/10.1186/ 19. Hall CS, Fottrell E, Wilkinson S, Byass P. Assessing the impact of mHealth s12884-017-1344-2. interventions in low- and middle-income countries – what has been shown 38. Atnafu A, Otto K, Herbst CH. The role of mHealth intervention on maternal to work? Glob Health Action 2014;7:10.3402/gha.v7.25606. doi:https://doi. and child health service delivery: findings from a randomized controlled org/10.3402/gha.v7.25606. field trial in rural Ethiopia. mHealth; Sept 2017. 2017. http://mhealth. 20. Aranda-Jan CB, Mohutsiwa-Dibe N, Loukanova S. Systematic review on what amegroups.com/article/view/16495. works, what does not work and why of implementation of mobile health 39. Shiferaw S, Spigt M, Tekie M, Abdullah M, Fantahun M, Dinant G-J. The (mHealth) projects in Africa. BMC Public Health. 2014;14:188. https://doi.org/ effects of a locally developed mHealth intervention on delivery and 10.1186/1471-2458-14-188. postnatal care utilization; a prospective controlled evaluation among health 21. Adepoju I-OO, Albersen BJA, De Brouwere V, van Roosmalen J, Centres in Ethiopia. PLoS One. 2016;11:e0158600. https://doi.org/10.1371/ Zweekhorst M. mHealth for clinical decision-making in sub-Saharan journal.pone.0158600. Africa: a scoping review. JMIR mHealth uHealth. 2017;5:e38. https://doi. 40. Liu F, Maitlis S, Observation N. In: Mills AJ, Durepos G, Wiebe E, editors. org/10.2196/mhealth.7185. Encyclopedia of case study research. London: SAGE; 2010. p. 610–2. https:// 22. Gbadamosi OS, Eze C, Olawepo OJ, Iwelunmor J, Sarpong FD, Ogidi GA, doi.org/10.4135/9781412957397.n229. et al. A Patient-Held Smartcard With a Unique Identifier and an mHealth Platform to Improve the Availability of Prenatal Test Results in Rural Nigeria: Demonstration Study. J Med Internet Res. 20:e18. https:// doi.org/10.2196/jmir.8716. 23. Ghana Statistical Service. 2010 Population and Housing Census, Regional Analytical Report, Upper East Region. 2013. http://www.statsghana.gov.gh/ docfiles/2010phc/2010_PHC_Regional_Analytical_Reports_Upper_East Region.pdf. 24. Ghana Statistical Service. 2010 Population and Housing Census, District Analytical Report, Binduri District. 2014. http://www.statsghana.gov.gh/ docfiles/2010_District_Report/Upper East/BINDURI.pdf. 25. Feng L, Maitlis S. Non-Participant Observation. In: Mills AJ, Durepos G, Wiebe E, editors. Sage Encyclopedia of case study research. Los Angeles: Sage Publications Ltd; 2010. 26. Latulippe K, Hamel C, Giroux D. Social health inequalities and eHealth: a literature review with qualitative synthesis of theoretical and empirical studies. J Med Internet Res. 2017;19:e136. https://doi.org/10.2196/jmir.6731. 27. Agarwal S, Perry HB, Long L-A, Labrique AB. Evidence on feasibility and effective use of mHealth strategies by frontline health workers in developing countries: systematic review. Trop Med Int Heal. 2015;20:1003– 14. https://doi.org/10.1111/tmi.12525. 28. Lund S, Nielsen BB, Hemed M, Boas IM, Said A, Said K, et al. Mobile phones improve antenatal care attendance in Zanzibar: a cluster randomized controlled trial. BMC Pregnancy Childbirth. 2014;14:29. https://doi.org/10. 1186/1471-2393-14-29. 29. Sen A. Health: perception versus observation : self reported morbidity has severe limitations and can be extremely misleading. BMJ Br Med J. 2002; 324:860–1. http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1122815/ 30. Håland E, Melby L. Negotiating technology-mediated interaction in health care. Soc Theory Heal. 2015;13:78–98. https://doi.org/10.1057/sth.2014.18. 31. Schoen J, Mallett JW, Grossman-Kahn R, Brentani A, Kaselitz E, Heisler M. Perspectives and experiences of community health workers in Brazilian primary care centers using m-health tools in home visits with community members. Hum Resour Health. 2017;15:71. https://doi.org/10. 1186/s12960-017-0245-9. 32. Vedanthan R, Blank E, Tuikong N, Kamano J, Misoi L, Tulienge D, et al. Usability and feasibility of a tablet-based decision-support and integrated record- keeping (DESIRE) tool in the nurse Management of Hypertension in rural
BMC Pregnancy and Childbirth – Springer Journals
Published: Jun 5, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.
Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera