Lighthouse

Overprescription is a significant concern in the United States and across the globe. Studies indicate that approximately 20% of prescription medications are considered medically “unnecessary” by physicians themselves (Lyu et al., 2017). The dramatic increase in medication usage is notably prominent in certain subfields of medicine such as psychiatry. Particularly, psychopharmaceuticals are among the most commonly prescribed medications worldwide (Alduhishy, 2018). Today, an estimated 1 in 6 U.S. adults receive prescription psychiatric medications as part of their mental health treatment (CDC). The dramatic increase in the prescription of psychotropic medications is also evident in pediatric populations (Zito et al., 2021). In North America, Europe, and Africa, an estimated 25-42% of children take two or more such medications concurrently, a phenomenon referred to as psychotropic polypharmacy (Barnett et al., 2019). The practice of prescribing these drugs for off-label uses is also prevalent in other domains of medicine such as neurology (Vijay et al., 2018). 

Despite the widespread use of psychopharmaceuticals in clinical care, recent research underscores the importance of exercising greater caution in the initiation and continuation of these treatments (Damiano & Soares, 2022). For many patients, psychiatric medications not only lead to unbearable side effects, but also bring about intense and prolonged physical and psychological symptoms upon dosage reduction or discontinuation. In the case of antidepressants, for instance, the average withdrawal incidence is 56% with almost half of reports characterizing the experience as “severe” (Davies & Read, 2019). Another poignant example of iatrogenic harm from psychopharmaceuticals is Benzodiazepine-Induced Neurological Dysfunction (BIND), which encompasses “a constellation of functionally limiting neurologic symptoms (both physical and psychological) that are the consequence of neuroadaptation and/or neurotoxicity to benzodiazepine exposure” (Ritvo et al., 2023). This new research has shown that over 50% of individuals who are taking, tapering, or have discontinued benzodiazepines report experiencing enduring and life-altering symptoms, such as low energy, distractedness, memory loss, nervousness, and anxiety.

Despite the alarming statistics, the withdrawal experience has long been downplayed in medical practice and understudied in research. This trend is fueled by a combination of factors, including  systems-level issues such as pharmaceutical industry marketing practices, overmedicalization of the human experience, and inadequate medical training. These have had a tremendous impact on shaping the narrative around pharmacotherapy in psychiatric care. Multi-system regulatory failures, including short and gender-biased clinical trials, have only exacerbated the situation (Steinberg et al., 2021). For long, a reductionist biomedical model of mental health and existing treatments has overshadowed patients' testimonies and lived experiences. Unsurprisingly, clinical guidelines for reduction and discontinuation of psychiatric medications and recommendations for strategies to cope with the ensuing withdrawal symptoms are extremely underdeveloped as well. 

This dearth of guidance ultimately jeopardizes patient safety. Navigating a complex web of systemic issues in healthcare, patients frequently encounter skepticism and lack of support from providers. These individuals experience hopelessness and a profound sense of isolation, with their struggles left unaddressed and misunderstood. It is imperative to develop solutions that address overprescription and withdrawal, particularly to empower people undergoing pharmaceutical discontinuation.

We pursue the following objectives to deliver a 3-part solution that spans human-centered design research, technology development, and clinical engagement. 

Objective 1: Understanding Withdrawal

Characterize and fundamentally advance understanding around the patient experience of prescription medication withdrawal through both qualitative and big data approaches in order to develop computational models of symptoms, trajectories of suffering and healing, and discontinuation outcomes. We are currently collecting and analyzing patient-generated data from a multitude of platforms (e.g., social media, electronic health records) through state-of-the-art natural language processing techniques. 

Objective 2: Developing Personalized Informatics Interventions

Employ human-centered and participatory approaches to design and develop personal informatics tools and just-in-time adaptive interventions that leverage multimodal data and  personalized models as part of managing the withdrawal process in everyday life.

Objective 3: Creating Shared Decision-making Clinical Tools through AI

Apply machine learning and emerging AI techniques to develop predictive models for precision medicine and shared decision-making with clinical care providers. Our goal is to pilot this system with our existing network of physicians who are at the forefront of the de-prescribing movement in medicine. 

Our approach for understanding and improving the withdrawal experience is multifaceted and encompasses a synergistic approach where each objective draws upon and enriches the outcomes and insights of the others. With our AI engine and human-centered methodologies, we aim to create a comprehensive and integrated framework for addressing the complexities of psychopharmaceutical withdrawal.

Our platform serves three main populations: (1) adult patients, (2) caregivers, and (3) clinicians. 

(1) Adult Patients

Literature is scant regarding the withdrawal experience and psychiatric medication-induced neurological injury, resulting in a lack of clinical guidelines. Patients' accounts of enduring and intense symptoms are therefore frequently met with skepticism within the medical community. When the effects of drug exposure outlast the drug itself, patients grapple with the long-lasting consequences of psychopharmacotherapy without any support from the healthcare system.  In the case of antidepressants alone, at least 20 million US adults experience withdrawal symptoms after stopping psychiatric medications. Our solution supports such patients in mitigating the physiological and psychological challenges involved in discontinuation of psychiatric medications. In particular, we provide personalized, evidence-based guidelines to patients during the process of discontinuation from certain medications. Our solution also incorporates predictive models for empowering those who are navigating the withdrawal process. These models will help individuals foresee and respond to symptoms in a more informed and proactive manner, ultimately enhancing their ability to manage their condition and reduce the associated risks. 

(2) Caregivers

Caregivers, who number in the tens of millions in the United States, play a crucial role in the lives of individuals going through the withdrawal experience from psychiatric medications. However, they often find themselves in a perplexing and challenging position. Their loved ones may express experiencing intense and protracted symptoms after discontinuation, yet most healthcare professionals often dismiss or attribute these symptoms to a relapse. This leaves caregivers confused and uncertain about the best course of action as they witness the suffering of their loved one. Caregivers often lack the education needed to fully comprehend the complexities of the withdrawal process, and this knowledge gap can lead them to place trust in the recommendations of healthcare providers, even if it involves reinstating the original medication, increasing the dosage, or adding new medications to the existing cocktail of psychopharmacotherapies. To best support their loved ones, we empower caregivers with access to education and resources that enable them to advocate effectively for the well-being of those they care for.

(3) Clinicians

Our platform not only equips clinicians with the AI-driven capability to simulate withdrawal trajectories based on a patient's medical history but also fosters a culture of shared decision-making. It empowers clinicians to engage patients in meaningful discussions about their treatment options and discontinuation strategies. By providing comprehensive data and insights, our solution enables clinicians to collaborate with patients in making informed choices that align with the patient's values, preferences, and unique circumstances. This approach not only enhances the therapeutic relationship but also ensures that patients have an active role in shaping their treatment plans. Moreover, recognizing that the majority of clinicians lack training resources and opportunities to stay abreast of emerging findings on medication discontinuation, our platform offers a wealth of educational materials co-designed with physicians. These resources provide critical guidance in practicing safety mitigation strategies for medication discontinuation. By keeping clinicians well-informed and up-to-date with the latest research and best practices, our solution ensures that they are well-prepared to deliver the highest standard of care and promote patient safety throughout the psychiatric medication withdrawal process.

The impetus for founding this organization stems from our team lead's lived experiences undergoing the withdrawal journey as a patient and navigating the psychiatric system as a caregiver. These personal encounters, while primarily lived by our team lead, have profoundly shaped our collective understanding. Witnessing the trying predicament of protracted withdrawal from psychiatric medications has ultimately deepened our empathy for the communities we serve. While our team’s extensive academic and professional background in health technologies is tremendously helpful in scaffolding our approach to the problem space, our personal proximity to the work is actually our greatest asset. 

Further, we recognize the importance of equity-centered, participatory approaches to technology development, particularly in sensitive health contexts. We therefore actively engage with those we serve, fostering collaborative partnerships to ensure that our solutions are not only informed by personal experiences but also shaped by the diverse perspectives within the populations we aim to support. Specifically, we are leveraging our network of clinical collaborators to connect with patients who are tapering off psychiatric medications as well as their caregivers and physicians. For example, the insights and experiences of these individuals directly inform our design of our patient-facing, caregiver-facing, and physician-facing resources and tools. Employing an iterative human-centered design methodology, we integrate these participatory approaches throughout stages of solution development, including needfinding, prototyping, and testing. 

Finally, our approach extends beyond individual interactions with our target populations. In particular, we have been intentional about forging partnerships with local and international peer-support community leaders and members. Our in-depth research in the psychiatric medication withdrawal space has highlighted the pivotal role played by grassroots movements, nonprofits, and patient advocacy groups. Recognizing their pioneering efforts, our organization is deeply involved in such initiatives in order to better understand the community’s unique needs and concerns.

Our solution innovates by addressing critical problems in the psychopharmaceutical space that have significantly been overlooked in academia, medical practice, and industry for decades. While case reports of the psychiatric medication withdrawal experience have organically emerged within patient communities, there has been little systematic effort to comprehensively understand and mitigate iatrogenic harm from psychopharmacology. Our partnerships with patients living with withdrawal, their caregivers, and clinicians affirm that attention is urgently needed in this space. 

Particularly innovative is our data-driven methodology, a significant departure from the current deprescription landscape. This convergent approach that involves integrating empirical data, AI-driven modeling, and human-centered insights has the potential to catalyze broader positive impacts in several ways. Firstly, by spotlighting the magnitude of the issues associated with psychiatric medication withdrawal and the knowledge gaps that exist (e.g., full understanding around the range of effects experienced during withdrawal, in both the short and long term, and the ways that such burdens are disproportionately experienced by groups who commonly experience health disparities), our work can inspire much-needed rigorous scientific research, including randomized controlled trials (RCTs) to examine these complex pharmacological effects in diverse populations. In turn, these insights will help to bridge the existing gaps in academic literature, medical practice, and industry strategies. Our novel data framework and AI engine can act as a catalyst for increased collaboration and investment in this space, revolutionizing how the broader healthcare sector perceives and addresses the challenges associated with patient safety, informed consent, and psychiatric medication withdrawal.

In essence, our solution not only empowers patients who have long been dismissed and underserved, but also could change the market dynamics by providing a more comprehensive understanding of psychiatric medication withdrawal. With our pioneering data-driven approach, we not only tackle existing problems but also lay the foundation for more systematic and evidence-based work in the future.

Our work tightly aligns with the United Nations Sustainable Development Goal 3, which seeks to ensure good health and well-being for all. In particular, we connect strongly with Target 3.4 in our mutual goal to reduce premature mortality from non-communicable diseases through novel technologies for prevention, treatment, and general promotion of mental health and well-being. 

By (a) gaining a deeper understanding of the factors contributing to protracted psychiatric medication withdrawal and (b) developing data-driven withdrawal models, personalized coping tools, and decision-making interfaces, we aim to intervene proactively to reduce potential medication risks, enhance treatment strategies, and overall promote mental well-being. In this way, our innovative, data-driven approach serves as a powerful tool for prevention and is well-positioned to contribute significantly to reducing the suicide mortality rate associated with psychotropic medications. 

We underscore that understanding the nuanced factors that contribute to suicide risk during and after medication exposure is paramount. Recent studies show that medication-induced suicide has long been systematically underestimated due to publication bias and financial conflict of interest. For instance, exposure to Selective Serotonin Reuptake Inhibitors–the most commonly prescribed class of antidepressants–is associated with higher suicide risk in adults (Hengartner et al., 2021). Case reports of suicidal behavior during the withdrawal period has also been reported in the literature (Waldvoge & Akira, 2022).

 Our research methodology includes analyzing real-world data from diverse sources to comprehensively capture the safety profiles of medications and their associated risks. Through this proactive approach, we not only seek to identify potential risk factors but also to develop personalized informatics tools that can be instrumental in preventing tragic events such as suicide. These tools will not only enhance the monitoring of individuals during the withdrawal process but also provide timely interventions and support based on individualized risk assessments.

Our solution harnesses the power of AI to collect and analyze diverse datasets. We employ a multi-faceted approach, combining existing public datasets with novel data we harvest through direct patient engagement. 

To elaborate, one key aspect of our data strategy involves leveraging existing public datasets (e.g., social media, public patient forums). These platforms serve as rich repositories of patient-generated insights, enabling us to capture a wide range of perspectives related to medication withdrawal. Through advanced natural language processing (NLP) techniques, we extract valuable information from these sources which then drive our predictive models for interactive patient-facing and clinical tools.

The second distinctive aspect of our data strategy involves direct patient engagement to capture novel and firsthand information. We employ multimodal data collection methods, incorporating passive measures alongside self-reported information about behavior, physiology, and experience. This direct engagement with patients promotes a more holistic and personalized understanding of the withdrawal process.

In summary, our strategy combines the ethical use of existing public datasets together with direct patient engagement to curate a comprehensive and diverse dataset. By leveraging NLP and AI techniques, we then extract and provide meaningful insights that contribute to a deeper understanding of psychiatric medication withdrawal and facilitate the development of innovative user-centered solutions to support individuals navigating this challenging journey.

In our commitment to ethical and responsible AI use, we have implemented a comprehensive approach to address potential risks associated with our solution. Our focus encompasses privacy protection, security measures, and user empowerment. 

Privacy Protection

1. We prioritize the privacy of individuals involved in our datasets. At every stage of the project, our data is either already anonymous (e.g., anonymized public datasets), or we de-identify it during the collection process.

2. Informed consent is a cornerstone of our approach. We ensure that individuals providing data through direct engagement are fully informed about the purpose, usage, and potential impacts of their data.

Security Measures

1. Our data storage and processing systems incorporate robust security measures. We utilize encryption protocols and access controls to safeguard sensitive information. 

2. We conduct regular security audits and assessments  to identify and address potential vulnerabilities in our systems.

Use Empowerment 

1. Our solution is designed to empower users with control over their data. Individuals can choose their level of disclosure and the extent to which their information is captured and utilized.     

2. We recognize the delicate and sensitive nature of topics related to lived experiences with psychiatric medication withdrawal. This sensitivity is integral to our ethical approach. In working with our patient and caregiver partners, we prioritize trauma-informed design and research methodologies. Our principles are rooted in guidance from the CDC and SAMHSA, specifically emphasizing the following six focus areas: (1) Safety, (2) Trustworthiness & Transparency, (3) Peer Support, (4) Collaboration & Mutuality, (5) Empowerment, Voice, and Choice, (6) Cultural, Historical & Gender Issues.  

One Year

Impact Goal 1: Understanding Withdrawal

• Patient recruitment and interviews:

  • Develop a comprehensive recruitment strategy for diverse representation

  • Collaborate with healthcare providers and support groups for access

  • Conduct interviews to capture nuanced lived experiences

• Data analysis and publication:

  • Utilize advanced analytics to extract meaningful patterns from existing public patient-generated data

  • Collaborate with data scientists and statisticians for robust analysis

  • Submit research findings to top-tier academic journals to advance the scientific understanding of withdrawal

Impact Goal 2: Develop Resources and Personalized Informatics Systems for Patients, Caregivers, and Clinicians

• Prototype development:

  • Build a medium fidelity prototype of the patient-facing informatics tool

  • Develop prototypes of caregiver-centered resources as well as decision-making tools for physicians

  • Conduct iterative usability testing with patient, caregiver, and clinician feedback

• High-fidelity integration:

  • Draw insights from medium fidelity prototype testing

  • Integrate findings into the development of a high fidelity prototype

Impact Goal 3: Expanding Our Network

• Clinical collaborators:

  • Identify additional clinical partners through conferences and networking events 

  • Establish collaborations with reputable healthcare institutions

• Grassroot organizations and advocacy groups:

  • Strengthen ties with grassroots organizations through active participation

  • Engage in dialogue with patient advocacy groups to address specific needs 

  • Leverage partnerships to increase awareness and adoption of our solution

Five Years

Impact Goal 1: Advance Scientific Understanding

• Research publications:

  • Conduct ongoing studies to contribute to the body of knowledge on withdrawal 

  • Publish additional insights and findings in reputable academic journals

Impact Goal 2: Scaling Impact 

• User base increase:

  • Implement marketing and outreach campaigns to increase clinician and patient enrollment

  • Monitor user feedback for continuous improvement

• US and international expansion:

  • Pilot our solutions in testbed locations where uptake is more tractable given receptivity and existing efforts around deprescription (e.g., the United Kingdom, where we also have a strong collaboration established with a psychiatry partner)

  • Develop a comprehensive scaling strategy for the US market

  • Investigate regulatory requirements for international expansion

  • Establish partnerships with global healthcare organizations for widespread adoption

Impact Goal 3: Large Scale Research and Policy Advocacy 

• Fundraising and randomized controlled trials (RCTs):

  • Seek funding opportunities for RCTs to examine pharmacological effects in diverse populations including withdrawal

  • Collaborate with renowned researchers for rigorous scientific validation

• Root cause mitigation strategy development:

  • Establish partnerships with medical schools to integrate withdrawal-specific curricula

  • Use accumulated data and insights to craft evidence-based policy proposals 

  • Engage with policymakers at the federal level to present recommendations

Our full-time core team at Lighthouse comprises one full-stack developer, one data scientist, one project manager, and one psychiatrist as well as team leads Shirin Amouei and Elizabeth Murnane, who are responsible for setting and driving the organization's strategy and supervising all technical, scientific, and outreach work. Our collaboration with the Empower Lab, directed by Murnane, provides access to additional part-time staff who contribute expertise in AI and natural language processing, user interface design, information science, neuroscience, and behavioral science.

We have been working on our solution for approximately 10 months.

Promoting diversity, equity, and inclusivity is a foundational aspect of our organizational ethos. Our leadership team is committed to fostering an environment that values and celebrates diverse perspectives. We recognize the critical importance of reflecting the rich tapestry of the communities we aim to serve.

Leadership Team Diversity

Our leadership team is proudly composed entirely of female leaders, bringing a wealth of diversity that represents various educational, cultural, and economic backgrounds.

Recruitment Strategies 

We are implementing inclusive recruitment strategies to attract a diverse talent pool. This includes actively seeking candidates from underrepresented backgrounds. In line with our organizational mission, we are also particularly interested in hiring individuals that have experienced, or been personally impacted by, iatrogenic harm. For instance, our goal is to diversify our team by recruiting peer-support community moderators who perform those functions as unpaid volunteers. We aim to not only honor their heartfelt commitment and sincere dedication to the withdrawal community, but also to integrate their valuable insights and experiences in the work our organization pursues. 

Our operational model centers around a collaborative approach that bridges the strengths of an academic research lab and a small business venture, both dedicated to the mission of patient empowerment. 

Empower Research Lab

Situated at Dartmouth College, the Empower Lab serves as the hub for our methodological and technological development. The lab conducts research in human-centered design, digital therapeutics, and personal informatics. A key focus is on the inclusive development of intervention technologies that approach individual-, community-, and systems-level change through an empowerment approach.  In building, evaluating, and scaling our solution, we leverage the expertise of this interdisciplinary collective of technologists, designers, and behavioral scientists.   

Lighthouse

Our venture arm plays a pivotal role in translating academic research into real-world impact. Most importantly, we aspire to remedy the “testimonial injustice” in healthcare through which patients’ experiential accounts are routinely discredited and dismissed. 

Clinical Collaborators 

Our network of healthcare providers spans both local and international collaborators. Leveraging our robust connections with practicing physicians leading the deprescribing initiatives for psychiatric medications, we find ourselves strategically positioned to advance solutions requiring patient and clinician engagement. As an initial step, we aim to pilot our work in the United Kingdom. The rationale behind this approach is threefold: 

1. the UK's sociopolitical climate has led to a more receptive environment for the deprescribing movement compared to the US; 
2. the country's national health system has facilitated broader regulatory initiatives and increased understanding related to iatrogenic harm from psychiatric medications within the medical community; and 
3. significant progress has been made by research and nonprofit communities in exploring this space.

After gaining insights and refining technical infrastructure through our UK testbed, we will move into US deployments of our data framework and user interfaces, with an eye to identifying generalizable versus context-specific adaptations of our solution.

Patient Communities and Advocacy Groups

Community engagement is the cornerstone of our operational plan. We actively collaborate with patient advocacy groups and peer-support organizations to draw upon their rich expertise and insights. 

In summary, our operational model is multifaceted and collaborative, blending academic rigor with real-world impact. This approach not only substantiates the feasibility of our solution but also ensures its relevance and impact in addressing the multifaceted challenges associated with medication withdrawal.

Our financial sustainability plan has a two-pronged approach: (1) integrating grants and community-driven support for initial phases, and (2) actively seeking partnerships and exploring additional financing avenues for long-term scalability.

Grants

To kickstart our initiatives, we are strategically pursuing small business innovation grants from US federal agencies such as the NIH and the NSF. These grants serve as crucial seed funding, enabling us to secure the necessary resources for research, development, and early implementation phases.

Community Donations and Fundraising

Recognizing the vital role that community engagement plays in our mission, we are establishing avenues for donations and community-driven fundraising. This approach not only fosters a sense of shared ownership but also ensures that our initiatives resonate with the very individuals we aim to empower.

Financial Sustainability and Scaling 

We acknowledge the need for strategic partnerships and innovative funding models for long-term financial sustainability and scalability. Our collaboration with MIT Solve and Cure is a key element of our financial strategy development. We are excited to partner with both organizations and leverage their expertise and network to explore future financing schemes. 

Our current operating costs stand at $350,000, with the bulk primarily allocated to human capital expenses, including staff salaries and trainee tuition, amounting to $250,000. This investment in our team ensures that we can attract and retain top talent, which is crucial for the success of our initiatives. Additionally, $50,000 is allocated to capital resources, covering computational needs and other requirements essential for our research and development efforts. A further $50,000 is designated for patient engagement costs, reflecting our commitment to fostering meaningful interactions and collaborations with the communities we serve. Looking ahead to the next year, we project a similar budget structure.

We are seeking funding in the amount of $60,000 to support our work in 2024. This budget is carefully calculated to address key areas critical to the success of our initiatives. 

$35,000 is allocated to staffing and human capital. This covers a portion of staff salaries and trainees’ tuition. 

$15,000 will be dedicated to participant compensation and engagement resources. We recognize the importance of valuing and facilitating the involvement of individuals who are most likely in a vulnerable state. This portion of the fund will cover resources such as interview spaces, transportation costs, and any other unique needs the participants might have.

$10,000 is allocated for computational resources and supplies essential for intervention development. 

Overall, this $60k funding request is a balanced and strategic approach to sustain and enhance our ongoing efforts, considering the diverse needs of our team, participants, and the technological infrastructure necessary for our initiatives.

Seed Funding

• The funding support will enable us to partially cover our operational costs, affording us the crucial time needed to refine our solution and develop a sustainable financial growth model.

Mentorship Support

• Navigating uncharted territories in an area where there hasn't been extensive prior work necessitates expert guidance and insights. The mentorship component of the Cure Residency is invaluable to us at this stage. Having access to experienced mentors who understand the challenges and nuances of our field will accelerate our learning curve, refine our approach, and help us avoid potential pitfalls.

Networking Opportunities

• The network offers insights into potential funding opportunities, strategic planning, and collaborations. This expanded network is not only a resource for immediate needs but also a foundation for long-term sustainability and growth, including to help us scale to traditionally hard-to-reach and underserved populations.

Lab Space

• Being located in rural New Hampshire, accessing specialized expertise and cutting-edge resources can be a challenge. The lab space in NYC provides us with proximity to a wealth of experts, research institutions, and collaborative opportunities. This physical presence in a metropolitan area facilitates easy engagement with a diverse range of professionals, fostering collaborative innovation and the exchange of ideas.