Nurse-led HIV services and quality of care at health facilities in Kenya, 2014–2016

Abstract Objective To develop a novel measure to characterize human immunodeficiency virus (HIV) programme quality at health facilities in Kenya and explore its associations with patient- and facility-level characteristics. Methods We developed a composite indicator to measure quality of HIV care, comprising: assessment of eligibility for antiretroviral therapy (ART); initiation of ART; and retention on ART or in care, if ineligible for ART, for 12 months. We applied the comprehensive retention indicator to routinely collected clinical data from 13 331 patients enrolled in HIV care and treatment at 63 health facilities in the Eastern and Nyanza regions of Kenya from 1 January 2014 to 31 March 2016. We explored the association between facility- and patient-level characteristics and the primary outcome: appropriate staging and management of HIV, and retention in care over 12 months. Findings Of the enrolled patients, 8404 (63%) achieved comprehensive retention 12 months after enrolment in care. In univariate analyses, patients at facilities where nurses delivered HIV treatment services (including eligibility assessment, initiation and follow up of ART) had significantly higher comprehensive retention rates at 12 months. In multivariate analyses, after adjusting for both facility- and patient-level characteristics, patients at facilities where nurses initiated ART had significantly higher comprehensive retention in care at 12 months (relative risk, RR: 1.22; 95% confidence interval, CI: 1.00–1.48). Conclusion Nurse-led HIV services were significantly associated with quality of care, confirming the central role of nurses in the achievement of global health goals, and the need for further investment in nursing education, training and mentoring.


Introduction
The global response to the human immunodeficiency virus (HIV) epidemic has been remarkably successful. In low-and middle-income countries, the number of people living with HIV who have access to antiretroviral treatment (ART) rose from 400 000 in 2003 to 17 million in 2015. Modelling estimates from the Joint United Nations Programme on HIV/ AIDS (UNAIDS) suggest that annual deaths from the acquired immunodeficiency syndrome (AIDS) have dropped by 43% over the same period 1 and 7.8 million deaths have been averted by the scale-up of ART services. 2 Increased access to prevention and treatment has also led to an estimated 35% drop in new HIV infections since 2000, including a 58% decrease among children. 3 Despite these achievements, UNAIDS modelling highlights that only 46% of people living with HIV have initiated ART and 2.1 million new HIV infections occurred in 2015. 1 Scaling up HIV prevention and treatment services and ensuring their quality is important if the ambitious global goals are to be met. 4 The 2020 UNAIDS 90-90-90 targets call for 90% of people living with HIV to be aware of their HIV status, 90% of those diagnosed with HIV to be on ART and 90% of those on ART to achieve viral suppression. 5 To reach these targets, people living with HIV need to learn their HIV status, link to appropriate treatment, achieve virological suppression and remain in care and on treatment for their lifetime. Careful attention to each step along the treatment continuum is essential, 6 as sustained retention in care remains a challenge in many HIV programmes. 7 HIV programme quality is assessed via diverse process indicators in both low-and high-resource settings. 8 For example, the Site Improvement through Monitoring System approach used by the United States President's Emergency Plan for AIDS Relief (PEPFAR) assesses core essential elements for quality service delivery at the community, clinic (site) and above-site levels. HIV programme quality is often also assessed by examining outcomes along the continuum of HIV care, including: the proportion of patients testing positive for HIV who are linked to care; the proportion started on treatment; the proportion who remain in care; and the proportion who achieve viral suppression (in settings where viral load measurement is available). Because this approach only includes patients eligible for ART, more inclusive measures of the continuum of HIV care have been proposed that capture outcomes for all HIV-infected individuals enrolled in care. 9 We used a novel and comprehensive measure of comprehensive retention in HIV care to characterize programme quality at health facilities in Kenya, and explored the associations with patient-and facility-level characteristics.

Study design and setting
We conducted an observational, longitudinal analysis using routinely collected clinical information on 13 331 HIV-infected individuals enrolled at 63 health facilities in the Eastern and Nyanza regions of Kenya. Health facility staff provided clinical services according to Kenya's national guidelines. 10 All facilities were receiving technical assistance from the International Objective To develop a novel measure to characterize human immunodeficiency virus (HIV) programme quality at health facilities in Kenya and explore its associations with patient-and facility-level characteristics. Methods We developed a composite indicator to measure quality of HIV care, comprising: assessment of eligibility for antiretroviral therapy (ART); initiation of ART; and retention on ART or in care, if ineligible for ART, for 12 months. We applied the comprehensive retention indicator to routinely collected clinical data from 13 331 patients enrolled in HIV care and treatment at 63 health facilities in the Eastern and Nyanza regions of Kenya from 1 January 2014 to 31 March 2016. We explored the association between facility-and patient-level characteristics and the primary outcome: appropriate staging and management of HIV, and retention in care over 12 months. Findings Of the enrolled patients, 8404 (63%) achieved comprehensive retention 12 months after enrolment in care. In univariate analyses, patients at facilities where nurses delivered HIV treatment services (including eligibility assessment, initiation and follow up of ART) had significantly higher comprehensive retention rates at 12 months. In multivariate analyses, after adjusting for both facility-and patient-level characteristics, patients at facilities where nurses initiated ART had significantly higher comprehensive retention in care at 12 months (relative risk, RR: 1.22; 95% confidence interval, CI: 1.00-1.48). Conclusion Nurse-led HIV services were significantly associated with quality of care, confirming the central role of nurses in the achievement of global health goals, and the need for further investment in nursing education, training and mentoring.
HIV programme quality, Kenya Miriam Rabkin et al.
Center for AIDS Care and Treatment Programs (ICAP) at Columbia University via funding from PEPFAR; this included in-service training using Kenya's national training curricula, supportive supervision, and mentoring on routine data collection and analysis. For this analysis, we assessed the relationship between selected patient and facility parameters and the outcome indicator: comprehensive retention in HIV care.

Patient-level data
The Optimal Models of HIV care project is a multi-country project, funded by the Centers for Disease Control and Prevention, aimed at improving the use of routinely-collected HIV care and treatment data for programme monitoring, evaluation, operations research and implementation science. As part of this project, we collect de-identified patientlevel data from ICAP-supported health facilities in Ethiopia, Kenya, Mozambique, Rwanda and the United Republic of Tanzania. Details of the methods are described elsewhere. 11 Briefly, at each patient visit, clinic staff document routinely collected patient information on national forms. This information is regularly entered into electronic databases by data clerks, with data quality assessments performed at least annually at each clinic. This analysis focused on Kenya, chosen because of the variability in nurse-led ART services in the country and to restrict between-country variability as an explanation for any findings.

Facility-level data
ICAP staff collected facility-level characteristics in August 2015 using ICAP's Program and Facilities Characteristics Tracking System site checklist. This is a biannual site assessment of the HIV care and treatment services and staffing at health facilities supported by ICAP.
Facility-level factors included: facility type (public primary, public secondary or private); setting (urban, semi-urban or rural); availability of onsite CD4+ cell testing; patient volume (categorized based on number of new enrollees into HIV care per year); availability of specific services (community ART adherence groups; financial incentives for adherence; facility-based group counselling for adherence support); and types of health-care worker at the facility (physicians, nurses, clinical officers [non-physician clinicians] or outreach workers). In addition, the survey assessed the availability of nurse-led ART services (assessment of eligibility for ART; initial prescriptions for ART; and follow-up care for ART).

Outcome definition
The primary outcome of interest in this analysis was a comprehensive measure combining retention in HIV care for 12 months and receipt of key services during that time. The comprehensive retention indicator included variables relevant to patients who were eligible for ART as well as those who were not yet eligible for ART; all patients received clinical and immunological monitoring, prophylaxis for opportunistic infections and counselling.
Patients who started ART within 12 months of enrolment met the definition of comprehensive retention in HIV care if they were assessed for ART eligibility; initiated on ART or retained in care until becoming eligible for ART (based on country guidelines), 10 and retained on ART for 12 months after enrolment at the health facility. Patients also met the definition of comprehensive retention in care if they were assessed for ART eligibility; found to be ineligible based on country guidelines; and retained in care at the health facility for 12 months after enrolment. Evidence of assessment for eligibility was a record of CD4+ cell count or World Health Organization (WHO) HIV/AIDS clinical stage 12 before ART initiation.
Patients did not meet the definition of comprehensive retention if they were missing a recorded WHO stage or CD4+ cell count in their medical records; designated as ART-eligible (according to CD4+ cell count or WHO stage); and did not initiate ART. Patients not in comprehensive retention also included those who died or became lost to followup in the 12 months after enrolment.
For all analyses, patients were considered lost to follow-up if they had no recorded visit in the last 3 months (patients on ART) or 6 months (patients not on ART).

Statistical methods
Descriptive statistics were used to describe the demographic and clinical characteristics of the study population. In analyses investigating factors associated with achievement of retention in care at 12 months (as defined above), univariate analyses were conducted using generalized linear mixed log-binomial regression models with random intercepts to account for within-clinic correlation. Outcomes are presented as the relative risk (RR) and 95% confidence interval (CI) of achieving 12-month retention in HIV care.
For the multivariable analyses, two models were examined. The first model included only facility-level covariates, while the second model included facility-level and patient-level covariates. For models 1 and 2, facility type was included because it was a priori identified as a potential confounder of the relationship between the availability of nurse-led initiation/management of ART and retention at 12 months. Other facility-level covariates found significant at α < 0.05 in the univariate analysis were also included in the multivariable models. The three measures focusing on nurse-led initiation and management of ART (i.e. assessment of ART eligibility by nurse; initiation of ART by nurse; and follow-up of ART by nurse) were highly collinear with each other; therefore only one (initiation of ART by nurse) was included in the multivariable models. The latter was selected because it had the strongest association with our outcome of interest in the univariate analysis. In model 2, patient-level covariates included age, sex and disease stage at enrolment in care.

Ethical approval
Use of anonymized patient-level data from health facilities was conducted as part of the Identifying Optimal Models of HIV Care and Treatment study. All data were de-identified before analysis and the investigators had no access to identifiable patient information. Institutional review board approval was  Table 1 presents the facility-level and patient-level characteristics of the population. The majority of facilities in this study were public primary health-care centres (33 facilities, 52%), treating 5856 (44%) patients. Nearly all facilities were situated in urban or semi-urban settings. CD4+ cell testing was performed on-site in 18 facilities (29%) where 5661 patients were enrolled (42% of the patient population). For the remaining facilities and patients, blood drawn at the facility was transported off-site for CD4+ cell testing. Community ART adherence groups were available at 32 facilities (51%), financial incentives for ART adherence at 5 facilities (8%) and group counselling at 61 facilities (97%). At 5 facilities (8%) of physicians were available in addition to nurses and other health-care workers, providing care to 1436 patients (11%); other facilities were staffed only by nurses and clinical officers. Outreach workers were available at 29 facilities (46%). Most patients were enrolled at facilities where nurses assessed patients for ART eligibility (11 996, 90%), initiated ART prescriptions (11 092, 83%) and gave follow-up care to patients on ART (10 738, 81%).

Results
Of the patient-level characteristics investigated, median age at enrolment was 30.2 years (interquartile range, IQR: 23.7-38. 5 In contrast, patients at facilities with a physician were less likely to achieve 12-month retention than those at facilities without a physician (RR: 0.75; 95% CI: 0.57-0.98). In addition, patients at facilities with the smallest new patient enrolment had significantly lower 12-month retention in comprehensive care (RR: 0.78; 95% CI: 0.65-0.93) compared with sites with the medium level of enrolment. Lastly, 12-month retention was significantly lower among patients with more advanced HIV disease (RR: 0.75; 95% CI: 0.73-0.78) and those missing CD4+ cell count or WHO staging (RR: 0.67; 95% CI: 0.64-0.71). Table 3 shows the findings based on multivariate analyses. In model 1 (adjusting only for facility-level characteristics), patients at facilities where nurses initiated ART were significantly more likely to be retained in care for 12 months (RR: 1.25; 95% CI: 1.01-1.54). No other facility-level characteristics remained statistically significant in this model. In model 2, in which we adjusted for both facility-and patientlevel characteristics, 12-month retention was more likely for patients attending facilities with nurse-initiated ART (RR: 1.22; 95% CI : 1.00-1.48) and less likely among patients having advanced HIV disease at enrolment (RR: 0.65; 95% CI: 0.60-0.70).

Discussion
In our analysis of 13 331 HIV-infected patients enrolled in HIV care in Kenya, we found that the majority of clinical services were provided by nurses and clinical officers. Nurse-led ART services were associated with significantly higher comprehensive retention in HIV care and treatment and this was the only sitelevel characteristic significantly associated with retention in the multivariate analysis. In addition, when adjusting for both facility-and patient-level characteristics, nurse-led ART services remained significantly associated with higher rates of comprehensive retention of patients in HIV care.
Previously published studies have shown that nurse-managed ART does not compromise, and may improve, the quality of care for people living with HIV. In a systematic review of task-shifting, eight studies showed that nurse-managed HIV services in low-resource settings had outcomes similar to physician-led teams. 13 Two randomized studies from South Africa also support the effectiveness of nurse-managed ART. A study at two health facilities near Cape Town and Johannesburg demonstrated that patients randomized to ART initiated by a physician and then monitored by a primary-care nurse had outcomes that were not inferior to patients randomized to physician-managed treatment. 14 In the other study, a cluster-randomized trial showed equivalence of nurseinitiated versus physician-initiated ART at 31 health facilities in the Free State province of South Africa. 15 In comparison to the previously published literature, our study is distinguished by its use of data from a larger number of health facilities, inclusion of both patient-and facility-level characteristics, the use of a novel outcome measure of comprehensive retention and the use of routine programmatic data.
Our findings also highlight the scarcity of physicians at HIV care settings in Kenya, even in urban and peri-urban areas. Only 8% of facilities had physician-led care, and only 11% of patients were being managed by a medical doctor. This is consistent with the situation in most low-resource settings where nurses, clinical officers and other non-physician clinicians provide the majority of health-care services. [16][17][18] In sub-Saharan Africa, where the burden of HIV is the highest, nurses are the backbone of the health system. 19,20 The strengths of this study include its novel measurement of quality of care, which combines a measure of retention    and retained on ART for 12 months after enrolment, or retained in care for 12 months if ART ineligible. b Model 1 included only facility-level covariates. Model 2 included facility-level and patient-level covariates. Facility type was a priori identified as a potential confounder of the relationship between availability of ART initiation and management by a nurse and retention at 12 months. Other facility-level covariates found significant at α < 0.05 in the univariate model were included in the multivariable models. All three patient-level covariates were included in the models. c Relative risks of achieving 12-month retention in care were calculated using generalized linear mixed log-binomial relative risk regression with random intercepts to account for within-clinic correlation. d Because eligibility assessment by nurse, initiation of ART by nurse and follow-up of ART by nurse were highly collinear with each other, only nurse initiation of ART by nurse was included in the models. in care with indicators of appropriate patient monitoring and initiation of treatment. Other strengths include the large number of patients and health facilities, and the inclusion of both patient-and facility-level parameters. The limitations include the non-random sample of health facilities, which were also limited to one country. While we adjusted for multiple patient and facility-level characteristics, the findings could have been affected by other parameters that we did not measure in this non-randomized study. The study focused on retention of patients and provision of key services; other important quality parameters that were not routinely collected at the time, such as achievement of viral suppression, patient satisfaction and quality of life, were not assessed and require further research.
In conclusion, using a comprehensive multi-component measure of quality of care, we demonstrated that pa-tients at facilities offering nurse-led ART services were more likely to be retained in effective HIV care and treatment at 12 months after adjusting for facility-and patient-level characteristics. These findings confirm the central role of nurses in the achievement of global targets, including the 90-90-90 goals, and the need for further investment in nursing education, training and mentoring. ■ meses. En análisis multivariables, tras ajustar las características a nivel de pacientes y de centros, los pacientes de los centros donde las enfermeras iniciaron el TAR registraron unas tasas de retención completa mucho más altas en 12 meses (riesgo relativo (RR): 1,22; intervalo de confianza (IC) del 95%: 1,00-1,48).
Conclusión Los servicios contra el VIH dirigidos por enfermeras estaban significativamente asociados con la calidad de la atención, lo que confirma el importante papel de las enfermeras a la hora de lograr objetivos de salud globales, así como la necesidad de invertir más en la educación, formación y orientación de las enfermeras.