Predictors of early ART initiation among HIV+ infants in Kenya: A retrospective review of HITSystem data from 2013-2017.

Abstract

Objective:

Early initiation of antiretroviral therapy (ART, before 12 weeks of age) among infants living with HIV reduces infant mortality and slows disease progression. However, inefficiencies in early infant diagnosis processes prevents timely ART initiation among infants living with HIV in Kenya. This study assesses predictors of early ART initiation among infants living with HIV in Kenya.

Design:

We retrospectively reviewed data from 96 infants living with HIV born between January 2013 and June 2017 at 6 Kenyan government hospitals.

Methods:

The primary outcome was infant receipt of ART by 12 weeks of age. We assessed bivariable and multivariable predictors of ART initiation by 12 weeks of age.

Results:

Among 96 infants living with HIV, 82 (85.4%) infants initiated ART at a median infant age of 17.1 weeks. Of the 82 infants who started ART, only 17 (20.7%) initiated ART by 12 weeks of age. In multivariable logistic regression analyses, testing per national guidelines (< 7 weeks of age) (aOR 40.14 [3.96-406.97], p=0.002), shorter turnaround time for result notification (≤ 4 weeks) (aOR 11.30 [2.02-63.34], p=0.006), and ART initiation within 3 days of mother notification (aOR 7.32 [1.41-38.03], p=0.006) were significantly associated with ART initiation by 12 weeks of age.

Conclusion:

Current implementation of early infant diagnosis services in Kenyan only achieves targets for early ART initiation in one-fifth of infants with HIV. Strengthening services to support earlier infant testing and streamlined processes for early infant diagnosis may increase the proportion of infants who receive timely ART.

Introduction

Approximately 50% of the world’s 160,000 new pediatric HIV infections each year occur in sub-Saharan Africa (UNAIDS, 2019). Most of these children are infected through mother to child transmission (MTCT). In Kenya, MTCT of HIV accounted for approximately 8,000 new infant and child infections in 2018 (Ministry of Health, 2018; UNAIDS, 2018). In the absence of antiretroviral treatment (ART), 20% of infants with HIV will die within the first year of life (Newell et al., 2004). However, ART can significantly improve clinical, immunological, and virologic outcomes and survival in pediatric patients (Cotton et al., 2013; Goetghebuer et al., 2009; Innes et al., 2014; Persaud et al., 2013; Wamalwa et al., 2012), with the landmark CHER study indicating that early initiation of ART (by 12 weeks of age) reduces infant mortality by up to 76% (Violari et al., 2008). Based on these studies, the World Health Organization (WHO) recommended immediate initiation of ART for all infants with HIV, regardless of immunological and clinical stage (World Health Organization, 2010). In 2012, Kenyan adopted these guidelines (Ministry of Health, 2012) and in 2014 issued a Rapid Advice statement to encourage faster uptake of this practice (Ministry of Health & National AIDS and STI Control Program, 2014).

Early infant diagnosis (EID) services are critical for identifying infants living with HIV and linking them to care. In Kenya, EID guidelines recommend the first infant test at 4-6 weeks of age through HIV DNA PCR testing, with periodic retesting through 18 months of age. However, late infant testing (Goggin et al., 2016; Hassan et al., 2012; Nyandiko et al., 2010), long turnaround times for infant test results (Wexler et al., 2017), and delays in caregiver notification of test result (Sarah Finocchario-Kessler et al., 2018) hinder timely infant ART initiation, even with guidelines recommending immediate ART initiation. National Kenyan estimates suggest that only 61%-84% of eligible children (0-14 years old) received life-saving ART in 2018 (Ministry of Health, 2018; UNAIDS, 2018; World Health Organization, 2017).

Few studies have explicitly evaluated the timing of ART initiation in infants. Since the new guidance recommending immediate ART for infants, only 2 studies have reported on the timing of infant ART initiation in Kenya. One study found that only 10.6% of infants identified as HIV-positive were immediately started on treatment (Okwara & Oyore, 2015) The second study, a cluster randomized control trial evaluating the impact of the HITSystem on EID outcomes, found that the median age at infant ART initiation among infants enrolled at control sites (i.e., infants receiving standard of care EID, rather than infants receiving the HITSystem intervention) was 25.1 weeks of age (Finocchario-Kessler et al., 2018), well beyond the first 2-3 months of life when mortality among infants with HIV peaks (Bourne et al., 2009; Marston et al., 2011). These studies indicate that timely ART initiation remains a challenge in Kenya; however, their generalizability is limited by narrow geographic coverage (Okwara & Oyore, 2015), low rates of EID uptake (Okwara & Oyore, 2015), and small sample size (Finocchario-Kessler et al., 2018; Okwara & Oyore, 2015).

Early initiation of ART for infants with HIV is one of the primary goals of EID services; yet, data on the rates, timing, and factors associated with ART initiation among Kenyan infants is extremely limited. The objective of this study is to assess the proportion of infants with HIV who were initiated on treatment by 12 weeks of age from 2013 to 2017 and describe factors associated with early initiation of ART. The sample is drawn from infants enrolled in the HIV Infant Tracking System (HITSystem), a web-based intervention supporting prospective follow up of mother-infant pairs to optimize EID outcomes. Understanding infant ART initiation in the era of immediate ART is necessary for policy makers and healthcare providers to better tailor engagement efforts to maximize the benefits of EID.

Methods

Study Setting and Population

We conducted a retrospective review of HITSystem data from 6 government hospitals in Kenya to determine the timing of ART initiation and predictors of timely ART initiation (< 12 weeks of age) among infants with HIV. All infants enrolled in this study received the HITSystem intervention as part of their care. The HITSystem is an eHealth intervention that aims to improve the efficiency of EID services by linking hospitals, laboratories, and mothers of infants living with HIV. Algorithm-driven electronic alerts notify providers when time-sensitive interventions (infant testing, maternal result notification, ART initiation for infants with HIV) are delayed so they can follow up with patients and automated SMS notify mothers when they need to return to the hospital for services (for infant re-testing, result notification). Text messages do not relay infant test results to the mother but, rather, asked mothers to return to the hospital. The primary objectives of the HITSystem are to expedite PCR test results, facilitate rapid initiation of ART for infants with HIV, and increase retention of infants through the 18-months. The HITSystem has been described in detail, previously (Finocchario-Kessler et al., 2018; Finocchario-Kessler et al., 2014; Finocchario-Kessler, et al., 2015; Finocchario-Kessler, et al., 2015).

Three of the included hospitals were included as intervention sites in a cluster randomized control trial to evaluate the HITSystem (Finocchario-Kessler et al., 2018; Finocchario-Kessler et al., 2015), clinical registration NCT02072603 (“research hospitals”). However, there were too few infants diagnosed as HIV-positive at the 3 sites to draw conclusions, so we also included data from 3 hospitals that were implementing the HITSystem as part of their standard of care, but were not included in the cluster randomized control trial (“programmatic hospitals”). Programmatic hospitals 1-3 started using the HITSystem to supplement their standard of care EID services in 2012, while research hospitals 4-6 started using the HITSystem in 2014, Table 1. All hospitals were government run. They represented District, Sub-County, and Provincial-level hospitals and were in the western, central, and coastal regions of Kenya.

Table 1.

Description of study hospitals

Hospital ID	Study enrollment start date	Site type	Total number of HIV+ infants
Hospital 1	January, 2013	Programmatic	12
Hospital 2	January, 2013	Programmatic	5
Hospital 3	January, 2013	Programmatic	19
Hospital 4	July, 2014	Research	18
Hospital 5	February, 2014	Research	21
Hospital 6	June, 2014	Research	7

Infants living with HIV were included in analyses if: (1) their mother was 18 years of age or older at the time of her infant’s enrollment in EID, (2) the infant had an HIV DNA PCR positive result at the “6-week” testing occasion, (3) the infant was born between January 2013 and June 2017, and (4) the infant was enrolled in the HITSystem for EID at one of the six hospitals.

Procedures

During the first EID enrollment visit, an infant file was created in the HITSystem where maternal demographic and antenatal care data, infant gender and date of birth, and infant HIV DNA PCR testing date were recorded. Subsequent infant services (PCR test results, mother notification of test results, date of ART initiation, and/or date and reason for early discharge) were recorded in the HITSystem, as they became available. Data were exported from the HITSystem to an Excel spreadsheet for analyses.

Study Measures

The dependent and potential predictor variables are described in Table 2.

Table 2.

Study Measures

Measure	Definition	Scaling
Dependent Variable
Infant age at ART initiationa	Infant’s age (in weeks) when initiated on ART	Early: ≤12 weeks Late: >12 weeks
Independent variables
Mother’s age	Mother’s age at infant’s enrollment into EID	Continuous variable, years
Study site	Site where mother-infant pair enrolled	Hospitals 1–6
Study site type	Type of HITSystem implementation	Research or programmatic
Mother’s education	Highest level of education achieved by the mother	0 (No formal education), 1 (Partial/completed primary), or 2 (Partial secondary or beyond)
Infant sex	Infant’s sex	Male or female
Department of enrollment	Where the infant presented for care	MCH or other
Year of HIV+ diagnosis	Year that the infant’s positive HIV result was available to the hospital	2013, 2014, 2015, 2016, 2017
Timeliness of infant testinbb	Infant’s age (in weeks) when first HIV sample was obtained	Per guidelines (<7 weeks) Outside guidelines (≥7 weeks)
TAT for mother notificationc	Time (in weeks) from when sample was obtained from infant until mother was notified of infant’s results	≤ 4 weeks >4 weeks
TAT for ART initiationd	Time (in days) from mother notification of infant’s positive HIV result to infant initiation of ART	Immediate (≤ 3 days) Delayed (>3 days)

MCH = maternal and child health department, where caregivers would present for routine EID services; TAT = turnaround time

^aCutoffs based on the CHER study that indicated the benefits of infant ART initiation by 12 weeks of age in reducing mortality and slowing disease progression (Violari etal., 2008).

^bCutoffs based on national guidelines recommending infant testing within 6 weeks of birth (Ministry of Health & Programme, 2015).

^cCutoffs based on the WHO recommendation that results from virological testing in infants be returned to the caregiver within 4 weeks of specimen collection (World Health Organization, 2016a).

^d3 days was used as the cutoff for early ART initiation to account for hospital and patient-level factors that may slightly defer ART (i.e. clinical officer unavailable on day of notification, mother wanted to consult with partner prior to infant initiation) but did not represent prolonged, systematic delays to initiation.

Analyses

The goals of this analysis were to describe the proportion of infants with HIV initiated on ART by 12 weeks postnatal and to identify predictors of “early” versus “late” ART initiation. We assessed bivariable and multivariable predictors of ART initiation by 12 weeks of age. Discrete variables are described using frequency and percentage. Continuous variables are described using median and interquartile range. Bivariable analyses were conducted using Chi-square and Wilcoxon Rank Sum tests, as appropriate. A multivariable stepwise logistic regression model was used to select the significant predictors. Based on previous literature suggesting maternal age as a significant predictor of pediatric HIV care, we controlled for maternal age (Goggin et al., 2019; Sandbulte et al., 2020; Watt et al., 2019). Variables significant at 0.05 from the bivariable analysis were selected for inclusion in the model. Given that differences in enrollment periods resulted in an uneven distribution of participants across years, year of diagnosis was excluded from the multivariable analysis, despite significance in the bivariable analysis. Statistical significance level was set at 0.05. All statistical analyses were conducted using SAS version 9.4 (SAS Institute, 2002-2012).

Ethical Considerations

This study was approved by Institutional Review Boards at the Kenya Medical Research Institute (SERU No. 3118) and University of Kansas Medical Center (Study 13793). All mothers enrolled in the cluster randomized control trial provided written, informed consent prior to their enrollment. Mothers enrolled at programmatic sites were informed about the HITSystem intervention that was implemented as part of the hospital’s standard EID services. Mothers provided oral informed consent to enroll in the HITSystem or were given the option to use only the paper-based registry without HITSystem support. Oral consent was deemed appropriate for programmatic participants since participation involved minimal risk and a signed consent form would have been the only document outside of medical records that linked participants to their HIV status.

Results

A total of 2,603 infants born between January 2013 and June 2017 were enrolled in the HITSystem at the six facilities included in this analysis, of whom 96 (3.7%) were HIV-positive. Among these 96 infants, a total of 82 (85.4%) initiated ART at a median infant age of 17.1 weeks (IQR 13.0-23.9). Of the 82 infants that started ART, only 17 (20.7%) initiated ART by 12 weeks of age. Median infant age at first HIV DNA PCR test was 7.6 weeks (IQR 6.1-14.6). Median turnaround time (TAT) from sample collection to mother notification of results was 4.9 weeks (IQR 4.0-22.3). Median TAT from mother notification of test results to ART initiation was 6.5 days (IQR 1.0-28.0).

Table 3 outlines characteristics of infants who initiated ART early versus late. In bivariable analyses, early ART initiation was associated with infant testing per guidelines, more rapid TAT for results notification, and more rapid TAT for ART initiation. Among infants who received early ART, median infant age at first test was 6.0 weeks, compared to 8.1 weeks for infants who received late ART (p<0.0001). Infants who received testing per guidelines were significantly more likely to initiate ART early compared to infants who received testing outside of the guidelines (42.1% vs 2.3%, p<0.0001). Median TAT for mother notification of results was significantly shorter for infants who received early ART compared to late (3.7 weeks vs 5.8 weeks, p<0.0001). Among infants whose caregivers were notified of their positive result within 4 weeks of sample collection, 44.4% were initiated on ART by 12 weeks of age, compared to 9.6% of infants whose caregivers were notified >4 weeks after sample collection (p=0.0004). Median time from mother notification of test result to ART initiation was significantly shorter for infants who received early ART compared to infants who received late ART (2.0 days vs 11.5 days, p=0.002). Among infants who were initiated on ART within 3 days of mother notification, 36.7% were initiated on ART by 12 weeks of age, compared to 10.4% of infants who initiated ART greater than 3 days after mother notification (p=0.005). Rates of early ART initiation did not vary significantly by site, with no site able to achieve greater than a 41.7% rate of early ART initiation. Rates of early ART initiation also did not vary by mother age, mother education, infant sex, or department of EID enrollment. While rates of early ART initiation did vary significantly by year, no clear trend emerged from 2013-2017. Thus, this variable was not included in the multivariable regression.

Table 3.

Characteristics of mother-infants pairs initiating ART

		Infant Age at ART Initiation
	Total sample (n=82)a	Early (≤12 weeks) N (%)	Late (>12 weeks) N (%)	P-value
Mother’s age [median (IQR)]	82	30.0 (25.0-34.0)	29.0 (24.0-32.0)	0.205
Study site				0.17
Hosp 1	12	5 (41.7%)	7 (58.3%)
Hosp 2	5	2 (40%)	3 (60%)
Hosp 3	19	4 (21.1%)	15 (78.9%)
Hosp 4	18	4 (22.2%)	14 (77.8%)
Hosp 5	21	1 (4.8%)	20 (95.2%)
Hosp 6	7	1 (14.3%)	6 (85.7%)
Study site type				0.052
Research	46	6 (13.0%)	40 (87.0%)
Programmatic	36	11 (30.6%)	25 (69.4%)
Mother’s education				0.28
No formal education	4	0 (0%)	4 (100%)
Partial/completed primary	42	11 (26.2%)	31 (73.8%)
Partial secondary or beyond	28	4 (14.3%)	24 (85.7%)
Infant sex				0.42
Male	46	11 (83.3%)	35 (16.7%)
Female	36	6 (76.1%)	30 (23.9%)
Department of EID enrollment				0.67
MCH	41	8 (19.5%)	33 (80.5%)
Other	20	3 (15%)	17 (85%)
Year of HIV+ Diagnosis				0.023
2013	6	4 (66.7%)	2 (33.3%)
2014	26	2 (7.7%)	24 (92.3%)
2015	23	4 (17.4%)	19 (82.6%)
2016	21	6 (28.6%)	15 (71.4%)
2017	6	1 (16.7%)	5 (83.3%)
Timeliness of infant testing				<0.0001
Per guidelines (<7 weeks)	38	16 (42.1%)	22 (57.9%)
Outside guidelines (≥7 weeks)	44	1 (2.3%)	43 (97.7%)
TAT: Mother notification				0.0004
≤ 4 weeks	27	12 (44.4%)	15 (55.6%)
> 4 weeks	52	5 (9.6%)	47 (90.4%)
TAT: ART initiation				0.0052
Immediate (≤ 3 days)	30	11 (36.7%)	19 (63.3%)
Delayed (> 3 days)	48	5 (10.4%)	43 (89.6%)

^aDenominators change due to missing values

In multivariable logistic regression analyses (Table 4), there were three statistically significant predictors of early initiation of ART. Timeliness of infant testing was the strongest predictor, with infants tested per guidelines having forty times the odds of early ART initiation compared to infants tested outside of the guidelines (aOR 40.14 [3.96 – 406.97], p=0.002). The odds of early ART initiation were about 11 times higher for infants with a <4 week TAT for mother notification compared with infants with a ≤4 week TAT (aOR 11.303 [2.02-63.34], p=0.006) and about 7 times higher for infants who initiated ART immediately compared to infants with delayed ART initiation (aOR 7.32 [1.41-38.03]).

Table 4.

Regression results from model for early versus late ART initiation

Variable	aOR	95% CI	P value
Mother’s age	1.07	0.91-1.25	0.41
Timeliness of infant testinga	40.14	3.96-406.97	0.002
TAT: mother notificationb	11.30	2.02-63.34	0.006
TAT: ART initiationc	7.32	1.41-38.03	0.018

^aReferent is per guidelines (<7 weeks)

^bReferent is ≤4 weeks

^cReferent is immediate (≤3 days)

A total of 14 (14.6%) infants did not receive ART. Of these infants who did not begin treatment, 6 (42.9%) passed away prior to ART, 3 (21.4%) transferred to another facility, 1 (7.1%) mother refused infant treatment, and 4 (28.6%) infants were lost to follow up. Median age of infant death among deceased infants was 27.1 weeks (IQR: 15.3-35.4).

Discussion

This study explored rates and predictors of early versus late ART initiation. Similar to other studies in Kenya (Finocchario-Kessler et al., 2018), the median infant age at ART in this study (17.1 weeks) was well past both the WHO recommended 12-week target and the peak mortality among HIV+ infants at 2-3 months of age (Bourne et al., 2009). Less than one-quarter (20.7%) of infants with HIV were initiated on ART within the recommended age range. Low rates of timely ART initiation among infants were seen across sites and throughout the 5 years of study implementation, with no site achieving >50% early ART initiation. While two-third of infants received early ART in 2013, this rate dropped significantly in later years and ranged from 7.7%-28.6%. Although not statistically significant, research sites (hospitals 4-6) trended towards lower rates of early ART initiation. The decline in rates of early ART observed after 2013 could be a result of the addition of these hospitals in the analysis in 2014. Guidelines prior to 2012 required baseline physical and psychosocial assessment that could take 2-4 weeks to complete prior to infant ART initiation (Ministry of Health, 2011). Hospitals 5 and 6, which contributed a combined 28 infants to the analysis in these later years, continued to require prolonged counseling prior to infant ART initiation through 2015, which presented significant delays to infant ART and likely contributed to this trend.

A study in the Democratic Republic of Congo and Malawi found that only infant age at HIV diagnosis was associated with ART initiation by 6-months of age (Sheahan et al., 2017). Using the more stringent cutoff of ART by 12-weeks, timeliness of infant testing remained a significant predictor of early ART initiation; however even when tested per national guidelines, 58% of infants still failed to initiate ART on time. This finding indicates that infant testing at 6-weeks is too late; thus, earlier testing may be required in order to achieve targets for timely ART initiation. Recognizing the barriers that low resource settings face in delivering timely EID services, the World Health Organization conditionally recommended at-birth testing for HIV, where feasible, in 2016 (World Health Organization, 2016a). The same year, Kenya began piloting at-birth infant HIV PCR testing prior to anticipated national rollout (Ministry of Health & National AIDS and STI Control Programme, 2016). In a pilot study in Lesotho, adding at-birth HIV DNA PCR reduced the median age at ART initiation from 14.8 weeks to 6.4 weeks (Gill et al., 2017). Adding at-birth testing to national guidelines may help achieve similar improvements in Kenya and other low resource settings. However, in the absence of universal at-birth testing, it is critical that EID programs are strengthened to increase the number of infants perinatally exposed to HIV who are tested within 6 weeks of age. Learning about EID from providers during pregnancy and perceiving less HIV stigma have been associated with on time infant HIV testing among infants exposed to HIV (Goggin et al., 2016); thus, engaging pregnant women through frequent and judgement-free counseling during antenatal care will be necessary to facilitate earlier infant testing and more timely ART initiation.

Shorter TAT for each key step (mother notification of results and ART initiation) was significantly associated with ART initiation by 12 weeks. However, even when notified within 4 weeks or initiated on ART within 3 days of mother notification, 55% and 63% of infants still failed to initiate ART by 12 weeks of age, respectively. Furthermore, despite guidelines in place since 2012 recommending immediate ART initiation for all infants with HIV (Ministry of Health, 2012), only 30 out of 78 of infants (38.5%) were initiated on ART within 3 days of their mother being notified of a positive diagnosis. Together, these findings suggest that the observed difficulty in achieving ART by 12 weeks is a multifaceted issue, compounded by late infant testing, long TAT for sample processing and caregiver notification, and delays with ART initiation. Median TAT from sample collection to caregiver notification in our study was 4.9 weeks, while TAT from notification to ART was 1 week. HITSystem data from other recent studies, with different sample populations, in Kenya have found longer TAT. Among all infants exposed to HIV, TAT from sample collection to result availability ranged from 3.6 – 5.4 weeks (Finocchario-Kessler et al., 2018; Wexler et al., 2017) and from result availability to caregiver was 2.0 - 3.2 weeks (Finocchario-Kessler et al., 2018). Though infant age at ART has decreased since the release of guidelines suggesting immediate ART for all infants with HIV, when median age of ART ranged from 3.7-45 months (Asbjornsdottir et al., 2016; Sridharan et al., 2017; Wamalwa et al., 2012) and the time from a positive diagnosis to ART initiation ranged from 2-9 weeks (Finocchario-Kessler et al., 2014; Wamalwa et al., 2012), current TAT still result in substantial delays and creates missed opportunities for early intervention.

In order to more widely achieve ART initiation by 12-weeks of age and optimize outcomes for infants with HIV, innovative strategies to streamline the testing process are needed. One emerging strategy is the use of point-of-care (POC) testing systems in low-resource settings. In 2016, the WHO approved point of care systems for infant HIV testing (World Health Organization, 2016b, 2016c). Such devices could facilitate same day infant testing, result availability, caregiver notification, and ART initiation, thus reducing a multi-month process to a matter of hours or days (Jani et al., 2017). These POC technologies are currently being evaluated for feasibility in Kenya and other similar settings (Ibrahim et al., 2017; Jani et al., 2017; Jani et al., 2014; Technau, Kuhn, Coovadia, Murnane, & Sherman, 2017). However, to maximize the potential benefits of these innovations, frequent engagement with key stakeholders in the EID process (clinicians and parents/caregivers) is still needed to ensure that infants are brought on time for testing and caregivers of infants with HIV are offered and prepared to accept ART initiation when they are first informed of their infant’s result.

While this study provides novel data regarding the timing of ART initiation and factors associated with timely ART initiation among infants, it has several limitations. First, the data presented here were collected in hospitals using the HITSystem intervention. Thus, these data may not reflect standard of care in EID settings. Even in the presence of this intervention, there were gaps in ART services for infants, which are likely amplified in non-intervention settings. Second, all mother-infant pairs assessed in this study did present for EID care and all analyses regarding timing of ART included only infants who started ART. We have no further information on mothers who never sought services or for the 8 infants who either transferred to another facility, were lost to follow up, or whose caregiver refused treatment. Future studies should explore ART initiation among infants diagnosed outside of EID care, confirm linkage to care amongst patients who transfer after a positive diagnosis, and assess outcomes for patients who were lost to follow up or transferred. Third, like the few studies that have previously reported on infant ART initiation, the generalizability of our results is limited by narrow geographic coverage and a relatively small sample size. Forth, the staggered start dates of study hospitals resulted in an uneven distribution of samples across sites and years; however, we believe that the benefit of a larger sample justified these decisions. Despite these limitations, considered together with previous literature our data strengthens the argument that current systems for EID in Kenyan are not achieving targets for infant ART initiation.

Linking infants living with HIV to appropriate treatment is one of the primary goals of EID. Despite this, the rates and timing of infant ART initiation and factors that impact these outcomes are critically understudied areas. National and international estimates for treatment initiation usually stratify by child (≤ 14 years of age) or adult (15+ years) only. Due to the unique challenges associated with diagnosing infants (PCR testing) and linking them with care (weight-based dosing, dependence on caregivers), better infant-specific national estimates (< 2 years) are desperately needed. Such data can gauge the effectiveness of EID and ART programs and will help understand and address barriers to infant HIV treatment.

Conclusion

This study found that only one-fifth of infants initiated ART by 12-weeks of age. Earlier age at infant testing, shorter TAT of sample processing and mother notification, and shorter TAT for ART initiation increased the likelihood of initiating treatment by 12 weeks of age. These findings provide evidence that earlier testing and streamlined processes may facilitate more timely ART initiation in Kenyan infants.

Significance.

Early initiation of ART for infants living with HIV is one of the primary goals of EID services; yet, data on the rates, timing, and factors associated with ART initiation among Kenyan infants are extremely limited. No infant-specific national estimates for ART initiation are available and few studies have evaluated the timing and predictors of ART initiation in infants. This study assesses the proportion of infants initiated on treatment by 12 weeks of age and describes factors associated with early initiation of ART. This data can help policy makers and healthcare providers better tailor engagement to maximize the benefits of EID.