A Randomized Clinical Trial to Assess the
Impact on an Emergency Response System
on Anxiety and Health Care Use among
Older Emergency Patients after a Fall
Jacques S. Lee, MD, MSc, Mary Jane Hurley, BScN, Debra Carew, MScN, Rory Fisher, MD, Alex Kiss, PhD,
Neil Drummond, PhD
Objectives: Personal emergency response systems (PERSs) are reported to reduce anxiety and health care
use and may assist in planning the disposition of older patients discharged from the emergency department
(ED) to home. This study measured the impact of a PERS on anxiety, fear of falling, and subsequent health
care use among older ED patients.
Methods: This study was a randomized controlled trial comparing PERS use with standard ED discharge
planning in subjects 70 years of age or older discharged home after a fall. Outcome assessors were blinded
to the study objectives. Anxiety and fear of falling were measured at baseline and 30 days using the Hospital
Anxiety and Depression Scale anxiety subscale (HADS-A) and modified Falls Efficacy Scale (mFES). Return
to the ED, hospitalization, and length of stay were recorded after 30 and 60 days.
Results: Eighty-six subjects were randomized and completed follow up (43 per group). There was no im-
portant difference in mean reduction in anxiety (mean change treatment ? control, +0.35; 95% confidence
interval [CI] = ?1.5 to 0.76; p = 0.55) or fear of falling (mean change, +4.5; 95% CI = ?6.7 to 15.7; p = 0.70).
Return visits to the ED occurred in eight of 43 patients in both the control and treatment groups (risk dif-
ference, 0.0%; 95% CI = ?16% to 16%). Hospitalization occurred in six of 43 in the control group versus
three of 43 in the treatment group (risk difference treatment ? control = ?7.0%; 95% CI = ?19.8% to 5.9%).
Conclusions: In contrast to previous studies, there was no evidence that a PERS reduced anxiety, fear of
falling, or return to the ED among older persons discharged from the ED.
ACADEMIC EMERGENCY MEDICINE 2007; 14:301–308 ª 2007 by the Society for Academic Emergency
Keywords: personal emergency response system, anxiety, fear of falling, randomized controlled
trial, geriatric emergency medicine
a fall. Falls account for 49% of the approximately three
million annual injury-related ED visits by persons aged
65 years and older.1Only 5%–15% of these falls result
in an injury that would require hospitalization.2–4How-
ever, older patients have higher rates of adverse out-
mergency physicians frequently make difficult
disposition decisions about older persons pre-
senting to the emergency department (ED) after
comes after discharge, including recurrent falls, return
to the ED (RTED), and subsequent hospitalization.5Fear
of falling is an important, underrecognized complication
of falls, affecting between 45% and 75% of older persons
with a recent fall.2,3,6,7This has the potential to create a
vicious cycle of a fall leading to fear of falling, restricted
activity, decreased conditioning, and increased risk for
subsequent falls, injuries, and hospitalization.8–11
From the University of Toronto and Sunnybrook Research Insti-
tute Toronto (JSL), Sunnybrook Health Sciences Center (MJH,
DC, RF), and Institute for Clinical Evaluative Studies (AK), Tor-
onto, Ontario, Canada; and Department of Epidemiology and Bi-
ostatistics (ND), University of Calgary, Calgary, Alberta, Canada.
Received August 31, 2006; revisions received November 17,
2006, and November 20, 2006; accepted November 20, 2006.
Presented at the Canadian Association of Emergency Physicians
annual meeting, Montreal, Quebec, Canada, June 2004.
Supported by the Physicians Services Incorporated Foundation
of Ontario (grant PSIF 02-13), which provided peer-reviewed
funding for this study.
Contact for correspondence and reprints: Jacques S. Lee, MD,
MSc; e-mail: firstname.lastname@example.org.
ª 2007 by the Society for Academic Emergency Medicine
PII ISSN 1069-6563583
Currently, the care of older persons in the ED requires
more time and resources and results in poorer patient
outcomes and satisfaction than in younger patients.5,12
This is particularly concerning for the practice of emer-
gency medicine, because the proportion of people older
than 65 years will double in the next 25 years.13,14Thus,
studying interventions such as the personal emergency
response systems (PERSs), which can be initiated in the
ED and may assist frail elders in continuing to live inde-
pendently, are crucial; such ED-based interventions have
been identified as a top research priority by the Ameri-
can Geriatric Society Research Agenda Setting Pro-
cess.15PERSs are designed to summon help should the
user fall in his or her home and not be able to get up
and have been proposed as one such intervention.16–23
In 2003, more than 500,000 North Americans subscribed
to a PERS,24at a cost of between $200 and $600 per client
per year. An increasing number of third-party payers
cover PERSs, and many systems are paid for by public
Previous hospital-based, nonrandomized studies of
PERSs have shown that their use reduced subsequent
health care use.19,20,23Anxiety is known to be an inde-
pendent predictor of frequent RTED; therefore, an inter-
vention that reduces anxiety has the potential to reduce
RTED and other health care use.28,29One previous
community-based randomized clinical trial assessed the
impact of a PERS on anxiety.23This study found a nonsig-
nificant trend toward reduced anxiety in the intervention
group. The results of this trial have been cited in several
subsequent reports as providing evidence of the psy-
chological benefits of PERSs, specifically anxiety reduc-
The primary goal of this study was to determine if the
addition of a PERS to ED discharge planning by a geriat-
ric-trained nurse reduces anxiety and fear of falling.
Reducing anxiety and fear of falling are important goals
by themselves but could have the additional benefit of
reducing deconditioning, falls, and subsequent health
care use. Therefore, we also assessed whether the use
of a PERS resulted in reduced subsequent RTED and
other health care use.
We conducted a randomized, controlled effectiveness
trial32of the impact of a PERS on anxiety and fear of fall-
ing. The treating emergency staff and outcome assessors
were blinded to subjects’ randomization status. We did
not blind subjects to the intervention, because use of a
sham PERS device might place participants at increased
risk in the event of an emergency. The research ethics
board of our institution approved the study, and all
participants provided written informed consent.
Study Setting and Population
We conducted this study in the ED of a large metropoli-
tan university teaching hospital (Sunnybrook Health Sci-
ences Center, Toronto, Ontario, Canada) with a census of
45,000 patients per year. Approximately 27% of patients
seen in our ED are 70 years of age or older. All persons
70 years and older who presented to the ED after a fall,
lived in their own home before their fall, could communi-
cate in English, had telephone service, lived within our
institution’s geographic catchment area, were mentally
competent, and could be discharged home in the opinion
of the attending emergency physician were eligible for
study participation. We excluded patients who already
had an ERS system, who lived in nursing or assisted-liv-
ing homes, or who lived in their own homes but required
16 hours of assistance per day or more. We recruited
patients between January 2003 and May 2004.
The study coordinator attempted to contact consecutive
eligible patients either in the ED or at home. We included
patients who were contacted within 72 hours of their
discharge from the ED and provided informed consent
(see Figure 1). A list assigning participants to either the
control or treatment groups was randomly generated by
computer, balanced in blocks of four. The study coordi-
nator opened the randomization envelope for consenting
participants only after all baseline data were collected, so
that group assignment would not impact baseline anxiety
Patients randomized to the control group received
conventional discharge planning, which in our institution
included assessment by a geriatric emergency man-
agement (GEM) nurse. GEM nurses were available eight
to ten hours per day, seven days a week, but would at-
tempt to contact patients who were discharged outside
their work hours by telephone. The main role of the
GEM nurses was to perform a focused geriatric assess-
ment and determine whether subjects needed specialized
outpatient services. These services included geriatric out-
reach, geriatric day-hospital, fall prevention, and home
care services. GEM nurses did not continue ongoing
case management of patients at home after they com-
pleted their initial assessment and referrals. To blind
the GEM nurses to intervention status, randomization
occurred after they had completed their assessment
and referrals. Control patients were requested not to
install a PERS during the study period.
Patients randomized to the treatment group received
conventional discharge planning described previously
plus free use of a PERS (Lifeline Systems Canada,
Toronto, Ontario, Canada) for two months. The PERS
included three elements: a portable alarm button worn
as a pendant or wristband, a speaker-microphone unit
connected to a standard telephone line, and a central
monitoring station that was staffed around the clock.33
If the client needed assistance, he or she could depress
the alarm button, sending an FM signal to the speaker-
microphone unit, alerting the central operator, and open-
ing a two-way voice channel in the client’s home. This
would allow the central monitoring station to communi-
cate with the patient and determine the type of response
required, ranging from contacting a neighbor or family
member to calling 9-1-1. Each user established a custom-
ized protocol designating whom to contact (e.g., neigh-
bor or relative) if the system was activated and no one
answered the PERS operator. If the client did not re-
spond after the system was activated and no designated
responder could be reached, 9-1-1 was activated. At the
time of installation, use of the device was explained to
Lee et al.?IMPACT OF PERS ON ELDERS AFTER A FALL
the patient by a trained technician. The study was de-
signed as an effectiveness trial32; therefore, we did not
monitor subjects’ compliance with use of the PERS sys-
tem. After completion of the study, patients were free
to continue or stop use of the PERS at their own discre-
tion and cost.
The study coordinator and a trained research assistant
collected all data. Baseline assessment was performed ei-
ther in the ED or at the subjects’ homes within 72 hours
of discharge from the ED. Participants underwent a
structured telephone interview between 30 and 37 days
after their discharge from the ED to assess the primary
outcome and between 60 and 67 days to assess subse-
quent health care use. In addition, health care use was
assessed by searching the hospital’s database for any
record of return visits for one year after the index ED
encounters. Any discrepancy between participants’ self-
reports and the hospital database was resolved by
reviewing participants’ health records and by consulting
family members or the family physician. All data were
double entered into a computerized database by trained
data entry personnel, and the data were screened for
missing, extreme, and impossible values using an
automated algorithm before closing the database (SAS
version 9.1; SAS Institute, Inc., Cary, NC).
The primary outcome, anxiety, was measured using the
mean change in Hospital Anxiety and Depression Scale
anxiety subscale (HADS-A) score from baseline to 30
days after the ED visit. The HADS-A has been extensively
validated and found to be reliable and responsive to the
effects of a variety of treatments for anxiety in elder med-
ical patients.34It consists of seven Likert-type scale items,
each with possible scores ranging from 0 to 3. Thus,
HADS-A scores range from 0 to 21, with higher scores
indicating more anxiety. Fear of falling was measured
using the previously validated modified Tinetti Falls Effi-
cacy Scale (mFES) at baseline and again between 30 and
37 days.35To complete the mFES, subjects were asked to
rate their confidence in performing ten everyday tasks on
a 0–10 verbal numeric rating scale, yielding scores rang-
ing from 0 to 100. The mFES differs from the original
FES only in that higher scores indicate more confidence
and less fear of falling. In addition to the mFES, patients
were asked to rate their fear of being incapacitated on a
Figure 1. Study flowchart. PERS = personal emergency response system.
ACAD EMERG MED?April 2007, Vol. 14, No. 4?www.aemj.org
ten-point verbal numeric rating scale (higher scores indi-
cate less fear of being incapacitated). Baseline mental
status was measured using the Folstein Mini-Mental
State Examination.36We used the five-item Triage Risk
Stratification Tool to assess baseline risk of RTED.37We
recently validated the ability of the Triage Risk Stratifica-
tion Tool to predict RTED in our setting.38Health care
use that was captured included subsequent visits to any
physician, RTED, hospitalizations, and total length of
Data were analyzed blinded to the intervention code us-
ing SAS version 9.1 (SAS Institute, Inc.). For our primary
outcome (HADS-A score) and the secondary outcomes
(mFES score, fear of incapacitation, and hospital length
of stay), we compared the estimated mean difference
(treatment ? control groups) using the Wilcoxon rank
sum test and report the 95% confidence intervals (CIs)
around the mean differences. For the proportion of pa-
tients who RTED or were hospitalized within 60 days,
we report the estimated relative risk and 95% CIs.
The sample size estimate was based on the ability to
detect a difference in anxiety, our primary outcome,
measured by the HADS-A. There is no definition in the
literature for the minimal clinically important difference
for the HADS-A. Two previous large studies of commu-
nity-dwelling elders found that a three-point change in
HADS-A score represented two standard deviations
from their baseline assessment.39Using three points as
the minimal clinically important difference yielded a sam-
ple size requirement of only 11 subjects per group. Such
a small sample size would not permit analysis of second-
ary end points and might be judged to lack face validity.
Thus, a more conservative definition of the minimal clin-
ically important difference for HADS-A of 1.5 points was
chosen. Because of uncertainty regarding the distribu-
tion of HADS-A scores in our target population, we
used the Wilcoxon rank sum test to estimate the required
sample size. Given a standard deviation of 3.4 and an a of
0.05, with 43 subjects per group, the study would have
>95% power to detect a between-group reduction of
1.5 points from 6.0 to 4.5 in the mean HADS-A score.
Other proportions were tested using the chi-square sta-
Patient flow into the study is described in Figure 1. A
total of 107 subjects were randomized (53 in the PERS
group and 54 in the control group), but 21 subjects did
not complete the study. One subject in each group was
randomized but did not receive the allocated inter-
vention, because the emergency physician had initially
planned to discharge him or her but he or she was actu-
ally admitted to the hospital. Three subjects in the inter-
vention group and one subject in the control group
were lost to follow-up. Fifteen subjects did not complete
the study (nine in the PERS group and six in the control
group). The reasons given for dropping out are pre-
sented in Figure 1. One patient in each group died, but
all others were alive and living independently at one
month. Participants were similar to nonparticipants in
mean age, gender, and presenting complaint. The study
their ED visit. The overall mean age of participants was
79.7 years (range, 70–93 years, SD ? 5.88), 72% were
female, and their mean Mini-Mental State Examination
score was 27.9 (SD ? 2.81). Participants randomized to
treatment were similar to controls at baseline in terms of
demographic characteristics, as well as baseline anxiety
score and fear of falling (see Table 1).
There was no statistically significant difference be-
tween groups for our primary outcome. The difference
between the treatment and control groups in mean
change in anxiety scores from baseline to 30 days was
+0.35 (95% CI = ?1.5 to 0.76; p = 0.55; see Table 2). We
performed a post-hoc subgroup analysis to determine
whether high-risk subjects might benefit from a PERS.
Addition of a PERS did not appear to be beneficial
among patients with either higher baseline anxiety
(HADS-A score >7) or increased risk for RTED (Triage
Risk Stratification Tool scores >2; see Table 2).37
Fear of falling decreased slightly more in the treatment
group compared with the control group (corresponding
to an increase in mFES score of 5.0 in the treatment
group and 0.6 in the control group). However, the differ-
ence between the two groups was not significant (mean
change in mFES score in the treatment group ? mean
change in mFES score in the control group = 4.5; 95%
CI = ?6.7 to 15.7; Wilcoxon rank sum test = 1826.0;
p=0.70). Therewasalsonosignificantdifference between
by 0.4 in the treatment group and increased by 0.4 in the
control group (mean change in the treatment group ?
mean change in the control group = ?0.8; 95% CI = ?2.3
to 0.7; Wilcoxon rank sum test = 1,691.0; p = 0.31).
There was also no important difference between the
groups in subsequent health care use. Eight of 43 patients
(19%) returned to the ED within 60 days of the fall in both
the treatment and control groups (risk difference treat-
ment ? control, 0.0; 95% CI = ?16% to 16%; p = 1.0). In
the treatment group, three of 43 (7%) were hospitalized
by 60 days versus six of 43 (14%) in the control group
(risk difference treatment ? control = ?7.0%; 95% CI =
Baseline Characteristics of Patients in the Treatment and Control
(N = 86)
(n = 43)
(n = 43)
Mean age (yr)
Lives alone (%)
Mean Triage Risk Stratification
Mean Mini-Mental State
Initial HADS-A score >7 (%)
Initial mean HADS-A score
Initial mean Falls Efficacy
All p-values were nonsignificant.
HADS-A = Hospital Anxiety and Depression Scale anxiety subscale.
Lee et al.?IMPACT OF PERS ON ELDERS AFTER A FALL
was 0.2 days in the treatment group versus 1.7 days in the
control group (length of stay, treatment ? controls = ?1.5
days; 95% CI = ?3.4 to 0.3 days; Wilcoxon rank sum test =
1,939.0; p = 0.27).
Recurrent falls occurred in nine of 86 subjects (10.5%;
95% CI = 4.0% to 16.9%), with four in the control group
and five in the treatment group (chi-square = 0.124; p =
0.742). Of subjects who fell again, eight of nine (88.9%)
were injured (one humeral fracture, one pelvic fracture,
two lacerations, and four contusions). Four patients in
the treatment group were unable to get up without assis-
tance compared with two in the control group (p = 0.40).
To the best of our knowledge, this study is the first ran-
domized trial to assess the impact of PERS availability
on anxiety and fear of falling using validated outcome
measures. In contrast with previous literature, this study
did not demonstrate any psychological benefits or re-
duced health care use among older persons discharged
from the ED after a fall.
Two previous Canadian studies have found that intro-
duction of a PERS reduced health care use.19,20The first,
based in British Columbia, studied a convenience sample
of 106 community-living subjects with a median age of
79.0 years who subscribed to a hospital-based PERS
and retrospectively compared their health care use for
one year before and after subscription to the system.20
They found a statistically significant decrease in the
number of hospital admissions and total inpatient days
between the one-year before and after periods but no
reduction in ED visits. Similarly, a study conducted in Al-
berta found reduced hospitalization rates and lengths of
stay during a 17-month follow-up period compared with
a one-year historical control period.19Users and their
caregivers in both studies reported heightened feelings
of security after subscribing to the PERS, but anxiety
was not formally assessed. Problems with use of histori-
cal controls have been documented elsewhere.40
Only one previous randomized clinical trial, conducted
among residents of a subsidized community housing
project in the Boston area, has assessed the impact of
a PERS on anxiety.23The investigators interviewed resi-
dents who responded to a mailed survey letter and en-
rolled consenting subjects if they fit into one of three
groups. Group 1 consisted of subjects who were judged
to be severely functionally impaired but not socially iso-
lated, group 2 consisted of subjects who were severely
functionally impaired and socially isolated, and group 3
consisted of subjects who were socially isolated but not
functionally impaired. Patients randomized to receive
the PERS were less likely to have an emergency to which
no one responded (e.g., being unable to get up after a
fall). The investigators reported a reduction in anxiety
in group 1 only, which they described as significant
(p = 0.09), even though the difference between groups
in anxiety was not statistically significant at the standard
0.05 level. Despite this, the results of this sole previous
randomized trial have been cited in several subsequent
reports as providing evidence for the psychological ben-
efits of PERS.16,18,20,30,31
The present study was specifically designed to evaluate
the impact of a PERS on anxiety using the HADS-A, a
reliable, valid, and responsive measure of anxiety.39,41,42
No decrease in anxiety associated with use of the PERS
It is unlikely that the lack of a statistically significant re-
duction in anxiety was due to a ‘‘floor’’ effect because our
patients, who had all fallen before presenting to the ED,
demonstrated significant baseline anxiety. The mean ini-
tial anxiety score of patients in this study was 5.1, com-
pared with a mean HADS-A score of 2.5 among 2,158
healthy community-dwelling Dutch persons older than
65 years.39As another comparator, patients in a highly
anxiety-provoking situation (newly diagnosed breast
cancer) reported mean HADS-A scores between 5.9
and 7.1, similar to our patients.42Thus, subjects in the
present study had relatively high anxiety at baseline,
with room for improvement. The 95% CI around the dif-
ference in anxiety between the treatment groups sug-
gests that any potential impact that PERS might have
on anxiety would be extremely small. Finally, post-hoc
subgroup analysis did not reveal any impact of the
PERS among the 36 subjects with baseline HADS-A
scores >7 (see Table 2), which is a commonly used cutoff
to define high risk for anxiety disorders.
Regarding our secondary outcomes, the hypothesis of
the study was that use of a PERS might reduce fear of
falling, based on previous research demonstrating that
both anxiety and social isolation are predictive of fear
of falling.8,35,43As expected, both treatment and control
patients exhibited substantial fear of falling. The mean
baseline mFES score in this study (76.3) is similar to the
mean mFES score (78.0) for 158 community-dwelling
Analysis of Intervention: Primary Outcome and Effect among Subgroups with High Anxiety and TRST Scores
Mean Change in HADS-A,
Day 30 ? Day 1,
Control Group (n)
Mean Change in HADS-A,
Day 30 ? Day 1,
Treatment Group (n)
Treatment ? Control
TRST score <2
TRST score >2
HADS-A score <7
HADS-A score >7
?0.12 (43)0.35 (?0.76, 1.5)*
0.06 (?1.4, 1.5)
0.68 (?1.1, 2.5)
0.53 (?0.56, 1.6)
?0.08 (?2.2, 2.0)
TRST = Triage Risk Stratification Tool; HADS-A = Hospital Anxiety and Depression Scale anxiety subscale.
*Wilcoxon rank sum = 1,801.0, p = 0.55. All other p-values were nonsignificant.
ACAD EMERG MED?April 2007, Vol. 14, No. 4?www.aemj.org
subjects who stated they were afraid of falling and sub-
stantially lower (i.e., indicating greater fear) than the
mean mFES score (89.4) for 370 subjects who were not
afraid of falling.44We found a nonsignificant trend to-
ward reduced fear of falling in the treatment group;
however, the CIs suggest that any difference between
the groups was too small to be clinically important.
We did not find that use of a PERS significantly im-
pacted rates of RTED. The hospitalization rate and
mean days admitted to the hospital were lower in the
PERS group (0.2 days vs. 1.7 days in the control group),
but this difference did not achieve statistical significance,
possibly due to low power. Of interest, the previous Brit-
ish Columbian19,20study also found no impact on RTED
but did demonstrate a statistically significant decrease
in the mean number of hospital admissions per year
(from 1.18 to 0.88) and a decrease in the mean number
of inpatient days per person per year (14.37 to 5.85).
Future larger randomized trials would be needed to def-
initely assess the impact of a PERS on hospitalizations
and mean length of hospitalization.
Limitations of this study include potential selection bias.
We were surprised by the number of subjects who
declined participation (285) or withdrew after randomi-
zation (15). This may have been due to our broad inclu-
sion criteria; any person older than 70 years who had
fallen was approached for consent. If these eligible sub-
jects who declined or withdrew from participation were
likely to benefit from the intervention, this might have
negatively biased our estimate of the impact of the
PERS. However, participants, nonparticipants, and those
who dropped out had similar mean age, gender, and pre-
senting complaints. The most common reason why eligi-
ble subjects refused to participate was because they were
too busy or did not believe they required a PERS.
Conversely, this study could be criticized because we
did not restrict enrollment to patients most likely to ben-
efit (i.e., those with high baseline levels of anxiety, func-
tional limitations, or social isolation). Selection of such
high-risk groups would involve measures not routinely
used in the ED. Instead, we used the presenting com-
plaint of a fall requiring an ED visit among patients older
than 70 years to identify potentially frail elders, because
we were interested in the pragmatic question of whether
addition of the PERS benefited an easily identifiable ED
population. However, it is possible that a PERS may ben-
efit specific high-risk groups. Future studies should de-
velop screening criteria that are capable of identifying
subjects most likely to benefit and that can realistically
be applied in the ED setting.
Another potential study bias may have arisen if the
GEM nurse assessment and discharge planning had a
significant therapeutic impact, operating as an ‘‘active
control.’’ Randomization and blinding of the GEM nurse
to group assignment to a large degree protected against
any differences in the way discharge planning was ap-
plied to each group. However, if GEM nurse discharge
planning considerably reduced anxiety for both groups,
this could hide the impact of the PERS on anxiety. How-
ever, the assessment by the GEM nurse was aimed at
finding previously unrecognized geriatric conditions, in-
quiring about fall hazards in the home, educating the
patient to avoid potential risky activities, or identifying
medications that might increase the client’s risks of fall-
ing and was not focused on patient reassurance. There
was no patient follow-up by the GEM nurse after dis-
charge that might have reduced patient anxiety. The
high baseline anxiety levels, and the fact that there was
no significant reduction in anxiety among either the
treatment or control subjects, further argues against an
active control effect. However, we cannot definitively ex-
clude the possibility that PERS use might be beneficial in
settings without access to GEM nurse–assisted discharge
planning. The follow-up period of 30 days is another po-
tential limitation of the study. We chose this time frame
because we were interested in the short-term impact on
anxiety and potential health care use. It is possible that
anxiety reduction might accrue over longer periods.
This study was inadequately powered to detect poten-
tially clinically important impacts of the PERS on the sec-
ondary outcomes of RTED, subsequent hospitalization,
and length of hospital stay. In fact, there was a nonsignif-
icant trend toward reduced lengths of hospital stay. A
larger study is required before definitive conclusions
can be made about the impact of a PERS on length of
Finally, this study did not assess the major benefit of
the PERS: to provide assistance in the event of incapacita-
tion or a health emergency. Previous data demonstrate
that these events are rare (2.9% out of 1,931 activations
over a 29-month period required treatment),23so a much
larger study would be needed to evaluate this outcome.
Given that PERS systems cost between $200 and $600
per client per year, future studies should also assess the
cost benefit of PERS use.
In contrast to previous literature, this study found no ev-
idence that addition of a PERS to assessment by a GEM
nurse reduced patient anxiety or fear of falling among
older persons discharged from the ED to home. This
study did not assess the fundamental role of the PERS,
to provide assistance in the event of a health emergency,
and was not sufficiently powered to exclude an impact on
hospitalizations or length of hospital stay. However, the
findings of this study raise questions about the purported
psychological benefits arising from the addition of a
PERS to the discharge planning of older persons dis-
charged from the ED after a fall.
The authors thank Lifeline Corp. for providing use of the
personal emergency response systems during the trial and
Dr. Gary Naglie for his careful revision of this manuscript.
1. Burt CW, Fingerhut LA. Injury visits to hospital emer-
gency departments: United States, 1992-95. Vital
Health Stat 13. 1998; (131):1–76.
2. Vellas B, Cayla F, Bocquet H, de Pemille F, Albarede
JL. Prospective study of restriction of activity in old
people after falls. Age Ageing. 1987; 16:189–93.
Lee et al.?IMPACT OF PERS ON ELDERS AFTER A FALL
3. Tinetti ME, Speechley M, Ginter SF. Risk factors for
falls among elderly persons living in the community.
N Engl J Med. 1988; 319:1701–7.
4. Tinetti ME. Factors associated with serious injury
during falls by ambulatory nursing home residents.
J Am Geriatr Soc. 1987; 35:644–8.
5. Aminzadeh F, Dalziel WB. Older adults in the emer-
gency department: a systematic review of patterns
of use, adverse outcomes, and effectiveness of inter-
ventions. Ann Emerg Med. 2002; 39:238–47.
6. Nevitt MC, Cummings SR, Kidd S, Black D. Risk fac-
tors for recurrent nonsyncopal falls. A prospective
study. JAMA. 1989; 261:2663–8.
7. Walker JE, Howland J. Falls and fear of falling among
elderly persons living in the community: occupational
therapy interventions. Am J Occup Ther. 1991; 45:
8. Howland J, Lachman ME, Peterson EW, Cote J,
Kasten L, Jette A. Covariates of fear of falling and
associated activity curtailment. Gerontologist. 1998;
9. Howland J, Peterson EW, Levin WC, Fried L, Pordon
D, Bak S. Fear of falling among the community-dwel-
ling elderly. J Aging Health. 1993; 5:229–43.
10. Arfken CL, Lach HW, Birge SJ, Miller JP. The preva-
lence and correlates of fear of falling in elderly per-
sons living in the community. Am J Public Health.
11. Tinetti ME, Mendes de Leon CF, Doucette JT, Baker
DI. Fear of falling and fall-related efficacy in relation-
ship to functioning among community-living elders.
J Gerontol. 1994; 49:M140–7.
12. Grief CL. Patterns of ED use and perceptions of the
elderly regarding their emergency care: a synthesis
of recent research. J Emerg Nurs. 2003; 29:122–6.
13. Center for Disease Control and Prevention. Data
warehouse on trends in health and aging. Available
trenddata.htm. Accessed Nov 23, 2006.
14. Rice DP, Fineman N. Economic implications of in-
creased longevity in the United States. Ann Rev Public
Health. 2004; 25:457–73.
15. Wilber ST, Gerson LW. A research agenda for geriat-
ric emergency medicine. Acad Emerg Med. 2003; 10:
bound elderly. J Geriatr Psychiatry. 1981; 14:111–3.
17. Dibner A. Personal emergency response systems:
communication technology aids elderly and their
family. J Appl Gerontol. 1990; 9:504–10.
18. Dibner AS. Personal response services present and
future. Home Health Care Serv Q. 1992; 13:239–43.
19. Koch WJ. Emergency response system assists in dis-
charge planning. Dimens Health Serv. 1984; 61(11):
20. Roush RE, Teasdale TA, Murphy JN, Kirk MS. Impact
of a personal emergency response system on hospital
utilization by community-residing elders. South Med
J. 1995; 88:917–22.
21. Dibner AS. Linking frail elderly to the help network.
Nurs Homes. 1982; 31(6):12–5.
22. Ruchlin HS, Morris JN. Cost-benefit analysis of an
emergency alarm and response system: a case study
of a long-term care program. Health Serv Res. 1981;
23. Sherwood S, Morris J. A study on the effects of an
emergency alarm system for the aged: a final report.
Boston, MA: Hebrew Rehabilitation Center for Aged,
1980. Grant no. HSO1788.
24. Interbank fx. Available at: http://www.edgar-online
0001193125-04-040562&glm = 1&nav = 1. Accessed Mar
25. Benson WF. Public financing for personal response
systems: a federal viewpoint. Home Health Care Serv
Q. 1992; 13:223–8.
26. Hearing held on Medicare coverage of personal
emergency response systems. Caring. 1983; 6:6.
27. Federal Trade Commission. Facts for consumers:
personal emergency response systems. Available at:
.htm. Accessed Nov 8, 2006.
28. Ford JD, Trestman RL, Steinberg K, Tennen H, Allen
S. Prospective association of anxiety, depressive, and
addictive disorders with high utilization of primary,
specialty and emergency medical care. Soc Sci Med.
29. Ford JD, Trestman RL, Tennen H, Allen S. Relation-
ship of anxiety, depression and alcohol use disorders
to persistent high utilization and potentially problem-
atic under-utilization of primary medical care. Soc Sci
Med. 2005; 61:1618–25.
30. Hyer K, Rudick L. The effectiveness of personal emer-
gency response systems in meeting the safety moni-
toring needs of home care clients. J Nurs Admin.
31. Bernstein M. ‘‘Low-tech’’ personal emergency re-
sponse systems reduce costs and improve outcomes.
Manag Care Q. 2000; 8:38–43.
32. Cochrane A. Effectiveness and Efficiency: Random
Reflections on the Health Services. London, England:
Nuffield Provinical Hospitals Trust, 1972.
33. Pocock SN, Grossfield K. A socially acceptable emer-
34. Herrmann C. International experiences with the
Hospital Anxiety and Depression Scale—a review of
validation data and clinical results. J Psychosom
Res. 1997; 42:17–41.
35. Tinetti ME, Richman D, Powell L. Falls efficacy as a
36. Folstein MF, Folstein SE, McHugh PR. ‘‘Mini-mental
state.’’ A practical method for grading the cognitive
state of patients for the clinician. J Psychiatr Res.
37. Meldon SW, Mion LC, Palmer RM, et al. A brief risk-
stratification tool to predict repeat emergency de-
partment visits and hospitalizations in older patients
discharged from the emergency department. Acad
Emerg Med. 2003; 10:224–32.
38. Lee J, Langevin M, Schwindt G, et al. Validation of
the Triage Risk Stratification Tool (TRST) scale to
identify elders at risk for returning to the emergency
department. J Am Geriatr Soc. 2005; 53:S131.
39. Spinhoven P, Ormel J, Sloekers PP, Kempen GI,
Speckens AE, Van Hemert AM. A validation study
ACAD EMERG MED?April 2007, Vol. 14, No. 4?www.aemj.org
(HADS) in different groups of Dutch subjects. Psy-
chol Med. 1997; 27:363–70.
40. Nading JH. Historical controls for extracorporeal
membrane oxygenation in neonates. Crit Care Med.
41. Zigmond A, Snaith R. The hospital anxiety and de-
pression scale. Acta Psychiatr Scand. 1983; 67:361–70.
42. McArdle JM, George WD, McArdle CS, et al. Psycho-
logical support for patients undergoing breast cancer
surgery: a randomised study. BMJ. 1996; 312:813–6.
43. Friedman SM, Munoz B, West SK, Rubin GS, Fried
LP. Falls and fear of falling: which comes first? A lon-
gitudinal prediction model suggests strategies for
primary and secondary prevention. J Am Geriatr
Soc. 2002; 50:1329–35.
44. Cumming RG, Salkeld G, Thomas M, Szonyi G. Pro-
spective study of the impact of fear of falling on activ-
ities of daily living, SF-36 scores, and nursing home
admission. J Gerontol A Biol Sci Med Sci. 2000; 55:
Lee et al.?IMPACT OF PERS ON ELDERS AFTER A FALL