Process Improvement in an Oncology Day Hospital: Design of a

Traceability and Indicator System

Eduardo Cañada

, Enrique Casado

and Ana M. Bernardos

Hospital Universitario Infanta Sofía, San Sebastián de los Reyes, Madrid, Spain

Universidad Politécnica de Madrid, Madrid, Spain

Keywords: Process Improvement, Traceability, Oncology Day Hospital, Information System.

Abstract: Day hospitals (DH) are organizational structures that enable the supervision of patients who must go through

diagnosis methods or treatments taking several hours, but not requiring inpatient hospitalization. Oncology

Day Hospitals (ODH) are a particularly complex subset of DH, due to the variety and type of pathologies that

are treated in them, the characteristics of cytostatic drugs, the involvement of different hospital units and

professional profiles, the number of stages of the care procedure and the cost. In this paper, we describe the

design of a traceability and indicator system for ODH, which aims at improving the performance and quality

of service, providing three-folded benefits for patients, practitioners and hospital managers. The system is

currently being tested in a public hospital in the Autonomous Community of Madrid. Their users perceive

that they have access to a much more accurate fingerprint of everyday workflow, thus facilitating the design

of improvement actions.

1 INTRODUCTION

Since their introduction in the 70s, Oncology Day

Hospitals (ODH) have played a key role in the

treatment of cancer: it is in this organizational

structure where the oncology patients receive their

chemotherapy treatment, with specifically tailored

drugs. ODH are target of continuous improvement

measurements: the disease impact in patients’ quality

of life, the nature of cytostatic drugs, the usually long

visit protocols, the number of professionals involved

in the treatment workflow and the cost of the attention

require that ODH operations are revisited and

optimized to guarantee safety, efficiency and quality

of service. For example, in the last years, a particular

effort has been done to deploy traceability systems for

cytostatic drugs over all the clinical workflow, to

provide maximum protection in drug administration.

The works of Queralt et al. (2015), Kergosien et al.

(2011) or Sini et al. (2011) focus on this issue, while

the preparation and manipulation of the drugs

themselves is also a matter of interest (Masini et al.,

2014). From a more holistic approach, some other

experiences, such as the one described by Galligioni

et al. (2009), examine the hindrances and benefits of

the use of specific tools to manage electronic

oncological patient records.

This paper describes the process improvement

analysis that has been carried out to have an accurate

fingerprint of the activity and performance of the

ODH at Hospital Universitario Infanta Sofía. Hospital

Universitario Infanta Sofía is a public hospital in the

Autonomous Community of Madrid (Spain), active

from 2008. In 2015, 35515 care sessions where

handled in its Day Hospital (SIAE, 2015); 7085 of

those were oncology sessions (approximately 20%).

As a result of the procedural analysis, a

Traceability and Indicators System (TIS) for the ODH

has been designed; it is composed by a real-time

visualization interface and a business intelligence tool

(dashboard). On one hand, the visualization interface

retrieves real-time timestamps at the different stages

of the ODH operation workflow, so it can provide

real-time data and alerts for health workers and

managers, at the same time that facilitates the

integration of information services for patients. On

the other hand, the business intelligence tool enables

the retrieval of a complete set of activity,

performance, quality of care and procedure indicators

that aims at providing information to design

continuous improvement strategies.

CaÃ

sada E., Casado E. and Bernardos A.

Process Improvement in an Oncology Day Hospital: Design of a Traceability and Indicator System.

DOI: 10.5220/0006278506010608

Figure 1: ODH workflow for a patient.

The TIS has been built on already available

information that was not being retrieved in an easy-

to-visualize way. The system aims at delivering three-

folded benefits: for the patient, to reduce and lighten

the time at the ODH; for the clinician, to facilitate the

retrieval of real-time information about the workload

and improved agenda management thanks to the

existence of accurate and reliable indicators; for the

manager, to facilitate the design of actions to raise of

the perceived quality of service and to handle

bottlenecks and process problems, so the impact of

the investments agenda can be better evaluated.

In this context, the paper is structured as follows.

Section II describes the operation workflow at the

ODH and the subsequent TIS’ functional and non-

functional requirements. Section III describes the

deployed architecture and the customized real-time

visualization interface. Section IV describes the

indicators. Finally, Section V concludes the work,

stating the impressions gathered about the system and

defining further steps.

2 OPERATIONAL CONTEXT

2.1 Workflow Analysis

A standard visit of a patient to the ODH is as follows:

on arrival, the patient goes directly to take the 1)

preliminary blood tests. After that, they head to the 2)

medical consultation, where the practitioner

examines the overall situation and prescribes the

cytostatic drugs. The cytostatic drugs are then 3)

prepared at the Pharmacy Service and sent to the

ODH for dispensation. When the cytostatic drugs are

prepared and delivered at ODH and the needed

resources (seat, bed, pump, etc.) are ready, the 4)

patient is admitted to the ODH for the 5) drug

administration. When the treatment is over, the

administrative staff in charge proceeds to the 6)

patient discharge, who leaves the hospital. Figure 1

summarizes the full workflow.

In practice, during all the process, it is needed to

do a follow up of the patients, personnel, resources

and drugs. This tracking involves four Hospital Units:

ODH, Laboratory, External Consultations and

Pharmacy. The professional profiles participating in

the workflow are oncologist (FAC), pharmacologists

(PHAR), health staff (NUR), administrative staff

(ADM) and ancillaries (ANC).

As our objective is to build a traceability and

indicators system (TIS) that may report a real-time

picture of the ODH operation and a full view of the

service performance, it is important to analyze which

milestones may be automatically retrieved through

the APIs provided by the commercial information

systems that are already deployed in the Hospital. In

the particular case of Hospital Universitario Infanta

Sofía, these systems include the Hospital Information

System (Selene from Siemens), the Pharmacy

Information System (FarmaTools from Dominion)

and the Laboratory Information System (LIS-

ServoLab from Siemens). Apart from those, an

Appointment and Queue Tracking System for

consulting rooms (AQ-Quenda from Plexus

Technologies) is also in use.

Table 1 summarizes all the events in the operation

workflow. As the reader will notice, there are events

that are not being registered (e.g. when the patient

enters the hospital or the NUR takes blood), and some

others that are still being registered in manual way,

thus their timestamps cannot be directly integrated

into the TIS.

2.2 System Requirements

On this workflow analysis, the functional

requirements for the TIS have been defined (i.e. what

the system has to do in practice):

 Show the status (situation in the encounter

workflow and particularities) of all the

scheduled patients, both for regular visits and

ODH.

 Include the non-scheduled patients that may

appear during the day.

 Execute the automatic capture of information

from all the hospital IS that provide

information about the patient whereabouts and

health records.

 Facilitate the input of relevant information in

manual mode, in case it is not recorded at the

HIS.

 Provide user management, so managers,

clinicians, administrative staff may access to

the information that is useful for them.

 Define specific rules for alerts and generate

visual feedback when these rules are not

satisfied.

 Generate alert messages for patients and

clinicians to manage specific situations, both

manually and automatically.

 Provide a dashboard of indicators that may

provide an overview of the activity and

resources occupancy.

 Generate and submit a daily report.

Table 1: Events in the ODH workflow.

Unit Event Info. Sys.

ODH PAT enters the hospital -

ODH PAT goes to ODH services for blood

extraction.

ODH NUR prints tags. HIS

ODH NUR takes blood. -

ODH NUR puts blood sample in the

pneumatic tube.

LAB LAB receives blood sample. * HIS

LAB LAB does blood analysis. -

LAB LAB provides the report. * LIS+HIS

EXC PAT takes the turn ticket. * AQ

EXC FAC calls PAT when LAB report

ready. *

AQ+HIS

EXC FAC provides subsequent

appointments and lab requests. *

HIS

EXC FAC fills in pharmacy prescription. * PhIS

EXC FAC fills in a form at HIS, updates

the PAT record and copies the report

in an unassigned note. The note is

always assigned to protocols (colon,

pulmonary, tumor committee). *

HIS

EXC FAC finalizes consultation. * HIS

PHAR PHAR checks the prescriptions

through the IS, following a stage-gate

predefined process. *

PhIS

PHAR The treatment is verified. * PhIS

PHAR PHAR prints the Report of

Preparation to Administer. This report

is signed by NUR and PHAR and it is

taken to the clean room.

Manual

PHAR Once the drug is ready, an ANC is

called.

Manual

PHAR ANC collects preparation at FAR. Manual

ODH ODH receives preparation. Manual

ODH ADM registers the patient and prints

the identification bracelet.

HIS

ODH PAT is admitted in ODH. * HIS

ODH NUR assigns a seat. HIS

ODH The administering process starts,

guided by the pharmacy system for

secure administration. *

HIS

ODH PAT discharge. * HIS

Events with asterisk are those that can be automatically retrieved

from IS. PAT: patient. LAB: laboratory. EXC: External

Consultation. ADM: Admissions. In the very specific case of ODH,

patients go straight to the blood draw service without going through

the reception desk.

Regarding the non-functional requirements (how

the system has to be), the TIS must perform

satisfactorily with respect to:

 Availability: it must be accessible and easy to

configure from any connected workplace at the

hospital, not needing any specific additional

software.

 Concurrency: its performance must to be not

penalized by the simultaneous use from

different workplaces.

 Security: the system must manage and control

every access and keep trace of them.

 Performance: the system response must be real-

time (not above 3 seconds).

 Usability: the system must be easy and

comfortable for the users. Not more than 10

minutes training should be necessary for the

users to work with the system.

2.3 Design Methodology

The methodology utilized to design the system

follows an iterative approach, in which iterations are

composed by analysis, development and testing

phases. In this case, the main users of the systems are

practitioners and managers, so their permanent

contribution on three prototyping stages that have

been necessary to come out with a first stable version

has been crucial.

3 SYSTEM ARCHITECTURE

The core of all the hospital information systems is the

HIS (Selene from Siemens), which is used in

Emergency Care, Hospitalization, External

Consultations and Day Hospital. Through it, any

patient appointment or request is managed

(diagnostic tests, subsequent consultations, inter-

consultations, follow-up notes, etc.). The HIS

facilitates the elaboration of forms, the generation of

reports and the visualization of the patient’s

Electronic Health Record.

Figure 2: System architecture.

Departmental applications are deployed to cover

the specific needs of a given Department or Service.

For example, FarmaTools is the departmental tool for

Pharmacy (drugs purchase, store distribution, and in

this case, cytostatic drug prescription, verification

and management). The appointment system (Quenda)

avoids voice calls and guarantees privacy overall the

hospital consultations. It enables to put in order the

waiting rooms and provides indicators for

consultation management (arrival time to the hospital,

consultation call time, finalization time). The

traceability and indicator system for ODH connects to

the APIs provided by these three tools to

automatically retrieve the data of interest.

Figure 2 shows the TIS architecture, which is in

practice deployed over Linux in a virtual machine.

The traceability and indicator system is composed by

several modules developed in PHP and HTML; these

modules retrieve real time information from the

mentioned systems. For integration with HIS Selene,

HL7 messaging is used, through a channel in the

integration engine MirthConnect. For Quenda,

FarmaTools and Selene’s mirror DB, direct access to

the databases is implemented (SQL Server and Oracle

DB in the last two cases). The TIS is composed by

four main elements: 1) a MySQL database, 2) a main

module in PHP/HTML that shows the current state of

the ODH patients, 3) eleven processes that update the

information in the MySQL database through cron

programmed jobs (Table 2) and 4) the indicators

module.

Table 2: Description of cron jobs.

Job Description Provider

TICKET

Patient takes the turn ticket on

arrival the hospital.

Quenda DB

LAB-

SENT

Patient goes to ODH for blood

test. Blood draw is notified.

HL7

ORU^R01,

HIS Selene

LAB-

RES

Lab result is received. HL7

EXC-

START

Timestamp when the oncologist

calls the patient for consultation.

Quenda DB

PHAR-

CONF

Timestamp for confirmation of

the drugs for the day, enabled by

the oncologist. When PHAR is

notified with the treatment

confirmation, the pharmacologist

in charge verifies the order and

submits it to the technicians, who

start working on the preparation.

FarmaTools

Oracle DB

EXC-

END

Timestamp when the oncologist

finishes the consultation.

Quenda DB

ODH-

ADM

After leaving the consultation, the

patient goes to the ODH and

admission verifies the

appointment and provides the

identification wristband.

HL7

ADT01

PROTOC

The drug protocol is obtained

from the PHAR database,

together with the number of

components to administer to the

patient.

FarmaTools

Oracle DB

When the administration of a

component is completed, it is

registered in the positive

dispensation module.

FarmaTools

Oracle DB

ODH-

DIS

When the administration is

completed, the patient is

discharged.

HL7

ADT^03

ALERTS

It checks if the alerts’ conditions

are fulfilled and generate the

defined alerts if so.

MySQL

Figure 3: a) Main dashboard for traceability. From the fourth column on, it is possible to find the timestamps for the 10 jobs

in Table 2. The 11th job generates the alerts, marked as blocks in orange. In the top right corner of the Figure, a shortcut to

filter the available agendas is provided.

Everyday at 7:00 am, a cron job initiates the

patients’ registry for the day, by using the agendas of

the ODH and the monographic agendas (related to the

oncology specialties) or nominative ones (related to

specific practitioners). Cron jobs are executed each 5-

10 minutes (configurable time), updating the data for

real-time visualization and feeding the indicators’

database.

Access and interface personalization is carried out

through IP control. ODH or Consultation users can

access the ODH TIS through any browser in an

authorized computer. The main interface, in Figure 3,

is daily initialized and dynamically completed

throughout the day, it starts “growing” when the ODH

activity begins. The interface provides visual alerts

(in orange) indicating when issues occur in the

workflow. These alerts are informative up to now,

being generated e.g. when “the laboratory results are

taking too long” or “the patient is not in the next step

of the workflow”. In the future, it is desired that these

alerts may trigger automatic or supervised responses,

such as “recall results from the laboratory” or “send a

notification to the patient”.

In the right side of the real-time interface, there

are several options that enable to parametrize the TIS,

showing information from a) all the agendas, only

from the ODH agendas or the Oncologists

Consultations agendas, b) detailed information about

the patient and c) day indicators.

4 REPORTING

Apart from the visualization interface, the TIS

provides a business intelligence tool that summarizes

a set of relevant key indicators, which can be

classified into four different groups: a) activity, b)

performance, c) procedural and d) quality of care

indicators.

Activity indicators reflect the day dynamics at the

Oncology Service (both at ODH and consultations).

These six indicators are directly accessible from

the interface in Figure 3 and include aspects such as

snapshot of number of active patients in ODH and

patient distribution per agenda (in ODH and

consultations), protocol, visit type and treatment type.

Table 3 shows an example of a protocol indicator, in

particular the one in which each chemotherapy

protocol for a given cancer type is retrieved (an

example of protocol is FOLFOX – oxaliplatin,

fluorouracil, folinic acid - for gastric cancer).

Table 3: Activity indicator example.

Activit

indicator no. 4

Scope

ODH, Oncology Consultations

and Pharmacy.

Definition Patient distribution per protocol.

Info. source Selene Replica, FarmaTools

Formula

Query that retrieves the total

number of patients in the agendas

of a) ODH (non-scheduled

patients included) and b)

Oncology Consultations,

classified by protocol.

Goal

To be defined by the evaluator /

manager.

Table 3: Activity indicator example (Cont.).

Activit

indicator no. 4

Responsible

Exploitation: Hospital

Management.

Evaluation: Oncology Service

Head

Visualization

The rest of indicators are retrievable from a

specific interface. There are nine performance

indicators, which aims at providing a view of the

service efficiency. This group of indicators include

e.g. the average number of admissions per hour, the

appointments’ status (cancellations, completed, not

registered), the number of non-scheduled patients, the

resource use or the real duration of the treatment. The

information available about an example of

performance indicator, the average duration of the

patient’s stay at ODH, is showed in Table 4.

There are six quality of care indicators that are

focused on compiling information about how the

patient’s perception may be. These indicators include

issues such as the delay of the patient with respect to

the appointment for blood extraction, the delay at the

oncology consultation, the time between the patient’s

appointment and the call time in the oncology

consultation or the delay in the start of the treatment,

etc. An example is available in Table 5.

Finally, there are some specific procedural

indicators that are focused on measuring the

evolution of procedure errors (e.g. admissions

without discharge time, number of never ended

consultations, number of non-called patients, etc.).

Table 4: Performance indicator example.

Performance indicator no. 8

Scope ODH

Definition Avera

e sta

duration at ODH.

Info. source Selene Replica, MySQL indicators

database

Formula Query, average duration of

admitted patients in ODH.

Goal To be defined by the evaluator /

mana

er.

Responsible Exploitation: Hospital Managmnt.

Evaluation: Oncology Service

Hea

Statistics Mean: 3.85 h.; Standard

deviation:0.17; Min: 3.51; Max:

4.21.

Visualization

Table 5: Example of Quality of care indicator.

Qualit

of care indicator no. 4.

Scope ODH

Definition Delay between the patient’s

scheduled appointment time and the

call time to consultation.

Info.

source

Selene Replica, Oncology

Consultation A

endas.

Formula Query, mean delay between the

scheduled time and the real call time

to consultation, filtered for days,

months or years.

Goal 0 minutes.

Responsible

Exploitation: Hospital Managmnt.

Evaluation: Oncolo

Service Hea

Statistics Mean: 22.97 m.; Standard deviation:

4.77; Min: 14.21; Max: 31.25.

Visualization

Figure 4 shows the components for the extract,

transform and load (ETL) process necessary to

generate the indicators. All these data are stored in the

MySQL database and queried from a PHP service that

uses Google Chart libraries to generate the final

interface. Everyday, a cron job captures the main

dashboard screen of the traceability system and

emails it to the designated receivers (e.g. Director of

the Oncology Service).

Figure 4: Extract, Transform, Load process for indicators.

DB stands for database.

5 CONCLUSIONS

The Traceability and Indicators System described in

this paper facilitates the retrieval of significant

information in an ODH, with the purpose of

improving daily operation and patient satisfaction.

Traditionally, part of the information retrieved by the

TIS has been manually recorded and processed to

obtain indicators, and only significant deviations,

detected. The TIS has been designed upon request of

the Oncology Service Management to have a better

view of the ODH workflow, so specific

organizational and technical actions can be designed

on a solid informational input.

The TIS has been designed in tight collaboration

with practitioners and administrative staff, with the

objective that it can fulfil the real requirements of

day-to-day operation in the Oncology Service of a

hospital, providing added value and avoiding

problems in the organizational change that the

availability of this tool may suppose.

The TIS relies on information that can be

automatically retrieved from the existing systems in

the hospital. This entails that the deployment does not

require complex technology or significant

investments, while providing relevant information of

the workings of the oncology day hospital. The design

process has been useful to identify information gaps

(e.g. part of the processes that are not automatized,

still done in manual ways), so their improvement can

be added to the strategic agenda of investments.

The system has been technically evaluated against

the design requirements in Section 2.2. Although it

fulfils them reasonably well, there are several aspects

that need to be improved. For example, the

management of both alerts and notifications still need

to be better implemented, and the set of indicators,

polished. Additionally, although the available

presentation interfaces have been incrementally

improved with the help of the users, different

presentation formats should be benchmarked. This

can also be applied to indicators; it is also necessary

to systematically analyse the causes of the detected

deviations to feed the TIS.

Up to now, practitioners and nurses have partly

tried the system in oncology consultations, ODH and

Service Head offices, but their experience with the

TIS is still brief to extract conclusive comments. In

any case, from their feedback it is possible to say that:

 The system gathers real-time information in a

reliable way, although it is still necessary to

make some procedural changes to better

contextualize some data. For example, due to

specific instructions given to ODH patients, the

time when these patients take the turn ticket for

consultation is stored, but the entry time at the

hospital is not (ODH patients are told to go

directly for blood draw service).

 It is necessary to provide better traceability at

resource level (seat or bed), with the purpose of

optimizing its occupation. This traceability is

manually done and may not be accurate enough

due to human errors. A technical solution

involving RFID or barcodes is currently being

designed to address this issue, and it will be

integrated at the TIS when ready.

 Alerts in the real-time visualization interface

need to include specific management options,

to track if the alert has been handled. Up to

now, they are merely informative and no action

is triggered from them.

 One of the possibilities to manage alerts is to

make them trigger SMS both to patients or care

professionals, as there is a corporate existing

platform for this purpose. Other options, such

as instant messaging, involve technology and

organizational changes and may delay the

integration with the service. In any case, it is

necessary to study to which extent these SMS

may be effective and useful for their recipients

(misleading or spam effects). In particular, it is

necessary that involved users express their

opinion about the best means and configuration

of alerts.

An important issue is related to human factors: to

assure the full exploitation of the visualization tool

for real time patient-flow management purposes wide

acceptance within the health team is needed. All

health staff members need to understand the ultimate

instrumental goal, focused on improving delivery of

quality of care and not staff monitoring. Specific

communication initiatives are needed to guarantee

that this fact is correctly understood. In this sense, the

involvement of navigator nurses in key steps control

has been shown of great help. In the next months, it is

expected that the tool is integrated in the daily activity

of the ODH and improved following the users

suggestions.

Taking into consideration this TIS, another tool is

currently being designed to monitor the workflow in

the 200 hospital consultations, so delay times and

bottlenecks can be identified.

ACKNOWLEDGMENT

Authors want to acknowledge the Hospital

Universitario Infanta Sofía Management for the

support to this initiative.

REFERENCES

Queralt M., Frías M.C., de Pedro, V., Ramírez, J., Mesalles,

M. 2015. Sistema de trazabilidad integral de

administración de citotóxicos. Boletín de Prevención de

Errores de Medicación de Cataluña, Vol. 12, no.4.

Kergosien, Y., Tournamille, J.F., Laurence, B. and Billaut,

J.C., 2011. Planning and tracking chemotherapy

production for cancer treatment: A performing and

integrated solution. International journal of medical

informatics, 80(9), pp.655-662.

SIAE (2015). Información estadística de hospitales.

Estadística de Centros Sanitarios de Atención

Especializada (National statistic information about

hospitals). Available at: https://www.msssi.gob.es/

estadEstudios/estadisticas/estHospiInternado/inforAn

ual/homeESCRI.htm.

Sini, E., Torresani, M., Locatelli, P., Restifo, N. and

Facchini, R., 2011. Information and communication

technologies enable a comprehensive governance of

patient safety in chemotherapy. An Italian framework

project. In 24th International Symposium on Computer-

Based Medical Systems (CBMS) (pp. 1-7). IEEE.

Masini, C., Nanni, O., Antaridi, S., Gallegati, D., Marri, M.,

Paolucci, D., Minguzzi, M. and Altini, M., 2014.

Automated preparation of chemotherapy: quality

improvement and economic sustainability. Am J Health

Syst Pharm, 71(7), pp.579-85.

Galligioni, E., Berloffa, F., Caffo, O., Tonazzolli, G.,

Ambrosini, G., Valduga, F., Eccher, C., Ferro, A. and

Forti, S., 2009. Development and daily use of an

electronic oncological patient record for the total

management of cancer patients: 7 years' experience.

Annals of Oncology, 20(2), pp. 349-352.