University of California, San Francisco Logo

University of California, San Francisco | About UCSF | Search UCSF | UCSF Medical Center

Paediatric HIV management at primary care level: an evaluation of the integrated management of childhood illness (IMCI) guidelines for HIV
Global Health Sciences Literature Digest
Published June 03, 2010
Journal Article

Horwood C, Vermaak K, Rollins N. Paediatric HIV management at primary care level: an evaluation of the integrated management of childhood illness (IMCI) guidelines for HIV. BMC Pediatr. 2009 Sep 22;9(1):59.

In Context

The Integrated Management of Childhood Illnesses (IMCI) guidelines were developed by the World Health Organization (WHO) and UNICEF to improve child survival in resource-limited countries.(1) For care of ill children, the IMCI provides guidelines that are designed to be used by trained health workers. South Africa adopted the IMCI in 1997. IMCI guidelines were amended to include methods to identify and manage HIV-infected and HIV-exposed children. Previous reviews of IMCI implementation have shown that IMCI training improves the performance of the health worker but before this study, implementation of the HIV guidelines had not been evaluated.


To determine how the IMCI HIV guidelines are used in clinical settings by trained and expert health care workers and to measure the validity of the HIV algorithm in clinical practice


First-level health care facilities in Limpopo and KwaZulu-Natal (KZN) provinces, South Africa

Study Design

Cross-sectional observational study


IMCI-trained health workers in primary health care clinics and all ill pediatric patients aged two months to five years.


1) The sensitivity, specificity, positive predictive value, negative predictive value, and likelihood ratio of the IMCI algorithm for HIV classification; and 2) the proportion of children classified by health care workers compared to IMCI experts as having symptomatic HIV disease and being HIV exposed.


A random sample of all IMCI-trained health workers working in primary care clinics was selected. All health workers had been informed that a child health practice survey was to be conducted, but selected health workers were not informed of their selection nor were they aware that IMCI was being evaluated. Consent for the children to have their health worker evaluated and for HIV counseling and testing (for children not known to be infected) was obtained from mothers or caregivers.

Two expert IMCI practitioners conducted observation of health workers and observed the interaction with the patient; the experts recorded findings without intervening with the care. The same IMCI expert observed all observations of a particular health worker. A second IMCI expert then assessed the same child separately and recorded the findings using a standard format. The second IMCI expert was not aware of the findings of the first IMCI expert's observation of the health worker. Identifying information about the clinics or health workers was not recorded. If the recommendations of the IMCI expert for management of the child differed from that of the health worker, the recommendations of the expert were used. Each health worker was observed for up to 20 consultations with ill children.

Interviews with the health workers were conducted to ascertain their training in child health and IMCI. Resources available at the clinics were also recorded. Data collected during the observation of consultations included the activities undertaken and counseling provided. The observer recorded the information made on the child's health card after the health worker had completed the work. Feedback from the expert to the health worker was provided after observation of all consultations was complete. The assessments from the health workers were compared to those of the IMCI expert, whose assessments were considered to be the gold standard.


There were 77 health workers observed at 74 clinics between May 2006 and January 2007. Each health worker was observed for a mean of 2.7 days and 17.7 observations. The mean age of the children observed was 19.6 months.

Among the health workers observed, 55 (71.4%) had no health care training for children beyond the IMCI. One worker was trained as an IMCI supervisor and one as an IMCI facilitator, seven had a diploma in primary health care, and four attended a short course in pediatric health care. Sixty-six (85.7%) workers had received training in HIV/AIDS counseling. Fifty-one (66.2%) had received the recommended follow-up after IMCI training. There was no ongoing IMCI supervision.

There were 1064 children with an HIV test result available. Using the IMCI algorithm, the IMCI experts classified 171 children as suspected symptomatic HIV and 201 as HIV exposed. Fifty-four (16.1%) of the 171 children with suspected symptomatic HIV were confirmed as HIV-infected, resulting in a sensitivity of the IMCI algorithm of 71%, specificity of 88.1%, positive predictive value of 31.6%, negative predictive value of 97.5%, and likelihood ratio of 5.99 (95% confidence interval [CI]: 4.79-7.48). Forty-three of the 201 children identified as HIV exposed were confirmed to be infected with HIV, resulting in a sensitivity of 56.6%, specificity of 84%, positive predictive value of 21.4%, negative predictive value of 96.2%, and likelihood ratio of 3.54 (95% CI: 2.77-4.51).

Thirty-one (40.3%) of the health workers did not classify HIV in any of the children and nine did not classify any child correctly for HIV. None of the health workers correctly classified every child for HIV. Health workers made a classification for symptomatic HIV in 428/1357 (31.5%) children; of these, 342 (25.2%) were classified correctly compared to the IMCI expert. Among all observed children, 202 (14.9%) were classified as suspected symptomatic HIV by the IMCI expert; of these 37 (18.3%) were correctly classified by the health worker. Health workers classified 84 children as suspected symptomatic HIV, including 28/76 (36.8% CI: 24.7%-51.0%) HIV-infected children. Health workers correctly identified 161/223 (72.6% CI: 65.3%-79.3%) children classified as HIV exposed by the IMCI expert.


The authors conclude that the IMCI HIV algorithm is valid when correctly applied, but few IMCI-trained health workers use the algorithm in clinical practice.

Quality Rating

This is an important and reasonably well-conducted study. However, the presence of the IMCI expert observer may have influenced the performance of the health worker under observation.

Programmatic Implications

Implementation of the IMCI has been shown to improve health worker performance,(2) but its usefulness is limited by poor adherence by health workers.(3, 4) Based upon the findings from this study, use of the HIV algorithm by health care workers will require additional training and supervision. With the increased availability of HIV testing, routine, opt-out testing of children is likely to be more effective than screening with the IMCI.


  1. Grove S. Integrated management of childhood illness by outpatient health workers: technical basis and overview. The WHO Working Group on Guidelines for Integrated Management of the Sick Child. Bull World Health Organ 1997;75 (Suppl 1):7-24. Abstract not available.
  2. Gouws E, Bryce J, Habicht JP, et. al. Improving antimicrobial use among health workers in first-level facilities: results from the multi-country evaluation of the Integrated Management of Childhood Illness strategy. Bull World Health Organ 2004;82:509-15.
  3. Pariyo GW, Gouws E, Bryce J, Burnham G. Improving facility-based care for sick children in Uganda: training is not enough. Health Policy and Plan. 2005;20 (Suppl 1): i58-i68.
  4. Horwood C, Vermaak K, Rollins N, Haskins L, Nkosi P, Qazi S. An evaluation of the quality of IMCI assessments among IMCI trained health workers in South Africa. PLoS One 2009;4:e5937.