A modified Delphi process was completed with eight Indigenous experts, and focus groups were conducted with 18 Indigenous women about culturally safe measures for Infant respiratory health. To our knowledge, this is the first consensus-based study on measures for detecting respiratory illness in Indigenous Australian infants. Measures that reached consensus included 15 measures at birth, 17 measures from 1 to 6 months of age, and three questionnaires to be used at 6 months of age. The preferred mode for data collection differed for the different time points. Consensus was reached that birth measures should to be collected via a hospital discharge summary, 1 to 6 month measures via parent report with mode decided by woman i.e. phone call, mobile phone application, or online survey and 6 months of age measures collected using parent report questionnaires completed with a trusted health professional in conjunction with clinical notes.
All measures at birth (15/15) were accepted for inclusion. The high rate of inclusion might be due to the standard nature of measures and minimal burden to participating women, as data would be collected via discharge summary. Measures collected from 1 to 6 months consisted of mainly acute respiratory symptoms (7/17) and health care utilisation (6/17). The result was a monthly self-report survey with a minimum of 13 and maximum 17 items. Of the 17 items, only five items were accepted for inclusion in rounds two and three. These five items were ‘wheeze/whistle’, ‘moist/wet/cough’, ‘dry cough’, ‘reasons for seeing a doctor’ and ‘change in exposure to tobacco smoke’. The remaining 12 items were included based on a rule to combine ‘very essential’ and ‘somewhat essential’ votes. The five initial items may indicate items the panel considered most essential to measure; they also are well aligned with the literature. Wheeze is the most reliable symptom to detect asthma (36) and wet cough for bronchiectasis (4, 6, 37). Seeing a doctor may indicate severity, and exposure to environmental tobacco smoke during infancy doubles the risk of hospitalisation for respiratory illness in infancy (38), so an important variable to collect.
Two potential respiratory questionnaires for use at 6 months of age were presented to the panel, 1) a 50-item questionnaire developed for British infants (39) and 2) a 20-item questionnaire adapted for Torres Strait Islander infants (32). It was consistent between the panel and women in the focus groups that Torres Strait Creole is not suitable for most Indigenous women, though a questionnaire with fewer items was preferred. The language of the 50 item questionnaire was largely understood and accepted by women, which is unsurprising as it stems from the ISAAC protocol which has been tested in 97 countries (40). The 50-item questionnaire was ultimately shortened to 33 items based on feedback. A developmental screening measure, the Ages and Stages questionnaire (41) as well as the adapted version for remote Indigenous communities (34) were also presented to the panel. Interestingly all panel members indicated inclusion of a measure on child development, when not typically measured in studies on respiratory health. The strong interest to include a developmental measure raises the question of what other measures may be important, and perhaps more meaningful to Indigenous communities. Other less commonly reported measures in child respiratory studies include child parent quality of life (42, 43) and child functioning (44).
This study had several limitations. The involvement of Indigenous women was limited. Women participated in one focus group to provide feedback on one type of measure (6 months of age respiratory questionnaires); we did not obtain final feedback from women on changes made to the questionnaire recommended by the expert panel (removal of 17 items). The measures identified in this study may be more confidently used if greater end user involvement had occurred (45). While we strongly acknowledge the importance of end-user involvement, the focus here was to gain expert consensus from Indigenous academics and clinicians on essential respiratory measures, future studies should place emphasis on pre-testing the identified measures with end-users from a range of communities. A second limitation was that findings may not be generalisable to the diversity of Indigenous peoples of Australia. While panel members were from different regional, remote and urban communities, the number of panel members was relatively small (8) and women were from NSW communities only. A third and important limitation was that the measures identified focus on a rather short period in a child’s life, birth to 6 months of age. The 6 months age range was of focus as it is the follow-up period of the larger trial. As many chronic respiratory illnesses only develop later in childhood and are uncertain in infancy, e.g. asthma and bronchiectasis, accepted measures for use throughout childhood are needed. Lastly, if further rounds of consensus were completed the number of items may have been reduced, which can result in higher response rates for trials (46). An important consideration to be examined if pre-testing of measures.
The strength of this study was the engagement of Indigenous experts from several disciplines to work together and identify a comprehensive set of respiratory measures in the context of cultural safety for Indigenous infants. Knowledge was generated with Indigenous academics, clinicians and women to optimise the cultural safety of data collection in a trial examining infant respiratory outcomes. The measures identified are for a number of time points in the first 6 months of life using a range of sources (medical records, parent report and observation). A range of sources is important given the known pitfalls of relying on any one of these sources alone (15).
A modified Delphi process may be a useful method to systematically involve Indigenous people in decisions for trials. The Delphi has been used in other areas of Indigenous health research including to develop mental health guidelines (21) and data collection strategies for maternity experiences (22). Other high-level consultative methods to develop measures for use with Indigenous people have also been used. A recent example is the development of a survey for the Mayi Kuwayu Study, a national longitudinal study on adult Indigenous Australian well-being (47). Consultation was completed with 165 Indigenous peoples attending 24 focus groups across Australia from 2014 to 2017. Pilot testing of the survey was completed with 160 and 209 Indigenous participants. A second example is the Healing the Past by Nurturing the Future study, a study in part to develop a measure to identify complex trauma experienced by Indigenous parents (48). Consultation includes four large-scale co-design workshops across three States with Indigenous parents, service providers, community leaders, researchers and wider community members. Comprehensive consultation is expected from conception to conclusion in research with Indigenous peoples (49). With varying methods and approaches for consultation, a Delphi methodology is one approach that can provide a systematic, transparent and feasible process for expert consensus in trials.
The Indigenous panel that participated in the consensus process made two important unexpected recommendations that may aid more accurate data collection and increase recruitment and retention in trials. The first was to provide education to participating families and health providers on respiratory symptoms and management pathways. This recommendation aligns with a recent qualitative study with 40 Indigenous community members reporting 70% considered chronic cough normal in children (50). By providing culturally appropriate definitions on respiratory terms such as wheeze and wet cough, and information on the importance of seeking treatment, the accuracy of parent report may improve and lead to better disease detection and optimal treatment (51). The second recommendation was to provide adequate follow up of participating infants. Cough guidelines recommend children aged 14 years or less with a chronic cough of 4 weeks should have a chest radiograph and spirometry test (when age appropriate) (52). In research studies on infant respiratory health, we have opportunity and ethical responsibility (49) to ensure that children receive adequate treatment during and on study completion. Studies designed with a reciprocal approach including assured access to quality treatment may improve retention rates, as in a recent study on incidence of respiratory illness in Queensland (12).
This is a preliminary step in developing a set of standard measures to detect respiratory illness in community based Indigenous infants. Future research is needed to test the validity and the reliability of the identified measures for use in trials and practice. Additional considerations for pre and pilot testing these measures may include information for families to combat the normalisation of respiratory illness (53), flexible mode of delivery given the many other needs and problems Indigenous families experience (20), and trusted and skilled interviewers to ensure cultural safety.