Health literacy and type of education intervention predicting post-education metered-dose inhaler misuse

Abstract

Background:

Asthma and chronic obstructive pulmonary disease (COPD) guidelines recommend self-management assessment and inhaler education at all care-visits. Assessment is vital for identifying inhaler misuse. Whether age-related factors impede the efficacy of educational interventions for inhaler technique among older patients is unknown. We aimed to study factors associated with metered-dose inhaler (MDI) misuse pre/post-inhaler education among younger (< 65) and older inpatient populations (≥ 65).

Methods:

Adult inpatients with asthma or COPD enrolled across five studies between 2007 and 2017, who were eligible for, consented, and assigned to one of three education interventions (Brief Intervention [BI], Teach-to-Goal [TTG], Virtual Teach-to-Goal [V-TTG]) were included. Participants’ visual acuity, health literacy, and MDI technique pre/post education were assessed using validated assessments. Binary logistic regression was used to investigate factors that increased odds of inhaler misuse.

Results:

Across the five studies, 394 unique participants were enrolled with a mean age of 51.9 years (SD±15). There was no significant difference in baseline MDI misuse by age, vision, or health literacy levels. Post-education misuse use was lower among patients with better baseline MDI technique, those who received TTG or V-TTG education, and those with high health literacy. Neither age nor visual acuity were significantly associated with increased rates of misuse, although age was correlated with low health literacy.

Conclusion:

MDI education with teach-to-goal modalities is more effective than brief intervention; however, patients with low health literacy (disproportionately affecting older patients) may benefit less from these interventions. Further investigation into tailored inhaler education is needed.

1. Introduction

Millions of individuals live with asthma or chronic obstructive pulmonary disease (COPD) nationally [1,2], both of which carry high morbidity and mortality [1,2]. COPD is the fourth most common cause for disability [3], fourth leading cause of death [3], and the fifth leading cause of 30-day hospital readmissions [4]. Discharge is a vulnerable time for patients as they are often expected to shoulder the burden of their care and know how to self-manage their disease [5]. Guidelines recommend self-management assessment and education at all points of care, including hospitalization [6]. Hospital-based education can be a teaching moment and can be especially critical due for high-risk patients who have elevated rates of inhaler misuse [5–8].

Older patients have a much higher prevalence of COPD compared to younger patients, though they share the same burden of asthma [1,2]. Currently, the United States is experiencing growth in its geriatric population, which includes a growing population at-risk for COPD and/or asthma [8–10]. Unfortunately, inferior asthma outcomes are evident amongst older patients as they experience more asthma-related urgent care than middle-aged or younger adults [10,11]. Additionally, inhaler misuse is among one of the most significant self-management issues for both asthma and COPD disease control among older patients [7].

Therefore, it is imperative to determine whether there are unique barriers facing the geriatric population that would prevent them from using, or learning how to use, their inhalers correctly. Currently, inhaler misuse is preventing these patients from receiving the full benefit of their medications [12]. Health literacy is known to be a risk factor hindering older populations’ self-management of chronic disease and is known to affect 60% of adults over the age of 65 [13]. A number of studies have shown that older adults tend to have more cognitive and functional impairments and therefore, are more likely to be affected by low health literacy [14,15]. Additionally, health literacy has been noted to be markedly lower in older patients even after adjusting for differences in visual acuity, health status, and differences in cognitive function [16].

The objectives of our study were two-fold. The first was to evaluate intrinsic and extrinsic risk factors that lead to the misuse of metered-dose inhalers (MDIs) among an older, urban inpatient population diagnosed with COPD, asthma, or both. The second objective was to compare the impact of different MDI-based educational interventions on decreasing post-education MDI misuse among younger and older patient populations.

2. Methods

2.1. Study design

This was a secondary analysis using data from five previously conducted prospective interventional studies at two urban academic hospitals, which evaluated barriers to, and education for, effective MDI technique for hospitalized patients with asthma or COPD who were eligible for, approached, and consented to participate in the research studies. Both baseline MDI technique and the ability to learn correct MDI technique were assessed and compared among older patients (aged 65 and older) to their younger counterparts. Enrollment in these prior studies spanned from September 2007 to November 2017 [6,8,17–19]. Research staff reviewed electronic medical records to screen for adult hospitalized patients with a diagnosis of asthma or COPD, who used anMDI, and who did not meet any exclusion criteria (unable to consent for themselves, non-English speaking, etc.). Treating physicians of the participants gave assent for their patients to be approached, and hospitalized patients who agreed to participate in one of the studies provided written consent. These studies were approved by either the University of Chicago Biological Sciences Division or by both the University of Chicago Biological Sciences Division and Mercy Hospital and Medical Center Institutional Review Boards (15729A, 16900A, 11–0248, IRB12–1844, IRB13–1139).

2.2. Data collection

Research assistants administered surveys to collect participants’ self-reported demographics (e.g., age, gender, race). Vision was assessed using a bedside pocket Snellen Card; sufficient vision was identified as equal or greater than 20/50 vision in at least one eye (with vision corrective lenses such as glasses or contacts if they were needed and available) [20]. Participants with sufficient vision were then screened for health literacy, which was measured using the Short Test of Functional Health Literacy in Adults (STOFHLA) [20]. The STOFHLA measures reading comprehension by giving participants 7 min to read sentences from medical scenarios with missing words and selecting the correct word from a list of multiple-choice options to complete the sentence. Each correct answer receives a score of 1, for a total possible score of 36. Participants were classified as having adequate health literacy if they scored 23 points or higher and were classified as having low health literacy if they scored less than 23 points [20].

In all five studies, baseline and post-education MDI technique was addressed using a validated 12-step checklist; MDI misuse was defined as ≤ 9/12 steps correct. (Appendix) [6,8] The studies varied in terms of the educational interventions evaluated; however, in all cases participants assigned to receive Teach-to-Goal (TTG), Brief Intervention (BI), or Virtual Teach-to-Goal (V-TTG); in several of the studies the assignment was randomized [6,17,19]. In each study where more than one educational strategy was assigned, the educator was masked to the other strategy to prevent contamination [17]. TTG is a patient centered teach-back method using a trained educator who completes rounds of assessment and education using teach-back and demonstrations, repeating for a maximum of three cycles [6,8,17,19]. V-TTG is similar to TTG in that it uses the same approach using both assessment and demonstration repetition, though using a virtual platform [18,19]. By comparison, the “Brief Instruction” (BI) intervention is a less resource-intensive education method in which the participant receives a single round of simple verbal instructions without any pre- or post-education assessments or demonstrations [6,17]. TTG takes approximately three times longer than BI (~6 vs. ~2 min) to teach MDI technique [6]. All educational strategies evaluated for and taught used of a valved holding chamber (“spacer”).

2.3. Data analysis

All quantitative outcomes were summarized using basic descriptive statistics. Pearson’s χ² and ANOVA were used to test for bivariate associations to assess our primary objectives: (i) how age and age-related factors affect MDI misuse and (ii) the interplay of age (<65 or ≥ 65) and education intervention. A multivariate analysis was completed via a binary logistic analysis to study the following:

$$\begin{array}{l}{Y}_{\mathit{\text{misuse}}}={\beta }_0+{\beta }_1{X}_{\mathit{\text{under}}65}+{\beta }_2{X}_{\mathit{\text{pre}}\text{–}\mathit{\text{intervention}}\mathit{\text{score}}}+{\beta }_3{X}_{\mathit{\text{Educational}}\mathit{\text{Intervention}}}\\ +{\beta }_4{X}_{\mathit{\text{Visual}}\mathit{\text{Acuity}}}+{\beta }_5{X}_{HL}+{\beta }_6{X}_{\mathit{\text{sex}}}\end{array}$$

The misuse rate, under 65, and sex were all binary variables, pre-intervention score was continuous, whereas the remaining variables were categorical. All statistical analyses were performed using STATA 15 (College Station, Texas: StataCorp LP), and significance was determined by a p-value of <0.05.

3. Results

There were 426 unique participants enrolled across the inpatient five studies, of which 394 completed their participation and were included in this secondary analysis. Thirty-two participants were lost due to attrition, either because they were discharged prior to completing their study activities or because they withdrew their participation. The majority of participants self-identified as female (67%) and Black (90.6%). Almost 20% were 65 years or older; the mean age was 51.9 (SD±15.0) years (Table 1).

Table 1.

Demographic data of the patients in the study.

Variable, N (%)	All Enrolled (N = 394)
Female	264 (67.00)
Race
White	30 (7.61)
Black, African American	357 (90.61)
Asian	1 (0.25)
Other	1 (0.25)
Unknown	5 (1.27)
Age: Under 65	321 (81.47)
Visual Acuity *
Good Vision (Under 65)	270 (84.1)
Health Literacy *
Adequate HL (Under 65)	211 (68.3)

^*Good visual acuity was determined by the participant having at least 20/50 vision using the Snellen card. Adequate health literacy was determined by a participant receiving over 23 points on the STOHFLA. Both visual acuity and health literacy were analyzed as binary variables by age (over 65 vs. under 65).

Abbreviations: HL = health literacy.

3.1. Misuse of MDIs

Baseline Misuse:

Most of the 394 participants misused their MDIs at baseline (88.8%) with an average MDI score of 6.49 (Fig. 1). There was no difference in baseline misuse between younger (n = 321) or older age groups (n = 73; 88.8% vs. 89.0% respectively, p = 0.9), those with sufficient vision (n = 318) compared to those with insufficient vision (n = 76; 87.5% vs. 93.1% % respectively, p = 0.2), and those with high health literacy (n =239) compared to low health literacy (n =87; 87.4% vs. 90.8 %respectively, p = 0.3).

Fig. 1.

Mean MDI Scores pre- and post-intervention.

The histogram above depicts the spread of MDI scores both pre- and post-intervention as a whole. Average baseline score was 6.49, with a standard deviation of 2.29 (passing is 10 or higher). The average score post-intervention was 10.30 with a standard deviation of 2.00.

Post-Education Results:

Among the 394 participants included in the analysis, almost one-fifth were assigned to BI (63/394, 15.9%), just under a half were assigned to TTG (189/394, 47.9%), and the remaining participants were assigned to V-TTG (142/394, 36.0%). Post-education, there was a global improvement in participants’ MDI scores, with misuse rates decreasing to 29.4% and an average score of 10.3/12 (Fig. 1). However, when looking at the interventions independently, significantly more participants in the BI group (74.6%) continued to misuse their MDIs compared to the TTG or V-TTG (16.4% vs. 26.7%, respectively; p < 0.001) groups.

3.2. Potential risk factors

Vision:

Among the 394 participants included, about one-fifth (19%; 76/394) had low visual acuity. A larger proportion of participants over 65 had insufficient vision (25/73) compared to their younger counterparts (16%; 51/321; p < 0.001).

When stratifying the post-education misuse rates by participants’ visual acuity and educational intervention, both high and low visual acuity groups had significant differences in post-education MDI misuse rates. For patients with high visual acuity, misuse rates were lower for TTG (16.1%) and V-TTG (25.4%) than BI (71.1%, p < 0.001). Similarly, significant findings were found among those with low visual acuity (TTG 18.2%, V-TTG 38.1%, BI 90.9%, p < 0.001), indicating that both TTG and V-TTG were effective educational interventions even for those with low visual acuity.

Health Literacy:

Among the 394 participants included, nearly one-fifth of them had low health literacy (19.3%). Of note, just over one-fifth of the participants (21.6%) could not have their health literacy levels evaluated due to having inadequate vision for administering the STOHFLA; therefore, the effect of health literacy on post-educational MDI misuse could not be evaluated for this subset of participants. In comparing the cohort under 65 versus over 65, a larger portion of participants under 65 (76.2%) had adequate health literacy compared to those over 65 (57.1%).

While there was no significant difference inhaler MDI at baseline, after all education was completed, about half (46.7%) of those with low health literacy continued to misuse their MDIs, whereas only a quarter (23.9%) of those with high health literacy continued to do so (p < 0.001). Each educational intervention was also evaluated individually through the lens of health literacy. Among those who received TTG education, those with low health literacy continued to misuse their MDIs at 3 times the rate of those with adequate health literacy (33.3% vs 11.6%, p = 0.001). A similar pattern was noted for participants that received V-TTG; those with low health literacy misused their MDIs at 2.5 times the rate of those with adequate health literacy (50.0% vs 19.7%, p = 0.004). However, those educated with BI did not show a significant difference of misuse rates based on health literacy (90.9% vs. 65.8%, p = 0.1).

Age:

Within each age group, there was a significant difference in post-education misuse between interventions. For the older (over 65) cohort, those who received BI misused their MDI more (100%) than those who received TTG (26.2%) or V-TTG (37.7%, p < 0.001). For participants under 65, a similar pattern was noted in those who received BI (71.4%) vs. TTG (13.6%) vs. V-TTG (24.5%, p < 0.001). (Fig. 2).

Fig. 2.

MDI scores by age-group and intervention.

Pre- and Post-results grouped by education intervention and by age-group. For the Over 65 cohort, only V-TTG was able to bring the mean score of MDI above the minimum passing score of 10. On the other hand, both TTG and V-TTG were able to improve the average scores of MDIs for the Under 65 cohort.

However, when evaluating each education intervention individually among older and younger participants, misuse rates did not differ significantly. Among participants who received BI education, both those over 65 and under 65 continued to misuse their MDIs post-intervention (100% vs 71.4%, p = 0.10). For those participants who received TTG or V-TTG education, both those over and under 65 continued to misuse their MDIs; the difference between groups was not quite significant as they oscillated slightly above p = 0.05 (TTG: over 65–26.2% v under 65 13.6%, p = 0.05; VTTG: over 65 37.7% vs. 24.5%, p = 0.05).

Multiple variable risk factors:

A binary logistic analysis was conducted of 379 participants, with 15 participants omitted who did not have a health literacy score due to having either too low visual acuity scores to receive the STOHFLA assessment or simply not having their score recorded. The regression analyzed various factors that could affect the odds of MDI misuse, which was studied as a binary variable. Our results showed that having high health literacy and a higher baseline MDI technique score decreased the odds of MDI misuse after education by 0.29 (95% CI: 0.15–0.56) and 0.68 (95% CI: 0.58–0.79), respectively, compared to having low health literacy or a poor baseline MDI technique. In addition, participants who received either TTG or V-TTG education had a decreased odds of MDI misuse by 0.05 (95% C: 0.02–0.12) and 0.08 (95% CI: 0.03–0.18), respectively, compared to BI. Notably, being under 65, male, or having good visual acuity did not significantly decrease the odds of MDI misuse (Table 2).

Table 2.

Binary logistic regression odds ratios for predicting inhaler misuse.

Variables OR (95% CI)	All Participants (N = 326)
Pre-Intervention MDI Score*	0.68 (0.58–0.79)
Age
Over 65	Ref
Under 65	0.59 (0.27–1.31)
Education Intervention
BI	Ref
TTG*	0.05 (0.02–0.12)
V-TTG*	0.08 (0.03–0.18)
Visual Acuity
Low Visual Acuity	Ref
High Visual Acuity	0.44 (0.16–1.25)
Health Literacy (HL)
Low HL	Ref
High HL*	0.29 (0.15–0.56)
Gender
Women	Ref
Men	0.80 (0.42–1.49)

^*This indicates a significant association with inhaler misuse. Binary logistic regression was run to predict a participant’s inhaler misuse given age, gender, age-related factors, and the various educational interventions. The n in the regression is a smaller 326 compared to the 394 in this study due to missing STOHFLA scores of those whose visual acuity was too low to test for their HL. Two sensitivity analysis were run – by imputing the median and by multivariate imputation. In both instances, there was no significance found and the study team is confident that omitting those without HL scores did not negatively affect the outcome of the regression in any significant way.

Binary logistic regression was run to predict a participant’s inhaler misuse given age, age-related factors, and the various educational interventions. The smaller n in this regression is due to missing STOHFLA scores of those whose visual acuity was too low to test their HL. A sensitivity analysis was conducted via median imputation and multivariable imputation, but no significance was found.

Abbreviations: OR = odds ratio; CI = confidence interval; MDI = metered dose inhaler; BI = brief intervention; TTG = teach-to-goal; V-TTG = virtual teach-to-goal; HL = health literacy.

Due to the ineligibility of participants with insufficient visual acuity to complete the STOFHLA, we conducted a sensitivity analysis by imputing the STOHFLA scores for those who were missing this value. The sensitivity analysis allowed us to assess whether the data were skewed among those with sufficient visual acuity. Two imputation methods were conducted, including median imputation and multiple variable imputation method; no significant differences in the results were noted using either method (Appendix).

4. Discussion

In our analysis, health literacy was the only age-related factor that predicted MDI misuse, whereas receiving BI and low baseline MDI technique were most associated with ongoing MDI misuse post-education. Importantly, those who received more tailored educational interventions, such as TTG or V-TTG, were more likely to improve their MDI technique, particularly among older participants. As shown, older adults are more vulnerable than their younger counterparts based on their increased risk for having lower health literacy, which can impact MDI use and thereby hinder disease self-management.

For a general hospital population, this study demonstrated that TTG was the most effective intervention, followed closely by V-TTG. Both of these interventions showed significant correlation in improving participants’ MDI technique, whereas BI education did not significantly improve participants’ technique, and no participant older than 65 years had improved MDI technique with the BI approach. TTG and V-TTG are patient-centered interventions in that they utilize pre-education assessment and tailored education approaches. Further, due to the “testing effect” whereby memory is enhanced by retrieving information during the learning process, there is potential for longer retention of MDI skills using both TTG and V-TTG [21]. With both of these interventions, participants were provided up to three rounds of teach-back education [6,8,17–19]. This approach reinforces the steps mastered on the first-round while also correcting steps that they missed during subsequent rounds, allowing the education to be tailored to the participants’ knowledge and ability to learn. Alternatively, BI does not provide tailored feedback or opportunities for the participant to improve their skills during education. This is notable because BI is intended to mimic what is often done in outpatient settings, where there may be limited resources for educational interventions – a limitation of the TTG and V-TTG strategies.

Low health literacy also affected about a fifth of participants and disproportionally affected those over 65. Despite no significant difference in misuse between health literacy groups prior to education, participants with low health literacy continued to misuse their MDIs at twice the rate of those with high health literacy post-education. When each educational strategy was evaluated separately, it was notable that there were significant differences post-intervention between health literacy levels for those who received either TTG and V-TTG. For instance, among participants with high health literacy, misuse was much lower, at only one-third of those with low-health literacy. That said, participants with low health literacy still benefited from TTG or V-TTG interventions compared to BI education, as again, misuse of MDIs among those with low health literacy receiving TTG or V-TTG was lower, about half as low, as those who had received BI. This demonstrated that TTG and V-TTG, despite being the most effective strategies overall, may still not fully address the educational needs of those with low health literacy. This indicated that more patient-centered educational interventions and/or more frequency of educational interventions may better address the skills-based learning needs of patients with low health literacy to obtain necessary MDI technique.

Our model demonstrated that vision itself was not a predictor of MDI misuse, but there was a proportionately higher percentage of participants 65 years and older with inadequate vision. Both TTG and V-TTG are effective regardless of a participant’s vision levels. In contrast, those who received BI education had very high rates of misuse that were twice as high as V-TTG and triple that of TTG; this was particularly pronounced in participants with low vision, in which misuse rates after BI were quadruple that of those who received TTG or V-TTG education. This is likely due to the tactile nature of TTG and V-TTG education, in which the participant demonstrates and re-demonstrates MDI technique steps to the educator. Adding another layer of memory to these steps may help overcome the barrier that poor vision. Thus, despite TTG’s reliance on visual demonstration, TTG and V-TTG are still useful methods in improving a patient’s MDI technique over BI in participants with poor vision.

Despite that age itself was not significant in our model, there was an effect between the usefulness of the education in patients older than 65. However, when looking at the interaction between age and type of MDI education the patients received using ANVOA, there was no significant finding (Table 3). Therefore, we addressed age and education as independent factors. Age itself was not significant in our model; however, low health literacy was a larger portion of participants over 65 had low health literacy as compared to those under 65. Thus, older patients can be at higher risk for MDI misuse after education due to low health literacy. TTG demonstrated to be the most effective education amongst participants of either age cohort. However, even among those who received TTG education, those who were 65 or older were more likely to continue misusing their MDIs compared to those who were under 65. Age-related differences between BI and V-TTG were not significant. Regarding BI, the insignificant finding could be due to the intervention simply not being as effective for any age group; however, this could also pertain to the relatively small sample size of this group compared to TTG.

Table 3.

Regression results with interaction of age and educational intervention.

Variable	Coefficient Estimate	Std Error	P-value
Under 65	−0.32	2.14	0.15
TTG	−0.69	0.22	0.01*
V-TTG	−0.63	0.22	0.01*
Under 65 TTG	0.15	0.23	0.51
Under 65 V-TTG	0.17	0.24	0.47
F-Stat: 14.97 (df = 5; 320)

^*On the individual level, the type of education the participant received was a significant predictor of inhaler misuse at the 95% confidence level.

There was no significant interaction between being under 65 and the type of educational intervention that was received.

Abbreviations: Std Error = standard error; TTG = teach-to-goal; V-TTG = virtual teach-to-goal; F-Stat = comparison of the variance between two groups in ANOVA regression; df = degrees of freedom.

In addition to the small sample size of some sub-populations, other limitations are notable. For instance, an important limitation of this study is the lack of generalizability given that the patient population enrolled in these studies was comprised of mostly urban, female, Black patients. However, many studies of asthma and/or COPD have lower recruitment of women and/or Black individuals; therefore, these findings contribute to the evidence base which reflects the diversity of the US population. Another important note is that the analysis in this study was a secondary analysis of data across five studies, and the original data were not focused solely on older patients. Therefore, the power was limited for some age-specific analyses. For instance, physical limitations of the older participants were not assessed as this was not a measure collected in the preliminary studies. Future studies should ensure that older participants are enrolled to produce robust data and sufficient power for age-specific analyses. Another limitation was that all of these studies used the STOHFLA to assess health literacy, which requires adequate vision for administration, prohibiting those with poor vision from assessing health literacy. Given that participants with poor vision comprised nearly ten percent of the total population, obtaining their health literacy would have been valuable to the analysis of these data. In the future, use of a validated tool that can evaluate health literacy without need for visual acuity could address this limitation. Finally, since the time these studies were conducted, additional MDI type inhalers have become available, for which the checklist used in these studies may or may not be a reliable assessment. Future studies including these types of MDIs would need to update the checklists to accurately reflect the steps needed for appropriate use of these MDIs.

In summary, our study reinforces how widely prevalent MDI misuse is, despite respiratory inhalers being the mainstay for delivery of critical medication needed for the management of asthma and COPD. It is imperative to recognize that most hospitalized patients require inhaler education to reduce rates of inhaler misuse and the hospital setting should be used to provide this education. TTG education was found to be the most effective, followed closely behind by V-TTG, indicating that V-TTG may be a useful tool in the current era of increased telemedicine visits and social isolation. Our data also showed that low health literacy may still be a barrier to learning proper inhaler technique, even using patient-centered interventions as they are currently implemented. Therefore, there is a need to develop and test new tailored, patient-centered inhaler education strategies and/or ensure multiple time points for education are provided for patients with low health literacy. Future work should explore the relationship between age, health literacy, and vision in more detail with a larger proportion of patients 65 and over, compared to our data. Further development with other multivariate logistic models may help explain the overlap between age, vision, and health literacy to better understand this finding. Lastly, creating patient-centered education specifically for patients with poor visual acuity and poor health literacy should be a priority, as they are the most vulnerable to post-education inhaler misuse and thus continued clinical consequences from inadequate medication administration.

Abbreviations:

BI

Brief Intervention

COPD

Chronic Obstructive Pulmonary Disease

HL

Health Literacy

MDI

Metered-Dose Inhaler

MMSE

Mini-Mental State Examination

OR

Odds Ratio

STOHFLA

Short Test for Functional Literacy in Adults

TTG

Teach-to-Goal

V-TTG

Virtual Teach-to-Goal