Strategic Planning of Hospital Service Portfolios

The DRGee Viewer

Dominique Brodbeck

, Markus Degen

, Andreas Walter

, Serge Reichlin

and Christoph Napierala

4,5

University of Applied Sciences and Arts Northwestern Switzerland, Basel, Switzerland

Inselspital, Bern University Hospital, Bern, Switzerland

University Hospital Basel, Basel, Switzerland

University of Lucerne, Luzern, Switzerland

Siemens Schweiz AG, Z

urich, Switzerland

Keywords:

Strategic Hospital Planning, DRG, Visual Analytics.

Abstract:

In 2012 inpatient ﬁnancing for hospitals in Switzerland was changed from a system based on cost per case

to a system based on a ﬁxed fee per case. The ﬁxed-fee model makes medical services comparable from

a ﬁnancial point of view. Characterizing medical service portfolios in this way, creates large amounts of

high-dimensional data. In order to operationalize this information and use it as a factual basis for decision

support, we developed a visualization tool and a methodology to support strategic planning of hospital service

portfolios. The method centers around a visual metaphor that provides the basis for strategic thinking. It is

complemented by a visualization tool that allows visualization, analysis, and modiﬁcation of service portfolios.

Extensive support is provided for visual comparison of different scenarios. Special features enable the tool to

be used during live planning sessions. The system was used in strategy workshops in over forty hospitals, and

has contributed to infrastructure planning, reorganization, and resource optimization decisions.

1 INTRODUCTION

The Swiss healthcare system is continuously under-

going change since the implementation of the new

Swiss health insurance law in 1996. Most signiﬁ-

cantly, in 2012 inpatient ﬁnancing was changed from

a system based on cost per case to a system based on a

ﬁxed fee per case. Such a signiﬁcant change has a far-

reaching impact on the overall system, especially on

the hospital sector. Hospital management, but also na-

tional and regional policy makers, are forced to plan

and manage in different dimensions than before.

The idea behind a ﬁxed-fee model is that hos-

pital cases can be classiﬁed into groups of similar

cases, and that these groups can then be treated like

products that are comparable, and that are reimbursed

with the same ﬁxed amount of money, because it

is assumed that they have the same cost structure.

The classiﬁcation rules are based on diagnoses, medi-

cal procedures, demographic patient information, and

other case-speciﬁc data. The combination of all these

groups, or products, makes up the medical service

portfolio of a hospital. This comparability of medi-

cal intervention from a ﬁnancial point of view is new

and opens a variety of new possibilities. In particu-

lar, it allows policy makers and hospital management

to make decisions based on factual information from

the analysis of current service portfolios, as well as

the simulation of medical service portfolios into the

future. Such ﬂexible analysis possibilities are crucial

to make sound infrastructure decisions that will meet

patient needs in the next years.

Characterizing and comparing service portfolios

in such a way quickly produces large amounts of

high-dimensional data. The portfolios need to be vi-

sualized, compared, interpreted, and modiﬁed by an-

alysts, as well as by groups of managers in live work-

shop settings. Common spreadsheet programs are too

general to cover these tasks well. In order to take full

advantage of the possibilities offered by the new reim-

bursement system, there is a need for tools that allow

the various stakeholders to analyze and communicate

the data in a ﬂexible and efﬁcient way.

This paper describes a method and a correspond-

ing visualization tool that supports management lev-

els in discussing strategic decisions and future devel-

Brodbeck D., Degen M., Walter A., Reichlin S. and Napierala C..

Strategic Planning of Hospital Service Portfolios - The DRGee Viewer.

DOI: 10.5220/0005210100820090

In Proceedings of the International Conference on Health Informatics (HEALTHINF-2015), pages 82-90

ISBN: 978-989-758-068-0

 2015 SCITEPRESS (Science and Technology Publications, Lda.)

opments based on current restrictions and conditions.

2 BACKGROUND

Diagnosis-related groups (DRGs) is a patient classiﬁ-

cation system that links similar types of cases that a

hospital treats, to the resources that the hospital uses

for the treatment. DRG systems were pioneered in

the USA, but are now in use in many other countries

(Schrey

ogg et al., 2006).

The SwissDRG system is a variation of German

DRG but refers to Swiss reference values (Swiss

DRG, Web). Like in all DRG systems, single cases

are classiﬁed into a speciﬁc class according to their

diagnoses (using ICD10), treatment types (e.g. hip

surgery), and other case attributes. Each group has a

unique code, which enables a precise attribution of a

case’s revenues to its costs. This results in a shift from

a more cost-oriented daily allowance to a product-

oriented view, where optimal management along the

patient pathway gains in relevance (e.g. in (Gocke

et al., 2002) and (Rohner, 2012)). This shift opens

the potential for providing a basis for evidence-based

decisions in an operative and above all in a strategic

perspective.

The dimensions of interest for strategic analysis

are: average length of stay, cost weights, and the over-

all base rate. The average length of stay (ALOS) is

used in a DRG system as an output measurement for

complexity of the case (Luke, 1979). The other di-

mension is the cost weight, a relative indicator for

the severity of the case, which serves as a measure

for resource intensity. Cost weight is recalculated ev-

ery year based on cost data provided by hospitals to

the Case Mix Ofﬁce (i.e. SwissDRG in Switzerland).

Each DRG is assigned speciﬁc values for these two in-

dicators, and all the values for all the DRGs are stored

in a reference catalog.

The base rate (which is basically determined by

negotiation between the hospital and insurance com-

panies) is multiplied with the cost weight attribute to

determine the monetary value of a case that is clas-

siﬁed into a particular DRG. Besides revenues from

private insurance and specialized pharmaceuticals or

technologies for special treatments (e.g. specialized

oncological treatments) these revenues represent the

main income that a hospital can generate from their

patients.

As theoretically the cases behind any DRG code

should on average represent similar cases in each hos-

pital, these elements make hospitals comparable and

allow a benchmark-oriented approach. Comparisons

can be achieved by using one hospital’s cases, and

then comparing them either with the reference values

of the DRG catalog, or to the portfolio of a peer hos-

pital, both at the service and at the cost level. Most

importantly, this allows comparing the efﬁciency of

hospitals, but it will also enable quality or other as-

sessments.

The services that a hospital offers are inﬂuenced

by many factors. In most administrative districts (i.e.

in our Swiss case corresponding to cantons), service

portfolios or service requirements are deﬁned by pol-

icy makers. However not all districts decide to pro-

vide a full scope of medical services, but might del-

egate the remaining activities to other providers that

can be situated beyond their control or area of di-

rect responsibility. This is complicated by the fact

that with the new reimbursement system, patients also

gained the freedom to choose their hospital of treat-

ment independently of their canton of residency.

On top of that, there are a number of national poli-

cies that impose further restrictions, e.g., the highly

specialized medicine act that restricts the number of

centers for very complicated treatments, or the chang-

ing outpatient health care provision that still heav-

ily depends on resident physicians, who are however

overaged and struggle to ﬁnd replacing practice hold-

ers. All these factors push hospitals towards stratify-

ing patient portfolios and focusing on selected medi-

cal areas, in order to improve their economic situation

actively.

Modern hospitals must engage in strategic discus-

sions about visions, cooperations, specialization, cen-

tralization and further infrastructure or organization-

ally relevant questions. These decisions affect their

market position and help them to cope with growing

and dynamic competitors both at administrative dis-

trict or regional levels.

For example, hospitals that today provide maxi-

mal service levels, will have to decide whether they

will focus on more speciﬁc therapies that generate

high cost weights (L

ungen and Lauterbach, 2002), or

whether they will continue offering a broad range of

health care services, focusing on a me-too strategy.

The analysis of pancreas and esophagus carci-

noma surgery in Germany as another example has

shown that it can be highly beneﬁcial for hospitals to

treat a high number of benign cases instead of focus-

ing on the malign cases with high cost weights but bad

average length of stay management. This is the case

even if the minimal numbers are reached that would

be required by the ofﬁcial rules and regulations.

Ultimately, Swiss hospitals will have to adapt their

current management models. In order to take the right

decisions in this complex environment, hospital man-

agers need to rely on data, and models that are based

StrategicPlanningofHospitalServicePortfolios-TheDRGeeViewer

Figure 1: The visual metaphor of the strategic model: a bubbleplot that plots the case mix index (CMI) of a medical service

vs. how well the length of stay compares to the country average (ALSO CH / ALOS). This creates four quadrants A, B, C,

and D, with different strategic interpretations. The size of the bubble is proportional to the case count for this service. The

color shows Major Diagnostic Category MDC.

on that data. Collecting the data is usually not a prob-

lem anymore, as modern hospital information systems

are well equipped for that job and are commonly in

use today. Projecting the data into the strategic mod-

els and operationalizing it however, requires new ap-

proaches.

3 METHODS

3.1 Strategic Model

As outlined above, the key dimensions of a hospital

case are its cost weight - as determined by the DRG

into which the case was classiﬁed - and the length of

stay. Since the analysis of individual cases is too low-

level for the kind of strategic questions that need to

be supported, the cases are aggregated into groups ac-

cording to their DRG code. These groups can be con-

sidered as the services that a hospital performs.

Working with inpatient service portfolios from

various hospitals of various sizes, types, and different

geographic areas has shown that plotting the normal-

ized average of the cases’ cost weights (also called

case mix index CMI) of such a service, versus the

deviation of the average of the lengths of stay (also

called ALOS) from the DRG catalog reference (CHA-

LOS), produces a graph that is easy to interpret in the

context of strategic questions (Figure 1). Each ser-

vice is plotted as a bubble with the size of the bub-

ble proportional to the number of cases, and the color

mapped to any of the other available service attributes

(e.g. Major Diagnostic Category MDC, proﬁt, cost,

department) depending on the focus of the analysis.

The plot can be separated into four quadrants that

each have a distinct strategic meaning, similar to the

BCG growth-share matrix used in a strategy or mar-

keting context (Boston Consulting Group, Web). This

allows to analyze the strategic positioning of a hospi-

tal’s inpatient portfolio using the four quadrants (A-

D). The quadrants have the following interpretation:

Quadrant A represents the area where the

grouped cases generate on average a high cost weight,

and result in an average length of stay that is better

(i.e. shorter) than the DRG catalog value. We there-

fore expect the bubbles in this quadrant to be prof-

itable and to represent the services where the hospitals

portfolio performs better than the benchmark. Strate-

gically, these activities support the hospital in creating

a brand, where the hospital can distinguish itself for

quality and performance, and thus should aim at in-

creasing case numbers and building relationships to

referring medical doctors. Further options could be to

HEALTHINF2015-InternationalConferenceonHealthInformatics

create a good infrastructure that would attract above-

standard surgeons, and centralize the activities of sur-

rounding hospitals in this area. In general, these kind

of services should be expanded and require focus.

In quadrant B the case mix indices (CMIs) are

lower than the average of the hospital, but the aver-

age length of stay in the hospital is better than the

catalog value. The cases in this part of the grid re-

quire detailed analysis of costs. The right infrastruc-

ture is of special importance because economic mar-

gins are tight. Strategic decisions like outsourcing or

PPP (private public partnerships) models should be

considered. In general, this area is expected to be at

least cost covering and the portfolio overall should be

kept constant.

At the bottom left, quadrant C represents cases

where ALOS ratio is below 100%, i.e. worse than the

benchmark and where a low CMI is generated. Be-

cause the CMI is low, these activities are potentially

loss generating. The hospital has to thoroughly in-

vestigate its patient pathways and focus on workﬂow

management, trying to cooperate closely with rehabil-

itation or care management. Furthermore, direct mea-

sures to lower costs need to be taken through imple-

menting for example intermediate care units (IMC)

or similar. Additionally, cooperations with resident

medical doctors and other hospitals that could take

over those cases, need to be investigated. In sum-

mary, this area marks activities where the hospital has

to reﬂect, why its activities cannot be executed cost-

effectively, or if they could be provided in a day-care

or outpatient management setting.

The last quadrant D is characterized by high case

mix indices and low ALOS ratios. As in quadrant

C, an internal focus needs to concentrate on patient

pathways and workﬂow management. From a long

term perspective, the number of cases could then be

increased. This can be realized through the creation

of specialized competence centers, with the aim of at-

tracting cases from surrounding hospitals and refer-

rals from resident physicians. Interesting improve-

ments can also be generated from applying a sound

case management that addresses high-outlier issues

(i.e. cases that remain in hospital above the high trim

point length of stay according to the respective DRG

code). Nevertheless this area has a dualistic perspec-

tive. Either processes and cost structures are opti-

mized, or the number of cases are reduced in order

to improve the overall economical situation.

While the above description is necessarily rough

and exemplary, it clearly indicates the added value of

such a graphical representation. It serves as a map of

reference for discussions without the need to refer to

quantitative tables with many dimensions.

3.2 Tool Support

In order to make it possible to apply this strategic

model in practice, we developed an interactive visual

tool called DRGee (as in ”(DR)Gee! Look at this!”)

that supports the planning process. The tool uses a

dynamic bubble chart, similar to (Gapminder, Web),

but with a focus on dynamically modifying the un-

derlying data, instead of temporal trend analysis. It

provides the following core functionality:

• Load a collection of portfolios (a theme)

• Visualize a portfolio in a standardized way

• Allow selection of parts of a portfolio

• Calculate and display characteristic indicators and

summary values of a portfolio or selected parts of

• Allow the modiﬁcation of a portfolio by creat-

ing a copy and editing individual services to play

through what-if scenarios

• Provide the possibility to compare the differences

between two or more portfolios

Figure 2 shows the main interface of the tool. The plot

that was introduced with the strategic model is fea-

tured prominently in the center of the interface. The

case-mix index is plotted against the ratio of catalog

length-of-stay and actual length-of-stay. The bubbles

represent medical services that a hospital performs,

and the size of the bubbles is proportional to the num-

ber of cases for the service.

Along the left side we ﬁnd from top to bottom:

• The list of portfolios that is loaded (simple selec-

tion of a portfolio in the list displays it and makes

it active, portfolios can be cloned here in order to

modify them)

• A table of all the services deﬁned in the selected

portfolios, showing any number of attributes that

characterize each service (services can be sorted,

selected, and colored here to make them active

and visible, deselected services are shown as

ghosts in the background to preserve overall con-

text)

• A display of characteristic indicators and sum-

mary values of a portfolio (e.g., total number of

cases, revenue, average case-mix index, average

length of stay, etc.). There is a choice between the

whole portfolio or just the selected subset.

All the necessary controls are contained in these three

interface elements. The menus in the menu bar are

only used for high-level conﬁguration of the tool (e.g.,

StrategicPlanningofHospitalServicePortfolios-TheDRGeeViewer

Figure 2: The DRGee tool. The 4-quadrant plot is at the center stage. Services (bubbles) can be directly resized and repo-

sitioned with the mouse to simulate what-if scenarios. Controls along the left are for selecting portfolios, services within

portfolios, and for displaying summary information. Tabs along the top are plugins that extend the basic functionality, mainly

for comparing the various scenarios and documenting the analysis process.

base rate, currency, etc.) and infrequently used func-

tions (e.g., data export, printing, etc.). This was a de-

liberate design decision in order to ensure the discov-

erability of the main functionality without having to

resort to user manuals.

The software was implemented in Java for plat-

form independence, and uses only a handful of third

party libraries, mainly for the look&feel, logging, and

PDF generation.

The architecture provides a dynamic extension

mechanism, where modules (plugins) can be com-

piled into self-contained (Java) jar ﬁles that are loaded

automatically and independently at starttime. Every

plugin loaded this way is placed into a separate tab in

the main window of the application, as visible along

the top of the plot in Figure 2. Encapsulating different

features into plugins allows to incrementally increase

the functionality of DRGee without having to touch

the overall architecture and core functionality of the

system. In addition, it supports the easy packaging

and delivery of different versions of DRGee with tai-

lored functionality.

3.3 Working with Scenarios

One of the key features of the tool is the possibility

to modify existing portfolios in order to play through

what-if scenarios, and simulate how the characteristic

values of a portfolio change. Each service (bubble)

has two dimensions that can be changed: length-of-

stay compliance (y-axis) and case count (size of bub-

ble). The case mix indices (x-axis) can not be mod-

iﬁed, as they are ﬁxed and predeﬁned by the DRG

reference catalog.

To modify a bubble, it can simply be manipulated

with the mouse pointer (drag to new position, drag ra-

dius to new size). The values in the table are adjusted

accordingly. Direct manipulation techniques have the

advantage that they are intuitive and efﬁcient if high

precision is not required (Shneiderman, 1983).

Comparing the newly created scenario with the

original portfolio is challenging. There are three

different approaches for visual comparison (Gleicher

et al., 2011): juxtaposition (showing objects side-by-

side), superposition (showing objects overlaid in the

same space), and difference (showing the difference

between the objects).

The DRGee tool supports all three approaches,

HEALTHINF2015-InternationalConferenceonHealthInformatics

Figure 3: Comparing portfolios by juxtaposition. Signif-

icant differences and overall trends are perceived quickly

(e.g., 901 moved and shrank, 540 moved and grew).

Figure 4: Comparing portfolios by superposition. Individ-

ual differences of only the modiﬁed objects are shown as

small multiples. This provides a complete and precise rep-

resentation of all the changes, at the disadvantage of losing

the overall context.

Figure 5: Showing the differences between two portfolios

explicitly. Changes in position are marked by arrows. The

size of the new bubbles is shown transparently so as to not

hide the original conﬁguration in the background.

each implemented as a separate plugin. Figure 3

shows two portfolios juxtaposed. This view works

mainly well for detecting a few strong outliers. Su-

perposition was implemented by using a small mul-

tiple (Tufte and Graves-Morris, 1983) representation

of the plot per bubble that changed. Each small rep-

resentation then only shows the changes to that par-

ticular bubble superimposed in the same plot (Figure

4). The third approach represents the differences be-

tween two plots explicitly as arrows between the orig-

inal and the new position (Figure 5). This technique

allows not only to identify changes in the portfolio but

also shows trends (e.g., ”DRGs in quadrant C tend to

be smaller”, ”DRGs mostly move to quadrant A”, etc.)

3.4 Workshop Use

A special requirement for the tool was that it had to

support use in a live workshop setting. As a con-

sequence the following additional aspects had to be

taken into consideration:

Animated Transitions

One of the main differences between operating a tool

oneself and watching someone else perform the oper-

ation is that not all the intentions and actions of the

operator are visible, but very often only the results of

the actions. Even with close attention, changes can

be missed by the audience, a phenomenon known as

change blindness (Rensink et al., 1997), leading to

mental disconnect and discomfort. In order to address

this problem, all the transitions in the tool (zooms,

pans, size and position changes of bubbles, change of

portfolios, etc.) are smoothly animated or designed as

drag and drop interactions, leading the viewers to the

next state of the visualization in a continuous way.

Simple Interface

Operating a tool live in front of an audience puts a cer-

tain pressure on the operator. It is therefore essential

that the user interface is robust and the functionality

reduced to a set of minimal yet powerful enough fea-

tures. All the controls need to be visible so that the

audience can perceive the series of actions performed

by the operator.

The controls for navigating the bubble plot for in-

stance are not implemented as continuous sliders, but

are constrained to just a few discrete buttons (see Fig-

ure 2 bottom, below the plot). The ﬁrst button zooms

to a ﬁxed standardized view, and as such serves as

the ”home” button. The second button zooms to the

full range that shows the complete portfolio, which

provides the overview. From these two well-deﬁned

positions, in-between views can be obtained by using

the ”+” and ”-” buttons that zoom the view by 20% in

the respective direction.

StrategicPlanningofHospitalServicePortfolios-TheDRGeeViewer

Documentation of the Decision Process

The strategic development of a hospital portfolio is an

iterative process, and it is often necessary to go a step

back and try another path. Traditional tools only pro-

vide linear undo/redo stacks, and only the ﬁnal result

is saved. In an analytical process however, especially

if it is a collaborative effort shared between partic-

ipants, the intermediate steps are important as well.

Making this path visible and navigable is important to

support the analytical reasoning process (Shrinivasan

and van Wijk, 2008).

Figure 6: The history of the analysis process is documented

as a tree (bottom). Each node represents a different editing

step of the portfolio. Nodes marked with a star have a com-

ment attached. Nodes can be rated (green - relevant, red -

not relevant). Any two nodes can be compared and their dif-

ferences visualized (top left) and numerically characterized

(top right).

In the DRGee tool, we implemented a history in

the form of a tree where every modiﬁcation to the

portfolio results in a new node (Figure 6). Users can

navigate to any node in the tree, see the portfolio’s

state at that point, and branch out from there. A com-

ment can be added to a node to describe the reasons

for the modiﬁcation. Any two nodes in the tree can

be compared, and the differences in the two states are

shown both visually and numerically as the relative

difference of the characteristic values.

The complete history is saved, and the history plu-

gin therefore also serves as a documentation feature.

Having access to an automatically generated, naviga-

ble and interactive documentation is very useful in the

consolidation phase after the workshop is conducted.

A viewer-only version of the DRGee tool can be dis-

tributed, loaded with the data and the history, to a

wider audience for review.

4 RESULTS AND DISCUSSION

The methodology, together with the supporting tool

described in the previous section, was used in over

forty different hospitals all around Switzerland over

the course of four years. The type of hospitals ranged

from smaller regional hospitals to hospital groups

within a larger geographic region.

Over the course of this time, many workshops

with hospital management were performed. Both the

method and the tool were reﬁned continuously, but

typically the workﬂow looked like this:

1. Collect the basic data from the hospital and mas-

sage it into a set of portfolios relevant for the

strategic theme that is of interest. Since the na-

ture and quality of the data varies considerably

between hospitals, we used a range of tools for

data cleaning and consolidation. Typically these

included a primary clean-up in delivery ﬁles (i.e.

most of the time incomplete rows), load and en-

richment of data in a database where consistency

checks were performed, and a ﬁnal validation us-

ing known performance indicators for the respec-

tive hospital in order to ensure the consistency of

the full data load. The ﬁnal step consisted in cre-

ating the DRGee Viewer load ﬁles.

2. Analysis of the portfolios by healthcare experts.

The tool’s analysis capabilities proved very use-

ful to ﬁrst gain the general overview of the data.

Following this, the portfolio comparison function-

ality allowed to check for outliers, or for yearly

comparisons of two speciﬁc portfolios (e.g. yearly

comparison of the same two clinics). A deeper

analysis was then performed by using the super-

position comparison. It was possible to directly

display single DRG groups for example over time,

and compare them in one overview allowing for a

direct comparison. The latter oftentimes provided

the basis for the simulation within the existing or

a new portfolio.

3. Perform a strategy workshop with hospital man-

agement in which the tool is used to present the

ﬁndings of the analysis, explore different scenar-

ios, and discuss future strategies. These work-

shops typically included senior hospital manage-

ment, controllers, or infrastructure planners, and

are lead by a healthcare consultant that is familiar

with the data. Typical questions revolved around

what-if scenarios. For example, what happens if I

alter the ALOS Ratio of a speciﬁc group by adapt-

ing my processes considerably? Or what are the

effects of increasing the number of cases in a spe-

ciﬁc DRG or group? Participants found it very

useful that they could directly see the effect of

changing parameters (i.e. either the ALOS ratio

or the number of DRGs per group) and validate

and further discuss the results either in the various

HEALTHINF2015-InternationalConferenceonHealthInformatics

portfolio comparison views, or see the proﬁtabil-

ity effects in the summary values.

4. Provide the data used in the workshop and

any resulting modiﬁcations to the portfolios to

hospital management, together with a limited-

functionality version of the tool for further dis-

cussion and communication. The recipients were

able to proﬁt by continuing the work beyond the

guided workshops. The wider distribution of the

tool, also to non-expert users, did not turn out

to be a problem. The tool proved to be intuitive

enough so that the required support was not sig-

niﬁcant.

In the workshops, the tool is used in front of a group

of people. This prompts the question, if there is an op-

portunity to extend the tool to provide more explicit

support for collaboration. There are various scenar-

ios in which collaborative visualization can occur. In

the space-time matrix that is used by (Isenberg et al.,

2011) to categorize the design space, our implementa-

tion of DRGee currently corresponds to a co-located

synchronous scenario, where a group of experts in the

same room interact socially to create a common un-

derstanding of the data. Large interactive walls or

multi-touch tabletop displays are technologies that are

of interest in this context. However these devices are

not yet widely available at the places where hospital

decision makers work, which is why the practical po-

tential of such approaches is still limited.

A web-based version of DRGee would facilitate a

distributed asynchronous scenario, with access to the

tool and the analysis results from arbitrary locations

at any time. The focus would not be on social aspects

of a large (mostly lay) audience such as investigated

by systems like Many Eyes (Viegas et al., 2007), but

more on the asynchronous aspect of supporting col-

laborative analysis across time.

Examples of the strategic insights and decisions

that resulted from this work are:

• A group of hospitals reorganized their main op-

erative theaters, because services were delivered

far more efﬁciently in one hospital than the other.

Consecutively the costs were reduced substan-

tially, leading to an overall beneﬁt within this sec-

tor of activity.

• The tool helped a smaller regional hospital to rec-

ognize its strategic fallacies, and thus proceed to a

strategic partnership by giving up a certain part of

their activities and re-focusing the hospitals over-

all market approach.

5 CONCLUSION

We have developed a tool to visualize and edit hospi-

tal service portfolios in live workshop settings. The

tool is embedded in a strategy methodology that is

used by hospital management and healthcare consul-

tants. The method is actively used and has been suc-

cessfully applied in more than forty hospitals around

the country so far.

Strategic planning is difﬁcult since the set of vari-

ables to consider is multidimensional and complex.

Having a tool at hand that supports modelling, visu-

alization, evaluation, and comparison of various ap-

proaches while self-documenting the individual steps,

proved to be extremely helpful and often acted as a

catalyzer in the planning workshops.

The supporting tool goes beyond graphics towards

an interactive toolset that allows to understand com-

plex situations intuitively and discuss strategic chal-

lenges. In this way the DRGee Viewer distinguishes

itself from purely operative tools that are available in

the market, as it is embedded in a logical framework.

This makes it robust enough (by using the same port-

folio structures at all levels) to provide a stable, com-

parable and evidence-based foundation for decisions.

The approach has the potential to be used at a

more complex level, beyond a single hospital groups

perspective, towards a more public health oriented di-

mension. This is where the value of such a tool could

be useful for health service planning and provision.

Other directions in which the system will be ex-

tended are twofold:

• On the technical side, a migration to a web-based

system is planned, to allow concurrent and collab-

orative work.

• It is planned to incorporate more data sources,

such as quality data, demographic data and geo-

graphic conditions, and a comprehensive simula-

tion within this new data landscape will be devel-

oped.

In summary, the method supported by the DRGee

Viewer reﬂects claims in literature (Geissler et al.,

2011) that DRGs are not solely a tool for ﬁnancing

hospital services, but are also well suited for increas-

ing transparency, inducing efﬁciency and supporting

the hospital management in strategic decisions.

REFERENCES

Boston Consulting Group (Web). The Product Portfolio.

Available online at http://tinyurl.com/prodportfolio,

visited on 08/26/2014.

StrategicPlanningofHospitalServicePortfolios-TheDRGeeViewer

Gapminder (Web). Gapminder Homepage. Available online

at http://www.gapminder.org/, visited on 11/16/2014.

Geissler, A., Quentin, W., Scheller-Kreinsen, D., and Busse,

R. (2011). Introduction to DRGs in Europe: Com-

mon objectives across different hospital systems. Di-

agnosis Related Groups in Europe: Moving Towards

Transparency, Efﬁciency and Quality in Hospitals,

pages 9–21.

Gleicher, M., Albers, D., Walker, R., Jusuﬁ, I., Hansen,

C. D., and Roberts, J. C. (2011). Visual comparison

for information visualization. Information Visualiza-

tion, 10(4):289–309.

Gocke, P., Debatin, J. F., and Drselen, L. F. J. (2002).

Prozessmanagement und Controlling in der Diagnos-

tischen Radiologie im Krankenhaus. Der Radiologe,

42(5):332–343. 00012.

Isenberg, P., Elmqvist, N., Scholtz, J., Cernea, D., Ma, K.-

L., and Hagen, H. (2011). Collaborative visualization:

Deﬁnition, challenges, and research agenda. Informa-

tion Visualization, 10(4):310–326.

Luke, R. D. (1979). Dimensions in hospital case mix mea-

surement. Inquiry, 16(1):38–49.

ungen, M. and Lauterbach, K. W. (2002). F

uhren DRG

zur Spezialisierung von Krankenh

ausern? Gesundh

okon Qual manag, 7(2):93–95. 93.

Rensink, R. A., O’Regan, J. K., and Clark, J. J. (1997). To

see or not to see: The need for attention to perceive

changes in scenes. Psychological Science, 8(5):368–

373.

Rohner, P. (2012). Achieving impact with clinical process

management in hospitals: an inspiring case. Business

Process Management Journal, 18(4):600–624. 00003.

Schrey

ogg, J., Stargardt, T., Tiemann, O., and Busse, R.

(2006). Methods to determine reimbursement rates

for diagnosis related groups (drg): A comparison of

nine european countries. Health Care Management

Science, 9(3):215–223.

Shneiderman, B. (1983). Direct manipulation: a step be-

yond programming languages. Computer, pages 57–

69.

Shrinivasan, Y. B. and van Wijk, J. J. (2008). Supporting

the analytical reasoning process in information visu-

alization. In Proceedings of the SIGCHI Conference

on Human Factors in Computing Systems, CHI ’08,

pages 1237–1246, New York, NY, USA. ACM.

Swiss DRG (Web). Swiss DRG Homepage. Avail-

able online at http://www.swissdrg.org/, visited on

08/25/2014.

Tufte, E. R. and Graves-Morris, P. (1983). The visual dis-

play of quantitative information, volume 2. Graphics

press Cheshire, CT.

Viegas, F., Wattenberg, M., van Ham, F., Kriss, J., and

McKeon, M. (2007). Manyeyes: a site for visual-

ization at internet scale. Visualization and Computer

Graphics, IEEE Transactions on, 13(6):1121–1128.

HEALTHINF2015-InternationalConferenceonHealthInformatics