Study design
The semi-structured interviews were part of a multi-center, prospective, descriptive cohort study on the natural course of MPS IIIA. Standardized clinical and observer-reported neurocognitive, developmental, and behavioral measures were captured at baseline and every 6 months for up to 24 months. These measures included the Bayley Scales of Infant Development, Third Edition (BSID-III).
All parents of children included in the study were invited to participate in interviews at baseline and at Month 12; participation was voluntary. Parents needed to have completed a baseline interview to be eligible for the Month 12 interview. Interviews were conducted in France, the Netherlands, Germany, and the UK and took place within three months following the baseline and Month 12 site visits.
Up to 25 MPS IIIA children were anticipated to be enrolled. To be eligible, children had to be nine years of age or younger, have a documented MPS IIIA diagnosis, be medically stable, and have never received an investigational product for MPS IIIA.
Interview conduct and transcription/translation
Interviewers were trained on the background of MPS IIIA and the natural history study, the use of the interview guide, and interviewing methodology.
Face-to-face parent interviews took place at a location of the parent's choice. All interviews were done by local interviewers in the parents’ native language; all Month 12 interviews except for two were conducted by the same interviewer as the baseline interview. Both baseline and Month 12 interviews lasted one hour to one hour and half.
During the interviews, parents were asked to identify which symptoms were the most challenging to manage; to discuss how caring for their child and managing their symptoms affects the family; and to describe how their child's symptoms have changed since diagnosis for the baseline interview or since the baseline interview for the Month 12 interview.
Interviews were audio-recorded with parents' consent and were transcribed word-for-word in the source language and de-identified. At baseline, German and Dutch transcripts were translated into English for analysis; English and French interviews were analyzed in the source language. At Month 12, no translations of transcripts were done; all interviews were analyzed in the source language.
Qualitative analysis
Conceptual models describing symptoms and impacts of children and parents
The qualitative analysis of the baseline interviews that led to the development of the conceptual models was an exploratory, iterative process. Concepts were organized into themes representing higher-order categories of analysis. The themes were then sorted into one of two conceptual models: either the model for children’s symptoms and impacts or the model for parent impacts.
Analyses for conceptual models were performed using a qualitative software package, ATLAS.ti Version 8. ATLAS.ti is designed to facilitate the storage, coding, analysis, and retrieval of qualitative data [8].
Qualitatively-derived severity stages
The qualitatively-derived severity stages were developed exclusively based on the symptoms described by parents of children with MPS IIIA during the qualitative interviews. No other data on children (such as age, gender, country, clinical or questionnaire data from the natural history study) was used to inform their development.
The development proceeded by inductive reasoning using the conceptual models as a starting point. The most common symptoms found in the sample were considered as the core of the disease and were grouped in Stage 1. Stages 2 and 3 were conceived as a progression of symptoms already present in Stage 1 as well as the accumulation of new symptoms. Stage 4 was conceptualized as reflecting the most severe, and often most uncommon, symptoms found in the sample. As a whole, the stages were developed to reflect MPS IIIA symptom progression (the worsening of a current symptom, such as loss of language), symptom accumulation (the addition of new symptoms alongside prior ones, such as motor function issues in addition to trouble self-feeding), and the emergence of critically severe symptoms (such as trouble with nutrition).
Following the definition of these four stages, each child was sorted based on an in-depth review of the child’s baseline transcript and an examination of the codes that were used in that transcript. Children were sorted into stages in a holistic, qualitative fashion. In other words, a child did not need to systematically have all of the symptoms in a stage to be sorted into that stage description.
Mixed methods: triangulating qualitatively-derived severity stages with BSID-III Development Quotient and child’s calendar age
We compared the sorting of children according to the qualitatively-derived severity stages, the child’s calendar age at baseline (6 years and older vs. less than 6 years old), and the child’s baseline BSID-III development quotient (DQ). DQ cognition was defined as mild/normal if 75–100%; moderate if 50–75%, and severe if < 50%. DQ is calculated by dividing the development age into which test scores place the child by the child's chronological age and multiplying by 100 [9]. The statistical associations were tested with Mantel-Haenzel Chi2 exact test.