*2.4. Data Analysis*

Data from the focus groups were audio recorded, which were later translated from Spanish to English and transcribed by a bilingual member of the U.S. research team. Translations were then checked by a second bilingual member of the research team, an independent bilingual reviewer (not part of the research team), and a bilingual member of the data collection team in Mexico. The English language transcripts were entered into NVIVO, a qualitative software program (QRS International, Melbourne, Australia) [34]. We followed the steps outlined by Braun and Clark [35] in first familiarizing ourselves with the data, open coding the transcripts inductively to note salient and reoccurring patterns, and categorizing these patterns into collapsed and meaningful themes via a preliminary codebook. That is, the codebook was developed in order to operationalize themes which had arisen from the data in a manner of frequency, specificity of discussion (e.g., examples and elaboration), and emotionality. This was in alignment with Krueger and Casey's [36] best practice recommendations for analysis of focus group data. As is common in qualitative research (see Crabtree & Miller [37]; Padgett [38]), we were sensitized through our preliminary coding process to the overlap between initial themes and those outlined by Massey et al.'s [29] framework for studying adolescent HL (e.g., our theme pertaining to "barriers to positive relationships with doctors" with theirs of "patient–provider relationships"), and thus moved from an inductive to a deductive codebook in order to assess fit of our data with this useful framework and to outline findings in a salient context to healthcare literature. (For an example of this inductive to deductive template approach, see Linton & Rueda [39].) The fit of our preliminary codebook to Massey et al.'s framework was excellent to the extent that only the titles and order of themes were changed. An inter-rater reliability kappa was calculated on the final codebook with a fourth and independent researcher who coded the data in its entirety. We had excellent reliability between the initial coding and that of the independent researcher who was unfamiliar with the study or its aims (*κ* = 0.93). Throughout the analysis process, the rigor of the study was further enhanced by use of observer triangulation and member checking. That is, three separate researchers corroborated to analyze the data [38] and the final themes of the study were sent to the school psychologists in Mexico for their feedback. Their feedback resulted in some minor changes to wording within the results section to capture accurate translations stemming from local colloquialisms. We report the following themes from our data in alignment with Massey's [29] framework of adolescent HL: (1) navigating the health system, (2) patients' rights and responsibilities, (3) preventative care, (4) information seeking, and (5) patient–provider relationships.
