The primary objective of this study is to is to develop a conversational AI service (chatbot) in Hong Kong to assist patients and caregivers with inquires related to Hospital Authority (HA)-prescribed medications, and to evaluate its effectiveness in answering medication-related questions. The main questions it aims to answer are:

1. How satisfied are the patients and caregivers with this pharmacy chatbot?
2. How will the pharmacy chatbot impact eligible patients or caregivers on their (or the patient they are caring for) medication adherence, knowledge, and the consultation time with HA pharmacists?

There will be an intervention group and a control group:

1. The intervention group will be invited to use the AI Pharmacy Chatbot online for 7 days through WhatsApp, a commonly used social media platform in Hong Kong. They can inquire about their medications prescribed under HA and get instant, validated answers from the chatbot.
2. The control group will not use the chatbot during the intervention period.

To evaluate the chatbot's usability, researchers will measure patient satisfaction through usability questionnaires issued to the intervention group after their intervention period. Differences in medication adherence, medication knowledge, and HA pharmacist consultation time will also be measured between the intervention and control groups after the intervention period, to determine the chatbot's impact.

In this study, we aim to develop an artificial intelligent (AI) pharmacy chatbot and evaluate its effectiveness in answering questions related to prescribed-medications.

We will first develop an AI pharmacy chatbot prototype utilizing advanced technologies such as AI and natural language processing. By implementing Hospital Authority (HA) drug database and suggestions from Chief Pharmacist's Office (CPO), a Proof-of-Concept phase will be conducted to validate the chatbot's ability to address medication-related questions for locally prevalent chronic conditions in Hong Kong: hyperglycaemia, hypertension, and hyperlipidaemia. After completing the prototype development, we will invite pharmacists from HA for testing and evaluation of the prototype, further finetuning the chatbot based on their feedback.

Then we aim to enhance the usability of the chatbot by incorporating feedback from patients and caregivers. We aim to recruit a minimum of 280 patients receiving HA-prescribed medication on hyperglycaemia, hypertension, and / or hyperlipidaemia, or their caregivers. According to a previous study, the medication adherence of patients with chronic conditions in Hong Kong was about 61.63%. For an increase to 80% of adherence would require a minimum of 140 participants per group, with a significance level (α) of 0.05, a test power (1-β) of 0.8, and a cluster design effect (D) of 1.5.

Recruitment of patients and caregivers will involve collaboration with HA to identify eligible participants. Designated pharmacists and other healthcare professionals from HA will help distribute recruitment materials to eligible patients and caregivers and provide researchers a list of participants who express the willingness to participate.

Participants will be randomized in a 1:1 ratio to either the control group or the intervention group. Before the intervention period, they will all receive a consent form and an information sheet which provides details on the study. They will also be asked to complete a pre-study survey, which includes questions on socio-demographics, drug usage habits, medical history and experience with chatbots.

Participants in the intervention group will then be invited to use the pharmacy chatbot and receive a usability test guideline, with an emphasis on the chatbot's role as a support tool for medication inquiries. The intervention duration will last for 7 days and the control group will have no exposure to the chatbot during the period. After 7 days, a post-study survey will be conducted to all participants in both the intervention and control groups. Participants in the intervention group will also receive a usability questionnaire to feedback on the chatbot.

A convergent mixed methods design, combining qualitative and quantitative approaches, will be used to evaluate the intervention effect of the chatbot.

Inclusion Criteria:

• Patients aged 18 years or older diagnosed with either hyperglycaemia, hypertension, or hyperlipidaemia, and currently receiving medication prescribed by the Hospital Authority in Hong Kong.
• Caregivers aged 18 years or older for patients described above.

Exclusion Criteria:

• There is no exclusion criterion, but they should meet the inclusion criteria above.

Individual participant data that underlie the results reported in the article, after de-identification (text, tables, figures, and appendices), will be shared. A written data-sharing request for meta-analysis should be submitted by email with a methodologically sound proposal. To gain access, data requestors will need to sign a data access agreement.

Beginning 9 months and ending 36 months following article publication

Researchers who provide a methodologically sound proposal for meta-analysis. To gain access, data requestors will need to sign a data access agreement.