URINE OUTPUT MONITORING

A Simple and Reliable Device for Monitoring Critical Patients’ Urine Output

Abraham Otero

Department of Information Systems Engineering, University San Pablo CEU, 28668 Madrid, Spain

Teodor Akinﬁev, Andrey Apalkov

Center of Automation and Robotics, Technical University of Madrid, Spanish Council for Scientiﬁc Research

(CAR UPM-CSIC), La Poveda, Arganda del Rey, 28500 Madrid, Spain

Francisco Palacios

Critical Care Unit, University Hospital of Getafe, Getafe, Carretera Toledo KM 12.500, 28901 Madrid, Spain

Jes´us Presedo

Department of Electronics and Computer Science, University of Santiago de Compostela

15782 Santiago de Compostela, Spain

Keywords:

Biosensors, Urine output, Critical care, Patient monitoring.

Abstract:

Currently, critical care units are equipped with sophisticated commercial monitoring devices capable of sens-

ing most of the patient’ physiological parameters, and of automatically supervising whether the values of the

parameters lie within a preestablished range set by the clinician. The automation of these tasks has discharged

the healthcare staff of a considerable workload. It also avoids human errors, which are common in repetitive

and monotonous tasks.

In all likelihood, urine output is the most relevant physiological parameter that has yet to be sensed or su-

pervised automatically. This paper presents a patent-pending device capable of sensing and supervising urine

output. The device uses reed switches that are activated by a magnet that is attached to a ﬂoat in order to

measure the amount of urine collected in two containers. When a container ﬁlls, it is emptied automatically

using a siphon mechanism and urine begins to collect again. An electronic unit sends the state of the reed

switches via Bluethooth to a PC. From this information, the PC calculates the urine output and supervises the

achievement of therapeutic goals. The end result is a fully automated, simple, inexpensive and accurate urine

meter.

1 INTRODUCTION

Current critical care units are equipped with sophisti-

cated commercial monitoring devices that allow clini-

cians to sense nearly any physiological parameter that

may provide information relevant for interpreting the

patient’s state. In most cases, these devices can also

supervise that the values of the physiological param-

eters they sense remain within a preestablished range

set by the clinician. This range represents the values

considered as normal for each parameter. If a param-

eter does not fall within its acceptable range, the cor-

responding sensing device alerts the healthcare staff

by means of an audible warning (Otero et al., 2009b)

(Hande et al., 2006).

These devices discharge the healthcare staff of a

considerable workload, since they need not continu-

ously supervise if the physiological parameters of ev-

ery patient lie within the acceptable range. They also

avoid human errors, which are common in any repet-

itive and monotonous task such as the one at hand

(Jungk et al., 2002) (Mora et al., 1993).

Otero A., Akinﬁev T., Apalkov A., Palacios F. and Presedo J..

URINE OUTPUT MONITORING - A Simple and Reliable Device for Monitoring Critical Patients’ Urine Output.

DOI: 10.5220/0003125500050013

In Proceedings of the International Conference on Biomedical Electronics and Devices (BIODEVICES-2011), pages 5-13

ISBN: 978-989-8425-37-9

 2011 SCITEPRESS (Science and Technology Publications, Lda.)

Arguably, the most relevant physiological param-

eter which still is measured and supervised manually

by healthcare staff is the patient’s urine output. Urine

output is the best indicator of the state of the patient’s

kidneys. If the kidneys are producing an adequate

amount of urine it means they are well perfused and

oxygenated. Otherwise, it is a sign that the patient

is suffering from some complication. Urine output is

required for calculating the patient’s water balance,

which is essential in the treatment of burn patients

(Mitra et al., 2006). Finally, it is also used in multi-

ple therapy protocols to check whether the patient re-

acts properly to treatment (Rivers et al., 2001). When

urine output is too low the patient is said to have olig-

uria. If the patient does not produce urine at all, then

he/she is said to have anuria. Sometimes, urine out-

put can be too high; in these cases the patient is said

to have polyuria.

Currently, in critical care units, urine is collected

in a graduated container that is connected to the pa-

tient’s bladder through a Foley catheter. Periodically

the nursing staff manually records the reading of the

container of every patient, and operates a valve which

releases the urine into a larger container. In critical

care units of ﬁrst world countries, this procedure is

usually performed every hour, 24 times a day, 365

days a year. In the case of emerging countries, often

only the burn patients –for whom urine output moni-

toring is of paramount importance– have this param-

eter registered every hour, while the remaining criti-

cal patients have it recorded every 2 or 3 hours. This

disparity in criteria is due to the more reduced avail-

ability of healthcare staff in the critical care units of

emerging countries.

It can even be argued that the monitoring interval

currently employed in ﬁrst world countries –once ev-

ery hour– also is a compromise between avoiding risk

states for the patient and relieving the nursing staff

of an excessive burden. A system capable of auto-

matically monitoring urine output would decrease the

workload associated with this task and, at the same

time, it would permit supervision to take place on a

more continuous basis.

This paper presents a patent-pending device capa-

ble of sensing and supervising urine output (Akinﬁev

et al., 2010). Section 2 reviews related work. Our

device is described in Section 3. It uses reed switches

that are activated by a magnet that is attached to a ﬂoat

in order to measure the amount of urine collected by

two containers. An electronic unit checks the state

of the reed switches and sends it to a PC, which su-

pervises the achievement of the therapeutic goals that

have been established for urine output. Section 4 dis-

cusses the results of this work. Finally, a series of

conclusions and lines of future extension are given.

2 RELATED WORK

Automating the supervision of urine output is a

problem that has only begun to be addressed by

the biomedical engineering community very recently.

There are several problems that have contributed to

this delay. On the one hand, it requires devices capa-

ble of measuring very small amounts of ﬂowing liq-

uid (up to 3-5 milliliters per hour). This precludes the

use of common industrial solutions such as ultrasound

sensors (Johnson, 1978) or commercial ﬂowmeters.

On the other hand, any component of the device that

is in contact, or may be in contact, with the urine

cannot be reused on other patients and must be re-

placed approximately every 4-7 days for hygienic rea-

sons. Therefore, any component of the device that is

or may be in contact with the urine must be easy to

dispose of, and should have a low price. This pre-

cludes the use of expensive high precision laser based

solutions (Ishida, 1990). Furthermore, contact with

the urine also means that the component is in indi-

rect contact with the patient’s bladder through a Foley

catheter. Therefore, the components that come into

contact with the urine have to be sterilized before use.

Finally, urine contains Uric acid, Sodium, Potassium,

Chlorine and other components which make it corro-

sive, especially for metals.

The ﬁrst device proposed to automatically mea-

sure urine output was Urinfo 2000, developed by the

Israeli company Medynamix (Hersch et al., 2009).

Urinfo 2000 was designed to automate the hourly

urine output measurement, but not to take more fre-

quent measures. Its operation is based on counting

the number of drops of urine produced by the patient,

and from this count it estimates urine output. The av-

erage error of the device when used to take hourly

measurements was 8% (±25 ml). Urinfo 2000 cannot

supervise therapeutic goals and its readings are not

transmitted to a central station/PC. Thus it requires

the nursing staff to take the urine output measures

from the device’s display, placed next to the patient’s

bedside.

The authors of this paper proposed a device to au-

tomate the supervision of urine output in (Otero et al.,

2009c). Several technical and legal problems pre-

cluded us from moving this device beyond the lab-

oratory validation phase. Using the knowledge we

gained building it, we developed a second device with

the main objective of conducting a series of clinical

studies based on more continuous and accurate mon-

itoring of urine output throughout the stay of patients

BIODEVICES 2011 - International Conference on Biomedical Electronics and Devices

in a critical care unit (Otero et al., 2010) (Otero et al.,

2009a). This device uses a high precision scale to

measure the weight of a commercial urine meter. On

the scale’s pan there is a support frame made up of

Bosch proﬁles that isolates the scale from force trans-

mission from the patient’s bed, and guarantees that

the urine ﬂows properly through the urine meter’s in-

put tube. The maximum measurement error of this

device is under 1.5%.

Currently this device is in use at a research unit

associated with the University Hospital of Getafe, in

Spain. In this research unit a series of experiments

aimed at the study of sepsis in an animal model –pigs–

are being conducted. The aim of these experiments is

twofold. On one hand, we would like to gain a better

understanding of the pathophysiological mechanisms

underlying systemic and renal hemodynamics during

sepsis –hence the interest in a continuous and accu-

rate monitoring of urine output. On the other hand,

we would like to deﬁne the optimum monitoring in-

terval for urine output, and to determine the level of

accuracy required for this task.

The high accuracy and acquisition rate of this de-

vice make it ideal for carrying out clinical studies.

However, this device was not intended to be used in

the clinical routine. Its size and operation make it

somewhat tedious to be used in a critical care unit.

Conversely, in using this device we have learned that

its accuracy –1.5%– and measurement interval –up to

10 sec.– are superior to what is required for properly

supervising urine output. Our experience shows that

an accuracy of 5% and a monitoring interval of 5-10

minutes is enough. Therefore, these parameters can

be relaxed to obtain a cheaper and simpler device.

That is the goal of the device presented this paper.

3 THE DEVICE

In this section we ﬁrst describe the general operating

principles on which our device is based. Although

we will use the device to measure urine output, the

principles on which it is based are general and could

be used to measure the volume of any liquid ﬂow-

ing through a tube. Then we describe the prototype

we have built, and the calibration procedure we have

carried out to determine its operating parameters. Fi-

nally, we describe how the supervision of therapeutic

goals is carried out.

3.1 General Operating Principles

In our device the urine arising from the patient’s blad-

der is collected through an input tube in a container.

Output

Inputtube

Float

Limiter

Magnet

Topopening

Reedswitch

Controlunit

Floatmovement

tube

Figure 1: Diagram of the small container of our prototype.

This container is equipped with an output tube that

empties its contents when it gets ﬁlled by taking ad-

vantage of the siphon principle (see Figure 1). One

end of the output tube will be placed in the bottom

wall of the container. Ideally the bottom wall of the

container will be inclined or conical and the tube will

be in the lowest part. This favors a more complete

evacuation of the container. Both ends of the output

tube are always open. When the container begins to

receive liquid, the liquid goes up inside of the out-

put tube until it reaches the elbow located in its top

part. When it reaches the elbow, liquid begins to fall

down the portion of the tube located on the exterior

of the container. By the principle of the siphon, the

liquid will ﬂow down the tube and the container will

be emptied.

For the system to function properly, the output

tube should not be too wide. If the tube is too wide,

when the water starts ﬂowing, air could rise through

the tube towards the container, breaking the siphon

principle. On the other hand, to drain the container as

quickly as possible, it is desirable that the tube be as

wide as possible. A compromise between these two

goals can be achieved using a variable diameter tube,

so that its diameter grows (staggered or progressively)

from the elbow of the tube to the end which is outside

the container (see Figure 1). In this way, the principle

of the siphon will work properly because the tube is

narrow enough in its upper part, but when the water

begins to fall it ﬁnds a wider tube which provides less

resistance, thereby increasing ﬂow rate.

The container must also be equipped with a ﬁl-

tered top opening that is used to equalize the internal

and external pressures without risk of bacterial con-

tamination. This opening should be located at a height

URINE OUTPUT MONITORING - A Simple and Reliable Device for Monitoring Critical Patients' Urine Output

Open

(a)

(b)

Figure 2: Operation of a reed switch. In the presence of

a magnetic ﬁeld, the switch is closed. In the absence of

magnetic ﬁeld, it is open.

above the elbow of the output tube. The liquid

will not start ﬂowing immediately when it reaches the

elbow height; but it will continue to accumulate in

the container until the height of the column of liq-

uid which is above the elbow height overcomes the

surface tension of the liquid against the walls of the

output tube. The surface tension increases as the di-

ameter of the output tube decreases. The height H

and the effective volume of the container V

(the vol-

ume of water contained just before the draining starts)

must be determined experimentally.

Ideally, the upper wall of the container will also

be inclined or conical in shape. The top opening to

equalize the internal and external pressures should be

placed in the uppermost part of the wall (see Figure

1). This prevents bubbles from forming in the upper

wall of the container, bubbles that would occupy vol-

ume and distort the measures.

One or more reed switches are placed on the out-

side of the container that will receive the urine. A

reed switch contains two or more magnetizable, ﬂexi-

ble, metal reeds hermetically sealed in a tubular glass

envelope whose end portions are separated by a small

gap. Under these conditions, the switch is open (see

Figure 2a). A magnetic ﬁeld properly applied will

cause the reeds to bend, and the contacts to pull to-

gether, thus closing the switch (see Figure 2b).

A ﬂoat within a structure designed to limit its

movement so that the ﬂoat can only move vertically

is located inside of the container, near the container

wall where the reed switches are placed (see Figure

1). The ﬂoat has a magnet attached which interacts

with the reed switches closing them when the mag-

net is approximately at the same height as each of the

reed switches.

An electronic unit is connected to the reed

switches, continuously checking their state. For the

general case in which there are N reed switches on

the outside wall of the container, the procedure for

measuring the volume of liquid ﬂowing into the con-

tainer is as follows. At least one reed switch should be

located in a position such that when the container is

empty, the magnet located on the ﬂoat closes the reed

switch. As the liquid begins to ﬂow into the container,

the ﬂoat, and therefore the magnet, begins to raise. At

some point, the magnet will stop interacting with the

ﬁrst reed switch and, therefore, it opens. At that point,

a volume V

of liquid has ﬂowed into the container.

When enough liquid has been accumulated in the con-

tainer, the magnet will close the second reed switch.

At this point an additional volume V

of liquid has

ﬂowed into the container, being the total amount of

liquid accumulated V

. When the magnet moves

higher, the second reed switch opens again and an ad-

ditional volume V

of liquid has ﬂowed –being the to-

tal volume of liquid V

In general, when the reed switch N is closed, we

will add the volume V

2N−2

to the the volume of liquid

that has ﬂowed, and when the reed switch N is opened

again, we will add the volume V

2N−1

. When the con-

tainer is emptied, the ﬂoat with the magnet will go

back to the bottom of the container, and therefore it

will close the ﬁrst reed switch. At this point, the ef-

fective volume of the container –V

– has ﬂowed, and

the measurement procedure is resumed.

The volumes V

2N−2

and V

2N−1

for each reed

switch, and the effective volume V

must be deter-

mined experimentally.

3.2 Our Prototype

The general operating principle described in the pre-

vious section suffers from a problem: the volume of

liquid that ﬂows into the container from the time the

container begins to empty through the siphon mecha-

nism, until it is completely empty, will not be mea-

sured. Depending on the speciﬁc application and

characteristics of the container, this may or may not

be tolerable. The problem that concern us, the mea-

surement of the amount of urine produced by a pa-

tient, requires an early warning about deviations from

the therapeutic goals. Thus, a small volume container

must be used –approximately 5 ml in the prototype

we have built. A container of such a small size needs

to be emptied a large number of times, which may

make the inability to measure the liquid that ﬂows

during the discharge of the container intolerable if the

patient has polyuria, i.e., the patient is producing a

large amount of urine. If the patient is producing nor-

mal amounts of urine, or if he/she has oliguria, the

amount of urine that will ﬂow during discharge of the

BIODEVICES 2011 - International Conference on Biomedical Electronics and Devices

BLUETOOTH

Figure 3: Diagram of our prototype device.

container will be nil or negligible.

It could be argued that in the case of patients that

have polyuria it is not required to have a high degree

of accuracy in the measurement of the urine output.

The more urine is produced, the less important it is

to have an accurate measure; only when the patient

is producing small amounts of urine it is important to

measure accurately (Otero et al., 2010). Therefore,

the operation of the device as is described in the pre-

vious section could be acceptable. In any case, this

problem can easily be solved using two different size

containers, each of them working according to the

principle presented in the previous section. The out-

put of the smaller container is connected to the input

of the larger container. Thus, although the volume of

liquid that ﬂows during the discharge of the ﬁrst con-

tainer is not measured in the small container, it will

be measured in the larger one.

Chaining two containers as described here causes

another problem. It can happen that when the small

container releases its content, the larger one is nearly

full. In the worst scenario, the ﬁrst drop that falls

from the small container would trigger the draining

of the large container. Thus, the content of the small

container would not be measured in the large one.

The maximum error this may cause is (V

)%,

being V

and V

the effective volumes of the small

and large container, respectively. On average, the

large container will start to drain when the small con-

tainer is emptied halfway. Therefore, the average er-

ror caused by this effect will be ((V

/2)/V

)%. As

long as V

>> V

, this error will be small.

The prototype we have built uses two containers.

The ﬁrst one was built to have a volume of approx-

imately 5 ml, and the second a volume of approxi-

mately 375 ml (see Figure 3). The ﬁrst container is

equipped with a single reed switch that is activated

when the ﬂoat is on its lowest position. The stag-

gered diameter of its output tube increases in three

steps. The output tube is connected to the input tube

of the larger container, which is equipped with two

reed switches. The tube of the large container is con-

nected to a 2.5 liters plastic bag that collects the liquid

once it has been measured.

An electronic unit continuously checks the status

of the reed switches, and reports any change via Blue-

tooth to a Java application that runs on a PC. From

these changes the application calculates the volume

of urine that has ﬂowed and displays a chart with this

information.

3.3 Device Calibration

The calculation of the effective volumes of the con-

tainers –V

and V

– must be determined experimen-

tally. By effective volume we mean the volume of liq-

uid that triggers the emptying of the container through

the siphon mechanism. This volume will be slightly

higher than the volume corresponding to the height of

the top of the siphon mechanism because the liquid

does not begin to ﬂow until the pressure overcomes

the surface tension force of the liquid against the walls

of the output tube. The smaller the diameter of the

output tube, the greater the height in the column of

liquid required to overcome the surface tension will

be.

The effective volume of each container was deter-

mined separately. A saline solution with properties

similar to those of urine and an dropper were used

to simulate the ﬂow of urine (see Figure 4). Each

of the containers were placed so that they would re-

lease their content on a bowl placed on the plate of a

high-precision industrial scale –a PGW 4502e, built

by Adam Equipment Inc. This scale has an accuracy

guaranteed by the manufacturer of 0.01 g. Given that

the density of the saline solution was known, we can

determine the volume of liquid that the container re-

leases into the bowl from the weight.

URINE OUTPUT MONITORING - A Simple and Reliable Device for Monitoring Critical Patients' Urine Output

Saline

solution

Dropper

Plasticbag

Largecontainer

Smallcontainer

Electronicunit

Figure 4: Picture of the prototype device with the saline

solution and eye dropper used in its validation.

The PGW 4502e is equipped with a serial port that

permits querying for readings. We built a program

that periodically takes measurements from the scale.

The program together with the dropper allow us to au-

tomate the process of carrying out multiple measures

of the volume of liquid released by the containers.

Using this set up we took 200 measurements of

the volume of liquid released by the small container,

and 50 measures of the volume of liquid released by

the large container. From these measures we calcu-

lated the effective volumes of the containers: 5.87 ±

0.32 ml and 376.72 ± 1.11 ml (mean ± standard de-

viation). At the time of writing, the authors are work-

ing to determine the volumes V

, V

and V

, i.e., the

volumes corresponding with the opening of the ﬁrst

reed switch of the large container, with the closing

of the second reed switch, and whit the opening of

second reed switch, respectively. This is challenging

because it requires measuring with high accuracy the

volume of liquid in the container before the liquid is

released, without interfering with the normal opera-

tion of the device. These volumes are not necessary

for the system to work correctly, but they would al-

low us to correct the measurements obtained from the

small container –which are less precise–, not just one

time but four times for each release cycle of the large

container.

3.4 Therapeutic Goals Supervision

The state of the reed switches placed in the containers

is sent from the electronic unit to a Java application

installed on a PC. From these states and from the val-

ues of V

and V

, the amount of liquid ﬂow can be

calculated. The Java application allows the health-

care staff to inspect a graph showing urine output,

and to set the therapeutic goals for the urine output.

These therapeutic goals are represented with the aid

of the Fuzzy Set Theory, a tool which has proved its

value for handling and representing medical knowl-

edge (Barro et al., 2001).

We shall introduce some basic concepts of the

Fuzzy Set Theory. Given a discourse universe R we

deﬁne a fuzzy value C by means of a possibility distri-

bution π

deﬁned over R (Zadeh, 1975). Given a pre-

cise value x ∈ R, π

(x) ∈ [0,1] represents the possibil-

ity ofC being precisely x. A fuzzy number (Kaufmann

and Gupta, 1984) is a normal (∃x ∈ R, π

(x) = 1)

and convex (∀ x, x

′

, x

′′

∈ R, x

′

∈ [x,x

′′

], π

′

) ≥

min{π

(x),π

′′

)}) fuzzy value. Normality and con-

vexity properties are satisﬁed by representing π

, for

example, by means of a trapezoidal representation. In

this way, C = (α,β,γ,δ), α ≤ β ≤ γ ≤ δ, where [β,γ]

represents the core, core(C) = {x ∈ R| π

(x) = 1},

and ]α,δ[ represents the support, supp(C) = {x ∈

R|π

(x) > 0}.

A fuzzy number C can be obtained from a ﬂexi-

ble constraint given by a possibility distribution π

which deﬁnes a mapping from R to the real interval

[0,1]. A fuzzy constraint can be induced by an item

of information such as “x has a high value”, where

“high value” will be represented by π

C=high

. Given

a precise number x ∈ R, π

C=high

(x) ∈ [0,1] represents

the possibility of C being precisely x; i.e., the degree

with which x fulﬁlls the constraint induced by “high

value”.

Physicians are accustomed to expressing the ther-

apeutic goals for urine output in milliliters of urine

produced per kilogram of patient body mass per hour

–ml/kg · h. Our tool allows them to indicate the

weight of the patient -P- and the therapeutic goals rep-

resented by the trapezoidal possibility distribution π

can be interpreted as a computational projection of

the piece of clinical knowledge “adequate UO”. The

minimum and maximum values acceptable for urine

BIODEVICES 2011 - International Conference on Biomedical Electronics and Devices

output are the beginning and the end of the support of

the distribution, respectively. If the patient produces

less urine than the amount corresponding with the be-

ginning of the support, the patient is clearly in olig-

uria. If he/she produces more urine than the amount

corresponding with the end of the support, the patient

is clearly in polyuria. The beginning and end of the

core are the limits of the interval within which ideal

values of urine output lie.

If u

is the urine output in ml/kg· h, the degree to

which the therapeutic goals established by the clin-

ician are being met is π

). If π

) = 1, the

urine output is within the range of ideal values. If

) = 0, either the urine output is less than the

minimum acceptable value -the patient has oliguria or

anuria-, or greater than the maximum acceptable –the

patient has polyuria. In both cases, the program pro-

duces an audible warning. The closer π

) is to 1,

the closer the amount of urine produced by the patient

is to the ideal value, and the closer π

) is to 0, the

closer the patient is to oliguria or polyuria.

In the tool π

) is represented by a color code

used when drawing the graph of urine output. Red

represents the null compatibility –the patient is clearly

in oliguria or in polyuria–, followed by orange, yel-

low, green and black, which represents the total com-

patibility -the urine output lies within the range of

ideal values. Therefore, the urine output graph pro-

vides instant visual feedback on the patient’s state.

The tool also generates an audible warning when

the small container is not ﬁlled within the maximum

time allowed by π

. This time is given by α · P/V

α being the beginning of the support of π

, P the pa-

tient’s weight, and V

the volume of the small con-

tainer.

4 DISCUSSION

Given that the effective volume of the small container

is 5.87 ± 0.32 ml (mean ± standard deviation), in

95% of cases the amount of urine that is really re-

leased when the container is emptied will be in the

range 5.87 ± 0.64 ml (mean ± 2·standard devia-

tion). Therefore, 95% of the measures derived from

the small container will have an error of 10.9% or

less. The purpose of this small container is to pro-

vide an early warning if the patient’s urine output is

not within the therapeutic goals: depending on the pa-

tient’s weight and state this container should be ﬁlled

within 5 and 12 minutes, if the patient is within the

limits of the therapeutic goals. If this does not occur,

the healthcare staff is alerted by an audible warning.

To obtain more accurate measures of the total

urine output, we rely on the second container. In 95%

of cases, the amount of urine released by this con-

tainer will be within the range 376.72 ± 2.22 ml. On

the other hand, on average half of the effective volume

of the small container will not be registered. Thus,

the error in the measures in 95% of the cases will be

5.87/2+2.22=5.16; 1.37% of its volume. This error is

low, especially when compared to errors committed

by the nursing staff when they take measures visually,

which has been reported to be as high as 26% (Hersch

et al., 2009).

Our device is capable of providing feedback on

the status of the patient’s kidneys more frequently

than is currently available in critical care units –

hourly. For a patient of 80 kilos, which should pro-

duce at the very least least 40 milliliters of urine per

hour, our device could warn of a deviation from the

therapeutic goals in less than 10 minutes – the time

in which the small container should be ﬁlled if the

patient produces urine within the therapeutic goals.

Therefore, it has the potential of improving patient

outcome. Given that it provides feedback on the pa-

tient’s state at shorter time intervals, it can allow the

clinician to react more promptly to complications in

the state of the patient.

The containers needed to build our device are not

more complex than the containers used in commer-

cial urine meters, which often require a small con-

tainer embedded within a large container, valves that

communicate both container with each other and with

a plastic bag, mechanisms to prevent the containers

from overﬂowing, etc. However, the price of our man-

ufacturing our device is slightly higher than the price

of the commercial urine meters because of the addi-

tion of four pieces: two ﬂoats, and two magnets. The

rest of the pieces that are part of our solution do not

have to be discarded because they are not in direct

contact with the patient’s urine, nor do they suffer sig-

niﬁcant degradation caused by its operation. Thus,

their cost can be amortized over a long period of time

and their impact on the overall cost of the solution is

negligible.

Despite its higher cost, the device has the potential

of producing signiﬁcant economic savings for the in-

stitutions that provide healthcare services. Currently,

a nurse must visit each of the patients’s beds of the

critical care unit to manually record urine output ev-

ery hour. The nurse must put on gloves since he/she is

going to manipulate body ﬂuids, walk to the patient’s

bed, take the measure visually, write it down, open

the valve that releases urine from the graduated con-

tainer to the plastic bag, wait for the urine to drain,

close the valve and check if the plastic bag needs to

be emptied. This procedure takes at least 2 minutes.

URINE OUTPUT MONITORING - A Simple and Reliable Device for Monitoring Critical Patients' Urine Output

In a small critical care unit with only 10 patients, this

means 20 minutes per hour; 8 hours a day. Typically,

nurses work in shifts of six hours. Therefore, each day

one and a third nurses’ shift are required only for tasks

related to monitoring urine output. Even a partial au-

tomation of these tasks has the potential of yielding

signiﬁcant economic savings.

5 CONCLUSIONS AND FUTURE

WORK

We have built a device capable of automatically sens-

ing and supervising the urine output of critical care

patients. The device comprises two containers of dif-

ferent volumes, a small one that receives the urine

coming from the patient’s bladder, and a greater vol-

ume container in which the ﬁrst container releases

its content when it gets full. Both containers release

their content automatically when they are ﬁlled using

a siphon mechanism.

The containers are equipped with reed switches

that are activated by a magnet that is attached to a ﬂoat

located inside the containers. These reed switches al-

low us to identify the instants at which they get ﬁlled

with urine. An electronic unit sends via Bluethooth

the information provided by the reed switches to a PC

which calculates the urine output from the switches’

state, and supervises the achievement of the thera-

peutic goals established by the clinician. The error

in measuring the patient’s urine output is under 2%.

The large container is the one which allows us to ob-

tain this high accuracy, while the small one permits

an early warning of deviations from the therapeutic

goals.

The cost of our device is slightly higher than the

cost of the commercial devices currently used in mon-

itoring urine output. However, the device has the po-

tential to save costs for the institutions that provide

health services by freeing a considerable amount of

time for the healthcare staff. Furthermore, it pro-

vides a more continuous supervision of the urine out-

put than is currently carried out in critical care units,

which may help improve patient outcomes.

As future work, we intend to take advantage of

all the state changes of the reed switches of the large

container to correct the urine output measures while

the large container is been ﬁlled. Currently, this cor-

rection is only performed when the large container re-

leases the urine. We also will start to use our device

in animal tests conducted in a research unit associated

with Getafe University Hospital. After this phase, we

intend to use it in a pilot test in the Intensive Care Unit

of this hospital.

ACKNOWLEDGEMENTS

We would like to acknowledge the support by the

Ministry of Science and Innovation of Spain, the Eu-

ropean Regional Development Fund of the European

Commission under the grant TIN2009-14372-C03-

03. T. Akinﬁev acknowledges the ﬁnancial support

received from CSIC under the project “New actua-

tors with high efﬁciency and control algorithms for

automation and robotics”. A. Apalkov acknowledges

the ﬁnancial support from Ministry of Science and In-

novation of Spain under Juan de la Cierva Program.

REFERENCES

Akinﬁev, T.; Apalkov, A.; Otero, A.; Palacios, F. Device for

measuring the amount of liquid that ﬂows and proce-

dure for its measurement. Patent Pending, 2010. ES

P201031227.

Barro, S., Mar´ın, R., Palacios, F., and Ru´ız, R. (2001).

Fuzzy logic in a patient supervision systems. Artiﬁ-

cial Intelligence in Medicine, 21:193–199.

Hande, A., Polk, T., Walker, W., and D.Bhatia (2006). Self-

powered wireless sensor networks for remote patient

monitoring in hospitals. Sensors, 6(9):1102–1117.

Hersch, M., Einav, S., and Izbicki, G. (2009). Accuracy and

ease of use of a novel electronic urine output monitor-

ing device compared with standard manual urinome-

ter in the intensive care unit. Journal of Critical Care,

24:629–633.

Ishida, S. (1990). Liquid level indicator using laser beam.

us patent 4938590.

Johnson, S. J. (1978). Liquid level measurement device.

United States Patent 3693445.

Jungk, A., Thull, B., and Rau, G. (2002). Fuzzy Logic

in Medicine, chapter Intelligent Alarms for Anaes-

thesia Monitoring Based on Fuzzy Logic Approach.

Physica-Verlag.

Kaufmann, A. and Gupta, M. (1984). Introduction to Fuzzy

Arithmetic. Van Nostrand Reinhold Company Inc.

Mitra, B., Fitzgerald, M., Cameron, P., and Cleland, H.

(2006). Fluid resuscitation in major burns. ANZ Jour-

nal of Surgery, 76:35–38.

Mora, F., Passariello, G., Carraylt, G., and Pichon, J.

(1993). Intelligent patient monitoring and manage-

ment systems: A review. IEEE Engineering in

Medicine and Biology.

Otero, A., Akinﬁev, T., Fern´andez, R., and Palacios, F.

(2009a). A device for automatic measurement of crit-

ical care patient’s urine output. In 6th IEEE Inter-

national Symposium on Intelligent Signal Processing,

pages 169–174.

Otero, A., Akinﬁev, T., Fern´andez, R., and Palacios, F.

(2010). A device for automatically measuring and su-

pervising the critical care patients urine output. Sen-

sors, 10(1):934–951.

BIODEVICES 2011 - International Conference on Biomedical Electronics and Devices

Otero, A., F´elix, P., Barro, S., and Palacios, F. (2009b). Ad-

dressing the ﬂaws of current critical alarms: a fuzzy

constraint satisfaction approach. Artiﬁcial Intelligence

in Medicine, 47(3):219–238.

Otero, A., Panigrahi, B., Palacios, F., Akinﬁev, T., and

Fern´andez, R. (2009c). Recent Advances in Biomed-

ical Engineering, chapter A prototype device to mea-

sure and supervise urine output of critical patients,

pages 321–324. Intech.

Rivers, E., Nguyen, B., e Havstad, S., Ressler, J., Muzzin,

A., Knoblich, B., Peterson, E., Tomlanovich, M., and

Group, E. G.-D. T. C. (2001). Early goal-directed ther-

apy in the treatment of severe sepsis and septic shock.

New England Journal of Medicine, 345:1368–1377.

Zadeh, L. (1975). The concept of a linguistic variable and

its application to approximate reasoning. Information

Science, 8:199–249. Part 1.

URINE OUTPUT MONITORING - A Simple and Reliable Device for Monitoring Critical Patients' Urine Output