MOBILE TIMELINE

Mobile Charting System that Provides

a Graphical Summary of a Patient’s Medical Record

Keisuke Ogawa

, Kazunori Matsumoto

, Masayuki Hashimoto

, Tatsuaki Hamai

Akiko Shibuya

2,3

and Yoshiaki Kondo

2,3

KDDI R&D Labs Inc., 2-1-15, Ohara, Fujimino-shi, Saitama, Japan

Nihon University School of Medicine, 30-1, Oyaguchi Kami-cho, Itabashi-ku, Tokyo, Japan

Tohoku University Graduate School of Medicine, 1-1, Seiryo-cho, Aoba-ku, Sendai-shi, Miyagi, Japan

Keywords: TimeLine, XML, Mobile, Electronic medical records, Event merge, Medical support system, Physician.

Abstract: In this paper, we introduce a novel medical information system that works on smartphones. This system

provides chronological graphs of a patient’s medications and medical examinations over an extended period

of years as graphical summaries of the patient’s medical history so that physicians can gain a clear

understanding of the patient’s status and develop treatment plans easily. In our system, we implemented the

original algorithm that reduces the amount of medical data by merging adjacent data hierarchically when the

time span for displaying the data is changed. By implementing this algorithm, the system proposed here was

shown to be about 30 times faster than the system with a conventional method. Also, we evaluated our

system through the experiment in medicine using real medical records. The results indicate that physicians

would benefit significantly by using our system especially in situations where they cannot use another

medical information system through their PC, such as the patient’s bedside.

1 INTRODUCTION

In recent years, the use of mobile phones as a client

device in medical information systems that display a

patient’s medical data has become increasingly

popular (G.Benelli et al. 2010). Implementation of

such systems promises significant advances in

patient care because of their mobility and flexibility.

This is because today’s mobile devices, such as

smartphones, now have the capacity for a variety of

medical applications and can be used from any

location (M. Watson 2006, J. Sammon et al., 2006,

A. Kumar et al., 2009). For example, a system that

allows physicians to see and share the electronic

health records of their patients (G. Benelli et al.,

2010, F. Andry et al., 2011) has been proposed.

However, in order to diagnose a patient with a

chronic disease such as diabetes, physicians often

need not only recent medical data from the patient’s

medical record but also a chronological graph of the

medications and medical examinations that the

patient may have undergone over an extended period

of years as a graphical summary of the patient’s

medical history. Since their overall treatment time is

very long, physicians often use these graphical

summaries to analyze and ascertain a patient’s status

and develop a treatment plan. While mobile devices

are very flexible, it is extremely difficult for them to

display a graphical summary because they do not

have sufficient computing and processing capacities.

In view of these problems, we propose a brand new

mobile medical information system that overcomes

these shortcomings (K. Ogawa et al., 2011). In this

paper, we introduce actual implementations of the

system and results of a small-scale experiment

conducted at Tohoku University Hospital.

2 CONVENTIONAL MEDICAL

INFORMATION SYSTEMS

In this section we give an overview of an EMR

system and an electronic patient chart as examples

of conventional systems used in medicine.

Ogawa K., Matsumoto K., Hashimoto M., Hamai T., Shibuya A. and Kondo Y..

MOBILE TIMELINE - Mobile Charting System that Provides a Graphical Summary of a Patient’s Medical Record.

DOI: 10.5220/0003708300230029

In Proceedings of the International Conference on Health Informatics (HEALTHINF-2012), pages 23-29

ISBN: 978-989-8425-88-1

 2012 SCITEPRESS (Science and Technology Publications, Lda.)

2.1 Example of EMR System

Fig.1 shows an example of a conventional EMR

system. This system has a user interface similar to

legacy paper medical records and displays the

patient’s medical information for one day.

Figure 1: EMR system.

Such systems that present first-hand medical data

obtained from a patient’s medical record are also

implemented on smartphones (F. Andry et al., 2011).

However, physicians often need to analyze a

patient’s medical history over several years (H.

Smith et al., 2006, G. L. Warnock et al., 2008). For

example, to determine a treatment plan for diabetes,

physicians must analyze medical data that have been

accumulated over several years because the disease

often progresses over a prolonged period. Therefore,

the problem with this system is that it displays the

medical data for only one day whereas the medical

data physicians need to perform a proper analysis

varies according to the patient and the disease.

2.2 Example of Electronic Patient Chart

Fig. 2 shows an example of an electronic patient

chart system. This system allows physicians to

analyze the medical information such as a patient’s

body temperature and blood pressure for a fixed

period.

However, different kinds of medical data are

generated over different intervals. For example, if a

physician needs to analyze and determine the

medications for a patient with diabetes who has the

flu, the doctor must examine the changes in the

patient over a long time span and investigate any

contraindications to a possible drug therapy over a

short time span. So the problem with this system is

that physicians cannot analyze medical data and

change the time span freely.

Figure 2: Electronic patient chart.

2.3 Limitations of Conventional

Systems

To summarize, conventional systems are unable to

sufficiently support the medical analysis that

physicians need to conduct. This is because these

systems do not have the ability to display the

patient’s chart for various diseases that occur over a

range of time spans.

3 PROPOSED TIMELINE

INTERFACE

As described in section 2, conventional medical

systems have a problem in that they are unable to

display the patient’s medical chart over a range of

time spans.

Furthermore, there is an increasing need

to be able to see patient medical data in locations not

restricted to those where they can access a

conventional medical system via a PC (M.Watson

2006, J.Smith et al. 2006, A.Kumar et al. 2009).

3.1 System Overview

In view of these problems, we propose a brand new

system that consists of a server and an application on

a mobile device. In this section, we introduce the

system implementation that we developed for use in

actual clinical practice. In our system, the server

transmits the medical data from a conventional EMR

system’s server and optimizes it. An application on a

mobile device displays the patient’s medical chart

with the TimeLine interface.

3.2 TimeLine Interface

The TimeLine interface has a function that displays

a chronological chart of the patient’s medical data

HEALTHINF 2012 - International Conference on Health Informatics

over various time spans. By means of this function,

medical data can be divided into categories called

medical items. Medical items are, for example, types

of medication or types of medical tests and so on.

Also, under each medical item, there are numerous

medical events that represent the medical services

the patient has received as shown in Fig.3.

Figure 3: TimeLine interface.

Also, TimeLine has a multistage time scale of years,

months, days, hours and so on. With the TimeLine

interface, physicians can change the time scale and

its length to display with a single easy operation

such as pinch-in/pinch-out operations on a screen of

a smartphone. For example, the time scale can be

changed from hours to days or to months. By

changing to a short time scale, physicians can

analyze medical data over a short period in more

detail. Conversely, by changing to a longer time

scale, physicians can analyze medical records over a

long period. Fig.4 shows the operation of time scales.

Figure 4: Time scale in TimeLine interface.

In this manner, physicians can analyze a range of a

patient’s medical history.

3.3 Adaptive Event Merge Algorithm

However, TimeLine requires the processing of a

large volume of data and high processing ability to

draw the charts because physicians tend to record a

large number of medical events. Take the case of a

patient with a chronic disease; medical events are

generated over a period of time measured in decades.

It takes a lot of time to process such large volumes

of data; consequently, the response speed of our

system goes down. So this is the most serious

problem of TimeLine interface. The best way to

solve this problem is to reduce the volume of data to

process. In order to do this, we developed the

adaptive event merge algorithm in TimeLine. The

purpose of this algorithm is data reduction. Because

of the resolution limit, if the user expands the time

span to be shown, all the data plots are not displayed

on the screen. Making use of this feature, the system

can reduce the volume of data to process by

reducing the data for each time scale. In other words,

if the system can process only the visible data plots

for display on a chart over a long time span, such as

the time scale for a year or ten years, we can reduce

the workload on the system. Now, the medical

events that our system processes tend to consist of

the data from medical inspections or medications

generated once a day and once every month in

average. So, the visible number of total data plots

tends to vary for a time scale of a year or longer.

Figure 5: Typical medical events in our system.

So, we created several DB tables for various

expansion rates across longer time spans, as well as

the minimum number of data tables for shorter time

spans. By using these DB tables, the system can

process only visible data for each time span. In fact,

MOBILE TIMELINE - Mobile Charting System that Provides a Graphical Summary of a Patient's Medical Record

we created the 20 data tables below:

Figure 6: DB tables and expansion rates.

The lower numbered tables store visible data for

shorter time spans (i.e. smaller expansion rates), on

opposite, while the higher numbered tables store

visible data for larger time spans. Also, the 3

column means the expansion rates from lower

numbered tables to higher numbered tables. As can

be seen, these DB tables are designed so that the

system can process the smaller number of data when

the system represents the chart of larger time spans

(i.e. the time spans that the number of visible data

tends to change).

Now, we briefly describe the steps of this algorithm

below:

1. The algorithm represents a medical event as

structured data that includes the start time, end time,

the value, and other factors.

2. For each expansion rate, the system calculates the

distance to adjacent data objects using the formula

below.

In the formula, we use for the “nth” data object of

a patient, for the start time, for the end time, C

for a coefficient, and E for the expansion rate

between each DB table.

Formula:

3. If the distance is smaller than the threshold, the

system generates a new data object that represents

the original two data objects. This threshold

represents the distance that a person can visually

recognize. In general, this threshold is one dot on the

screen. Fig. 7 shows an example of the merging of

two medical events in XML format.

Figure 7: Event Merge in XML.

4. The system registers the data in each of the DB

tables.

5. The system generates the chart using new data

objects. Fig.8 shows how it looks like in the

TimeLine interface.

Figure 8: Adaptive Event Merge in TimeLine.

As can be seen, if the time scale changes, several

events merge adaptively. In addition, to reduce the

time required to draw the charts, we had the server

draw the chart and transmit the image to the client

device. In this manner, the system reduces data

processing and improves responsiveness.

tets

⋅

−

HEALTHINF 2012 - International Conference on Health Informatics

4 EXPERIMENTAL RESULT

AND EVALUATION

To evaluate our system, we held a small-scale

experiment in medicine. So in this section, we

present the evaluation results obtained by physicians.

4.1 Experimental Conditions

First, the experimental conditions are introduced.

4.1.1 Evaluators

Fourteen physicians from six departments were

chosen as evaluators (Table.1).

Table 1: Evaluators.

Department Number of ph

sicians

Pediatrics 7

Circulator

Medicine 1

General Internal Medicine 1

Respirator

Medicine 3

Fixed Prosthodontics 1

Dentistry 1

Total 14

A smartphone and PC was used as the system’s

client device and the system’s server, respectively.

The specifications are presented in Table.2.

Table 2: Specifications of the device.

Client Mobile

Device

Server PC

OS Androi

Windows 7 32-

CPU

Snapdragon 1

GHz

Intel Core i7-

620UM 1.06 GHz

Memory 256 MB 4 GB

4.1.2 Form of the Questionnaire

All physicians answered a questionnaire designed to

evaluate the system. For each question, there were

five choices. A score of 5 represents the most

positive answer. Conversely, a score of 1 was the

most negative answer. Fig. 9 shows an example of

the questionnaire we used.

Figure 9: Example of the questionnaire.

4.1.3 Evaluation Items

The main evaluation items are listed below:

1. The situation in which physicians used the

system:

We asked where the physician used the system for

analyzing inpatients’ and outpatients’ medical data.

This result corresponds to the TimeLine evaluation

in our system.

2. The response speed of the system:

The response speed of the system is an important

factor when determining the system’s usability.

This result corresponds to the adaptive event merge

algorithm evaluation in our system.

4.2 Results

The experimental results for each evaluation item

are presented in this section.

4.2.1 Location of Use

Fig.10 shows the evaluation result for the location

where physicians analyzed outpatients’ medical data.

Each horizontal bar represents the evaluation scores

for each answer listed on the left side. Also, average

evaluation scores for each answer are shown on the

right side of the bar. (Fourteen physicians, including

a clinician from outside the hospital participated in

the survey, so the answers “other rooms” and “ward”

have 13 responses.)

Question: Do you want or need to use the system to

analyze outpatients’ medical data in these places?

Figure 10: Evaluation result for outpatients.

Fig.11 shows the evaluation result for the places

where inpatients’ medical data are analyzed.

Question: Do you want or need to use the system to

analyze inpatients’ medical data in these places?

MOBILE TIMELINE - Mobile Charting System that Provides a Graphical Summary of a Patient's Medical Record

Figure 11: Evaluation result for inpatients.

As can be seen, there is a certain need to use our

system anywhere. In particular, there is a remarkable

need to use it at other hospitals, at the bedside of

patients, and in intensive care units. These are the

places where physicians cannot use conventional

medical systems in spite of the fact that they need to

refer to the patient’s medical data.

Also, these physicians can be divided into two

groups according to their department.

・Category1: Physicians who have to analyze the

medical data over a long time span (e.g., pediatrics).

・Category2: Physicians who have to analyze the a

patient’s recent medical data for short time span

(e.g., respiratory medicine).

According to these categories, evaluation results

reveal weaknesses for the situations below:

・Situation 1 (usual working situation):

ward, examination room, other rooms, patient’s

bedside, intensive care unit, treatment room, nursing

station

・Situation 2 (non-standard working situation):

working in other hospitals, academic meetings, at

home

Table 3 shows the average evaluation score for each

category.

Table 3: Average score for each category.

Situation1 Situation2

Category1

4.41 3.91

Category2

3.67 4.73

These results indicate that physicians in category 1

have a greater need to use our system in usual rather

than non-standard working situations. Conversely,

physicians in category 2 have a greater need to use

our system in non-standard working situations.

Physicians in Category 1 often need to analyze

medical information covering a long time span in

order to make a medical diagnosis. For example, the

reference values differ depending on the patient’s

age, particularly children aged less than 20 years.

Thus, pediatricians often need to analyze and

compare medical history information when creating

a medical treatment plan especially at a place where

they see their patients without access to a

conventional medical system, such as at the bedside.

On the other hand, physicians in Category 2 often

need to analyze medical information covering a

short time span in an emergency. For example, if the

condition of a patient suddenly changes, they have to

make their decisions by accessing the

contraindications for newly released medications

especially where they cannot see the patient’s

medical history with a conventional medical system,

such as outside the hospital.

4.2.2 System Response Time

In order to evaluate the performance of our system,

in addition to the physician evaluations, in another

experiment, we measured and compared the

response time of our system with adaptive event

merge algorithm and the system without it. In the

experiment, the system randomly chooses the time

span and draws a graph using medical records for

the selected time span. Data used in the experiment

are presented in Table.4.

Table 4: Number of medical events for typical patients

case (Experimental data).

Patien

umber of medical events

Diabetic 2374

ephrotic syndrome 7439

Dwarfis

8935

For each set of medical data, we repeated the

experiment 54 times and measured the average time

for drawing graphs using medical events from each

patient’s medical records. Table.5 shows the result.

Table 5: Averaged response time of proposal method

(system with adaptive event merge algorithm) and

conventional one (system without adaptive event merge

algorithm).

Patient Proposed Conventional

Diabetic 75.3(ms) 1945.98(ms)

ephrotic

ndrome

82.3(ms) 2966.9(ms)

Dwarfism 116.9(ms) 3216.8(ms)

HEALTHINF 2012 - International Conference on Health Informatics

The result indicates that our proposed method was

about 30 times as fast as the conventional method.

Additionally, we conducted a subjective evaluation

of our system’s responsiveness. Table 6 shows the

number of physicians who chose each score in

response to the question “What was your impression

of the responsiveness of our system?”

Table 6: Physicians’ evaluation for responsiveness.

Score Number of physicians

5 (excellent) 3

4 (good) 6

3 (acceptable) 4

2 (not good) 0

1 (poor) 1

As can be seen, at least 9 of 14 physicians evaluated

the responsiveness of our system as acceptable. Thus

it can be asserted that the responsiveness of our

system is sufficient for clinical use.

5 CONCLUSIONS

In this paper, we presented a new medical system.

The system has the three features listed below:

1. The client application of our system works on

mobile devices like a smartphone and can be used

anywhere.

2. The client application has a timeline interface

that visually displays the medical records of the

patient.

3. Via the adaptive event merge algorithm, the

client application responds quickly.

Due to these features, our system satisfies the

physician’s need to be able to make medical

diagnoses regardless of where they are. Through the

experiment that compared our system to the

conventional method, we showed that our system

using the adaptive event merge algorithm enables a

response at least 30 times as fast as the conventional

system. Also by conducting a qualitative evaluation,

we showed that the performance of our system is

acceptable for clinical use. Through the experiment

and the analysis of the result, we showed that there

are different usage patterns according to the

specialty of the physicians.

REFERENCES

M. Watson, 2006, Mobile Healthcare Applications: A

Study of Access Control Proceedings of the 2006

International Conference on Privacy, Security and

Trust.

J. Sammon et al., 2006, Enabling Communications for

Mobile Workers within Healthcare Environments:

MobileHCI '06, Proceedings of the 8th conference

Human-computer interaction with mobile devices

and services

A. Kumar et al., 2009, ELMR: Lightweight Mobile Health

Records: A Study of Access Control SIGMOD '09

Proceedings of the 35th SIGMOD international

conference on Management of data.

F. Andry, L. Wan. and D. Nicholson, 2011, A Mobile

Application Accessing Patients’ Health Records

Through A Rest Api:4th International Joint

Conference on Biomedical Engineering Systems and

Technologies.

K. Ogawa, K. Matsumoto, M. Hashimoto, T. Hamai, Y.

Kondo, 2011, Mobile TimeLine EMR System: 4

International Joint Conference on Biomedical

Engineering Systems and Technologies.

H. Smith et al., 2006, The Impact of Prescribing Safety

Alerts for Elderly Persons in an Electronic Medical

Record: Arch Intern Med/Vol 166

M. DesRoches et al, 2008, Electronic Health Records in

Ambulatory Care - A National Survey of Physicians:

The new England journal of medicine

G. L. Warnock et al., 2008, A Multi-Year Analysis of

Islet Transplantation Compared With Intensive

Medical Therapy on Progression of Complications in

Type 1 Diabetes: Clinical Transplantation Volume 86

- Issue 12 - pp 1762-1766

G. Benelli and A. Pozzebon, 2010, Turning a mobile

phone into an interactive multipurpose assistant in

healthcare scenarios: Biomedical Engineering Systems

and Technologies: International Joint Conference,

volume 52 of Communications in Computer and

Information Science, pages 356–368.

MOBILE TIMELINE - Mobile Charting System that Provides a Graphical Summary of a Patient's Medical Record