The I-Score Study: Developing a New Patient-Reported Tool for the Routine HIV Care of Patients on Antiretroviral Therapy

This 24-month study aims to develop and validate a theoretically founded patient-reported outcome measure for use in routine HIV care with patients on antiretroviral therapy (ART). Eight clinical sites will be involved (4 in Canada and 4 in France). The measure will centre on patients' reported barriers to ART adherence and will be adapted for administration via electronic tablet. The study has four main steps. The psychometric validation of the tool and core of the study will be conducted in Step 3. The Observational -Prospective design applies to the field test and retest phase of Step 3. The four steps are nevertheless described below.

Step 1 -Development of the measure's conceptual framework (Start: October 2015)

1. A scoping review of existing HIV-specific patient-reported outcome measures
2. A meta-ethnography of qualitative research on HIV-positive persons' reported barriers to antiretroviral therapy (ART) adherence
3. A content analysis of qualitative interviews with HIV-positive patients on ART about their perceived barriers to adherence (n = 30 patients) The meta-ethnography and content analysis will serve to develop the conceptual framework on which the tool will be based. It will not only define the tool's conceptual domain but outline its relationship to other concepts that will be verified statistically as a part of the tool's validation. Once the framework is completed, the measure will be specified and a bank of items will be generated.

Step 2 -Linguistic validation of the preliminary measure The tool's linguistic validation will involve the professional harmonization of the English and French versions and the creation of a universal French version adapted to all French sites.

Step 3 - Exploration and validation of the measure's psychometric properties

1. During the pilot phase, we will administer our measure and other measures to a small sample size (n = 80) and conduct a number of " cognitive interviews " with participants (n = 8). This will mainly help improve our measure before the field test phase (e.g., by examining missing data, incoherent responses, any problems with question comprehension).
2. The field test (n = 160) will be done with the revised measure of the pilot phase and will include exploratory and confirmatory analyses to examine, among other aspects, the correspondence between the collected data and the conceptual framework developed in Step 1. Here we will also examine the sub-dimensions of our measure and verify the concurrent (e.g., viral load, intention to adhere to ART), convergent (e.g., beliefs about antiretroviral medication) and discriminant (e.g., quality of life) validities of the tool.
3. For the retest phase, the field test participants will be recalled either 1 week or 4 weeks later for another round of data collection to assess the measure's reliability. In this phase an additional measure not included in the previous phase will be included in the questionnaire (adherence to ART). This new measure will be used to verify the predictive validity of the tool.

Step 4 -Pilot the clinical use of the instrument Here, the finalized tool will be applied in a clinical setting with a small sample of HIV-positive patients under ART (n = 24) and their clinicians (n = 8).The aim is to collect data on the tool's perceived utility from the perspective of both populations in order to provide an initial assessment of the feasibility of the tool's wider implementation.

Statistical analyses planned Statistical analyses of the quantitative pilot test phase (n = 80)

• Response coherence
• Analysis of item score distributions
• Analysis of missing data
• Principal components analyses Statistical analyses of the field test phase (n = 160)
• Exploratory analyses
• Confirmatory analyses
• Verification of the measure's convergent, discriminant and concurrent validities Statistical analyses of the retest phase (a random sample of n = 160)
• Principal components analysis and internal consistency through Cronbach's alpha in order to verify whether the factors obtained at the retest phase are the same as those obtained from the field test phase
• Descriptive statistics and multivariate analyses will be carried out in order to compare score means and to verify whether the factors associated with the I-score are the same in phases 2 and 3 Statistical analyses of the Pilot of the instrument's clinical use (n = 24 patients; n = 8 HIV clinicians)
• The statistics produced for this step of the study will be descriptive (e.g., percentages, averages) and comparative. For comparative analyses between patients and their doctors (e.g., on the tool's appreciation), tests will be adapted according to whether the variables used are continuous (e.g., t-tests) or categorical (e.g., chi-square).
• Furthermore, patient and clinician data will be analyzed with a multi-level model to examine, for example, if patient appreciation (or non-appreciation) of the tool is associated or not with differences in the clinicians' perceptions (estimation of individual and structural factors associated with appreciation or non-appreciation).