Skip to content

Meetings

Jump to bottom
Insa Suchantke edited this page Jan 18, 2020 · 14 revisions

Meetings with Tobias - Notes

2020-01-13 Meeting with Maria "Forschungswerkstatt"

  • Palliativversorgung instead of Palliativmedizin

  • Ethics commission

    • 20th of Jan (or later)
    • complete design
    • also questionnaire

  • for publication

    • not publishing twice the same
    • Congress
      • submit abstract --> national and international
      • naming results in titel
    • Abstract not always exactly same title
    • name country/region (not in magazine, congress may be possible (study just in Erlangen))
    • Citation: "Satzung von der FAU" (gute wissenschaftliche Praxis)

2019-12-16

  • Publication

    • 2 Abstracts (one for the UX study and one for the data analysis) -> maybe EAPC (European Association of Palliative Care), 2021 in Helsinki (submission for 2020 is already closed)
    • 1 Paper with both UX study and data analysis -> Palliative Medicine, Palliative Care, Zeitschrift für Palliativmedizin,Palliative Supportive Care ?
    • Are there any requirements regarding the journal? Impact factor, peer reviewed?

  • UX Study

    • ~ 10 participants (Tobias said that it shouldn’t be a problem to find 10 people)
    • we need an ethics approval, but that shouldn’t be a problem (according to Tobias), deadline for the application is 21.01.2020 and we would get the approval on 4./5.02.2020, more information here
    • Tobias invited us to a meeting of the “Forschungswerkstatt” (13.01.2020) of the palliative care department, with Maria, they can give us information about how to interview participants, how to create questionnaires, …
    • getting more scales so that more patients can use the scale (Tobias offered to buy them, or we buy them ourselves)

  • Data Analysis

    • new dataset with psychometric data gathered over time -> analyze the progression of the symptoms
    • psychometric dataset -> predict needed nursing care level in the e.g. next 5 month
      • logistic regression with rule mining
      • cross-sectional correlation
      • predict with the same rule mining approach and predict progressions
    • predict dehydration level based on scale data

  • App

    • MIDOS:
      • How will it be managed with the input of the two more sufferings?
      • Patient has to reenter their sufferings each time (headache, ...)
      • Or the old sufferings are saved and next time patients can select them?
        • yes, this works --> should be displayed again
        • but don't showing result from last time
      • Medical qualified person: makes 3 more questions with same choices as in the MIDOS (extended MIDOS)
      • IOPS is international (Palliative outcome scale)
    • idea: think about integrating emergency call services (Hausnotrufdienst, Rotes Kreuz, …) in the region
      -> the patient clicks a button and the emergency call service gets an alert and then they call the patient, if nobody answers the phone, they will go there and check on the patient or will call a deposited contact (neighobour, nok) (or maybe call an ambulance)
    • if the questionnaire is filled out by someone else who is somehow medical staff or medically qualified there should be follow-up questions (like overburdening of the family)
    • our system makes communication and interaction with the patient possible (which is one big advantage of our system) with the individual feedback and the emergency call functionality
      -> integration video calls would be a possibility to further improve this

  • Manual input

    • who much water the patient drinks (<500ml, 500 ml – 1 l, 1 l – 1,5 l, …., >2 l, and “I don’t know” should also be possible)

  • Doctor’s web app

    • green, yellow, red cards only with name -> what should happen after the doctor clicks on one of these cards? -> all parameters, the critical ones first and the other ones below,
    • additional information: name, age, gender, address, date of birth, diagnosis, also info which doctor added him/herself, relevant secondary diagnosis
    • if the doctor look at one patient’s data, it should still be possible to see if the condition of another patient changes/gets critical, so that the doctor doesn’t miss that
    • while in a view with one patient, doctor should also see when other patient cards get red (alarm)
    • there should be a possibility for data input from the doctor and that data should also be displayed for each patient (after the doctor selects one patient in the overview screen)
    • want to talk to them --> Video App Call from SAPV Team

  • Is PalliCare a registered trademark? Are we allowed to use it?

  • PalliDoc (they enable interaction) --> for palliative providers to manage that

2019-11-25

  • Dataset from Tobias ("Patients in palliative care - Development of a predictive model for anxiety using routine data")

    • dataset is publicly available, see "Supporting info 4"
    • Illness and psychometric data
    • Related to anxiety
    • More than 10.000 patient data
    • Also ECOG, if having brain tumor, ...
    • What can we do with it?! Data-mining approach?!

  • Biometric data

    • Getting simulated database from Tobias which contains realistic but not real data, maybe 10-20 patients
      (one of Tobias HiWis will take care of that)
    • Some kind of rule based analysis of the patients' biometric data -> these "rules" would be provided by experts
    • Intra-individual trends: for example, this param changed significantly over a defined period of time
    • NOT: Database of Prof. Prokosch -> this would be the "normal hospital data", for palliative patients it's only psychometric data, otherwise almost text -> which is basically the same as the public database from the anxiety paper, so we can just use that one
    • Delphi-Method: systematic, multi-stage survey procedure with feedback and is an estimation method which serves to be able to estimate future events, trends, technical developments and the like as well as possible (https://de.wikipedia.org/wiki/Delphi-Methode)
      -> this method would be used to generate rules/trends (e.g. if that value increases that much in that period of time, then the patient's status critical) which we then use as base to give patients feedback about their data
      -> Tobias would take care of that and the just provide us the results

  • Desktop-App for Doctor/SAPV-team (spezialisierte ambulante Palliativversorgung)

    • Doctor and SAPV-Team using same Interface (which should ideally be a desktop or browser app)
    • Overview which shows all patients, the status of the patient should be clearly visible (traffic light colors: everything normal -> green (or maybe just white), something wrong -> yellow, critical -> red)
    • View must be arranged in a way that it is immediately visible which patients are ranked as critical --> data security (?)
    • if doctor clicks on one patient shown in the overview, he sees more information and it's highlighted what param is critical

  • Next of Kin

    • should have the same interface as the patient (the nok is able to do and see the same things as the patients)
    • data can be entered by the next of kin for the patient (and obviously by the patients themselves) -> it has to be logged if this value was entered by the patient or the nok, this could be important regarding data analysis as the nok is presumably biased)
    • nok should see if patient is currently not feeling good (also like in the patients overview screen)

  • Outlook

    • for the SAPV-Team -> system could provide fastest/shortest/"best" route from patient to patient, also regarding the patients' current status

2019-11-08

Tobi´s Overview

  • Medical Devices —> both in our database
  • Psychometric Data
  • Patient / Next of Kin
  • HCP/ Ph
  • Graphic UI: Appetite, Anxiety, … --> In- or Output Voice
  • Information (suggestion) --> for Patienten
  • HCP: only Information (suggestions: medical device) --> to patient & the other way around
  • Motor: would use some kind of algo looking for some rules/ patterns --> only scientific

Our Presentation

  • Tobi likes the logo —> little changes

  • Overview: Next of Kin

    • Proxy Assessment (when patient is too sick) -> proxy ist ein Stellvertreter, Bevollmächtigter, kann zB ein family caregiver sein oder auch ein Arzt
    • Patient/by proxy
    • Educational aspect
    • desktop version would be preferred for doctors, ambulatory care maybe phone/tablet ok

  • Backend implementation

    • Alternative: Server within frame of hospital
    • Decentralized storage of data --> PC at patient´s home (because of data security)
    • Data goes from smartphone to Raspberry Pi & then can be transferred

  • 1.User Story

    • Psychometric Data: only if yes/higher number is ticked, then an additional questions will be added (additional questions will be provided by Tobias) (not if zero is ticked)
    • use validated version of midos questionnaire for the psychometric data, maybe later include ipos questionnaire (IPOS)
    • for first sprint: only one parameter and focussing on it
    • Bloodpressure & pulse and bodyweight & bodywater measuring together in one medical device
    • Oxygen saturation --> manual input
    • We decide which device we want to have —> Tobi will buy it
    • Biometric data: drinking (volume of fluid) (manually), how often goes to toilette (frequency of stools), how mobile a patient is (backlog)
    • ECOG: 0 to 5 (dead) stages, describes how mobile a Patient ist (nr. 3: more than 50 % in bed than outside) https://www.kompetenznetz-leukaemie.de/content/aerzte/scores/performance_status/

  • No yesterday option (depends on patient) --> Algo decides by itself

  • Reminding Patient (call or ringing sound)

  • Scan paper -> not so useful

  • Next of kin with alarm

  • Patient / by proxy

  • 3.User Story —> later on!

    • automatic alert when biometric values are in critical range -> doctor and patient define critical ranges together as they are highly dependent on the patients e.g. Patient always has high blood pressure —> Alarm goes on at different values than at other patients (scientific evaluation)

MockUp App

  • Big letters & icons
  • Patients aged 68 and older
  • Bright and light is good but contrast should be clear
  • Icons must be easy to differentiate
  • AlertButton with cross or SOS
  • Medical devices: show if they are still connected and if there is still battery left
  • Educational Video how to use it (Walkthrough), everything kind of guided
  • people do not try out what happens if they want to click the button, they want now before what will happen -> how to do that?
  • maybe other hints on the buttons (suggestions were something like “Wie geht es mir heute” for the questionnaire and “Was habe ich heute gemessen” for the biometric data)
  • Want to see my history: should always be displayed
  • print single scores of midos, not the overall score (because it is not that meaningfull), don’t show zero symptoms

—> Offer that 2 elderly people look over MockUp and make suggestions / improvements

Clone this wiki locally