The PACT (Patient Activation Through Conversations) Study

Study Methodology

This is a cluster randomised controlled trial comparing the PACT program to Usual Care. Each cluster consists of a Teamlet in a NHG Polyclinic, consisting of a regular group of Family Physicians, a Care Manager, and a Care Coordinator, that provide chronic care for a regular group of about 5000 patients.

Study Population

The study population will include patients with suboptimal control of diabetes (HbA1c >8%) between the ages of 21 to 64. Patients with a suboptimal control of diabetes frequently require a combination of both pharmacological and lifestyle measures for improvement of their glycaemic levels.

The inclusion criteria include the following:

• Individuals with a diagnosis of Type 2 Diabetes Mellitus who are enrolled in Teamlets in NHGP
• Age between 21 to 64 years old
• HbA1c level of more than 8% in the last 6 months

The exclusion criteria includes the following:

• History of acute coronary syndrome in the past 3 months
• History of stroke or transient ischaemic attacks in the past 3 months
• History of end stage organ failure (liver cirrhosis, cardiac failure, latest eGFR <30 mL/min/1.73m2 in the last 6 months)
• Cancer requiring treatment in the past 5 years.
• Active psychiatrist follow up
• Social issues that are significant enough for follow-up or referral with a medical social worker within the last 12 months
• Women who are pregnant or breastfeeding
• Patients who are enrolled in other current diabetes-related interventional studies
• Inability to provide written consent for the study protocol and able to commit to the study duration.

These patients are excluded as they may have conditions that could be of a higher priority for management as compared to Diabetes Mellitus, and may not be able to provide informed consent or participate actively in shared decision making.

Summary of Study Design

The unit of randomisation will be each cluster, as represented by a single Teamlet. Randomisation will be carried out on all Teamlets that are involved in this research study from NHG Polyclinics. The study team aims to involve two to three polyclinics that includes six Teamlets in each polyclinic. A total of 12 to 18 Teamlets will hence be involved, subject to resource and recruitment considerations.

Teamlets in each polyclinic will be randomised to either the PACT Program condition or Usual Care condition. Patients who belong to Teamlets that are randomised to the PACT Program condition will be recruited for participation in the PACT Program, and patients who belong to Teamlets that are randomised to the Usual Care condition will be recruited and provided with routine care. Due to the nature of the PACT intervention, blinding of the Teamlet members (Doctor, Care Manager, Care Coach) and patients will not be feasible. The data analyst will be blinded to the group allocation.

The PACT Program intervention will last 3 months. Four participant assessments will be conducted, and this will take place at baseline and at the 3-month, 6-month, and the 12-month mark.

Recruitment of participants will take place over 12 months and follow-up of participants will be carried out over 24 months. Further analysis will be undertaken for 12 months, so the total length of the study will be 4 years and 0 months.

PACT program procedures The PACT program lasts 3 months and is a health coaching program for people with type 2 diabetes.

Patients have an in-person visit with a Care Coach at the baseline visit, where review of their metabolic parameters, self-care behaviors and health motivators is carried out, supported by use of a PACT form. This is followed by goal setting.

The PACT form consists of:

1. Results of patients' latest clinical readings are shared with the patient
2. Patient fills in their key motivators for managing their diabetes well
3. Patient fills in a survey that assesses their own self-care behaviors (diet, exercise, medication adherence), diabetes distress
4. Goal setting with the care coach to set SMART (specific, measurable, achievable, relevant and time-sensitive) goals

The Care Coach has been trained to apply a bio-psychosocial approach, motivational interviewing skills and provide diabetes care support. The discussion will take approximately 15 to 30 minutes.

Supportive follow-up is provided over the following 3 months. The Care Coach will support patients in reviewing their goals, assist in problem solving, and provide motivational support. Follow-up will be conducted 2 weekly via telephone calls and/or WhatsApp messages, according to the patient's preference. Care coaches will minimally send patients a whatsapp message or conduct a phone call, however, we acknowledge that some patients may not reply or pick up the phone. This will take approximately 10 to 20 minutes each time. After 3 months, the patients will be reverted back to routine clinical care.

Outcome Assessments

Patient outcomes will be assessed at baseline (12 months before study period), and at regular intervals over 24 months for clinical data. Patient reported outcomes will be assessed at 3, 6, and 12 months.

1. Clinical Outcomes

• Clinical outcomes will be extracted from the NHGP Business Intelligence database, from routinely collected data.
• The outcomes that will be assessed include:
• Clinical outcomes at baseline and follow-up: HbA1c, systolic and diastolic blood pressure, body mass index, Total cholesterol, HDL-cholesterol, LDL-cholesterol, and Triglyceride levels, serum creatinine level, estimated glomerular filtration rate, urine albumin-creatinine ratio, and the urine protein-creatinine ratio
• At baseline: Presence of co-morbidities, which include hypertension, hyperlipidaemia, ischaemic heart disease, stroke, peripheral vascular disease, previous lower limb amputations, chronic kidney disease, diabetic neuropathy, and diabetic retinopathy
• Patients usually make clinic visits every 3 to 6 months. Clinical outcomes data will be obtained for baseline and over a period of 24 months.

1. Patient reported outcomes

• Patient reported outcomes will be assessed using a questionnaire at baseline, at the 3-month mark, 6-month mark, and at the 12-month mark.
• The questionnaire will be self-administered with interviewer assistance on patient request.
• The baseline and follow-up questionnaires include the following measures:
• Sociodemographic information - Only in baseline questionnaire
• Self-care behaviours, measured with the Summary of Diabetes Self-Care Activities scale (SDSCA)
• Diabetes-related distress, measured with the 2-item Diabetes Distress Scale (DDS-2)
• Diabetes empowerment, measured with the short form of the Diabetes Empowerment Scale (DES-SF)
• Health-related quality of life, measured with the EQ-5D-5L
• Patient satisfaction, as measured by 5 questions from the Patient Satisfaction Questionnaire-III- Only in the PACT 3-month questionnaire.

1. Referral to community exercise programs. As part of comprehensive lifestyle management, the patients may be directed to community exercise programs, such as those run by SportSG. Attendance to such programs will be recorded via patient-reported information in the 3-month, 6-month, and 12-month questionnaires.
2. Adverse events Improvements in self-care behaviors (diet and physical activity) may be associated with hypoglycaemic symptoms. Monitoring for serious hypoglycaemic episodes requiring emergency department visit and/or admissions to hospital will be recorded via patient-reported information the 3-month, 6-month, and 12-month questionnaires.
3. Drugs Prescribed Drugs could affect clinical outcomes such as HbA1c. Information such as drug prescribed, dosage, and cost of medications will be extracted from NHGP Business Intelligence database.
4. Resources and Cost

Costs and resource utilisation will be analysed, including:

• Health-service utilisation costs including Medical (Doctor), Nursing (Care Manager), Allied Health (Dietician, Psychologist, Medical Social Worker) consultation sessions
• Training and utilisation costs for the Care Coaches, including the number of Telephone/WhatsApp follow-up calls by the Care Coaches
• Laboratory tests
• Medication costs

Sample Size and Statistical Methods

Estimation of Sample Size A minimal clinically important difference of 0.4% in HbA1c is expected between participants of the PACT program compared with Usual Care. In estimating the sample size, a standard deviation of 1.2% was used, as estimated from a local study. We aim to recruit between 12 to 18 Teamlets in the study. If 12 Teamlets are recruited (6 Teamlets are randomised to PACT and 6 to the control group), 34 subjects are required for each Teamlet (total 204 participants in the PACT program and 204 participants in Usual Care), at 80% power with one-sided alpha at 5%, and setting the intracluster correlation coefficient at 0.01 with an assumed drop-out rate of 20%. If 18 Teamlets are recruited (9 randomised to PACT and 9 to the control group), 24 subjects are required for each Teamlet (216 participants in the PACT program and 216 participants in Usual Care), at 80% power with one-sided alpha at 5%, and setting the intracluster correlation coefficient at 0.01 with an assumed drop-out rate of 20%. Sample size calculations were conducted using SampSizeV2.

Statistical and Analytical Plans

1. Clinical Outcomes evaluation Categorical data will be summarized as counts (percentage), while continuous data will be summarized as mean (standard deviation) if they are normally distributed or median (interquartile range) if they are skewed. Baseline characteristics will be compared between participants in the PACT program and Usual care. Differences in the outcomes of interest (clinical and patient-reported) at different data points will be compared between participants in the PACT program and Usual Care. As this is a cluster-randomised trial, a mixed model will be used to evaluate the longitudinal difference-in-difference outcomes between the PACT program and Usual Care, incorporating Teamlets as the random effect. Relevant baseline covariates will be included in the model to account for potential confounding factors.
2. Resource and cost evaluation This study will analyse the cost-effectiveness and resource utilisation of the PACT program. The potential economic impact of new program implementation will be studied by estimating the incremental cost effectiveness ratio (ICER) between PACT and Usual Care (incremental direct medical cost per unit improvement in clinical outcome).

Effectiveness outcomes may include differences between PACT and controls in terms of HbA1c, DM-related complications, DM-related deaths, or quality-adjusted life years (QALYs).