PARTNERSHIPS FOR ENHANCED ENGAGEMENT IN RESEARCH (PEER) HEALTH
Cycle 1 

Project Overview

Uganda’s under-five mortality is high, currently estimated at 90/1000 live births. Poor referral of sick children that seek care from the private sector is one of the contributory factors. The proposed intervention aims to improve timely referral and uptake of referral advice of children that seek care from private facilities. The private sector in this project will include private clinics and registered drug shops. The project will be implemented in Mukono district, central Uganda. This region was selected because a recent concluded trial in the district showed that drug shop vendors (DSVs) adhere to diagnostic test results, treat appropriately and refer sick children; although uptake of referral is poor. The main reasons attributed to the observed poor referral were negative attitude towards referral forms from drugs shops by the health workers at referral facilities, perceptions of poor quality of care at referral facilities and costs involved (Hutchinson. 2012). This project is a follow up to address these factors with the aim to improve uptake of referral. The primary objective of the project is to assess the effect of strengthening the referral system on timely uptake of referral of sick children who seek care in the private sector. The secondary objectives are: 1) to assess appropriate case management for malaria, pneumonia and diarrhea; 2) to explore factors which influence the referral or non-referral of sick children from the private sector; and 3) to assess the cost effectiveness of uptake of referral of sick children who seek care in the private sector. A cluster randomized design will be applied to test the intervention in Mukono District, Central Uganda. In the intervention arm, village health teams will sensitize communities on the importance of referral. Private outlets in both arms will be trained to diagnose, treat, and refer children. The proposed intervention will have three components: i) village health trainers (VHTs) will be trained to do community sensitization and initiate community discussions aimed at identifying community support mechanisms for financial hardship (to be community led and managed) ii) training and supervision of providers in the private sector to diagnose, treat and refer sick children, iii) regular meetings between the public and private providers (convened by the district health team) to discuss the referral system. The data generated from this study will contribute to an understanding of factors of importance for strengthening the referral system, including optimal training required, supervision activities, community participation and the costs involved. The lessons learned are likely to inform programming at a national and district level to improve referral of children from the private sector.

Final Summary of Project Activities

The primary objective of this project was to assess the effect of strengthening the referral system on the timely uptake of sick children who seek care in the private sector. The secondary objectives were: 1) to assess appropriate case management for malaria, pneumonia, and diarrhea; 2) to explore factors which influence the referral or non-referral of sick children from the private sector; and 3) to assess the cost effectiveness of the uptake of referral of sick children who seek care in the private sector.

A cluster randomized design was used to test the intervention in Mukono District, Central Uganda. A sample of study clusters implemented the intervention with ten randomly allocated to the intervention and ten to the control arm. The primary outcome was the proportion of sick children referred from the private sector were seen at higher level facilities:

• Intervention Arm: Community awareness on referral and training private providers using RDTs/ICCM to treat and refer sick children; supervision and regular meetings between the private and public sector.
• Control Arm: No community awareness on referral (trained private providers using RDTs/ICCM to treat and refer sick children).

Below are some of the findings from the trail:

• A total of 10,809 children aged <5 years were recruited in the study, (5955 intervention, and 4854 control). Less than 20% of the children were aged less than one year. Over 88% of the children had slept in a mosquito net the previous night. Most baseline parameters were comparable across the two arms. The majority of children presented with fever (78.1% in the control and 84.0% in the intervention); about a half of children. 46.5% in the control and 56.2% in the intervention arm presented with cough; while 25.4% presented with diarrhea in the control arm and 36.7% in the intervention arm. In the control arm, 58.5% mRDT were performed compared to 86.9% in the intervention arm. Of those mRDTs performed, 60.1% were positive in the control arm and 30.1% in the intervention arm.
• Overall, 38.7% of the children in the control arm were given Artemether/lumefantrine and 33.1% in the intervention arm. More children in the intervention arm, 43.7% were given amoxicillin compared to 19.1% in the control arm. Zinc tablets were given to 22.2% of the children with diarrhea in the control arm compared to 35.6% in the intervention arm. Similarly few children, 25.1% in the control arm were given ORS and 35.5% in the intervention arm. A high proportion of providers adhered to the mRDT test (83.0% providers in the control arm, and 94.3% in the intervention arm gave ACT to children who were mRDT positive), P=0.04.
  
• Of these 4128/4854 (85.0%) and 5262/5955 (88.4%) sought care within 24 hours. There were very few referrals in both arms of the study. In the control arm, only 20 (1.8%) of the children with negative mRDT and 21 (1.2% with mRDT positive were referred. In the intervention arm, 33 (0.9%) mRDT negative and 21 (1.4%) mRDT positive children were referred. There was no significant difference in referrals in both arms, P=0.53
  

In conclusion, adherence to mRDT test was better in the intervention arm than the control arms. Overall referrals were very few, with no significant differences between the two arms. The improved case management in the intervention could have affected the referrals in that arm.

The team also designed protocols and tools to conduct a community survey to assess uptake of referral and costs involved in taking up referral. They conducted 24 qualitative interviews: i) six focus group discussions (FGDS) among care takers who recently had sick children (three among mothers and three for fathers) to assess their views why caretakers take up referral or not; ii) four FGDs among health workers (private and public health workers together) to explore their perceptions about the referral system; iii) eight key informant interviews (KIIs) with village health teams (VHTs) and community leaders (local council leaders) to understand whether the VHTs conducted community sensitization on referral as they were meant to; iv) six FGDs community members (three for women and three for men) to explore whether a financial scheme is acceptable to the population and if they think it can enable uptake of referral. The team is currently analyzing the information gained from these interviews