Background Frequent attenders (FAs) at emergency department (ED) in Singapore hospitals have been increasing over the years. More than half of the FAs are reported to be alcohol-related frequent attenders (ARFA) and they were found to be using EDs unnecessarily. We aim to assess if there will be a difference in patient outcomes in terms of ED usage and cost-effectiveness by implementing an assertive community treatment (ACT) program to manage AFRAs.

Methods This is a prospective, multi-centre, before-and-after, superiority and cohort study to assess the impact of ACT from 4 study sites. 200-300 patients will be recruited and followed up for 12 months. The primary objective of the study is to investigate whether there will be a reduction in AFRA ED attendances. The secondary objective is to estimate the change in total cost utilization.

Conclusion/Significance All patients who are on ACT programme will be enrolled in this study. The study intervention will be used as a new mode of care at participating hospitals. We expect to see reduced alcohol addiction level, reduced isolation level, improved motivation and better overall health. With reduced alcohol-related hospital visits, we would also expect to see improved healthcare utilization by ARFAs which will lead to increased cost savings to the healthcare systems and decreased social costs.

BACKGROUND Singapore has seen a steady growth in the use of emergency department (ED) services. ED visits increased from 752,122 in 2007 to 1,006,800 in 2013. Despite efforts to redirect low acuity ED patients to primary care, EDs continue to be busy and with long waits for patients. Data suggest that frequent attenders to ED take up a disproportionate share of the resources. Based on a set of pre-determined criteria, frequent attenders were found to be more likely to make inappropriate and unwarranted visits to ED. These ED visits would be better served by a different part of the healthcare system.

Patients with alcohol misuse problems are known to make up a significant proportion of frequent attenders. Such patients are referred to as Alcohol-Related Frequent Attenders (ARFAs).This population of patients exist in the community and tend to cycle through the healthcare system through emergency departments.

ARFA's are characterized by poor physical and mental health and a high level of unmet social needs, contributing to markedly poor engagement with appropriate services. They belong to the segment of patients with severe alcohol misuse disorders that have low motivation for change.

It is known that traditional clinic-based alcohol services have difficulty engaging ARFAs. Traditional treatment focuses on abstinence-based therapies. Its emphasis has been on patients who are ready to change their drinking and can organize their time sufficiently for on-site clinic treatment. Traditional treatment excludes ARFAs who are the most severe alcohol misuse patients. Such an approach has been supported by the argument that it is therapeutically appropriate to concentrate on patients that want to change while others are allowed to reach their personal rock bottom.

The traditional approach has been ineffective in treating the ARFA population. The majority of ARFAs do not present for treatments to reduce their drinking. Amongst alcohol misuse patients that present for treatment, the drop-out rate is high. Local data from the National Addictions' Management Service (NAMS) Specialist Outpatient Clinic shows that up to 75% of such patients drop-out of follow-up with a counsellor in just 3 months. Patients who engage poorly with addictions treatment tend to present instead to emergency services and have unplanned hospital admissions for alcohol-related consequences.

With the failure of the traditional, clinic-based, abstinence approach, an alternative means to treat the ARFA population is needed. Inappropriate and substantial utilization of public emergency medical resources by ARFAs provides a strong impetus to focus on an alternate strategy to mitigate demand. Community-based strategies may provide the answer. Treatment in the community may mean that abstinence remains a lofty goal. With a paradigm shift towards using harm-reduction strategies, evidence shows that reducing morbidity and mortality is achievable.

ASSERTIVE COMMUNITY TREATMENT Assertive community treatment (ACT) refers to a model of community service provided to people with severe mental illness originally pioneered by Stein and Test20. Key components include: i) rapid access to services, ii) a small case load per case worker, iii) a high ratio of community to office-based appointments, iv) assertive engagement - with multiple attempts, v) a shared care approach with care coordinators working within a multidisciplinary team that meets frequently.

ARFAs have multiple medical issues that may be related to alcohol misuse such as liver cirrhosis, gouty arthritis, alcoholic gastritis etc. In Singapore, local hospitals' community health services have access to specialist support and are embedded with community social services. Hospitals' community health services are well-positioned to serve ARFAs within their area of coverage. With the guidance and support from addictions specialists from the NAMS, ACT teams from hospitals' community health services would be able to tackle the alcohol addiction, manage the multiple medical problems, apply synergies with local community services in helping these patients.

ACT for ARFAs is well-established in the UK. With differing burden of disease, demographics, culture and drinking habits, it is unclear if this treatment would be effective here. Singapore had previously implemented ACT for non-alcohol misuse patients with severe and persistent psychiatric illness which demonstrated improved outcomes and reduced healthcare utilization. There is a need to evaluate the impact of ACT on healthcare utilization and clinical outcomes in Singapore.

Our study will utilize ACT with the aim of reducing ED use, usage of other emergency services and decrease the addiction to alcohol in the ARFA population.

IMH NAMS - ACT TEAM PILOT This study will be an extended pilot, utilizing and building upon the system and intervention as established by the NAMS ACT team, which commenced its service in April 2018.

The NAMS ACT pilot (for its first year of service) led to a 45% reduction in alcohol-related misuse in a group of 14 patients, for whom 6 months of ACT was provided. The pilot is currently undergoing write-up and data analysis, with the intention for publishing by the year-end (Study Title: Alcohol Assertive Community Treatment in Singapore: Reducing alcohol-related A&E presentations and impact on psychosocial functioning).

HYPOTHESIS The primary objective of the study is to investigate whether the implementation of ACT can reduce ED attendances by ARFAs. A secondary objective is to estimate the change in total cost utilisation from the implementation of ACT.

EXPECTED RISKS AND BENEFITS It is intended that all patients on assertive community treatment will be enrolled in this study. The intervention will be implemented as part of a new standard of care of the hospitals involved - Khoo Teck Puat Hospital (KTPH), Tan Tock Seng Hospital (TTSH), Singapore General Hospital (SGH) and Ng Teng Fong Hospital (NTFGH).

Expected risks include poor responses to the intervention. However, based on the experience with ACT in the UK and in Singapore among non-alcohol misusers, we believe this presents an acceptable low risk of occurrence. The collection of data poses the usual minimal risk of privacy breach. Safeguards for study participants will be put in place. Prior to data analysis, all patients' identifiers will be removed from the dataset to safeguard the patients' privacy and confidentiality. Data will be securely stored.

Benefits to study participants include reductions in the level of addiction to alcohol, reduction of isolation, improved motivation and better overall health. Benefits to society include reduced unnecessary healthcare utilisation by the ARFA population leading to cost savings to the system and reduced social cost when the level of alcohol misuse is controlled.

STUDY DESIGN AND PROCEDURES/METHODOLOGY This is a prospective, multi-centre, before-and-after, superiority, cohort study to assess the impact of ACT. We adhered to the SPIRIT (Standard Protocol Items: Recommendations for Interventional Trials) guideline for a clinical trial protocol.

As identifiable data would be collected and intervention in the form of ACT is proposed, consent for participation will be obtained from study subjects. Consent for participation will include consent to release retrospective data of 5-years from time of recruitment for evaluation.

STANDARD TREATMENT PRIOR TO IMPLEMENTATION OF ASSERTIVE COMMUNITY TREATMENT Standard treatment includes episodic treatment for individual visits depending on the reason for the ED presentation. Presently, Singapore does not have national clinical guidelines on treatment of alcohol misuse. Treatment of the ARFA population varies as per local hospital policies and may include follow-up with psychiatrists, clinical psychologist, counsellors, nurses, social workers and community support workers. Referrals may be made to the national referral center, National Addictions Management Service (NAMS) for addictions follow-up. NAMS is a clinic-based service managed by a multidisciplinary team of specialists, providing individual counselling and psychotherapy for alcohol addictions.

For patients undergoing ACT, the new standard of care will supersede local hospital policy.

INTERVENTION: ASSERTIVE COMMUNITY TREATMENT Assertive Community Treatment (ACT) is part of a 6-month program ARFAs will undergo.

ACT teams will engage with ARFAs on a dedicated schedule. Each visit would last an average of 60 minutes. There will not be a re-arrangement of no-show visits. ACT teams will re-engage the ARFA over the phone to plan for the next visit as per the schedule.

In keeping with principles of ACT, ARFA's will be assertively sought out in the community and provided intensive engagement, with holistic and supportive care alongside a harm reduction approach. Care will also involve case management and involvement of SSA's.

Each interaction during the treatment period will focus on various aspects covering physical health, mental health, and addiction 14. ACT team members will explore for problems across 4 life domain categories; Medical, Psychiatric/Psychological, Social, and Substance, also referred to collectively as an acronym - MPSS. Weekly Multi-Disciplinary Team (MDT) meetings will be conducted and problems will be discussed according to the MPSS format.

COMPOSITION OF ACT MULTI-DISCIPLINARY TEAM Each hospital's ACT team will be comprised of nurses and/or allied health practitioners the backbone of the team. Doctors have received the requisite training by the NAMS will be involved in leadership of these teams. In view of the intensity of the program, in line with international practice, each ACT team would manage 16 active patients over a period of 6 months.

INTERVENTION FIDELITY ACT workers providing care to participants will complete a contact log detailing the care provided for each patient following each contact. There will not be a replacement for non-contact visits. The contact log will include details about the member of staff involved, the mode (i.e. telephone or face-to-face) and setting of the contact (i.e. patient's home, community setting etc).

SAMPLE SIZE AND STATISTICAL METHODS Determination of Sample Size As this is a real-life implementation study, we aim to recruit all eligible patients in the ARFA program.

Based on a single hospital's data, the ARFAs utilize Emergency Department services at a mean of 30.15 visits/year (Standard deviation of 23.47). Assuming a 40% decrease in usage, a standard deviation of 23 visits, a power of 80% and alpha of 0.05, we calculate the required sample size to be 29 patients. Assuming a 15% drop-out rate, we will require 34 patients.

Statistical and Analytical Plans Using complete set data analysis, cases undergoing ACT will be compared to their own historical outcomes as controls. ED, EMS utilization, variables from SPF, CISS and UCLA 3-point loneliness scales will be compared. Using the paired sample t-test or Wilcoxon Signed Rank test, the post-intervention utilization of ARFAs undergoing ACT will be compared to the pre-intervention utilization. To adjust for potential confounders linear regression analysis will be undertaken. Normality of the variables included in the model will be assessed using PP plots. Interactions will be tested for [Variables] that are clinically relevant and if found statistically significant. Model adequacy will be assessed through coefficients of determination (R2) and by plotting error terms against predicted values. Presence of multicollinearity will be evaluated by VIF statistics.

Should complete set analysis show efficacy, we will proceed to analyse data using an intention-to-treat analysis.

The incremental costs of delivering the intervention will be estimated and compared to any observed cost savings from reduce usage of ED services and other health services. Uncertainties will be modelled and an estimate of the change to total costs made.

Inclusion Criteria:

• • Patients who are not on active follow-up at the National Addictions Management Service at the Institute of Mental Health (IMH).

  • Patients who speak English or Mandarin.
  • Patients who are 21 years of age or older.
  • ≧2 SNOMED alcohol-related ED visits in last 12 months AND ≧5 Emergency Department visits (All-cause)
  • Patients who are cognitively intact and able to understand and appreciate the contents of the consent.

Exclusion Criteria:

• • Patients who are on active follow-up at the NAMS at the Institute of Mental Health (IMH).

  • Patients who are uncontactable by phone AND through opportunistic recruitment.
  • Patients who do not give consent.
  • Patients who do not speak English or Mandarin.
  • Patients who are under 21 years of age.

Singapore has a population of 5.7 million, served by 7 adult EDs. Based on unpublished data, we estimate a population of 200-300 patients who would meet the criteria to be classified as ARFA in Singapore. Our study will implement the community-based intervention in ARFAs from 4 adult EDs.

Once patients are recruited, individual, identifiable 5-year historical data from the point of study enrolment will be obtained from the electronic medical records system. This data will include dates of ED visits and ED diagnosis.

Each intervention site will recruit up to 32 patients per year. The ACT teams will see a maximum of 16 patients over a period of 6 months (Appendix 1). Four hospitals will participate in the study over a period of 4 years with different initiation dates at each site. The setting up of such a service requires on-site specialist guidance and the specialist addictions team from NAMS is only able to guide one hospital's ACT team at a time for a period of 12 months.

As this involves a vulnerable patient group (Alcohol misuse patients) and a sensitive outcome (Alcohol addiction levels), there are no plans to share this individual-level data with other researchers