Preference for daily (1HP) vs. weekly (3HP) isoniazid-rifapentine among people living with HIV in Uganda

SUMMARY BACKGROUND Both 1 month of daily (1HP) and 3 months of weekly (3HP) isoniazid-rifapentine are recommended as short-course regimens for TB prevention among people living with HIV (PLHIV). We aimed to assess acceptability and preferences for 1HP vs. 3HP among PLHIV. METHODS In a cross-sectional survey among PLHIV at an HIV clinic in Kampala, Uganda, participants were randomly assigned to a hypothetical scenario of receiving 1HP or 3HP. Participants rated their level of perceived intention and confidence to complete treatment using a 0–10 Likert scale, and chose between 1HP and 3HP. RESULTS Among 429 respondents (median age: 43 years, 71% female, median time on ART: 10 years), intention and confidence were rated high for both regimens. Intention to complete treatment was rated at least 7/10 by 92% (189/206 randomized to 1HP) and 93% (207/223 randomized to 3HP). Respectively 86% (178/206) and 93% (208/223) expressed high confidence to complete treatment. Overall, 81% (348/429) preferred 3HP over 1HP. CONCLUSIONS Both 1HP and 3HP were highly acceptable regimens, with 3HP preferred by most PLHIV. Weekly, rather than daily, dosing appears preferable to shorter duration of treatment, which should inform scale-up and further development of short-course regimens for TB prevention.

Improving uptake for TB preventive treatment (TPT) is crucial to address the TB pandemic; however, there is little evidence on treatment preferences among people living with HIV (PLHIV) in high TB incidence settings. 1,2Although 6 months of isoniazid preventive treatment (IPT) has been accessible for decades, shortercourse TPT is not widely available in routine care. 3,4ecently, 3 months of weekly isoniazid-rifapentine (3HP) has been recommended in many high TB burden countries, including Uganda, while 1 month of daily isoniazid-rifapentine (1HP), another WHOrecommended short-course TPT regimen, is not yet available. 57][8] In addition, the cost of rifapentine has dropped since 2019, and fixeddose combinations have become available, making 1HP and 3HP viable alternatives to IPT. 3,9,10 Despite the emphasis on patient-centered care, which takes into account the preferences and values of PLHIV in the WHO End TB strategy, 11 it remains uncertain whether patients would prefer 3HP or 1HP.Patient preferences could inform TPT regimen recommendations and scale-up, but there may be trade-offs with cost and the feasibility of offering several regimens. 12Treatment preference impacts patient adherence, satisfaction, and outcomes. 13Evidence on TPT preferences among PLHIV is especially sparse. 2 Studies in low TB incidence settings found that TPT regimens with higher effectiveness, fewer side effects and shorter duration were preferred. 14,15rior work to inform TPT policy has measured perceived confidence to complete treatment (self-efficacy), 16 which refers to people's beliefs about their individual capabilities to execute behaviors necessary to achieve important goals. 17In addition, perceived intention (commitment to act in a certain way) is influenced by self-efficacy, as well as patients' attitudes and subjective norms. 187][18] Therefore, in this cross-sectional survey, we aimed to estimate the acceptability of 1HP and 3HP among adult PLHIV in a high TB incidence setting using perceived confidence and intention, as well as preferences between the two regimens, to inform public health decisions.

Study design and participants
We conducted a cross-sectional, randomized, noninterventional survey among adult PLHIV (age �18 years) attending the Mulago Immune Suppression Syndrome (ISS) clinic under the Makerere University Joint AIDS Program (MJAP) between July and November 2022.Mulago ISS clinic is the largest specialized outpatient HIV/AIDS clinic in Uganda, with over 16,000 clients active in care.In 2022, 5,019 PLHIV initiated TPT at the clinic; 79% received IPT and 21% 3HP.Mulago ISS clinic started using 3HP in July 2020 through the 3HP Options Trial (clinicaltrials.govNCT03934931), a type 3 effectiveness-implementation trial of optimized 3HP delivery strategies. 19In June 2022, 3HP was rolled out at the clinic as part of routine national HIV programming.Although recommended by national HIV guidelines, 5 1HP was not in use at the clinic when this study was conducted.
We invited consecutive adult PLHIV attending their routine HIV/AIDS clinic appointments who had not started TPT within the past year, were not currently being treated for active TB, and were not incarcerated to participate in the survey.All interested and eligible PLHIV were included.All participants provided written informed consent in English or Luganda.Given that many clients had previously taken TPT, excluding them may have led to selection bias.Moreover, retreatment with TPT could be considered in the future.Thus, we included PLHIV with prior TPT (3HP or IPT) experience who had not been treated within the last year.

Data collection
Participants were randomly assigned (using simple randomization) to hypothetical scenarios of receiving either 1HP or 3HP via the in-built randomization feature of Sawtooth's Lighthouse studio offline survey app, which also facilitated survey data collection. 20llocation was concealed for the first part of the survey until the interviewer had to present either 1HP or 3HP information to elicit confidence and intention for the respective regimen.Participant education flipbooks, available in English and Luganda, were used to inform participants about TB, latent TB infection (LTBI), as well as 1HP or 3HP regimen-specific information on efficacy, number of tablets per dose, dosing frequency, duration of treatment, potential drug-drug interactions, including with antiretroviral therapy (ART), and potential side effects (Supplementary Data 1).Participant education materials and the questionnaire design were refined based on structured feedback from the first 29 participants and interviewers during a pilot phase.This feedback led to minor word changes and a change in the order of questions.All data were included in the final analysis.
Following this education, participants were asked to rate their level of perceived confidence and intention to complete the assigned hypothetical regimen, using a visual 11-point Likert scale ranging from '0' (strongly disagree) to '10' (strongly agree).Participants assigned to 1HP were asked to respond to the following standard statements: "I feel confident that I would be capable of completing all 28 daily doses of this treatment" and "I would intend to complete all 28 daily doses of this treatment".Participants assigned to 3HP responded to equivalent statements with 12 weekly doses.After rating confidence and intention for one regimen, participants were given a brief comparison of key features to the alternative regimen (1HP or 3HP) using illustrations in the participant education flipbook.They were then asked to state their preference between 1HP and 3HP.
Participants' sociodemographic and clinical baseline characteristics were self-reported (age, sex, education level, employment status, current use of hormonal contraception) or extracted from electronic medical records (ART status, regimen, and duration, viral load status, prior TB and prior TPT).Participants' selfreported age was confirmed using electronic medical records.Participants also self-reported household characteristics with which we derived the global multidimensional poverty index (MPI) -a measure of poverty that examines deprivations across 10 indicators in the dimensions of health, education, and standards of living; people deprived in one-third or more of all indicators are considered multidimensionally poor. 21he questionnaire was administered by six trained Ugandan research assistants (JN, FW, JK, CN, LA, and AN), who were fluent in both English and Luganda.They reviewed flipbook content with participants and encouraged and answered any questions prior to administering the survey.

Outcomes
The main outcomes were 1) confidence and 2) intention to complete 1HP or 3HP (11-point Likert scale), and 3) preference for 1HP compared to 3HP (binary).We explored associations with sociodemographic and clinical characteristics.Participants' confidence and intention were also considered predictors of preference for 1HP vs. 3HP.

Statistical analysis
All data were analyzed using both STATA v14.Patient preferences for 1HP vs. 3HP summarized as low (rating 0-3), moderate (4-6), or high (7-10).We used logistic regression to estimate associations between participant characteristics with preference for 1HP vs. 3HP.We used ordinal logistic regression to estimate associations between participants' confidence and intention ratings (using ratings from 0 to 10) and their sociodemographic/clinical characteristics.As prior 3HP experience could impact preference for 1HP vs. 3HP, we analyzed the main outcomes excluding participants who had previously taken 3HP in sensitivity analyses.

Ethical approval
This
In an ordinal logistic regression using the 0-10point Likert scale ratings, confidence was rated higher by those randomized to the hypothetical 3HP regimen compared to the hypothetical 1HP regimen (adjusted odds ratio [aOR] 1.72, 95% confidence interval [CI] 1.12-2.65).Participants using hormonal contraception rated their confidence lower than those who were not using hormonal contraception (aOR 0.53, 95% CI 0.29-0.97).Confidence ratings did not differ by age, sex, education status, MPI category, employment status, time on ART or prior TPT (Figure 3A).In a sensitivity analysis excluding participants with prior 3HP experience, results were similar (Supplementary Figure S1A).
In the ordinal logistic regression for the 0-10-point Likert scale rating of intention, randomization to the hypothetical 3HP regimen was associated with higher intention to complete treatment (aOR 1.78, 95% CI 1.14-2.78).Intention ratings did not differ by age, sex, education status, MPI category, contraceptive use, employment status, time on ART or prior TPT (Figure 3B).Results were similar in a sensitivity analysis excluding participants with prior 3HP experience (Supplementary Figure S1B).
After adjusting for other variables, the participant groups most likely to prefer 1HP were those assigned to the hypothetical 1HP scenario (27% preferred 1HP over 3HP: aOR 3.01, 95% CI 1.78-5.08)and those with no history of prior TB (21% preferred 1HP over 3HP: aOR 3.27, 95% CI 1.32-8.08).In no group did more than 27% of participants prefer 1HP over 3HP (Table 2).Participants' preferences for 1HP vs. 3HP did not differ by prior TPT experience.Our findings on participants' preferences for 1HP vs. 3HP did not change significantly in the sensitivity analysis that excluded participants with prior 3HP experience (Supplementary Table S1).

DISCUSSION
In this cross-sectional survey of 429 PLHIV randomly assigned to hypothetical TPT scenarios of receiving 1HP or 3HP at a large urban HIV/AIDS clinic in Kampala, Uganda, we found that 3HP was more often preferred, but both regimens were highly acceptable.When given a choice, over 80% preferred 3HP (88% of those randomized to the hypothetical 3HP regimen, 73% of those randomized to hypothetical 1HP).These findings suggest that health systems should elicit PLHIV preferences before scaling-up both 3HP and 1HP, and that in Uganda, greater focus should be on 3HP.
Acceptability was high for both regimens, indicating that either regimen, if offered as the only option, would likely yield high TPT coverage.This high acceptability is consistent with findings of high treatment Multidimensional poverty was included as a bivariate: we compared the group which was poor or severely poor to the reference group which was defined as vulnerable or not vulnerable (i.e., not poor).TPT ¼ TB preventive treatment; OR ¼ odds ratio; CI ¼ confidence interval; 3HP ¼ weekly isoniazid-rifapentine for 3 months; 1HP ¼ daily isoniazid-rifapentine for 1 month; ART ¼ antiretroviral therapy.
acceptance and completion for both regimens demonstrated by several randomized clinical trials. 6,8,22,23lthough shorter treatments are often assumed to be preferred, we found the opposite in this studyperhaps reflecting preferences for fewer medication days or lower pill burden.We informed participants that 1HP may require adjusting their daily ART dosage by adding an extra dolutegravir pill to cater for drug-drug interactions with rifapentine. 24Moreover, weekly dosing may be preferred over daily dosing.A qualitative study in Peru found that participants perceived less frequent dosing as less toxic. 25Although we presented potential side effects as equivalent for both regimens, this perception could have influenced our participants' choice.
In addition, preference for 3HP might also reflect greater familiarity with the regimen, although we did not find an association with prior TPT experience and preference.About one third of participants had previously taken 3HP, most of them likely in fixed-dose combinations.However, participants may have been familiarized with 3HP indirectly through the experiences of peers at the clinic, through its rollout in a randomized trial, 19 and programmatically through the HIV program, 5 or via the clinic's routine health education talks.
No prior studies have directly compared the preferences of adult PLHIV for 1HP vs. 3HP.In pediatric populations, reported drivers of TPT preference were pill burden (treatment duration and frequency), medication fatigue due to prolonged daily dosing, and capacity to remember to take their medicines. 263HP was deemed more preferable to IPT due to its weekly dosing schedule and minimal disruption to daily lives. 25ualitative evidence from the Mulago ISS clinic collected prior to the roll-out of 3HP, demonstrated that weekly dosing was a key facilitator for patient acceptance and completion of 3HP. 27igh acceptability of a short-course TPT regimen is key to achieving public health goals.Although we found that 3HP was associated with higher confidence and intention, both regimens were highly acceptable (regardless of patient characteristics, including education), and it is unclear whether small absolute differences of confidence or intention would translate to a relevant increase in TPT coverage and treatment completion compared to 1HP.These differences may also be affected by scale biases, where some groups may be more or less likely to select the end of the scale.
Our study has some limitations.First, although being offered TPT was a realistic scenario for this population, the scenarios were nevertheless hypothetical.Observation of real choices (i.e., revealed preferences) under actual treatment implementation would provide more direct evidence of patient preference for 1HP vs. 3HP.In addition, our measurement of preferences using a quantitative scale may not have been contextually suitable to the diverse population of HIV-positive Ugandans, with varying literacy gaps and sociocultural influences.However, this survey was administered by experienced research staff, both TPT scenarios were well explained, and piloting suggested that all elements of the study were well understood, including the quantitative scale.Finally, this survey was done in an urban, highly ART-experienced population.Future studies may consider replicating this assessment in populations with less ART literacy and in rural settings with less TPT experience.In summary, this study highlighted that both 1HP and 3HP were highly acceptable TPT regimens among 429 PLHIV in Kampala, Uganda, with 3HP being the preferable option to many, despite the common assumption that its longer duration would make it less preferable.Our findings suggest that health systems should not simply replace 3HP with 1HP based on the untested perception that a regimen of shorter duration will necessarily be more acceptable or result in higher levels of completion.These findings strongly support the roll-out of short-course TPT (both 3HP and 1HP) in high TB-HIV burden settings.Where feasible, patients should be given choices between regimens; in Uganda, if providing such choice is not feasible, priority may be given to 3HP.

R É S U M É C O N T E X T E
regimens.Intention to complete treatment was rated at least 7/10 by 92% (189/206 randomized to 1HP) and 93% (207/223 randomized to 3HP).Respectively 86% (178/206) and 93% (208/223) expressed high confidence to complete treatment.Overall, 81% (348/429) preferred 3HP over 1HP.C O N C L U S I O N S : Both 1HP and 3HP were highly acceptable regimens, with 3HP preferred by most PLHIV.Weekly, rather than daily, dosing appears preferable to shorter duration of treatment, which should inform scaleup and further development of short-course regimens for TB prevention.K E Y W O R D S : tuberculosis preventive treatment; acceptability; values; human immunodeficiency virus

Figure 1 .
Figure1.Summary of study participant eligibility screening and random assignment to the two hypothetical TB preventive treatment (TPT) scenarios, daily isoniazid-rifapentine for 1 month (1HP) or weekly isoniazid-rifapentine for 3 months (3HP).

Figure 2 .
Figure 2. A) Participants' level of perceived confidence and B) intention scores on an 11-point Likert scale ranging from 0 ¼ strongly disagree to 10 ¼ strongly agree, for completing the hypothetical TPT regimen to which they were randomized.The bars and corresponding numbers show the frequency of the Likert scale score for 3HP (black) or 1HP (grey).Percentages are shown in brackets at the top for the proportions of participants who expressed low (0-3), moderate (4-6) or high (7-10) confidence/intention for 3HP vs. 1HP.C) Participants' preferences for 1HP vs. 3HP are shown for all participants and by regimen shown first at random assignment (1HP or 3HP).The bars indicate preference for 3HP (black) or 1HP (grey).TPT ¼ TB preventive treatment; 1HP ¼ daily isoniazid-rifapentine for 1 month; 3HP ¼ weekly isoniazid-rifapentine for 3 months.

Figure 3 .
Figure 3. Variables associated with A) confidence and B) intention to complete TPT (n ¼ 429 participants).The circle indicates the adjusted OR with a line to indicate the 95% CI.The vertical dotted line shows the null.The null denotes no association with confidence/intention.A circle to the right of the null whose line does not cross the null indicates higher confidence/intention.A circle to the left of the null whose line does not cross the null indicates lower confidence/intention.A circle whose line crosses the null indicates no association with confidence/intention.The ORs compare subgroups of assigned hypothetical TPT scenario, 3HP vs. 1HP, as well as baseline variables.Multidimensional poverty was included as a bivariate: we compared the group which was poor or severely poor to the reference group which was defined as vulnerable or not vulnerable (i.e., not poor).TPT ¼ TB preventive treatment; OR ¼ odds ratio; CI ¼ confidence interval; 3HP ¼ weekly isoniazid-rifapentine for 3 months; 1HP ¼ daily isoniazid-rifapentine for 1 month; ART ¼ antiretroviral therapy.

Table 1 .
2 (Stata Corp, College Station, TX, USA) and R v4.1.2(R Project for Statistical Computing, Vienna, Austria).Confidence and intention Likert scale ratings were Participants' sociodemographic and clinical characteristics, overall and by TPT scenario assignment.
survey was approved by the School of Public Health Research Ethics Committee at the Makerere University College of Health Sciences (Kampala, Uganda), the Uganda National Council for Science and Technology (Kampala, Uganda), and the University of California San Francisco Institutional Review Board (San Francisco, CA, USA).

Table 2 .
Variation in preference for 1HP vs. 3HP by participants' characteristics.