Family Support-Based Intervention Using A Mobile Application By Pharmacists For Older Adults With Diabetes: A Randomised Controlled Trial

Background Family support is crucial in the care of elderly patients with diabetes. Pharmacists have proven to play an important role in assisting patients to achieve goals of drug therapy. However, more information is needed to evaluate the potential benets of family-based intervention through a mobile application by pharmacists in the elderly diabetes patients. Objective To evaluate the inuences of family support-based intervention using a mobile application on glycaemic control and diabetes control-related outcomes in the older adults with diabetes. Setting A hospital in Method A randomised controlled trial was conducted with 9 months of follow-up. Family members in the intervention group (n = 78) received diabetes educational courses and encouragement via a mobile application from pharmacists to help their elderly relatives with diabetes in self-management tasks. The control group received usual care (n = 79). Main outcome measure Glycosylated haemoglobin (HbA 1c ). Results As compared to baseline, signicant improvements were observed in the intervention patients for HbA 1c , blood pressure, family behaviour in diabetes care, diabetes knowledge, self-management practices and medication adherence (P < 0.001). The intervention group showed greater decline in HbA 1c levels relative to the control group (-0.97% vs. -0.12%; P = 0.001). Signicant differences between groups for changes in blood pressure levels including scores of family behaviour, diabetes knowledge, self-management and medication adherence were found (P < 0.001), with the intervention group showing greater improvement. Conclusion Family support intervention via a mobile application by pharmacists is benecial to diabetes care for the elderly.


Introduction
Diabetes is a major health problem worldwide and may affect an estimated 700 million in 2045. The highest prevalence of diabetes in numerous countries including low and middle income countries was found in elderly people (aged 65 years and over) [1,2].
The elderly patients with diabetes frequently have multimorbidity and higher premature mortality than those without this disease [2]. In addition, geriatric syndromes, such as impaired cognition, depression and multiple medications are the major presentations related diabetes in advanced age. These diabetes-associated conditions of the elderly may change a patient's self-management ability that results in adverse diabetes outcomes [2,3].
The approach to diabetes care in the older adults should be individualized according to heterogeneity of their coexisting illnesses and geriatric syndromes [4].
Evidence suggests that social support promotes better health outcomes in chronic diseases including diabetes [5,6]. Low and middle income countries tend to have strong kinship networks. Family members have a key role in the care of the elderly [7]. They may have insu cient awareness of patient care. Previous studies provided educational interventions for both family members and patients and reported improvement in diabetes control [8][9][10]. Family members represent as considerable sources of support for diabetes patients particularly the older adults [11]. The impact of family involvement on self-management practices was found to be signi cant. The success of self-management is the key to positive diabetes outcomes [2]. A family caregiver may or may not accompany patients during their hospital visits; he or she is unlikely to obtain complete information about the care of patients from the health professional for various reasons. The intervention strategy is critical to help family caregivers in acquiring regular education from health experts.
Nowadays, mobile phone has been engaged in daily life of people globally. Mobile applications (mobile apps) are used in not only managing many tasks in everyday life but also providing healthcare delivery beyond hospital borders [12]. Published articles have indicated that mobile apps are advantageous to assist self-management in diabetes [13,14]. For health providers, pharmacists can play a meaningful role in advising patients on drug therapy. Likewise, they have been found to help patients in achieving lifestyle and therapeutic goals [15,16]. An earlier research reported that pharmacist-led intervention was helpful to support family members in patients' selfmanagement tasks and improved diabetes outcome [10]. To our knowledge, there is a lack of reports of models incorporating mobile apps into family support intervention by pharmacists in elderly diabetes care.
We designed a randomised controlled trial to examine the impact of family support intervention using mobile app to improve glycaemic control and other diabetes control-related outcomes targeted for older adults with type 2 diabetes. The educational intervention was delivered via mobile app by pharmacists to family members. Afterwards, the family members would inform the contents received to the patients. Using mobile app as a tool to deliver information could overcome cognitive overload found in face-to-face education where large amount of information is provided in a short time. Short interesting and easy-to-digest information provided through mobile app over a long time period is more likely to attract attention and facilitate learning, compared to traditional methods of education.

Aim of the study
The aim of this study was to evaluate the effectiveness of family support-based intervention via a mobile app by pharmacists on clinical outcomes, family behaviour, diabetes knowledge, self-management practices and medication adherence in older adults with type 2 diabetes.

Ethical considerations
Ethical approval for the study was granted from the Research Ethics Committees of Faculty of Pharmaceutical Sciences, Prince of Songkla University (ST. 0521.1.07/1105). Written informed consent was obtained from all participants and their family caregivers.

Study design
The study was designed as a two-arm randomised controlled trial. The setting was outpatient diabetes clinic at one hospital in the southern part of Thailand.

Participants
A research pharmacist recruited outpatients from the diabetes clinic at the study site.
Inclusion criteria were (1) ≥ 65 years of age, (2) a diagnosis of type 2 diabetes, (3) inadequate glycaemic control (glycosylated haemoglobin (HbA 1c ) level > 7% or 53 mmol/mol), (4) receipt of oral antihyperglycaemic agents, and (5) having a family member as a caregiver. The eligible family members were (1) ≥ 18 years old, (2) a spouse, child or relative of the patient, and (3) able to participate in this study by using mobile phone. Patients who received insulin therapy were excluded from the study.
All participants were randomly assigned to either the intervention group or the control group. The allocated sequence was generated by strati ed randomisation method based on age of patients, HbA 1c levels and duration of diabetes.
To calculate sample size, determination performed using HbA 1c from the study by Wichit et al. [9]. At the end of the study, HbA 1c levels in the intervention and control groups were 7.0 ± 1.2% and 7.3 ± 1.4%, respectively. With 0.05 for type 1 error and 80% for power of test, a sample size of 74 patients within each group was needed.

Control group
The control group received usual care provided by health professionals at the diabetes clinic. In the usual care, pharmacists normally examined the prescription validity and identi ed drug-related problems.

Intervention group
Beyond usual care, study patients also received family-based intervention via a mobile app. The goal of the intervention was to facilitate family members' assistance in the patients' self-management tasks for achieving the diabetes goals. Physicians were blinded to the intervention.
The research pharmacist delivered diabetes education every day for 3 months. At the end of each session, 1-3 quizzes were released to family caregivers. Additionally, medication reminders were distributed 4 times a day: 6.00 am, 12.00 pm, 6.00 pm and 8.00 pm, and physician appointment reminders sent 1 week before the appointment date. If family caregivers had any diabetes-related problems, they were able to send their queries to the research pharmacist anytime that they wanted to. After that, the feedback would be tailored to the questioner directly.

Outcomes
The primary outcome was change in HbA 1c of the intervention group relative to the control group. Secondary outcomes were changes in (1) family behaviour (supportive and obstructive behaviour) in the diabetes care measured by the revised version of the Diabetes Family Behaviour Checklist [18], for which Cronbach's alpha was 0.84 for this study, (2) diabetes knowledge evaluated by the General Knowledge of Patients with Diabetes [19], for which Kuder-Richardson Formula 20 (KR-20) was 0.7 for this study, (3) self-management practices assessed by the modi ed version of the Summary of Diabetes Self-Care Activities Scale [20], for which Cronbach's alpha was 0.54 for the current study, (4) medication adherence, and (5) blood pressure. Except for diabetes knowledge, which was measured in both patients and their families, all outcomes were assessed in patients. The outcomes were determined at baseline and at 9 months after the intervention. Nonetheless, adherence to medications was assessed at every visit, 3 month apart, until the end of the study.
Adherence to diabetes medication was based on pill count during hospital visit and calculated using the formula: [(number of pills received -number of pills remaining)/number of pills received] x 100.

Statistical analysis
Participant characteristics were presented as frequency with percentage for categorical data and mean with standard deviation (SD) for continuous data. Differences in demographic characteristics between the intervention and control groups were analysed using chi-square test for categorical variables and independent samples t-test for continuous variables. To compare primary and secondary outcomes of the two groups, between-group differences were examined using independent samples t-test and within-group differences were tested using paired t-test. Comparisons of changes in medication adherence scores between groups at different time were evaluated using split-plot ANOVA. Statistical signi cance was considered at P-values < 0.05. The analyses were conducted with SPSS version 22.0 (SPSS Inc., Chicago, IL, USA).

Results
The patient enrolment began in January 2019 and the nal follow-up of the study participants nished in January 2021.
The ow chart of participants throughout the study is illustrated in Fig. 1. The 166 eligible patients who agreed and took part in the study, were allocated to the intervention group (n = 83) and the control group (n = 83). However, 5 intervention patients and 4 control patients were dropouts during the follow-up.
For characteristics of patients and their family caregivers, no signi cant differences in patients' characteristics were apparent between the two groups at baseline (P > 0.05) ( Table 1), whereas their relatives' characteristics were signi cantly different between groups in gender, age, education, patient's relationship and living in the same house as patients (P < 0.05). At baseline, clinical outcomes (HbA 1c and blood pressure) were not signi cantly different between groups ( Table   2). Within the intervention group, HbA 1c and blood pressure signi cantly declined from baseline to 9 months (P < 0.001). At the end of the study, changes in HbA 1c and blood pressure were signi cantly different between the intervention and control groups (P ≤ 0.001). There were no signi cant differences between groups in diabetes control-related outcomes (scores of family behaviour, patients' knowledge, caregivers' knowledge and self-management) at baseline (Table 3). Compared to baseline, all items of diabetes control-related outcomes signi cantly increased in the intervention group (P < 0.05). Signi cant increase in patients' knowledge and self-management scores also occurred in the control group (P = 0.014). After the 9-month intervention, changes scores in all diabetes control-related outcomes differed signi cantly between groups (P ≤ 0.001). Table 3 Comparisons of changes in diabetes control-related outcomes within and between groups (mean + SD) Medication adherence scores between groups showed no signi cant differences at the beginning of the study (P = 0.750) ( Table 4). The intervention group had greater increases in the adherence scores than the control group at different time from 3 months to 9 months, with interaction of group x time and between-group differences in change patterns being signi cant (P < 0.001).

Discussion
This randomised controlled trial demonstrated that family support-based intervention delivered through a mobile app had positive effect on glycaemic control for older adults with diabetes. The improvement in family support, diabetes knowledge, self-management behaviour and medication adherence was also in accordance with the family-involved intervention.
Signi cant improvement in glycaemic control was seen in the family-based intervention group compared with the control group. The results of this study were agreement with those obtained from studies by Shi et al.
[8] and Withidpanyawong et al. [10] which indicated that family support was bene cial for diabetes control. The previous studies had family caregivers accompanied with the patients to take part in the intervention courses, whereas the present study had the family members participate in the intervention sessions only via mobile app. Unlike the current study, a trial by Wichit et al. [9] found lack of difference between groups in glycaemic control when using a family-involved method. However, the follow-up period in the prior trial [9] was only 13 weeks, which is shorter than that in this study (9 months follow-up). Wichit et al. [9] study also provided additional education to participants at week 9. Nevertheless, after this education delivery no further data were gathered 3 months later, that is the optimal timing for HbA 1c measurement [2]. Accordingly, it may not be possible to measure the small changes in HbA 1c . Findings here con rmed the results of an earlier meta-analysis [21] which indicated that family involvement in patient care was a substantial resource for the control of diabetes outcomes. Adequate glycaemic control is a key aspect of prevention of diabetes-related microvascular and macrovascular complications, particularly in the elderly [22,23].
Apart from diabetes content of the intervention in this study, the education for family caregivers focused on hypertension-related lifestyle modi cation. Accordingly, the results showed positive impact on systolic and diastolic blood pressure in the intervention group with signi cant differences from those in the control group. Lee et al. [24] reported that family assistance was a strong factor for self-regulation in the elderly patients with uncontrolled hypertension. The American Diabetes Association [2] recommends blood pressure targets < 140/90 mm Hg for elderly patients with diabetes. Although baseline blood pressure levels in the intervention group (141/77 mm Hg) were slightly higher than the guideline target, family members' participation presented as the considerable factor for improvement in these outcomes (132/72 mm Hg). Strict blood pressure control leads to diminished risks of mortality and diabetes complications including nephropathy and cardiovascular diseases [2,22].
The present ndings pointed out that signi cant increases for supportive and obstructive family behaviour were observed in the intervention group compared to baseline, similar to those of Withidpanyawong et al. [10] study.
Helpful family involvement was associated with better self-management of diabetes, on the other hand harmful involvement was related to worse self-management adherence. However, the intervention group experienced signi cantly greater positive care from family members than did the control group. This may make the participants manage their diabetes more effectively and lead to the improved glycaemic control that was seen in our results.
Elderly diabetes patients are at high risk for developing geriatric syndromes including cognitive dysfunction [2]. A low education level was a signi cant factor for the insu cient knowledge regarding diabetes self-management in the elderly [25]. The majority of study participants here completed their education at primary school or lower (intervention group 81% vs. control group 75%). Furthermore, the usual hospital education for patients with diabetes was similar in both the intervention and control groups. Nevertheless, the intervention group had a larger improvement in knowledge than did the control group that may be attributed to information acquired from family caregivers using mobile app. Higher diabetes knowledge was signi cantly associated with better selfmanagement practices and lower HbA 1c [26].
Effective self-management behaviour is essential to meet glycaemic goals [2]. Results here observed that selfmanagement practices improved signi cantly more in the family intervention group than in the control group.
The results were consistent with ndings in the study by Wichit et al. [9]. Previous research by Getie et al. [27] supported our ndings that having family members as caregivers positively affected diabetes self-management. Intervention family caregivers here were predominantly spouses or children (88.5%) and in-home partners (71.8%). Caregivers who were spouses or children acted as the key providers for assistance in terms of selfmanagement [28]. Likewise, more frequent contact or being in-home supporters was correlated with better control of diabetes [29]. However, signi cant increments in self-management activities were also detected in the control group when compared to baseline. There may be contamination bias from the intervention patients and/or Hawthorne effects from intensive follow-up that may change the behaviour of the control patients.
The intervention group had signi cantly better medication adherence than did the control group. The family intervention was effective for improving adherence as shown in a trial conducted by Withidpanyawong et al. [10].
Mayberry et al. [29] suggested that greater caregiver closeness had an association with adherence to diabetes medications. Accordingly, the intervention patients who received more support from their family members may increasingly adhere to medication taking. In contrast, patients who perceived taking medications to be unimportant were more likely to have non-adherence [30]. The current study informed the intervention patients about the importance of taking prescribed medications through family-based education via mobile app. This matter may enhance patients' awareness and improve their behavior in taking medications. Medication adherence was an important factor related to glycaemic target (HbA 1c < 7%) in patients with diabetes [2,31].
The utilisation of mobile app incorporating with family support for elderly diabetes care is the strength of this study. The mobile app has capacity to provide diabetes education as well as facilitate interactions between family caregivers and health providers such as pharmacists. This mobile technology saves family members time and expense by decreasing travel to clinics, which is especially helpful for persons who have hectic schedules or distance barriers [32]. However, there are several limitations to the current study. The participants were recruited from a single hospital in Thailand, hence there is a limitation on generalisability of the ndings. In addition, differences in demographic characteristics of family caregivers between the intervention and control groups may affect the research ndings. Finally, this study investigated only patients who received oral antidiabetic agents; the results may differ for those who used insulin injections, where the administration is more complicated than that of oral medications.

Conclusion
The present study demonstrated that family-involved intervention based on a mobile app can help older adults with diabetes to achieve better glycaemic control, the primary outcome, than the traditional care. Additionally, favourable changes related to the intervention were detected in secondary outcomes including blood pressure, family behaviour, diabetes knowledge, self-management activities and medication adherence. Further research is important to evaluate the long term effects of family intervention and to design the superior delivery of diabetes care services for the elderly.
Declarations Figure 1 Study participant ow diagram