AUTHORS: C.I. Gotsche1,2, M. Simwinga2, A. Muzumara2, K.N. Kapaku2, L. Sigande2, M. Neuman1, M. Taegtmeyer3, E. Corbett1, C. Johnson4, A. Schaap1,2, A. Mwinga2, K. Hatzold5, H. Ayles1,2

HIV self-testing in Zambia: User ability to follow the manufacturer's instructions for use

CONCLUSION The OraQuick® HIV rapid test, though validated under ideal conditions, is shown to be challenging to use under real life conditions even after step-by-step demonstrations. Cognitive interviews and video recording showed that further improvements of pictorial/written instructions for use and the way the HIVST kits are designed is required to decrease user errors and to enable people to follow the IFU reliably. Following these findings the CPS study was redesigned to include a detailed demonstration of the correct methods for using the test. Whether the personal demonstration given in this study can be replaced by a video clip, a phone-app or by public demonstrations should be scrutinised in future studies.

The users ability to perform the HIVST correctly, is important to obtain a valid test result. Incorrect usage usually results in a negative result rather than an invalid one, which may falsely reassure the user. Therefore, the discrepancy between the participants perception of their test performance and their actual performance should be considered while scaling-up HIVST in that setting. Further studies with larger sample sizes are required to examine which steps of the test if conducted incorrectly are associated with a reduced sensitivity and should be addressed when modifying the IFU.

CONTACT Dr. Caroline Gotsche, MSc

Caroline.Gotsche1@alumni.lshtm.ac.uk

+49 15781571515

1) London School of Hygiene and Tropical Medicine, Faculty of Infectious and Tropical Diseases, London, United Kingdom, 2) Zambart Project, University of Zambia, Lusaka, Zambia, 3) Liverpool School of Tropical Medicine, Pembroke Place, Liverpool, United Kingdom, 4) World Health Organisation, Geneva, Switzerland, 5) Population Services International, Harare, Zimbabwe

VIDEO ANALYSISVideo recordings were analysed descriptively and common user errors were detected using a predetermined standardised checklist (Table 1). Scores were transferred into a database (Microsoft Excel Version 15.24 ©Microsoft) and subsequently analysed with a general-purpose statistical software package (Stata 14.0 ©StataCorp). Quantitative data was analysed for frequencies.

RESULTS Between June and August 2016, video recordings of 17 participants (13/17, 77% male) performing unsupervised HIVST (76.5% male, 88% rural) were obtained. Most participants (15/17, 88%) were living in rural communities and conducted the test at home. 2/17 (11.76%) participants performed the test in urban health facilities. Most of the participants have had an HIV test before (13/17, 76%). Almost half of the participants perceived their risk of having HIV as high or medium (8/17, 47%), 6/17 (35%) saw no chance at all of having HIV, and 3/17 (18%) did not know.

Video analysis showed that only 4/17 (24%) participants were able to conduct all steps correctly. Common user errors are shown in Table 2. Women were more likely to perform all steps of the test correctly (OR 5.5, 95% CI 0.46-65). Only 8/17 (47%) participants read the IFU before the test, despite explicit instructions to do so. There was a significant association (p< 0.05) between participants who read the instructions and their ability to correctly collect the oral fluid specimen. 12/17 (71%) of the participants collected the specimen correctly. The predominant difficulty with the swabbing seemed to be the movement of the pad along the gums.

Perception of the test: The vast majority (14/17, 82%) of participants found it easy to understand the IFU. Furthermore, (15/17, 88%) did not report believing that they committed any errors, which clearly contradicts the findings of the video analysis.

Participant read the instructions before testing

Opened the pouch with a picture of the vial first

Opened the vial correctly

Placed the opened vial in the khaki cardboard stand

Opened the pouch with a picture of the test device

Did not touch the test device flat pad

Did not dip the test device into the developer solution before collecting specimen

Collected the specimen correctly

Put the flat pad of the test device into the vial with developer solution and left the vial with the test device in the cardboard standWrote down or noted start time/set watch for 20 minutes

Did not remove the test device from the vial while test was running

Did not spill developer solution

Read results at 20 minutes

Comments related to interpreting results or reaction to results

Requested assistance from research staff

Table 1: Overview of the used observation checklist (further details were recorded in the video analysis)

User errors n/N %

Test preparation

Did not read the instructions before testing 9/17 53 %

Read the instructions after specimen collection 6/17 35 %

Placed the cardboard stand on an uneven surface 2/17 12 %

Spilled the developer solution 3/16 18 %

Specimen collection

Touched the pad of the test device 2/17 12 %

Swabbed teeth instead of gums 2/17 12 %

Swabbed the gum incompletely 1/17 6 %

Swabbed several times 2/17 12 %

Timing

Did not set an alarm 5/16 31 %

Did not wait 20 minutes 1/8 13 %

Table 2: List of common user errors (selection)

METHODS The study population included individuals 15 years or older who were able to give informed consent. The population targeted resided in settings with a generalised HIV epidemic (>1% prevalence). Urban and rural populations have been included. Before the start of the study, maps of one urban and one rural community were developed with the help of google maps and Zambart data sources. Sample enumeration areas were randomly selected and each household within the given area was visited for the

PSI/ UNITAID STAR Project. Two approaches were used to investigate the user ability to perform the HIVST. Cognitive interviews were conducted to assess understanding of the IFU. Video recordings were obtained in a second nested study of participants conducting the OraQuick® HIV rapid self-test.

References 1. Wong V, Johnson C, Cowan E, Rosenthal M, Peeling R, Miralles M, et al. HIV self-testing in resource-limited settings: regulatory and policy considerations. AIDS Behav. 2014 2. Zachary D, Mwenge L, Muyoyeta M, Shanaube K, Schaap A, Bond V, Barry Kosloff, Petra de Haas and Helen Ayles, Field comparison of OraQuick ADVANCE Rapid HIV-1/2 antibody test and two blood-based rapid HIV antibody tests in Zambia. BMC Infect Dis. 2012

POSTER NUMBER MOPED1167

Cognitive interviewing revealed that participants struggled to open the test kit easily. The most difficult instructions to understand were those related to the collection of oral fluid by swabbing the gums. Adolescents were more likely to swab accurately and to rely on both images and written instruction compared to adults. Understanding and interpreting images and particular terms (e.g. pouch, press firmly) was perceived challenging (Table 3)

COGNITIVE INTERVIEWSFor one nested study, cognitive interviews were conducted with 17 purposely selected adults and adolescents to assess understanding of the IFU. Working with community representatives, written instructions on the IFU were translated into local languages used in the regions the study was conducted. Participants were interviewed using a structured guide and the IFU. The structure of the guide mirrored the steps (pictures and accompanying written instructions) on the IFU. Participants were asked to do four major things: 1) read the instructions 2) reflect on the pictorial and written instructions and explain

back to the researcher 3) perform the actions depicted in the picture 4) reflect again on how easy or difficult other members of

the community like them would find the written and pictorial instructions.

RESULTS

Lay-out of the IFU All participants were comfortable with the lay-out and were able to point out that the front contained information on how to conduct the test, and the back instructions on how to interpret results.

Images/ pictures Some images did not relate well to the context and thus were wrongly interpreted. Cutlery was used on IFU to warn people not to eat or drink 15 minutes before testing but people in Zambia use hands; a red line was used to warn people not to do something but Zambians are used to a red crossed (X) image for ‘do not’. Clustered images were confusing.

Phrases/ words Meaning was lost in some words and phrases and participants suggested use of much simpler words used in everyday discourse; pocket for pouch, swab for slid along.

Interpreting results All cognitive participants correctly read and interpreted their own results, even though some had difficulties understanding the instructions. However, some suggested that people with poor sight may not see faint lines. Generally, participants use both the written and pictorial instructions to perform the test and read results whichever was clear at a given stage/ moment.

Table 3: User’s reflections on the IFU

“…others have never even used a fork and a knife, so it would probably be difficult for some others (people) to understand. At least you put, let’s say, vegetables or a bowl of nshima” (female, 29 years old)

BACKGROUND HIV self-testing (HIVST) is a new approach to increase testing uptake.(1) Although evidence demonstrates that supervised users can accurately perform HIVST, the ability of unsupervised users to do so, using manufacturer's instruction for use (IFU) requires further investigation.(2) The Clinical Performance Study (CPS) within the PSI/ UNITAID STAR Project provided participants, living in two communities within Lusaka district of Zambia, with the OraQuick® HIV rapid self-test (OFT) and IFU, and asked them to perform the test (See poster TUPECO842). In the pilot phase of the CPS (June 2016), participants were provided manufacturer’s IFU only. The IFU (Picture 1) had already been adapted following pilot work in Zimbabwe and Malawi and had pictorial symbols added to aid comprehension by partially literate or illiterate participants. The IFU had also been translated into locally used languages. This study investigated the ability of intended users of the OraQuick® HIV rapid self-test in Zambia to understand and follow the IFU.

Picture 1: Manufacturer’s IFU (Nyanja/English)