Factors related to falls, weight-loss and pressure ulcers – more insight in risk assessment among nursing home residents

O R I G I N A L A R T I C L E
Factors related to falls, weight-loss and pressure ulcers – more insight
in risk assessment among nursing home residents
Christina Lannering, Marie Ernsth Bravell, Patrik Midl€ov, Carl-Johan €Ostgren and Sigvard M€olstad
Aims and objectives. To describe how the included items in three different scales,
Downton Fall Risk Index, the short form of Mini Nutritional Assessment and the
Modified Norton Scale are associated to severe outcomes as falls, weight loss and
pressure ulcers.
Background. Falls, malnutrition and pressure ulcers are common adverse events
among nursing home residents and risk scoring are common preventive activities,
mainly focusing on single risks. In Sweden the three scales are routinely used
together with the purpose to improve the quality of prevention.
Design. Longitudinal quantitative study.
Methods. Descriptive analyses and Cox regression analyses.
Results. Only 4% scored no risk for any of these serious events. Longitudinal risk
scoring showed significant impaired mean scores indicating increased risks. This
confirms the complexity of this population’s status of general condition. There were
no statistical significant differences between residents categorised at risk or not
regarding events. Physical activity increased falls, but decreased pressure ulcers. For
weight loss, cognitive decline and the status of general health were most important.
Conclusions. Risk tendencies for falls, malnutrition and pressure ulcers are high in
nursing homes, and when measure them at the same time the majority will have several
of these risks. Items assessing mobility or items affecting mobility were of most impor-
tance. Care processes can always be improved and this study can add to the topic.
Relevance to clinical practice. A more comprehensive view is needed and prevention
can not only be based on total scores. Mobility is an important factor for falls and
pressure ulcers, both as a risk factor and a protective factor. This involves a challenge
for care – to keep the inmates physical active and at the same time prevent falls.
Key words: falls, frail older, malnutrition, nursing homes, pressure ulcers, risk
assessment
What does this paper contribute
to the wider global clinical
community?
 The complexity to risk group cat-
egorise frail older persons. Risk
tendencies for falls, malnutrition
and pressure ulcers are high
among older people living in
nursing homes and the majority
will have several risks. The total
scores, which constitute basis for
risk grouping, are not always
sufficient information for the
preventive work as a more com-
prehensive view is needed.
 Care processes can always be
improved. The results from this
study can contribute to the knowl-
edge on how to assess older frail
persons. Maybe there are other
ways than using several assessing
scales. Mobility remains an impor-
tant factor, both as a risk factor
and a protective factor and that is
challenge for care to manage.
Accepted for publication: 5 November 2015
Authors: Christina Lannering, RN, PhD Student, Unit of Research and
Development in Primary Care, Futurum, J€onk€oping; Marie Ernsth
Bravell, PhD, RN, Associate professor, Institute of Gerontology, School
of Health Sciences, J€onk€oping University, J€onk€oping; Patrik Midl€ov,
MD, Associate Professor, Department of Clinical Sciences in Malm€o,
General Practice/Family Medicine, Lund University, Malm€o; Carl-
Johan €Ostgren, MD, Professor, Department of Medical and Health
Sciences, General Practice, Link€oping University, Link€oping; Sigvard
M€olstad, MD, Professor, Department of Clinical Sciences in Malm€o,
General Practice/Family Medicine, Lund University, Malm€o, Sweden
Correspondence: Christina Lannering, PhD Student, Unit of
Research and Development in Primary Care, Futurum, SE-551 85
J€onk€oping, Sweden. Telephone: [phone].
E-mail: [e-mail]
A study aimed to provide more insight in how different scale items
interact with each other and how they are associated to severe out-
comes. It is not a prediction study or a study of diagnostic accu-
racy, but a study that can contribute to the field of knowledge of
assessments in older persons.
© 2016 John Wiley & Sons Ltd
940 Journal of Clinical Nursing, 25, 940–950, doi: 10.1111/jocn.13154

Introduction
In Sweden, like in the rest of Western societies the numbers
of older people are increasing, and also the number of older
people needing care and services (WHO 2012). The munici-
pality has the responsibility to offer care in nursing home
facilities when the older person can no longer manage at
home. According to national statistics Sweden had approxi-
mately 87,600 persons at the age 65 and over permanently
staying in nursing home facilities. From these, 80% were
aged above 80 years and 69% were women (Board of
National Health and Welfare 2012). Old age care in Sweden
requires a need assessment performed by a special trained
social worker. The need assessment is based on the amount
of functional problems performing activities in daily life
(ADL). As a ‘stay-in-place’ policy is applied, home care ser-
vice is offered as long as possible. Moreover, the number of
beds in nursing homes has decreased by 20% during the last
10 years in Sweden (Board of National Health and Welfare
2012). These circumstances indicates that individuals mov-
ing in to nursing homes today are frailer and in more
extended need of care than previously.
To provide optimal care and foremost preventive actions
it is essential to know risk status, and helpful tools to
establish action policies can be assessing scales. Accord-
ingly, scoring risk for falls, malnutrition and pressure ulcers
are common preventive activities in the care of older people
and numerous scales and assessments are used for this pur-
pose. In Sweden the most common tools are Downton Fall
Risk Index (DFRI) to assess fall risk (Downton 1993), the
short form of Mini Nutritional Assessment (MNA-SF)
(Rubenstein et al. 2001) to assess risk for malnutrition and
the Modified Norton Scale (MNS) (Ek 1987) to assess risk
for developing pressure ulcers.
Background
Falls by older people in nursing home facilities are common
events. The prevalence of falls in institutionalised older peo-
ple is reported at 53–62% of the inmates (Rosendahl et al.
2003, Meyer et al. 2009). Risk factors have been described
as gait and balance instability, cognitive and functional
impairment, sedating and psychoactive medications (Ruben-
stein et al. 1994) and number of diseases (Damian et al.
2013). Some falls may be caused by a single factor, but the
majority of falls are caused by a combination of factors
(Cameron et al. 2010).
Older persons are also considered to be at high risk of
malnutrition. Several studies in nursing homes populations
have shown both high risk and high prevalence of malnutri-
tion; a recent review showed that approximately 14% were
classified as malnourished and more than half were at risk
of malnutrition (Kaiser et al. 2011). A follow-up study in
Swedish nursing homes showed that nutritional status was
improved, but still 63% were assessed at risk, and 30% of
those were malnourished (Torma et al. 2013).
A third major and serious event that is common among
older persons in nursing homes is pressure ulcers. A recent
systematic review of pressure ulcers risk factor studies iden-
tified three primary risk domains; mobility/activity, perfu-
sion and general skin status. However, no single factor can
predict pressure ulcer risk, which is caused by a complex
interplay of factors (Coleman et al. 2013). A Swedish nurs-
ing home study showed a prevalence of pressure ulcers at
14% and according to risk assessment, a risk between 26–
30% (Gunningberg et al. 2013) which is similar to other
European studies (Meesterberends et al. 2013).
One must also consider ageing as a risk factor for these
outcomes, knowing that biological ageing increases the vul-
nerability and decreases the reserve capacity (Fried et al.
2001, Rockwood & Mitnitski 2007).
In Sweden, DFRI, MNS and MNA-SF are routinely used
together to assess risks in older persons living in nursing
homes. The scales are included in the quality registry Senior
Alert which is a national investment aimed to increase the
quality of the preventive work. The widely used MNA was
developed and validated for the assessment of older, frail
persons. MNA has a long history (Secher et al. 2007) and
seems to be well suited for nursing home residents (Diek-
mann et al. 2013). Further validation has shown that the
short-form can be used as a stand-alone unit (Bauer et al.
2008, Salvi et al. 2008, Dent et al. 2012). DFRI was vali-
dated in a Swedish study (Rosendahl et al. 2003) and
appeared to be a useful tool for predicting falls among
older people in residential care facilities. However, a com-
parison with DFRI and nurses judgement alone showed no
clinical benefit for DFRI (Meyer et al. 2009). MNS is
tested, recommended and well known in Sweden (Gunning-
berg et al. 2013) and it is validated to its actual content
(Ek & Bjurulf 1987).
It is reasonable to believe that frail older persons have
several risks and that general decline increases serious
events, but using several different instruments can be time
consuming and increase the workload as results must be
documented and interventions should be planned and
followed. Therefore, it is important to put knowledge to
this topic so that nurses can reflect upon the usefulness.
One problem when using the three scales together is that
several functions are assessed repeatedly as they exist in
more than one scale. Mobility and cognition are, for exam-
© 2016 John Wiley & Sons Ltd
Journal of Clinical Nursing, 25, 940–950 941
Original article Risk assessment among nursing home residents

ple, assessed in all three scales but in different ways and
with different grading. The ambition with the present study
is to gain knowledge about the relations among the out-
comes and the included scale items.
Aim
This study aims to find patterns of associations among scale
risk items in MNA-SF, DFRI and MNS, with the outcomes
falls, pressure ulcers and weight-loss.
Method
Study population
Data from this study were collected from a longitudinal
cohort study of older people living in nursing homes in
Sweden; The Study on Health and Drugs in Elderly
(SHADES). The SHADES study was launched in 2008
and completed in 2011 and the overall aims were to
describe and analyse morbidity, health-conditions and
drug-use among older people in nursing home facilities.
A convenience sample of 12 nursing homes including 443
beds was included in the SHADES study. The nursing
homes were located in three different regions in southern
Sweden and were all in the public sector. As participants
were included during the whole study period, the partici-
pants had different durations which consequently led to
varying number of follow-up assessments. When the study
nurse returned for a follow-up visit, all new inmates were
asked to participate, not just those who moved in where
a deceased participant had lived. Figure 1 describes the
inclusion flow. Exclusion criteria were palliative care or
language problems. All together 664 were asked to par-
ticipate and 429 were included.
As 98 individuals participated only once, 331 of 429
participants were available for prospective analysis. For
some of the statistical analyses sub-samples were used,
which are described under the heading of ‘Statistical analy-
ses’.
The study, which complied with the Declaration of Hel-
sinki, was approved by the Regional Ethical Review Board,
Link€oping, Sweden (M150-07). Written informed consent
was obtained from all participants. If the patient could not
understand the information and give informed consent, it
was obtained from a proxy. For those who declined to par-
ticipate, information of age, gender, diagnoses and reason
for not participating was recorded. There were no differ-
ences between the participants and those who declined to
participate according to age, gender and diagnoses.
Data collection
A total of six visits at every sixth month were performed by
specially trained nurses. The same nurse visited the same
nursing homes at each follow-up visit for examinations,
data collecting and to include new residents. Risk-assess-
ment tools were performed with support from the staff, that
is, each individual’s contact person. Fall risk was assessed
with DFRI, which includes 11 risk items concerning previ-
ous falls in the last six months, use of drugs (tranquillisers/
sedatives, diuretics, antihypertensives, antidepressants and
antiparkinsonian drugs), sensory deficits (visual- and hearing
impairment), limb abnormalities (hemiparesis), cognitive
dysfunction and walking ability (Downton 1993). Each item
is scored 1 point and added to give a total score range of
maximum 11. A score of 3 and more indicates an increased
risk to fall. Falls were defined as ‘an event which causes the
patient to come unintentionally to the ground or some lower
level, regardless of the cause’ (Lamb et al. 2005).The num-
ber of individual falls was not taken into account, but only
if they had fallen or not.
Risk for malnutrition was assessed by using MNA-SF,
which is a six-question short-form of MNA (Rubenstein
et al. 2001). MNA-SF covers the past three months andEligible to inviten = 664Refused to parcipate n = 100Proxy denied parcipaon n = 87Excluded (palliave care, language problem or diedbetween consent signing and baseline examinaonn = 48Included toSHADES n = 429Baseline6 monthsfollow-upn = 331Died n = 50Included at last visit and parcipated only once(baseline), not possible to follow up n = 48Died n = 38Included at visit 5 and participated at baseline andone follow up, not possible to follow up at 12 monthsn = 3812 monthsfollow-upn = 255
Figure 1 Inclusion to the SHADES study.
© 2016 John Wiley & Sons Ltd
942 Journal of Clinical Nursing, 25, 940–950
C Lannering et al.

addresses decreased food intake (0–2 points), estimated
weight loss (0–3 points), mobility (0–2 points), acute diseases
or psychological stress (0 or 2 points), neuropsychological
impairment (0–2 points) and BMI (0–3 points). The higher
the value the lower the risk. The maximum MNA-SF score is
14 points. A score of 7 points or less indicates malnutrition,
8–11 indicates risk of malnutrition and 12–14 points indi-
cates no risk for malnutrition.
Risk for pressure ulcers was assessed with MNS. In addi-
tion to the more internationally known Norton Scale, MNS
also includes two items assessing nutrition status. MNS
consists of seven items; mental condition, activity, mobility,
food intake, fluid intake, incontinence and general physical
condition. Each item is assessed with a range from 1 (lack
of function) to 4 (normal function). The maximum score is
28 and a score at 20 or lower indicates an increased risk
for pressure ulcers (Ek 1987).
The scales internal consistency in this study, measured by
Cronbach’s alpha showed 05 for DFRI, 066 for MNS
and 045 for MNA-SF.
In the SHADES study the nurses examined the partici-
pants in many ways regarding different assessment scales,
blood testing, use of drugs, different measurements etc. For
this present study we used data from DFRI, MNA-SF,
MNS, data of weight and data of eventual presence of pres-
sure ulcers. Pressure ulcers were graded as: (1) persistent
discoloration, with intact skin surface; (2) epithelial dam-
age; (3) damage to the full thickness of the skin without a
deep cavity and (4) damage to the full thickness of the skin
with deep cavity. In this study, all kinds of pressure ulcers
were taken into account, but not gradated, only counted as
existing or not. Description of the study population con-
cerning prevalence and risks are presented in Table 1.
Statistical analyses
Descriptive statistics were used for baseline characteristics
and to describe outcomes at 6 and 12 months. For analys-
ing outcomes in relation to risks at six and 12 months a
subsample of 255 was used, that is, participants with at
least three occasions of measure (baseline and at least two
follow-ups). The proportions of being at risk or not were
compared using a two-sided Chi square test. To analyse the
longitudinal changes of the total scores in the three assess-
ments scales, General Linear Model (GLM with repeated
measures) was used. For the GLM analyses a subsample of
116 was used, that is, those who participated during the
whole study period (30 months). Cox regressions were per-
formed to analyse factors related to the outcomes falls,
weight loss and pressure ulcers. Three outcome variables
were created ‘Falls last six months’ (coded as 1), ‘Weight
loss with at least 10% or at least 6 kg’ (coded as 1) and
‘Presence of pressure ulcers’ (coded as 1). Those who even-
tually already had suffered from any of these events at
baseline were included in the analyses with the purpose to
increase the size of the population. Exclusion of these indi-
viduals would have decreased the population and made it
less representative for a nursing home population. How-
ever, the time variable for those who already suffered from
some kind of pressure ulcer at baseline was time coded with
0. For the other two outcomes the time variable was
determined as time from baseline to the follow-up visit date
when the outcome variable first was detected. If no occur-
rence of the outcome variables during the two follow-ups
was detected, the time variable was determined to be
approximately one year, or, if the participant was deceased,
time to death was calculated. Three Cox regression analyses
in two steps were performed for each outcome. Baseline
data on age, gender, number of drugs and number of diag-
noses were included as covariates in the first step to control
Table 1 Description of study population at baseline
n = 331
Age (mean) 84
Mean DFRI (SD) 48 (16)
DFRI risk % 93
Falls % (≥1 fall last six months) 62
Mean MNA SF (SD) 103 (25)
Mna-SF risk % 58
MNA-SF item Weight
loss > 3 kg last three months %
3
MNA-SF item BMI < 19 (%) 6
BMI 19–20 (%) 13
BMI 21–23 (%) 18
BMI > 23 (%) 63
Mean BMI (SD) 25 (476)
Mean MNS (SD) 23 (345)
MNS risk % 27
Presence of pressure ulcers % 10
Mean number of medication (SD) 685 (304)
Mean Number of diseases (SD) 29 (133)
Mobility
Wheelchair bound (%) 30
Walking with assistance (%) 15
Walking with or without aid (%) 55
Bedridden (%) 0
Mean MMSE (SD)* 17 (63)
MMSE<24 points (%) 68
Hospital care last six months % 24
Emergency care last six months % 6
*MMSE was not performed on all participants. Due to cognitive
dysfunction, blindness or impaired hearing, questions could not be
understood for all, which made 83% eligible to assess.
© 2016 John Wiley & Sons Ltd
Journal of Clinical Nursing, 25, 940–950 943
Original article Risk assessment among nursing home residents