Article Text

Use of quality improvement methodology to improve care of women with hypertensive disease in pregnancy and haemorrhage in Yemen (low-income, high-insecurity setting)
  1. Sabrina Das1,
  2. Emtiaz Abdulwahid2,
  3. Audrey Moisan3,
  4. Anna Kathrina De Jesus4
  1. 1Obstetrics and Gynaecology, Imperial College Healthcare NHS Trust, London, UK
  2. 2Taiz Houban Mother and Child Hospital, Medecins Sans Frontieres Operational Centre Amsterdam, Amsterdam, The Netherlands
  3. 3Maternity Department, Centre Hospitalier de Mayotte, Mamoudzou, France
  4. 4Paediatrics, San Juan De Dios Educational Foundation Inc, Pasay City, The Philippines
  1. Correspondence to Dr Sabrina Das; sabrina.das{at}


This safety improvement project used quality improvement methods in a Comprehensive Emergency Obstetric and Newborn Care facility in Yemen, managed by an international humanitarian non-governmental organisation. It is responsible for about 6500 deliveries per annum. Following a local review of maternal deaths and serious incidents in 2020–2021, care for women with severe hypertensive disease in pregnancy and postpartum haemorrhage were highlighted as potential areas for improvement. These are also the two most common reasons recorded for maternal mortality in Yemen (and in many low-income countries worldwide). We also wanted to foster an open and honest safety culture within the department that encouraged learning from error.

We used an inclusive approach in designing the improvements, with change ideas collated via frontline doctors, midwives and nurses working in the maternity unit. Data were collected via manual audit, and through routinely collected data. We focused on the following measures: number of incidents reported per month, documentation quality of antihypertensive prescriptions, fluid restriction practices in women with severe pre-eclampsia, number of minutes taken to control severe hypertensive episodes, postpartum haemorrhage identification rates and tranexamic acid stock usage. We evaluated the efficacy of team simulation training through precourse and postcourse questionnaires.

We found demonstrable improvement in our measures relating to treatment of women with hypertensive disorders of pregnancy, and in postpartum haemorrhage identification and treatment. Team simulation training was a difficult intervention to sustain but was received well with positive results during our test session. Incident reporting showed a temporary increase, but this effect was not sustained.

We concluded that quality improvement methodology is a valuable tool even in challenged healthcare settings such as this one, in an active conflict zone. Behaviour change in team culture and safety culture is harder to sustain and demonstrate without a long-term strategy.

  • quality improvement
  • obstetrics and gynecology
  • global Health

Data availability statement

All data relevant to the study are included in the article or uploaded as supplementary information.

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See:

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


  • Complications of hypertensive disease in pregnancy and from postpartum haemorrhage are common in low-income countries and form the leading causes of maternal death worldwide.

  • Chronic war and political instability can result in healthcare settings operating in conflict areas for years, with a subsequent need for clinical governance structures and opportunities to apply continuous quality improvement.


  • This study applies quality improvement methods in a low-income, high-insecurity context, aiming to improve the adherence to established clinical protocols in the management of pregnancy complicated by haemorrhage and hypertensive disorders.

  • We have shown the feasibility of doing quality improvement work in conflict zones, and the importance of applying rigour of method.


  • Even in low-resourced healthcare settings and those situated in challenging contexts such as conflict zones, basic principles of what is known on leadership, engagement and change management can and should be applied.

  • Investment in quality improvement training and expertise among clinical leadership in humanitarian settings may be helpful in establishing culture and rigour of continuous quality improvement.


Since March 2015, Yemen has been embroiled in a civil war. The chronic effects of war have left the health system collapsed and there is a reliance on international humanitarian agencies to deliver basic medical care.1 Taiz-Houban Mother and Child Hospital (MCH) is a maternity and child health secondary referral hospital managed by an international humanitarian non-governmental organisation. It has provided maternity care to the local population since the start of the conflict.

The socio-economic situation due to the war has only worsened due to the COVID-19 pandemic. Up to 80% of the population live below the poverty line. There are about 4 million internally displaced persons in Yemen, 92% of whom report not having any income at all or living with <25 000 Yemeni Riyals (US$ 40) per month.2

The poor primary care infrastructure and deprioritisation of public health needs during the war, together with barriers in access to care such as insecurity, affordability of transport (all care provided at MCH is free of charge) and cultural factors (use of traditional healers and delayed presentation) mean that the population of pregnancy women who attend MCH tend to have little or no antenatal care and often present in extremis.

There is no indication that the international non-governmental organisation will be able to withdraw support from Taiz Houban region, and thus there is a need to assess and maintain clinical quality standards at MCH. The short-term (4 months) availability of expertise in quality improvement (QI) at MCH through the arrival of an International Obstetrician and Gynaecologist (ObGyn) with a QI background presented an opportunity to apply this methodology to reduce variation of practice, raise standards of care and identify systemic safety issues in a socially, politically and economically challenging context.


The two main direct causes of maternal death in Yemen are haemorrhage (33%) and hypertensive disease of pregnancy (13%). Together, they comprise 46% of all maternal deaths.3

The rates of maternal mortality studied in 2016, after the onset of the war, have shown an increase in all Yemeni governorates. The national average of 213.4 deaths per 100 000 live births in 2016 had increased by 1.3% since 2013. In Taiz governorate, there was a 2.4% increase from 2013 to 2016.4

Measuring quality in the low-income, high-insecurity setting is challenging, let alone attempts at improving quality measures.5 6

MCH is a Comprehensive Emergency Obstetric and Newborn Care (CEmONC) facility. It is responsible for about 6500 deliveries per annum, with an approximate caesarean delivery rate of 13%. There is an en site neonatal unit that operates at the level of a special care baby unit (no invasive respiratory support), a therapeutic feeding centre and an emergency department that sees patients with trauma and injuries, as well as children and pregnant women. At the start of the COVID-19 pandemic, an alongside respiratory unit that sees children and pregnancy women with respiratory symptoms was opened. Patients who are thought to be unwell from COVID-19 are transferred to a referral hospital in the region. Intensive care facilities for mothers and babies are available in a nearby private hospital.

In our local review of maternal deaths at MCH, occurring between February 2020 and March 2021, we identified six maternal deaths; 4/6 were complicated by maternal haemorrhage and 2/6 were complicated by hypertensive disease. There were recurrent themes identified in review of these sentinel cases:

  • Delay in recognition of unwell women and subsequent delay in escalation of care.

  • Communication problems between team members and delayed involvement of the multidisciplinary team.

  • Poor fluid management.

Inadequate documentation, a lack of discipline with drug prescribing practice and slow response to acutely raised blood pressure were also highlighted through case review of serious incidents reported in 2020–2021 and were norms that were observed in the department.


As a maternity service, MCH is a mature service, and has been running over the past 7 years since the start of the conflict. The operational team in 2015 had been focusing on getting the structural elements in place (pharmacy, operating theatre, ward spaces, equipment, energy supply, water sanitation, human resources, etc). These are largely running smoothly now.

Regarding outcome measures for quality, the project has a data team that compiles outcome data for maternal mortality, neonatal mortality, admission rate from the maternity unit to the neonatal unit, transfer rate to intensive care and other outcome measures. These broad measures are difficult to use in the context as a measurement of quality as the hospital only receives high-risk cases. There is a strict admission criteria and obstetric triage system at the door to identify women who are low risk, who are redirected to other local health centres.

Thus, we have chosen to focus on key process measures where we feel maximum impact can be felt downstream in eventually improving maternal morbidity and mortality. We used a maternal death local review and root cause analysis process, and a similar process applied to clinical incident case reviews to determine which process measures we wished to focus on. In line with the Institute for Healthcare Improvement (IHI) Framework for Safe and Reliable Healthcare,7 we attempted to measure staff culture and key human factor indicators alongside our clinical process measures.

The IHI model for improvement8 is a simple and effective methodology for QI well established in healthcare settings. With our desire to shift the focus onto quality of care through evaluation of key process measures, we decided to use the IHI model and successive Plan-Do-Study-Act (PDSA) cycles to accomplish our objectives.

Global aims

  1. Fostering a safety culture among staff, with an open and honest environment that encourages learning from incidents, use of structured communication tools and accountability for unprofessional team behaviour.

  2. Improve the management of pre-eclampsia and eclampsia.

  3. Improve the management of postpartum haemorrhage (PPH).

Specific aims

  1. Increase incident reporting.

  2. Commence a programme of multiprofessional obstetric emergency drill training (with postcourse participants reporting increased confidence in key human factors skills such as communicating concerns and working well in a team).

  3. Decrease errors in drug prescription.

  4. Increase number of women with severe pre-eclampsia with hourly fluid input and output recorded, and fluid restriction carried out (in accordance with local protocols).

  5. Increase number of women with severe pre-eclampsia with acute hypertension controlled in a stepwise manner and in a systematic way (in accordance with local protocols).

  6. Increase number of women with PPH who have treatment in a systematic way, including use of tranexamic acid for estimated blood loss (EBL) >500 mL (in accordance with local protocols).

  7. Increase PPH identification to enable effective management.

All measures had to be manually collected as there is no electronic patient record. PPH rates and stock levels of tranexamic acid are routinely collected and reported data were pulled from monthly reporting. Other measures had to be collected via manual auditing, apart from measuring team safety culture, which was reliant on staff filling in a questionnaire before and after the multiprofessional obstetric emergency skills simulation training day.

  1. Number of incidents reported.

  2. Staff confidence in key safety behaviours (eg, “I feel confident to communicate safety concerns”, “I feel confident in working within the team”), self-rated on a Likert scale.

  3. Per cent of antihypertensive prescriptions with dose, timing and frequency clearly documented.

  4. Per cent of women with severe pre-eclampsia who were fluid restricted.

  5. Number of minutes taken to control severe hypertensive episode.

  6. PPH rates (via clinical coding data).

  7. Tranexamic acid stock usage.


The team involved in this improvement project comprised the International ObGyn who was present from March 2021 to July 2021, the International Midwife who was present from June 2021 to January 2022, the Hospital Director (International staff) who was present from June 2021 to February 2022 and the Medical Activity Manager (Yemeni national substantive staff member at MCH).

We tried to be inclusive in our approach in co-designing changes with staff, and constantly providing two-way feedback during daily ward rounds. Systemic issues were identified,

such as the difficulty in interpreting the drug chart due to its design, resulting in delayed or missed antihypertensive medication and the use of an out-of-date chart for magnesium sulfate prescription. Many inadequate practices were due to norms developing in the department and cultural challenges. For example, women were left on the floor on stretchers in the maternity department following transfer from the emergency room (ER), as the male stretcher carriers felt a pressure to leave the all-female maternity department as soon as possible. Following team conversations explaining the importance of getting women onto a trolley or into a delivery room, and explicit permission for male stretcher carriers to stay in the maternity department until this occurred, we observed some improvements.

Some clinical policies were introduced following consultation of the team leaders:

  • Pink (20-gauge) cannulae were frequently used for maternity patients despite the local guidelines stating that only large bore cannulae should be used in the context of PPH. These were removed from the department and replaced with large bore cannulae.

  • EBL was not routinely documented after every birth. EBL awareness posters aiding visual blood loss estimation were put on walls to try and remind staff to initiate the PPH protocol, including the administration of tranexamic acid9 once EBL exceeded 500 mL.

BRASS-V drapes are blood collection drapes with a calibrated pouch to assist in more accurate blood loss estimation and were suggested by staff who had used them in other facilities.

Finally, as with many other safety improvement initiatives, we found that an education programme was a key driver to create a culture of learning and improvement from clinical cases. We trialled multiprofessional simulation training, weekly case presentations, circulation of safety messages using WhatsApp (most staff did not have an email address) and used teaching sessions as opportunities to generate engagement, change ideas and ownership of interventions. We wanted to increase incident reporting, and hence simplified the reporting form, rebranding them as ‘Quality Reporting’ forms that frontline staff were encouraged to fill in only basic details of cases that managers could then investigate.

The driver diagram in figure 1 summarises the drivers and intervention ideas tried.


Online supplemental table 1 shows the results of our PDSA cycles.

Supplemental material

Quantitative data were not collected following July 2021 as this is when the International ObGyn left the project. The International Midwife remained till January 2022 and feedback when she left the project was that EBL estimation was a sustained practice and that tranexamic acid use continued but there were continual issues with inadequate stock as the ordering practices would take time to catch up with prescribing practices.

Blood pressure control and fluid management were improved subjectively but we do not have numerical audit data to confirm sustainment of initial improvements.

Feedback from the Hospital Director in February 2022 was that incident reporting fell back to pre-intervention rates, with only one incident reported (at the request of the Hospital Director rather than from a frontline staff member). Behavioural interventions such as encouraging systematic handover using the Situation-Background-Assessment-Recommendation (SBAR) tool and ensuring safe transfer of women from ER to maternity, were also difficult to measure improvement, and subjective assessment reported ongoing challenges with culture change.

We used run charts to plot and observe the effects of interventions on:

  • Number of incidents reported.

  • Tranexamic acid usage.

  • PPH rates.

We used a statistical process control chart to plot and observe the effects of interventions on the control of blood pressure (number of minutes taken to control acute hypertensive event).


The results of each intervention described in our ‘Study of the Intervention’ and PDSA cycles show that the 4-month presence of the International ObGyn created educational opportunities, forums for team discussions and some structure for a risk management strategy in the department. The subsequent arrival of an International Midwife to do further work in midwifery education reinforced messages that led to demonstrable improvement in the management of acute hypertension and in PPH identification (see the embedded charts in online supplemental file 1 for more details).

Staff were very receptive to the improvement programme, and it was not difficult to engage staff in providing improvement ideas. Weekly case presentations were usually carried out at 16:00–17:00 hours, and the night shift staff (who started at 17:00 hours) would come in early before their shifts to attend the presentation. Some staff members attended teaching sessions on their days off.

Interventions that had a clear message had more success in sustained improvement. For example:

  1. Give tranexamic acid when the EBL is >500 mL and start using mechanical methods to control bleeding.

  2. When the systolic blood pressure is ≥160 or the diastolic blood pressure is ≥110 (in the red zone on the vitals chart), give an antihypertensive and recheck in 30 min. If still in the red zone, give another antihypertensive. Escalate treatment according to flow chart.

These messages were already well established and written down in the local protocol, which all staff had access to. Transcribing them onto simple flowcharts and emphasising the importance of following them on daily ward rounds and through critical case presentations helped staff follow the clear instructions that they all already knew they should be doing.

We struggled more with problems that required a cultural shift, or a behavioural change. For example, introduction of the SBAR tool for handover. Despite the SBAR tool being an organisation-wide adopted team communication tool, the training for this was not well attended (being a virtual lecture delivered in English via video link) and staff found this form of communication difficult as it did not translate well into Arabic. Cultural issues, such as the barrier between men and women speaking to each other, are other challenges to consider when trying to improve team communication.

The idea of using multiprofessional obstetric emergency drill simulation training to open conversations between team members surrounding team working, handover and communication was limited by the ongoing lack of expertise in facilitation, simulation training and debrief. The International ObGyn had planned two sessions a month apart, with the aim to handover the work (including scenarios) to the International Midwife, and other national staff. The skills required to facilitate and conduct simulation training were harder to teach over this time period, and we only managed to do one session as the other had to be cancelled due to staffing shortages over Ramadan. The reflections from the single session carried out was documented in a blog article, which details the potential benefit of this approach, given more resource.10

BRASS-V drapes were a change idea proposed by one of the Yemeni national doctors, but we were unable to procure these. There were challenges particular to this context, where the supply chain is limited in a conflict area. Our issues with supply of tranexamic acid and other drugs were ongoing challenges.

Lessons and limitations

Our experience in doing this work found that broad principles of change management and leadership in more affluent contexts can also apply in this context. Psychologist Jonathan Haidt’s rider and the elephant analogy,11 and Chip and Dan Heath’s triad for behavioural change12 can be applied here:

  1. Direct the rider: clear flow charts, visual blood loss estimation posters and simple messaging helped staff reduce the time taken to control acute hypertensive episodes, increased used of tranexamic acid in the management of PPH and increased identification of PPH.

  2. Motivate the elephant: critical case presentations detailing the importance of blood pressure control and early identification of haemorrhage, including details of aftermath of incidents, built a clear case for change.

  3. Shape the path: creating an environment where change was easier, through re-design of drug charts and vital signs charts, meant that error was less likely.

Overwhelming feedback from Yemeni national staff was that they desired opportunity for continued professional development, hence the engagement with any clinical educational sessions, and that they valued on the floor presence by senior management. They had many ideas for improvement and required a forum to air these ideas, and permission to enact them. The idea of the “Leader as a Coach”,13 where leaders and managers should move away from “command and control” and instead use coaching and empowerment to drive improvement and performance, was found to be even more relevant in this context where the national staff are the stable human resource, while the international staff (mostly in management and leadership roles) are rotational, staying only 4–12 months. There are also many vacancies in the international roles due to visa processing difficulties, gender limitations (only female staff were allowed to work in the maternity department) and the COVID-19 pandemic.

Regarding project limitations, patients were not included in the design of this project. Patient and public involvement was deemed to be too challenging given the language and cultural barriers. We did use real patient stories to motivate and engage staff during the simulation training and during case presentations.

On reflection, patient involvement may have been feasible via the Health Promotion Team, who were a team of Yemeni staff members who collected feedback and complaints from patients.

Another study limitation was that we only have numerical audit data for the 3 months when the interventions were studied, and thus long-term sustainment cannot be confirmed.

Our only team member who had prior experience and a formal background in improvement was only present in the project for 4 months. Thus, when she left, team members continued to sustain the improvement initiatives, but were not able to conduct data collection. Subsequent feedback was given to the project only in subjective terms.

Although clinical expertise was of a high level both among our international team and our national team of staff, there was a lack of expertise in education and use of simulation training. The ambition to commence an obstetric drills team training programme did not take off due to the lack of continued expertise in education and simulation training.


This project attempted to use improvement methods to improve specific elements of clinical care in this very specific and challenging context.

With regard to our specific aims:

  1. We showed a short-term increase in incident reporting through the presence of the International ObGyn and simplification of the reporting process. This increase was not sustained after the International ObGyn left the project.

  2. The single multiprofessional obstetric emergency drills simulation was received well with a small increase in self-reported confidence in participants in the key safety behaviours via survey. However due to the lack of expertise and confidence in conducting simulation training, this was unable to continue as an intervention. It is unlikely that the single episode of drills simulation would affect any meaningful long-term culture and communication improvement, but our work showed that this type of educational intervention is feasible and with investment of expertise and time, could be a valued multiprofessional teaching space for action orientated learning.

  3. The new drug chart co-designed by staff showed an improvement in drug prescription practices.

  4. Fluid restriction in women with severe pre-eclampsia was more reliably carried out following our intervention, but we do not have long-term audit data to confirm sustainment. Subjective reporting from staff 6 months on suggests sustainment of this practice.

  5. Blood pressure management was improved over the 3 months of intervention, but we do not have long-term audit data to confirm sustainment. Subjective reporting from staff 6 months on suggests sustainment of this practice.

  6. Tranexamic acid usage was increased over the intervention period, indicating that PPH management was probably improved following better identification. We would have needed more data points regarding tranexamic acid consumption in following months to see if the improvement sustained. PPH identification improved over the intervention period, although more data points would be needed to confirm sustainment of improvement.

Our quantitative and qualitative data have shown that broadly speaking, improvement work in this context is feasible and beneficial. In a humanitarian context, the lack of formal risk management and governance structures, as well as an often traumatised and burnout workforce, can mean that improvement work can seem overwhelming. The benefits of using a structured improvement methodology is that the scope of the work can be focused into specific aims, with specific measures that can provide assurance that improvement is happening.

Although there are language and cultural differences, the general principles of engagement, encouraging ownership of the improvement project by the frontline staff and benefits of a coaching leadership style were applicable in this low-income, high-insecurity context. Clear messaging, supportive leadership and staff development would be essential strategies in advancing on this work moving forwards. Expertise is lacking locally in improvement methodology, debrief and facilitation skills and risk management skills including root cause analysis following clinical incidents. Conversely, clinical expertise is reliably present among the national Yemeni ObGyn staff, but they are mostly not given remit for improvement, nor feel empowered. Of course, being in a country currently engaged in an active civil war (with no end in sight), there is a resource implication and instability in the workforce. However, in long-term humanitarian projects such as this one in Taiz Houban MCH (which has been going on for 7 years, since 2015), a robust strategy for the long term in providing resource for improvement expertise, staff support following critical incidents and analysis of these incidents to make robust improvement recommendations is needed.

In conclusion, the lesser heard contexts such as this one, operating in conflict zones around the world, should benefit from attempts to apply basic principles of safety improvement in healthcare. This work is feasible, necessary and as important as in any other healthcare delivery system.

Data availability statement

All data relevant to the study are included in the article or uploaded as supplementary information.

Ethics statements

Patient consent for publication


We would like to acknowledge the assistance in data presentation and in the use of statistical process control charts provided by Ms Amy Cruickshank, and expert input in the local review of maternal death by Drs Ana Montoya and Arvind Subramaniam. We would also like to acknowledge all the doctors, midwives, nurses and other staff members working in the Taiz Houban Mother and Child Hospital, whose engagement with this improvement project is testament to their commitment to serving Yemeni women in their time of greatest vulnerability and need—during pregnancy and birth.


Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.


  • Twitter @drsabrinadas

  • Contributors SD is responsible for overall content as guarantor. SD and EA contributed to the conception and design of the work. SD, AM and AKDJ acquired and interpreted the data. SD analysed the data. SD drafted the work. AM, AKDJ and EA revised the work critically for important intellectual content. All four authors approved the final version to be published and agreed to be accountable for all aspects of the work.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.