Journal of Nursing Management, 2007, 15, 683692

Benchmarking in nursing care by the RAFAELA patient

classification system a possibility for nurse managers
M
LISBETH FAGERSTRO
1

PhD, RN, Dean1

and AUVO RAUHALA

MD2

Associate Professor and Head of the Conservative Division, Vasa Central Hospital, Vasa, Finland

Correspondence
Lisbeth Fagerstrom
Svenska Yrkeshogskolan University
of Applied Sciences
Seriegatan 2
Vasa 65320
Finland
E-mail: lisbeth.fagerstrom@syh.fi

F A G E R S T R O M L . & R A U H A L A A . (2007) Journal of Nursing Management 15, 683692

Benchmarking in nursing care by the RAFAELA patient classification
system a possibility for nurse managers

Aim The aim of the study was to explore the possibilities of benchmarking with the
RAFAELA system. In this study, comparisons are made between: (1) costs for
one nursing care intensity point; (2) the nursing care intensity per nurse; (3) the
relationship between nursing care intensity per nurse and (4) the optimal nursing
care intensity.
Background During the period from 1994 to 2000 a new system for patient
classification, the RAFAELA system, was developed in Finland.
Methods 86 wards from 14 different hospitals in Finland took part in the study.
Results The costs for one nursing care intensity point on the adults wards were on
average 7.80. The average workload was 25.2 nursing care intensity points per
nurse. The optimal nursing care intensity was exceeded during 49.5% of the days
and under during 20% of the days.
Conclusions The study shows that benchmarking with the RAFAELA system
provides many opportunities for the nurse managers resource allocation and their
personnel administration.
Keywords: benchmarking, nursing care intensity, patient classification system, personnel
staffing and scheduling
Accepted for publication: 25 August 2006

Introduction
Benchmarking is a fashionable word within many
branches today, but within nursing care and nursing
research, systematic benchmarking is fairly undeveloped and has not been researched. The basic idea with
benchmarking as a method for developing an organization is to learn from others and preferably from the
best of the organizations that can show the best results
within the area (Camp 1993, Kaivos et al. 1995). What
could be a good resource allocation within nursing
care?
During the period from 1994 to 2000 a new patient
classification system (PCS), the so-called RAFAELA
system, was developed in Finland. Already at the end
DOI: 10.1111/j.1365-2934.2006.00728.x
 2007 The Authors. Journal compilation 2007 Blackwell Publishing Ltd

of the 1990s, the RAFAELA system had spread to

approximately 15 hospitals. To standardize, test the
credibility of the system, the Association of Finnish
Regional and Local Authorities decided to start a
large research project (Fagerstrom & Rauhala 2001).
One important aim during the project time was to
develop and commence a systematic benchmarking
activity within nursing care (Fagerstrom & Rauhala
2003). Today, information is collected biannually
from the units participating (about 15 hospitals and
250 wards) in the national benchmarking system,
according to the concept presented in this study.
Benchmarking reports are made and are then
available for the users, executives and politicians on
different levels of the organizations.
683

L. Fagerstrm and A. Rauhala

The interest for careful monitoring of costs and personnel resources has increased and the need for computer systems for cost and activity monitoring is also
growing (Ruland & Ravn 2003). The patients need of
good care and fair resource allocation are current
international questions (Needham 1997, Partanen
2002). In Finland, the benchmarking activity started
within nursing care for the elderly with the Resident
Assessment Instrument (RAI) system, which has been
developed for the planning and evaluation of long-term
nursing care (Noro et al. 2001). This benchmarking
concept includes patient-related information, for
example, patient structures, dementia, depression,
ability to function, quality indicators as well as information about costs and personnel recourses for each
unit.
In this study, a concept will be presented for
systematic and national benchmarking that has
been developed for secondary health care with the
help of the RAFAELA system in Finland. Results
from a sub-study concerning resource allocation with
the help of the RAFAELA system will also be presented here.

The RAFAELA system

The RAFAELA system consists of three parts: (1)
patients nursing care intensity (NCI) measured daily by
the Oulu Patient Classification/Qualisan (OPCq) and
(2) the daily nursing resources that have directly or
indirectly been allocated to patients nursing care. By
using these two sets of data, a measure of nurses
workload as Nursing Care Intensity points per nurse
(NCIP/N) is obtained. (3) The optimal level of NCI
workload is then established simultaneously using the
Professional Assessment of Optimal Nursing Care
Intensity Level for several weeks (PAONCIL; Fagerstrom et al. 2000, Rauhala & Fagerstrom 2004). The
basic idea of the RAFAELA system is that the workload
expressed as NCI/N is compared with the optimal NCI
level for the ward. For a more detailed description of
the system earlier research is recommended for the
reader.
The validity of the OPCq instrument has been tested
several times, with good results. Content validity has
been tested using expert panels of nurses (Fagerstrom
2000). Construct validity has been tested by using
PAONCIL method as a gold standard (Fagerstrom
et al. 2000). A patient perspective has also been used in
validity testing (Fagerstrom et al. 1999). The perquisites
for achieving reliable results have been determined
(Rauhala & Fagerstrom 2004). The RAFAELA system
684

has now been developed to include also nursing care on

outpatients departments, psychiatric nursing care and
primary health care.
The RAFAELA system has been developed based on a
holistic view of human, a view of leadership based on
human resource management and on the idea that
nursing care consists of complex nursing care units
(Fagerstrom 2000).
Nursing care intensity is supposed to be a large part
of the nurses workload, but there are also many other
factors that simultaneously affect the nurses total
workload, for example, the organizing of work, skill
mix, organizational factors, the total workload
(Fagerstrom et al. 2002, Adams & Bond 2003, Rauhala
& Fagerstrom, unpublished results).

Benchmarking with the RAFAELA system

According to Karlof et al. (2000), the first phase of a
benchmarking process is a survey, which is followed by
a comparison between organizations. During the third
phase the aim was to try to understand the reasons for
the performance gaps between the organizations. The
fourth and last phase consists of improvements and
development to achieve better results. During recent
years, benchmarking as a method has developed into
benchlearning, i.e. the increase in knowledge taking
place by trying to learn from others is emphasized
(Pleijel 2004). According to this theory, every longlasting durable development must be based on increasing knowledge. Benchlearning can be understood as an
attitude, a learning approach. A learning organization is
seen to be a condition for efficiency and long-term
success (Karlof et al. 2000). Good examples or competition increase ambition and drive the learning process, which in turn pushes development forward and
encourages personnel to make better results (Karlof
2004).
Benchmarking as a process commences with the
identification of an area that could be improved and
finding critical indicators that can describe the
improvement area (Camp 1993, Karlof et al. 2000).
The primary aim of PCS has been and still was to be
able to estimate the need for personnel beforehand,
based on the patients caring needs (Giovannetti
1979). The basic idea in the RAFAELA system is that
the setting of personnel resources would be optimal,
i.e. the personnel resources and patients caring needs
are in balance and that there are realistic possibilities
to meet these needs (Fagerstrom 2000). PCS provides
the base for personnel planning, cost calculations and
analyses of the activities. These areas are central for

2007 The Authors. Journal compilation 2007 Blackwell Publishing Ltd, Journal of Nursing Management, 15, 683692

Benchmarking by the RAFAELA patient classification system

benchmarking activities and subsequently they are the

improvement areas within benchmarking.
The aim of systematic benchmarking, developed
through the RAFAELA system, was to produce exact
results concerning NCI, personnel resources and factors
influencing costs, in order to be able to compare the
productivity, quality and costs of nursing care between
organizations/units. The critical indicators consisted of
two larger data units, i.e. ward-related information and
information on NCI (see Table 1).
The above-mentioned information is submitted twice
a year to the unit/office on the national level that is
responsible for analyses and reporting. Benchmarking
reports are published biannually and are available on
the Internet for the units included in the benchmarking
activities (Fagerstrom & Rauhala 2003). The reports
present analysis of the above-mentioned indicators and
different combinations between the indicators. Comparisons between hospitals, specialized areas and wards
are presented in the form of both Tables and box-plot
Figures.

Aims
The overall aim of the study was to explore descriptively if benchmarking using the RAFAELA system can
be an administrative tool for nurse managers within
resource allocation, as the information from the abovedescribed benchmarking concept provides possibilities
for comparisons between specialized areas and between
hospitals (university hospitals/regional hospitals and
local hospitals).

How large are the variations in salary costs for the

NCI, measured in NCI points between special areas
and between hospitals?
Are there any differences in workload, measured in
NCIP per nurse between specialized areas and
between hospitals?
What is the relationship between NCI/N and the
optimal NCI level on the ward, i.e. how well have
the resource allocation and planning succeeded?

Sample
The study was performed within the frames of the
national research project Finnish Nursing Care Intensity benchmarking within nursing care (in Finnish:
Finnhoitoisuus hoitotyon benchmarking) and the data
were gathered from the period of 01.01.2001 to
31.12.2001. A total of 86 wards from 14 different
hospitals in Finland participated in the benchmarking
study. These 14 hospitals were divided into three different categories: (1) university hospitals (E, L, M); (2)
regional hospitals (A, B, C, I, K, N); (3) local hospitals
(D, F, G, H, J). Finland is divided into five university
hospital districts that are responsible for the highly
specialized and most expensive care. These districts are
then divided into 21 health care districts. Within each
health care district there is a regional hospital that is
responsible for the specialist-led health care. Many
smaller local hospitals have been closed down during
the past 10 years, mainly because of economic reasons,
as small units have been seen as expensive to run.
The following specialized areas were represented: dermatology (three wards), geriatric (one),

Table 1
Critical indicators for benchmarking with the RAFAELA system
Indicators
(A) Ward-related information
Type of activity/specialized area
The ward activity character
Central patient groups
Organization of the nursing care
Personnel employment structure/skill mix
Nursing personnel wage costs, both according
to budget and the accounts
Reliability percentage for nurses classifications
for each ward
(B) Information on NCI
The daily NCI
The NCI/N and per ward
The hospital stays related to each patient

Examples

Geriatric, surgery, neurology, paediatrics, internal medicine etc.

Beds per ward, opening hours, possible closures, weekday ward
Medical diagnoses
Module working, primary nursing
Head nurse, registered nurses, practical nurses, assisting personnel
Divided into regular salaries, salaries for extra personnel, extra compensations
>70%

NCI per patient and calendar day, NCI per areas of needs (AD),
the patient's sex, age, etc.
Optimal NCI level, exploratory power of the PAONCIL study, NCI/N
and per calendar day
DRG and length of the hospital stay

NCI, nursing care intensity; DRG, diagnosis-related group; PAONCIL, Professional Assessment of Optimal Nursing Care Intensity Level.
2007 The Authors. Journal compilation 2007 Blackwell Publishing Ltd, Journal of Nursing Management, 15, 683692

685

L. Fagerstrm and A. Rauhala

gynaecology-obstetric (five), surgery (11), neurology

(five), ophthalmology (one), oncology (four), orthopaedic-traumatology (seven), paediatrics (20), lung
diseases (five), rheumatology (five), internal medicine
(13) and mixed wards (six).
The participants in the study had to fulfil certain
criteria, i.e. daily use of the OPC, for at least
36 months, tested reliability through parallel classification (degree of agreement between two nurses
measurements of the same patient >70%) as well as
systematic registration of daily personnel resources. The
reliability levels of the participating wards were on
average 77%.

Data collection and data analysis

The data material from the 86 wards consisted of
576 883 NCI classifications for approximately 100 000
patients and 85 000 periods of hospital stay. Along
with NCI and associated data much additional data of
the characteristics of the wards, economic data, etc.
were collected (see Table 1).
The data collection took place during 01.01.2002
31.03.2002. Nurse managers were responsible for
compiling ward-related information concerning
personnel resources, ward activities and costs in an
electronic questionnaire in Excel format. The hospitals
IT-contact persons compiled the information concerning NCI in a specified text format, according to detailed
instructions. These above-mentioned persons were
assessed to be able to guarantee the best possible quality
of the data.
The adequacy of the raw data received were first
checked by the researchers both manually and using
diverse algorithms to detect missing and obviously
incorrect data. Later, during the analysis from April
2002 to August 2002 questions aroused concerning the
correctness of various parameters produced in every
hospital or ward. Thus, active communication between
the researchers and the contact persons of the participating hospitals and wards was needed to correct
missing or erroneous data.
Ethical guidelines for nursing research in the Nordic
countries published by The Northern Federation of
Nurses (1995) provided the guidelines for the planning
and implementation of the empirical study. To protect
patient confidentiality the data were treated entirely
anonymously.
The statistical program package S P S S for Windows
10.0 and the MS Excel 2000 spreadsheet program were
used to analyse the material. All data were compiled in
a large Excel table. The established parameters in the
686

benchmarking concept (see Table 1) were analysed

and calculated for each ward. These results were
then transferred to a statistic program called S P S S
for further analyses and the making of Tables and
Figures in the form of box-plots, in which the median
and spread according to percentage distribution can be
seen (25% and 75% percentiles).

Results
Presentation of wage costs for NCI
The information of each patients NCI and period of
hospital stay was combined with wage costs for each
ward. This made it possible to analyse carefully the
costs per NCIP. Wage costs in relation to the number of
beds were analysed and the average wage cost for one
bed was 30 028 on adults wards and 53 431 on
childrens wards. It was also possible to calculate the
wage costs for the nurses per patient according to the
benchmarking concept. On the adults wards the wage
costs per patient and calendar day were 96.50 and on
childrens wards these costs were more than double, i.e.
206.90.
The average costs for one NCIP on the adults wards
varied from 6.30 to 8.40, and the average was 7.80
(see Table 2). The average price for one NCIP on childrens wards was remarkably higher, i.e. 13.20, which
is explained through a much higher personnel density
on childrens wards compared with adults wards (see
Figure 1).
A cost comparison between the hospitals E, I and N
that were participating with 1315 wards shows
remarkable differences. At the university hospital E one
NCIP cost 9.60, when the regional hospitals I and N
have a cost of 8.90 and 7.20.
A closer analysis of the cost level for nursing care load
on a regional and local hospital level (B, G, H, J, K)
with 47 participating wards, gives us an average price
of 8.80 for one NCIP. The average price for one NCIP
at the participating university hospitals was 10.40, i.e.
18.2% higher costs for highly specialized nursing care.

Presentation of nurses workload in the form of

NCIP per nurse
The analyses show fairly large variations concerning the
workload of the nurses both between specialized areas
and between hospitals (see Figure 2). The NCI/N was
on average lowest on childrens wards (mean 15.1 for
paediatrics A, B and C; see Table 3) and highest on
mixed wards (27.1). These mixed wards can be found at

2007 The Authors. Journal compilation 2007 Blackwell Publishing Ltd, Journal of Nursing Management, 15, 683692

Benchmarking by the RAFAELA patient classification system

Table 2
The wage costs per nursing care intensity (NCI) point per specialized area and per hospital (AN), in
Specialized areas

Dermatology
Geriatrics
Gynaecology-obstetrics
Surgery
Neurology
Ophthalmology
Oncology
Orthopaedic-traumatology
Paediatrics A
Paediatrics B
Paediatrics C
Pulmology
Rheumatolgy
Mixed wards
Internal medicine
Mean

6.1

10.5

10.6

6.0
7.6
15.2

9.1

6.3

9.1

9.5
10.8
10.8
17.2

10.2

14.1

9.1

8.4

10.2
2.3

7.5

5.2
13.1

16.4
7.5

7.1
7.1

7.9
8.6

Mean

5.4

6.6

18.5

6.7
14.0
9.5

8.0

6.0
8.0

7.3
6.9
10.6

16.3
9.3
5.9
7.4

7.2
8.1
8.9

7.8
14.5

4.6

8.0

9.2

6.2
7.0
7.2

6.8
9.4

12.2

8.5
8.7
8.2
7.5
7.0
8.4
7.3
7.4
13.7
9.7
16.2
8.3
8.4
6.3
7.7
9.0

50

20

40

15
Mean NCI per nurse

The price of 1 NCI points,

9.5
9.5
8.6

7.4
8.5
9.6

8.9

6.7

8.3

J
10.1
8.7

8.4
6.7
8.4
5.2

14.1

10

30

20

5
10

n= 3

10 5

11

13

Derm gyn-obst neur oncol

pedA pedC reuma int med
ger
surg
oft
ortop-trau pedB
pulm mixed

Figure 1
The wage costs per nursing care intensity (NCI) point per specialized
areas, in (total number of wards 83).

smaller hospitals and they have patients from various

special areas, mostly within conservative medicine. The
average workload on adults wards was 25.2 NCIP/N.
Within internal medicine the NCIP/N was relatively
constant, the values varying from 24.0 to 28.9
(+20.4%). However, much larger variations could be
noticed within surgery, from 20.7 NCIP/N to 38.6
NCIP/N+ (86.5%).
A comparison between the hospitals E, I and N that
participated with 1315 wards shows that the variations in nurses workload were between 22.3 NCIP/N
and 25.1 (mean 23.9 NCIP/N). The workloads at university hospitals E and M were fairly close to each
other, i.e. 24.4 NCIP/N and 23.5 (mean 24). From the
third university hospital L, only paediatric wards

0
n= 3
A

6
B

1
C

2
D

10
E

1
F

4
G

4
H

14
I

4
J

7
K

2
L

7
M

13
N

Hospital

Figure 2
Mean nursing care intensity (NCI) per nurse, per hospital (78 wards).

participated and the workload there was remarkably

lower (12.9 NCIP/N). When making closer analyses
of the workload at regional and local hospitals (B, G,
H, J, K) that were participating with 47 wards, the
average for these was 24.5 NCIP/N. Subsequently, the
workload on the regional and local hospital level was
generally 10.8% higher than on the university hospital
level.

Presentation of NCI/N in relation to optimal NCI

levels
From a total of 86 participating wards, the optimal NCI
level had been decided for 53 wards (62%, see
Figure 3). The workload, expressed in the form of

2007 The Authors. Journal compilation 2007 Blackwell Publishing Ltd, Journal of Nursing Management, 15, 683692

687

L. Fagerstrm and A. Rauhala

Table 3
The mean nursing care intensity (NCI) per nurse, in points per specialized areas and per hospital (AN)
Specialized areas
Dermatology
Geriatrics
Gynaecology-obstetrics
Surgery
Neurology
Oncology
Ortopaedic-traumatology
Paediatrics A
Paediatrics B
Paediatrics C
Pulmology
Rheumatology
Mixed wards
Internal medicine
Mean

26.9
24.5

16.1

24.1

29.1

20.7

24.1

25.4

15.0
18.7
10.3

20.4
22.8
23.8

20.3
38.6

26.0

33.6
14.3

27.1

12.2
25.3

23.9

27.6
17.1

21.0

11.7

20.6

24.0
24.4

12.9

23.5

27.1
27.1

24.7
23.2

25.5

22.8
25.0
22.3

27.1
13.5
15.6

27.0
24.3
25.9
26.5
14.3

15.4
21.4
27.5
23.8

25.5
17.7
37.1
28.9

12.9

28.9
23.5

29.1
28.4
25.1

Mean
24.4
26.4
23.0
25.9
24.1
24.9
26.7
14.0
19.3
12.0
25.4
23.6
27.1
26.0
22.8

NCIP/N, was 25.2 on the adults wards. On the childrens wards the average was 15 NCIP/N. The optimal
NCI/N, which had been decided earlier with the help of
the PAONCIL-method, was on average 2027 NCIP/N
on the adults wards and 1320 on the paediatric wards.
The NCI/N was analysed in relation to the fixed
optimal NCI level of each ward and Figure 3 shows the
number of days the NCI/N exceeded the optimal NCI
level (its upper limit). The variations between wards
and hospitals are shown in Table 4. The situation on
the five neurological wards participating in the study
was extremely problematic (89%); the work situation
was also problematic for the nurses on the internal
medicine (65%) and on the mixed wards (62%). In
general, the optimal NCI/N was exceeded on the childrens wards on 39.7% of the days and on the adults
wards 49.5% of the days. Finally, it can be stated that

120

Number of days over optional level, %

28.6

11.7

22.5

20.0
26.4

25.8
19.2
24.8
24.3
12.7

100
80
60
40
20
0
20
5
6
3
3
6
7
2
2
1
3
4 10
n= 1
derm
surg
oncol
pedA pedC
reuma inter med
gyne-obst
neur orthop-trau pedB
pulm
mixed

Figure 3
Number of days in percentage when the nursing care intensity (NCI)
per nurse were above optimal level and per specialized areas (total
number of wards 53).

Table 4
Number of days in percentage when the nursing care intensity (NCI) per nurse were above optimal level, per specialized areas and per hospital
Specialized areas
Dermatology
Gynaecology-obstetrics
Surgical
Neurology
Oncology
Ortopaedic-traumatology
Paediatrics A
Paediatrics B
Paediatrics C
Pulmonary
Rheumatology
Mixed wards
Internal medicine
Mean

688

63

41
39

42

84

41

19
49
8
18

34

24
40
93

84

67

34
52

78
3

42
18

22
43

28

43
70
26

32

58

63

68
33
56

43

73
98
98

32
46

50

73
58

41

47

67
65
52

Mean
42
26
43
89
38
55
34
67
18
42
33
62
65
48

2007 The Authors. Journal compilation 2007 Blackwell Publishing Ltd, Journal of Nursing Management, 15, 683692

Benchmarking by the RAFAELA patient classification system

Table 5
Number of days in percentage when the nursing care intensity (NCI) per nurse were below optimal level, per specialized areas and per hospital
Specialized areas
Dermatology
Gynaecology-obstetrics
Surgical
Neurology
Oncology
Ortopaedic-traumatology
Paediatrics A
Paediatrics B
Paediatrics C
Pulmonary
Rheumatology
Mixed wards
Internal medicine
Mean

87

22

29
14
2

23
19

14

21
16

19

11

32
14
49
44

9
74

18
57

37
16

34

74

24
4
25

35

19

2
31
18

17

the NCI/N was above the recommended optimal level

during 48% of the days.
A corresponding analysis concerning the NCI/N in
relation to the lower limit of the optimal level shows
that the workload was below the optimal level on
average during 20% of the days (see Table 5). NCI/N
was below the optimal level only during 1% of the days
on the neurological wards. Concerning the internal
medicine wards, the number was 6% and on mixed
wards it was 10%.

Discussion
The base for benchmarking should be comparable data,
which requires clearly defined terminology and variables as well as methods (see Holcomb et al. 2002). The
reliability concerning patient classification was decided
based on the reliability degree for parallel classification,
which in this study was 77%. Concerning the credibility
of the study, it should also be taken into consideration
that the data collection according to the developed
concept was made for the second year in a row and the
material from 2001 was estimated as reliable.
The part of the workload depending on the NCI was
10.8% higher on the regional and local hospital level
than on the university hospital level. To be able to
reliably discover more in detail if these results are general in Finland would require that the same type of
wards and enough number of patients per specialized
area would participate, i.e. similar wards with regard to
their special features.
Comparisons between specialized areas have their
limitations in practice, for example, wards that are
called surgical wards can be different from each other in
certain areas. This can be explained through the size of
the hospital and its opportunities for specialization and

7
16
15

23

4
11

32

23

8
4
14

Mean
19
37
16
1
24
15
28
13
54
18
38
10
6
20

the co-operation between close-by hospitals as well as

the possibilities primary care has, to receive patients for
continued treatment and rehabilitation, for example.
During the past few years, a common Nordic version
of the DRG-groups, NordDRG, has been used by a
growing number of hospitals in Finland. Analyses based
on DRG-information were not made, because information on the patients DRG was available only from
two university hospitals and five regional hospitals. The
NordDRG-system in Finland does not consider how
personnel resource-demanding the patients are: instead
the personnel costs are divided equally over all nursing
care days. This criticism has been presented by a number of researchers (Halloran & Kiley 1987, Johnson
Lutjens 1993). Today, there is a clear aim on an international level to connect the DRG-system to the NCI
and at some hospitals in Finland this is already possible
today.
A more careful analysis of the connection between
NCI, personnel resources and the length of hospital stay
is also possible based on the connection to Nord DRGinformation. Some studies have shown that the NCI
explains much more of the variations in the length of
hospital stay than the weight coefficient of the DRGgroup (Halloran & Kiley 1987, Johnson Lutjens 1993).
The length of the hospital stay is important for the total
costs per patient. Earlier studies have shown that the
patients NCI per day increases if the hospital stay is
shortened (Coben 1991, Shamian et al. 1994, Blegen
et al. 1998). Analysis of the benchmarking material
now available from 2001 to 2005 enables more careful
analysis of NCI and its relationship to the length of
hospital stay.
The much lower NCI/N on the childrens wards was
remarkable (mean 15.1) compared with the adults
wards (mean 25.2). However, this should be compared

2007 The Authors. Journal compilation 2007 Blackwell Publishing Ltd, Journal of Nursing Management, 15, 683692

689

L. Fagerstrm and A. Rauhala

with the set optimal levels of the wards, which is much

lower for the childrens wards.
In this study, remarkable differences were noticed in
salary costs for the production of one NCIP, i.e. within
the highly specialized nursing care on a university hospital level the salary costs for one NCIP were 10.40,
and the salary cost level on a regional and local hospital
level were only 8.80. Subsequently, the cost level follows the nurses workload expressed in NCIP, but in
reverse order, i.e. higher level of nursing care load per
nurse gives a lower price per NCIP.
In this study, there was a clear imbalance between
NCIP/N in relation to the set optimal NCI levels on
almost half of the days. Corresponding results have
been reported from Australia, where 3050% of the
nurses think that they cannot satisfy the patients need
for nursing care, due to a lack of personnel resources
(Hegney et al. 2003). What is the result of NCIP/N
exceeding the upper limit of the optimal level during
48% of the days? The overall aim of the RAFAELA
system was that the personnel resources should be in
balance with the patients caring needs and this is supposed to be guaranteed when the NCIP/N is on the
optimal level for the ward.
Today, there is a need to connect nursing staffing
with patient outcomes in the Nordic countries. An
imbalance between NCI and personnel resources clearly
affects the nursing care quality and the care results.
Current international research shows that there is a
clear connection between nurse staffing and key outcomes, such as complaints, urinary tract infections,
pneumonia, length of stay, upper gastrointestinal
bleeding, thrombosis, shock and failure to rescue
(Blegen et al. 1998, Kovner & Gergen 1998, Needleman et al. 2002). Aiken et al. (2002) large survey
showed that poor nurse staffing associated with higher
30-day mortality and failure to rescue.
High workload and poor staffing ratios are also
associated with factors, such as nurse burnout, low job
satisfaction and increased nurse stress. Absence due to a
number of reasons is a growing problem within nursing
care (Beil-Hildebrand 1996). A recent analysis of sick
leave and nursing care load data from the RAFAELA
system showed that when NCI/N exceeded the optimal
level with 15%, the amount of long-term sick leave
increased (>3 days) with 30% and short-term sick leave
(13 days) with 4050%.
Workplace stress in nursing is caused not only by high
workload and leadership/management issues, but also
by ethical dilemmas because of high professional goals
(Mylott 2005). The nurses working situation has been
described as a struggle between what one wants and
690

what one can actually achieve, or with other words, a

struggle between being or not being a good nurse
(Fagerstrom 2006, p. 8). When personnel resources do
not match the patients caring needs the nurses experience bad working conditions, lower job satisfaction and
the quality of the nursing care is getting worse, i.e. they
have feelings of inadequacy (Fagerstrom et al. 2002).
Also nurse managers suffer from ethical dilemmas between administrative goals and the caring mission of
nursing (Nyberg 1990). It is a challenge for both the
nurses and the managers to be able to make a synthesis
of the dialectic tension between the administrative goals
of the organization and the professional and ethical
goals in nursing, i.e. a synthesis between the administrative anti-thesis and the caring thesis (Fagerstrom
1999).
Constant organizational changes and new and changed nursing care and treatment methods are supposed to
increase the NCI and the nurses workload (Levenstam
& Bergbom 2002). A continuous monitoring and analysis of different factors is therefore important (Spence
et al. 2006). According to the theory on learning
organizations, benchmarking activity should create a
breeding-ground for learning organizations (Karlof et
al. 2000). A learning organization aims towards a search
for new knowledge. This would subsequently facilitate
organizational development to occur. Benchlearning is
an attitude, i.e. a will to learn from experiences of others
(Pleijel 2004). In order for the benchmarking activity to
be really effective and useful, it should be developed into
benchlearning. According to literature, a development
from benchmarking to benchlearning requires systematic working methods (Pleijel 2004). Transferred into
nursing care administration, one could ask if the nurse
managers of today have the conditions for a maximum
use of the possibilities of the benchmarking concept.
One risk is that the important information that is a result
of the benchmarking system remains unused. In order to
prevent this from happening, training of the opportunities of benchmarking should be provided for nurse
managers.
It is worth asking to what extent nurse managers have
a preventive approach in their leadership. Research
shows the importance of an active and inclusive style of
management (McVicar 2003) with the aim being to
reduce the experience of stress. In many countries, there
is nowadays a shortage of nurses not only due to lack of
interest for the nursing care branch among the youth,
but also due to the fact that trained nurses work abroad
or within other areas. As a clear connection has been
noted between a manageable work situation, job
satisfaction and an adequate number of nurses (Adams

2007 The Authors. Journal compilation 2007 Blackwell Publishing Ltd, Journal of Nursing Management, 15, 683692

Benchmarking by the RAFAELA patient classification system

& Bond 2003, Hegney et al. 2003), one could ask if the
negative picture of the nurses work situation that quite
often appears in mass media has already influenced the
shortage of personnel. With more old people in society
and an increased need for nursing care, the question
about personnel resources is already a burning issue
today, which in turn speaks for benchmarking with the
RAFAELA system as a necessary tool for the nurse
managers.

Conclusions
The RAFAELA system has been evaluated as userfriendly and the opportunity to use a national benchmarking system provides added value to the branch,
and this in turn has increased the interest for the system
among nurse managers and decision-makers in Finland.
Benchmarking with the RAFAELA system provides
nurse managers with many opportunities in their
decision processes concerning cost and productivity
analysis and resource allocation. Resource allocation
based on NCI is a condition for qualitative nursing care.
Based on this study, it can be stated that benchmarking
with the RAFAELA system is a well-functioning
administrative tool for nurse managers and decisionmakers on different levels of the health care and hospital organizations of today.

References
Adams A. & Bond S. (2003) Staffing in acute hospital wards: Part
1. The relationship between number of nurses and ward organizational environment. Journal of Nursing Management 11,
287292.
Aiken L.H.J., Clarke S.P., Sloane D.M., Sochalski J. & Silber J.H.
(2002) Hospital nurse staffing and patient mortality, nurse
burnout, and job dissatisfaction. JAMA 288, 19871993.
Beil-Hildebrand M. (1996) Nurse absence the causes and the
consequences. Journal of Nursing Management 4, 1117.
Blegen M.A., Goode C.J. & Reed L. (1998) Nurse staffing and
patient outcomes. Nursing Research 47 (1), 4350.
Camp R.C. (1993) Lar av de basta! Benchmarking i tio steg.
Studentlitteratur, Lund (in Swedish).
Coben E.L. (1991) Nursing case management. Does it play?
Journal of Nursing Administration 21 (4), 2025.
Fagerstrom L. (1999) The patients caring needs to understand
and measure the unmeasurable. Doctoral thesis, Abo Akademi
University Press, A60.
Fagerstrom L. (2000) Expert validation of the Oulu Patient
Classification a phase in the development of a new system of
patient classification, RAFAELA. Nordic Journal of Nursing
Research and Clinical Studies 20, 1521 (in Swedish).
Fagerstrom L. (2006) A hermeneutic study of nurses workload
the dialectic tension between to be or not to be a good nurse.
Nursing Ethics 13 (6), 111.

Fagerstrom L. & Rauhala A. (2001) Finnhoitoisuus hoitotyon

benchmarking Pilottiporjektin raportti vuodelta 2000. Suomen
Kuntaliitto, Helsinki (in Finnish).
Fagerstrom L. & Rauhala A. (2003) Finnhoitoisuus hoitotyon
benchmarking Projektin loppurapportti 20002002. Suomen
Kuntaliitto, Helsinki (in Finnish).
Fagerstrom L., Eriksson K. & Bergbom I.E. (1999) The patients
perceived caring needs measuring the unmeasurable. International Journal of Nursing Practice 5, 199208.
Fagerstrom L., Rainio A.-K., Rauhala A. & Nojonen K. (2000)
Validation of a new method for patient classification, the Oulu
Patient Classification. Journal of Advanced Nursing 31 (2),
481490.
Fagerstrom L., Nojonen K. & Akers A. (2002) Methodological
triangulation as a validation method in the testing of the
PAONCIL method. Journal of Nursing Science 14, 180191 (in
Finnish).
Giovannetti P. (1979) Understanding patient classification systems. Journal of Nursing Administration 9, 49.
Halloran E.J. & Kiley M. (1987) Nursing dependency, diagnosisrelated groups, and length of hospital stay. Health Care
Financing Review 8 (3), 2735.
Hegney D., Plank A. & Parker V. (2003) Nursing workloads: the
results of a study of Queensland Nurses. Journal of Nursing
Management 11, 307314.
Holcomb B.R., Hoffart N. & Fox M.H. (2002) Defining and
measuring nursing productivity: a concept analysis and pilot
study. Journal of Advanced Nursing 38 (4), 378386.
Johnson Lutjens L.R. (1993) Determinants of hospital length of
stay. Journal of Nursing Administration 23 (4), 1418.
Kaivos P., Laamanen K., Salonen L. & Valpola A. (1995)
Benchmarking Huipputasosta opiiminen suomalaisia kaytannon kokemuksia. Metalliteollisuuden Kustannus Oy, Helsinki (in Finnish).
Karlof B. (2004) Inledning och bakgrund. In Benchlearning 2
Erfarenheter fran sju praktikfall (C. Pleijel ed.), pp. 925.
Ekerlids forlag, Falun (in Swedish).
Karlof B., Lundgren K. & Edenfeltdt Froment M. (2000)
Benchlearning forebilder for havstang for utveckling. Ekerlids
forlag, Falun (in Swedish).
Kovner C. & Gergen P.J. (1998) Nurse staffing levels and adverse
outcomes following surgery in U.S. hospitals. Journal of Nursing Scholarship 30 (4), 315321.
Levenstam A.-K. & Bergbom I. (2002) Changes in patients need
of nursing care reflected in the Zebra system. Journal of Nursing Management 10, 191199.
McVicar A. (2003) Workplace stress in nursing: a literature
review. Journal of Advanced Nursing 44 (6), 633642.
Mylott L. (2005) The ethical dimension of the nurses role in
practice. Journal of Hospice and Palliative Nursing 7, 113118.
Needham J. (1997) Accuracy in workload measurement: a fact or
fallacy? Journal of Nursing Management 5, 8387.
Needleman J., Buerhaus P. & Mattke S. (2002) Nurse-staffing
levels and the quality of care in hospitals. New England Journal
of Medicine 346 (22), 17151722.
Noro A., Finne-Soveri H., Bjorkgren M. et al. (2001) RAI-tietojarjestelman kayttoonotto ja pitkaaikaishoidon benchmarking. RAI-raportti 1/2000. Stakes, Helsinki (in Finnish).
Nyberg J. (1990) Theoretic explorations of human care and
economics: foundations of nursing administration practice.
Advanced Nursing Science 13, 7484.

2007 The Authors. Journal compilation 2007 Blackwell Publishing Ltd, Journal of Nursing Management, 15, 683692

691

L. Fagerstrm and A. Rauhala

Partanen P. (2002) Hoitotyon henkiloston mitoittaminen erikoissariaanhoidossa. Doctoral Thesis, Kuopio University Publications E, Social Sciences 99, Kuopio (in Finnish).
Pleijel C. (2004) Benchlearning metodoversikt. In I Benchlearning
2 Erfarenheter fran sju praktikfall (C. Pleijel red ed.), pp. 27
57. Ekerlids forlag, Falun (in Swedish).
Rauhala A. & Fagerstrom L. (2004) Determining optimal nursing
intensity: the RAFAELA method. Journal of Advanced Nursing
45, 351359.
Ruland C.M. & Ravn I.H. (2003) Usefulness and effect on costs
and staff management of a nursing resource management

692

information system. Journal of Nursing Management 11,

208215.
Shamian J., Hagen B., Hu T.W. & Fogarty T.E. (1994) The
relationship between length of stay and required nursing care
hours. Journal of Nursing Administration 24 (7/8), 5258.
Spence K., Tarnow-Mordi W., Duncan G. et al. (2006) Measuring
nursing workload in neonatal intensive care. Journal of Nursing
Management 14, 227234.
The Northern Federation of Nurses (1995) Ethical Guidelines for
Nursing Research in the Nordic Countries. Typotequet.

2007 The Authors. Journal compilation 2007 Blackwell Publishing Ltd, Journal of Nursing Management, 15, 683692