# Medical Innovation: Protocol

## 8.11.23

### List of product ideas

* Use LLMs to translate medical diagnostics into something normal patients understand
* Translate patient comments into a possible diagnosis
* AI-powered preliminary diagnosis of mental health issues ("assistant")

### Idea: Diagnosis assistant (Timm)

* What is exactly my product/service? What does it deliver to the customer rather than what technology is it based on?
* Take patient response and translate into a diagnosis (Virtual assistant)
* Additionally, determine any extra information that needs to be asked to determine a better diagnosis (helps the therapist, not entirely replace it)
* What is the core value provided by my product / service / …?
* Ease of access and improved diagnosis by helping the person who does the diagnosis (possibly fully replace?)
* Essentially, faciliate medical diagnosis (from point of view of therapist/doctor/medical practitioner)
* Who are my customers?
* Therapist/doctor, not the patient! They are the ones who have to adopt this service/product, so we should cater to them. From POV of patient, virtual assistant or real therapist does not matter so long as diagnosis is improved.
* But keeping patient happy through ease of use is probably a valuable consideration
* What are the needs/desires of my customers?
* Easier and faster diagnosis
* Fact access to large knowledge base
* Not intrusive, does not impede work of therapist (1% lows should be acceptable, i.e. wrong diagnosis)
* 2nd opinion that can be instantly accessed
* What do my competitors do?
* [Assistant that helps therapists](https://moveforwardvirtualassistants.com/): not really what my idea is about, they do paperwork for therapists but a human seems to be doing it
* [Symptomate](https://symptomate.com/): this is directed towards the patient, not assisting the therapist
* [17 online diagnostics](https://techcult.com/best-ai-medical-diagnosis-apps/): this is a very good list of competitors, should use later
* What is my differentiation?
* It should be easy to use and also 'better', this is a question we would answer ocne we have a solid idea
* What relevant and differentiated value do I offer?
* TBD
* What are customer benefits from my provided value?
* Faster and better diagnosis
* Patients are happier because the diagnosis is better -> improved business?
* Saving costs, one therapist can do the work of many (maybe?)
*
* What are my capabilities to deliver this value consistently/repeatedly?
* This is more of a technical question, should focus once we have a better idea

### Idea: Diagnosis assistant (Ruochen)

**What is exactly my product/service? What does it deliver to the customer rather than what technology is it based on?**
* A convenient chatbot that can interact with patients at any time and has a certain psychological background. Gather information and filter valid info to help doctors to diagnose or monitor patients' healthy state.
* Can allow patients to have an electronic psychologist anytime, anywhere, to relieve anxiety and psychological emergencies. At the same time, the model can collect the dialogue information of patients in real time and put forward effective information to achieve real-time evaluation, which can effectively prevent emergencies, or automatically contact the doctor or alarm when an emergency occurs.

**What is the core value provided by my product / service / …?**
* Real-time conversation, expert chatbot, anytime, anywhere.
* Real-time assessment. If necessary, trigger the alarm， to get the help
* provide doctors with a large amount of effective data and auxiliary diagnosis.

**Who are my customers?**
* People with mental problems, lonely people who want to talk to, therapists
* What are the needs/desires of my customers?
* People with mental problems：Psychologist, necessary/efficient treatment, assessment of state (intervention)
* lonely people who want to talk：Comfortable chat object, there is a certain degree of professionalism, can relax emotions
* therapists：Collect effective information to reduce workload and better understand patients

**What do my competitors do?**
* https://woebothealth.com/At the use of cognitive behavioural therapy (CBT) techniques that enable users to track themselves by gradually guiding them to correct bad automatic thoughts and by regularly sending them smart mood charts. Wobot also comes with 100 evidence-based stories from Dr. Darcy's clinical team, guided by evidence-based stories that make it easy for users to learn the skills to reduce stress and live a happier life, thus enjoying a healthy lifestyle.

**What is my differentiation?**
* More professional: The trained data are actual medical cases and medical papers
* Assess the user's mental health in real time and intervene if necessary (contact doctors, police, family)
* It provides doctors with more information related to patients, which helps doctors to better evaluate patients' psychological conditions
* Convenient, anytime, anywhere, personalized customization

**What relevant and differentiated value do I offer?**
* Different inputs (audio, video, text) to meet different needs

**What are customer benefits from my provided value?**
* patient：Low price, anytime, professional, personalized, secure
* doctor: Objective, low cost to collect effective information, at any time to monitor the psychological state of patients, reduce the workload.

**What are my capabilities to deliver this value consistently/repeatedly?**
* Collected patient information + doctor's diagnosis =medical cases. This can be the training data
* with more data, the mdoel can perform better

### Andrei idea

#### Use LLMs to translate medical diagnostics into something regular patients understand

**What is precisely my product/service? What does it deliver to the customer rather than what technology is it based on?**

Empathetic and easy to consult medical diagnostics. Diagnostics can be intimidating and lengthy, making them challenging to consult over time. A patient might want to check back on a diagnostic for things such as diet restriction or undesired interactions with food, other medicine, etc.

**What is the core value provided by my product/service?**

Empathy: medical situations/check-ups are stressful, and most people would be unfamiliar with the terms, how medicine works, etc. While doctors provide a diagnosis once, this could be forgotten, and it’s challenging to check.

**Who are my customers?**
B2B2C: We charge health providers for the utility/ comfort they offer to their patients.
What are the needs/desires of my customers?
Easier access to medical services for patients needs more time for busy doctors.

**What do my competitors do?**
* Actual medical doctors?
* Brazilian startup NeuralMed uses proprietary AI algorithms for real-time medical text analysis. Harpy, the startup’s AI application programming interface (API), uses natural language processing (NLP) to extract information from radiology reports to create structured medical records. Harpy is deployable in hospitals’ existing systems for analysing discrepancies in medical reports and creating alerts during critical findings. It also creates a database of previous patient examinations. Hospitals use NeuralMed’s solution to better allocate resources and improve the accuracy and structure of diagnosis data. [SOURCE]

***[NEED EXPLANATION] I have a big issue/ misunderstanding with this idea: is it legal for a doctor to automate the diagnosis? Likely not, and in that case, what value are we exactly providing? Are we providing a summary of the disease/ diagnosis, with indications on how to treat or live with it?***

**What is my differentiation?**
// Not sure how to answer this. See note above

**What relevant and differentiated value do I offer?**
// Same as previous questions

**What are customer benefits from my provided value?**
Easy access to an easy-to-understand diagnosis

**What are my capabilities to deliver this value consistently/repeatedly?**

The business could charge for each diagnostic.

#### Translate patient comments into a possible diagnosis.

**What is precisely my product/service? What does it deliver to the customer rather than what technology is it based on?**
An assistant interface that can turn a conversation into a preliminary diagnosis. Patients get easy diagnoses, while health practitioners are offered more time to do their job and skip boring parts.

**Who are my customers?**
B2B2C: We charge health providers for the utility/ comfort they offer to their patients.
What are the needs/desires of my customers?
Be healthy, avoid appointment times, and get diagnosed on their schedule. Doctors want to reduce menial tasks and have as much time as possible.

**What do my competitors do?**
* Talkie.ai: Calls patients using AI-based virtual assistants [SOURCE](https://blog.google/outreach-initiatives/entrepreneurs/30-ai-startups-changing-the-future-of-healthcare/) They are going in the opposite direction of calling patients. This validates the market need.
* BetterHelp: Online mental health sessions [SOURCE](https://www.betterhelp.com/get-started/)

**What is my differentiation?**

**[IDEA]: Some startups create transcribes and meeting summaries so people can return to them. What if we combined this with the assistant-based diagnostic to get one source of truth for patients? All conversations with in-person doctors or AI assistants add up to a “patient file”. When I was doing theraphy, I was keeping a shared journal with my theraphist for each day. Every month or so, I would get new or different questions that I would have to fill in. I have the feeling that practicioners of mental healthy or others in the medicine area could use better digitalisation overall.``**

**What relevant and differentiated value do I offer? What are customer benefits from my provided value?**
* Ease of access to medicine + digital system for patient management?
* What are my capabilities to deliver this value consistently/repeatedly?
* The solution would be available throughout the patient’s case and ideally for as many patients as possible.

#### AI-powered preliminary diagnosis of mental health issues (“assistant”)
It is the same as the last idea but smaller and specialised for mental health issues. Mental health issues or personality are diagnosed using standard quizzes with multiple answers, which would be nice to integrate into the diagnosis part. They are time-consuming, and a lot of therapists compute the scores manually.

### General notes (what we observed during conversation)

* What does the patient want?
* Therapist/service should be friendly and empathetic, this is important in user satisfaction
* Transcribing patient conversations seems to be important, where we go from there is another point of discussion
* Transcription can be used by both the doctor and the service/assistant
* Ruochen alternative idea: electronic diary or coach, not a full diagnosis?
* [Dei link: imaginary friend](https://replika.com)
* Common idea among all 3 ideas: doctors are not enough to give a full diagnosis, need more
* Timm: knowledge base available through assistant
* Ruochen: Service is always with the patient, continuous monitoring/ data collection
* Dei: Everything about patient accessible in one place, something needs to be able to read, process and interpret all of this information. Can help find correlations and possible relationships. *Digital Patient Record*
* Use friendly AI assistant that transcibes and summarizes information collected during a session, this is added to the record
* A core aspect of our ideas is to be able to comprehensively collect information from the patient (digitalizing the data collection and retention).
* Because there is so much information, there needs to be some kind of summarizing at the output stage
* Primary customer: medical practitioner, not the patient (because patient does not care about records, only diagnosis).
* Other customers could be the healthcare system, or really anyone interested in maintaining the patient record
* For the Friday feedback session, better to keep it general than come up with a formalized idea (based on our 3+ ideas)
* We need to collect feedback from more sources before coming up with a formal idea
* People to interview in order:
* Doctors, nurses, health practicioners
* People doing research in the medical area
* LMU students
* Healthcare entreprenerus
* TUM students radiology/ health sciences?

## 10.11.23

* Revenue sources: insurance can pay for stuff too
* Need to deal with the digital information acts / privacy protection

## 14.11.23

* Transcript ideas are best
* Voice to text
* Need to keep transcript somewhere (possibly patient file)
* Use transcript to make a diagnosis
* Alternative: digital diary that the patient keeps with them at all times, summarized so that the therapist can use it during therapy
* Need something to differentiate us (killer feature):
* Remove manual work (bookkeeping) for the therapist
* Would like feedback from external sources
* Start small
* Stay away from enterprise level
* Phone / Laptop app? Easy to integrate, fix any problems (basically only a software product)
* Healthcare moves slow, features should be deliberate and not super innovative. Improve what already exists
* Dei Therapy idea:
* Discussions are more free flowing, so AI has a better usage here because it allows being able to extract key points that the practitioner misses or forgets
* Customer: insurance is probably the customer because they would pay for it, patient and therapist are stakeholders
* We market to therapists, but the insurance pays
* Want therapists to recommend to the insurance that our product is useful
* But maybe we show it insurance, and they suggest it to the therapists?
* Need to also insure that the therapist (customer) is happy with the product (i.e. this will probably go into our pitch, convince people why they need this product):
* Why do therapists need our product? -> Market research
* Reduce bookkeeping through automatic transcription and summarizing
* Does not impede work of the therapist, should be additional / assisting
* Product features that can be interesting:
* Automatic transcribing of therapy session
* Questionnaire generation
* Creating notes/summary from the transcript
* Access to a large knowledge base
* Ease of setting up (phone / laptop is basically all that is needed)
* Low barrier of entry
* Market research:
* Start planning it, but does not have to be ready for the presentation pitch
* Send emails to therapists, need to draft letters

* Next meeting: Friday 17th 17:00

## 17.11.23

### Timm notes

* [Bing Image Creator (DALLE 3)](https://www.bing.com/create?toWww=1&redig=FF54E3DB455E4F6296915E92F74323C6)
* [Kapwing](https://www.kapwing.com/): AI video generator, needs signup, we need to provide input
* [Invideo](https://invideo.io/make/ai-video-generator/): Similar to Kapwing, also needs signup. 10 minutes of AI generation, whatever that means
* [Canva](https://www.canva.com/): Can create talking heads, maybe we want this?
* Imo, think it is better to just have a PowerPoint presentation with nice AI generated images, and give a convincing speech. We don't have a lot of time to mess around with AI tools that may not work well.
* We could also add small pieces of AI generated material to our presentation, like talking faces.
* Just need to figure out good prompts that are consistent with DALLE 3
* We could, if we wanted to, also make the PowerPoint into a video.
* We might want to have a bit of music

### Script Draft

* [Initial Script for video, by Dei](https://pastebin.com/AjGv7XC0)

### Discussion notes

* We need voice generation to narrate the script that Dei has provided
* Ruochen will look at a good text to voice generator that can narrate our presentation
* [Preliminary Final Script](https://docs.google.com/document/d/1QOz9WlHYnHr8OiSnpgUoY2YvaabhpCu48MCTCUbDcmY/edit?pli=1)
* Next meeting on Sunday

## 18.11.23

* This is just for Timm to write notes to figure out how the generators work
* Hotshot:
* Does not seem to be worth our time, only generates very short GIFs that have a lot of poor facial expressions and wonky shapes
* Lensgo:
* Stick to short prompts, it craps out when you have more than one sentence
* Only use anime style, realistic is garbage
* Does not seem to like 16:9 dimensions, stick to the more square ones
* Don't use camera movement, very hit or miss from what I've seen

## 24.11.23 Feedback on presentation

* Need to be careful with how we handle sensitive data
* GDPR (but not a good place to start)
* Maybe a medical expert article?
* Read up on the specific German law
* [Digital Patient Act](https://www.bundesgesundheitsministerium.de/elektronische-patientenakte), will come into full effect in the next few years
* Might be a good idea to just keep our model /software locally on the network of the therapist, instead of dealing directly with the DPA
* Our idea may have a large technical and legislative hurdle, more than other teams, because the data is sensitive
* Timm: we should narrow down what we do exactly, be a bit more precise.
* Need market research/feedback before we can fine tune our idea

## 02.12.23

* Other stuff:
* Ask therapist if he knows how to contact many of them for interviews
* Decide how to contact customers in general
* Allocate workloads for this week

* Customer: Therapist
* Gains:
* Better F2F sessions by focusing on the patient instead of multitasking
* More time of the workday allocated to consulting people
* An organized way to store patient information automatically (audio + transcription that are synced, summaries, searcheable notes)
* Automatic questionnaire administration for the patients
* Pains:
* Therapist is forced to multitask by taking notes and cannot listen to the patient fully
* Relistening to a recording is tedious: the therapists that record their sessions have to go through a lot of material and manually seek certain bits
* Not enough time to consult patients, therapists tend to be overbooked
* Storing information is a task in itself: written notes can be lost and moving written notes to digital is laborious
* Therapists may not be tech savvy enough to create a helpful workflow, e.g. by moving notes online, using 3rd party transcription services for their audio recordings
* Admintesring questionnaire by hand is tedious: scores have to be computed by hand, the therapists sends the test via mail and receives a different or edited file with the patient's answers
* Customer Jobs:
* Have F2F sessions listening to patients
* do the medical record (notes&recording) during F2F sessions
* assess their clients' mental wellbeing through questionaires
* reorganize notes & transcribe recordings at the end of the day, after patient session
* check medical reference and past cases for a current patient
* ensure patient information is safely stored if held on digital devices
* handling financial paperwork/ obligations e.g. patient insurance, forms etc.
* Checking in on patients' journals or take-home assignments
* Gain Creators:
* Digital notes can be searched intelligently, e.g. text or tags attached by therapist (idea: identify tags from the AI transcription)
* Use AI to summarise and extract core ideas from a transcribed session
* A single patient file with all recording, notes and questionnaire answers in one place; automating the centralisation process
* Shared digital diary between the patient and the therapist
* Pain Relievers:
* Use AI to transcribe audio recordings, allowing therapists to focus on the F2F session
* Digitalise the questionnaires so scores can be attached to
* Products & Services:
* Multi platform/ web application with transcription as a core feature
* AI summarizer for the patient file/ case
* Questionnaire tracking/ administration system, i.e. a digital form

* Tasks for Sunday meeting:
* Timm: write template email we use to contact therapists
* Dei:
* Ruochen:

* Tasks for the next seminar:
* Contact people
* Each can do this for the native country so we can parralize work
* Ask on reddit or forums
* Write cold emails to people online
* How to interview people
* Don't take too much of their time & emphasize that you will be brief
* Questions brainstorming:
1. What bothers you the most in your work as a therapist?
2. What is the most time-consuming activity?
3. Would you pay for a service that trascribes your patient session automatically?
4. What is a tedious task during therapy that you wish were automated?
5.
* Final interviewing scenario [TODO, PICK 2]:
* Ruochen suggests 3 questions, because we want to get a little bit more information.

## 03.12.23

Contact: [e-mail]

Greetings!
We are students from the Technical University of Munich (TUM) studying entrepreneurship in healthcare. We want to build better tools to help therapists care for their patients, be it during or in-between sessions. Only the first 3 questions are mandatory, so feel free to leave the others blank if you only have 5 minutes. Short answers, of about 2-3 sentences, are perfect! Thank you for your precious time!
Question 1: What are the 3-5 tasks you spend the most time on during therapy sessions? How much time do you spend on them (in percentage and minutes)?
Question 2: Which of the above tasks do you deem the most tedious and/or frustrating? Why?
Question 3: How do you transcribe and organise notes from therapy sessions?
(Optional) Question 4: Imagine you have a digital companion with you during sessions. Which of the following features would be most helpful (you can select multiple or even add your own): transcribing conversations to text, a shared journal with the patient, digital questionnaires, automatic summarizer of previous patient sessions.
(Optional) Question 5: Assume the digital companion from the previous question saves you a few hours per week. Would you use the extra time to better care of your current patients, or would you onboard new ones?
(Optional) Question 6: If you are available for a 15-20 minutes interview, leave your email address below and we'll schedule you at your convenience.

Thank you for your time,
Timm Plagge

* Ruochen has time for the next meeting on Monday and Thursday (past 12)
* Next meeting, we need to finalize the Value Proposition Canvas

## 04.12.23

Contact details: [e-mail]
Greetings!
We are students from the Technical University of Munich (TUM) studying entrepreneurship in healthcare. Our task is to build a tool to help therapists care for their patients, be it during or in-between sessions. We would appreciate it if you could spend a few minutes to answer our questions. If you are short on time, you can just answer the first **three questions** (2-3 sentences would suffice for each, there are 6 questions in total). Thank you for your precious time!
Question 1: What are the 3-5 tasks you spend the most time on during therapy sessions? How much time do you spend on them (in percentage and minutes)?
Question 2: Which of the above tasks do you deem the most tedious and/or frustrating? Why?
Question 3: How do you transcribe and organize notes from therapy sessions?
(Optional) Question 4: Imagine you have a digital companion with you during sessions. Which of the following features would be most helpful (you can select multiple or even add your own): transcribing conversations to text, a shared journal with the patient, digital questionnaires, automatic summarizer of previous patient sessions.
(Optional) Question 5: Assume the digital companion from the previous question saves you a few hours per week. Would you use the extra time to better care of your current patients, or would you take on new ones?
(Optional) Question 6: If you are available for a 15-20 minute interview, please leave your email address below and we will contact you in the future.

## 07.12.23

Value Proposition Canvas

* Customer: Therapist
* Product: MindScribe, the digital companion
* Gains (that the therapist wants):
* Dedicate more of the the F2F session time to the patient, instead of multitasking
* Automatically store and organize patient information (audio and transcriptions synced, searchable and automatically summarized)
* Questionnaires automatically administred to patients
* Pains (that the therapist has):
* Therapist forced to multitask by taking notes, cannot pay full attention to the patient
* Listening to recordings of sessions is tedious: a lot of material to go through, hard to find specific snippets manually
* Therapists overbooked, not enough time to care for individual patients
* Storing information is a task in itself: might not write down everything, written notes can be lost
* Therapists may not be tech savvy enough to create a helpful workflow, e.g. by moving notes online, using 3rd party transcription services for their audio recordings
* Administering questionnaires by hand is tedious: scores have to be computed by hand, therapists send tests and receives a different or edited file with the patient’s answers
* Customer Jobs (what the therapist does):
* F2F sessions with patients
* Take notes and recordings of the F2F sessions, add to the medical record
* Reorganize notes & transcribe recordings after patient sessions
* Assess mental wellbeing of the clients, e.g through questionnaires
* Check medical references and past cases, compare to current patient
* Ensure patient information is safely stored
* Other paperwork, e.g finances, insurance etc.
* Checking in on patients’ journals or take-home assignments
* Gain Creators (what can our product do):
* Intelligent search of digital notes, e.g text or tags attached by therapist
* AI summarizes and extracts key information from transcribed sessions
* Single patient file with all recordings, notes and questionnaires in one place
* Shared digital diary between patient and therapist
* Pain Relievers (what our product solves):
* Use AI to transcribe audio recordings, allowing therapists to focus on the F2F session
* Digitalise the questionnaires so scores can be attached to
* Products & Services (what our product offers):
* Multi platform/ web application with transcription as a core feature
* AI summarizer for the patient file/ case
* Questionnaire tracking/ administration system, i.e. a digital form

## 13.01.24

* We can mention that we considered other customers, but that it is difficult for us to assess their needs, etc. because we don't know much about them (insurance, patients for example)
* 2 BMC, one before feedback, one after. That way we can show how we iterated our thoughts, and how things changed
* Timm has another interview next week
* Ruochen can look at some papers that could be relevant for us (basically, we are looking for market research that we can cite that describe patient and/or therapist needs), would be Chinese market
* So basically we will do 3 pieces of research:
* Therapist interviews (Timm)
* Chinese market research & papers (Ruochen)
* Research for what therapists spend a lot of time on
* What therapists want
* WHat competitors have done
* Legal compliance, digital patient act & maybe GDPR (Dei)
* Meeting schedule:
* Thurs 18:00 : Short meeting on Thursday 18:00 to make sure we are on track
* Sat 18:00 : Proper meeting to do our 2nd BMC iteration & plan our presentation

## 18.01.2024

### Business Plan

##### Customer Segments

* Therapists: Clinical Psychologists, Counselors, Psychiatrists
* Patients (Stakeholders): Individuals seeking therapy. They must be insured in order to be able to store assistant output in ePA
* Insurance companies (Payer): We can only provide services to insured patients since our app falls under ePa and DiGA

##### Value Propositions

* **Faster and Easier Diagnosis for Therapists:**
* Accelerated diagnostic processes through AI assistance.
* Streamlined session note-taking with AI-powered transcription.
* **Improved Patient Experience:**
* Digital questionnaires for automatic diagnostics enhance the patient experience.
* Centralized patient files provide a comprehensive view for therapists.
* **Operational Efficiency:**
* Reduced bookkeeping efforts with automatic transcription and summarization.
* Facilitated discussions allow AI to extract key points, aiding therapists.

#### Channels

* **Online Platforms**: Availability on web and mobile platforms for easy access. The application would fall under DiGA reglementations for encryption

#### Customer Relationships

* **Continuous Support and Training**: Ongoing assistance and training programs for therapists.
* **Feedback Loops**: Regular feedback collection from therapists for product improvement.
* **Collaborative Relationship with Insurance Companies**: Joint initiatives for improving mental health services.

#### Revenue Streams

* Fixed price per session that will be discounted by the insurance provider

#### Key Resources

* **AI Technology**: Advanced algorithms for transcription, summarization, and diagnostics.
* **Knowledge Base**: Up-to-date medical information for accurate diagnostics. used to correlate notes with specialty literature. Correlation between patient cases using reccomendation systems (unlikely to be allowed by regulation)
* **ePA/ DiGA Compliance**: Actively engage and mantain relationships with insurance companies and audit firms to affirm proper handling of patient data.
* **Cloud Computing Providers**: Infrastructure support for scalable and reliable services. Training of AI models since we must rely on our own models by regulation.

#### Key Activities

* **Continuous Improvement of AI Algorithms**: Research and development to enhance transcription and diagnostic capabilities.
* **Sales and Training of Therapists**: Outreach efforts to mental health professionals. Resources dedicated to ongoing support and training programs teaching therapists to properly use and integrate the product in their workflow. These could be digital workshops and wikis available to therapists.
* **Partnerships and Collaborations**: Establishing and nurturing relationships with insurance providers.

#### Key Partnerships

* **Therapists and Medical Professionals**: Collaborative efforts for product development and improvement.
* **Insurance Companies**: Strategic partnerships for endorsement and integration. Source of income.
* **Cloud Computing Providers**: Strategic partnerships for reliable and scalable infrastructure. Resources will have to be localised in Germany and we will serve our own models.

#### Cost Structure
* **Research and Development**: Investment in AI technology enhancement.
* **Product Development**: UX interviews, product development and refinining
* **Marketing and Sales Efforts**: Budget allocation for promotional activities.
* **Cloud Computing Services**: Expenses related to utilizing cloud infrastructure for scalability.

### ePA+DiGA Research
* https://gesund.bund.de/en/apps-on-prescription-content-and-benefits
* https://gesund.bund.de/en/the-electronic-patient-record#future-functions
* https://gesund.bund.de/en/apps-on-prescription-content-and-benefits
* https://www.bundesgesundheitsministerium.de/elektronische-patientenakte#collapse-control-6281
* https://www.myrasecurity.com/en/knowledge-hub/diga/
* https://www.bfarm.de/SharedDocs/Downloads/EN/MedicalDevices/DiGA_Guide.pdf?__blob=publicationFile
* https://www.regdesk.co/bfarm-guidance-on-fast-track-process-for-digital-health-applications-data-security/
* https://www.extrahorizon.com/navigating-the-complexities-of-data-security-in-healthcare-a-deep-dive-into-encryption

### Meeting

* Timm interview:
* Was decently useful, will write it up so we can directly put it into a PowerPoint slide for our presentation
* What the person said -> how we changed our idea

---

* Quick outline of presentation:
* Introduce our initial business canvas (PowerPoint) (30s)
* Show our research (but only vaguely discuss, just show we did it) (30s)
* Show our final BMC, discuss in more detail (4 min)
* Every person talks about one of the sections (can be roughly cut into 3 sections)
* We can use the sticky notes to discuss the canvas one step at a time
* Final discussion & feedback ()
* Note that lecturer mentioned the timing will be quite strict this time, so we should practice our speeches

### Timm feedback on 1st BMC

* Customer segments:
* 3 choices good, maybe we explicitly state we focus more on therapists and insurance? For patients, it only has to not be annoying
* Ultimately our goal mnust be to satisfy the insurance, but at start we can focus mainly on therapists
* Value Proposition:
* Maybe explain more about *why* these things lead to more value, not just *how*. E.g. happier patients, reduced costs etc.
* Show how value proposition links back to therapist and/or insurance
* Channels:
* No offline mode? In theory, we could provide hardware to do transcription locally (but less flexible and more costly to develop)
* Maybe we should brainstorm a second channel for insurance
* Customer Relationships:
* I think we should keep our relationship with therapists more personal at the start, because we will be looking for a lot of feedback. I.e. direct support line and training, later we could automate this.
* We should focus on feedback from therapists and insurance, but only ask for patient feedback through the therapists, to filter out a lot of noise.
* Revenue Streams:
* Looks good, not much to say really
* Key resources:
* Think we need to define knowledge base a bit more precisely, to differentiate ourselves
* Cloud Computing is only a resource for us, we will not have much of a relationship with them outside of using their services (doubt we would grow big enough to need direct support service)
* Key activities:
* Updating system is easy if we are purely software based, just need to be compatible with different OS (security may be difficult)
* Partnerships:
* Insurance can give use feedback on legal, and possibly statistics on how our product affects mental wellbeing across large number of patients (different to therapist, who only has a small insight but more personal)
* Focus on personal relationships with therapists
* Cost structure:
* Our 2 biggest cost drivers will be development of AI system and computing costs, should probably explicitly state (although maybe conforming to privacy guidelines may also be expensive)
* Since we are not directly catering to patients, probably do not need something like a public ad campaign, so advertising not a big cost. We can try to spread through word of mouth

## 22.01.24

* [Initial BMC draft from Dei](https://docs.google.com/document/d/14Dg1zj2jIDUzoX66GZM_8hd50dc67915b4SYixIaipc/edit)
* Ruochen feedback
* Need to monitor wellbeing of therapist themselves (therapist needs), as can be very stresseful
* In China patients prefer to use online platform, instead of going in person. This has fee associated with it, such as renting video service or a room to do meetings.
* Patients cannot always control their own feelings, patients and therapists may feel very tired after sessions. -> burnout
* Patients often look for senior therapists (1500+ hours of sessions), instead of going to younger therapists. Demand for therapists is lopsided
* Having access to a large knowledge base would maybe reduce this imbalance?
* Therapists also need people to talk to about their work, such as about stress, feedback, questions etc.
* Develop a platform that allows exchange of information
* Strategy to help therapist and patient relax
* Suggestion from app on how to use different techniques/tools to improve session?
* Product could put together information recorded across many sessions to figure out patient's overall problems and solutions.
* Encourage patient to write down more of their personal thoughts, to give a more accurate diagnosis.
* Product could process information to figure out alternative forms of therapy, make suggestions.

* Meeting to finalize the BMC and presentation: 9:00 at Stammgelande
*

## 24.01.24

Script for presentation (5 min)

* **Slide 1 (Title page) (Ruochen)**
* Today we will present our business canvas for MindScribe, the digital therapy platform.
* **Slide 2 (Itineray) (Ruochen)**
* Our presentation will start with our final canvas, and if we have time we will discuss our research and initial canvas, which was based on our value proposition canvas
* **Slide 3 (complete BMC) (Andrei)**
* MindScribe is a digital platform for helping therapists and counselors provide better care in the digital age. We rely on AI to automate tedious tasks, freeing them up for patients.
* **Slide 4 (Therapist BMC) (Andrei):**
* We believe therapists need an IT department to centralise the many ways they take notes: paper, password-locked files, physical recorders. Most of them have little time to worry about security, which is the reason they avoid the digital.
* To this end we propose a unique digital platform that works with the devices they already own. The platform helps them record and organise notes, monitor patient progress and diagnose.
* The main feature of the platform will be automatic transcription and summarization Therapists will use the platform to record sessions, and notes magically appear to their patient file case.
* Once the notes are in, therapists will leverage a digital knowledge base, allowing them to correlate the patient case over multiple session with specialty literature, suggesting treatment approaches and diagnostics.
* Besides the knowledge base, we will digitalise often-used questionnaires, further aiding the diagnostic.
* Compliance is of upmost importance. Just like IT departments, we will manage the devices remotely, ensuring they are up to date, have firewalls turned on or screen lockdown enabled. If the device is not secure, data recording will not be possible.
* **Slide 5 (Insurance BMC) (Timm):**
* This is the BMC showing value flow to the insurance customer.
* Insurance is interested in ensuring that patient data is safely stored and processed. We achieve this by encrypting data and communication on our end, and offering IT support for therapists showing them how to safely use our app.
* Insurance is also interested in how our app makes a difference during therapy, and we can do this by having a metric board. This would not show patient information, but general statistics across many therapy sessions, such as how much time is being spent on various activities during therapy.
* **Slide 6 (Cost and Revenue) (Ruochen):**
* About our Key Partnerships: The public health system determines what our products can do;publishers give us access to their literature that we use to feed our knowledge base, and can also provide to therapists.
* Our main costs are Computing Resources to train our model, literature licenses to add to our knowledge base; Marketing to therapists at Professional Events
* We monetize our product by offering a subscription service for therapists based on how many hours they us it.
* **Slide 7 (Product Research) (Timm, Ruochen, Dei):**
* **Ruochen**: For product research, I read a lot of articles about different types of counseling services in China, and I found that therapists commonly face challenges: Shortage of New Visitors, Financial Stress, Burnout, and the integration of AI+Internet, which allows patients to contact therapists anytime, anywhere, and reduces therapist expenses (e.g., rental fees), replaces trivial things, and thus makes therapist's work easier!

Translated with www.DeepL.com/Translator (free version)
* **Dei**: The product can be iterated in a lean manner. It is not mandatory to integrate with ePA, and we are not under the DiGA provisions. The AI models must run on our servers and we must have a paper trail and storage policies for the data.
* **Timm**: I conducted interviews with therapists in training to see what kinds of problems they are facing. The biggest take from the interviews is that therapists are running out of time: they don't have enough for every patient, and they also have to rush their notes at the end of sessions, not during. We think our transcribing & summarization services are a big help there.
* **Slide 8 + 9 (Initial BMC) (Timm):**
* A quick comparison between our initial and final canvas: red is removed, orange we kept and green on the next slide is what we added. For the most part, we kept our customers the same, but made our value propositions and key activities more specific, such as how exactly we would ensure compliance.

## 29.01.24

* Final presentation (do you guys think someone else is going to evaluate us???)
* 2 minutes sales pitch
* 6 minutes rest of business plan
* Flyer (for potential investors, marketing 101)
* Who we are
* What we are
* Why you should support us
* Mission statement and vision (more general)
* Print it out and give it to them at the start of the presentation
* Things that need to be done:
* Should agree on who we are (made up profiles/team)
* Realization plan (a bunch of estimates):
* The actual product itself (list of features, how it should look, very idealised but we kind of need something at this point)
* How we will penetrate the market (market entry, talk to therapy associations?)
* How we will expand and grow
* Amount of initial capital needed
* Time needed to launch MVP
* Pricing:
* cost estimates on our side
* cost for the therapist and insurance
* return on investment for the investors (long and short term)
* how long
* estimates of our growth (look at other companies)
* Benchmarking? Mentioned in the lecture slides
* Customers
* Who are we catering to
* Competitors:
* From our basic knowledge, there doesn't seem to be a proper transcribing system in the therapy industry yet.
* If we are pioneers, are there other companies who could easily catch up and surpass us quickly?
* Maybe look at business model of ChatGPT and OpenAI?
* Market:
* Barriers to entry (customers, capital, regulation)
* Possible resistance in adoption of product
* Who we need to work with
* What is our market niche to begin with?
* Sales pitch (this is probably last thing we do)
* Cool name and logo
* Cool presentation
* Mission statement and vision (more specific than flyer)
* SWOT analysis:
* Probably do this in the middle of our work, after we have initial business plan, and then use SWOT analysis to do final touches and work out how we will solve major problems or capitalize on our strengths
* Quick work division:
* Ruochen:
* Will look at templates for presentation and flyer (busy until Friday)
* Find a common theme (color scheme)
* Timm:
* Realization plan
* Competitors
* Dei:
* Pricing
* Competitors
* Together:
* Initial SWOT analysis
* Work days:
* we will sadly have to pull through on Saturday and Sunday (this is a lot of work!!!)

## 04.02.24

* Some small notes that I think we have missed:
* We need to talk more about difficulties:
* Customer resistance
* Ease of use?
* How much time to accept our product?
* Need to discuss solutions as well
* Can we patient or protect our product? Investors would be interested in this
*