Can a Quality Improvement Intervention Improve Person-Centred Maternity Care in Kenya?

"While improving the quality of women's experiences during childbirth is a critical component to ensure comprehensive, high-quality maternity care experiences and outcomes, further research is required to understand which intervention methods may be most appropriate...in resource-constrained settings."

Preventable maternal mortality is a major concern, especially in sub-Saharan Africa. This study aimed to understand whether quality improvement collaboratives (QICs) could improve person-centred maternity care (PCMC) experiences for women delivering in three government-run health facilities in Kiambu and Nairobi counties, Kenya. PCMC places the woman and her family at the centre of care, involving her in decisions and respecting and responding to her needs, values, and preferences.

Each of the three intervention facilities was asked to establish a QI team consisting of nurses and midwives, as well as non-clinical support staff (e.g., data clerks). External QI experts engaged facility leadership at the outset and worked to generate buy-in to support the project activities. QI teams used the Model for Improvement (MFI) to set aims for improvement, establish common measures to capture improvement data, identify change strategies that appeared likely to improve PCMC behaviours, and test these using Plan-Do-Study-Act (PDSA) cycles within their respective facilities. Two out of the three facilities focused on two PCMC sub-domains: (i) dignity and respect and (2) communication and autonomy, while the third facility also worked on the PCMC sub-domain of (iii) supportive care. QI teams conducted exit interviews of patients continuously throughout the study to assess whether change strategies led to improvement in PCMC experience.

A pre–post design was used to examine changes in PCMC scores across three intervention and matched control facilities at baseline (n = 491) and endline (n = 677). Difference-in-difference analyses using models that included main effects of both treatment group and survey round were conducted to understand the impact of the intervention on PCMC scores.

Findings suggest that intervention facilities' average total PCMC score decreased by 5.3 points post-intervention compared to baseline (95% confidence interval (CI): -8.8, -1.9) and relative to control facilities, holding socio-demographic and facility variables constant. In addition, the intervention was significantly associated with a 1.8-point decrease in clinical quality index pre-post intervention (95% CI: -2.9, -0.7), decreased odds of provider visits, and less likelihood to plan to use postpartum family planning.

All intervention facilities found these negative results surprising because of the general sense that the QI approach was helpful; however, they (along with the researchers) also offered a number of potential explanations as to why the intervention may not have been successful.

• The most common explanation for the lack of intervention impact was related to health system constraints, such as lack of resources from the government (e.g., inconsistent supply of essential medicines) and insufficient staff-to-patient ratios. For example, behaviours that could lead to PCMC improvements within the subdomain of dignity and respect, such as staff introductions or learning a patient's name, may not have been viewed as a priority, despite the ease with which these approaches may be implemented.
• A few staff believed that providers may have found it challenging to embrace the new concept of PCMC as critical to their responsibilities and that the time perceived to be taken to implement these activities could impede their ability to carry out clinical duties successfully and efficiently. The focus on patient experience only as opposed to patient experience combined with clinical trainings may have contributed to these perceptions. Future studies may want to coordinate QICs that include both a provider clinical training and person-centred care model. (Because women reported higher levels of person-centred care with midwives as opposed to doctors, specific trainings aimed towards doctors to provide respectful care may be needed.)
• Almost all staff discussed how the requirement to obtain weekly feedback from patients on aspects of their experiences through exit interviews was time-intensive - both for the QI team and for women who had just delivered. These circumstances may have contributed to burnout and declining participation by QI team members.
• There were two national strikes of doctors and nurses that interrupted the intervention's momentum and potentially reduced enthusiasm among remaining active QI team members, as well as among facility and county leadership.
• Across both control and intervention facilities, staff reported broader QI interventions, including effective communication and respectful maternity care, occurring over the duration of the project from a variety of stakeholders, including government. It is notable that government-led QI initiatives had strong buy-in by the facilities.

In conclusion: "While QICs are often cited as viable approaches to improve health outcomes and processes in resource-constrained settings because they focus on working within the means and context of a health system, these approaches may not be appropriate to improve PCMC behaviours among clinical staff in all contexts."