A review of risks, adverse effects and mitigation strategies when delivering mental health services using telehealth

Abstract

Background

This paper presents a scoping review of the peer-reviewed literature regarding reported risks, adverse effects and mitigation factors related to providing mental health services using telehealth.

Aims

The paper aims to describe risks and risk management strategies.

Methods

Publications were included if they reported upon risks, adverse events or mitigation factors experienced, hypothesised or discussed for: any population (any country, any age), service (any mental health services), intervention (telehealth), English language, 2010 to 10 July 2021, any publication type (commentary, research, policy), excluding protocol papers, and self-help tools. The following databases were searched: PsycINFO (from 2010 to 10 July 2021), MEDLINE (2010 to 10 July 2021) and the Cochrane Database from 2010 to 10 July 2021.

Results

The search strategy resulted in 1,497 papers and after exclusions a final 55 articles were selected. Results of this scoping review are presented in terms of types of risk, risk by client population, risk by modality (eg group therapy using telehealth) and risk management.

Conclusions

Recommendations for future research include gathering and publishing more detailed information regarding near-miss and actual adverse events when delivering mental health assessment and care using telehealth. In clinical practice, training is required for potential adverse events, and to prevent them and reporting mechanisms in place to collate and learn from these.

Keywords:

• Telehealth
• mental health
• COVID-19
• risk
• adverse effects
• adverse events
• mitigation strategies
• risk management
• quality
• safety

Introduction

Telehealth can be defined as the remote provision of healthcare using technology. Telehealth can increase access to mental health services (Nelson & Bui, 2010; Nelson & Sharp, 2016). Telehealth services for mental health care has increased during COVID-19. Telehealth has a role in all settings including in rural, as well as well-serviced (typically urban) areas because of opportunities to reduce travel (time, money), reduce carbon footprint, and increase anonymity amongst other benefits.

While telehealth is gaining momentum there remain factors which are hindering development of the sector, and these include: limited postgraduate training using telehealth, attitudes of some mental health clinicians and clients, gaps in evidence and until recently the previous lack of government funding rebates or other funding for services provided using this modality (Simpson & Reid, 2014). Prior to COVID-19, countries like Australia had initiated government funded rebates (Medicare) for telehealth care in rural areas. With COVID-19 has come further funding for mental and health services to be provided in this way.

Quality and safety of care provided using telehealth

While most studies report high levels of client and provider satisfaction with mental health services provided over telehealth, along with demonstrating the reliability of clinical assessment and treatment for various diagnoses, relative to face-to-face assessments – there will always be a need to provide guidance, training, and governance to ensure high quality and safe care is provided over telehealth for all health professions (Egede, 2009; Kramer & Luxton, 2016).

Since organisations and individuals providing mental health services using telehealth will differ in multiple ways there will be multiple approaches for providing high quality, safe care and appropriate risk management. For instance, in Australia standards exist including: the Australian National Safety and Quality Digital Mental Health Standards (NSQDMH Standards, 2020) as well the broader Australia/New Zealand Standard on Risk Management (AS/NZS 4360:1999) which should be applied to telehealth care. There are also guidelines from professional bodies regarding clinical care provided online (see Appendix A).

In the UK, the UK Care Quality Commission regulates online consultations by clinical providers. Their framework for inspections asks five main questions about online providers: Are they safe? Are they effective? Are they caring? Are they responsive to people’s needs? Are they well led? To date the UK Care Quality Commission have identified six main safety concerns regarding care provided online, several of these are specific to health practitioners who provide prescriptions, and several are relevant to any online health care provider (Care Quality Commission, 2020). These safety concerns are: appropriate information sharing after online consultations with the regularly treating clinician(s), and verifying the identity of the client. Further guidance on safety reporting in research studies is provided in the UK by the NHS Health Research Authority, and the National Institute for Health and Care Research and by similar bodies in other countries. To ensure safe clinical care is provided online an assessment of the client’s mental capacity and ability to consent and tailoring safety mechanisms for telehealth are needed (Renaud-Komiya N, 2018).

Telehealth changes

Telehealth can encompass a range of changes to clinical assessment and delivery of therapy. For instance, clients may take photographs or videos and share them with their mental health clinician to better demonstrate an adult or child’s symptoms outside of the clinical encounter and setting. Assessment tools are now often online with faster scoring (Maheu, 2012). Sensors on the client or in the room might be used for assessment, monitoring or biofeedback. Shared online games or other educational material can be used with video-conferencing modes during sessions with a clinician as well as used independently to maintain therapy goals (Maheu, 2012). Multi-person videoconferencing can be used for group therapy, training or supervision (Commonwealth of Australia, 2018). As telehealth becomes more common, new sets of ethical, legal, and risk management issues may arise and these are likely to be both broad, across telehealth modalities as well as specific to some forms of telehealth. Telehealth relevant consensus guidelines and relevant regulatory standards are emerging, many were updated or came urgently in the wake of COVID-19 when telehealth practice increased. For instance, the 27 July 2020 Telehealth Guidance for Practitioners document by the Australian Health Practitioner Regulation Agency (AHPRA) (AHPRA, 2020).

Adverse events

Health systems and clinicians are risk averse (Kickbusch, 2003). Adverse events occur in health care and are defined as unintended events or effects which cause harm. Adverse events include problems or incidents, and side effects of therapeutic care, programs, products or devices. They impact negatively on the client, the provider or the system (Walshe, 2000). Through identifying the nature and rate of adverse events, initiatives to improve care can be developed. Some adverse events are so compelling that action and improvement follow briskly but others need reporting and compilation to be better understood and improvements instigated.

Broadly, the risk management process includes identifying and assessing risks, treating risks and then monitoring to prevent adverse events. A simple example of an adverse event in health care could be a wrong pharmaceutical or dose given to a patient. Adverse events should be categorised by their severity, degree of expectedness, contributing factors and potential relatedness to the care, program, product or device. Several unique risks have been identified as relevant for mental health care. These include (1) vulnerability of the client/patient due to deficits in capability or resources eg severely mentally unwell and/or homeless; (2) self-harm/suicide risk; (3) mental instability increasing risk to self or others eg hallucinations; and (4) Violence/dangerousness (O’Rourke & Hammond, 2000). Unique risks in telehealth delivery of mental health care might include: issues relating to privacy, loss of internet or other communications, unclear physical location of client during a crisis, or health provider risks such as providing telehealth care into a location without having an approved license to practice there. Of course, there are positive and negative risks associated with all conditions, interventions, and modalities.

Rationale for this review

While reporting of adverse events using telehealth is growing, there are currently few standards for adverse event reporting in telehealth. Even in the wider quality improvement realm there remains a lack of consensus on how to collect and measure adverse events (Rafter, 2015).

Many countries have implemented COVID-19 specific telehealth policies easing legal, logistical, safety and reimbursement concerns which previously existed with telehealth. Some of these COVID-19 specific telehealth policies have defined end-dates. For clinicians to continue to use telehealth to deliver mental health services and for these telehealth policies and ease of reimbursement to continue long-term, further demonstrations of effectiveness and safety of telehealth may be required. While standards and guidelines for telehealth set forth the aims and structure for future safe clinical encounters, there is no document or publication which summarises the safety of telehealth for mental health services in terms of risks, adverse events, or near-misses.

Objective

Thus, this paper aimed to scope and describe the existing published literature regarding the safety of telehealth for mental health services. This review focused on telehealth modalities which included engagement with a clinician.

Methods

This review extracted data from the existing peer-reviewed literature using the following inclusion criteria: risks (adverse events, incidents, sentinel events, complaints, “near-misses”) and mitigation factors when using telehealth for mental health assessment and/or therapy for all ages. Literature was sought for inclusion regarding: risk management, context of the telehealth service; how risk is avoided; papers which identified system level risks (for instance using root cause analysis); any papers which analysed risks using for instance risk analysis matrixes to consider the sources of risk, consequences and likelihoods of those risks; any papers which evaluated or prioritised risks (for instance by comparing risk in the local telehealth context with any previously established risk criteria) and those papers describing how risks were treated (for instance: management, changing procedures or work practices, providing education) and how compliance with risk management is monitored. Literature was included for: any population (any country, any age), service (any mental health services), intervention (telehealth), English language, years (2010 to 10 July 2021), any publication type (commentary, research, policy) but excluding protocol papers, and excluding self-help tools where a clinician was not involved, and telephone delivered services. Due to the progress in telehealth services over the last decade the review focused from the year 2010 to 10 July 2021.

This review was guided by the PRISMA Extension for Scoping Reviews (PRISMA-ScR) statements. Excel and COVIDENCE software were used for information management. The databases searched included: PsycINFO (from 2010 to 10 July 2021), MEDLINE (2010 to 10 July 2021) and the Cochrane Database from 2010 to 10 July 2021. Detailed search terms are in Box 1. The grey literature was not searched. A medical librarian from the University of Sydney reviewed and approved of the search strategy. All titles and abstracts arising from the search were independently assessed by two researchers, discrepancies for inclusion were discussed with all three researchers until a consensus to include or exclude was made.

Box 1. Search terms:

(("telemedicine"[MeSH Terms] OR "telemedicine"[All Fields] OR "telehealth"[All Fields]) OR ("telemedicine"[MeSH Terms] OR "telemedicine"[All Fields]) OR telecare[All Fields] OR telemental[All Fields]) AND (("psychology"[Subheading] OR "psychology"[All Fields] OR "psychology"[MeSH Terms]) OR ("psychiatry"[MeSH Terms] OR "psychiatry"[All Fields]) OR ("emotions"[MeSH Terms] OR "emotions"[All Fields] OR "emotional"[All Fields]) OR mental[All Fields]) AND (("risk"[MeSH Terms] OR "risk"[All Fields]) OR adverse[All Fields] OR “sentinel”[All Fields] OR “incident”[All Fields] OR “complaint”[All Fields] OR “avert” [All Fields] OR “near-miss” [All Fields]) AND “English”[lang]

One researcher extracted data from the full-text of the identified studies (Table 1). A second researcher cross checked all of the extracted data to ensure accuracy and completeness, and a third researcher reviewed the completed data extraction table for consistency and re-checked with several original papers to ensure accurate data capture. There were no disagreements regarding the data extraction phase but for three articles, additional detail was extracted from the full-text paper into Table 1. Data extracted included: first author last name, year of publication, study design, study aims, setting, participant number (or study number if a review paper) and characteristics, main findings, risk, adverse events or mitigation factors encountered or implemented. As well, bias or other methodological concerns were recorded. Our original plan to assess for risk of bias using The Cochrane Collaboration’s Risk of Bias tool for included RCTs and to critically assess the quality of the evidence using the GRADE criteria for each publication, did not lend well to the data available in existing papers as no papers sought to answer questions specifically about the safety of telehealth. As well, scoping reviews typically do not include quality assessment of included literature. The review protocol was not registered on PROSPERO because initial data extraction for several papers had occurred prior to finalising the study protocol. Data extraction for these papers was checked for completeness and consistency once the protocol was finalised, but this precluded registration on PROSPERO. These decisions are in keeping with Scoping Reviews.

Table 1. Articles included in the review (2010 to 10 July 2021) (N = 55 total articles included). Bold author name signifies usual care context not research.

Author, Year (Review)	Title	Design & Aims	Setting/Clients/N Digital technology type	Findings	Risks/adverse effects or mitigation of	Critical analysis
Aikens, 2015	Potential Impact of Incorporating a Patient-Selected Support Person into mHealth for Depression	Patient preference trial. To investigate benefits of integrating a patient-selected support person into mobile health (mHealth) for depression	Adults Patients (Depression, or on anti-depressants) enrolled 2010–2012 from primary care. N = 221 (57%) met criteria for being at risk for nonadherence (Morisky Medication Adherence Scale) Mobile phone app	Patients who chose to integrate a support person into an mHealth program for management of depression would show greater improvement in adherence and depression symptoms than those who participated in the same program without a support person.	Patients reporting suicidal ideation advised to discuss with physician and given the number of a 24-hour suicide hotline. Those meeting additional criteria for suicidal risk were also given information on emergency help and offered immediate transfer to a suicide hotline. Clinician alerts were triggered if elevated suicide risk, rarely or stopping taking antidepressant as prescribed.	This was assessing an intervention but was a non-experimental design so results may lack internal validity. Interestingly it was the more severely ill patients who chose to have a support person and those who did gained the most benefit from the program. Authors hypothesised that support people want to help but typically lack the support structure and tools to do so, this program may help in that regard.
Al-Asadi, 2014	Pre-treatment Attrition and Formal Withdrawal During Treatment and Their Predictors: An Exploratory Study of the Anxiety Online Data	Analyse data in the Mental Health Online system to identify predictors of attrition from self-directed, online anxiety treatment	Adults Anxiety Online funded by Australian Federal Government Delivered by National eTherapy Centre Swinburne University of Technology. N = 3880 of which 142 individuals formally dropped out during treatment (4.25%) Online platform	Those reporting a good quality of life were more likely to drop out of treatment. Those who stayed in were more likely to indicate they had poor QOL and were motivated to change this.	The risks described in this paper relate to patient/client attrition. Which in this study healthier people, and those not ready to make changes were more likely to drop out. As well, those who had a mental health concern other than anxiety were more likely to drop out (likely that the online anxiety program was not meeting their needs).	Low attrition rate of 4.25% may be because authors were not able to ascertain if others dropped out (ie didn’t complete the 12 online modules). They were only able to count those who formally requested to drop out.
Andren, 2019	Therapist-guided and parent-guided internet-delivered behaviour therapy paediatric Tourette’s disorder: a pilot randomised controlled trial with long term follow-up	A pilot, single-blind, parallel group randomised controlled trial to evaluate the feasibility of adapting habit reversal training (HRT) and exposure and response prevention (ERP)) to therapist/parent guided online platform	Children A specialist outpatient clinic in Sweden. Participants N = 23 children 8–16yrs with Tourette’s (TD) or chronic motor or vocal tic disorder (PTD) Online platform	Internet-delivered BT has the potential to greatly increase access to evidence-based treatment for young people with TD/PTD.	Adverse events were registered using an adapted version of the Safety Monitoring Uniform Report Form (SMURF) at mid-treatment (via telephone) and post-treatment (at the clinic). N = 12 self-reported adverse events via SMURF. None of the adverse events were rated as serious.	RCT design but only a pilot, high retention, 12 month follow-up. Supplementary table S5 linked to this paper indicated the adverse events included: depressive symptoms, irritation/anger, increased tics, anxiety/worry, stress, pain, obsessions, concentration problems and tiredness. Depressed mood and irritation/anger were the most common adverse events reported.
Bauer, 2016	Team-Based Telecare for Bipolar Disorder	Program evaluation of videoconference-based collaborative care program	Adults N = 400 Department of Veterans Affairs, USA. Video telehealth	No participant characteristics predicted non-engagement. Increased lithium use, decreased antidepressant use, and appropriate increased prazosin use	Program completers demonstrated significant improvements in several clinical indices, without evidence of compromise in patient safety. Although there were no suicides during program enrolment, two suicide attempts were identified equalling a suicide attempt rate of 2.2%/yr vs known suicide attempt rate at 2%/yr.	The suicide attempt rate in the Telecare program was the same as the suicide attempt rate estimates for this population generally.
Bucci, 2018	Actissist: Proof-of-Concept Trial of a Theory-Driven Digital Intervention for Psychosis	A proof-of-concept, single blind RCT of Actissist, compared to a symptom-monitoring control to assesses the feasibility and acceptability of Actissist, a digital health intervention in cognitive model of psychosis targeting key early psychosis domains.	Adults UK N = 36 early psychosis patients randomised Mobile phone app	Actissist was feasible (75% participants used Actissist) 97% participants remained in the trial; 90% recommend. Treatment effects large on negative symptoms, psychotic symptoms and mood – better than treatment as usual.	Safe, 0 adverse events	To minimize risk, we did not store identifying data on either the app or the server. All data transmitted to and from the servers was encrypted over https with strong ciphers according to Approved Cryptographic Algorithms Good Practice Guidelines. Program included regular clinician meetings, medication and risk monitoring, and psychosocial interventions. Standalone apps that do not link with services. Use of the app was incentivised.
Chavira, 2014	Feasibility of two modes of treatment delivery for child anxiety in primary care	Design: Feasibility study comparing in person CBT with telephone supported CBT Aims: examine feasibility of cognitive behavior therapy (CBT) for children with anxiety in primary care, using two modes of treatment delivery.	Children N = 48 parents and children (ages 8–13 yrs) Program was Cool Kids and delivered by Masters’ degree Counselling clinicians. USA. Video telehealth with telephone	No adverse effects in either group were reported.	7 families dropped out of treatment. Reasons for drop Outs: scheduling, child non-compliance, child improvement, and initiation of new dosage for ADHD medication (n = 1). None reported a worsening of their symptoms during participation.	No other items to add.
Cheng, 2018	Telepsychiatry for stable Chinese psychiatric outpatients in custody in Hong Kong: a case-control pilot study	Case-control pilot study Compare face-to-face with teleconsultations	Adults Hong Kong N = 86 male out-patients in custody receiving psychiatric teleconsultations compared with 249 receiving face-to-face psychiatric Consultations Video telehealth	Teleconsultation group had significantly better results in Chinese General Health Questionnaire scores compared to face-to-face group	No significant adverse events were identified for the teleconsultation group.	Selection bias may have played a role in the outcomes as the males receiving teleconsultations had volunteered for that mode of delivery. Only stable out-patients were enrolled.
Chen, 2017	Phone-Based Interventions in Adolescent Psychiatry: A Perspective and Proof of Concept Pilot Study With a Focus on Depression and Autism	Design: pilot study Aim: explore the utility of text message interventions in adolescent psychiatry	Children N = 3 adolescents (depression) N = 6 adolescents (autism) Mobile phone text messages	Adolescents with depression were willing to disclose suicidal ideation through text-messaging before telling their case manager or families in person.	Adolescents had message fatigue with daily prompts. Found daily questions about their mood and sleep tiresome, unhelpful to their care.	Willingness of adolescents to disclose more about their mental health to electronic devices, which feel less judgmental than face-to-face disclosure.
Childs, 2020	Rapid design and deployment of intensive outpatient, group-based psychiatric care using telehealth during coronavirus disease 2019 (COVID-19).	Design: 2 case studies Aim: describe processes and early results of telehealth services in 2 adolescent treatment programs	Adolescents with psychiatric conditions USA N = 5 to 60 telehealth consultations weekly for multiple weeks	Telehealth to deliver group-based treatments in high-risk populations is both possible and necessary during COVID.	Found no adverse events. COVID policies allowed for some typical regulations to be waived – for telehealth to remain: electronic medical record integration, improved knowledge re safety, and quality will be needed to ensure long standing policy changes are made to support reimbursements	Streamlining integration of telehealth into existing electronic medical records and broadening patient care types (individual and group) across different levels of care will help ensure sustainability of telepsychiatry
Chisolm & Sarkar, 2015	E-Health Use in African American Internet Users: Can New Tools Address Old Disparities?	Design: Survey (2010 Pew Internet and American Life Health Tracking Survey) Aims: Explore predictors of e-health behaviours and activities for African American Internet users	Adults and children N = 395 African Americans Online platform	Significant differences in e-health information-seeking related to personal and family health concerns and experiences. Targeted online e-health resources and interventions can educate and empower a significant subset of the population.	Finding high-quality sites can be difficult, and health information Web sites may have problems with completeness and accuracy of information. Users may not have sufficient health literacy to effectively evaluate or use the information they find.	Creating well prepared Internet users is imperative, and culturally informed interventions may have particular value in the African American population, who use e-health resources but still retain significant concerns regarding cultural adequacy, visual style, and trustworthiness of online information.
Christensen, 2014 REVIEW	E-Health Interventions for Suicide Prevention	Review Of online screening for suicide, effectiveness of e-health interventions to manage suicidal thoughts, and interventions when suicide risk identified in social media postings.	Adults and children Review Online platforms and social media	Online screening may have a role, but need RCTs to determine if it can reduce suicide and in what settings. Internet interventions may be most effective if they target suicidal thoughts vs conditions (eg depression). Using social media posts remains uncertain.	Online screening for suicidal thoughts and behaviours acceptable among young people, screening for suicide risk does not increase the risk of suicidal thoughts or behaviours. The direct referral of at-risk students to in-person services, such as campus counsellors, was found to be effective, although only 10%–20% agreed to referral. It is unknown if sharing thoughts and feelings on social media is beneficial to an individual or whether they are placed at further risk. Online therapeutic interventions addressing risk factors, such as depression through online CBT do not appear to cause harm.	Very few of the studies included a control group. Most research conducted in USA with university students or young adults.
Daker-White & Rogers, 2013 REVIEW	What is the potential for social networks and support to enhance future telehealth interventions for people with a diagnosis of schizophrenia: a critical interpretive synthesis	Design: Literature review with interpretive synthesis Aims: answer: how is digital health framed in the psychiatric literature? How do the benefits of telehealth translate to people with schizophrenia? What do users think about the internet, self-management and peer support?	Adults and children N = 2088 assessed and 40 papers included Papers about telehealth, internet use for people with schizophrenia UK Telehealth various	Internet has considerable potential in terms of accessing and utilising lay support, the potential of communication technologies in mobilising of resources for personal self-management or peer support was a relatively absent.	Concerns with surveillance and control for people with schizophrenia, focusing on the Internet as a potential site of risk and danger	No other items to add.
Davies et al., 2014 REVIEW	Computer-Delivered and Web-Based Interventions to Improve Depression, Anxiety, and Psychological Well-Being of University Students: A Systematic Review and Meta-Analysis	Review To systematically review trials of Web-based interventions to improve depression, anxiety, psychological distress, and stress in university students	Adults Review 17 studies Review of online platforms	Web-based interventions can be effective in improving students’ depression, anxiety, and stress outcomes compared to inactive controls.	Attrition was the only risk written about in this review of 17 studies. 12/17 studies reported on attrition which was common. Of the trials which examined it, illness severity was not associated with attrition. In the two trials which assessed it, adherence was not related to outcomes.	The 17 included studies were generally small samples, and included mild/moderate clinical presentations. No studies indicated reasons for attrition.
Deady, 2016	An Online Intervention for Co-Occurring Depression and Problematic Alcohol Use in Young People: Primary Outcomes From a Randomized Controlled Trial	RCT To evaluate the feasibility and efficacy of Web-based self-help intervention (DEAL Project) in treating co-occurring depressive symptoms and problematic alcohol use in young people.	Adults (18–25yrs) N = 104 Australia Recruited via social media Online platform	Statistically significant improvement in depression symptom severity and reductions in alcohol use quantity and frequency, however at 6 mos improvements no longer significant.	Attrition was the only risk mentioned. One third of the sample completed the program. Reasons for attrition are difficult to interpret: dropout due to dissatisfaction versus dropout due needs being met.	Data was self-report.
Deady, 2017 REVIEW	eHealth interventions for the prevention of depression and anxiety in the general population: a systematic review and meta-analysis	Review of RCTs of eHealth interventions aimed at the prevention of anxiety and depression in the general population published between 2000 and January 2016. To evaluate effects of eHealth interventions to prevent anxiety and depression.	Adults Review 10 trials included in the review. Online platforms	Small but positive effect of eHealth interventions. Inadequate evidence regarding effect on long-term disorder incidence rates.	Attrition/adherence was the only risk mentioned. Adherence was moderate across the studies and similar to that of clinical studies.	Short time period of trials included in the review. English only studies.
Di Carlo, 2021 REVIEW	Telepsychiatry and other cutting-edge technologies in COVID-19 pandemic: Bridging the distance in mental health assistance.	Narrative review Aims: examine advantages and risks of telepsychiatry, limitations and ethical concerns	Review of 45 articles about telepsychiatry	Uncertainty whether telehealth appropriate for those in acute crisis, with psychotic symptoms, those at risk for suicide or self-harm. Many articles state the perceived loss of therapeutic alliance may be a risk	Privacy and confidentiality remain risks for telepsychiatry. COVID led USA to waive information protection laws to ease use of telemedicine during the pandemic but the need to protect data as well as the secure connection during telehealth remains.	Clients with auditory, visual or cognitive impairments may also not be suitable for telehealth
Drum & Littleton, 2014	Therapeutic boundaries in telepsychology: Unique issues and best practice recommendations	Commentary / policy paper discusses boundaries in telepsychology settings, suggests best practice recommendations.	Commentary / policy paper. Need to further refine boundaries for telehealth to help prevent clinicians from engaging in ethically risky behaviors. Telehealth various	Extensions of sessions may be more common as both parties’ schedules may be more flexible causing them to adhere less to structured time frames. Due to the flexibility of telehealth clinicians and clients may interact therapeutically while in public settings which is not appropriate. Practicing outside of one’s jurisdiction could lead to ethical and legal difficulties. Casual space and clothing may undermine the professional relationship. Ensuring the privacy of others in the client’s space, as well as the clinician’s space is important (including eg family photos)	Increased risk for lack of punctuality due to failures in technology. Telepsychology in non-business hours may lead to expectation of instant feedback or view the therapist’s availability as unpredictable. Potential for excessive communication. Potential unintentional favouritism to certain clients by interacting more often or more quickly with some. Therapist’s use of emoticons, excessive punctuation (eg !), messaging acronyms may feel less professional. Working at all hours (ill, vacation, weekends) can tax the therapist and may lead client to feel guilty.	Advertising services on social media (eg Facebook) not ideal as clients may end up “liking” the page, sending “friend requests.” Or “following” may lead to complications related to unintentional client self-disclosures and makes the therapeutic relationship less formal.
Egede et al., 2009	Psychotherapy for depression in older veterans via telemedicine: a randomised, open-label, non-inferiority trial	RCT To establish non-inferiority of behavioural activation therapy for major depression delivered via telemedicine to same-room care in largely male, older adult veterans.	Adults veterans (aged ≥58 years) with major depressive disorder from clinics in USA in 2007–2011. N = 100 patients telemedicine group and N = 104 in same-room group. Video telehealth	Results from the intention-to-treat population were similar. No significant differences between outcomes on either arm. The criteria for non-inferiority were met.	There were no adverse events. 16% attrition in the telemedicine group at 12 month follow up compared to 14% attrition in the same room group at 12 months. The safety measure in place in this study was having direct telephone number of the police department in which patient resides in case an emergency dispatch is needed.	This study had good adherence, large rural population enrolled (70%) as well as minority population (African American 40% of the enrolled sample).
Engel, 2016	Centrally Assisted Collaborative Telecare for Posttraumatic Stress Disorder and Depression Among Military Personnel Attending Primary Care	RCT To determine effectiveness of a telecare intervention for PTSD and depression in military primary care.	Adults Military Primary care 2012–2014 Telecare (n = 332) or usual care (n = 334). Video telehealth	Telecare with stepped psychosocial management modestly improved outcomes of PTSD and depression among military personnel attending primary care.	This study apriori planned to report upon adverse effects. There were no study-related deaths, psychiatric emergencies, or hospitalizations.	Utilization data were self-report and this may have affected quantification of adverse events.
Fisher et al., 2019 REVIEW	Psychological therapies (remotely delivered) for the management of chronic and recurrent pain in children and adolescents.	Design: Systematic review (Cochrane Database) Aims: to determine the efficacy of psychological therapies delivered remotely for management of chronic pain in children and adolescents Pain, disability, depression, anxiety, and adverse events were extracted as primary outcomes – at end of treatment, 3 and 12 month follow-up.	Children N = 8 RCTs (n = 371) Of a psychological therapy delivered remotely via the Internet, smartphone device, computer-based programme, audiotapes, or phone Children (0–18yrs) Online platform Mobile apps	Benefit was conferred from the trials which used either behavioural or cognitive behavioural therapies for children with chronic pain.	None of the included studies reported adverse events. All trials had dropouts either post-treatment or at follow-up, or both. 3/7 RCTs gave reasons for these drop-outs. None of the included studies reported adverse events so we do not know whether children entered into any treatment or comparison group experienced adverse events, and we have no data to inform a judgement about the safety of the treatments.	Adverse events need to be measured and reported in future publications.
Flodgren, 2015 REVIEW	Interactive telemedicine: effects on professional practice and healthcare outcomes	Design: Systematic review (Cochrane Database) Aim: assess effectiveness, acceptability, and costs of interactive telemedicine, delivered in addition to, or as alternative to, usual care as compared to usual care alone	Adults and children N = 93 RCTs total N = 7 RCTs of mental health Video telehealth	Seven studies that recruited participants with different mental health and substance abuse problems, reported no differences in the effect of therapy delivered over video-conferencing, as compared to face-to-face delivery.	For the 7 mental health studies included – there was no mention of risks or adverse effects except for adherence and attrition from the program being offered.	Only one of the studies included children.
Gilmore & Ward-Ciesielski, 2019	Perceived risks and use of psychotherapy via telemedicine for patients at risk for suicide	Survey of Providers regarding demographics, caseload of suicidal patients, perceived risks for using telemedicine with patients at risk for suicide, attitudes towards telemedicine, and use of telemedicine with patients at risk for suicide.	Mental health providers surveyed N = 52 1/5 providers have provided care to individuals at high risk for suicide via telemedicine. Video telehealth	No association between experience treating suicidal patients and use of telemedicine with patients who are at high risk for suicide. Main concerns are: assessment, lack of control with patient, and triage. Suicide risk assessments can be completed via telemedicine yet providers indicating they are not sure they can do this effectively.	Authors state that protocols on suicidal risk and behaviours and telemedicine are urgently required. Unfortunately this paper has no information on experienced risks or their mitigation.	The sample was self-selected.
Glozier, 2013	Internet-Delivered Cognitive Behavioural Therapy for Adults with Mild to Moderate Depression and High Cardiovascular Disease (CVD) Risks: RCT	RCT to evaluate the effectiveness of internet-delivered Cognitive Behaviour Therapy (iCBT) on depressive symptom severity and adherence to medical advice and lifestyle interventions	Adults Australia N = 487/562 (88%) participants completed the endpoint assessment Online platform	In people with mild to moderate depression and high levels of CVD risk factors, a freely accessible CBT program produced small improvement in depressive symptoms, adherence and some health behaviours.	After enrolment 44 participants reported a risk of deliberate self-harm: 21 in the iCBT arm and 23 in the control arm. All were contacted by clinicians, as part of the risk management protocol. None were deemed acute clinical risks and none were removed from the study. There were no withdrawals due to study related adverse events.	Participants were recruited from a large health survey (45 and up study) and this may have led the participants to be healthier than the average population.
Green, 2018	A preliminary trial of an online dissonance-based eating disorder intervention	RCT Of women with eating disorder symptoms randomised to an online program to dissuade objectification and maladaptive social comparison compared to control (no services).	Adult Women N = 82 USA Online platform	Reduced eating disorder symptoms and risk factors, but effect disappeared at 2 month follow up and beyond.	Cardiac function of participants was monitored in-person. Participants demonstrating atypical cardiac function were referred to their primary care physicians for further evaluation.	Participants were not grouped into diagnostic sub-types.
Grist, 2018	Acceptability, Use, and Safety of a Mobile Phone App (BlueIce) for Young People Who Self-Harm: Qualitative Study of Service Users’ Experience	Design: Mixed methods Phase 1 trial of BlueIce Aim: explore the acceptability, use, and safety of BlueIce, a mobile phone app for young people who self-harm	Children N = 40 Young people aged 12–17 years attending specialist CAMHS in the UK – who were self-harming or who had a history of self-harm Mobile phone app	BlueIce was judged to be a helpful and safe way of supporting adolescents to manage thoughts of self-harming. Young people frequently described having BlueIce as reassuring and that it made them feel safer that there was a tool they could use for instant support.	No adverse effects of using BlueIce were reported. No clinician withdrew any person from the study due to an escalation of risk or the emergence of suicidal planning or attempt. No participants had reported calling the emergency numbers provided on BlueIce. 2 people reported concerns that negative feelings might be triggered by reading about their own negative feelings or events or seeing many sad faces on the mood diary. However, both these young people also reported benefits from reflecting on their mood diary entries.	The mood diary and mood lifter sections were particularly helpful in offering a way to track their moods and offering new strategies to manage their thoughts to self-harm. No participants believed BlueIce would increase their thoughts of self-harming.
Guise, 2014 REVIEW	Patient safety risks associated with telecare: a systematic review and narrative synthesis of the literature	Systematic review To identify patient safety risks with Telecare in homecare services and how risks have been addressed in training	Review 22 articles in the review. Video telehealth	Training initiatives related to safe telecare use are not described in the literature. Efforts to improve identification of safety and quality issues should increase knowledge of patient safety risks. It will also facilitate learning and competency building and development of best clinical practice for further mitigation of potential harm.	Risks associated with telecare are sometimes shared as challenges with implementation. Reframing these using the human factors systems approach to quality and safety in healthcare will help to categorise and explicate these issues. Lack (across provider organisations) of recognised standards and procedures for service delivery.	This review focused on homecare use of telehealth – not psychological services. While some telehealth tools are subject to regulation as medical devices, where reporting of adverse incidents is mandated, enforcing such regulations in the homecare setting is challenging – dependent on the voluntary actions of home healthcare providers and patients.
Hubley, 2016 REVIEW	Review of key telepsychiatry outcomes	Systematic review on telepsychiatry (2000–2016) 452 articles	Adults Articles were about: (1) satisfaction; (2) reliability; (3) treatment outcomes; (4) implementation outcomes; (5) cost effectiveness; and (6) and legal issues. Video telehealth	Providers express concerns about therapeutic rapport when using telepsychiatry yet patients are less concerned about this and most of the literature on equivalency to in-person therapy shows equivalence with respect to outcomes.	Notwithstanding legal concerns about loss of confidentiality and limited capacity to respond to psychiatric emergencies, the authors uncovered no published reports of these adverse events in the use of telepsychiatry.	Providers cite difficulties incorporating telepsychiatry into their practice, difficulty accessing training for telepsychiatry, and lack of resources.
Iorfino, 2017	Using New and Emerging Technologies to Identify and Respond to Suicidality Among Help-Seeking Young People: A Cross-Sectional Study	Design: cross-sectional study Aim: examine the use of a suicidality escalation protocol to respond to suicidality among help-seeking young people	Adolescents, young adults N = 232 (16–25 year olds) recruited from primary mental health care service or online (eg through HeadSpace) NSW, Australia Online platform	153 people (66%, 153/232) reported suicidality and provided an alert online – Further escalation (email or phone and clinical review) initiated for 35 people (15%, 35/232) Higher psychological distress (p < 0.001) and current alcohol or substance use (p = 0.02) predicted suicidality.	Online assessments are preferred and accurate for identifying suicidal thoughts and behaviors Online screening can increase access to treatment especially when integrated with professional services. There is a significant need for effective suicide escalation protocols that can appropriately and efficiently manage risk. A real-time mapping system to (local) mental health services might be useful. Further follow-up through partnerships with specific local or national suicide prevention organizations may be needed.	This was using the “Synergy Online System” has systematic assessment, automatic escalation of an individual’s data, and assisting clinical team review and decision-making processes. Ongoing development of the Synergy Online System would benefit from using longitudinal outcomes to improve the existing algorithms’ accuracy for identifying individual cases of suicidality that should be escalated and followed up immediately by a clinician and service. This paper is co-authored by Ian Hickie and Jane Burns who both have an equity stake in Innowell.
Kramer & Luxton, 2016 REVIEW	Telemental Health for Children and Adolescents: An Overview of Legal, Regulatory, and Risk Management Issues	Review of existing laws and regulations to summarise risk management issues relevant to providing telemental healthcare to children Aims: Highlight the legal and regulatory issues involved in providing this form of care to children and adolescents	Review of existing laws. Washington USA Telehealth various	Licensure: Telemental health allows care across state lines so need multiple licenses (in USA). Malpractice liability issues: Best practice guidelines and standards for in-person care apply to telemental health, although telehealth specific guidelines are emerging. This risk is unclear still as malpractice suits adjudicating what constitutes appropriate standard of care in telehealth are just emerging.	Informed consent: Review informed consent regulations to understand what is required e.g informing of potential for tech failure, confidentiality limitations etc. Also consider age of consent laws. Emergency Management: Safety planning ahead of initiating telemental health treatment e.g. local civil commitment laws, specific emergency procedures, child protection laws, having a local support person. Security and privacy: Ensure privacy, complying with laws on appropriate technology.	Other considerations: Non-traditional settings which do not have clinical staff available raise additional concerns. Undertaking a risk assessment is important.
Kreuze, 2017 REVIEW	Technology-enhanced suicide prevention interventions: A systematic review	Review of studies using technology-enhanced interventions to address the determinants of suicidal behaviour.	Review 16 studies Telehealth various	Participants in telephone, Internet, text and telephone programs experienced significant reductions in suicidal ideation whereas CDROM and Internet groups experienced non-significant reductions. Studies providing education had variability, with video interventions reducing suicidal ideation but not online programs	9–70% attrition rates depending on the study	Intervention effectiveness was variable and few studies enrolled a control group.
Lee, 2016 REVIEW	Web-based interventions for prevention and treatment of perinatal mood disorders: a systematic review	Review of web-based interventions including telehealth for prevention and treatment of mood disorders in the perinatal period, defined as the start of pregnancy to 1 year post-partum.	Review 4 included articles Online platforms	All studies conducted in the post-partum period. All reported improvement in maternal mood after intervention.	Attrition within studies ranged from 13 to 61%.	Attrition could be due to the program or the research itself.
Litz, 2014	A randomized controlled trial of an internet-based therapist-assisted indicated preventive intervention for prolonged grief disorder	RCT To assess the feasibility, acceptability, tolerability, and efficacy of an Internet-based therapist assisted cognitive-behavioral intervention for prolonged grief disorder	Adults N = 84 randomised to two groups (intervention and wait-list control) Online platform	Reductions in prolonged grief, depression, anxiety and posttraumatic stress	Non-completers did not differ on any variable except work status. With those working more likely to drop out.	Sample was mainly white, well-educated and grieving loss of a spouse.
Lowery, 2014	Distributing Medical Expertise: The Evolution And Impact Of Telemedicine In Arkansas	Design: A descriptive report of the Arkansas telemedicine system, and its evolution since 2003 Aims: Describe the evolution and impact.	Adults Descriptive report Video telehealth	Describes the evolution of the Arkansas telemedicine program.	Risks: technology limitations, inadequate user training, concerns about credibility of experts consulted about technology, accuracy of services provided. Institutional pressures (e.g. financial disincentives, workflow readjustment). Grants can be used to fund set-up but sources of funding are needed to maintain the service.	Technical issues: The centre had full-time technical support staff. To communicate the value of telehealth, the centre established relationship with communities to generate trust.
Lovell, 2017	Clinical effectiveness, cost-effectiveness acceptability of low-intensity interventions in management of obsessive compulsive disorder: the Obsessive–Compulsive (OCD) Treatment Efficacy RCT (OCTET)	RCT To determine clinical effectiveness and cost-effectiveness of two low-intensity CBT interventions (supported computerised cognitive–behavioural therapy (cCBT) and guided self-help compared with waiting list controls.	Adults NHS trust in the UK N = 475 adults with OCD randomised over 2011–2014 Online platform	Low-intensity interventions did not lead to clinically significant improvements in OCD symptoms, economic analysis over 12 months suggests low-intensity interventions are cost-effective and may have an important role in OCD care pathways.	13 serious adverse events occurred during the RCT; 10 were unplanned hospitalisation two of these due to self-harm, suicidality. 12 other events were deemed unrelated to the program/RCT. One of these was a participant leaving employment. 173 other non-serious adverse effects were reported which related to change in mental health status, hospitalisation, medication change.	Detailed reporting of adverse events was present in this paper.
Luxton, 2010 REVIEW	Safety of Telemental Healthcare Delivered to Clinically Unsupervised Settings: A Systematic Review	Design: Systematic review (of published papers until 2009) (adults) (majority for depression, several for veterans) Aim: provide an overview of safety issues	Adults N = 9 studies that evaluated delivery of telemental healthcare to unsupervised settings. Most of the studies used telephone as the technology. The 9 included studies were conducted in USA (7), Spain (1) and Germany (1) Video telehealth	Common safety plans described were exclusion criteria to prevent participants from becoming at risk. Other studies reported safety plans requiring monitoring of symptoms during treatment. Another potential safety concern for telehealth is that patients might experience increased feelings of social isolation compared in-person. Clinicians could consider the evaluation of social isolation and availability of social support.	2/9 of the studies reported events requiring use of safety procedures. The first was a depression screening study in which one participant indicated suicidal ideation and the local police were called to check on this person. The other study involved home health monitors and one participant stopped recording any data, the local clinician was called and the team contacted the participant who had forgotten to enter data and all was well. Another study described connection difficulties for one client because of his employer’s network firewall. Overall, the data suggest that the use of safety plans in these studies, including those focused on prevention of adverse effects, was effective.	If the telehealth infrastructure is not reliable and there are not any redundancies built in, patients may be at risk if the system unexpectedly fails at critical times. Planned use of telephone at these times is wise. Some technical knowledge to fix times of challenge is recommended. Technical malfunction has the potential to frustrate the patients and thus reduce treatment satisfaction and adherence. The setting might also create risks. Poorly lit rooms, small video monitors, or inadequate audio can cause eye strain or hearing problems. Safety plans are defined as Pre-determined procedures for reducing risk, preventing adverse reactions, and for responding to adverse events when they occur. These include appropriate screening processes for risk (e.g. suicidality), monitoring of patients during the course of treatment, and the establishment of safety protocols to ensure that the best methods for resolving adverse events are followed when they do occur.
Luxton, 2014	Suicide risk management during clinical telepractice	Commentary/policy paper	Commentary paper Video telehealth	Suicide risk assessment should be an ongoing process as risk levels are fluid. Telehealth sessions should begin with: verification of patient location and contact information. Patients are also asked whether they feel that their home environment is safe and private. Additional questions assess whether the patient appears intoxicated or otherwise dishevelled, distressed, or upset; suicidal desire and ideation. If patient is deemed high-risk coordination for possible hospitalisation is started.	Collect alternate methods of contact in case videoconferencing connection is lost, identify a support person who can assist in an emergency and ensure “release of information form” are signed for use in an emergency, tailor safety plans to the region. The screening protocol, safety protocol and telehealth session procedures developed by Luxton et al. are presented in full in this paper.	Maintaining a risk management or safety protocol that guides the assessment process, encourages consultation with colleagues and supervisors, and informs clinical decision making can reduce patient and clinician anxiety, enhance accuracy and reliability of the assessment, and thereby, supports patient safety. For continuity of care, provide a list of local community mental health resources that they may keep as a reference.
Luxton, 2015	An Evaluation of the Feasibility and Safety of a Home-Based Telemental Health Treatment for Posttraumatic Stress in the U.S. Military	Feasibility study Aim: to assess feasibility of telehealth into home for veterans with PTSD. Safety data collected: adverse events, psychiatric hospitalizations, suicides and nonfatal suicide-related behaviors, number of times the patient support person was utilized during treatment, treatment adherence, and frequency of requests for patient or therapist technical support.	Adults N = 10 veterans USA Video telehealth	Clinically significant reductions in posttraumatic stress symptom severity and depression symptoms were observed. Soldiers indicated high levels of satisfaction with the treatment, and there were no adverse events requiring activation of emergency safety procedures. Technical problems associated with the network were observed but successfully mitigated.	There were no safety concerns that required activation of the emergency protocol. Several participants did indicate suicidal ideation but risk was assessed, they remained in the service and improved. Suicide assessment and risk management Standard Operating Procedure (SOP) to assess and document risk including: current ideation, presence of a plan, suicidal intent, history of previous attempts, and degree of impulsivity. Risk correlates (e.g. recent loss, financial problems, preparatory behavior (e.g. available means), and other risk factors (e.g. substance dependence)- at baseline and at each session.	SOP also included assessment of clinical factors: intoxication, disorientation, severe emotion dysregulation, questions of home environment, such as “Is anyone else at home today?” and “Do you feel that your environment is safe and private?” This checklist is also used to document telehealth equipment and connectivity status, adequate lighting, and any disruptions to session.
Marzano, 2015	The application of mHealth to mental health: opportunities and challenges	Commentary paper	UK Mobile phone apps	Some apps/telehealth have collected data without explicit consent, for marketing, commercial, and other purposes. Consent and confidentiality concerns need to be addressed and compliance to governance such as: UK Data Protection Act, European Data Protection Regulation, copyright, and licensing laws.	Only five apps tested for clinical effectiveness among more than 3000 mental health apps available for download when this paper was published. The NHS Health Apps Library contains (as of 11 September 2015) 26 apps classified under mental health, having been reviewed by a clinical assurance team.	Guidelines on the development of commercial health and wellbeing apps by the British Standards Institution is a step in the right direction yet it remains to be seen whether these voluntary standards will ultimately mean compliance with EU Medical Device Directives.
Maheu, 2012	Future of Telepsychology, Telehealth, and Various Technologies in Psychological Research and Practice	Design: Overview/ commentary article USA and Australia	Adults and children Potential risks mentioned: Cloud computing and storage of data. Video telehealth	Cloud computing may be more secure than other options. Revision of licensure to help ease cross jurisdictional practice is needed.	Clients pay for life coaches and others who practice online. These people may be inadequately trained for telemental health care and are unlicensed to practice in areas they service publicly but are without regulation.	Providers need to stay abreast of technological developments in privacy, security, and accuracy of their communications, and evaluate ethical and practice guidelines for informed consent. Provider needs to know how to access local resources and emergency backup teams where client is based. Evidence of practitioner errors in judgment in online behaviours is beginning to grow. For instance: use of profanity, boundary infractions. There is a need for psychologist training in telehealth, in self-assessment, and in continuing education.
McGinn, 2019 REVIEW	Recommendation for Using Clinical Video Telehealth with Patients at High Risk for Suicide.	Design: narrative review of telehealth with high-risk patients Aim: review literature and provide recommendations on telehealth with high-risk patients	Number of articles reviewed not reported, years not reported, search strategy not reported.	None of the literature indicates that patients at high-risk for suicide should be excluded from telehealth services. Telehealth mental health services for this population can increase access to care and reduce hospitalisations.	The literature provides evidence that emergency situations that arise can be successfully managed over telehealth. Telehealth has been used successfully to assess and treat people experiencing psychiatric emergencies. Telehealth can be particularly helpful for rural areas.	There is evidence that telehealth mental health services with this high-risk population can increase treatment adherence in rural patients following psychiatric hospitalisation. Recommendations: location obtained, educate family members re signs of risk, stay on telehealth while arranging emergency services if needed.
McLean, 2013 REVIEW	The Impact of Tele-healthcare on the Quality and Safety of Care: A Systematic Overview	Design: Systematic review of systematic reviews Aim: scrutinise the evidence regarding the benefits, risks and costs of tele-healthcare	N = 80 reviews were included (1997–2011) Video telehealth	Many studies showed no differences in outcomes between tele-healthcare and usual care. Telehealth: reduced hospitalisations particularly for people with most severe disease. Evidence on the cost-effectiveness of tele-healthcare remains sparse.	“Patient safety considerations were absent from the evaluative tele-healthcare literature.”	Telehealth research continues to contain a large number of short-term (<1 yr) feasibility studies with under 20 participants. Many studies do not adequately describe the intervention. This makes it difficult to disentangle the contributions of technological and human/ organisational factors on the outcomes reported.
McLean, 2021 REVIEW	Exploring the Efficacy of Telehealth for Family Therapy Through Systematic, Meta-analytic, and Qualitative Evidence	Design: Narrative and systematic reviews with one meta-analysis Aim: Multiple aims. Review professional guidelines for family therapy on telehealth, review of studies on family telehealth	N = 20 studies in the systematic review and 13 studies in the meta-analysis	Existing telehealth guidelines focus on operational matters not the therapeutic process while on telehealth. Meta-analyses showed telehealth to be equivalent to face-to-face for child behavioural problems and parental depression.	Authors suggested caution about suitability of certain cases for telehealth therapy – by individual risk or complexity including history of trauma, assault or suicidal ideation. Suggested accommodations for clients with these kinds of risk profiles included: “slowing the work down’ spending more time with the client to understand their background, current situation and priorities to determine potential risk and to arrange risk management plans prior to starting any telehealth therapies.	Telehealth interventions to prevent an at-risk client from walking out of a telehealth consult include: establishing permission and processes for discussing “hot issues” during the session; indicating when the therapist would need to intervene, providing emotion regulation strategies and checking in on the client’s well-being after the session. Typical procedures for contacting police remain for telehealth as they do for in-person sessions.
Morland, 2015	Home-based clinical video teleconferencing care: Clinical considerations and future directions	Design: Commentary paper focused on adults Veterans affairs, includes youth USA	Adults and children Telehealth limited in individuals with moderate to severe cognitive difficulties or sensory impairment (such as severe hearing loss). Patients with impaired reality, severely psychotic or paranoid individuals, those with poor impulse control, uncontrolled anger or self-injurious tendencies may be more appropriate for office-based face-to-face care. However, evidence to date indicates telehealth services are well-accepted and not causing exacerbation in symptoms among these patients. Video telehealth	Risk management protocols are completed as part of the informed consent process. After initial phone contact, a packet of information that must be returned prior to the first treatment session is sent out to each patient. This packet contains the written informed consent, the risk management protocol, and other treatment information.	An initial phone call is used to assess if the client is appropriate for telehealth treatment. The provider is responsible for determining telephone numbers for law enforcement and emergency services in the patient’s location. It is important to discuss the emergency plan protocol with a patient prior to the first session, as well as during ongoing sessions, to normalize procedures and make modifications as necessary (e.g. new address of location, new primary support person).	The risk management protocol includes a review of the patient’s presenting problem and background that includes their history of suicidal ideation and/or attempts, homicidal ideation/ attempts, psychiatric hospitalizations, and current substance use. In addition, the protocol asks patients to verify the address and telephone number where the sessions will be occurring and identify other individuals who will be in the home or location during the sessions. An additional safety measure that is recommended, but not necessary, is to have a patient complete a release of information for a primary support person. This person could be notified in case of emergency and could even help if emergency procedures are implemented during a session. The provider will stay with the patient via telehealth until transport to safety has occurred.
Nielssen, 2015	Procedures for risk management and review of crisis referrals from MindSpot Clinic, a national service for remote assessment and treatment of anxiety and depression	Description of the procedures used to manage risk, and an audit of case summaries of patients who were urgently referred for crisis intervention.	Adults Australia N = 2599 people 2013–2014 Video telehealth	Suicidal thoughts reported by 2366 (26.1 %) and suicidal plans reported by 213 (2.4 %). There were 51 acute referrals, of whom 19 (37.3 %) lived in regional or remote locations. The main reason for referral was the patients’ self-report of imminent suicidal intent. The police were notified in three cases, and in another case an ambulance attended after the patient reported taking an overdose. For the remaining acute referrals, MindSpot therapists were able to identify a local mental health service or GP.	Most of the crisis referrals were made at the time of assessment, with only three patients referred for urgent care as a result of crises emerging while in treatment. 0.6 % of those seeking assessment were referred to local mental health services for urgent face to face care and this occurred successfully for every case.	The average amount of therapist time to make an acute referral was 4 hours (telephone calls with patient, discussions with supervisors and time spent locating and contacting local health services).
Olden et al., 2017	Pilot Study of a Telehealth-Delivered Medication-Augmented Exposure Therapy Protocol for PTSD	Pilot RCT aimed to investigate the feasibility of conducting a research trial with exposure therapy delivered via videoconferencing.	N = 11 adults at risk for PTSD due to their occupation (eg firefighter, military) USA Video telehealth	Treatment completers demonstrated decreases in PTSD and depressive symptomatology, high therapeutic alliance, treatment satisfaction, and telehealth satisfaction.	There were no significant technical, medication, or safety issues, and no clinical emergencies. 36% of participants dropped out. Those who dropped out were more likely to have depression or sexual trauma – which is in keeping with the literature.	Small sample
Russo, 2017	What drives attitude towards telemedicine among families of pediatric patients: A survey	Design: Survey in 2015 With Logistic regression to understand relationship amongst factors Aim: investigate sociodemographics, clinical, technology, attitude towards telemedicine, perceived advantages and fears	Children and their parents N = 751 families Tertiary care children’s hospital Rome, Italy Video telehealth	Results show an interest towards telemedicine: large proportion of families were keen on using an app issued by the hospital and more than half were willing to do a televisit.	Lack of trust towards telemedicine was most common concern (30%) followed by fear of excessive responsibilities for the family (28%). Fear for privacy was not a strong concern (57% considered it of little importance or not important at all).	There was a willingness to use telemedicine amongst all families, unrelated to disease type, severity or distance to hospital. Previous research has shown sociodemographic variables are not predictive of interest in telemedicine and confirmed in this study. Paediatric context is particularly favourable for telemedicine services: parents and patients are young and familiar with new technologies
Salisbury, 2016	Effectiveness of an integrated telehealth service for patients with depression: a pragmatic randomised controlled trial of a complex intervention	RCT Compare telehealth services for depression compared with in-person	England Patients enrolled for 43 GP practices N = 609 participants 2012–2013 Video telehealth	Compared with usual care, intervention participants reported improvements in anxiety, better access to support and advice, greater satisfaction with the support they received, and improvements in self-management and health literacy.	During the trial, 70 adverse events were reported by participants (equal across intervention and control arms), one of which was related to the intervention (increased anxiety from discussing depression) and was not serious.	Few details on the 70 adverse events.
Sansom-Daly et al., 2016 REVIEW	Consensus Among International Ethical Guidelines for Provision of Videoconferencing-Based Mental Health Treatments	Review of best practice guidelines for telehealth for mental health	Review of 19 guidelines or position papers from 7 countries. 11 firm and 123 other recommendation regarding the appropriateness of e-mental health, competence, legal and regulatory issues, confidentiality, consent, professional boundaries, and crisis management. Video telehealth	Crisis management guidelines include recommendations: support person be nominated, continual assessment of risk throughout therapy, phone numbers of eg police on hand, stay online with client until emergency services arrive in cases of crisis, discuss crisis plan in advance, discuss availability of therapist in case of crisis occurring between sessions, discuss potential misunderstandings due to technology.	Provider needs to be familiar with mandatory reporting in the jurisdiction in which the client lives. Provider needs risk management plan in place and plans for departure of client or dropping of telehealth signal in times of crisis. Provider to confirm location of client at start of each session.	Valuable review of guidelines with detailed supplemental files. Only 6/19 guidelines discussed e-mental health services for high risk populations such as those with psychosis, although there is no evidence about who e-mental health services is best for, or could harm. 70% of the guidelines were for psychologists providing telehealth.
Santesteban et al., 2020 REVIEW	Telehealth interventions for schizophrenia spectrum disorders and clinical high-risk for psychosis individuals: A scoping review	Review of studies using telehealth for clients with a diagnosis of schizophrenia-spectrum disorder and those who may be considered to be in the very early stages of psychosis (clinical high risk)	Review 14 studies included Video telehealth	There was no study reporting on videoconferencing interventions for individuals at clinical high risk. All the studies reported that videoconferencing implementation was feasible, and most of them described high acceptance by individuals with a schizophrenia-spectrum disorder.	Studies addressing risk management reported that videoconferencing was safe and there were no risk issues during treatment. Anecdotal cases of inpatients in acute phase of psychotic illness reported less satisfaction with videoconferencing than previous face-to-face interventions (i.e. suspicion regarding the videoconferencing device).	Individual studies reported the need to repeat instructions more times in videoconferencing and feeling more fatigued after each interaction than in their face-to-face interactions due to the long exposure to the screen.
Sasangohar, 2020	Adapting an Outpatient Psychiatric Clinic to Telehealth During the COVID-Pandemic: A Practice Perspective.	Case study of one art therapy group program in the USA Aims: to describe the strengths and challenges of telepsychiatry implementation	USA. For patients with complex and persistent psychiatric conditions. No sample size reported.	Patients were well-engaged, view into patient’s home lives was useful for clinicians. Need more safety planning, need to work on interprofessional team collaboration when telehealth is used, allow for delays and interruptions	Managed risks, even with suicidal patients, without any adverse effects – by having emergency contact details, and offers to directly admit patient to inpatient ward after telehealth consult if needed	Authors stated that managing risk in telehealth requires more time, effort and creative problem-solving than for in-person mental health care. Need to ensure health professional doesn’t use personal web platform. Physical environment in telehealth does require thought (eg hide laundry)
Thabrew, 2018 REVIEW	E-Health interventions for anxiety and depression in children and adolescents with long-term physical conditions	Review of trials To assess the effectiveness of e-health interventions vs controls for treating anxiety and depression in children and adolescents with long-term physical conditions.	Searched inception to 2017 Included 5 trials of 3 interventions Telehealth various	Only five low quality trials were found. The very low-quality of the evidence means the effects of e-health interventions are uncertain at this time for this population, especially in children aged under 10 years.	No adverse effects were reported by any authors.	Interventions were designed and evaluated by the same people, this may have introduced bias.
Torous, 2014	Promise and perils of digital psychiatry	Design: discussion paper Harvard, USA	Adults Concerns include: clinician may not detect observable traits or signs; difficult to verify the patient and clinician are communicating with one another and no one else. A concern that may produce both ethical and privacy violations. Connection must be secure and patient data must be safe. Telehealth various	Text messages, email have their own risks for data security and privacy as well as possible misunderstanding of tone and attitude. Clinicians should never friend patients on social media, as such encounters may present complex boundary violations and may be interpreted by medical boards of registration as invitations of inappropriate closeness.	Clinicians and patients should establish a safety plan for psychiatric/ psychological emergencies, such as suicidal ideations or threats. Many telepsychiatry clinics train onsite staff to be able to respond when patient safety issued arise.	Clinicians must limit how patients may contact them though their website, as they risk creating an actual but “inadvertent” patient-clinician relationship. The more interactive the website, the higher the risk of creating this unintended patient physician relationship. Clinicians are responsible for information posted on their websites: inaccurate or out of date health data may also increase risk of liability if harm results. Clinicians must also clearly define any conflicts of interest on their websites, if they are, for example, promoting a certain therapy. Breaches of confidentiality on blogs written by clinicians have led to the termination of clinicians in the past.
Wade, 2012	A qualitative study of ethical, medico-legal and clinical governance matters in Australian telehealth services	Qualitative study to understand how Australian telehealth service providers perceived and addressed ethical, medico-legal and clinical governance matters arising from service delivery.	Qualitative interviews and analyses of N = 37 telehealth clinicians and managers perceptions. In 6 jurisdictions of Australia. 68% to rural areas. 11 medical specialities, surgery, mental health, paediatrics, nursing and allied health. Video telehealth	Those interviewed reported some problems with privacy and security, and variable informed consent practices. No examples of malpractice were raised, although there was a common misperception that distant providers were not responsible for clinical care. With respect to clinical governance, telehealth was seen as enabling improved quality, integration and implementation of evidence-based care, and to be a major support for the rural health workforce.	Potential medico-legal risks: no local service to deal with adverse events, making video recordings, which could be subpoenaed in court, and difficulties in obtaining indemnity insurance. Risk mitigation strategies include: limiting type of work done, selecting appropriate patients, and having a lower threshold of concern. No malpractice proceedings were reported. Others said telehealth reduced medico-legal liability by avoiding transport and hence increasing patient safety, and by providing additional expertise by having second opinions.	Lack of risk management policies was seen as a risk in itself.
Whaibeh, 2020	Reducing the Treatment Gap for LGBT Mental Health Needs: the Potential of Telepsychiatry	Design: narrative review Aim: describe the potential of telepsychiatry to better meet the needs of the LGBT community	Narrative review	Lesbian, gay, and bisexual individuals regularly report discriminatory behaviour and dissatisfaction with their providers. Finding culturally competent psychiatrists can be difficult particularly for people living in rural settings.	Telepsychiatry offers the opportunity of providing higher quality of care from culturally competent psychiatrists, regardless of the patient’s geographical location. The paper states that by linking LGBT patients to culturally sensitive providers, telepsychiatry may reduce the negative attitudes of certain providers.	Rather than increasing risk, this paper describes how telepsychiatry can reduce the overall risks for a person who identifies as LGBT and who is seeking mental health care.
Whittaker, 2017	MEMO an mHealth intervention to prevent the onset of depression in adolescents a double‐blind, randomised, placebo‐controlled trial	Design: RCT Intervention: cognitive behavioural therapy-based programme (MEMO CBT) to be delivered via multimedia mobile phone messages to teens Aim: prevent onset of depression	Children N = 855 students (13–17 years old) 15 high schools in Auckland, NZ Mobile phone text messages	No evidence of benefit from the mobile phone CBT intervention compared with a control. Depression prevention remains a challenge.	Risk protocol outlining referral and management of students identified as high risk of depressive disorder and/or suicidal ideation agreed before the commencement of the study. Serious adverse events (SAEs) were identified by questionnaire administered to the participants at the post-intervention and follow-up interviews. School guidance counsellors were advised to alert study team to any adverse events. None occurred.	Enrolment into the study excluded students with current diagnosed depression or self-harm. Students identified at high risk of a depressive disorder or suicide ideation/behaviour during the study (after randomisation) were referred to the school guidance counsellor for additional support and intervention but were not excluded from the study. Researchers stated in future research they plan to be “be less stringent in our exclusion of participants.” Meaning they would opt to include teens with depression and self-harm in future.

Quantitative data synthesis was not planned due to the heterogeneity of study questions, designs and findings. Instead, a qualitative data synthesis was employed using our completed data extraction file as a foundation. We initially identified and coded information about each study in terms of its methods, settings, population, findings, safety data and critical appraisal. In a separate working document, we organised the coded information into potential categories and these categories were then grouped into themes and these became our sub-headings in the Results section.

Results

The search strategy (after removing duplicates) resulted in 1,497 papers (Supplementary PRISMA Flow Chart Figure 1). For all 1,497 papers the titles and abstracts were searched and those papers which did not fit the inclusion criteria were removed. The remaining 174 articles were assessed for eligibility using the full text of the article. At this stage, 119 articles were removed as they were about in-person services or were virtual reality interventions or were focused on self-help mental health websites or apps without engagement with a clinician. One of these papers excluded was a duplicate paper on the same study, so it was removed, leaving a final 55 articles for full review and data extraction (Table 1). Articles were retained which focused on psychology, psychologists, psychiatrists, social work, occupational therapists as well as nurses, counsellors and other primary care providers providing mental health services using telehealth. Articles in English, for all age groups (adults and children) were retained. Articles were obtained from the University of Sydney online library. The four articles not available at the University of Sydney were all obtained from other online libraries in Australia.

During our scoping it became clear that risks were not systematically reported, nor were there any papers specifically aiming to report risks or adverse effects. Across all papers, there is the possibility of publication and reporting biases in that authors/organisations may not publish the risks or adverse effects encountered with telehealth in peer-reviewed papers, even if this was reported appropriately to the respective safety boards during a research study. This may be particularly true for commercial entities. There is little literature regarding adverse events with respect to mental health care provided using telehealth, nor analyses of adverse events occurring during mental health care provided using telehealth which classify events regarding severity, expectedness, contributing factors and potential relatedness to care, program, product, or device. Attrition bias is also possible within the papers included in this review. Clients/patients may not adhere to treatment via telehealth and those who attend or remain in treatment may differ and may be clients with whom less risk or adverse effects are encountered.

Findings

Summary of findings

Of the 55 included articles, 20 were reviews, 6 were papers providing a commentary/overview or a framework and 29 were original research articles (Table 1). The papers in this review can be divided into several categories. Multiple papers sought to report risks of mental health services provided over telehealth and observed no risks. Other papers reported that attrition or drop-outs from therapy programs provided via telehealth were the greatest risk. One paper reported on mental health/psychiatric emergencies and the role of telehealth. There were several papers reporting the risks of mental health services provided over telehealth by diagnosis including several papers reporting specifically on risks and risk management of suicidality and self-harm. There is an emerging literature about risks related to about mobile phone apps and text messages in relation to mental health care. Three papers reported on risks associated with telehealth services provided as group therapy. Multiple papers were case studies of one organisation’s experience with delivering mental health care using telehealth. Last, there were publications with suggestions for risk management when providing mental health services over telehealth.

No observed risks

Several papers reported observing no risks associated with providing mental health services using telehealth (Bucci, Cheng, Childs, Egede, Engel, Iorfino, McLean, Olden, Sasangohar, Thabrew, Whittaker). Bucci’s (2018) study on a digital intervention for people with psychosis reported zero adverse events. Similarly, for Cheng’s (2018) study of psychiatric telehealth services which reported no adverse events. Child’s (2020) psychiatric telehealth group therapy study found no adverse events. Egede et al.’s (2009) RCT of psychotherapy for depression, and Engel’s (2016) RCT for PTSD, in older veterans via telemedicine reported no adverse events. Another paper reports on the Australian “Synergy online system” which aims to screen youth for suicidality and no adverse events were found (Iorfino, 2017. McLean (2013) included 80 reviews (published between 1997 and 2011) on telehealth in general and found none of these 80 reviews mentioned safety, risks or adverse effects. Olden et al. (2017) studied telehealth for people with PTSD and reported no adverse events. Sasangohar (2020) described an art therapy program provided over telehealth during COVID-19. They found they could manage risks with suicidal patients by having emergency contact details and offers to directly admit patient to an inpatient ward if needed. Thabrew (2018) reviewed e-health interventions for anxiety and depression and observed no adverse events. An RCT in New Zealand to prevent onset of depression in adolescents, by Whittaker (2017) gathered data on adverse events via questionnaire and none were reported. A final paper was an overview article which highlighted risks regarding mental health workforce accreditation, regulation and the growth of life coaches and others who practice online (Maheu, 2012).

Attrition/drop-outs

Other papers reported the most frequent and/or significant risk experienced with providing mental health services using telehealth was attrition or drop-outs of clients/patients from the therapy program (Davies Deady, Fisher, Flodgren, Kreuze, Lee, Litz). None of these papers reported on attrition by study arm if they had a control arm, three papers examined factors related to attrition and found illness severity (Davies et al., 2014), and outcomes (Deady, 2017) were not associated with attrition, but that working outside the home did increase attrition (Litz, 2014). Most authors stated attrition was similar to that experienced in face-to-face care but did not quantify this. The 2014 review by Davies and the 2017 review by Deady found that attrition was common when mental health services were being provided by telehealth (12/17 studies in the Davies review) but attrition or drop-outs were not found to be related to outcomes. Of the trials included in Davies which examined it, illness severity was not associated with attrition. Another review (Fisher et al., 2019) of 8 RCTs of psychological therapies for children with chronic pain reported drop-outs as the only risk observed. Another review of telehealth for mental health care (Flodgren, 2015) included 7 RCTs total (1 for children, 6 for adults) and reported upon lack of adherence to therapy and attrition from the trials. Kreuze (2017) reviewed technology-enhanced suicide prevention programs and found attrition rates to be the most common risk, ranging from 9 to 70% attrition depending on the study. A review by Lee (2016) similarly found attrition rates to be the biggest experienced risk (up to 60% attrition) for programs aiming to prevent and treat perinatal mood disorders. Attrition was the only risk mentioned in the research study by Litz (Litz, 2014) and this was found to be higher in those working versus those not working. Interestingly, the paper by Whaibeh (2020) offered the opposite experience, stating that telehealth provides an opportunity for people who are lesbian, gay, bisexual or transgender (LGBT) to connect with mental health services which are more respectful -and this may lead to less attrition.

Mental health emergencies

Hubley (2016) wrote about legal concerns with loss of confidentiality and the risk of limited capacity to respond to psychiatric emergencies during telehealth services. McLean et al. (2021) conducted a review of family therapies provided over telehealth. McLean found a difference in the recommendations in different papers. In some when a mental health urgency arose (for instance, serious conflict between family members), some therapists felt it was more difficult to manage when family members were on different screens in different locations as they felt this removed the ability of the other family members to detect the early signs of conflict. Yet other therapists felt there was a greater sense of safety when family members joined from separate devices in separate spaces as they felt the risk of physical conflict was reduced.

Suicide/self-harm

Several papers specifically addressed risks of providing telehealth psychology services with respect to suicide and self-harm (Christensen, Grist, Iorfino, Luxton and Luxton, McGinn, Sasangohar). One review (Christensen, 2014) gathered evidence on the effectiveness of e-health interventions to manage suicidal thoughts, and interventions when suicide risk is identified in social media postings. Authors wrote that it is unknown if sharing thoughts and feelings on social media is beneficial to an individual or risky. This review concluded that online interventions to prevent suicide do not cause harm. Grist (2018) evaluated the safety of a mobile phone app for adolescents who self-harm. No adverse events were reported however two of the 40 participants felt the app made them feel more sad. Iorfino (2017) reported on screening youth for suicidality through the “Synergy online system” and found no adverse events. The review by Luxton (2010) found that two of the 9 telehealth studies reported events requiring the use of safety procedures. Noting that telehealth was typically telephone only in these studies. One event was suicidal ideation and another participant stopped responding and safety measures were acted upon. Luxton’s 2014 paper is about suicide risk management during clinical telepractice and contains screening and safety protocols to manage suicide risk during telehealth services. McGinn (2019) writes that none of the existing literature indicates that patients at high-risk for suicide should be excluded from telehealth services. Telehealth mental health services for this population can increase access to care and reduce hospitalisations, and any emergencies which arise during a telehealth session can be managed appropriately in that session. Sasangohar (2020) stated they were able to manage risks, even with suicidal patients, without any adverse effects. They stated this was easier for patients they had met in-person prior to telehealth consultations, however, due to COVID-19 several patients were not able to be seen in-person and continued with telehealth without any adverse effects.

By diagnosis

Several papers were about the risks of mental health care provided over telehealth with respect to specific diagnoses; including: psychosis, schizophrenia, depression, PTSD, eating disorder, OCD, Tourette’s, and bipolar disorder (Santesteban, Bucci, Daker, Salisbury, Glozier, Olden, Green, Lovell, Andren and Bauer respectively). Santesteban et al. (2020) reviewed 14 papers, concluding that telehealth is safe for people with schizophrenia and psychosis. Bucci’s (2018) study on a digital intervention for people with psychosis reported zero adverse events. Another paper, by Daker-White and Rogers (2013) discussed clients with schizophrenia who had concerns with surveillance and control with the internet often seen as a potential site of risk and danger. Salisbury (2016) studied telehealth services for people with depression and during the trial, 70 adverse events were reported by participants (equal across intervention and control arms), one of which was related to the intervention (increased anxiety from discussing depression) and was not serious. Glozier’s (2013) RCT of online treatment for people with depression reported that after enrolment 44 participants reported a risk of deliberate self-harm: 21 in the intervention arm and 23 in the control arm. All were contacted by clinicians, as part of the risk management protocol. None were deemed acute clinical risks, and none were removed from the study. Olden et al. (2017) studied telehealth for people with PTSD and reported no adverse events. Green (2018) evaluated online care via an RCT for people with an eating disorder. As part of its risk management, this study had a protocol to refer individuals with atypical cardiac function to primary care physicians for further evaluation. Lovell (2017) reported 13 serious adverse events in their RCT of online treatment for obsessive compulsive disorder (OCD); 10 were unplanned hospitalisations with two of these due to self-harm, suicidality and 12 other events were deemed unrelated to the program/RCT. The study by Andren et al. (2019) investigating a therapist guided online intervention for children with Tourette’s disorder reported 12 adverse events which were outlined in a supplemental file and included: depressive symptoms, irritation/anger, increased tics, anxiety/worry, stress, pain, obsessions, concentration problems and tiredness. Depressed mood and irritation/anger were the most common adverse events reported by Andren. There were two suicide attempts in the study by Bauer (2016) which was an evaluation of telehealth services for bipolar disorder. However, this represented a rate of 2% which is similar to the 2.2% reported rate for suicide attempts in this population, according to the authors.

Apps and text messages

There were two papers specifically about mobile phone technologies and mental health services. One paper was about mobile phone applications (apps) (Marzano) and the other about text messages (Chen). Marzano (2015) highlights there were only five apps tested for clinical effectiveness among 3000 mental health apps available in 2015, with an additional 26 mental health apps reviewed by the clinical assurance team in the National Health Service (NHS) in the UK. The other paper, a pilot study (Chen, 2017) of text message-based interventions for adolescents with depression or autism spectrum disorder (ASD) reported that the adolescents found daily questions about their mood and sleep tiresome and unhelpful to their care.

Group therapy

Three papers discussed unique risks regarding group therapy over telehealth (Childs, Gettings and Sasangohar). Childs et al. (2020) described the safety and feasibility of intensive outpatient psychiatric group interventions through COVID. Gettings et al. (2015) trialled telehealth support groups for siblings and stated risks included technical problems, the lack of visual cues, higher potential for misunderstanding, and no opportunity for individuals to communicate without it being heard by entire group. The third paper (Sasangohar et al., 2020) about an art therapy group using telehealth did not mention any risks encountered but authors recommended more careful safety planning for high-risk patients and wrote that professional boundaries need to be attended to as it felt like a much more informal virtual setting.

Risk management

Several authors wrote about risk management when using telehealth for mental health services (Aikens, Di Carlo, Drum, Guise, Kramer, Luxton, 2014 and Luxton, 2015, McGinn, Morland, Sansom-Daly, Torous and Wade). Aikens (2015) examined the effect of having a support person involved in a mobile phone treatment intervention. The more severely ill patients chose to have a support person and those who did gained the most benefit from the program. The authors hypothesised that support people want to help but typically lack the support structure and tools to do so and so this program may help in that regard. Drum and Littleton (2014) wrote about therapeutic boundaries in telepsychology to reduce risks such as: lack of punctuality due to failures in technology, the use of telepsychology in non-business hours which may lead to an expectation of instant feedback, the potential for excessive communication, the potential for unintentional favouritism to certain clients by interacting more with some than others. Drum and Littleton (2014) also wrote about the potential risks during telehealth when it involves communication in writing, suggesting that the health provider should exercise caution in using emoticons, excessive punctuation (eg!), or messaging acronyms as these may make the professional relationship less clear. The review by Guise et al. (2014) suggested that reframing risks using the human factors systems approach to quality and safety in healthcare will help to categorise and explicate risk issues and further develop standards and procedures. Kramer and Luxton (2016) states that the risks are unclear as malpractice suits (in the USA) regarding what constitutes appropriate standard of care in telehealth are just emerging. Obtaining consent may require additional features such as informing the client of the potential for the technology to fail and potentially of the limitations in technology for confidentiality (Kramer & Luxton, 2016). The Luxton, 2014 paper provides a suicide screening protocol, safety protocol and telehealth procedures to manage suicide risk during telehealth. The 2015 Luxton paper shares the authors’ telehealth-based suicide assessment and Standard Operating Procedures (SOPs). They used these SOPs to assess and document risk including current ideation, presence of a plan, suicidal intent, history of previous attempts, and degree of impulsivity. They also gathered data on risk correlates (e.g. recent loss, financial problems, preparatory behavior and other risk factors e.g. substance dependence). This was done at each session. In this study, several participants did indicate suicidal ideation, but risk was assessed, they remained in the service, and they improved. Morland (2015) describes a risk management protocol for telehealth services for clients with a clinical history of suicidal or homicidal ideation/attempts, hospitalisation, substance use and asks clients to verify the address and telephone number where the session is occurring (each time). Ideally a client will complete a release of information for a primary support person who would be called in an emergency. Sansom-Daly et al. (2016) summarised international ethical guidelines for mental health services over telehealth. This paper makes 123 recommendations including, for example: that the provider needs to be familiar with mandatory reporting in the jurisdiction in which the client lives, the provider needs risk management plans in place and plans for departure of client or dropping of telehealth signal in times of crisis, the provider needs to confirm location of client at the start of each session amongst other recommendations. Torous (2014) writes that clinicians must limit how patients may contact them as they risk creating an inappropriate patient-clinician relationship. Last, Wade (2012) conducted a qualitative study of the ethical and legal aspects of telehealth delivery in Australia. Wade summarised potential medico-legal risks including not having any local service to deal with adverse events, making video recordings, which could be subpoenaed in court, and difficulties in obtaining indemnity insurance. Risk mitigation strategies suggested by Wade et al. (2012) include: limiting the type of telehealth work, selecting appropriate patients, and having a lower threshold of concern.

Discussion

COVID-19 has accelerated the uptake and use of telehealth, due to necessity and eased by policy and billing changes during this time. The longevity of these policy and billing changes may be dependent upon the on-going evidence regarding the suitability, effectiveness, integration, and safety of telehealth including for certain populations and indications. As well, the co-authors of this paper, through their leadership of research as well as leading teams providing routine clinical mental health care provided over telehealth have been told that mental health clinicians are concerned about their ability to ensure client and caregiver safety when care is being delivered over telehealth and practicing clinicians cite a lack of collated knowledge regarding the risks experienced by other clinicians in the past as well as how to manage risks when mental health care is being delivered over telehealth. Regarding the nature of risk with telehealth, risks may be due to the health condition, the mental health care itself, or risks may be due to the modality (ie unique due to telehealth) or combinations of these. Hence, adverse events should be categorised by their severity, degree of expectedness, contributing factors and potential relatedness to care, program, product, or device. There are unique opportunities and risks associated with the modality itself which will benefit from collating, understanding, and mitigating – ultimately to improve the safety of mental health care provided over telehealth.

In the studies included in this scoping review, risks and adverse events were most often reported anecdotally. Noting that many of the papers were research studies, or commentary papers and not reports of usual tele-mental health care. Few reported risks or adverse events using systematic approaches. For instance, very few papers reported the range of possibilities regarding what could go wrong, extreme cases, or discussed factors in terms of patient, staff, task, communication, equipment, control actions, organisational or environmental systems (Kaya, 2019). Few papers reported adverse events and if reported, few rated them according to severity or likelihood. Few papers were able to compare adverse event rates observed during telehealth compared to face-to-face care.

In summary, this review found multiple papers which aimed to report upon risks or adverse events of mental health services provided over telehealth but observed few to no risks. Attrition, or drop-outs, from mental health services provided using telehealth were deemed to be one of the most common risks. There is growing evidence regarding the risks of screening, assessment and treatments provided via telehealth for a range of mental health diagnoses, and for people experiencing self-harm and suicidality. As well, data are newly emerging regarding the safety and procedures for providing group and family therapies over telehealth.

Our paper also reviewed several papers which make specific recommendations for risk management when providing mental health services over telehealth. There are unique challenges of delivering tele-mental health care, as well as reporting and recording risks and adverse events during telehealth consultations. For instance, clinicians may be more likely to work independently and from home, when providing telehealth care as opposed to face-to-face care. Mental health care over telehealth will have unique risks such as: potential internet disruptions or poor-quality during crises, privacy concerns such as a video consultation being “hacked”, challenges physically attending to a telehealth client if in crisis or violent situation, for providers: regulation and supervision, health and safety, skills to fulfill scope of practice using telehealth and so on.

In conclusion, literature to date indicates telehealth is a safe modality for providing mental health care, even for high-risk clients. While managing risks via telehealth is typically viewed as more delicate than managing risk in person, the current view is that it is largely congruent with what would be done to manage risks in an in-person encounter. However, we would suggest that data are limited with respect to adverse events due to telehealth and this area needs further understanding.

This review builds from Luxton (2010). Luxton included papers published before 2009 whereas our review covers the literature published between 2010 and July 2021. As well in the Luxton (2010) review, most telehealth services provided were telephone services and almost all of the included papers enrolled only adults. Telehealth is a fast growing and changing field and thus this current review paper is crucial in providing contemporary knowledge regarding risks and adverse effects, and prevention of risks, for mental health care provided via telehealth for adults, and children. Through knowledge of past adverse events, we can devise improved guidelines and policy to ensure greater safety for mental health services over telehealth.

A limitation of this review is that publication bias may have played a part in the inability to find papers discussing adverse effects or risks of telehealth for mental health assessment and therapy. This systematic review of risks will ideally guide additional work in this area. Limited safety data currently constrains the development of telehealth training, standards, and policies unique to mental health care delivered using telehealth. A rich reporting culture must be created to capture accurate and detailed data about nuances of care in telepsychology. For instance, gathering data for near misses is well known to contribute fruitful data for quality improvement in health systems, along with reporting adverse events (Barach & Small, 2000). To gain improvements in the health system (in general, but also for telehealth) we must plan, implement, and constantly analyse standardized and systematic measures of adverse events and near-misses. Ideally this will include also collecting data on the apparent causes and underlying factors for the adverse event. This should serve to inform the improvement of telehealth making it rare for adverse events to occur while also minimising the harm of any new adverse events. Adverse event reporting can highlight issues occurring locally, or broader (nationally/internationally) and can assist in informing state or national policies around telehealth. As well, ideally measurement of adverse event rates over time can assist in evaluating whether improvements are occurring (Rafter, 2015).

Of course, in the reporting of adverse events one must balance between accountable reporting and a culture where no individual is blamed unless there are negligence issues (Rafter, 2015). Several examples of ways to improve adverse event and near-miss reporting include incident reporting, processes for complaints, audits, screening patient series or annual reviews or observations (Rafter, 2015; Walshe, 2000). These represent improvements over the more unsystematic method of voluntary reporting (Rafter, 2015). The severity of events can be recorded, as well as the expected standard of care. These require professional judgements and will likely continually mature in their definitions, particularly for mental health services provided using telehealth as it gains ground as a service delivery method. We should anticipate seeing a greater number of adverse events and near-misses being reported in telehealth as reporting of these adverse events improve. This should not be taken to mean that telehealth is becoming more dangerous but more the paradoxical situation caused by better records. In 2011 a positive step within the research world to record adverse events related to eHealth was proposed; this being the writing of the 2011 CONSORT statement modified for e-Health interventions. This CONSORT e-Health framework suggests that RCTs of eHealth interventions should include reporting of attrition, why the RCT was stopped early (if it was stopped early) as well as reporting of harms and unintended effects (Eysenbach, 2011). There are 10 papers published since 2011 which report planning to use or using the CONSORT e-Health framework, none of these are for mental health services delivered using telehealth, however.

Conclusions

Important findings of this review include the lack of standard operating procedures, training, and reporting guidelines for preventing and reporting adverse events associated with telehealth. Specifically, there is a need to improve assessment of suitability of clients for telehealth mental health services (McLean et al., 2021) and the safety of telehealth for those with mental illness (Daker-White & Rogers, 2013). This review indicates data are emerging and to date telehealth for mental health services shows a promising safety profile. Knowing more about the unique risks related to using telehealth for mental health services may assist educators, regulatory bodies, third-party payers, and insurers to devise training, approaches, rules, and procedures to improve the safety and quality of services (Di Carlo, 2021; Vanderpool, 2017). Having a greater sense of risks and risk management related to telehealth is likely to give technology developers greater confidence to create new products which better suit existing needs (Maheu, 2012).

Importantly, when expanding into new communities – these are often urban to rural – telehealth services for mental health needs should complement the existing system of in-person care in that community. The establishment of relationships with existing service providers not only ensures a more sustainable service, but also an improved ability to manage risks for clients and greater choice of settings if telehealth consultations are deemed inappropriate.

For private psychiatrists/psychologists or for mental health practice groups, internal policies and procedures should be reviewed to mitigate the unique risks associated with telehealth. These may include pre-screening client risk, emergency plans, tailored informed consent and safe and appropriate use of technology. Finally, staff training is essential to ensure technical and professional competence is maintained. Mental health practitioners may need to partner with external technical service providers to ensure a gold-standard service is delivered. Ensuring technical and professional competence is maintained has been important in COVID times as health professionals have been delivering telehealth services from home.

We must first identify risks, then analyse, evaluate and treat these risks to ultimately to ensure a safer health service environment. Those providing telehealth services might start by developing incident reporting systems that focus on near misses, and/or risks averted; as well, organisations could provide incentives for voluntary reporting of adverse events, while ensuring confidentiality. Future data collection on the unique risks and adverse effects experienced using telehealth for mental health services should use a systems approach for data collection, analysis, and consequent actions toward improvement. Ideally these future data will enable an examination of risks and adverse events during telehealth services compared to face-to-face services to distinguish unique risks and events due to the modality of care rather than due to the population and/or mental health service itself.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

Martiniuk was salary funded by Australian National Health and Medical Research Council (NHMRC) Fellowships during the writing of this work. Specifically an NHMRC TRIP Fellowship in 2017-2019 and an NHMRC Investigator Grant starting Jan 1, 2021 for 5 years.

Appendix A

  

Guidelines:

There are also guidelines from professional bodies regarding clinical care provided online. For instance, those by the Australian Psychological Society (APS) entitled “Telehealth measures to improve access to psychological services for rural and remote patients under the Better Access initiative” (APS, 2017 and APS, 2011a). Other relevant guidelines sourced during the writing of this paper include those released by the American Psychology Association (joint task force) (American Psychology Association, 2013), the Royal Australian and New Zealand College of Psychiatrists Professional Practice Standards and Guides for Telepsychiatry (The Royal Australian and New Zealand College of Psychiatrists, 2013) and the RANZCP Telehealth in Psychiatry COVID-19 Update Guidelines and Standards (2020), and the American Telemedicine Association (American Telemedical Association, 2017).

Methods detail:

Box 1: Search terms:

(("telemedicine"[MeSH Terms] OR "telemedicine"[All Fields] OR "telehealth"[All Fields]) OR ("telemedicine"[MeSH Terms] OR "telemedicine"[All Fields]) OR telecare[All Fields] OR telemental[All Fields]) AND (("psychology"[Subheading] OR "psychology"[All Fields] OR "psychology"[MeSH Terms]) OR ("psychiatry"[MeSH Terms] OR "psychiatry"[All Fields]) OR ("emotions"[MeSH Terms] OR "emotions"[All Fields] OR "emotional"[All Fields]) OR mental[All Fields]) AND (("risk"[MeSH Terms] OR "risk"[All Fields]) OR adverse[All Fields] OR “sentinel”[All Fields] OR “incident”[All Fields] OR “complaint”[All Fields] OR “avert” [All Fields] OR “near-miss” [All Fields]) AND “English”[lang]

We planned to assess risk of bias using The Cochrane Collaboration’s ‘Risk of bias’ tool and to assess the quality of outcomes using the GRADE criteria. The review protocol was not registered on PROSPERO because initial data extraction for several papers had occurred prior to finalising the study protocol. Data extraction for these papers was checked for completeness and consistency once the protocol was finalised, but this precluded registration on PROSPERO.