You don’t drink? You die!

Dehydration (the lack of water within the body) is a serious threat. You may feel thirsty if your body wants to get hydrated, but your grandma probably doesn’t. As the feeling of thirst decreases with age, the need of water intake rises, due to reduced kidney functionality and more medication which needs to get washed out of the body. High demand of water with low liquid intake = problem. In in 2015, in Germany alone 100.764 people, mostly older than 70, got hospitalized with the primary diagnosis “acute dehydration”, leading to direct costs of more than 275 mio. € for the health system. Not speaking of the > 5000 deaths, and all the suffering and costs related to the possible implications of dehydration: paradox diarrhea, fall, cardiovascular diseases, kidney malfunction.

Why doesn’t anybody do something against dehydration?
Well, people do. Risk patients are monitored with a so-called hydration protocols, a table where the nurse notes down the volume of drinks given to the person of concern, a daily sum is calculated and – if too little was drunk – alarms are raised.

The problem with this approach: it takes the nurse some minutes every day. Time in nursing homes is expensive, and for that reason hydration-protocols are usually not applied in all, but only few cases of higher risk. But it is not these high-risk cases, which result in all these deaths and millions of costs. It’s the sum of those individuals, who are not monitored.

Solving the problem by automation

If in the current state of care liquid intake is not monitored because of the additional time needed for this task, then there is the need to automate fluid intake monitoring.

In other words: we want to minimize the time of human interaction needed for this task. Today, human interaction is needed in two major categories of liquid intake monitoring:

-> We need a sensor device, that automatically measures the drunk volume. To reduce errors (e.g. spilling), it would be best to do so as close to the process of drinking as possible.

-> The measured data should be analyzed and processed automatically. The human should only be needed to act if the hydration level is low.

With these two points and the requirements in mind, the following architecture seems legit:


Intended way of use

  1.  When preparing drinks, the caregiver in the nursing home fits a suitable version of the mouthpiece to the drinking container. The mouthpiece comes in different forms: e.g. the tip for a regular straw, the lid of a feeding cup or the screw cap for a coke bottle.
  2. When handed the drink, the individual clicks in her/his personal sensor unit. Now, the assembly measures the volume he/she drinks and saves the value within the sensor unit.
  3. When finished, the sensor unit is detached from the mouthpiece and remains at the individual. The mouthpiece is collected and put into the dishwasher, with many others.
  4. In the end of the day, a nurse walks by with a phone or tablet, which automatically collects all the data saved on the sensor units, pops up warnings if neccessary and hands over the data for automatic documentation

Example of use: mouthpiece (1) with attached sensor unit  (2) mounted on a regular straw

Additional requirements

Edit: so far, all requirements are met. Even a demo app is developed, to showcase the usability (see log 10). There is an issue with carbonated liquids, they can not be measured accurately. For tap water, the device is tremendously accurate (log 8).

Origin of the idea, IP and dedication to the open source movement: 

This project is a private developement as follow-up of research which was done by J. Kreutzer, J. Deist and myself at the Institute of Micro Technology and Medical Device Technology (MIMED) at the Technical University of Munich. Results of the research were published here. HydrObserve as private project is fully permitted by Prof. T. Lüth, the chairholder.

Personally, I think the problem of dehydration is too simple to be left as it is, especially if the research mentioned above already guides a possible solution. My first thought was to do this with an entrepreneurial approach, but quit doing so for good reasons:
a) I don't have time for that
b) At least in Europe, we have the problem that the ones who apply the product to be (nursing homes) are not the ones who pay the bill if they don't (health insurance does). As insurance does not have any authority to make nursing homes use the product, and the latter tend to be not so progressive, a market entry might be tough.
c) Bosch GmbH holds a very broad patent in the topic (DE102010041112A1) without ever having developed a product (a.k.a. patent troll?). I'm sure this patent can be bypassed, but I didn't feel like wrestling with a multi without proper legal advice. Which was too expensive for my cause.

In this few entrepreneurial weeks, I protected my idea of a modular system (-> see 2nd build log) for Germany with a "Gebrauchsmuster", a kind of unsafe micro patent (Nr. 20 2017 000 580, IPC G01F 22/00). With this open source project here on Hackaday, I declare all my IP that I hold with it under the open source license of this project.

If you see a chance to do transform this sketch of mine  into a real product, preventing all the suffering of dehydration: do it and get make as much money as you wish!

All progress will be documented intensively in the build logs. Only major advances are added here in brief comments.

UPDATES:

  1. Found the perfect sensor for this application: a tiny magnetic turbine + hall sensor (log 1).
  2. Used materials of Chinese sensor pass basic food safety test (log 2).
  3. The architecture of the device will be modular and with this avoid problems of available solutions (log 3).
  4. The housing/mechanical parts of the device are considered fine for now in the second iteration
  5. Electronics are working as they should in the second iteration
  6. Firmware working. See https://github.com/jflaschberger/hydrobserve_firmware.git
  7. Performance test was very positive! See log 8 for details.
  8. Prototypes of various mouthpieces successful (feeding cup lid and screw cap). See log 9.
  9. Android demo app now available! See log 10 for details and download apk file in download section. Source code available here: https://github.com/jflaschberger/hydrobserve_demo_app_android.git