Bulletin of the World Health Organization

Optimizing paediatric HIV care in Kenya: challenges in early infant diagnosis

Peter Cherutich a, Irene Inwani b, Ruth Nduati c, Dorothy Mbori-Ngacha d


Kenya’s population of 30 million people has nearly 1.2 million people living with HIV/AIDS with at least 250 000 in need of immediate antiretroviral therapy (ART). In 2003, the Kenyan Ministry of Health started a programme to rapidly scale up ART with the goal of progressively delivering effective ART, according to standardized regimes, to HIV-infected people in need of treatment, reaching 50% by the end of 2005 and 80% by the end of 2008.1 As a medium-term goal, the Kenyan government adopted WHO’s “3 by 5” initiative, which aimed at providing 95 000 people with ART by end of 2005, of whom 10% were to be children. By July 2005, 45 000 adults and more than 2000 children were on treatment, most of them older children. The obvious inequitable access to HIV treatment is of major concern considering that 90% of the HIV infections are due to mother-to-child transmission of HIV and without treatment, half of the infected children do not live to celebrate their second birthday and therefore need to be identified as early as possible to benefit from the newly availed antiretroviral (ARV) therapies.2

There was therefore a need to conduct a survey to identify the gaps existing within the current delivery of paediatric HIV care. This paper specifically reports systems and structural barriers to early identification and staging of children exposed to, or infected with, HIV.


The existing Kenyan Ministry of Health policies on HIV diagnosis in children were reviewed. Institutions providing ARV were identified using the national ART register in seven of the eight provinces, and a convenient sample of 58 government and mission facilities/nongovernmental organizations providing ART to children was chosen, where a rapid assessment was conducted in the months of August and September 2005. A standard tool of structured questions and checklists was used to obtain information. In addition, qualitative information was obtained through face-to-face and/or telephone interviews. The key respondents were medical superintendents and matrons in charge of the facilities; physician(s) and heads of pharmacy, laboratory, mother and child health clinics, as well as ART clinic personnel. All possible entry points for paediatric HIV care were assessed and included maternal and child health clinics, paediatric inpatient wards and outpatient services as well as the comprehensive HIV clinics. Voluntary counselling and testing centres located within the 58 facilities were also surveyed.

The African Network for the Care of Children Affected by AIDS (ANECCA) 10-point package for comprehensive paediatric HIV care was used as the benchmark for monitoring and evaluating sites providing ART to children.3 We also used the integrated management of childhood illnesses (IMCI) chart to asses the facilities’ capacity to screen and stage HIV-infected children. However, this paper focuses only on barriers to the early confirmation of HIV status and staging of HIV-infected children.


Policy environment

In 2003, the National AIDS/STD Control Programme (NASCOP) developed a five-year strategic plan to guide the scaling-up of ART. A review of this strategic document showed that identification of HIV-infected/exposed children was not specifically provided for and the models of care adopted by the document precluded maternal and child health services as entry points for care for HIV-infected children. In 2004, NASCOP developed guidelines for HIV testing in clinical settings.4 These guidelines provided for universal HIV testing in all paediatric wards and for children infected with, or suspected to have, tuberculosis.

Health facility survey

A total of 58 facilities were identified to be providing paediatric HIV services. These facilities were broadly classified as government, faith-based and nongovernmental organizations, and included two national referral hospitals, 11 district hospitals, six health centres sponsored by the government, 12 mission hospitals, 18 mission health centres and four children orphanages. In the 42 (72%) that provided information on number of children in care, 6639 children were being followed up in the paediatric HIV clinics and 1988 (30%) were on ART. There were only seven sites having more than 100 children on ART on follow-up.

Identification of HIV-infected/exposed children
Follow-up of mother–infant pairs in the maternal and child health clinics

Only six facilities had well established systems of referral and follow-up of HIV-exposed infants identified in the prevention of mother-to-child transmission of HIV (PMCT) programme. None of the facilities had a mother–child hand-held clinical record that clearly identified HIV-exposed or infected children. The existing mother’s cards had a slot for antenatal HIV status of mother but this information was rarely transferred to the child’s immunization/growth monitoring card. In most facilities, no attempt was made to ask the mother her HIV status or to check the antenatal clinic card that might indicate the child’s HIV status.

Testing of children

We did not find any copies of the national guidelines for HIV testing in clinical settings in these facilities and none of the facilities had in-house policy guidelines on HIV counselling and testing for children. Most health caregivers were not clear on testing guidelines for children. However, many recognized the inadequacy of the voluntary counselling and testing guidelines in addressing HIV testing in children.5

In almost all the facilities assessed, children attending child welfare clinics were not routinely assessed using the IMCI criteria, and HIV testing was only offered on rare request by the clinicians.

A large number (31%) of the 58 health facilities did not provide on-site HIV testing to children. Children were not routinely tested in the voluntary counselling and testing centres located within the facilities, in keeping with the existing policy that the individual who is being tested must be legally competent to give informed consent. Children brought by parents or guardians were referred to the other clinics for diagnostic testing where testing was done at a cost denying many access to this service. In all the facilities, contrary to the 2004 policy, there was no routine testing of children whose parents or siblings were identified as HIV infected or those in high-risk clinics, for example in tuberculosis clinics.

Only four facilities were capable of performing a diagnostic deoxyribonucleic acid (DNA) polymerase chain reaction (PCR) on children under 18 months, and even then only as research or pilot studies. None carried out routine processing of HIV DNA samples for routine clinical care of HIV-infected children. We did not find any facility that had staff trained in dry blood spot techniques of specimen collection. All other facilities had antibody-based tests for HIV diagnosis in children. None of the facilities assessed had an algorithm for early infant diagnosis.

Staging of HIV-infected children

None of the facilities had job aids to guide health-care workers on appropriate clinical staging for HIV-infected children. The facilities surveyed were able to stage HIV disease in children in areas where appropriately trained personnel were deployed; however, this was not done routinely. Clinical staging was best done at the paediatric wards. Five facilities did not have the capacity to perform basic tests required pre-ARV initiation, such as haemogram, renal and liver function tests. There were 26 facilities with various types of CD4 machines reflecting the different sources of support but reagents for the CD4 machines were either lacking or their supplies were erratic, therefore universal capacity for clinical and immunological staging of children was lacking.

Laboratory human resource and infrastructure
Human resource capacity

Except for four high level facilities that had university-trained laboratory personnel, all other laboratories visited were under the supervision of a diploma-level laboratory technologist. All laboratory staff were capable of performing at least one form of HIV testing.

All laboratories reported a lack of adequate personnel to cope with the HIV workload in addition to carrying out other routine laboratory assays.

Laboratory infrastructure

Most of the facilities had adequate laboratory space as well as HIV diagnosis and treatment-monitoring capacity. The most common HIV tests available were the Rapid Tests. Determine® and SD Bioline HIV-1/2 were the commonly used tests for HIV testing in children, while Uni-gold™ was used as the tiebreaker.

Referral system for specimens and return of results

Only one laboratory visited had systems in place for referring patients or transporting samples from one site to the other. For tests that were not performed within a specific site, patients were given laboratory request forms and sent to have the tests done elsewhere. Such patients were frequently referred to private hospitals where the cost of most of the tests was likely to be higher than in the public hospitals, therefore inaccessible to many children. Even where such tests were done there was no system for delivery or return of test results, and patients had to wait until their next appointment to have their test result reviewed. All specimens from patients had identification tags on tubes, which included patient name and inpatient/outpatient number and hence there was the potential for involuntary HIV disclosure as patients moved to different laboratories seeking services. Although all laboratory requests are supposed to be on a standard form, in most of the facilities this was being done on ordinary paper and/or improvised forms. Virtually each facility had their own mechanism of recording laboratory results. None of the laboratories was involved in an organized network of well equipped facilities that provided routine support for lower-level facilities.

Good laboratory practice

Some of the hospitals had made efforts to develop standard operating procedures (SOPs) for some of the laboratory investigations done. However, none of the hospital visited had any SOPs for preventive maintenance of the existing equipments.

Quality control procedures

Except for internal quality assurance (QA), which was being followed where necessary, there were no other quality control (QC) methods in place, except for two major referral hospitals. Even where internal QC was followed, there was no documentation seen during this survey. Where tests were carried out at different service provision points and by different cadre of staff, the main laboratory was given every 10th sample to retest without disclosure of initial results and both the results were later compared. One provincial hospital had an elaborate research laboratory that also provided essential QC for the service provision laboratory.

Data management

Only 80% of these facilities had records indicating the number of children tested for HIV. However, there was a wide discrepancy in facilities’ ability to report certain indicators. One district hospital’s database was only able to record completed years and therefore children under 18 months were generally categorized as being under two years. Therefore, any indicator capturing HIV testing at 18 months was meaningless.

The majority (72%) of the facilities used the recommended Ministry of Health registers to capture data, but the original registers were missing in most facilities forcing the health-care providers to improvise using hardcover black books or photocopied registers. Some registers were modified to accommodate the indicators required by partners supporting the sites. In summary, there was no uniform reporting system and the structures to support reporting, for example registers and computers, were insufficient.


There are several key observations from this study. The policy framework to support identification of HIV-infected children that had been established narrowly focused on sick children and did not address identification of HIV-infected children in PMCT settings. PMCT is a major entry point to care for HIV-exposed and infected children. Children of women who had accessed PMCT services were not able to benefit from co-trimoxazole prophylaxis, early HIV diagnosis and regular follow-up beyond the prescribed “well child” visits for immunization and growth monitoring. The existing tools were not supportive of a systematic follow-up of the children and policies were not adequately disseminated.

Both HIV antibody and DNA PCR testing capability were available in-country, even though the latter was not available for routine clinical care settings. The assessment identified the weaknesses in the structures and systems available and the lack of systematic diagnosis of HIV through laboratory or clinical-based algorithms. Existing policies excluded children from the free HIV-testing services offered through voluntary counselling and testing, and therefore HIV-antibody testing which was widely available could only be accessed at a cost for the children. There were serious lapses in management of the laboratories. The laboratories lacked systems for referring samples and instead referred patients but not necessarily to specific facilities. Preventive maintenance of the equipment and good laboratory practice were not in place in the majority of the laboratories.

IMCI guidelines and WHO HIV-staging clinical tools that required minimal laboratory testing were also available. However, these algorithms were rarely used and staff lacked basic job aids to support their use.


Several recommendations can be made from this study. The first is the need to revise the existing national HIV-testing policies to embrace testing in the early infancy as a pillar of early infant diagnosis. These policies need to be widely disseminated to ensure health providers are aware of the content. Second, there is a need to develop a clear standard system of follow-up of children identified as HIV-exposed through PMCT services. Issuing the mother–child welfare card during the initial antenatal visit or at first contact for those delivered at home, with a clear indication of HIV-exposure status, should facilitate this referral.

Health-care personnel should be trained to routinely use a standard tool like IMCI to identify children suspected to be HIV infected, stage them, test and link them to care. Subsequently, national tools for paediatric HIV care should be developed and supplied to all facilities offering care. All health facilities and community organizations should have directories that will assist in referring children to appropriate care, psychosocial support and primary health care services, including immunization. All children who are enrolled in HIV programmes should be staged using clinical criteria where CD4 machines are not available. Consequently, simple staging tools should be developed and availed to frontline health-care workers.

Early identification and diagnosis is vital for survival of HIV-infected children. There is an urgent need to develop a system to avail diagnostic facilities such as PCR testing and P24-antigen testing for infants in partnership with the existing facilities. We propose that a referral laboratory is established at provincial level and is mandated to provide training and supervision of lower-level facilities.

There should be a national consensus on the indicators necessary to monitor identification and treatment of HIV-infected children. We suggest the indicators be reported at the following levels: health facility, district, province and national. At the same time, national tools for paediatric HIV care should be developed and supplied to all facilities offering care.

In conclusion, policies, guidelines and clinical and laboratory structures are required to accelerate diagnosis of HIV-infected and exposed children. Health-care workers need to be trained, supervised and provided with tools to ensure that HIV-exposed or infected children are identified and enrolled into care.

Limitations of the study

This assessment had several limitations, including the fact that we only focused on sites already offering ART. We did not gain access to some sites and for logistical reasons we did not visit far-flung health facilities. It would also have been desirable to understand community barriers (e.g. costs) to access to HIV testing for children. However, we believe that the information we collected from the 58 sites generally reflects national challenges in paediatric HIV and the findings would be broadly applicable to most health-care settings.

Impact of the study on paediatric HIV care in Kenya

The findings of this survey were disseminated to the National Paediatric HIV Steering Committee in March 2006. A national network for collection and referral of blood samples for DNA PCR has been established with close to 10 000 specimens being processed between January 2006 and March 2007. All the four laboratories identified as capable of performing DNA PCR and three others that were not part of this survey have either joined the national network for early infant diagnosis or are in process of joining. In addition, a national algorithm for the early infant diagnosis has been developed (Fig. 1) and has been disseminated widely and is included in all trainings on paediatric HIV care and PMCT. The policies on paediatric HIV testing have been updated to include routine offer of HIV testing to all children accessing the health-care system and routine offer of early infant diagnosis for all six-week-old infants on their first immunization clinic visit.6 A combined mother–child card has also been finalized and will be piloted soon. ■

Fig. 1. Algorithm for early infant diagnosis, 2006<sup>a</sup>
Fig. 1. Algorithm for early infant diagnosis, 2006a



  • National AIDS/STD Control Programme (NASCOP), Ministry of Health, Nairobi 00202, Kenya.
  • Kenyatta National Hospital, Nairobi, Kenya.
  • NARESA, Nairobi, Kenya.
  • CDC Kenya, Nairobi, Kenya.