JOURNAL OF TELECOMMUNICATIONS, VOLUME 8, ISSUE 2, MAY 2011

10
2011 JOT
http://sites.google.com/site/journaloftelecommunications/
Performance Analysis of a Nationwide
Telemedicine Call Center

Ishtiaq Al Mamoon and Shahriar Khan

Independent University, Bangladesh


Abstract- The object of study was a GSM medical call centre or service point where patients are connected over mobile
phone to doctors, who provide emergency services, medication, and drug information. The GSM call centre is equipped with
a class IV public switch as Inter exchange carrier, and a PBX. Subscribers can dial 789 from anywhere in Bangladesh, with
a phone under the operator. The call hit rate is indicative of patient response, and generally showed an upward trend during
its existence. The survey conducted among responding agents showed the strengths, and areas requiring improvement of
this system. A significant number of callers were assisted during medical emergencies. At this time, the call patterns indicate
that this project is likely to be sustainable, and that medical tele-conferencing over cellular phone has a potentially large
market.

Index Terms- Telemedicine, GSM, Call pattern, Bangladesh, Call center

1 INTRODUCTION
A major mobile phone operator in Bangladesh and
Telemedicine Reference Centre Ltd (TRCL) have jointly
started a mobile communication based on medical
teleconferencing project. This is a GSM-based medical
call centre where patients are treated by doctors over their
mobile phones. The Short Message Service is used for
transmitting medical, diagnosis, and lab reports to the
patient. This paper briefly discusses the incoming call
properties, call ratio, per hour patterns, and other issues
from TELES.iSWITCH and PBX generated Call Detail
Record, CDR [1].

2. CDR FROM TELES AND PBX REPORT
The CDR contains call index, caller number (OID), call
duration, and call setup time according to ISUP, port &
channel info, connection status and CAU (call summary).
These data provide indications for financial, human
resources and strategic issues [1] [2]. The project under
study has been successfully operational from November
4, 2006. For testing purposes, the Telecom (Technical)
Dept of TRCL analyzed the weekly call record data from
Nov 4th to 10th Nov 2006 and generated the following
reports [1] [3]:

A. TELES.iSWITCH report from CDR
The numbers of calls as well as the total duration is
important for financial reasons, and project planning. The
total number of call and time can be generated from the
Call Detail Record (CDR).

The first 20 sec was free, for calls to 789. Charging starts
after 20 Sec. The calls that try to access the 789 from GP
phone are considered as hits, and among them, the call of
duration of more than 20 sec are treated as served call. A
served call means the call was received by agents [4].




Table 1. Call Congestion Status














JOURNAL OF TELECOMMUNICATIONS, VOLUME 8, ISSUE 2, MAY 2011
11
2011 JOT
http://sites.google.com/site/journaloftelecommunications/













Fig. 1 Hits vs. Answered Calls


Analyzing the above table and graph, it is clear that the
answered call ratio in terms of hits in TELES.iSWITCH
is very low. This is because in teles switch the inlet
capacity is 60 channels and outlet capacity was only 15.
This means that the call center capacity was limited.
When the calls tried to terminate to the PBX end, a
bottleneck was created, and most calls failed to terminate.
It was also observed that the number of hit calls, and the
served calls increased proportionally [1][5].

The PBX announcement time and system time were
configured such that, in the period of 20 seconds, 16 sec
was allocated for announcement and 3 sec or 1 cycle was
used by the PBX system. In this 20 sec, users can avail
the facility of DTMF.

In the above table, it was found that the average call
congestion was 26 percent, meaning that 74 % of calls
were not served properly.





























Table 2. Call Numbers Generated in 24 hours (source:
Primary data).


A call pattern from the CDR was generated during the
test trail period. Table 2, which was derived from the call
pattern findings, represents the number of calls per hour
in seven days. A decision was taken by strategic
management that to improve call congestion rate, the first
20 seconds will not be free. After 19th Nov to till now the
congestion rate was above 75%. [5].

Analyzing the following graph, it was seen that the peak
time or busy hour call attempt (BHCA) was 14:00-16:00
and 19:30- 22:00 hours. The call rate remained the same
from 14:00 to 23:00 hours. From 01:00 to 08:00 hours, it
was very low.

0
5 00 0
10 00 0
15 00 0
2 0 00 0
2 5 00 0
3 0 00 0
3 5 00 0
4 0 00 0
No. of Calls Hit
to Call Centre
No. of Calls
(Duration more
than 20s) Served
JOURNAL OF TELECOMMUNICATIONS, VOLUME 8, ISSUE 2, MAY 2011
12
2011 JOT
http://sites.google.com/site/journaloftelecommunications/

Fig. 2. Call Pattern of HealthLine Project

This graph indicates how human resources should be
planned for proper utilization. The above pattern could be
different depending on natural disaster, political
imbalance or any other social issues.

B. PBX provided Call pattern:
The following graph shows the average received call
from 4th to 10th November. From the graph it was found
that the Busy hour was 17:00 to 19:00. It was also found
that the call rate was increasing from 16:00 hours.



Fig. 3 Received call pattern form PBX report server

C. PBX channel congestion:
Only 4 channels (Voice channel) were opened in DSS1
line to check the call congestion. Five GP numbers were
used to check the call congestions. Five cell calls were
attempted at the same time.

The following table was extracted from PBX server
(ACD). To derive this table it was found that the average
answered call number was 47.14%. The total average
unanswered time was 81.71% [1] [4] [5] [6].







Table: 3 call congestion data from PBX



D. CDR Indications
CDR can indicate different issues about call connection
and termination process. One can find CAU values from
D Line in CDR. Some effects are described in the
following [7].
CAU_NCC means Normal Call Clearing in ISUP
protocol. That means the call was connected, set up and
terminated properly. It was found that during the 7 days
of trial period, 87.67% of call volume value was
CAU_NCC.
CAU_NUNSP indicates which call is normal but
unspecified. It may show call volume but cant trace the
channel no. or inlet or outlet port no. It was found that
during the test trial 11.09% of call volume call volume
was CAU_NUNSP.
CAU_SEQC means switching equipment congestion, or
call traffic blocking where the calls are torn. Only 1.24 %
of total call volume value was found as CAU_SEQC [1]
[3] [5] [7] [8].

3. FINDINGS OF SURVEY
Survey questionnaires were filled by a total of 10 agents.
The findings of the survey are summarized below:
HealthLine service information:
90% of the agents are medical personnel.
On an average 200 calls received by the agents per
day.
The average duration of consultation is 2-3 mines/
call.
Caller information (according attended call)
Average age of the patients is 25-32.
Initially 1-2% of caller registered for HealthLine
service
Daily more less 80-90 members (registered) seek for
medical information and service by dialing 789
30% of the member calls twice a day for the same
specific reason.
For emergency medication (press 5) purposes 70% of
total caller call in 789, among them 25%-30%
problem are female or female related issues.
The peak hour is 11:00-13:00 at day time and 21:00
to 23:30 for both members and non members (only
consider Press 5).
Average 5% to 8% of callers seek drug information
service.
At least 10 to 12 Subscribers Seeking help for
understanding Lab report.
JOURNAL OF TELECOMMUNICATIONS, VOLUME 8, ISSUE 2, MAY 2011
13
2011 JOT
http://sites.google.com/site/journaloftelecommunications/
Everyday, numerous calls are made, which have
nothing to do with medical issues. Most callers are
wrongly informed about the project. A few callers
with psychiatric problems call the 789 system.
GP cell holder village doctors or village quack are
practicing Telemedicine and Tele-consultation based
medical treatment with their patient. In that case they
consult and share different health & medical follow-
up related issues with HealthLine agents.

According to the Technical survey:
The existing system is not user friendly to doctors;
they expected a more equipped call centre.
80% of doctors think that the company needs to
upgrade the Telecommunication system such as the
Computer Telephony Interrogation (CTI) system
because then they are more flexible with their work.
Most of the patients are satisfied with the
consultation with doctors [1,5, 6].

4 FUTURE WORK
A future direction of this work would be to have doctors
and patients video conference through mobile.
International inbound medical call center can also be set
up by using VoIP technology. Here overseas calls are
received from abroad in IP format, and then converted
into ISDN/SS7/RS323 to route the call. Nowadays,
foreign patients can consult with local doctors at low cost.
Reverse facilities will also be possible through this
technology [1].

5. CONCLUSIONS
From the technical issues and the Call patterns of the
switch and PBX, it is found that the dial 789 system
works up to expectations. The call pattern from
TELES.iSWITCH and PBX are similar. The switch can
provide the list of HIT data and PBX provides the
attended calls. There is also some inconsistency in the
pick up time, possibly because Bangladeshi mobile users
tend to give missed calls. The upward trend of this
service among its users is clearly indicated. The survey
conducted among inbound center doctors showed the
strengths and areas of improvement for this system.

REFERENCES
[1] I. A Mamoon, Implementation of Telemedicine in
Bangladesh, through DAB, GTC Telemedicine Pilot
Project and GP-TRCL Healthline Project as case studies.,
MSc Thesis, Independent University, Bangladesh, 2007.
[2] I A Mamoon, S Khan, "Implementation of Telemedicine in
Bangladesh, with Grameen Phone HealthLine 789 Project
as Case Study," A-098 P-184, Proceedings of the National
Conference on Electronics, Information and
Telecommunication, Rajshahi University, 29-30 June, 2007
[3] I A Mamoon, S Khan, "An Intial Architecture for GSM
Medical Call Center 789," 605-036, Proceedings of
IASTED International Conference on Advanced in
Computer Science and Technology (ACST 2008) in
Langkawi, Malaysia, April 02,2008.
[4] I A Mamoon, S Khan, An implementation of TDM Based
GSM Medical Call Center HealthLine 789 C-033,
International Conference on Electronics, Computer and
Communication(ICECC 2008) . 27-29 June, 2008 Rajshahi
University.
[5] Unpublished service level data of HealthLine Project.
[6] Different type of call detail record (CDR) from
TELES.iSWITCH and Siemens HiPath 3750.
[8] I A. Mamoon, Shahriar Khan, "Telemedicine Services at
the Diabetic Association of Bangladesh with Technical
Collaboration of Grameen Telecom," Proceedings of the
Workshop on Prospects and Problems of Mobile and Land
Phones in Bangladesh, Independent University,
Bangladesh, 8 December, 2005.
[9] D.L. Paul, K.E. Pearlson, R.R. McDaniel Jr., "Assessing
technological barriers to telemedicine: technology-
management implications," Engineering Management,
IEEE Transactions on, vol. 46 , iss. 3, pp. 279 - 288, Aug.
1999.
[10] A. Z. Malik, "Telemedicine Country Report-Pakistan," e-
Health Networking, Application and Services, 2007 9th
International Conference on, pp. 90 - 94, Taipei, Taiwan,
19-22 June 2007
[11] A. El Matri, Telemedicine in Tunisia: links with European,
Arab and African countries, Information and
Communication Technologies: From Theory to
Applications, 2008. ICTTA 2008. 3rd International
Conference on, pp. 1 - 1, Location: Damascus, 7-11 April
2008.
[12] M. Kifle, V. W. A. Mbarika, C. Tsuma, D. Wilkerson, J.
Tan, "A TeleMedicine Transfer Model for Sub-Saharan
Africa," Hawaii International Conference on System
Sciences, Proceedings of the 41st Annual, pp. 244 - 244,
Waikoloa, Hawaii, 7-10 Jan. 2008.
[13] A-M. Croteau, D. Vieru, "Telemedicine adoption by
different groups of physicians," System Sciences, 2002.
HICSS. Proceedings of the 35th Annual Hawaii
International Conference on, pp. 1985 - 1993, 7-10 Jan
2002
[14] R. Komiya, "A proposal for telemedicine reference model
for future standardization," Enterprise networking and
Computing in Healthcare Industry, 2005. HEALTHCOM
2005. Proceedings of 7th International Workshop on, pp.
224 - 228, 23-25 June 2005
[15] M. T. Mijares, "Telemedicine in Ecuador: failure or a
learning experience?," Enterprise Networking and
Computing in Healthcare Industry, 2004. HEALTHCOM
2004. Proceedings. 6th International Workshop on, pp. 41 -
43, 28-29 June 2004
[16] A. Pal, V.W.A. Mbarika, F.Cobb-Payton, P. Datta,
S.McCoy, "Telemedicine diffusion in a developing
Country: The case of India (March 2004)," Information
Technology in Biomedicine, IEEE Transactions on, vol. 9 ,
iss. 1, pp. 59 - 65, March 2005
[17] J.E. Cabral Jr., K. Yongmin, "Multimedia systems for
telemedicine and their communications requirements,"
Communications Magazine, IEEE, vol. 34 , iss. 7, pp. 20 -
27, July 1996.
[18] C. LeRouge, M. J. Garfield, A. R., Hevner, "Quality
attributes in telemedicine video conferencing, System
Sciences, 2002. HICSS. Proceedings of the 35th Annual
Hawaii International Conference on, pp. 2050 - 2059, 7-10
Jan 2002
[19] Yajiong Xue, Huigang Liang, "Analysis of Telemedicine
Diffusion: The Case of China," Information Technology in
Biomedicine, IEEE Transactions on, vol. 11 , iss. 2, pp.
231 - 233, March 2007.