Financial Incentives for Pediatric HIV Testing (FIT): Caregiver Insights on Incentive Mechanisms, Focus Populations, and Acceptability for Programmatic Scale Up

Abstract

Children living with HIV experience gaps in HIV testing globally; scaling up evidence-based testing strategies is critical for preventing HIV-related mortality. Financial incentives (FI) were recently demonstrated to increase uptake of pediatric HIV testing. As part of this qualitative follow-up study to the FIT trial (NCT03049917) conducted in Kenya, 54 caregivers participated in individual interviews. Interview transcripts were analyzed to identify considerations for scaling up FI for pediatric testing. Caregivers reported that FI function by directly offsetting costs or nudging caregivers to take action sooner. Caregivers found FI to be feasible and acceptable for broader programmatic implementation, and supported use for a variety of populations. Some concerns were raised about unintended consequences of FI, including caregivers bringing ineligible children to collect incentives and fears about the impact on linkage to care and retention if caregivers become dependent on FI.

Introduction

Many children living with HIV (CLHIV) globally remain undiagnosed; children lag behind adults in achieving the UNAIDS 95–95–95 goals for HIV testing, treatment, and virologic suppression [1]. In Kenya, 21% of CLHIV were undiagnosed in 2018 [2], despite increased efforts to diagnose CLHIV, such as nationwide index case testing [3]. Early antiretroviral therapy (ART) is critical in reducing pediatric HIV-related mortality [4]. Scaling up evidence-based testing strategies for improving pediatric HIV testing is critical for preventing HIV-related mortality [5, 6].

Barriers to testing older children [2–15 years old] outside of PMTCT programs include fear of children testing positive for HIV, fear of discrimination, limited training of healthcare workers to conduct HIV testing for children, limited HIV testing infrastructure within facilities, lack of uniform data reporting systems, and financial barriers, including cost of child care, transportation, and lost salary [7, 8]. Multiple interventions exist to address these barriers, including expanding clinic hours, provider-initiated HIV testing and counseling, and community engagement [7, 8].

Financial incentives (FI) may address financial barriers that caregivers face in testing children. FI have been effective in increasing desirable health behaviors in numerous trials [9–13], including HIV testing in adolescents [9]. FI may influence decision-making by directly offsetting costs associated with transportation, missed work or school, and by increasing willingness to test [14–16]. FI can also work by nudging caregivers to take actions earlier [17]. Although FI are effective, there are concerns that FI may have unintended consequences when offered in a programmatic setting, such as raising expectations and creating dependency, leading to decreased willingness to continue in care if FI are not provided. In addition to concerns about unintended consequences, ethical issues, such as decreased voluntariness and concerns about equity, underlie potential acceptability of the use of FI to motivate health behavior [18, 19].

A better understanding of multiple stakeholder perspectives on the acceptability and feasibility of FI can inform potential programmatic use. A previous study on healthcare worker experiences with FI in a trial setting found high acceptability, but also noted concerns on ensuring feasibility and sustainability within programmatic settings [20]. Gaps remain in understanding caregiver experiences and beliefs about the use of FI to increase pediatric HIV testing and illuminate mechanisms of how FI influence HIV testing decisions. This study aimed to evaluate caregivers’ views on the motivational and social impact of offering FI in a programmatic setting.

Methods

Study Design and Population

This qualitative study assessed caregiver experiences and opinions on the motivational and social impact of FI on pediatric HIV testing. The study was informed by the transtheoretical model, which describes stages of health behavior decisions and key influences on those decisions [21].

This study is a qualitative follow-up study to the FIT trial (NCT03049917), which was conducted in three counties (Kisumu, Siaya, and Homa Bay) in Kenya with high HIV prevalence [22]. Adults living with HIV (caregivers) were eligible to participate in the FIT trial if they were ≥ 18 years old, receiving HIV care, and had at least one child < 13 years who was of unknown HIV status, defined as never having tested for HIV, or not completing testing after the cessation of breastfeeding. In the trial, caregivers were randomized to receive 2018 US $0 (control), US $1.25, US $2.50, US $5.00, or US $10.00 in 1:1:1:1:1 allocation; the trial outcomes were the proportion of caregivers completing testing for one or more children and the time to testing completion [21, 23].

We sampled participants based on whether they chose to test during the FIT trial and if they were exposed to the FIT intervention and included: (1) FIT study testers: participants enrolled in the FIT trial who completed testing for their children; (2) FIT study non-testers: participants enrolled in the FIT trial who did not complete testing for their children; and (3) FIT naïve: participants who were not enrolled in the FIT trial, but had children; unlike in the trial, these children did not need to be of unknown HIV status. This approach provided a variety of views in different exposure/uptake populations. Recruitment of qualitative participants occurred after the FIT trial had ended and was conducted in 6 of the 19 trial clinics.

Data Collection

Caregivers provided written informed consent to participate, completed a demographic questionnaire, and participated in an in-depth interview (IDI). Two trained Kenyan social scientists, who were not involved in the FIT trial and did not provide clinical care for caregivers, conducted interviews using a semi-structured interview guide. The interview guide was designed to elicit caregivers’ experiences with FI, how FI influenced their decision-making and actions, and their suggestions on future implementation of FI in a programmatic setting. Interviews lasted an average length of 21 min. Interviews were conducted in English, Kiswahili, or Dholuo depending on participant preference. All interviews were audio-recorded, transcribed, and translated by trained staff who were proficient in English, Kiswahili, and Dholuo. Interviewers wrote structured debrief reports following each interview to summarize key information obtained [24] and help inform when conceptual saturation had been reached.

Data Analysis

Conventional content analysis was performed on caregiver transcripts [25], with the goal of identifying caregivers’ experiences with and beliefs about the use of FI for pediatric HIV testing. Specifically, analysis focused on characterizing: (1) satisfaction and personal experiences with the FI program, (2) mechanisms underlying decisions to test or not test, (3) beliefs about FIs as appropriate motivators, (4) beliefs and experiences with randomization and intervention components, and (5) beliefs about scaling up the FI intervention.

A codebook was first developed deductively based on transtheoretical model domains and literature reviews on the use of FI to encourage health actions. Inductive codes were then developed based on caregivers’ responses in debrief reports and a review of a subset of full interview transcripts. The codebook was refined through an iterative process of reviewing additional transcripts and assessing the codebook for accuracy to capture emerging themes. Codebook development and the coding process used a modified constant comparative approach to ensure data accuracy and reliability. After the codebook was optimized to the data set, coders (JZ, DLA, KBS) completed one round of consensus coding to ensure consistency in code interpretation and application across transcripts. Coders were each assigned a group of transcripts for independent coding, and then exchanged transcript groups for review by another coder. During the review process, coders applied additional codes as necessary, and noted any disagreements in code applications. After all transcripts were reviewed, the coding team discussed and resolved all disagreements. Transcripts were coded using Atlas.ti (version 8). Queries were used to compare experiences between caregivers, and identify themes emerging from the transcripts.

Results

Fifty-four caregivers participated in individual interviews; 25 were enrolled and had tested their children during the FIT trial, 15 were enrolled in the FIT trial but did not complete HIV testing for their children, and 14 were not enrolled in the FIT trial. The majority of caregivers were female (n = 41, 76%) and reported primary education as their highest level of education (n = 39, 72%). Overall, less than half (n = 24, 44%) of caregivers were employed, but employment status varied between FIT testers (n = 14, 70%), FIT non-testers (n = 5, 33%), and FIT naïve caregivers (n = 5, 36%). Among FIT testers, most caregivers (n = 45, 84%) were parents of the children, and the median number of children tested was 2. For FIT non-testers and FIT naïve caregivers, we did not obtain data on relationships with children. FIT naïve caregivers had tested none (n = 6, 43%), some (n = 4, 29%), or all their children for HIV (n = 4, 29%) (Table 1).

Table 1.

Demographics

	FIT tester (N = 25) n (%) or median (IQR)	FIT non-tester (N = 15) n (%) or median (IQR)	FIT naïve (N = 14) n (%) or median (IQR)
Female	16 (64%)	12 (80%)	13 (93%)
Highest level of education completed
Primary	17 (68%)	13 (87%)	9 (64%)
Secondary/high school	6 (24%)	2 (13%)	4 (29%)
University/college	2 (8%)	0	1 (7%)
Employed	14 (70%)	5 (33%)	5 (36%)
Median household size	5 (5, 6)	4 (2.5, 5)	3 (2, 4)
Number of children tested	2 (1, 2)	0* (0, 0)	1 (0, 1)

^*N = 13; 2 of the 15 caregivers did not mention how many children they have tested for HIV

Caregivers identified several barriers to pediatric HIV testing including: financial challenges (e.g. finding money to pay for transportation), time constraints, lack of support from partners, fear of discrimination, fear of having infected their children, and disclosing their HIV status to children. FIT naïve caregivers identified similar barriers to HIV testing as FIT experienced parents (Table 2).

Table 2.

Example quotes related to FI mechanism and considerations about the use of FI in programmatic settings

Domain	Construct	Quotes
Challenges to pediatric HIV testing	Financial constraints	… when I came here I was helpless, … I had left one of my children in [village name] then I had another son in [high school] that in fact I could not raise fees …. I found it was going to be a challenge going to get that child all the way here because I didn’t have that fare (FIT non-tester, 37-year-old mother, $0 control arm)
	Time constraints	I always got so busy and I am also the one who had to take care of her, so this made her fall sick often… so when I was asked to bring her for testing, I went and talked to her and she agreed and said ‘it is okay, I will just go and ask for permission from school that you are taking me to the hospital’ so the teacher also agreed and we went (FIT tester, 30-year-old father, $2.50 arm)
	Lack of support from family members	The challenges because this girl has her father, I went to the father and spoke to the father but the father did not agree. he said ‘no, not on my child’ (FIT tester, 36-year-old-mother, $5 arm)
	Fear of having infected the child	I didn’t test her when she was younger, you see when I conceived her that is when I discovered that I had HIV and that time when I was diagnosed with HIV, I was still afraid … Because I didn’t give her that care that is given to children, I breast fed her at the same time gave her solid foods. So that is why I was afraid, I just knew that I already infected my child (FIT tester, 45-year-old mother, $10 arm)
FI mechanism	Direct offset: direct cost	I didn’t have transport money but it forced me to borrow transport because when I heard that I was coming to test and I already developed courage because of that push, I was courageous that even [if] he is sick, I will just start him on medications early because I am also on medication (FIT tester, 45-year-old mother, $10 arm)
	Direct offset: indirect cost	I felt good and I learnt that there is benefit if you bring your child, there is a benefit you get at least after that you will go and buy the child something. (FIT tester, 43-year-old mother, $2.50 arm)
	Money as a nudge	Coming here messed my work schedule however they had told me that they were going to give me transport reimbursement following the card that I scratched yes that is why I stopped going to work and decided to just bring her. The money encouraged me a bit (FIT tester, 39-year-old uncle, $5 arm)
	No incentives needed	You know the child is yours the child is yours. The doctor knows his children’s status now it is your responsibility, you can bring the child even without being given money, so that you can know their status because that child is yours. Therefore there is no need of being given money because the child is yours not his, you are the one who is supposed to look at the welfare of your child (FIT naïve, 22-year-old female caregiver, no FI)
Proposed focus population for FI	Primary caregivers	I think it should enroll more women than men because women are a little bit reliable and even in terms of responsibilities very few men are responsible (FIT tester, 39-year-old father, $10 arm)
	Enrollment on care	Those who have just been enrolled, they are still weak in treatment… they might be discouraged and stop treatment. But those who are ongoing know what they have been through and the steps that they had to take, so they are okay, they can be explained to slowly and they understand (FIT tester, 27-year-old mother, $10 arm)
	Financial challenges	…those who have been on care for a long time and they are following the rules of treatment well but they are not financially okay for sure even if you go to their home you will find that they are really [impoverished] and maybe even the government knows and have taken responsibility of paying fees for his children (FIT non-tester, 58-year-old male caregiver, $2.50 arm)
Potential negatives	Program manipulations	The negative thing that can happen is that he may start spreading to people that we only come here for money and not because we are sick and you know he might discourage even those who wanted to come when he says that we only come because of money and not because one is sick (FIT non-tester, 35-year-old female caregiver, $2.50 arm)
	Dependency on money	I’m already used to it because that money ended and it was renewed, I scratched and found out that I was continuing, don’t you see that is an addiction, like I know that if I go to [hospital’ s name] then I will get money, so when I don’t get that addiction it discourages me (FIT tester, 43-year-old mother, $2.50 arm)

HIV-related knowledge—including modes of transmission and common HIV symptoms in children—was accurate among most caregivers. However, caregivers had minor gaps in HIV transmission knowledge. Many caregivers believed sharing razor blades, toothbrushes, and other sharp objects were common contributions to HIV transmission among children and relatives. A few caregivers reported not being aware of the possibility of testing their children, while others reported confusion about whether a child without symptoms or born before receiving their positive HIV test result could be living with HIV, as reasons for not having tested their children in the past. One caregiver described thinking his child could not be living with HIV because the child was born before his HIV diagnosis:

…I thought he didn’t have [HIV] because [the child] was someone who was older…I remembered that when I was being tested and diagnosed with HIV [that] child had been born long before that. So I was sure he couldn’t be having HIV.

(FIT tester, 45-year-old father, $1.25 arm)

Although caregivers noted some concerns related to feasibility and sustainability, they viewed the use of FI as an acceptable and effective approach for improving pediatric HIV testing. Specific considerations influencing caregiver views on the use of FI centered around three main themes: (1) FI can directly offset costs or provide a nudge to encourage testing, (2) beliefs about FI values provided in research studies and programmatic settings were mixed, and (3) proposed focus populations for incentives reveal caregiver understanding of the goal and mechanism of incentives (Table 2).

FI Mechanisms Can Directly Offset Costs or Provide a Nudge to Encourage Testing

Caregivers reported distinct mechanisms by which FI enabled them to test their children, either by directly offsetting direct or indirect costs, or by nudging them to take action in order to receive a FI. Caregivers reported using the money to offset school fees, provide food for the family, pay debts, and provide treats for their child like a toy or snack.

…the money that I was given that day, it helped me a lot, because I didn’t [have] money that day that my child could go to school with, so I gave him transport and lunch and I also found a means of helping myself.

(FIT tester, 30-year-old father, $2.50 arm)

Caregivers also reported that FI encouraged them to take actions faster. Many caregivers reported having thought about testing their children for HIV, but postponing due to barriers or lack of perceived urgency. FI provided the nudge needed to motivate them to take action now rather than later.

[The money] changed my mind, I was a bit lazy and kept postponing this issue of coming to the clinic to test this child, but when I heard that I was going to be given money I said let me go try … So at least it encouraged me to reach the hospital with the child.

(FIT tester, 27-year-old mother, $10 arm)

Beliefs About FI Values Provided in Research Studies and Programmatic Settings were Mixed

Caregivers were asked to reflect on the value of FI they were randomized to receive in the trial, as well as potential values that might be used in a hypothetical future programmatic setting. Caregivers’ feelings towards the values of FI received in the trial were mixed, but there were no meaningful differences between caregivers from different incentive values. Many caregivers who participated in the trial and brought their children for testing said that they would still have brought their children for testing had they received bigger or smaller FI. While caregivers noted that smaller FI values were less likely to lead to immediate action, many caregivers said they would bring their children for testing faster if FI values were doubled:

Interviewer: If you found more money would that have encouraged you to come immediately or you would have just come back as you had planned?

Respondent: I would have come back immediately so that I can get it.

(FIT tester, 42-year-old aunt, $2.50 arm)

When asked whether they felt the value of the FI they received within the trial was fair, most caregivers said they were happy about the amount they received. The majority of caregivers recognized that the FI value was based on luck (e.g. randomization), and were therefore willing to accept whatever amount they received.

…whatever I got I just said that it is my luck, if I get good money or little money that was my luck so I just accepted whatever I got”.

(FIT non-tester, 43-year-old female caregiver, $5 arm)

However, a few caregivers were not happy with the value of FI they received, and said they wished to have received more money.

Given historical experiences with money being offered to members of their community, caregivers advised against providing large FI values within programmatic settings, as these might arouse suspicion. In general, caregivers felt that smaller value FI being given programmatically would be perceived as transportation money.

No, the kind of money that brings such kind of questions are big money like someone giving you a hundred thousand and you didn’t have any thoughts about that but money that is being given like the one that you gave us for transport, it is not money that can bring such thoughts.

(FIT tester, 39-year-old uncle, $5 arm)

Some caregivers mentioned potential negatives associated with offering FI, including the possibility of corruption, the possibility of people bringing in other people’s children, and dependency on FI. Caregivers offered possible strategies to overcome these potential negatives, including checking birth certificates.

…if you give me a lot of money then I can even start a business, so I just be taking other people’s children and bring them for testing… and now that can also bring another problem (FIT tester, 35-year-old mother, $5 arm)

Some caregivers felt that financial incentives were not necessary, and that motivation should come from intrinsic rather than extrinsic factors. These caregivers felt it was their responsibility to test their children and they would still bring their children for testing even if there was no money, and believed that intrinsic motivation should be encouraged for long-term care of the child, given the limited perceived sustainability of FI.

According to me, what should encourage someone is the health of their child… it is the health of the child that is important.

(FIT naïve, 30-year-old female caregiver, no FI)

Proposed Focus Population for Incentives Revealed Caregiver Understanding of Goal and Mechanism of Incentives

Caregivers had diverse views on the focus population for FI. Differences in beliefs depended on how they thought FI operated and what purpose FI fulfill. However, beliefs did not differ between FIT testers, FIT non-testers, and FIT naïve groups. Some saw FI as compensation for the person in charge of the children’s health, and felt FI should focus on primary caregivers, especially mothers:

Money, even if you give a father or a mother, it is the mothers who know the problems of her household, you know fathers will give you only half of the money and give you very little. So it is the woman, she will budget with it.

(FIT non-tester, 49-year-old female, $5 arm)

Some caregivers viewed FI as income generating or substituting and felt FI should focus on people with financial challenges. Other caregivers felt the focus population should be determined based on their time of enrollment on HIV care, but had diverse opinions on which population should be prioritized. Some felt FI could help motivate those newly diagnosed to take immediate action to test their children, while others felt FI could be the nudge needed for those who have been enrolled in care longer but have still not yet tested their children.

They need to encourage those who have just been initiated on care…otherwise they will find it challenging…motivating them will make them feel like they are not alone and that they can also make it with the medication and be like the others; because at that time when they are being initiated they might have a lot of stress.

(FIT naïve, 22-year-old female, no FI)

…[the one who is already established in the program] already know, she knows the importance, who knows that when I am called I must go and leave my job. For me, I think that’s the one who is supposed to be given.

(FIT tester, 35-year-old mother, $2.50 arm)

In addition, some caregivers felt that FI should not be based on who the intended recipient was. They believed everyone should be considered equal and people should receive FI regardless of their time enrolled in HIV care.

I think once someone is on care their issues are the same… those things are equal, there is nothing like that because you cannot speak of an older disease or a new disease, a disease is a disease.

(FIT tester, 38-year-old mother, $2.50 arm)

Discussion

FI have the ability to motivate health behaviors, including HIV testing for children of unknown HIV status. In our follow-up study to a trial that used FI to improve pediatric HIV testing, caregivers described high acceptability and belief in effectiveness of FI, and many recommended future implementation of FI programs. FI were found to work by directly offsetting costs and by nudging caregivers to make testing a higher priority, depending on the caregiver’s specific barriers to testing their child(ren). In addition to financial challenges, other barriers to testing children for HIV reported by caregivers were found to be similar to those identified in previous studies [7, 8].

Behavioral economics—which combines traditional economic theory with psychology theory—describes the mechanisms of FI to either operate through a “nudge”, by supporting an individual to take an action, even if the FI value itself does not fully offset their costs [11, 12, 17, 26], or through directly offsetting costs [17, 26]. Comparing the two mechanisms, FI values that nudge—or alter caregivers’ priorities—influence people’s decision-making through psychological processes. In contrast, FI values that directly offset costs operate mechanistically more similarly to traditional FI [27]. Within the FIT trial, the incentive arms that did not directly offset the average costs for the study population did not have a significant or appreciably different level of HIV testing uptake from the unincentivized control, suggesting that a nudge was not effective in this population [23].

Some caregivers in our study reported they would have been more likely to test rapidly if they received a higher FI value, while others reported that FI value would not have impacted their behavior. Within the parent trial, increasing FI values were associated with both higher uptake and faster time to test completion [23]. Other FI trials of HIV testing that included multiple fixed value incentive arms show a heterogeneous relationship with increasing value and uptake [10, 11], but a similarly structured trial on voluntary medical male circumcision for HIV prevention demonstrated a dose–response relationship between FI value and uptake [28]. FI are generally more effective in encouraging one-time actions—like HIV testing uptake—and less effective in maintaining long-term behaviors—like adherence for viral suppression [29, 30]. Therefore, the goals of FI use should be carefully evaluated before implementing FI programmatically.

Caregivers’ opinions on who should receive FI in a programmatic setting differed from those reported by health care workers in a prior analysis [20]. Health care workers suggested expanding the population to include adolescents and adult men, while male and female caregivers in this study preferred focusing on women. This finding existed despite a similar proportion of caregivers and health care workers being female. These differences in who the focus recipients should be within a programmatic setting may create additional challenges for health facility leaders deciding whether or how to implement FI within a programmatic setting. Both health care workers and caregivers raised concerns about ensuring that children brought for testing truly were eligible, citing concerns that caregivers could bring children that were already tested or did not belong to them. The solutions raised by caregivers included checking birth certificates or other time-intensive procedures, while solutions proposed by health care workers were more systems-oriented and included using routine program data to determine eligibility [20].

Previously we described health care worker perceptions on programmatic use of FI for pediatric testing [20]. Healthcare workers described concerns regarding limited resources available within their health facilities in relation to programmatic delivery of FI, and worried about increasing workload associated with increased number of people bringing their children for testing [20]. Unlike health care workers, such comments were rarely mentioned by caregivers. Similarly, health care workers had concerns about the cost of sustaining such a program [20] while caregivers did not raise many concerns about programmatic cost. Health care workers also discussed the option of offering non-monetary incentives [20], which was not raised by caregivers.

Limitations

It was not possible in this study to directly assess the concern of erosion of intrinsic motivation or the risk of caregivers bringing children who were not theirs for testing. This study included views from caregivers; however, it would be ideal to combine views from other stakeholders, such as policymakers, to inform a more comprehensive understanding of the potential use of FI in larger programmatic settings. Participants were recruited from participating clinics in three counties in Kenya, and the results may not be generalizable to other counties in Kenya and other low- and middle-income countries. Finally, children ≥ 13 years were not included in this study, limiting generalizability to adolescents.

Conclusion

FI were considered highly acceptable and feasible for programmatic delivery by caregivers, and many caregivers supported the idea of future scale up. Caregivers described two mechanisms of FI action, including directly offsetting costs and “nudging” action earlier. Caregivers generally perceived the range of FI values in the trial to be acceptable in programmatic settings, but differed in their views about who the intended beneficiaries of such a program should be.