Continuous quality improvement, linkage to peer supporters in prevention of mother-to-child human immunodeficiency virus transmission programs in Rwanda

Abstract

BACKGROUND

Peer supporters are instrumental for the successful navigation of prevention of mother-to-child human immunodeficiency virus (HIV) transmission services, including maternal viral load suppression and early infant HIV diagnosis. These outcomes are critical for the elimination of mother-to-child HIV transmission. We assessed the impact of continuous quality improvement (CQI) services on assignment to peer supporters and undetection of the viruses among pregnant women living with HIV (WLHIV) and mother-to-child transmission of HIV (MTCT) rates among HIV exposed babies in Rwanda.

AIM

To assess the impact of CQI services on assignment to peer supporters and undetection of the viruses among pregnant WLHIV and MTCT rates among HIV exposed babies in Rwanda.

METHODS

Between 2021 and 2022, CQI services were implemented in 18 of the 38 healthcare facilities included in this analysis. Healthcare workers in 18 facilities used CQI approaches to key predictors of, and contributors related to, the implementation of the prevention of MTCT (PMTCT) program to improve maternal and infant outcomes. This was a secondary data analysis that explored the association between CQI and assignment to peer supporters, undetection of the virus, and MTCT rates. To assess the impact of CQI on these outcomes, a multivariable logistic regression model was used to compute adjusted odds ratios (aOR) and corresponding 95% confidence intervals (CI).

RESULTS

A total of 1145 mother-baby pairs were included, of whom 558 (48.7%) were from facilities that implemented CQI. At the end of evaluation, 1043 (91.0%) either completed 24-months of follow-up or remained in care in the same facilities. Of 102 not available at the evaluation, 84 were transferred out, and 18 were dead or lost to follow-up. Overall, 405 (35.4%) women were assigned to peer supporters, 1004 (87.7%) had undetectable viruses, and 8 (0.7%) infants were infected with HIV. Compared to women from non-CQI facilities, those from CQI implementing facilities had 32% higher odds of being assigned to peer supporters (aOR = 1.32; 95%CI: 1.02-1.71). Similarly, disclosing HIV status (aOR = 1.53; 95%CI: 1.10-2.14) and a higher number of health care workers per 1000 active patients (> 4 vs ≤ 4) (aOR = 1.34; 95%CI: 1.05-1.75) were associated with higher odds of being linked to peer supporters. Having a formal education was associated with reduced odds of being linked to peer supporters compared to those with no education. Compared to women from non-CQI facilities, those from CQI implementing facilities had 51% higher odds of having undetectable viruses (aOR = 1.51; 95%CI: 1.05-2.18). CQI was associated with fewer transferred out compared to non-CQI facilities (6.0% vs 8.8%; P = 0.07).

CONCLUSION

Nearly 3 in 10 pregnant WLHIV in Rwanda were linked to peer support to support the successful navigation of PMTCT services, and CQI increased this linkage and aided the achievement of undetectable viruses. Implementing CQI and promoting HIV status disclosure is critical to facilitate peer support linkage and improve maternal and infant outcomes.

Key Words: Continuous quality improvement; Peer support, mother to child transmission of human immunodeficiency virus; Pregnant women living with human immunodeficiency virus; Human immunodeficiency virus; Rwanda

Core Tip: Peer supporters are instrumental for the successful navigation of prevention of mother-to-child human immunodeficiency virus (HIV) transmission services, including maternal viral load suppression and early infant HIV diagnosis. This retrospective cohort study assessed the effect of continuous quality improvement on the assignment of peer supporters, achievement of undetectable viruses among pregnant women living with HIV, and mother-to-child transmission rates among HIV exposed infants in Rwanda. The findings showed that compared to healthcare facilities that did not employ continuous quality improvement approaches, those that did improved assignment of peer supporters and achievement of undetectable viruses.

INTRODUCTION

Mother-to-child transmission of human immunodeficiency virus (MTCT) is the main method of human immunodeficiency virus (HIV) transmission in children. The chance of HIV transmission from the mother to child is as high as 42% if the mother is not treated for HIV[1]. Transmission of HIV from mother to child is reduced to less than 1% if the mother receives HIV treatment[2]. Although MTCT of HIV is low in high-income countries, it is still high in low- and mid-income countries. While between 2012 and 2014 the rates of MTCT in the United Kingdom was 0.46% and 0.27%[3], up until 2023, reported rates of MTCT was 20% in Western Africa, and 6% in Eastern Africa[4]. These data indicate regional disparities, and additional efforts need to be implemented to eliminate MTCT. Management of HIV in children is challenging due to several factors, including but not limited to caretaker capacities, drug formulations, drug adherence, nutritional needs, neurodevelopmental issues, stigma, and forgetfulness[5-7]. Due to these complexities of caring for a child living with HIV, prevention of MTCT (PMTCT) remains a cornerstone.

Rwanda is one of only seven African countries to have achieved the UNAIDS 95-95-95 targets. By 2025, of the estimated 230000 people living with HIV in the country, 96% knew their status; of these, 98% were on treatment, and of those on treatment, 98% were virally suppressed[8]. Furthermore, nearly 99% of pregnant women living with HIV (WLHIV) were receiving antiretroviral therapy (ART), in the mission of eliminating MTCT. These achievements are commendable, and through strong leadership and commitment, HIV epidemic control will be sustainable. Supporting PMTCT efforts is critical to maintain sustainability of the elimination of HIV in pediatric populations.

Peer supporters, particularly mentor mothers (HIV-positive women who have successfully completed the PMTCT cascade and are trained to support less-experienced pregnant mothers living with HIV), are instrumental for the successful navigation of PMTCT services, including maternal viral load suppression and early infant HIV diagnosis. These outcomes are critical for the elimination of MTCT. For example, data from Nigeria showed that, compared to women who were not linked to a structured mentor mother support, those who were linked had higher odds of being retained in care and being virally suppressed[9]. Impact of mentor mother support on viral load suppression and retention has been shown elsewhere[10,11]. Structured mentor mother support was also critical in facilitating timely uptake and timely early infant HIV diagnosis[12]. Therefore, any intervention that has the potential to link pregnant women to peer supporters for successful navigation of PMTCT services should be encouraged.

Recognizing the importance of peer support, this study assessed the impact of continuous quality improvement (CQI) services on assignment to peer supporters among pregnant WLHIV in Rwanda. In addition, based on the theme of undetectable virus equals untransmissible, the study also explored the impact of CQI on the achievement of undetectable viruses among these women and MTCT rates among HIV exposed infants (HEI). CQI refers to a systematic ongoing process that constantly uses data to improve outcomes. CQI identifies problems and implements solutions to achieve the desired outcome[13]. In this study, we hypothesize that CQI approaches improves assignment of peer supporters and aid the achievement of undetectable viruses among pregnant WLHIV and reduce MTCT among HEI in Rwanda.

MATERIALS AND METHODS

Study design, setting, and population

A retrospective study was done by using data collected from mother-baby pairs from 38 primary healthcare facilities in Rwanda as previously described[14]. In brief, the cohort consisted of adults 18 years and older, pregnant WLHIV. A total of 1145 mother-baby pairs were included, of whom 558 (48.7%) were from healthcare facilities that implemented CQI services to improve patient outcomes in PMTCT, mother-infant retention, and early infant diagnosis.

Inclusion and exclusion criteria

All pregnant WLHIV who were 18 years of age and older and their HEI from 38 healthcare facilities were eligible for the study. All 1145 women had information about peer support linkage and were included in this analysis.

Sample size estimation

This was a secondary data analysis using data that was primarily intended to assess the association between sustained maternal viral load suppression and cascade of infant HIV testing and mother-baby pair retention in Rwanda as previously described[14]. All participants from this database were evaluated, and those who met the inclusion criteria for the current analysis were included without a sample size calculation.

Definition of variables

The main exposure variable was utilization of CQI approaches, dichotomized as “Yes” and ”No” for facilities that used and those that did not use CQI approaches, respectively. The main outcome of interest was assignment/Linkage to peer supporters, also dichotomized as linked vs not linked for pregnant women who were linked to peer supporters and those who were not linked. Additional variables explored in this analysis included age (18-24 years, 25-34 years, 35-50 years), education level (none, primary, secondary, tertiary), marital status (single, married, cohabiting, widowed/divorced), timing of ART initiation (before pregnancy, during pregnancy), ratio of the number of healthcare workers per 1000 patients who were active on ART. This ratio was dichotomized at the median number as (> 4 vs ≤ 4), status of participants (completed 24-months of follow-up, active in care, dead, lost to follow-up, or transferred out), and disclosure of HIV status (disclosed, did not disclose). HIV disclosure was defined as sharing one’s HIV diagnosis with a partner, peer educator, friend, or any family member. The viral load results at the delivery time were used to identify proportions of pregnant WLHIV who had undetectable viruses. Those with viral load ≤ 20 copies/mL were considered to have undetectable virus, and those with viral load > 20 copies/mL were considered to have detectable virus. HIV status of HEI was ascertained up to 24 months of follow-up.

CQI intervention

Eighteen (18) healthcare facilities were selected to participate in a CQI intervention to improve PMTCT services. CQI facilities were selected to match non-CQI facilities based on the number of people living with HIV being served. The selection included medium volume facilities (200-349 people living with HIV per year) to high volume facilities (≥ 350 people living with HIV per year). Each facility nominated two to three staff members, including PMTCT focal persons, monitoring and evaluation officers, nurses, and medical officers, who completed a three-day standardized training based on the model for improvement and the plan-do-study-act cycle. Training focused on aim setting, PMTCT process mapping, use of routine data to identify gaps, root cause analysis, selection of change ideas, and rapid-cycle testing. The predefined set of PMTCT indicators targeted for improvement included: (1) Proportion of pregnant WLHIV linked to peer supporters; (2) Undetectable maternal viral load (≤ 20 copies/mL) at delivery; (3) Timely early infant diagnosis (< 2 months of age); (4) Mother-baby pair retention in care at 24 months postpartum; and (5) Prevention of mother-to-child transmission rate. Facility-level multidisciplinary teams, comprising clinical staff (nurses, medical officers), data managers, counselors, and peer supporters, met monthly to analyze facility-specific data, identify performance gaps, and test change ideas. Specific change ideas tested included: Streamlining peer support referral processes to reduce time from enrollment to linkage; implementing standardized PMTCT checklists at key clinical touchpoints; establishing dedicated peer supporter schedules at facilities; enhancing data quality systems for real-time tracking of outcomes; and strengthening integration of HIV disclosure counseling into routine care. Teams used rapid-cycle testing (plan-do-study-act cycles) to pilot these changes, assess impact, and refine approaches based on monthly performance data. Facilities received monthly mentorship and supportive supervision for three months from the University of Maryland, Baltimore, project staff, and were subsequently linked into a learning network to facilitate peer learning and improvement over 36 months. Mentorship intensity decreased over time, transitioning from biweekly on-site support to monthly district-level meetings during the spread phase. Quarterly learning sessions brought together all 18 facility teams to share successes, challenges, and lessons learned. Between quarterly sessions, multidisciplinary teams applied CQI methods at the facility level, supported by ongoing on-site mentoring from the University of Maryland, Baltimore.

Statistical analysis

This was a secondary data analysis that investigated the association between the use of CQI approaches and linkage to peer supporters, achievement of undetectable viruses among pregnant WLHIV, and MTCT rates among HEI in Rwanda. Categorical participants' characteristics were stratified by status of linkage to peer supporters, and comparison made by the χ² tests. Continuous variables were presented by median and interquartile range. To determine the association between the use of CQI approaches and linkage of peer supporters and achievement of undetectable viruses, binary logistic regression models were used. First, bivariate analysis was conducted. Then all variables assessed by bivariate analysis were entered into a multivariable regression model to adjust for the confounders. To account for clustering of study participants within healthcare facilities, random effects models were used in both bivariate and multivariable analyses. Two variables had missing data ranging from 1.1% to 5.5%. Due to the minimal amount of missing data from variables included for this analysis, a complete case analysis was conducted. Statistical analyses were conducted using SAS version 9.4 (SAS Institute Inc., Cary, NC, United States).

Ethical approval

The original study was approved by the Rwanda National Ethics Committee with approval certificate number No. 104/RNEC/2022.

RESULTS

Overall

A total of 1145 WLHIV were included, with the median age (Interquartile range) of 35 (30-40) years. Nearly half of these women, 558 (48.7%), were from facilities that implemented CQI approaches. About two-thirds, 744 (65.0%), had a primary school education, and 807 (70.5%) were either married or lived together. Eight hundred and eighty-eight disclosed HIV status to others, and 758 (66.2%) initiated ART during pregnancy. Regarding healthcare facility factors, about 537 (46.9%) women came from health care facilities with ≥ 4 healthcare workers per 1000 active patients (Table 1). A total of 1043 (91.0%) women either completed 24-months of follow-up or remained in care in the same healthcare facilities. Of 102 not available at the evaluation, 84 were transferred out, and 18 were dead or lost to follow-up. CQI implementing facilities were associated with fewer transferred out compared to non-CQI facilities (6.0% vs 8.8%; P = 0.07).

Table 1 Individual and facility characteristics, n (%).

Characteristics	Overall (n = 1145)	Assigned to peer support
		Yes (n = 405)	No (n = 740)	P value
CQI approaches implemented
No	587 (51.3)	189 (46.7)	398 (53.8)	0.021
Yes	558 (48.7)	216 (53.3)	342 (46.2)
Age
13-24	116 (10.1)	42 (10.4)	74 (10.0)	0.944
25-34	380 (33.2)	132 (32.6)	248 (33.5)
35-50	649 (56.7)	231 (57.0)	418 (56.5)
Education level
None	165 (14.4)	84 (20.7)	81 (11.0)	< 0.001
Primary	744 (65.0)	259 (64.0)	485 (65.5)
Secondary	222 (19.4)	59 (14.6)	163 (22.0)
Tertiary	14 (1.2)	3 (0.7)	11 (1.5)
Marital status
Single	233 (20.4)	72 (17.8)	161 (22.1)	0.225
Married or living together	807 (70.5)	299 (74.0)	508 (69.8)
Divorced/widowed	92 (8.0)	33 (8.2)	59 (8.1)
Missing	13 (1.1)	1	12
HIV status disclosed
No	257 (22.5)	69 (17.0)	188 (25.4)	0.001
Yes	888 (77.5)	336 (83.0)	552 (74.6)
Pregnant at the start of ART
No	324 (28.3)	91 (23.0)	233 (33.9)	< 0.001
Yes	758 (66.2)	304 (77.0)	454 (66.1)
Missing	63 (5.5)	10	53
Healthcare workers/1000 active patients
≤ 4	608 (53.1)	188 (46.4)	420 (56.8)	0.001
> 4	537 (46.9)	217 (53.6)	320 (43.2)

CQI: Continuous quality improvement; ART: Antiretroviral therapy; HIV: Human immunodeficiency virus.

Assignment to peer supporters

A total of 405 (35.4%) were assigned to peer educators. Healthcare facilities that implemented CQI assigned a higher percentage of women to peer educators compared to those that did not implement CQI (38.7% vs 32.2%; P = 0.021).

Factors associated with linkage/assignment to peer supporters

Bivariate analysis: Several individual and facility characteristics were associated with assignment to peer supporters. Compared to facilities that did not implement CQI approaches, those that did had higher odds of linking pregnant women to peer supporters [odds ratio (OR) = 1.33; 95% confidence interval (CI): 1.04-1.70] (Table 2). Individuals with a primary school education or higher had lower odds of being linked to peer supporters compared to those without formal education. Disclosing HIV status to others (OR = 1.66; 95%CI: 1.22-2.26) and ART initiation (during pregnancy vs before pregnancy) (OR = 1.90; 95%CI: 1.45-2.48) were all associated with increased odds of being linked to peer supporters. A higher number of health care providers per active patients was also associated with peer supporters’ linkage. For example, healthcare facilities that had > 4 healthcare workers per 1000 active patients had 52% higher odds of linking pregnant women to peer supporters compared to healthcare facilities that had ≤ 4 healthcare workers per 1000 active patients (OR = 1.52; 95%CI: 1.19-1.93).

Table 2 Factors associated with linkage to peer supporters among pregnant women living with human immunodeficiency virus in Rwanda.

	Bivariate		Multivariable
	OR (95%CI)	P value	aOR (95%CI)	P value
CQI approaches implemented
No	Reference		Reference
Yes	1.33 (1.04-1.70)	0.021	1.32 (1.02-1.71)	0.033
Age
13-24	Reference		Reference
25-34	0.94 (0.61-1.45)	0.772	1.21 (0.76-1.92)	0.422
35-50	0.97 (0.64-1.47)	0.898	1.10 (0.71-1.70)	0.676
Education level
None	Reference		Reference
Primary	0.51 (0.37-0.72)	< 0.001	0.52 (0.37-0.74)	< 0.001
Secondary	0.35 (0.23-0.53)	< 0.001	0.36 (0.23-0.56)	< 0.001
Tertiary	0.26 (0.07-0.98)	0.046	0.25 (0.06-0.97	0.044
Marital status
Single	Reference		Reference
Married or living together	1.32 (0.96-1.80)	0.085	1.21 (0.87-1.70)	0.258
Divorced/widowed	1.25 (0.75-2.08)	0.388	1.11 (0.65-1.90)	0.703
HIV status disclosed
No	Reference		Reference
Yes	1.66 (1.22-2.26)	0.001	1.53 (1.10-2.14)	0.011
ART initiation
Before pregnancy	Reference		Reference
During pregnancy	1.90 (1.45-2.48)	< 0.001	1.62 (1.22-2.16)	0.001
Healthcare workers/1000 active patients
≤ 4	Reference		Reference
> 4	1.52 (1.19-1.93)	< 0.001	1.34 (1.05-1.75)	0.019

CQI: Continuous quality improvement; ART: Antiretroviral therapy; HIV: Human immunodeficiency virus; CI: Confidence intervals; OR: Odds ratio; aOR: Adjusted odds ratio.

Multivariable analysis: After adjusting for covariates, the association between CQI implementation and peer support linkage persisted, with facilities that implemented CQI approaches having 32% increased odds of peer supporter linkage compared to facilities that did not implement CQI approaches [adjusted odds ratios (aOR) = 1.32; 95%CI: 1.02-1.71]. While individuals with primary school education or higher continued to be associated with reduced odds of being linked to peer supporters compared to those without formal education, those who disclosed HIV status to others and those who initiated ART during pregnancy continued to be associated with higher odds of being linked to peer supporters. Furthermore, the number of healthcare workers per 1000 active patients continued to be associated with peer supporters’ linkage with those facilities having > 4 healthcare workers, having higher odds of linking pregnant women to peers compared to those with ≤ 4 healthcare workers (aOR = 1.34; 95%CI: 1.05-1.75).

Viral detection and MTCT rates

One thousand and four (87.7%) women had undetectable virus at the time of delivery. Women from CQI implementing facilities had a higher percentage of undetectable viruses compared to women from non-CQI implementing facilities (89.8% vs 85.%; P = 0.027). A total of 8 (0.7%) HEI had incident HIV infection, and the rates of infections were not significantly different between CQI and non-CQI implementing facilities (1.1% vs 0.3%; P = 0.168), respectively.

Factors associated with un-detection of the virus

Bivariate analysis: Only CQI and timing of ART initiations were significantly associated with achievement of undetectable virus. Compared to women from facilities that did not implement CQI, women from facilities that implemented CQI had higher odds of achieving undetectable viruses (aOR = 1.49; 95%CI: 1.04-2.13) (Table 3). Women who initiated ART during pregnancy had lower odds of achieving undetectable viruses compared to those who were ART-experienced prior to becoming pregnant (aOR = 0.67; 95%CI: 0.47-0.97).

Table 3 Factors associated with undetection of the virus among pregnant women living with human immunodeficiency virus in Rwanda.

	Bivariate		Multivariable
	OR (95%CI)	P value	aOR (95%CI)	P value
CQI approaches implemented
No	Reference		Reference
Yes	1.49 (1.04-2.13)	0.028	1.51 (1.05-2.18)	0.028
Age
13-24	Reference		Reference
25-34	1.22 (0.67-2.21)	0.521	1.24 (0.67-2.31)	0.494
35-50	1.28 (0.72-2.25)	0.400	1.18 (0.66-2.11)	0.576
Education level
None	Reference		Reference
Primary	1.02 (0.60-1.74)	0.934	1.04 (0.61-1.78)	0.883
Secondary	0.63 (0.35-1.14)	0.126	0.62 (0.34-1.14)	0.123
Marital status
Single	Reference		Reference
Married or living together	1.51 (0.99-2.30)	0.053	1.48 (0.96-2.30)	0.079
Divorced/widowed	0.75 (0.40-1.40)	0.370	0.74 (0.39-1.41)	0.361
HIV status disclosed
No	Reference		Reference
Yes	1.11 (0.74-1.68)	0.612	0.88 (0.56-1.37)	0.576
ART initiation
Before pregnancy	Reference		Reference
During pregnancy	0.67 (0.47-0.97)	0.031	0.62 (0.43-0.91)	0.014
Healthcare workers/1000 active patients
≤ 4	Reference		Reference
> 4	0.91 (0.64-1.30)	0.605	0.83 (0.58-1.19)	0.312

CQI: Continuous quality improvement; ART: Antiretroviral therapy; HIV: Human immunodeficiency virus; CI: Confidence intervals; OR: Odds ratio; aOR: Adjusted odds ratio.

Multivariable analysis: After adjusting for covariates, the association between CQI and timing of ART initiation on achievement of undetectable viruses continued. Women from facilities that implemented CQI approaches had 51% higher odds of achieving undetectable viruses compared to women from facilities that did not implement CQI (aOR = 1.51; 95%CI: 1.05-2.18). Women who initiated ART during pregnancy had 38% decreased odds of achieving undetectable viruses compared to women who experienced ART prior to becoming pregnant (aOR = 0.62; 95%CI: 0.43-0.91).

Intra-cluster correlation coefficient: Model results showed an intra-cluster correlation coefficient of 0.39, indicating that 39% percent of the variability in the probability of assigning women to peer supporters was accounted for by healthcare facilities, leaving behind 61% percent of the variability in the probability of assigning women to peer supporters to be accounted for by individual, CQI intervention, or other unknown factors.

DISCUSSION

This was a retrospective study that involved 1145 pregnant WLHIV and their HEI in Rwanda. Slightly more than one-third of these women were linked to peer supporters to assist with the successful navigation of PMTCT services during pregnancy and post-partum periods. We noted that facilities that implemented CQI approaches had higher odds of linking these women to peer supporters compared to facilities that did not implement CQI approaches. In addition, disclosure of HIV status to others, initiation of ART during pregnancy, compared to before pregnancy were associated with higher odds of being linked to peer supporters. Individuals who had a formal education had reduced odds of being linked to peer supporters compared to those who had no formal education. Furthermore, besides individual characteristics, facility characteristics also played a role in peer support linkage, with facilities having a higher number of healthcare workers per 1000 active patients having higher odds of linking women to peers compared to facilities with a lower number of healthcare workers. Although the percentage of women who had undetectable virus was higher in facilities that implemented CQI compared to facilities that did not implement CQI, the proportions of MTCT rates among HEI were similar between facilities that did and those that did not implement CQI.

Our results provide critical insights into the interventions that facilitate PMTCT services. We have demonstrated that CQI approaches improved linkage of pregnant women to peer supporters to enhance navigation of PMTC services in Rwanda. As previously stated, peer supporters potentiated viral load suppression, retention of mother-baby pairs, and early infant diagnosis[9,12,15,16]. Therefore, any intervention that facilitates linkage to peer support in PMTCT efforts should be advocated. A previous study in Kenya showed that CQI approaches improved the integration of nutrition assessment, counseling, and support in PMTCT from the median of 15% to 88% and retention of mother-baby pairs from the median of 19% to 66%[17]. The success of the Kenyan program was facilitated by mentor mothers who supported the healthcare workers to provide services, including adherence counseling, routine home visits, tracing of mother-baby pairs, and establishment of support groups, which all safeguarded active participation of newly enrolled mother-baby pairs. These data indicate the capacity of peer support in the successful navigation of PMTCT programs to improve maternal and infant outcomes.

A study in South Africa and Kenya showed that CQI approaches improved maternal uptake of ART, leading to the reduction of the proportion of HEI testing positive, suggesting the importance of deploying CQI approaches in PMTCT programs[18,19]. In an attempt to understand the mechanism by which CQI approaches improved linkage of peer supporters to pregnant women, we triangulated the data further. This was done by exploring the distribution of pregnant women who disclosed HIV status, had informal education, initiated ART during pregnancy, and the number of healthcare workers between facilities that implemented CQI and those that did not. We noted that the distribution of these women who disclosed HIV status, had informal education, and the number of healthcare workers between the two groups was similar; however, there were more women who initiated ART during pregnancy in the CQI implementing facilities than in non-CQI implementing facilities (72% vs 60%). If it was expected that women who initiated treatment during pregnancy needed more peer support assistance than ART-experienced ones, it is likely that the effect of CQI on peer support linkage would be partly explained by this differential distribution. Following the uneven distribution of women who initiated ART during pregnancy, we further assessed if the association between CQI and peer support was modified by this uneven distribution. Results showed that the association between CQI and peer support persisted in both women who initiated ART before and during pregnancy, although statistical significance was lost because of the reduced sample size in this stratified analysis. This suggests that CQI intervention approaches are critical in improving peer support linkage irrespective of the factors evaluated in this study.

Several other factors were associated with peer support linkage. For example, women who had a formal education had lower odds of being linked to peer supporters. This observation could be because individuals with formal education are more knowledgeable of HIV and were more likely to take control of their personal health, and thus need less assistance for PMTCT navigation compared to those who had no formal education. It is likely that those with formal education need more privacy or programs targeted at linking peers to those with informal education. Although stigma against HIV could be a possible explanation, a systematic review of 40 studies showed that having higher education is associated with a lower likelihood of experiencing overall stigma[20]; therefore, we cannot conclude that lower peer support linkage among educated women was due to stigma. Disclosure of HIV status to others involved individual commitment and empowerment with minimal or no stigma[21,22]. These attributes suggest readiness to receive assistance in any possible ways in the care of HIV and may explain increased odds of being linked to peer support compared to those who did not disclose their HIV status. We also observed higher odds of peer support linkage among women who initiated ART during pregnancy compared to their counterparts who were ART-experienced prior to becoming pregnant. It is likely that those who were ART naïve prior to pregnancy require more supportive interventions than ART experienced to ensure higher levels of adherence and hence better maternal and infant outcomes. It is expected that the majority of ART-experienced individuals have passed the stages of managing the psychological burden of a lifelong condition and navigating various personal and social challenges related to treatment adherence and engagement in care.

We observed a higher percentage of women with undetectable viruses from facilities that implemented CQI approaches than from facilities that did not implement CQI approaches. Although these results were promising, proportions of MTCT between the two groups were similar. These observations could be due to several reasons. First, the number of HIV infection to HIV exposed babies in this study was too small to explore statistical significance. Second, we noted that the timing of ART initiation matters. Women who who were ART experienced had higher odds of achieving undetectable viruses than those who initiated treatment during pregnancy. Data triangulation showed that a higher proportion of women in facilities that implemented CQI initiated treatment during pregnancy compared to women from facilities that did not implement CQI (72.9% vs 58.9%). It is likely that MTCT of HIV has already occurred prior to ART initiation.

A few limitations are worth mentioning in this study. We assessed the effect of CQI on assignment to peer supporters and achievement of undetectable viruses among pregnant WLHIV and MTCT among HEI using a secondary data analysis that was not primarily intended to address this question. The use of secondary data analysis may underestimate the true effect size of the observed association. The non-randomized nature of the study design may be impacted by systematic differences such as mentorship intensity, infrastructure, and leadership engagement between facilities. These differences may limit the interpretation of the observed associations. Data analysis revealed an uneven distribution of women who disclosed HIV status between facilities that did and those that did not implement CQI approaches, and disclosure tended to influence peer support linkage. This observation may partly explain the association between CQI and peer support linkage. Despite this observation, data triangulation showed the association between CQI and peer support linkage persisted irrespective of HIV disclosure status. Furthermore, we also noted uneven distribution of the timing of ART initiation between facilities that did and those that did not implement CQI approaches, and the timing of ART initiation was associated with achievement of undetectable viruses. The use of routinely collected clinical data from 38 healthcare facilities led to a larger sample size, making this study stronger and generalizable within the Rwandan context.

CONCLUSION

Nearly 3 in 10 pregnant WLHIV in Rwanda were linked to peer support to support the successful navigation of PMTCT services. The study also demonstrated the impact of CQI approaches on the linkage of peer supporters and the achievement of undetectable viruses among these women. Moving forward, the strategic implementation of CQI approaches, improvement of health education, and handling of stigma to promote HIV status disclosure are critical to facilitate peer support linkage and improve maternal and infant outcomes in PMTCT programs.

ACKNOWLEDGEMENTS

The authors express sincere gratitude to the pregnant WLHIV who participated in this program and whose data made this analysis possible. Support for the implementation of PMTCT and quality improvement activities was provided by UNICEF and the Ministry of Health through the Rwanda Biomedical Center. The authors thank the healthcare workers, facility CQI teams, and district supervisors for their dedication to improving maternal and infant outcomes, as well as the data management teams for their support in maintaining high-quality routine clinical data. The views expressed in this article are those of the authors and do not necessarily represent the official policies or positions of UNICEF, the Rwanda Biomedical Center, or the Ministry of Health.