Efficacy of a structured audio-visual asthma care session during GP visit in Australia: a single GP centre intervention trial

Study type

This was an unblinded, randomised intervention trial.

Study setting

We conducted this study at Coomera Town Medical Centre, Queensland, Australia. This primary healthcare clinic provides bulk bill services to the community and treats approximately 1000 patients of all ages per month. They provide comprehensive health care to all individuals and their families in the community and specialised care for specific age groups such as children, women, and the elderly.

Study participants

We included participants aged 6–60 years, of either sex, confirmed as having asthma for the first time by an attending physician while visiting the clinic, English speaking, and who gave consent to participate in this study (Table 1). Asthma was diagnosed by a physician based on history, clinical examinations, and spirometry following NAC guidelines (National Asthma Council Australia 2022a). We chose this age group for our study participants as confirming the diagnosis of asthma below 6 years is difficult due to difficulty performing lung function testing (e.g. spirometry) below 6 years (Ducharme et al. 2015). In Australia, ~80% of older people (aged 65 years and above) have at least one chronic health condition (AIHW 2021). We limited our participants’ ages to 60 years to avoid possible chronic health conditions, including a window period of 5 years. We also had some exclusion criteria for selecting the participants, including having a known co-morbidity and other physiological or pathological condition that may have different pathophysiology of asthma and may hinder the intervention (Table 1). We purposefully excluded people with co-morbidity and pregnancy or lactation as these pathological and physiological conditions may interfere with our study intervention. A person with an unstable mental condition might not follow our intervention as it consisted of educational materials. We excluded tourists or people who were not residents of the area near the study site, as we had a follow-up visit on/after 90 days of the enrolment, and these people had a high chance of being lost to follow-up. We also purposefully excluded Aboriginal and Torres Strait Islander peoples as they have a significantly higher prevalence of asthma than the non-Indigenous population in Australia (AIHW 2020). They often presented with co-morbidities (Jenkins et al. 2009). Moreover, our proposed intervention is for the overall population. Aboriginal or Torres Strait Islander people need culturally tailored intervention because they respect cultural differences, build trust, and improve engagement with health services (Bainbridge et al. 2015).

Sample size

For sample size calculation, we considered a prior study by Reddel et al. (2015) where they calculated the average asthma score from a nationwide representative sample in Australia. The average asthma score was 19.20 [standard deviation (s.d.) 5.28] in that study. We assume, in our population, the baseline mean asthma score to be the same (but due to the small sample size, the s.d. would be double), and our proposed intervention may increase the asthma score by 20% from the baseline. To determine the appropriate sample size for our study, we aimed to detect a medium effect size with a power of 80% at a 5% significance level. Based on prior literature and theoretical considerations, we estimated an effect size (Cohen’s d) of 0.5. Therefore, the required sample size per group is approximately 78 participants in each group (156 in total).

Randomisation

After enrolment of the study, we randomised the participants in a 1:1 ratio to the intervention arm and control arm. We followed the permuted block (size of 4) technique (Broglio 2018) for randomisation to make the allocation equal at any time point of the study. The researchers of this study had no role in the randomisation and allocation process. Dedicated study nurses of the general practice clinic performed this task.

Study interventions

After randomisation and allocation of the study arms, the control group received the standard management for asthma by a GP. The standard intervention consisted of medical treatment as per their need, advice as to the use of a metered-dosed inhaler (no powder, for children with spacer), and general asthma education by the attending GP. The intervention group received one additional educational session by a trained study nurse in addition to the standard management for asthma by the attending GP. The education materials consisted of an audio-described video display in English adopted from the NAC (National Asthma Council Australia 2022b) and Nucleus Medical Media (Nucleus Medical Media 2022). The Nucleus Medical Media’s video described what asthma is, its symptoms, and its triggering factors. The NAC’s video demonstrated the standard procedure for using a metered-dose inhaler with and without a spacer. The graphical presentation of these videos showed how the children should use inhalers with a spacer, and how an adult can use an inhaler properly without a spacer. However, the NAC’s graphical presentation videos that we used in our trial have now been updated by human demonstration as role play videos. The total length of these two (Nucleus Medical Media and any of the two NAC videos) videos was ~5 minutes. We demonstrated the video in a dedicated room at the clinic to avoid cross-intervention with the control group participants. Other than these videos, no other health education related to asthma control was provided. However, all participants received general asthma education from the attending GPs as their standard practice during the GP consultations. In the case of child participants, the attending parents/carers received the study intervention on behalf of their children.

Data collection

We used a pre-tested case record form to collect all socio-demographic and clinical data of the participants. On enrolment, we collected several sociodemographic and clinical data from the participants. We considered the Asthma Control Test (ACT) (see Supplementary Table S1) as our outcome variable. We adopted this scoring system from the Australian Asthma Handbook which was originally developed by Asthma Australia, in association with the NAC and GlaxoSmithKline (Asthma Australia 2021). This scoring system has five asthma questions, five responses, and an overall score ranging from five to 25 points. Fewer points indicate the worst control of asthma. We assessed asthma scores for all participants on day zero (just after enrolment, before receiving any intervention) as our baseline score and on day 90 as our endline score. We set a 90-day follow-up to assess our endline measurements as it is recommended to have follow-up visits every 3–6 months for participants with asthma according to their severity/symptoms (Horak et al. 2016). Attentiveness to the intervention media was measured by direct observation by a dedicated study nurse.

Statistical analysis

Statistical analysis was performed on data from all randomised participants in the study on an intention-to-treat (Tripepi et al. 2020) basis. Descriptive statistics of the baseline characteristics are presented in percentage, mean, and standard deviations as appropriate. We compared baseline characteristics of both treatment and control arms and presented their P values. To assess the treatment effect, we conducted a ‘longitudinal analysis of covariates’ (Twisk et al. 2018) after adjusting for potential confounders (see Results section) to assess the independent effect of the intervention on asthma. In our study, we considered confounding factors that may alter the effect of the association between the intervention and outcome such as paticipant’s age, sex, carpet use at home or bedroom, exposure to smoking, furry pet at home, gas stove, and adverse neighbourhood characteristics (such as neighbourhood cleanliness; having adequate parks, playgrounds, and play spaces; neighbourhood traffic load). Several studies have reported significant associations of these factors with asthma (Shahunja et al. 2021). We also adjusted the baseline characteristics. Results are presented with mean changes (difference between endline score and baseline score) of outcome variables across the groups and the difference between their differences. A P-value less than 0.05 was considered to have statistical significance. All statistical analyses were done using a personal computer and statistical software STATA (StataCorp, USA, ver. 17.0).

Ethics

Participation in this study was voluntary. We only enrolled a participant in this study after having written informed consent. However, the participants had the full right to withdraw from the study at any time until the completion of this study. Participants received standard management for asthma by a GP from the centre if anyone denied taking part in this study. This study was approved by the Human Research Ethics Committee of the University of Queensland (Approval number: 2019002311).

Participants

From February 2020 to August 2021, we screened 9820 people for the eligibility criteria of this study and identified 595 (6%) potential participants with asthma, 156 were enrolled in this study and the other 439 were excluded due to one or more exclusion criteria as mentioned earlier. The regular pace of enrolment of the participants was hampered in April and May 2020 due to the outbreak of COVID-19. After randomisation, we assigned 78 participants to the intervention group and another 78 participants to the control group (Fig. 1). We had only one participant drop-out who was in the intervention group. We could limit the dropout to almost nil as we had strong inclusion and exclusion criteria, participants came from the usual catchment area, and there were regular patients of this general practice centre. A consort flow chart has described the enrolment process of our study (Fig. 1).

Fig. 1.

CONSORT flow diagram.

Baseline characteristics

Although age-stratified randomisation was not performed, 21% of our study participants were children (6–17 years). Overall, 60% of the participants were female. Half of the participants were overweight/obese (body mass index ≥25), and half of the overall participants had a furry pet at home (Table 2). Some of these characteristics were comparable across the groups (intervention vs control). However, the percentage of participants who did not have a spacious and clean neighbourhood, furry pet, carpet, or gas heater in the house and were exposed to active and passive tobacco smoking were not comparable across the groups that may occur by chance. Details about baseline characteristics are described in Table 2. The baseline asthma score was approximately 19.6 (±4.8) for overall participants (Table 2). Although our target was to measure the endline asthma score on day 90 of the enrolment, the mean days for endline measurement for the intervention group were 96.1 days and 98.2 days for the control group, which was statistically not significant (Supplementary Table S2).

Seasonality of the enrolment

June and July seemed the peak season for maximum enrolment of participants with asthma. Forty-five percent of the total participants were enrolled during these 2 months of the year (Supplementary Fig. S1). The highest enrolment in these months was possibily due to winter and the outbreak of COVID-19 as people visited the general practice centres more often due to respiratory symptoms at this time. On the other hand, we had fewer enrolments in April and May 2020, perhaps due to repeated lockdowns, quarantine, and relatively less movement of people outside the home.

Compliance with the intervention

In our study, most of the participants in the intervention arm showed good compliance with our intervention materials. More than 90% of participants watched the educational videos (intervention materials) at full length. Eighty-eight percent of participants watched the videos with full attention (Fig. 2).

Fig. 2.

Compliance with the intervention videos.

Intervention effects

Irrespective of age, intervention group participants had an increment of 1.9 units in their ACT score at the endline compared to the baseline. However, the control group had a decline in this score by ~0.5 units. Thus, the intervention group achieved a ~2.5-unit better increment in asthma score compared to the control arm, which was statistically significant. Children also showed a remarkable increment in ACT score in comparison to the control group in the age-stratified analysis. However, it did not show statistical significance due to a lack of sample (Table 3).

Strength and limitations

The main strength of our study is that it was a randomised intervention trial. Due to randomisation, we could avoid selection bias and control most of the baseline characteristics of the participants. Although asthma diagnosis and participants’ selection were made by the attending GP, there were dedicated nursing staff to deliver the intervention. The intervention video was short, simple, and in the participants’ native language, which could keep the participants’ attention while playing the video. The nursing staff communicated with the participants and gave reminders before 90 days to take the endline measurement. Attrition was limited to only one participant, which helped to retain our expected sample size and the power of the analysis. Asthma diagnosis by a physician also gives our study a precise technique to identify the study participants.

However, there are a few limitations as well. The study could not assess the participants’ compliance with the intervention while they were at home. The study was conducted during the COVID-19 pandemic. As such, repeated restrictions, including restricted movement and using face masks, might also influence the outcome of asthma. However, due to randomisation, we assume both groups were influenced equally so it did not affect the study outcome. We excluded several potential participants with asthma from participating in this study due to the presence of other co-morbidities, as outlined in our exclusion criteria. This exclusion may limit the generalisability of our findings. This study was conducted in a single general practice centre. Our sample population may represent only those people who regularly attend that clinic.