Research Article - Neuropsychiatry (2017) Volume 7, Issue 6

Supporting Older Adults with Mild Cognitive Impairment in Medication Taking: A 2-year Longitudinal Study

Corresponding Author:
Angela Yee Man Leung, PhD
Centre for Gerontological Nursing, School of Nursing, The Hong Kong Polytechnic University, Hong Kong, China
(+852) 27665587
Fax: (+852) 23649663



Objective: Community dwelling older adults with mild cognitive impairment (MCI) encounter challenges in many daily activities including medication taking. With inadequate health literacy, many of them have difficulty in understanding the instructions on drug labels and make decision on medication taking. This study aims to assess the effect of two interventions, verbal advice plus educational leaflet (VAEL) and comic books (CB), on medication taking among persons with MCI.

Methods: This is a two-year multi-site longitudinal study with pre-and-post design and four-point measures (Time 1, 2, 3, and 4). Subjects were aged 60 or above, scored <22 in Chinese version of the Montreal Cognitive Assessment (MoCA), and were taking ≥ 2 drugs regularly. Health literacy was measured by the 24-item Chinese Health Literacy for Chronic Care (CHLCC) and medication adherence was measured by the 8-item Morisky Medication Adherence Scale (MMAS). Generalized Linear Mixed Models (GLMM) and Generalized Estimating Equation (GEE) were used.

Results:Among the 131 subjects, the majority (62%) were females and mean age was 82.5 (SD 6.5). One-third of them received primary education (31%). Controlling age, there was a significant increase in CHLCC scores after the use of CBs (T4-T3: coefficient=2.742, p=0.003;T4-T1: coefficient=5.073, p=0.014). On the other hand, comparing to the baseline at T1 and controlling age, the chance of moving MMAS level one level higher than its original level at T4 was high (O.R.=3.316, S.E. 0.2979, 95% CI 0.650-1.747, p<0.001). Similar improvement in MMAS level was observed at T3 and T2.

Conclusion: The two interventions (VAEL and CBs) seem to be a good strategy to improve health literacyand medication adherence among community dwellers with MCI. More research is warranted to support medication taking in this specific population.



This study includes two outcome measures, medication adherence and health literacy. Medication adherence was measured by the 8-item Morisky’s Medication Adherence Scale (MMAS) – Chinese version [13,14]. The total MMAS score ranged from 0 to 8. There were 3 MMAS levels: MMAS level 1 referred to the MMAS score lower than 6, indicating ‘poor adherence’; MMAS level 2 referred to the MMAS score 6 to 7, indicating ‘medium adherence’, and MMAS level 3 referred to MMAS score 8, indicating high adherence. Health literacy was measured by the 24-item Chinese Health Literacy for Chronic Care (CHLCC) [15]. The total CHLCC score ranged from 0 to 48. Higher total score indicates higher health literacy level. In addition, we also assessed activity of daily living (ADL) and depressive symptoms. Activity of daily living was measured by Barthel Activity of Daily Living (ADL) scale. The higher the Barthel ADL score, the higher the capacity to carry out ADL activities independently [16]. Depressive symptoms were measured by Geriatric Depression Scale (GDS). GDS score higher than 8 indicates the likelihood of suffering from depressive symptoms.

▪ Samples

Only those who were aged 60 or above, were taking ≥ 2 drugs regularly, and with mild cognitive impairment (MCI) were included in this study. MCI was defined as ‘scored <22 in Chinese version of the Montreal Cognitive Assessment (MoCA) [17]. For those with 6 years or less in education, 1 score was added [17].

▪ Interventions

Two interventions are included in this study: Intervention 1 is a 90-minute intervention using verbal advice and a written educational leaflet (VAEL) while Intervention 2 is also a 90-minute intervention using health literacy-oriented comic books (CBs). Each intervention was carried out in two home visits to each subject. Medical/ nursing/pharmacy students were trained as volunteers to carry out these interventions according to the standard protocols. Both interventions aim to improve older adults’ health literacy and medication adherence.

▪ Schedule of measurements

A baseline assessment was made before the implementation of Intervention 1 using VAEL and this is considered as Time 1 (T1). A post-intervention assessment was made after VAEL at Time 2 (T2). One year later, another assessment was made at Time 3 (T3) before the implementation of Intervention 2 using CBs to educate the subjects. A postintervention assessment was made at Time 4 (T4) after CBs.

▪ Data analysis

We reported descriptive analysis of subjects’ demographics, ADL level and level of depression. Continuous variables were presented as means and standard deviations while ordinal or categorical variables were presented in numbers and percentage. An overall p value less than 0.05 was used to determine statistically significance. Generalised linear mixed models (GLMM) were used to compare the mean health literacy scores at different time points. Coefficient estimates and p values were reported. General Estimating Equation (GEE) was used to compare MMAS level over time. Odds ratios, standard errors and p values were reported.

▪ Ethical approval and consent

Both The University of Hong Kong/ The Hospital Authority’s Hong Kong Island West Cluster Institutional Review Board (UW 15-545) and The Hong Kong Polytechnic University’s Human Subjects Ethics Subcommittee (HSEARS20160913002-01) approved this study. All participants enrolled in this study provided a written informed consent.


A total of 131 subjects were included. The majority (82%) were females, and the mean age was 82.5 (SD 6.5) (Table 1). More than half (55%) received no formal education, 31% received primary education and 14% received secondary education or above. About one-third (31%) were living alone.

n %
Age M: 82.5 SD: 6.50
Female 81 61.8%
Education level
No formal education 60 55.0%
Primary 34 31.2%
Secondary or above 15 13.7%
Living status
Living Alone 26 20.6%
Living with Spouses 43 34.1%
Living with Children 26 20.6%
Living with Spouses and Children 20 15.9%
Others 11 8.8%
Barthel Activity of Daily Living M: 18.6 SD: 2.54
Geriatric Depression Scale M: 3.58 SD: 3.39

Table 1: Demographics.

▪ Change of health literacy

The mean of CHLCC score accelerated from 33.9 at T1, to 35.9 at T2, to 36.2 at T3, to 39.0 at T4 (Figure 1). Controlling age, the mean CHLCC score at T4 was significantly different from the one at T1 (coefficient=5.073, p=0.014) (Table 2). The mean CHLCC score at T4 was also significantly different from the one at T3 (coefficient=2.742, p=0.003) (Table 2).

Estimate P-value
Health Literacy
Time 4 – Time 1 5.073 0.014
Time 4 – Time 3 2.742 0.003
Time 2 – Time 1 2.062 0.407

Table 2: Comparison of mean health literacy scores over time using Generalized Linear Mixed Method (GLMM).


Figure 1: Changes of mean of health literacy scores over time.
Note. T1 = Time 1; T2 = Time 2; T3 = Time 3; T4 = Time 4

▪ Change of medication adherence level

Table 3 showed the distribution of MMAS scores and MMAS levels at different time. About 30 (23.4%) subjects were considered as poor medication adherence, and half (53.1%) were medium adherence (Table 3). In consideration of the possible ceiling effect of medication adherence, 30 subjects whose MMAS scored 8 were excluded from the GEE analysis. Comparing to the MMAS level at T1 and controlling age, the chance of moving MMAS level one level higher than its original level at T4 was high (O.R.=3.316, S.E. 0.2979, 95% CI 0.650-1.747, p<0.001) (Table 4). Similarly, the chances of improving medication adherence at T3 (O.R.=2.295, S.E. 0.2576, 95% CI 0.326-1.335, p=0.001) and T2 (O.R.=2.506, S.E. 0.2645, 95% CI 0.400-1.437, p=0.001) were about two times more than the one at T1 (Table 4).

Time MMAS level MMAS Score n %
T1 1 <6 30 23.44%
2 6-7 68 53.13%
3 8 30 23.44%
T2 1 <6 18 16.22%
2 6-7 73 65.77%
3 8 20 18.02%
T3 1 <6 30 23.62%
2 6-7 64 50.39%
3 8 33 25.98%
T4 1 <6 17 16.04%
2 6-7 56 52.83%
3 8 33 31.13%

Table 3: Distribution of medication adherence scores over time.

Parameter Std. Error 95% Confidence Interval Hypothesis Test
Odds Ratio Lower Upper Wald Chi-Square df p value
Threshold MMAS_level=1 0.169 1.2276 -4.185 0.627 2.101 1 0.147
MMAS_level=2 3.053 1.2214 -1.278 3.510 0.835 1 0.361
Age 0.985 0.0148 -0.044 0.014 1.036 1 0.309
Time 4 3.316 0.2797 0.650 1.747 18.363 1 <0.001
Time 3 2.295 0.2576 0.326 1.335 10.397 1 0.001
Time 2 2.506 0.2645 0.400 1.437 12.067 1 0.001
Time 1 ref.

Table 4: Estimating the chance of improving medication adherence by General Estimating Equation (GEE).


To the best understanding of the project team, this study is the first interventional study addressing medication taking among persons with MCI. It showed that the two interventions (VAEL and CBs) conducted by trained volunteers were effective to improve both health literacy and medication adherence among persons with MCI. Simply using VAEL did not make any change in health literacy; however, by adding CBs, MCI persons showed significant improvement in their ability to receive and interpret the relevant health information on drug labels.

Persons with MCI reported many barriers in medication taking. The most common reported barriers were: having difficulty in remembering the amount or time of each medication (49%) and difficulty in reading the instructions on prescription bottles (42%) [18]. These difficulties were not only encountered by persons with MCI, but also encountered by those with normal cognition [18]. However, to the persons with MCI, remembering word-for-word is even more demanding than those with normal cognition. Existing drug labels are all written in ‘words’ that do not support memory, understanding and interpretation.

This study introduced an innovative strategy – comic books to improve individuals’ understanding and interpretation of the instructions on drug labels, and improve their medication adherence. The use of pictures in comic books supports individuals’ gist memory that is, remembering the general meaning of things [19]. Gist memory encourages individuals to capture meaning, or remember the essential meaning, the “substance” of information, rather than exact words, numbers, or pictures. A recent experimental study showed that gist-based conceptual processing of pictures was preserved in persons with MCI, supporting the persons to extract and use gist information [20]. Therefore, pictures have superiority effect on understanding and retrieving information [20].

Misunderstanding and misinterpretation of dosage instructions on drug labels is common. A study showed that the rates of correct interpretation of the eight warning instructions on drug labels varied: from 0% to 78% [21]. This meant that some people knew nothing about the warning instructions on the drug labels after reading, and although some were able to interpret the information from the drug labels, they could understand only 78% of the content, at most. The majority of older adults in Hong Kong are illiterate or have a low education level, so they are at particular risk of misunderstanding information on drug labels and misusing medication.

This study has several limitations: Firstly, this study did not include a control group for comparison. More research with better design is warranted. Secondly, the two interventions were separated with a wash out period of one year. We assumed there were no significant changes in the participants’ physical and mental health status. Caution should be taken when interpreting these findings.

As a conclusion, the current study provided encouraging results to support medication taking among persons with MCI. Trained volunteers giving verbal advice according to the standardized protocol, using an educational leaflet and health literacy-oriented comic books seemed to effectively improve health literacy and medication adherence. Since inadequate health literacy was interconnected with poor medication adherence [22], an intervention to improve both health literacy and medication adherence is worthy to be supported.


Part of the findings was presented in the International Alzheimer’s Disease Conference 2017 in Hong Kong Science Park, Hong Kong on 25 May 2017.


The authors wish to acknowledge the invaluable contribution of the participants and the Medication Use Supportive Team (MUST) members. Special thanks go to Mr. Alex Kwan, Mr. Godfrey Lam, Ms. Cynthia Ng, Ms. Joyce Hui, Ms. Karen Tam and all estate managers of the 15 public estates for their kind support and assistance in data collection, and the anonymous reviewers for their thoughtful review and guidance.

Availability of data and materials

The datasets generated and/or analyzed during the current study are not publicly available as the release of the dataset in public domain may compromise research participants’ privacy but the datasets are available from the corresponding author on reasonable request.

Authors’ contribution

AYM Leung contributed to the study design, application for ethical approval, data collection, statistical analyses, and interpretation of the findings, intellectual input, and revision of the manuscript. S Lo contributed to data collection and interpretation of the findings. I Leung contributed to paper drafting, data collection and data analysis. J Liu and S Ting contributed to study design, data collection, and interpretation of the findings. All authors read and approved the final manuscript.


This project was supported by Hong Kong Housing Society (HKHS) 2016-2018.

Competing interests

The authors declare that they have no competing interests.

Consent for publication

Not applicable.

Ethics approval and consent to participate

Ethical approval was obtained from The University of Hong Kong/ The Hospital Authority’s Hong Kong Island West Cluster Institutional Review Board (UW 15-545) and the Hong Kong Polytechnic University Human Subjects Ethics Review Subcommittee (ref. no.: HSEARS20160913002-01). Informed written consent was obtained from all participants of the study.