An Information and Communication Platform Supporting

Analytics for Elderly Care

Orhan Konak

, Harry Freitas Da Cruz

, Marvin Thiele

, David Golla

and Matthieu-P. Schapranow

Hasso Plattner Institute, Digital Health Center, Rudolf-Breitscheid-Str. 187, 14482 Potsdam, Germany

Data Experts GmbH, Allee der Kosmonauten 33G, 12681 Berlin, Germany

schapranow@hpi.de

Keywords:

Nursing Care, Software Architecture, Decision Support, In-Memory Database.

Abstract:

Germany faces an increase of its elderly population and along with it the number of people reliant on nursing

care is also rising. In this context, access to reliable information is key for all actors involved, be they family

members or political decision makers. Currently, the country lacks a centralized platform on which such actors

can access and exchange relevant information, e.g. as concerns ﬁnding a suitable facility or identifying trends

on the demand for care spots. Existing solutions are based on regional data silos, which render information

exchange time-consuming and error-prone. As a result, nursing care actors lack access to timely, reliable

information to support strategic decision-making. In this paper, we introduce a software platform built upon

an In-Memory database that meets the information and communication needs of the different user groups

involved. The platform establishes the necessary framework for real-time data collection and information

exchange, laying the foundation for deriving key performance indicators and enabling data exploration and

prognoses.

1 INTRODUCTION

The increase in life expectancy in Germany has co-

incided with a sharp rise in the number of elderly

people who are dependent on specialized care. In

fact, the average amount of pensioners with substan-

tial care needs has increased threefold between 1995

and 2017 (Bundesministerium f

ur Gesundheit, 2018).

This phenomenon is not restricted to Germany. The

worldwide number of the elderly – those aged 60

years or over – is expected to more than double by

2050 and triple by 2100 (United Nations, 2017). Ag-

ing leads to a gradual decrease in physical and mental

capacity, a growing risk of disease, and further depen-

dence on nursing care (World Health Organization,

2018).

Today, acquiring information on nursing care ser-

vices in Germany is a laborious undertaking. Pa-

tients looking for a nursing care facility can com-

municate directly with it or consult with a nursing

care support center to obtain information on avail-

able care spots. Both alternatives are associated with

a signiﬁcant time expenditure as important informa-

tion such as availability must be explicitly requested,

often by phone. At the policy level, social plan-

Column Store

Lightweight

Compression

Partitioning

Multi-Core

Parallelization

OLAP Cubes

Geo DataClaims Data

Governmental

Data

Facility Data

Behavioral

Data

Facilities Waiting list Prognoses

Application

Layer

Data

Layer

Family

members

Social

workers

Facility

managers

Social

planners

R R

Master Data Transactional Data

Information

Reporting

Visualization

Tool

Incorporated

Data Sources

Figure 1: Software system architecture of the elderly care

information and communication platform depicted as a

FMC block diagram.

ners aim to improve a district’s nursing care offering

through demand forecasting and capacity planning.

This requires comprehensive up-to-date datasets com-

ing from different sources. Currently, these sources

are either scattered throughout different silos, thereby

presenting high heterogeneity, or are simply not avail-

able outside the respective institutions.

Konak, O., Freitas Da Cruz, H., Thiele, M., Golla, D. and Schapranow, M.

An Information and Communication Platform Supporting Analytics for Elderly Care.

DOI: 10.5220/0007717801970204

In Proceedings of the 5th International Conference on Information and Communication Technologies for Ageing Well and e-Health (ICT4AWE 2019), pages 197-204

ISBN: 978-989-758-368-1

 2019 by SCITEPRESS – Science and Technology Publications, Lda. All rights reser ved

197

A centralized software platform with reliable in-

formation on offer and demand for nursing care ser-

vices can help to close the gaps between silos. Such

an integrated platform can also be of interest for re-

lated actors, such as social workers and facility man-

agers. Fig. 1 shows the architecture of our software

solution in a Fundamental Modeling Concepts (FMC)

block diagram. The foundation is formed by data

coming from different sources, which are either pre-

deﬁned (master data) or continuously updated (trans-

actional data). The master data is only updated peri-

odically, consisting of governmental data, geograph-

ical data, and insurance claims data. In contrast,

transactional data is continuously updated, originat-

ing from nursing care facilities and users’ search be-

havior. The different sources are integrated and har-

monized in an In-Memory Database (IMDB), which

has been demonstrated to support ﬂexible and fast

analysis of extensive amounts of data (Plattner and

Schapranow, 2014). The application offers a wide

range of services such as a personalized waiting

list, recommendations for nursing services matching

given search criteria, ﬁne-grained data exploration,

and prognosis on different key performance indica-

tors (KPIs). The requirements of our software were

elicited following principles from design thinking, a

user-centric approach. The platform standardizes in-

formation exchange across the nursing care spectrum,

making it possible for the ﬁrst time to tap into exist-

ing data silos. It therefore establishes the basis upon

which to perform advanced analytics, ultimately en-

abling users to derive recommendations that inform

policy and to reduce the time needed to ﬁnd appropri-

ate care services.

The remainder of this work is structured as fol-

lows: Sect. 2 places the work in the context of already

existing initiatives in the ﬁeld. In Sect. 3 the software

architecture and the technical infrastructure are dis-

cussed. Our speciﬁc contributions are described in

Sect. 4, while they are critically analyzed in Sect. 5.

The paper concludes with an outlook on the next steps

in Sect. 6.

2 RELATED WORK

Decentralization of care delivery is currently gaining

more importance (Real et al., 2018). Information and

communication technologies in nursing care thus shift

towards decentralized networking. Honor (Honor,

2018) and careship (Careship, 2018) are two exam-

ples of this development. The idea is to bring together

people in need of care with suitable caregivers from

the neighborhood. This strengthens local care. Nev-

ertheless, compared to our platform, it does not cover

people who can no longer stay in their own home and

are dependent on a care facility.

Other platforms, such as AOK Pﬂege-

Navigator (AOK, 2018), Wohnen im Alter (Wohnen

im Alter, 2018) and Seniorplace (Seniorplace, 2018)

offer a complete overview of care facilities in the

search region and also provide additional information

on the facilities. Unfortunately, these platforms lack

information about capacity utilization. In contrast to

our solution, this information has to be requested.

The focus on a speciﬁc user group is another dis-

advantage of these approaches. In order to improve

the entire process of care, there is no holistic ap-

proach. All approaches concentrate only on one spe-

ciﬁc user group or the interaction of two user groups,

e.g. carers and people in need of care. On our plat-

form, on the other hand, the four actors care recipient,

nursing facility, social planner and social worker are

linked.

Linking the actors further leads to a homogeniza-

tion of data. This data collection and process stan-

dardization of information exchange in association

with the IMDB technology facilitates individual real-

time analysis, consistency, scalability and leads to

more transparency (Knawy, 2017). Current solutions

in this area, as in the case of Recom (RECOM, 2018),

only take place within a closed facility, such as a hos-

pital or a care facility.

The federal ministry of education and research in

Germany is promoting aging in place through the pri-

macy of outpatient rather than inpatient care (Bun-

desministerium f

ur Bildung und Forschung, 2017).

Ambient assisted living and e-health are current

methods that intend to promote this (Cedillo et al.,

2018). Still, people reliant on care facility is in-

creasing (Destatis, 2015). In order to meet increased

demand, social planning derives prognosis based on

static and aggregated data at a district or state level.

Short-term changes due to new technologies and pop-

ulation migration are not absorbed by this approach

and thus make it cumbersome for changes. In com-

parison, our centralized approach provides analysis of

real-time data even at a municipality level.

3 METHODS

In the following, we share details about the utilized

research methodology. To get an overall picture of

the existing challenges in the ﬁeld of nursing care

we used the design thinking approach. Based on its

guidelines, we elicited the software requirements via

subject-matter interviews with the identiﬁed stake-

ICT4AWE 2019 - 5th International Conference on Information and Communication Technologies for Ageing Well and e-Health

198

holders. To implement these requirements we made

use of selected IMDB building-blocks and web de-

velopment technologies.

3.1 Requirements Engineering

In order to understand the whole process of decision-

making in the nursing care ﬁeld, we followed the

problem-solving method design thinking (Lockwood

and Papke, 2017). In interviews with subject-matter

experts, we were able to pinpoint problems and chal-

lenges, e.g. concerning how they are connected to

each other and how they communicate. This enabled

us to derive a concise representation of the stakehold-

ers in terms of personas and user stories. Personas are

a generic entity that represents a typical user of the

solution. In particular, four personas have been iden-

tiﬁed. The ﬁrst is the family member who is looking

for an appropriate nursing care facility as soon and as

easily as possible. The second is the social planner,

responsible to carry out data analyses and draw up the

elderly care plan, which contains a number of recom-

mendations for policy making. The third is the social

worker in a nursing care support center, in charge of

counseling people on several topics related to nurs-

ing care. Finally, we identiﬁed the facility manager,

who needs to inform prospective patients on available

spots, submit key indicators to district administration

and ultimately optimize occupancy rates. A user story

helps to create a simpliﬁed description of a require-

ment (Cohn, 2004). A user story template often uses

the following format: ’As a Persona, I want feature

so that reason’. This led to a collection of user stories

that have guided the implementation of the platform.

They laid out the foundation for a set of functional and

non-functional requirements, which can be examined

as Supplementary Material

as per the ISO/IEC/IEEE

24765:2017 (International Organization for Standard-

ization, 2017).

3.2 IMDB Technology

Following the users’ non-functional requirements for

our software platform such as fast response times and

scalability, we decided to utilize the IMDB technol-

ogy. It uses the main memory for storage and enables

real-time data processing and analysis of the stored

data (Prassol, 2015). Increasing computing power and

recent advances in hardware allows real-time analysis

of massive amounts of data (Plattner and Schapranow,

2014). Our platform makes use of selected building-

blocks of the IMDB technology, which are described

in the following.

https://tinyurl.com/y4vt8b86

Column-oriented Data Layout. In classical rela-

tional databases data is stored in a row-oriented for-

mat (Halpin and Morgan, 2008). This characteris-

tic makes this approach particularly suitable for on-

line transactional processing (OLTP) systems. Row-

oriented data storage allows fast writing of data, while

retrieval takes longer. In contrast, column-oriented

databases are less able to cope with this frequent writ-

ing (Ordonez et al., 2018). However, the columnar

structure is often used in online analytical process-

ing (OLAP) systems. Since user queries are limited

to only a subset of the available attributes, for exam-

ple when a given district wants to access its KPIs, our

platform uses only columnar tables. Another advan-

tage of columnar format is that the data can be stored

in highly compressed form.

Lightweight Compression. Compression is tar-

geted at retaining the information content of a data

structure with a concomitant reduction of disk space

needed to represent it (Plattner, 2014). Column-wise

storage eases lightweight compression given that cer-

tain patterns in the data are often repeated (Svens-

son, 2008). As our platform shall cater to a wide

user base, acceptable performance must be ensured.

Performance gains can be enabled either by cache-

conscious algorithms (Rao and Ross, 1999) or by re-

ducing the amount of data transferred from and to

main memory, which can be achieved by data com-

pression approaches (Westmann et al., 2000).

Partitioning. Partitioning can be either vertical or

horizontal (Lightstone et al., 2007). Vertical parti-

tioning splits columns of database tables into multiple

column-wise subsets, which then can be distributed

on individual databases servers (Hellerstein et al.,

2007). In contrast, horizontal partitioning handles

large database tables by dividing them into smaller

pieces of row-wise data. This approach has the ad-

vantage that it enables parallel search operations and

improves scalability (Plattner, 2014).

Multi-Core Parallelization. Trends in multi-core

processing with multiple cores in each case can be

exploited by reducing the amount of sequential work

and developing parallized applications in order to

achieve highest processing speed. The IMDB system

used for our platform enables inter and intra operator

parallelism (Plattner, 2014). The IMDB handles par-

allelization autonomously so that we do not have to

explicitly manage it (Plattner and Leukert, 2015).

An Information and Communication Platform Supporting Analytics for Elderly Care

199

3.3 Web Technologies

For the development, we utilized the Model-View-

Controller (MVC) architecture. It is an architectural

pattern used to efﬁciently link the user interface to the

underlying data models (Goodrich and Lenz, 2016).

It separates the application into the components mod-

els, controllers and views. Separating our software

platform into these components leads to improved

scalability, ease of maintenance and reusability. In

particular, we used the Ruby on Rails framework

v3.2 along with the JQuery user interface library.

The geographical exploration of care services and the

overview of care utilization was developed with the

open-source JavaScript library Leaﬂet.js (Vladimir

Agafonkin, 2018).

3.4 User Experience Evaluation

For a ﬁrst assessment of this platform, we conducted

a survey for users in the role of a family member. A

survey was carried out among people with different

socio-demographic characteristics. In order to eval-

uate how well family members are supported by the

platform, they were asked to ﬁnd a nursing care facil-

ity matching pre-deﬁned criteria. Subsequently, they

answered a survey on the usability of the platform.

A total of 42 subjects took part in the survey, out of

which 36 gave an answer. Questions were answered

using a Likert scale with grades from ”not applica-

ble” (1) to ”fully applicable” (5).

4 CONTRIBUTIONS

In the following, we detail the contributions we made

concerning software requirements, software architec-

ture, incorporated data sources, and process standard-

ization.

4.1 Software Requirements

As per Sect. 3.1, the interviews conducted with the

different stakeholders formed the basis for a set of

user stories, from which we derived functional and

non-functional requirements.

4.1.1 Functional Requirements

Our platform provides the needed foundation for dif-

ferent stakeholders to exchange information, with

each user group having different sets of needs, which

might overlap. In the following, we brieﬂy outline

the identiﬁed functional requirements for each user

group.

Family Members. This user group is primarily

concerned with ﬁnding a suitable care spot which

meets speciﬁc criteria. Among others, geographi-

cal proximity is a key factor in the decision-making

process. Therefore, the platform shall enable fam-

ily members to deﬁne advanced search criteria and

explore the results with a geographical visualization

front-end. Since the desired facility is not always

available, family members shall be able to deﬁne

search agents which can trigger e-mail notiﬁcations

upon available spots matching the criteria deﬁned.

Furthermore, the platform shall provide centralized

waiting lists across all facilities.

Social Planners. Tasked with providing informa-

tion for policy-makers in the ﬁeld of nursing care

planning, social planners’ primary aim is access to up-

to-date KPIs on important developments in the ﬁeld,

concerning, e.g., availability of nursing care offers,

personnel infrastructure, occupancy rates, and demo-

graphic trends, including prognoses. These KPIs are

currently generated by means of manual data collec-

tion, i.e. via questionnaires directed at care facili-

ties, a time-consuming and error-prone task. There-

fore, the platform shall provide the infrastructure for

continuous data collection to ensure actuality. In ad-

dition, the platform shall enable the social planner to

explore available KPIs in an explorative fashion, so as

to enable ad-hoc queries to be answered. An overview

of selected KPIs can be examined in Tab. 1.

Social Workers. These stakeholders help family

members and patients in all questions related to nurs-

ing care, particularly assistance in ﬁnding a suitable

nursing care offer. In addition to the same services

also provided to family members, the platform shall

provide social workers with detailed prognoses on the

availability of a given nursing care offer in a given

district.

Facility Managers. These stakeholders often need

to answer information requests from social planners,

regulatory agencies, governmental entities, and fam-

ily members. While answering a phone call for ﬁve

minutes may not seem like a signiﬁcant amount of

time taken individually, collective they considerably

burden facility managers and employee who should

otherwise be engaged in patient care. To stream-

line this process, the platform shall enable facility

managers to report information required by the other

ICT4AWE 2019 - 5th International Conference on Information and Communication Technologies for Ageing Well and e-Health

200

stakeholders in a standardized fashion, e.g., regarding

service offering, available spots, personnel infrastruc-

ture, etc.

4.1.2 Non-functional Requirements

Non-functional requirements of the platform include

the same set characteristics which users expect from

modern web applications, concerning ease of use, fast

response time and device independence, i.e., ability to

use the software both from desktop and mobiles de-

vices. Additionally, the possibility to easily extend

the available KPIs shall be guaranteed by the plat-

form.

Table 1: Excerpt of KPIs. Abbreviations: SP=Social

Planner; FM=Family Member; FMa=Facility Manager;

SW=Social Worker.

Key Performance Indica-

tor

SP FM FMa SW

Care Utilization

x x x

Available Capacity

x x x x

Days of Capacity Over-

load

x x

Number of People in Care

Facility

x x

Covered Days

x x x

Number of Nursing Staff

x x

Nursing Staff Ratio

x x x

Nursing Staff Need

x x

Waiting Time per Care

Form

x x x x

Demand on Care Place

x x x

Care Migration

x x x

Forecast on KPIs

x x x

4.2 Software Architecture

The elicited requirements were mapped into soft-

ware components (application layer) within a web

platform built upon an IMDB database infrastructure

(data layer) which harmonizes the incorporated data

sources. The resulting software architecture is de-

picted in Fig. 1. In the following, we provide a de-

scription of each of the individual software layers in

top-down fashion.

Application Layer. The application components

were built using the MVC architectural pattern, en-

capsulating component responsibilities in ’Facility’,

’Waiting List’, ’Prognoses’, ’Information Reporting’

and ’Visualization Tool’. The ’Facility’ component

manages facility information, geographical explo-

ration, search capabilities, including search agent (F1-

F5). A core functionality of the platform resides

in the ’Waiting List’ component which manages re-

quests for nursing care offers in a centralized fash-

ion across different facilities (F6). The KPIs avail-

able on the platform can be forecast by means of the

’Prognoses’ component (F7). This component uti-

lizes analysis routines available in the IMDB, such as

time-series analysis. Facilities must provide differ-

ent stakeholders with information regarding available

spots, personnel availability, among others. This is

enabled by the ’Information Reporting’ component,

with which facilities can avoid having to answer in-

quiries by phone or e-mail manually (F8-F10). The

’Visualization Tool’ enables interactive exploration of

KPIs by means of an external software tool that is able

to consume OLAP cubes (F11-F13). Since the visual-

ization tool is directly embedded in the platform, it is

transparent for the user. Finally, security is provided

by means of an authentication and authorization role-

based access control components, implemented with

the Ruby gems ’devise’ and ’can-can’.

Application frontend was developed using

HTML5 and Javascript with JQuery using principles

of web usability in order to enable ease of use (NF1).

In particular, for the geographical exploration com-

ponent, we made applied caching strategies using the

user browser’s embedded off-line storage, allowing

to cache network requests and thereby avoiding

round trips to the web server, ensuring an interactive

web application (NF2). This is particularly relevant

with regards to shape data (GeoJSON) required to

display municipalities on a map. Furthermore, the

platform adapts to different devices ﬂexibly, desktop

or mobile, providing device-independence (NF3).

Data Layer. For the data layer, we utilized the

IMDB building blocks column-oriented storage,

lightweight compression, partitioning, and multi-core

parallelization and OLAP cubes. In our platform,

we utilize column-based storage for the data entities.

This approach storage is beneﬁcial both from a com-

pression perspective and for analytical purposes. Be-

sides, for database design, we strived to achieve the

third normal form for all entities, futher improving

disk utilization (Bernstein, 1976). This was imple-

mented, e.g., to keep track of available care spots

across all service types for a given facility. Further-

more, the partitioning and multi-core parallelization

capabilities of the IMDB enables data belonging to

different districts to distributed across different server

nodes, optimizing response times (NF2). To enable

data exploration, we developed virtual OLAP cubes

An Information and Communication Platform Supporting Analytics for Elderly Care

201

which meet the analysis needs of social planners. Fur-

thermore, virtual OLAP cubes provide the possibility

of extensibility by adding new measures or dimen-

sions without the need to recreate datamarts, as is

usual with traditional Business Intelligence systems,

ensuring ease of extensibility for KPIs (NF3).

Incorporated Data Sources. An important contri-

bution of our platform is integrating different data

sources within a central IMDB infrastructure. The in-

corporated data sources differ in the frequency with

which they are updated, termed either master data

(updated periodically) or transactional data (updated

daily). Belonging to the former category, are data

from health insurance claims, geographical data co-

ordinates for districts and municipalities (e.g. GeoJ-

SON), and statistical data, such as demographics for

a given region, which are typically available from

governmental sources. Transactional data sources

relate to data provided by the care facilities them-

selves, concerning real-time occupancy rates, i.e.,

how many spots are currently available for a given ser-

vice, and personnel infrastructure. Furthermore, the

platform can gather non-sensitive information about

users queries, e.g. geographical origin, services re-

quested, waiting times, etc. Regarding, the claims

data, we cooperated with a large German health in-

surance company which enabled us to obtain ﬁne-

grained information on nursing care utilization (Fre-

itas da Cruz et al., 2018).

4.3 Process Standardization

The various user interviews helped us to identify the

gap in the communication channels between the ac-

tors involved in nursing care. All four identiﬁed users

are connected to each other directly or indirectly.

Hence, all actors interact with each other by infor-

mation exchange and are part of the decision-making

process of all individual actors.

Family

Member

Social

Planer

Facility

Manager

Social

Worker

Family

Member

Social

Planer

Facility

Manager

Social

Worker

Figure 2: Image depicts the intersection of the different user

groups: disjoint ellipses with dashed lines (left) vs. joint

ellipsis with solid lines (right). The platform joins the user

groups in order to save time and share common data rather

than manually built bilateral communication channels.

As shown in Fig. 2 on the left, the family mem-

ber e.g. is in direct contact with the care facility and

nursing care support center for information-gathering

about the facility. However, they are in indirect con-

tact with the social planner. The social planner is in-

terested in relevant KPIs on nursing care in order to

derive recommendations from them. The collected

KPIs depend not only on information coming from

the direct interface with the facilities. They also de-

pend on the search queries of the family members.

The ﬂow of information between the facilities and

all other actors takes place on an individual level.

While family members’ search queries are always

answered by telephone or e-mail, queries coming

from the ministry via social planner about personnel

and care infrastructure are sent via non-standardized

channels and in individual consultation. The entire

data exchange between the individual actors does not

take place in a given format but after bilateral agree-

ment. With the help of our platform, as shown in

Fig. 2 on the right, the entire communication chan-

nel is standardized and all information exchange is

taking place at one single point. Standardization of

the process improves and promotes quality and saves

time (Knawy, 2017).

5 EVALUATION AND

DISCUSSION

Requirements imposed on the platform were met as

follows. From a functional perspective, the user is

provided with a geographical exploration tool of all

available care facilities in order to ﬁnd the most suit-

able one in real-time without the need for additional

help (F1-F5). A centralized waiting list helps both the

care recipient and the nursing care facility to stream-

line the process of ﬁnding a suitable care spot (F6).

In the future, we will provide linear regression mod-

els to deliver forecasts on the expected waiting time

and ﬁne-grained long-term forecasts on the demand

for nursing care spots (F7). Data harmonization and

process standardization are achieved by means of di-

rect communication channels and minimum informa-

tion reporting standards (F8-F10). Consistent infor-

mation reporting from care facilities results in ad-

ditional KPIs (F11-F13), which allow more precise

analyses and prognoses in social planning.

Our platform fulﬁlled non-functional require-

ments, such as ease of use, through the use of HTML5

and Javascript with JQuery (NF1). The incorporated

IMDB technology ensures fast response time (NF2)

and extensibility (NF4). Furthermore, the platform

can be accessed from different devices (NF3).

ICT4AWE 2019 - 5th International Conference on Information and Communication Technologies for Ageing Well and e-Health

202

In its current state, our platform has a number

of advantages over existing solutions. It supports

family members with additional information on nurs-

ing care facilities and thus helps them to save time.

In addition to the availability of real-time informa-

tion, the centralized collection and integration of var-

ious user groups and data sources, such as insurance

claims data, geo-information data, and statistical data

coming from governmental bodies, provides a suit-

able measuring instrument for KPIs. This data collec-

tion and process standardization paired with the used

IMDB technology facilitates the analysis and forecast

for social planning.

The survey presented in Tab. 2 suggested that the

platform is user-friendly and useful in ﬁnding care fa-

cilities and services. However, not all functionalities

needed were found by the users. For instance, the

possibility of entering a place name as a search cri-

terion was missing. Currently, the search function-

ality for ﬁnding a care spot is constrained to postal

code. A further question which we sought to answer

by the survey was if the platform provides better sup-

port than existing solutions from the users perspec-

tive. However, only one person stated to had experi-

ence with a similar solution, the response was consid-

ered as not sufﬁciently informative.

The approach chosen to evaluate the application

relied on a small sample size, thereby potentially pro-

viding biased the results. To address this limitation, a

more comprehensive user evaluation involving other

users and a wider sample is expected to be carried out

in the future when new features are implemented.

Table 2: Results of the survey of family members. Ques-

tions were answered using a Likert scale with grades from

”not applicable” (1) to ”fully applicable” (5). This can be

used to calculate a weighted mean.

Question

Weighted

Mean

I got along with the platform pretty well.

4.14

I got along with the map pretty well.

4.27

I was able to handle the ﬁlter settings.

4.14

I understood the icons on the map.

3.89

I have found all the functionality I need.

3.23

I like the information on the available ca-

pacity of the nursing care facility.

4.14

I felt supported by the platform.

4.14

I would like to use the platform when

searching for care places and offers.

4.11

6 CONCLUSION AND FUTURE

WORK

With the help of our software architecture, process

integration, and technology infrastructure, we devel-

oped a platform, that facilitates the searching process

for a care facility and helps to derive more detailed

analyses in order to improve the care offer of nursing

facilities. The data foundation is comprised of mas-

ter data by various institutions and transactional data

generated by the incorporated user groups. The soft-

ware architecture allows an easy extension for future

work, e.g. ﬁnding a kindergarten or school place.

Future work concerns deeper analysis of social

planning, incorporation of new user groups and new

proposals to try different methods. The following

ideas could be tested: 1) As we used linear regres-

sion models to forecast the waiting time and the de-

mand for nursing care offers, the application of more

sophisticated machine learning models could help to

improve the forecast quality. 2) It could be interesting

to include nursing care as a service. Offering help on

an outpatient basis leads to a larger data collection and

can thus further improve social planning. 3) Includ-

ing clinical data could also help to enhance demand

forecast accuracy.

ACKNOWLEDGEMENTS

Parts of this work were generously supported by a

grant of the German Federal Ministry of Economic

Affairs and Energy (01MD15005).

REFERENCES

AOK (2018). Pﬂege-Navigator - Suchmaske der Pﬂegeein-

richtungssuche. https://www.pﬂege-navigator.deOct

20, 2018.

Bernstein, P. A. (1976). Synthesizing Third Normal

form Relations from Functional Dependencies. ACM

Transactions on Database Systems (TODS), 1(4):277–

298.

Bundesministerium f

ur Bildung und Forschung (2017).

Altere Menschen. https://www.gesundheitsforschung-

bmbf.deOct 20, 2018.

Bundesministerium f

ur Gesundheit (2018). Pﬂege. https://

www.bundesgesundheitsministerium.de/themen/pﬂe

ge/pﬂegeversicherung-zahlen-und-fakten.htmlOct 20,

2018.

Careship (2018). Seniorenbetreuung & liebevolle Pﬂege f

Senioren. https://www.careship.deOct 20, 2018.

Cedillo, P. et al. (2018). A Systematic Literature Review

on Devices and Systems for Ambient Assisted Living:

An Information and Communication Platform Supporting Analytics for Elderly Care

203

Solutions and Trends from Different User Perspec-

tives. In 2018 International Conference on eDemoc-

racy & eGovernment (ICEDEG), pages 59–66. IEEE.

Cohn, M. (2004). User Stories Applied: For Agile Soft-

ware Development. Addison-Wesley signature series.

Addison-Wesley.

Destatis (2015). Pﬂege. https://www.destatis.de/

DE/ZahlenFakten/GesellschaftStaat/Gesundheit/Pﬂe

ge/Pﬂege.htmlOct 20, 2018.

Freitas da Cruz, H. et al. (2018). Interactive Data Explo-

ration Supporting Elderly Care Planning. eTELEMED

2018, page 68.

Goodrich, G. and Lenz, P. (2016). Rails: Novice to Ninja:

Build Your Own Ruby on Rails Website. SitePoint.

Halpin, T. A. and Morgan, T. (2008). Information Mod-

eling and Relational Databases. Morgan Kaufmann

Publishers.

Hellerstein, J. M. et al. (2007). Architecture of a Database

System, Foundation and Trends in Databases, vol-

ume 1. Now Publishers.

Honor (2018). In-Home Care Your Family Will Love.

https://www.joinhonor.comOct 20, 2018.

International Organization for Standardization (2017).

ISO/IEC/IEEE 24765:2017 - Systems and Software

Engineering – Vocabulary. https://www.iso.org/ stan-

dard/71952.htmlOct 20, 2018.

Knawy, B. (2017). Leading Reliable Healthcare. Taylor &

Francis.

Lightstone, S. et al. (2007). Physical Database Design: The

Database Professional’s Guide to Exploiting Indexes,

Views, Storage, and more. Morgan Kaufmann.

Lockwood, T. and Papke, E. (2017). Innovation by Design:

How Any Organization Can Leverage Design Think-

ing to Produce Change, Drive New Ideas, and Deliver

Meaningful Solutions. Red Wheel Weiser.

Ordonez, C. et al. (2018). A Survey on Parallel Database

Systems from a Storage Perspective: Rows Versus

Columns. In Database and Expert Systems Applica-

tions, pages 5–20, Cham. Springer International Pub-

lishing.

Plattner, H. (2014). A Course in In-Memory Data Manage-

ment: The Inner Mechanics of In-Memory Databases.

Springer.

Plattner, H. and Leukert, B. (2015). The In-Memory Revolu-

tion: How SAP HANA Enables Business of the Future.

Springer Publishing Company, Incorporated.

Plattner, H. and Schapranow, M.-P., editors (2014). High-

Performance In-Memory Genome Data Analysis:

How In-Memory Database Technology Accelerates

Personalized Medicine. Springer-Verlag.

Prassol, P. (2015). SAP HANA als Anwendungsplat-

tform f

ur Real-Time Business. HMD Praxis der

Wirtschaftsinformatik, 52(3):358–372.

Rao, J. and Ross, K. A. (1999). Cache Conscious Index-

ing for Decision-Support in Main Memory. Technical

report.

Real, K. et al. (2018). Using Systems Theory to Ex-

amine Patient and Nurse Structures, Processes, and

Outcomes in Centralized and Decentralized Units.

HERD: Health Environments Research & Design

Journal, 11(3):22–37.

RECOM (2018). Wissen f

ur das Gesundheitswesen.

https://www.recom.euOct 20, 2018.

Seniorplace (2018). Wohn- und Pﬂegeangebote.

https://www.seniorplace.deOct 20, 2018.

Svensson, P. (2008). The Evolution of Vertical Database

Architectures – A Historical Review. In Proceed-

ings of the 20th Int’l Conf on Scientiﬁc and Statistical

Database Management, pages 3–5. Springer-Verlag.

United Nations (2017). World Population Prospects.

https://population.un.org/wpp/publicationsOct 20,

2018.

Vladimir Agafonkin (2018). Leaﬂet - A Javascript Li-

brary for Interactive Maps. https://leaﬂetjs.comOct

20, 2018.

Westmann, T. et al. (2000). The Implementation and Per-

formance of Compressed Databases. SIGMOD Rec.,

29(3):55–67.

Wohnen im Alter (2018). Das Onlineportal f

ur Pﬂege

& Wohnen im Alter. https://www.wohnen-im-

alter.deOct 20, 2018.

World Health Organization (2018). Ageing and Health.

http://www.who.int/news-room/fact-sheets/detail/

ageing-and-healthOct 20, 2018.

ICT4AWE 2019 - 5th International Conference on Information and Communication Technologies for Ageing Well and e-Health

204