Factors associated with poor outcomes among people living with HIV started on anti-retroviral therapy before and after implementation of "test and treat" program in Coastal Kenya.

OBJECTIVES: To determine the factors associated with poor outcomes among people living with HIV (PLHIV) started on anti- retroviral therapy before and after implementation of "Test and treat" program in 18 facilities in Coastal Kenya.
METHODS: A retrospective cohort study design was used to study PLHIV aged > 15 years and started on ART in the periods of April to August 2016, and April to August 2017, then followed up for 24 months. Primary outcome was retention defined as being alive and on ARVs after 24 months. Death and loss to follow-up were considered as poor outcomes. Kaplan-Meier survival methods were used to describe time to primary outcome. Cox proportional regression analysis was used to determine factors associated with poor outcomes.
RESULTS: 86 patients (470 before test and treat, and 316 after test and treat cohorts) were enrolled. Overall, the median [IQR] age was 39.3 [32.5-47.5] years and 539 (69%) were female. After 24 months, retention rates for the before (68%) and after (64%) test and start groups were similar (absolute difference: -4.0%, 95%CI: -11-3.1, P = 0.27). There were 240(31%, 95%CI 27 to 34%) PLHIV with poor outcomes, 102 (32%) and 138 (29%) occurred among the test and treat group, and delayed treatment patients respectively. In multivariable regression model, test and treat had no significant effect on risk of poor outcomes (aHR = 1.17, 95%CI 0.89-1.54). Increasing age (aHR = 0.98, 95%CI 0.97-0.99), formal employment (aHR = 0.42, 95%CI 0.23-0.76) and not being employed (aHR = 0.53, 95%CI 0.34-0.81) were negatively associated with poor outcomes. The risk of poor outcomes was higher among males compared to female patients (aHR = 1.37, 95%CI 1.03-1.82) and among divorced/separated patients compared to the married (aHR = 1.44, 95%CI 1.04-1.99).
CONCLUSION: Retention patterns for the "test and treat" cohort were comparable to those who started ART before "test and treat". Patients who are males, young, divorced/separated, with poor socio-economic status had higher risks for poor clinical outcomes. Interventions targeting PLHIV who are young, male and economically disadvantaged provide an opportunity to improve the long-term outcomes.

Introduction

Significant progress has been made in the fight against HIV and AIDS with over 25.4 million people living with HIV (PLHIV) on Anti-Retroviral Therapy (ART) out of the 38 million PLHIV globally [1]. Majority of the PLHIV are in Eastern and Southern Africa, accounting for 53% (19.6 million) of the global burden, with about 1.5 million being Kenyans [2]. By 2017, about 12.9 million (66%) PLHIV in the region were accessing antiretroviral therapy, among them 1.16 million Kenyans [3]. Also, by 2017, the estimated percentage of people living with HIV who achieved viral suppression in the region was 52%, and 56–60% in Kenya [1].

The uptake of ART accelerated in the recent 5 years due to increased access to ART as a result of the World Health Organization (WHO) guidance released in 2015 on when to start ART and Pre-exposure prophylaxis (PrEP) [4], for countries to treat all HIV infected people with Highly Active Anti-Retroviral drugs (HAART) irrespective of their CD4+ levels or WHO stage. Kenya adopted the guidance in July 2016 with a campaign conducted to initiate ART to all the PLHIV who were in care but not started on ART [5]. This was implemented in the Coast counties from September 2016 after capacity building of health care workers and distribution of commodities. By March of 2017, all the clients in Mombasa, Kilifi and Kwale Counties who were on care had been started on ART, while newly identified PLHIV were immediately started on ART as per the new guidelines, as soon as they were identified and adequately prepared to continue with treatment. The new guidance was famously referred to as the “Test and treat” locally and required PLHIV to start ART immediately or within 14 days of HIV diagnosis. Prior to the “Test and treat”, the National AIDS and STI Control (NASCOP) guidelines of 2014 [6] were in use which required only PLHIV who met one or more of the following criteria to be started on ART: pregnant or breastfeeding women, children below 10 years, people with CD4 <500 cells/ml, WHO stage 3 or 4, TB/HIV con-infection and Hepatitis B co-infection.

Based on scientific evidence that was available then, a mathematical model done in 2009 on universal voluntary HIV testing with immediate antiretroviral therapy as a strategy for elimination of HIV transmission, showed that immediate ART for those identified as HIV positive would reduce HIV associated mortality to less than 1 case per 1000, and reduce HIV prevalence to less than 1% from the high of almost 5% in most Sub-Saharan Africa countries [7]. A prospective cohort study was done at 14 sites in 7 African countries (Botswana, Kenya, Tanzania, Uganda, Rwanda, South Africa, and Zambia) involving 3,381 sero-discordant couples who were followed up for a period of 24 months. In the cohort, 349 of the study participants who were HIV positive and eligible for ART were started on ART appropriately as per guidelines. At the end of the study, out of the 103 new HIV infections among the previously sero-negative partners, only one was linked to a client who was on ART, demonstrating a 92% reduction in transmission rates in the group who were on ART [8]. In Uganda, a study among sero-discordant couples showed similar findings with marked reduction in HIV transmission with the use of ART from an incidence of 9.2/100 person-years to zero infections [9]. The TEMPRANO ANRS 12136 study and the Strategic Timing of Antiretroviral Treatment (START) study provided further important evidence to support universal ART by demonstrating better clinical outcomes in HIV asymptomatic patients who start ART at an early stage of their disease, when CD4+ cell counts are above 500 cells per cubic millimeter [10, 11].

Although studies have demonstrated overwhelming evidence of the benefit of “Test and treat” on reducing HIV transmission [7], evidence on other outcomes like attrition from care have been inconsistence. The retention rate reported in a pooled analysis of 154 cohorts in Low- and Middle-Income countries was approximately 70%, two years after starting ART (2008 to 2013) [12]. Recent data after the universal test and treat policy suggest the retention has not improved, in Democratic Republic of Congo, the retention rate was 77% after two years of starting ART following the adoption of test and treat policy [13]. The same study found higher rate of attrition after the adoption of the universal test and treat policy. Another cohort in Masaka, Uganda, found PLHIV starting ARTs within seven days of HIV diagnosis had higher risk of lost to follow-up [14]. Systematic reviews report varying interventions to improve retention on care but with mixed outcomes [15]. Various studies have also reported different predictors of attrition for different follow-up times and in different settings, some before and others after the universal test and treat policy [13, 14, 16].

With the adoption of the “Test and treat”, new evidence is thus needed on the retention patterns and identification of individual and health system factors that are associated with clinical outcomes. In this study, we evaluated the effectiveness of the “Test and treat” program in eighteen [17] facilities spread across three coastal counties (Mombasa, Kilifi and Kwale) of Kenya, and provided evidence on factors associated with poor clinical outcomes among PLHIV on ART.

Materials and methods

Study setting

This study was conducted in 18 HIV treatment centers in 3 coastal counties of Kenya; namely Mombasa, Kilifi, and Kwale. It involved the 6 centers with the highest number of clients started on ART in the April to August 2017 period in each of the counties, which are also referral centres for the counties and therefore gave a representation of PLHIV from all corners of the three counties. The three counties have a combined population of 3 million and, by time of study, the HIV prevalence was 4.1% in Mombasa and 3.8% for both Kilifi and Kwale(17), with majority of the population living below the poverty line.

Study design

This was a retrospective observational cohort study in which attrition and retention for two cohorts of PLHIV was collected retrospectively, analyzed and compared. The first cohort included PLHIV who started ART in the period of April to August 2016 before the counties of Kilifi, Kwale and Mombasa implemented the WHO recommendations of universally treating all HIV positive people with Highly Active Anti-Retroviral Therapy (HAART). The second cohort included PLHIV started on ART in the period of April to August 2017 after the “Test and treat” guidance was implemented. The period between the two cohorts was excluded because it was assumed to be contaminated since the old clients who had not started ART based on the recommendations of the old guidelines were being transitioned to the new guidelines.

Study population

All PLHIV initiated on ART aged >15 years in the periods between April to August 2016, and April to August 2017 were recruited. Records documented in the ART registers formed the sampling frame for the study, which were 1623 and 1496 for before and after “Test and treat” respectively.

Sample size determination

The sample size was estimated on the basis of statistical power to show significantly higher hazard of Lost to Follow Up (LTFU) among the “test and treat” cohort, compared to the delayed treatment cohort. In a South African cohort, the adjusted hazard ratio of LTFU among the test and treat patients compared to the delayed treatment was 1.58 [17] while the proportion of LTFU among HIV patients was 33.6% in Kilifi, Kenya [18] and 34% in Nigeria [19].

Assuming a LTFU of ~30%, a two-tailed alpha of 0.05, with statistical power >80%, a sample size of at least 600 HIV patients (300 in each cohort) was enough to show a 58% higher risk of LTFU among the HIV patients starting ARTs under test and treat policy (aHR 1.58) with 207 expected LTFUs. However, the study collected and analyzed data from 786 patients (316 for “test and treat” and 470 for delayed treatment) [20].

Sampling techniques

Probability proportionate to size sampling was used where each facility contributed numbers proportional to the numbers started on ART in the study period. Then, within each facility, simple random sampling was used to select the sample.

Study variables

In this study, the outcome was retention defined as a state where patients were known to be alive and receiving ART at the end of the follow-up period [16], in this case at every 3 months for 24 months after starting ART for both cohorts. All PLHIV who died or were lost to follow up (LTFU) were considered to have poor outcome. Attrition was construed to have occurred in both cohorts if a client discontinued taking ART for any reason, including death, loss to follow-up, and stopping ARV medications [16, 21]. Socio-demographic factors (age, gender/sex, and marital status), socio-economic factors (education level, occupational status) and clinical features (nutritional status, WHO stage, presence of opportunistic infections, ART regimen) were the independent variables in this study.

Data collection tools and sources

Quantitative data was extracted through desk review from ART registers and patient files in the selected facilities using a data abstraction tool translated in the online Open Data Kit (Kobocollect©), downloaded and stored in Microsoft excel database then backed up externally.

Ethical considerations

In accordance with the principles governing research involving human participants, this study ensured that respondents’ ethical rights were upheld through submission, review and approval of the study proposal by the Pwani University Ethical Review Committee (ERC/MSc/032/2020). Only anonymized data were extracted from patients’ files. All data collected as part of this study was handled with utmost confidentiality.

Statistical analysis

Study data were extracted from patients’ records using standard questionnaire designed on Open Data Kit (Kobocollect©) and exported to STATA Version 16.1 (College Station, Texas 77845 USA) for analysis. Continuous variables were assessed for outliers by plotting visual aids like histograms, scatter plots and q-q plots for assessing normality. Outliers and illogical variables were flagged and corrected by checking correct values in patient records. Body Mass Index (BMI) was calculated as weight (Kg) divided by square of height in metres and grouped following WHO classification: <18.5, 18.5 to 24.9, 25.0 to 29.9 and ≥30.0. Data was assumed not to be missing at random, an extra category `missing’ was added to each variable to ensure all patients were included in the regression models.

Continuous variables were reported as mean (±SD) or median (with IQR), depending on underlying distribution. Categorical variables were reported as counts with their respective percentages. Categorical variables were compared using chi-square test or Fishers’ exact test between the two cohorts while continuous variables were compared using non-parametric Wilcoxon Rank Sum Test. The study main exposure was a binary variable classified as patients who were diagnosed with HIV and started on ARTs before the policy of “test and treat” which was introduced in 2016 and those diagnosed from 2017 onwards. This was a cohort design with 24 months of follow-up after starting ART, hence the ‘`test and treat” patients were those diagnosed and started on ART in the course of 2017 and followed up for 24 months ending in 2019. Other exposures explored in the regression analysis were demographic, socio-economic and clinical features at time of starting.

The study outcome was retention up to 24 months after starting ART. The retention rates in the study at all the time points were calculated as follows: where Co is all patients initiated on ART in the cohort;

Tt is all patients transferred out of care by time t;

Dt is all patients who died by time t; and

LTFUt is all patients lost to follow-up by time t.

%RTt will therefore be the proportion of all patients-initiated ART in the cohort who did not transfer out of care, are still alive and in care at time t. The retention rates were reported as proportions and the differences in all the outcomes between “test and treat” and delayed treatment patients compared using two-sample test of proportions and the absolute differences reported.

All PLHIV who died or were lost-to-follow-up (LTFU) were considered to have poor outcome. Time to poor outcome t was defined from date of starting ARTs to date of the events or completing 24 months of follow-up for those who were actively on ARTs and on follow-up after 24 months. Probability distributions of each event during 24 months of follow-up were calculated using the Kaplan–Meier survival approach and compared between the groups (delayed and “test and treat” groups) using log-rank test. PLHIV who transferred out of the study area were right censored at the time of leaving the cohort. To explore the effect of “test and treat” versus delayed treatment on poor outcome, Cox proportional hazard regression was performed with the main exposure (“test and treat” or delayed treatment) and adjusted for confounders collected. The proportional hazard assumption was tested using Schoenfeld residuals method. To account for HIV treatment care and other unobserved heterogeneity across the three counties (Kwale, Mombasa and Kilifi), shared gamma frailty Cox regression models were performed. A base model was run with the main exposure adjusted for age and sex with the three counties as random effect component in the shared gamma frailty Cox regression models. The final multivariable models included all other confounders collected at time of starting ART. CD4 counts were excluded in the regression models because a large proportion of patients (>50%) had no CD4 results at the time of starting ARVs. The measure of effect reported was adjusted hazard ratios and their respective 95% confidence intervals. Final multivariable discriminatory power was assessed using Area Under the Receiver Operating Characteristics curve (AUC).

We performed a sensitivity analysis including only the test and treat cohort PLHIV who were started on ARTs with fourteen days of HIV diagnosis as recommended by the policy. All the PLHIV in the delayed cohort were included in this sensitivity analysis.

Finally, retention was assessed in the ART programme after 24 months of starting ARTs versus the collapsed poor outcomes (deaths or LTFU). Transfer out were right censored at the time of leaving the cohort. A binary outcome was created; either being active on ARTs at the end of 24 months follow-up or having one of the poor treatment outcomes. Using the same approach as with other outcomes above, shared gamma frailty Cox regression models were run and reported as adjusted hazard ratios.

Results

Descriptive statistics of study participants

The study enrolled 786 patients with 470 (60%) being from the cohort before “test and treat” and 316 (40%) in the “test and treat” cohort. A total of 341 (44%) patients were from Kilifi County, 332 (42%) were from Mombasa County, and 113(14%) from Kwale County. Overall, the median (IQR) age was 39.3(32.5–47.5) years, 539 (69%) of the patients were females, and majority (423 or 54%) were married. Only 136 (17%) were economically dependent, 236 (30%) were unemployed and 321 (41%) had secondary level education (Table 1).

Table 1

Descriptive characteristics.

Characteristic	Total patients (N = 786)	Test and treat group (N = 316)	Delayed group (N = 470)	P-value
Demographic characteristics
Sex
Female	539 (69)	212 (67)	327 (70)	0.46
Male	247 (31)	104 (33)	143 (30)
Age in years: median (IQR)	39.3(32.5–47.5)	38.9(32.0–46.4)	39.4 (32.5–47.5)	0.57
Recruiting County
Mombasa	332 (42)	85 (27)	247 (53)	<0.001
Kwale	113 (14)	49 (16)	64 (14)
Kilifi	341 (44)	182 (58)	159 (34)
Marital status
Married	423 (54)	165 (52)	258 (55)	0.08
Single	142 (18)	48 (15)	94 (20)
Divorced/separated/widowed	221 (28)	103 (33)	118 (25)
Socio-economic characteristics
Education level
None	50 (6.4)	24 (7.6)	26 (5.5)	<0.001
Primary	279 (36)	136 (43)	143 (30)
Secondary	321 (41)	117 (37)	204 (43)
Tertiary	136 (17	39 (12)	97 (21)
Employment status
Self employed	201 (26)	89 (28)	112 (24)	0.02
Informal employment	84 (11)	34 (11)	50 (11)
Not employed	236 (30)	106 (34)	130 (28)
Economic status
Independent	250 (32)	99 (31)	151 (32)	0.53
Semi-independent	400 (51)	168 (53)	232 (49)
Dependent	136 (17)	49 (16)	87 (19)
Clinical characteristics
Body Mass Index (BMI)
<18.5	119 (15)	59 (19)	60 (13)	0.04
18.5 to 24.9	390 (50)	147 (47)	243 (52)
25 to 29.9	120 (15)	45 (14)	75 (16)
≥ 30	64 (8.1)	20 (6.3)	44 (9.4)
Missing	93 (12)	45 (14)	48 (10)
WHO Infection stage
Stage I	527 (67)	208 (66)	319 (68)	0.08
Stage II	193 (25)	86 (27)	107 (23)
Stage III	64 (8.1)	20 (6.3)	44 (9.4)
Stage IV	2 (0.3)	2 (0.6)	0 (0)
Starting ART regimen
TDF/3TC/EFV	732 (93)	301 (95)	431 (92)	0.05
AZT/3TC/NVP	21 (2.7)	4 (1.3)	17 (3.6)
TDT/3TC/NVP	11 (1.4)	6 (1.9)	5 (1.1)
Others*	22 (2.8)	5 (1.6)	17 (3.6)
CD4 level before ART initiation# (cells/mm^3) Median (IQR)	355 (172–514)	308 (172–440)	369 (173–558)	0.06
Number of adherence sessions before ART initiation
≤1	265 (34)	125 (40)	140 (30)	0.02
2	129 (16)	47 (15)	82 (17)
≥3	280 (36)	109 (35)	171 (36)
Missing data	112 (14)	35 (11)	77 (16)
Had opportunistic infection	62 (7.9)	29 (9.2)	33 (7.0)	0.21
Days to starting ARVs after HIV diagnosis, Median (IQR)	6 (0–118)	0 (0–8)	34 (12–567)	<0.001

*ABC/3TC/LPV/r, ABC/3TC/EFV, AZT/3TC/EFV, D4T/3TC/NVP and TDF/3TC/NVP,

#CD4 test were not systematically conducted for the “test & treat” group at starting ARVs because they were not required to decide when to start ARVS, therefore only 283 patients had CD4 data, 73/316 (23%) among those in test & treat group,

All results are N (%) unless where specified,

IQR; Interquartile range,

The P-values are from chi-square test or Fishers’ exact test (where any n<5) for categorical variables and Wilcoxon Rank Sum Test for continuous variables.

For clinical characteristics, approximately half of the patients had normal BMI (18.5 to 24.9), while 119 (15%) were underweight (BMI<18.5). A total of 732 (93%) patients were initiated on TDF/3TC/EFV as their starting regimen. Approximately two-thirds (67%) of the patients were classified as WHO stage I while only 2 (0.3%) were in WHO stage IV. Only 73 (9.3%) had ≥500 CD4 cells/mm^3 while 279 (36%) had <500 CD4 cells/mm^3. Overall, the CD4 count was not significantly different between the two cohorts (P = 0.06) including across the different WHO stages (Fig 1). There were only 62 (7.9%) patients with opportunistic infections; with 36 (59%) having TB, 9 (14%) with herpes simplex virus, 4 (6.3%) with sexually transmitted infections, 6 (9.4%) with bacterial infections and 7 (11%) with oral candidiasis. Overall, the median (IQR) days to starting ARVs after HIV diagnosis was 6 (0 to 118), it was 0 (0 to 8) among patients on test & treat compared to 34 (12 to 567) days among the delayed cohort (P<0.001). Among the 316 patients on test & treat cohort, 233/316 (74%) started ART within seven days of HIV diagnosis. (Table 1).

Fig 1

Comparing CD4 count levels for test and start and delayed cohorts across the WHO stages.

*All the Patients in WHO stage IV were missing CD4 counts.

Retention patterns

In the first three months after starting ART, the retention rates were 88% for the “test and treat” cohort and 84% for those started on treatment before “test and treat”, the absolute difference being 4.2%. However, this difference was not significant (95% CI -0.8–9.2%, P = 0.10). After two years of treatment with ARVs, the retention rates declined in both the “test and treat” and delayed treatment groups to 64% and 68% respectively, absolute difference being -4.0%. This difference was also not significant (95%CI -11-3.1%, P = 0.27) (Table 2). The retention rates for the two cohorts were not significantly different on any of the follow-up months (Fig 2).

Table 2

Cumulative retention rates from month 3 to 24 for “test and treat” and before “test and treat” cohorts.

Month	Test & start cohort (n = 316)	Before test and treat cohort (n = 470)	Absolute difference (95% CI)	P-value*
3	269 (88)	375 (84)	4.2 (-0.8 to 9.2)	0.10
6	243 (81)	358 (81)	-0.2 (-5.9 to 5.6)	0.95
9	237 (79)	340 (77)	1.6 (-4.5 to 7.6)	0.62
12	228 (76)	330 (75)	1.1 (-5.2 to 7.4)	0.74
15	210 (71)	316 (72)	-1.4 (-8.1 to 5.2)	0.67
18	204 (69)	312 (71)	-2.3 (-9.1 to 4.5)	0.50
21	191 (66)	305 (70)	-3.9 (-10.9 to 3.0)	0.26
24	183 (64)	295 (68)	-4.0 (-11.1 to 3.1)	0.27

*P-values from two-sample test of proportions

Fig 2

Comparing retention rates at 3 months intervals for 24 months for cohorts before and after "test and treat" with their 95% confidence intervals.

Patterns of outcomes

During follow-up for 1,144 person-years, 240/786 (31%, 95%CI 27–34%) PLHIV had poor outcomes (died or LTFUs), a rate of 211 (95%CI 186–240) poor outcomes per 1,000 person-years. Out of the 240 poor outcomes, 214/240 (89%) and 26/240 (11%) were LTFUs and deaths respectively. Of these 240 poor outcomes, 102/316 (32%) and 138/470 (29%) occurred among the `test and treat’ and delayed treatment patients: rates of 227 (95% CI 187–275) and 202 (95%CI 171–238) per 1,000 person-years respectively, age, sex and county adjusted HR 1.09 (95%CI 0.84–1.42) (Fig 3). Of the 240 poor outcomes, 52/240 (22%) occurred on the day of starting ART; with 23/102 (23%) in the ‘test and treat’ cohort and 29/138 (21%) from the delayed treatment cohort (P = 0.36).

Fig 3

Kaplan-Meier curve of not having poor outcomes for 24 months after starting ART.

The KM curve starts after one because of the poor outcomes that occurred at day zero.

Individual level factors associated with poor outcomes

In the multivariable regression model, “test and treat” had no significant effect on risk of poor outcomes (aHR = 1.17, 95%CI 0.89–1.54). However, increasing age was associated with protective effect on hazard of poor outcomes (aHR = 0.98, 95%CI 0.97–0.99). The hazard of poor outcomes was higher among male compared to female patients (aHR = 1.37, 95% CI 1.03–1.82). Compared to self-employed patients, those with lower risk of poor outcomes were the formally employed (aHR = 0.42,95%CI 0.23–0.76) and not employed (aHR = 0.53, 95%CI 0.34–0.81). Divorced/separated patients had significantly higher hazard of poor outcome (aHR 1.44, 95%CI 1.04–1.99) compared to married patients. Other features explored were not associated with poor outcomes (Table 3). The multivariable regression model AUC was 0.66 (95%CI 0.62–0.70).

Table 3

Multivariable analysis of individual level factors associated with poor outcomes.

Factors	Poor outcomes (N = 240) #	Adjusted HR (95% CI) *	P-value
Type of treatment
Delayed treatment	138 (29)	Reference
Test and treat	102 (32)	1.17 (0.89–1.54)	0.27
Age in years
<30	43 (18)	Reference
30 to 40	85 (35)	0.84 (0.56–1.27)	0.41
40 to 50	74 (31)	0.70 (0.44–1.10)	0.13
≥ 50	38 (16)	0.54 (0.32–0.91)	0.02
Sex
Female	150 (28)	Reference
Male	90 (36)	1.42 (1.07–1.88)	0.02
Marital status
Married	118 (28)	Reference
Single	46 (32)	1.01 (0.69–1.48)	0.97
Divorced/separated	67 (37)	1.39 (1.01–1.92)	0.04
Widowed	9 (23)	0.98 (0.48–2.02)	0.96
Education level
No school	18 (35)	1.51 (0.85–2.67)	0.15
Primary	82 (30)	0.91 (0.67–1.24)	0.56
Secondary & above	140 (31)	Reference
Employment status
Self employed	88 (33)	Reference
Informal employment	76 (38)	1.23 (0.89–1.69)	0.22
Formal employment	14 (17)	0.41 (0.22–0.74)	0.004
Not employed	62 (26)	0.52 (0.34–0.80)	0.003
Economic status
Independent	74 (30)	Reference
Semi-independent	124 (31)	1.10 (0.79–1.53)	0.59
Dependent	42 (31)	1.55 (0.93–2.58)	0.10
BMI group
<18.5	46 (38)	1.29 (0.90–1.85)	0.16
18.5 to 24.9	115 (30)	Reference
≥ 25	44 (24)	0.82 (0.58–1.18)	0.29
Missing	35 (38)	1.46 (0.95–2.23)	0.08
Type of ART
TDF/3TC/EFV	219 (30)	Reference
Others**	21 (39)	1.39 (0.87–2.22)	0.16
WHO stage
Stage I	155 (29)	Reference
Stage II	59 (31)	1.09 (0.78–1.54)	0.60
Stage III & IV	26 (39)	1.45 (0.86–2.45)	0.16
Adherence to counseling sessions before ART initiation
≤1	78 (30)	1.06 (0.76–1.48)	0.72
2	43 (33)	1.24 (0.84–1.83)	0.28
≥3	77 (28)	Reference
Missing	42 (38)	1.79 (1.18–2.70)	0.006
Had opportunistic infection	19 (30)	0.84 (0.47–1.50)	0.56

*Adjusted HR from shared gamma frailty Cox model with the county as a random intercept

**ABC/3TC/LPV/r, ABC/3TC/EFV, AZT/3TC/EFV, D4T/3TC/NVP and TDF/3TC/NVP

# Results are N (%).

Health system factors associated with poor outcomes

Having regular facility staff meetings was associated with significant lower hazard of poor outcomes (aHR = 0.66, 95%CI 0.47–0.91, P = 0.01). The cadres of health-care workers involved in initiating clients on ART did not influence the clinical outcomes of PLHIV.

In the sensitivity analyses, including only the 233 PLHIV among the test & treat cohort who started ART within fourteen days of HIV diagnosis according to the guidelines, the retention rates at all the follow-ups were not significantly different (S1 Table). After adjusting for all exposures in Table 3, test & treat per policy was not associated with poor outcomes (aHR 1.22, 95%CI 0.90–1.66, P = 0.20). After adjusting for the health system variables on Table 4, test & treat per policy was not associated with poor outcomes (aHR 1.15, 95%CI 0.86–1.54, P = 0.35).

Table 4

Multivariable analysis of health system factors associated with poor outcomes.

Factors	Poor outcomes N = 240 (%)	Adjusted HR (95%CI) *	P-value
Type of treatment
Delayed treatment	138 (58)	Reference
Test and treat	102 (43)	1.14 (0.87–1.50)	0.33
Cadre of health worker initiating ART
Medical officer	4 (1.7)	0.85 (0.30–2.40)	0.76
Clinical officer	169 (70)	0.81 (0.59–1.13)	0.22
Nurse	114 (48)	1.30 (0.90–1.86)	0.16
Counsellor	86 (36)	1.09 (0.73–1.62)	0.67
Peer educator	40 (17)	1.19 (0.78–1.82)	0.41
Community health volunteer	23 (9.6)	1.06 (0.66–1.71)	0.80
Mentor mother	17 (7.1)	0.98 (0.56–1.71)	0.95
Laboratory technician	29 (12)	1.41 (0.87–2.28)	0.16
Pharmaceutical technician	68 (28)	0.98 (0.63–1.52)	0.93
Health Records Information Officers	74 (31)	0.77 (0.53–1.13)	0.19
Presence of regular staff meetings
Quality work improvement team meetings	61 (26)	0.71 (0.47–1.05)	0.09
Regular facility staff meetings	119 (51)	0.66 (0.47–0.91)	0.01
CCC departmental clinical meetings	162 (70)	0.83 (0.59–1.18)	0.30
Multi-disciplinary team meetings	136 (59)	1.03 (0.74–1.41)	0.88

*Adjusted HR from shared gamma frailty Cox model with the county as a random intercept

#Results are N (%).

Discussion

Survivorship/Attrition patterns

In this study, the retention rates between the cohorts of PLHIV started on ART before and after implementation of test and treat were not significantly different. In both cohorts, the retention declined to a low of 64% and 68% among the test and treat and the delayed cohorts respectively after 24 months of starting ART. The retention at one year is comparable to that of 89% that was found in a 2017 study for test and treat patients in Kenya and Uganda [22]. A cohort study done in Uganda at almost the same time (January 2015 to December 2017) involving 646 patients concluded that “there was no significant difference (P = 0.231) in the mean retention times of patients initiated on ART based on CD4 cell count compared to those initiated under the “test and treat” strategy [23]. The retention rates are similar to other studies globally as was demonstrated by Fox & Rosen, who averaged retention to be 78% at 12 months, 71% at 24 months, and 69% at 36 months across all regions in a meta-analysis of 154 cohorts of PLHIV from 42 countries: 24 in Africa (114 cohorts), 10 in Asia (28 cohorts), and 8 in Latin American Countries (12 cohorts) [12, 24]. However, in a retrospective study in Malawi, immediate ART was found to be associated with low retention rates among PMTCT mothers. A multivariable analysis of 456 pregnant women on ART showed that “initiation of ART on the same day as HIV diagnosis, was independently associated with reduced retention in the first six months [25].

In this study, one in every ten cases of attrition were caused by death which is inconsistent with the study of Flynn et al., 2017 who observed death as the largest cause of attrition at 80% and 63% of the deaths which occurred in the first year of HIV therapy in a prospective cohort study in Uganda [26]. In our study, one in every five poor outcomes occurred on the day of starting ART and did not differ between the two cohorts, 23/102 (23%) for the test and treat cohort and 29/138 (21%) for the delayed treatment cohort underlying the fact that starting patients on ART immediately upon diagnosis did not lead to poor outcomes. This underscores the need to do more to keep these patients on treatment. Thorough adherence counselling and treatment preparation should be prioritized when initiating ART to improve outcomes.

Individual level factors associated with poor outcomes

Similar to previous studies, increasing age was significantly associated with protective effect on hazard of poor outcomes, with the risk of poor outcomes reducing by 2% for every year gained [27, 28]. This could be due to the reason that older people living with HIV are able to accept their HIV status and adhere to treatment, hence the better outcomes as opposed to adolescents who often struggle with self-stigma, relationships and self-identity issues.

Males had almost 40% higher likelihood of having poor outcomes compared to female patients. These findings are supported by other studies from Kenya, Uganda, Malawi and Nepal that found male gender to be associated with poor outcomes [29-32]. Among the reasons for the poor outcomes among men in Kenya are the poor health seeking behavior, less psychosocial support systems and higher levels of stigma compared to women. Divorced/separated patients were one and half times more likely to have poor outcomes compared to married patients which may be due to the lack of consistent support systems that may be lacking. This is in contrast to a study in Kilifi—Kenya [29] which did not find marital status to be associated with poor outcomes.

Compared to self-employed patients, formally employed patients had half the risk of having poor clinical outcomes, which is supported by other studies like the systematic review and meta-analysis led by Nachega which included 28 studies published between 1996 and 2014 involving 8743 HIV-infected individuals from 14 countries [33]. Inconsistent with our expectation, in this study, patients without any form of employment were found to have almost half the risk for poor outcomes when compared to those in self-employment. This is an area that requires further research to affirm and explain the findings.

This study found that patients who were economically dependent had more than 60% higher risk of poor outcomes compared to those who were economically independent. This is consistent with studies in Ethiopia [34] and Zambia [35] which concluded that “poor households are more likely to experience an AIDS death” and “low socioeconomic status in patients hospitalized for HIV/AIDS were more likely to die than high socioeconomic status inpatients” respectively. Such patients lack transport to access ART services, cannot afford nutritious food and other related services that would improve their clinical outcomes.

Patients who had missing documentation on the number of adherence counselling sessions before ART initiation had close to twice the risk for poor outcomes compared to those who had 3 or more counselling sessions. In most cases, patients without documentation of adherence did not receive any adherence counselling sessions at all. In a study among PMTCT clients in Ethiopia [36], women who received adherence counselling were more than 4 times likely to adhere to ART thereby leading to good clinical outcomes.

Health system factors associated with poor outcomes

Among the health system factors analyzed, only having regular facility staff meetings was associated with significant lower hazard of poor outcomes. Regular staff meetings lead to coherence, stronger integration and team work among staff members. Team work through involving different staff cadres in a facility improved patient outcomes in a Chicago facility, USA [37]. However, the settings and staffing patterns are not similar to the Kenyan context, therefore, it may not be entirely applicable to the Coast of Kenya facilities. A Kenyan human resources study in public service demonstrated that practices that enhance teamwork improved staff performance [38]. The type cadres of health care workers involved in imitating clients on ART did not influence the clinical outcomes of PLHIV.

Strengths and limitations of the study

The strength of the study is the systematically collected data for 24 months after starting ART in a real-world setting. Although this study followed all the laid-out guidelines including the STROBE guidelines [39] for reporting, it still has a few limitations. First, the study relied on data in patient files and registers. CD4 counts were not systematically tested especially for the test & treat cohort who did not need this information to decide whether to start ART or not. It is therefore likely the CD4 results presented are subject to bias and thus the comparison made on Table 1 should be interpreted cautiously. Second, follow up for patients who were found to be lost to follow up was not done. In some studies, it is reported that some of the lost to follow up patients restarted treatment in other facilities and could be active on ART while others could have died. The study did not have the capacity to identify clients that had been diagnosed earlier, started ART and stopped, and later came back as newly diagnosed PLHIV. Lastly, since this is a retrospective study, bias arising from unmeasured features cannot be entirely ruled out.

Conclusion and recommendations

Our findings suggest the “test and treat” program is as effective as the previous policy in retaining PLHIV on ARTs for at least 24 months after initiating ART. Interventions targeting PLHIV who are young, male and economically disadvantaged provide an opportunity to improve the long-term outcomes.

(TIF)

Click here for additional data file.

(CSV)

Click here for additional data file.

Cumulative retention rates from month 3 to 24.

(TIF)

Click here for additional data file.

17 Nov 2021

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Manish Sagar, MD

Academic Editor

PLOS ONE

Journal requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf.

2. Please provide additional details regarding participant consent. In the Methods section, please ensure that you have specified (1) whether consent was informed and (2) what type you obtained (for instance, written or verbal). If your study included minors, state whether you obtained consent from parents or guardians. If the need for consent was waived by the ethics committee, please include this information.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

Reviewer #2: Partly

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: I Don't Know

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Summary and impression:

This was a retrospective cohort study PLWH who started on ART in the periods of April to August 2016 (delayed treatment group) compared to those in periods of April to August 2017 (test and treat group), each followed up for 24 months. Primary outcomes were death or loss to follow up.

The study enrolled from 3 coastal counties in Kenya and had a fair sample size, with the two comparison groups enrolled about a year apart; before and after local institutionalizing the WHO recommendation of test and treat in 2017. The participants for the cohorts were obtained from records in ART registers to identify those started on ART (470 enrolled) before and after (316 enrolled) Test and treat initiated. Primary endpoints were lost to follow up, viral suppression and death. Sample size well justified by earlier studies and adequate in this study, with total of 786 patients for analysis.

The main findings showed no difference in retention rate, viral load suppression rate and death in first 3 months and after 2 years of follow up. Although no difference in poor outcome between groups, individual factors were identified that were associated with increased risk of poor outcome in the test and treat group when compared to the delayed treatment. Higher risk with younger age, male gender, not employed and divorced status. No difference with the varied health care workers involved with care, only finding regular facility staff meetings associated did not seem to make sense. The overall conclusions were two-fold; retention patterns for the “test and treat” cohort was comparable to those who started ART before “test and treat”. Patients who are males, young, divorced/separated had higher risk of poor outcome in the test and treat group compared to the delayed treatment group.

Overall, the study addresses an important topic that can inform regional efforts on increasing ART adherence and HIV care retention in the test and treat era. However, the study is not novel and findings are not new; many previously studies have been done in other African countries.

The positives of this cohort study include the fact that this was real-life data as test and treat rolled out in the region. The comparison groups had fairly comparable distribution to further support the validity of these comparison groups. They were also well distributed among the counties in this region of Kenya, therefore, can make generalization in that entire region of Kenya and not just a particular township.

The significance of the finding is clear from this study and in the paper - that among HIV regardless of those with early treat or delayed, that poor outcome is associated with lack of support, less of economical independence, male gender and younger age. These are high risk groups which has and continues to need particular attention and focus in implementation research and public health efforts to increase adherence and retention in order to decrease risk of poor outcome and death. Tailoring efforts on these high risk groups of poor outcome is an important public health message.

However, the study is not novel nor innovative in its question and design. Several studies in African have looked at retention rate based on ART vs. test and treat already and stated in the paper so the presented data is not adding anything new to the existing body of literature on this topic, except that it confirms the high risk groups in this particular Kenyan region/country. There are some major and few minor improvements suggested in this current form.

1. Because the data is from an ART registry, it would be possible that those who enroll in the latter group (test and treat) could have been started on ART earlier, had some increase in CD4 and then stopped at some point before restarting again in this region where registry is located. These patients which I imagine from the overall low CD4 count of the groups would not be rare occurrences, but would be incorrectly categorizes as treatment naïve and also placed in the test and treat group at the later time period. But in fact, these patients should technically in the delayed group. The authors should attempt to determine how many participants had this scenario, because this would substantially bias the result and minimize the beneficial effects in the true test and treat. This would also explain why the test and treat group had lower CD4 count below 500 when one would expect that the CD4 T cell level be higher that the delayed treatment group as they are given the option of treating earlier before reaching CD4 <500.

2. Introduction focused on the merit and efficacy of test and treat in reducing transmission and severe disease. But this study is really focused on the topic of retention on ART and care, so authors should consider reviewing the literature on this specific topic; touching on the feasibility and barriers in implement and doing test and treat strategy.

3. Retention could have been affected by the ART regimen and the two groups differ in regimen. EFV may overall be better tolerated but could cause early discontinuation in some due to neurocognitive side effects and similar for NVP, with its own side effect profile. Would be of value to parse out if EFV vs NVP had any effect on retention.

4. I would expect that CD4 be lower in the delayed group but the CD4 is in fact lower in the test and treat group. This lower CD4 level, however, is consistent with the higher rate of OIs and the similar rate of poor outcome occurred on the day of starting ARVs in both groups (22% vs 23%). It seems like the study did not show difference in poor outcome because immunologically the patients were in similar states and similar CD4 count, albeit the test and treat was not significantly lower. It is unclear from the study design and warrant explanation in the discussion on why the test and treat had such low CD4 count.

5. It would be helpful to include P values in Table 1 with patient characteristics or mention in text to highlight any differences. From review of the data it does not seem like the two groups had any significant differences except maybe for the lower BMI categories (could be indicative of their state of health at enrollment) and maybe those who received NVP (could be factor in differences in retention rate).

Reviewer #2: The manuscript does not present any new data or information, as it has been shown from numerous studies that delaying HIV treatment leads to poor outcomes and countries worldwide have adapted the ‘Test and Treat’ strategy for treatment.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

25 May 2022

Reviewer 1.

Reviewer #1: Summary and impression:

This was a retrospective cohort study PLWH who started on ART in the periods of April to August 2016 (delayed treatment group) compared to those in periods of April to August 2017 (test and treat group), each followed up for 24 months. Primary outcomes were death or loss to follow up.

The study enrolled from 3 coastal counties in Kenya and had a fair sample size, with the two comparison groups enrolled about a year apart; before and after local institutionalizing the WHO recommendation of test and treat in 2017. The participants for the cohorts were obtained from records in ART registers to identify those started on ART (470 enrolled) before and after (316 enrolled) Test and treat initiated. Primary endpoints were lost to follow up, viral suppression and death. Sample size well justified by earlier studies and adequate in this study, with total of 786 patients for analysis.

The main findings showed no difference in retention rate, viral load suppression rate and death in first 3 months and after 2 years of follow up. Although no difference in poor outcome between groups, individual factors were identified that were associated with increased risk of poor outcome in the test and treat group when compared to the delayed treatment. Higher risk with younger age, male gender, not employed and divorced status. No difference with the varied health care workers involved with care, only finding regular facility staff meetings associated did not seem to make sense.

Response:

Having regular staff meetings was taken to symbolize that the team had some level of teamwork and good leadership which could influence service delivery.

The overall conclusions were two-fold; retention patterns for the “test and treat” cohort was comparable to those who started ART before “test and treat”. Patients who are males, young, divorced/separated had higher risk of poor outcome in the test and treat group compared to the delayed treatment group.

Overall, the study addresses an important topic that can inform regional efforts on increasing ART adherence and HIV care retention in the test and treat era. However, the study is not novel and findings are not new; many previously studies have been done in other African countries.

The positives of this cohort study include the fact that this was real-life data as test and treat rolled out in the region. The comparison groups had fairly comparable distribution to further support the validity of these comparison groups. They were also well distributed among the counties in this region of Kenya, therefore, can make generalization in that entire region of Kenya and not just a particular township.

The significance of the finding is clear from this study and in the paper - that among HIV regardless of those with early treat or delayed, that poor outcome is associated with lack of support, less of economical independence, male gender and younger age. These are high risk groups which has and continues to need particular attention and focus in implementation research and public health efforts to increase adherence and retention in order to decrease risk of poor outcome and death. Tailoring efforts on these high-risk groups of poor outcome is an important public health message.

However, the study is not novel nor innovative in its question and design. Several studies in African have looked at retention rate based on ART vs. test and treat already and stated in the paper so the presented data is not adding anything new to the existing body of literature on this topic, except that it confirms the high-risk groups in this particular Kenyan region/country. There are some major and few minor improvements suggested in this current form.

1. Because the data is from an ART registry, it would be possible that those who enroll in the latter group (test and treat) could have been started on ART earlier, had some increase in CD4 and then stopped at some point before restarting again in this region where registry is located. These patients which I imagine from the overall low CD4 count of the groups would not be rare occurrences, but would be incorrectly categorizes as treatment naïve and also placed in the test and treat group at the later time period. But in fact, these patients should technically in the delayed group. The authors should attempt to determine how many participants had this scenario, because this would substantially bias the result and minimize the beneficial effects in the true test and treat. This would also explain why the test and treat group had lower CD4 count below 500 when one would expect that the CD4 T cell level be higher that the delayed treatment group as they are given the option of treating earlier before reaching CD4 <500.

Response:

Thanks for highlighting this important observation. We extracted data from the patients’ files, we had no opportunity to interview the patients to get some of the data being requested. We agree with the reviewer this is a plausible scenario. However, the test and treat policy does not anticipate such scenario and to be pragmatic, the ART initiator relies on reported information from the patient. It may be challenging to establish a case restarting ARVs unless reported by the patient. Such patient would then not be treated as ART naïve because their date of HIV diagnosis would be recorded. We have acknowledged this as a study limitation in the second last paragraph of discussion section.

We have checked the dates of HIV diagnosis versus the date of starting ARTs to establish those who started ARTs within the seven days of diagnosis. We have added the days to starting ARTs on Table 1. Additionally, we conducted a sensitivity analysis including only the test & treat patients who started ARTs within seven days (74%). The results are shown on the last paragraph of results section and supplementary Table 1 below.

Supplementary table 1. Cumulative retention rates from month 3 to 24 for “test and treat” and before “test and treat” cohorts.

Month Test & start cohort (n=233) Before test and treat cohort (n=470) Absolute difference (95% CI) P-value*

3 224 (87) 375 (84) 3.3 (-2.2 to 8.9) 0.25

6 172 (79) 358 (81) -2.3 (-0.9 to 4.3) 0.49

9 168 (77) 340 (77) -0.4 (-7.2 to 6.4) 0.91

12 162 (74) 330 (75) -0.9 (-7.3 to 5.5) 0.79

15 151 (69) 316 (72) -2.9 (-9.6 to 3.8) 0.40

18 145 (67) 312 (71) -4.1 (-10.9 to 2.7) 0.24

21 137 (66) 305 (70) -4.6 (-12.3 to 3.2) 0.24

24 131 (64) 295 (68) -4.9 (-12.8 to 3.0) 0.22

In the test & treat policy, there is no systematic testing for CD4 counts because they are not needed to decide when to start ARTs as it used to be in the delayed cohort. Therefore, very few patients in the test & treat cohort have CD4 (~23%). These would be likely the very sick ones that a clinician feel the need to know their CD4 counts. It is highly likely the 23% CD4 counts we reported in this cohort is biased. We have checked if the CD4 counts were any different across the WHO stages stratified by the two cohorts and added these in the results section (figure 1).

Figure 1: Comparing CD4 levels at starting ART for delayed and test and start cohorts

2. Introduction focused on the merit and efficacy of test and treat in reducing transmission and severe disease. But this study is really focused on the topic of retention on ART and care, so authors should consider reviewing the literature on this specific topic; touching on the feasibility and barriers in implement and doing test and treat strategy.

Response:

Although studies have demonstrated overwhelming evidence of the benefit of “Test and treat” on reducing HIV transmission(7), evidence on other outcomes like attrition from care have been inconsistence. The retention rate reported in a pooled analysis of 154 cohorts in Low- and Middle-Income countries was approximately 70%, two years after starting ART (2008 to 2013)(12). Recent data after the universal test and treat policy suggest the retention has not improved, in Democratic Republic of Congo, the retention rate was 77% after two years of starting ART following the adoption of test and treat policy(13). The same study found higher rate of attrition after the adoption of the universal test and treat policy. Another cohort in Masaka, Uganda, found PLHIV starting ARTs within seven days of HIV diagnosis had higher risk of lost to follow-up(14). Systematic reviews report varying interventions to improve retention on care but with mixed outcomes(15). Various studies have also reported different predictors of attrition for different follow-up times and in different settings, some before and others after the universal test and treat policy(13,14,16).

3. Retention could have been affected by the ART regimen and the two groups differ in regimen. EFV may overall be better tolerated but could cause early discontinuation in some due to neurocognitive side effects and similar for NVP, with its own side effect profile. Would be of value to parse out if EFV vs NVP had any effect on retention.

Response:

Among the PLHIV that we sampled, 95% (301/316) of the test and treat cohort and 92% (431/470) of the delayed treatment cohort were on EFV based regimen. We have formally explored the effects of the ART regimen on attrition in the multivariable regression model reported on Table 3. We found no evidence that attrition was affected by ART regimen. Even on Table 1 where we tested the differences using chi-square, the P-value indicates some borderline effect (P=0.05) which attenuated in the multivariable model.

4. I would expect that CD4 be lower in the delayed group but the CD4 is in fact lower in the test and treat group. This lower CD4 level, however, is consistent with the higher rate of OIs and the similar rate of poor outcome occurred on the day of starting ARVs in both groups (22% vs 23%). It seems like the study did not show difference in poor outcome because immunologically the patients were in similar states and similar CD4 count, albeit the test and treat was not significantly lower. It is unclear from the study design and warrant explanation in the discussion on why the test and treat had such low CD4 count.

Response:

Following the test & treat policy, initiation of ARVs does not need a CD4 threshold. Therefore, most of these patients were not tested for CD4 counts (we have data for only 23%). It is likely the reported CD4 count among the test & treat could be systematically different from those not tested. However, we have compared the overall CD4 counts and across the WHO stages and found no evidence of differences between the test & treat versus the delayed cohorts (Table 1 and Figure 1). This is a general limitation of a study using data extracted from patient files. We have acknowledged this limitation on the second paragraph of the discussion section.

We don’t have immunologically data to assess their differences between the two groups at baseline and during the follow-ups. Such data would be very useful to monitor the long-term effect of test & treat policy. We have mentioned this limitation and suggested future studies including ART program should systematically be collecting this data like the viral load.

5. It would be helpful to include P values in Table 1 with patient characteristics or mention in text to highlight any differences. From review of the data it does not seem like the two groups had any significant differences except maybe for the lower BMI categories (could be indicative of their state of health at enrollment) and maybe those who received NVP (could be factor in differences in retention rate).

Response:

Thanks for pointing out this. We have updated Table 1 by adding the P-values, adding some footnotes to explain some variables especially the CD4 counts and added a new variable; time to starting ARVs from day of HIV diagnosis. The new variable highlights the difference in days to starting ARVs as would be expected.

Reviewer 2

Reviewer #2: The manuscript does not present any new data or information, as it has been shown from numerous studies that delaying HIV treatment leads to poor outcomes and countries worldwide have adapted the ‘Test and Treat’ strategy for treatment.

Response:

We thank the reviewer for taking time to review our manuscript. We concur the test & treat policy has been rolled out worldwide. Our study is exploring the effectiveness of the test & treat post rolling out in a pragmatic setting. Although some data does suggest the policy is effectively, there are studies reporting the opposite. For example. our study did not find any differences in retention between the test & treat versus the delayed cohort. A more recent study in Masaka Uganda, reported higher lost to follow-ups among the test & treat compared to the delayed cohort (https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0217606 ). There are also fears that asymptomatic patients with higher CD4 cell counts would have poor adherence to ART so that early initiation would results in loss to follow up (https://pubmed.ncbi.nlm.nih.gov/26148280/ ). It is also likely widespread use of antiretroviral treatment at a population and individual level may lead to development of drug resistance.

There is thus need for more data from different settings assessing the long-term effect of the test & treat policy. Our study present data from different settings (to our best level of knowledge, no such study has been conducted across the three coastal counties of Kenya) and followed the PLHIV for a period of two years post ART initiation. Our findings that the retention was not any different between the two cohorts and that even the test & treat patients had attrition of more than one third after two years warrant more studies to understand the changing trends.

Submitted filename: Response to reviewers.docx

Click here for additional data file.

15 Jun 2022

Factors associated with poor outcomes among people living with HIV started on anti-retroviral therapy before and after implementation of “Test and treat” program in Coastal Kenya

PONE-D-21-23151R1

Dear Dr. Mwamuye,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Manish Sagar, MD

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

Reviewer #2: (No Response)

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: (No Response)

Reviewer #2: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: There were 5 recommendations for revisions and the authors have addressed them all. A few were beyond the scope of the data set reviewed but authors addressed these limitations in more detail. The study although not novel does add to the body of evidence showing the effectiveness of ‘test and Treat’ strategy and additional provides data on poor retention of this method, which seems to be a major concern. The effectiveness of this strategy does depend on the region of the world, the population and the socioeconimic status so this data shows data from this coastal area of Kenya.

Reviewer #2: The authors have addressed the comments from the reviewers and revised the manuscript accordingly.

The authors have amended the protocol and in the revised version they have:

- added in paragraph in introduction to discuss more about the background and HIV test and treat versus delayed treatment

- Updated Table 1 (with p-values)

- Updated the discussion on the study importance and value.

Hence, this revised draft of the manuscript appears to be scientifically suitable for publication.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

**********

8 Sep 2022

PONE-D-21-23151R1

Factors associated with poor outcomes among people living with HIV started on anti-retroviral therapy before and after implementation of “Test and treat” program in Coastal Kenya

Dear Dr. Mwamuye:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Manish Sagar

Academic Editor

PLOS ONE