Changes in Physical Therapy Providers’
Use of Fall Prevention Strategies
Following a Multicomponent
Behavioral Change Intervention

Background and Purpose. An abundance of evidence suggests that interven-
tions targeting fall risk factors are effective; however, it remains unknown
whether, or to what extent, this body of evidence has affected the clinical
practice of physical therapy providers. The purposes of this study were: (1) to
describe knowledge of, and attitudes toward, fall risk factors and fall reduction
strategies; (2) to assess self-reported use of fall reduction strategies with
patients; and (3) to identify factors associated with increased use of fall
reduction strategies with patients among physical therapy providers exposed
to a behavioral change strategy. Subjects and Methods. A cross-sectional survey
of physical therapy providers from hospital-based and freestanding outpatient
physical therapy facilities throughout north-central Connecticut was conducted
between October 2002 and April 2003. The participants were 94 physical therapy
providers who had been exposed to the Connecticut Collaboration for Fall
Prevention (CCFP) behavioral change effort. The CCFP program uses multicom-
ponent professional behavioral change strategies to embed fall risk factor
assessment and management, based on evidence from randomized controlled
trials, into the clinical care of older patients. A telephone questionnaire—
focusing on fall risk factor knowledge and attitudes and self-reported fall risk
factor assessment and management practices before and after exposure to the
CCFP efforts—was administered to consenting physical therapy providers. Results.
Environmental hazards and gait and balance deficits were named as fall risk
factors by 86 (91%) and 73 (78%) participants, respectively. All of the targeted
risk factors were mentioned by at least 30% of the participants. Sixty-four
participants (68%) reported increased fall reduction practice behaviors. The area
of multiple medications was noted most frequently, with 77 participants (82%)
noting new practices related to medication use. Only knowledge of fall risk factors
and pre-CCFP behaviors were associated with increased fall reduction practices.
Discussion and Conclusion. Physical therapy providers reported an increase in
practice behaviors in response to the multicomponent behavioral change strat-
egy. Knowledge of fall risk factors was associated with increased fall reduction
practice behaviors, most likely due to the focused nature of the education
strategy. [Brown CJ, Gottschalk M, Van Ness PH, et al. Changes in physical
therapy providers’ use of fall prevention strategies following a multicomponent
behavioral change intervention. Phys Ther. 2005;85:394 – 403.]

Key Words: Behavioral change, Fall prevention, Physical therapy, Rehabilitation.

Cynthia J Brown, Margaret Gottschalk, Peter H Van Ness, Richard H Fortinsky, Mary E Tinetti

394

Physical Therapy . Volume 85 . Number 5 . May 2005

Research

Report

䢇

ўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўў

ўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўўў

ўўўўўўўўўўўў

O

ne third of community-dwelling adults over
age 65 years fall each year.

1

People who

experience a fall are at increased risk for
subsequent falls.

2

Major injuries, including

fractures, head trauma, and soft tissue injuries, occur in
about 10% of individual falls.

1

Falls have been associated

with decreased physical and social functioning as well as a
3- to 6-fold increased risk for nursing home placement.

3

Observational studies

2,4,5

have shown falls in community-

dwelling older people to be associated with several risk
factors, including muscle weakness, gait and balance
deficits, polypharmacy, and postural hypotension. Fall-
ing appears to result from the accumulated effect of
these and other multiple risk factors.

2,6

In several ran-

domized controlled trials (RCTs),

1,5,7–11

both single and

multifactorial intervention strategies have proven effec-
tive in both decreasing these risk factors and reducing
falls. In particular, physical therapy interventions,
including gait training, progressive balance exercises,
and home evaluation for environmental hazards, have
proven effective, both as a single intervention and as part
of a multifactorial intervention strategy.

12

Although a

wealth of evidence suggests that interventions targeting
fall risk factors are effective, it remains unknown
whether, or to what extent, this body of evidence has

affected the clinical practice of outpatient physical ther-
apy providers.

The Connecticut Collaboration for Fall Prevention
(CCFP) program is an ongoing, community-wide effort
in the north-central Connecticut area to translate RCT
evidence into clinical practice. The objective is to embed
multifactorial fall risk factor assessment and manage-
ment throughout the health care system. The focus of
the CCFP effort is on health care providers caring for
ambulatory, community-living older adults, the group
for which evidence of effectiveness of fall prevention
efforts is the strongest. Physicians, nurses, discharge
planners, and physical therapists and occupational ther-
apists working in home health care, ambulatory settings,
and hospitals are involved in the CCFP effort. Out-
patient physical therapy providers, the focus of the
current study, represent one arm of the ongoing CCFP
effort.

In a sample of outpatient physical therapy providers
exposed to efforts to increase knowledge and promote
relevant risk factor assessment and management, the
aims of this study were: (1) to describe the physical
therapy providers’ knowledge of, and attitudes toward,
fall risk factors and fall reduction strategies; (2) to assess

CJ Brown, MD, is Investigator, Birmingham/Atlanta VA Geriatric Research, Education, and Clinical Center, and Assistant Professor, Department
of Medicine, University of Alabama at Birmingham. Address all correspondence to Dr Brown at University of Alabama at Birmingham, VAMC
GRECC 11-G Room 8225, 1530 3rd Ave S, Birmingham, AL 35294-0001 (USA) ([e-mail]).

M Gottschalk, PT, MS, is Staff Physical Therapist, Department of Rehabilitation Services, Yale-New Haven Hospital, New Haven, Conn.

PH Van Ness, PhD, MPH, is Lecturer, Department of Epidemiology and Public Health, and Associate Research Scientist/Senior Biostatistician,
Department of Internal Medicine, Yale University School of Medicine, New Haven, Conn.

RH Fortinsky, PhD, is Professor of Medicine, Center on Aging, University of Connecticut Health Center, Farmington, Conn.

ME Tinetti, MD, is Professor, Department of Epidemiology and Public Health, and Professor, Department of Internal Medicine, Yale University
School of Medicine.

Dr Brown, Ms Gottschalk, Dr Fortinsky, and Dr Tinetti provided concept/idea/research design. Dr Brown and Dr Tinetti provided writing and
project management. Dr Brown provided data collection, and Dr Brown and Dr Tinetti provided data analysis. Ms Gottschalk provided subjects.
Dr Tinetti provided fund procurement and institutional liaisons. Dr Brown, Ms Gottschalk, Dr Van Ness, and Dr Fortinsky provided consultation
(including review of manuscript before submission). The authors thank the physical therapy providers who participated in the study; Paula Clark,
RN, for assistance in data collection; Bridget Mignosa and Virginia Towle for assistance with data management; and Grace Jeng, MD, and Lisa M
Walke, MD, for assistance with coding.

This study was approved by the Institutional Review Board of Yale University School of Medicine.

This project was supported, in part, by a grant from the Donaghue Foundation and by a Yale Pepper Center grant (P60AG10469) from the National
Institute on Aging. Dr Brown was supported by a training grant from the National Institute on Aging (T32AG19134) and is a recipient of a John
A. Hartford Foundation/American Federation for Aging Research Academic Geriatrics Fellowship Program Award (R04191) and a training
support grant from the Hartford Foundation-funded Southeast Center of Excellence in Geriatric Medicine.

An abstract of this research was presented at the Annual Meeting of the American Geriatric Society; May 14 –18, 2003; Baltimore, Md.

This article was received December 15, 2003, and was accepted November 4, 2004.

Physical Therapy . Volume 85 . Number 5 . May 2005

Brown et al . 395

their self-reported behaviors and practices related to use
of fall reduction strategies for their patients; and (3) to
identify factors associated with an increase in use of fall
reduction strategies for patients in their practices.

Method

Setting and Participants
North-central Connecticut was the designated interven-
tion area and includes all of Hartford and surrounding
suburbs and rural areas. Of the more than 871,000
people residing in the north-central Connecticut area,
11% are 70 years of age or older.

13

All hospital-based and

freestanding outpatient physical therapy practices in this
area that see patients 70 years of age or older were
identified via 2 methods. First, a physical therapy pro-
vider familiar with the practices in the area compiled a
list, and second, telephone book listings were accessed
through several Web sites via the Internet. Practices were
contacted by letter to inform them of the CCFP effort
and were then individually contacted by telephone and
an outreach visit was scheduled. A total of 129 outpatient
rehabilitation practices were identified, including both
freestanding and hospital-based practices. Of those, 8
practices refused to participate and 2 practices did not
provide data on their physical therapy providers. The
sample of physical therapy providers, therefore, was
derived from 119 practices. Three hundred physical
therapy providers were employed in these practices.
Potential participants for the present study included all
outpatient physical therapy providers, including physical
therapists and physical therapist assistants, who had been
exposed to CCFP efforts, defined as having received one
outreach visit by April 1, 2003. One hundred forty-two
physical therapy providers met this criterion. Ninety-four
providers eventually participated in the study. Figure 1
presents the flow of potential participants and the num-
ber of individual physical therapy providers affiliated
with the targeted 119 outpatient rehabilitation practices.

The number of physical therapy providers interviewed
from each practice ranged from 1 to 5. Individual
participants were excluded only if they relocated prior to
completion of a questionnaire or if an interview could
not be scheduled.

Baseline characteristics of the cohort are shown in Table
1. Overall, the group was predominantly female (69%),
and the majority were physical therapists (91%). Direct
patient care accounted for the majority of their work
hours (mean hours per week

⫽31.0, SD⫽9.6), with older

patients (70 years of age and older) accounting for one
third of those hours.

Intervention
Based on the available literature that suggests that
multiple approaches are most effective when profes-
sional and organizational behavioral change is the objec-
tive,

14

the CCFP program uses a multifaceted approach

to encourage the adoption of fall risk factor assessment
and management in the care of older patients. The
professional and organizational behavioral change strat-
egies used include outreach visits; procuring support for
the CCFP effort by facility administrators and supervi-
sors; training manuals with instructions for implement-
ing components of fall risk assessment and management;
patient and physical therapy provider risk factor check-
lists and patient handouts; a Web site from which all
materials could be downloaded; working groups of local
physical therapy providers who adapted the fall preven-
tion protocols for practical implementation and who
worked with investigators in encouraging other physical
therapy providers to implement fall prevention prac-
tices; “opinion leaders,” defined as providers whose
opinion was valued by the rest of the group; encourage-
ment of early adopters, or those providers who are first
to adopt new ideas and put into practice new behaviors;
newsletter reminders to the physical therapy providers
about various aspects of the CCFP program; and media
attention to heighten community awareness of falls as an
important clinical problem.

14 –21

All of these strategies

were in place in the north-central Connecticut area
during the study. The CCFP efforts focus on 6 risk
factors based on RCT evidence that interventions target-
ing these risk factors are effective at reducing fall
rates.

1,5,7–11

These risk factors are: gait or balance impair-

ments, multiple medications, postural hypotension, sen-
sory and perceptual deficits, foot and footwear prob-
lems,

22,23

and environmental hazards.

The outreach visits were the primary strategy and
included a presentation, in the physical therapy provid-
ers’ offices, of the 6 risk factors along with the recom-
mended management and specific strategies for incor-
porating the assessment and management into their
clinical practice. A team that included a physician, a
nurse, and a physical therapist made the outreach visits,
and all physical therapy providers included in the study
were exposed to a visit at least once during the study
period. Sessions usually occurred before or after work or
during lunch and lasted for approximately 1 hour.
Strategies for fall-related assessment included examina-
tion of gait and balance; examination of the feet for
calluses, bunions, and nail problems; review of the
medications for number and types; and examination of
blood pressure in the supine and standing positions.
Fall-related interventions suggested for incorporation
into physical therapist management included progres-
sive balance exercises, gait training, referral to a podia-
trist if foot and footwear problems were noted or to the

396 . Brown et al

Physical Therapy . Volume 85 . Number 5 . May 2005

ўўўўўўўўўўўўўўўўўўўўўўўўўўўў

Figure 1.

Flow of potential participants. This figure shows the number of individual physical therapy providers affiliated with 119 outpatient rehabilitation
practices that did or did not ultimately receive a Connecticut Collaboration for Fall Prevention (CCFP) outreach visit. Asterisk indicates that of 129
target outpatient rehabilitation practices, 8 practices with 11 physical therapy providers declined participation in the CCFP effort and 2 outpatient
rehabilitation practices declined to provide information on the number of providers who were employed in the practices.

Physical Therapy . Volume 85 . Number 5 . May 2005

Brown et al . 397

primary care provider if the patient was noted to have
orthostatic hypotension or to be taking multiple medi-
cations. The physical therapy intervention was essentially
the same as that described for the Yale FICSIT trial.

24

Potential incentives and barriers also were addressed
during these visits, and easy-to-use materials were pro-
vided in the form of a training manual. The training
manual included the strategies for assessing and manag-
ing each of the risk factors and patient handouts. A
one-page evaluation and management form also was
developed for use in the patient charts. Because the
physical therapy providers practiced in the area where
the CCFP effort was occurring, they were exposed to the
additional behavioral change strategies described ear-
lier. During the outreach visit, names, addresses, and
telephone numbers of physical therapy providers were
obtained for the purpose of constructing a database. The
participants were informed that they would be contacted
at a later date to provide feedback about the CCFP
program.

Interview
Participants were contacted, by telephone or electronic
mail, at least 6 weeks after the outreach visit, with a range
of 6 to 24 weeks, and invited to complete an in-depth
telephone interview at their convenience. One of the
authors (CJB) or a trained research nurse administered
a telephone questionnaire to all consenting physical
therapy providers according to procedures approved by
the Institutional Review Board of Yale University School
of Medicine. The questionnaire was a mixture of open
and closed-ended questions designed to obtain qualita-

tive and quantitative information that focused on knowl-
edge, attitudes, and self-reported practice behaviors con-
cerning fall risk assessment and management. Closed-
ended questions were scored using a Likert-type scale,
with responses ranging from 1 to 4. For example,
attitude questions asked how important the physical
therapy providers thought the risk factor was for man-
aging patients in their clinical practice, and responses
for these questions were scored as: 1

⫽“not very impor-

tant,” 2

⫽“somewhat important,” 3⫽“moderately impor-

tant,” and 4

⫽“very important.” Responses varied

depending on the question asked, but all closed-ended
questions had 4 possible answers, with 1 being the lowest
rating and 4 being the highest rating. Demographic data
and practice-related information also were obtained.
The time required to complete the telephone question-
naire ranged from 25 to 40 minutes.

To assess interrater reliability, the 2 interviewers each
administered the questionnaire to the same 8 physical
therapy providers within a 48-hour period. Interviewers
were masked to each other’s questionnaire results. Test-
retest reliability data were assessed in 9 participants by
repeating the interview 7 days after the initial interview
without knowledge of the original responses. For both
interrater and test-retest reliability data, weighted kappa
statistics were calculated for the 17 four-level data ques-
tions (ie, all questions that asked participants to rate on
a Likert scale from 1 to 4). Weighted summary kappas of
.65 and .64 were obtained for interrater and test-retest
reliability, respectively, suggesting good reliability.

25

Answers that were either 2 points more or less than the
previous score on the ordinal scale were considered to
be outliers, an event that occurred in less than 4% of the
responses.

Outcome Measures
The primary outcome measures were the physical ther-
apy providers’ post-CCFP self-report of use of fall pre-
vention strategies with their patients and a change in use
of fall prevention strategies from before to after expo-
sure to CCFP outreach efforts. Participants were asked to
compare their fall prevention practice behaviors a year
ago (pre-CCFP) with their current behaviors (post-
CCFP). Specifically, they were asked to respond to the
following 2 questions: “Thinking back a year ago, how
often did you consider fall prevention in your routine
care of older adults?” and “In your clinical practice now,
how often do you consider fall prevention in your
routine care of older adults?” A change in behavior was
defined as changing one or more levels on the 4-level
ordinal scale, which ranged from “almost never or
never” to “almost always or always.” If changes had been
made, they were asked to give specific examples of the
behaviors that had changed. To examine the open-
ended questions about change in practice, the interview-

Table 1.

Characteristics of Study Sample (N

⫽94)

Characteristic

N (%) or X

ⴞSD

(Range)

Age (y)

37

⫾9.9 (23–60)

Sex

Male

29 (31)

Female

65 (69)

No. of years since graduation with most

advanced degree, median (range)

9 (1–38)

Employed

Full-time

75 (80)

Part-time

19 (20)

Type of provider

Physical therapist

86 (91)

Physical therapist assistant

8 (9)

Hours per week spent in direct patient care

31

⫾9.6 (8–50)

Hours per week spent caring for older adults

12.2

⫾8.5 (1–40)

Physical therapy providers by practice type

Hospital based

39 (41)

Freestanding

55 (59)

398 . Brown et al

Physical Therapy . Volume 85 . Number 5 . May 2005

ўўўўўўўўўўўўўўўўўўўўўўўўўўўў

ers recorded all answers. One of the authors (CJB) and
2 other researchers independently coded the practice
behaviors into broad categories. The coders discussed
the categories and reached a consensus about the cate-
gories to be used in the coding. Categories chosen by the
coders included an increase in assessment, education,
referral, or awareness; increased use of exercises or
training; and no change noted by the physical therapy
provider. The individual physical therapy provider’s
answers and the category in which the answers had been
placed were compared among the 3 coders. Any differ-
ences in coding were discussed and resolved by the
coders.

The secondary outcomes were knowledge of, and atti-
tudes toward, fall risk factor assessment and manage-
ment strategies. To assess knowledge, providers were
asked to list risk factors and available interventions that
had been presented during the outreach visit. The
number of risk factors and interventions listed were
summed to create a composite variable for risk factor
knowledge and knowledge of available interventions. To
assess attitudes, participants were asked to rate, using a
Likert-type scale, the importance of the 6 risk factors for
falls in patients in their clinical practice and their
confidence in the effectiveness of the strategies pre-
sented by the CCFP program, as previously described.

Data Analysis
Appropriate descriptive statistics, including frequencies,
proportions, means, standard deviations, and medians,
were developed for the characteristics of the study group
and for each of the questions addressing knowledge,
attitudes, and self-assessed behaviors. To investigate the
relationship among demographics, knowledge, atti-
tudes, and the self-reported change in fall prevention
behaviors, bivariate analyses using Kendall rank correla-
tion coefficients

26

were conducted. The Bowker test of

symmetry was used to test the equality of the self-rated
behaviors before and after exposure to the CCFP pro-
gram.

27

Subsequently, the independent contributions of

demographics (ie, age, sex, type of certification), knowl-
edge, and attitudes to change in practice behaviors were
explored using 2 logistic regression models. The first
model used a change score for self-rated behaviors from
before exposure to after exposure to the CCFP program.
Due to a skewed frequency distribution and small num-
bers in some categories, change scores were trichoto-
mized into those with a decrease or no change, an
increase of 1 point, or an increase of 2 or more points on
the ordinal scale. The second model examined only the
post-CCFP practice behavior score. In models with post-
CCFP behaviors as the outcome measure, self-reported
pre-CCFP exposure practice behaviors was included as an
independent variable. Models were fit using the backward
elimination method. Analyses were carried out using SAS

statistical software, version 8.01.* A probability value of less
than .05 was considered statistically significant.

Results
Table 2 presents the risk factors and available interven-
tions named by the physical therapy providers when
asked, in an open-ended fashion, to list any factors or
interventions they could. The most frequently named
risk factors were environmental hazards and gait or
balance impairments. Multiple medications was named
as a risk factor by 77% of the participants, a close third
in frequency of responses to gait or balance impair-
ments. Postural hypotension was named as a risk factor
by 30% of the physical therapy providers.

The participants rated the importance of each of the 6
risk factors emphasized during the CCFP outreach visits
and in the training materials to the patients in their
clinical practice. Impaired gait or balance was consid-
ered very important by 76 physical therapy providers
(81%). Multiple medications and environmental haz-
ards were both rated as “very important” by 48 partici-
pants (51%). Postural hypotension was rated as “very
important” by 43 participants (46%). The risk factor
receiving the lowest percentage of “very important”
ratings, by 36 participants (38%), was sensory and per-
ceptual deficits.

Figure 2 presents a comparison of the frequency of use
of fall prevention practice behaviors by examining the
physical therapy providers’ self-reported behaviors from
a year ago and their self-reported behaviors after expo-
sure to the CCFP behavioral change effort. Specifically,
participants were asked how often they considered fall
prevention in their clinical practice at the present time
and a year ago. The Bowker test of symmetry comparing
self-reported behaviors before and after exposure to the

* SAS Institute Inc, PO Box 8000, Cary, NC 27511.

Table 2.

Knowledge of Risk Factors and Interventions for Falls (N

⫽94)

a

Risk Factor

Named
Risk
Factor,
N (%)

Named an
Intervention
to Address
Risk Factor,
N (%)

Environmental hazards

86 (91)

67 (71)

Gait or balance impairments

73 (78)

90 (96)

Multiple medications

72 (77)

45 (48)

Sensory and perceptual deficits

54 (57)

10 (11)

Foot and footwear problems

44 (47)

35 (37)

Postural hypotension

28 (30)

14 (15)

a

Open-ended questions posed: “Name as many preventable risk factors for

falls in older adults as you can think of” and “Name as many interventions or
treatments that might help prevent falls in your older patients.”

Physical Therapy . Volume 85 . Number 5 . May 2005

Brown et al . 399

CCFP program showed the difference between the
scores was significant (P

⬍.0001) and is evidence that this

difference was not due to chance alone. Figure 3 pre-
sents the distribution of change in physical therapy
providers’ self-reported fall prevention behaviors from
before to after exposure to the CCFP program.

Table 3 presents the practice behaviors noted by the
physical therapy providers to have changed from a year
ago. We also analyzed the data excluding the data for the
physical therapist assistants and found the same results,
so data for the entire sample are presented in the table.
The area of multiple medications was noted most fre-
quently, with 77 participants (82%) noting a behavior
change related to medication use. Forty-one participants
reported referring patients to their primary care pro-
vider or pharmacist to review medications; 31 partici-
pants were now listing and reviewing patients’ medica-
tion lists at intake; and 32 participants were doing some
education, either verbally or with written handouts,
about medications and fall risk. The areas of foot and
footwear problems and sensory and perceptual deficits
had the smallest amount of practice behavior change,

with 50 and 58 providers noting no change in their
practice behavior, respectively.

Using multivariate logistic regression, we examined the
independent contribution of demographics (age, sex,
type of certification), post-CCFP knowledge, and atti-
tudes to the primary outcome of an increase in self-
reported fall prevention practice behaviors. In the multi-
variate model, using the change score from before
exposure to after exposure to the CCFP effort as the
outcome measure, only post-CCFP knowledge of the risk
factors for falls was associated with an increase in self-
reported fall prevention behaviors, with an odds ratio
(OR) of 1.5 (95% confidence interval [CI]

⫽1.1–2.2).

When post-CCFP behavior scores, which ranged from 1
(“almost never or never”) to 4 (“almost always or
always”), were used in the model as the outcome mea-
sure, similar results were demonstrated. Pre-CCFP
behaviors had an OR of 2.4 (95% CI

⫽1.6–3.8), indicat-

ing that those who were frequently using the fall reduc-
tion strategies before the CCFP program were 2.4 times
more likely to be using them after the program. Risk
factor knowledge had an OR of 1.4 (95% CI

⫽1.0–2.1),

Figure 2.

Self-reported frequency of use of fall prevention strategies with patients before and after Connecticut Collaboration for Fall Prevention (CCFP) program
(N

⫽94). For each answer, the striped bar corresponds to the pre-CCFP behavior; the dark bar corresponds to post-CCFP behavior.

400 . Brown et al

Physical Therapy . Volume 85 . Number 5 . May 2005

ўўўўўўўўўўўўўўўўўўўўўўўўўўўў

indicating that those who had greater knowledge of fall
risk factors were 1.4 times as likely to be using the
strategies. None of the other factors were associated with
post-CCFP practice behaviors in either model.

Discussion and Conclusion
The participants reported an increase in use of fall
reduction strategies with patients in response to the
multicomponent

behavioral

change

strategy.

As

expected, knowledge of and attitudes toward the fall risk
associated with gait or balance impairments and environ-
mental hazards were high. However, multiple medica-
tions, not under the usual purview of physical therapy
providers, also was recognized as a risk factor by 78% of
the participants, and 51% believed multiple medication
use to be a very important risk factor in their clinical
practice. Only 30% of the participants mentioned pos-
tural hypotension as a risk factor. After exposure to the
CCFP effort, however, 22 participants noted that they
routinely checked postural blood pressures in their
older patients.

More than two thirds of the physical therapy providers
reported increased frequency of use of fall reduction

strategies in their older patients. Results of the Bowker
test of symmetry provide evidence that the changes from
before exposure to the CCFP program to after exposure
to the CCFP program are not due to chance alone. More
importantly, the majority of the physical therapy provid-
ers had adopted strategies for reduction of fall risk
factors that they had not used in the past. These strate-
gies included an increase in the use of referrals to other
health care providers, increased use of exercises, and
increased education of patients about their fall risk
factors.

In multivariate models, only post-CCFP knowledge of the
risk factors for falls and fall-related practices prior to
exposure to the CCFP effort were significantly and
independently associated with an increase in self-
reported use of fall prevention strategies with patients.
Our findings are in agreement with those of previous
studies that suggest that knowledge alone does not result
in professional behavioral change.

18,28,29

Strengths of this study include the high level of partici-
pation by those physical therapy providers who had been
reached during this study. Only 15 (13%) of those

Figure 3

Distribution of change in self-reported use of fall prevention strategies with patients before and after Connecticut Collaboration for Fall Prevention
(CCFP) program (N

⫽94).

Physical Therapy . Volume 85 . Number 5 . May 2005

Brown et al . 401

physical therapy providers who were contacted refused
to participate either in the CCFP program or in this
study. This high participation rate lessens the likelihood
of selection bias. The telephone questionnaire itself is an
intervention that may further enhance the use of fall
reduction behaviors by physical therapy providers. Fre-
quently, after completion of the questionnaire, the par-
ticipants acknowledged they had not been assessing risk
factors and using management strategies as often as they
would have liked, and they verbalized a plan to increase
their use of the behaviors.

Several important caveats warrant comment. First, con-
cern may be raised that, by choosing to use retrospective
preintervention self-assessment, physical therapy provid-
ers may have overestimated their current fall reduction
behavior compared with previous behavior. However, it
has been theorized that there could be a change, due to
the intervention itself, in the standards used to judge the
preintervention and postintervention self-assessments.

29

For example, physical therapy providers may believe that
they are well versed in fall reduction strategies and
would rate themselves highly on a pretest. Yet, once they
receive an educational intervention, they may realize
they know less than they previously believed they knew.
In a posttest, they may rate themselves on a similar level
as they did before the intervention, but now their
assessment is based on what they learned. If, at the same
time, they did a retrospective pretest (ie, what did you
know before?), they have a better sense of what they
knew, compared with what they know now. This may be
particularly important when evaluating an educational
intervention, because participants may have an
increased understanding of or insight into the subject
they are rating. The change in the standards used to
rate the intervention can reduce the validity of the
self-ratings. Retrospective ratings have been well vali-

dated in this type of research, as they afford consistent
criteria for the preintervention and postintervention
self-assessments.

30,31

Similarly, the self-reported behav-

iors may have overestimated the actual behaviors. The
validity of self-report will need to be addressed to deter-
mine how well self-report reflects true behavior. The
addition of the open-ended questions about what behav-
iors had changed adds strength to the physical therapy
providers’ self-assessment. The open-ended questions
were used to encourage accountability for their answers
because participants were expected to justify the answers
they gave. In this study, the physical therapy providers
noted an increase in their use of fall prevention strate-
gies with patients and were able to give numerous
examples of the strategies they were consistently using in
their clinical practice.

In addition, the participants may have wanted to please
the interviewers. We attempted to minimize this by
having the interview occur by telephone, which theoret-
ically lessens the pressure of answering in a socially
desirable manner, as the interviewee never meets the
interviewer.

32

Of greatest importance in predicting who would adopt
the fall prevention strategies was the knowledge of the
risk factors for falls and, not surprisingly, pre-CCFP
practice behaviors. The education and materials pre-
sented during the outreach visit were tailored for imme-
diate use by the physical therapy providers, which may
have facilitated successful adoption of fall assessment
and management. Materials included the training man-
uals with instructions for implementing components of
the fall risk assessment and management, patient and
physical therapy provider risk factor checklists, and a
variety of patient-centered handouts. The outreach
effort used a “hands-on” approach that addressed barri-

Table 3.

Changes in Fall Reduction Practices Reported by Physical Therapy Providers (N

⫽94)

a

Practice Changes
Noted by
Physical Therapy
Providers

b

Fall Risk Factors

Gait or
Balance
Impairments

Multiple
Medications

Postural
Hypotension

Sensory and
Perceptual
Deficits

Foot and
Footwear
Problems

Environmental
Hazards

No change noted

41

17

50

58

50

40

Increased awareness

19

16

5

4

8

5

Increased assessment

18

5

22

16

21

18

Increased education/

recommendations/
handouts

26

32

17

13

29

43

Increased exercises/

training

35

Not applicable

4

15

3

Not applicable

Increased referral

1

41

10

4

10

1

List and review

medications

31

a

For each risk factor, question posed: “Now what do you do differently, if anything, to address this risk factor?”

b

Physical therapy providers could give multiple answers to the question.

402 . Brown et al

Physical Therapy . Volume 85 . Number 5 . May 2005

ўўўўўўўўўўўўўўўўўўўўўўўўўўўў

ers to change, allowed for problem solving, and permit-
ted ease of integration of the fall prevention strategies
into the physical therapy provider’s usual routine. The
results of this study suggest that the use of a multicom-
ponent change strategy can be successful in promoting
behavioral change, even in an area as complex
and multifactorial as fall risk factor assessment and
management.

References

1

Tinetti ME, Baker DI, McAvay G, et al. A multifactorial intervention

to reduce the risk of falling among elderly people living in the
community. N Engl J Med. 1994;331:821– 827.

2

Nevitt MC, Cummings SR, Kidd S, Black D. Risk factors for recurrent

nonsyncopal falls: a prospective study. JAMA. 1989;261:2663–2668.

3

Tinetti ME, Williams CS. Falls, injuries due to falls, and the risk of

admission to a nursing home. N Engl J Med. 1997;337:1279 –1284.

4

Tinetti ME, Speechley M, Ginter SF. Risk factors for falls among

elderly persons living in the community. N Engl J Med. 1988;319:
1701–1707.

5

Close J, Ellis M, Hooper R, et al. Prevention of falls in the elderly

(PROFET): a randomized control trial. Lancet. 1999;353:93–97.

6

Tinetti ME, Williams TF, Mayewski R. Fall risk index for elderly

patients based on number of chronic disabilities. Am J Med. 1986;80:
429 – 434.

7

Hornbrook MC, Stevens VJ, Wingfield DJ, et al. Preventing falls

among community-dwelling older persons: results from a randomized
trial. Gerontologist. 1994;34:16 –23.

8

Wolf SL, Barnhart HX, Kutner NG, et al. Reducing frailty and falls in

older persons: an investigation of Tai Chi and computerized balance
training. J Am Geriatr Soc. 1996;44:489 – 497.

9

Campbell AJ, Robertson MC, Gardner MM, et al. Psychotropic medi-

ation withdrawal and a home-based exercise program to prevent falls:
a randomized controlled trial. J Am Geriatr Soc. 1999;47:850 – 853.

10

Cumming RG, Thomas M, Szonyi G, et al. Home visits by an

occupational therapist for assessment and modification of environ-
mental hazards: a randomized trial of falls prevention. J Am Geriatr Soc.
1999;47:1397–1402.

11

Province MA, Hadley EC, Hornbrook MC, et al. The effect of

exercise on falls in elderly patients: a preplanned meta-analysis of the
FICSIT trials. JAMA. 1995;273:1341–1347.

12

American Geriatrics Society, British Geriatrics Society, and Ameri-

can Academy of Orthopaedic Surgeons Panel on Falls Prevention.
Guidelines for the prevention of falls in older persons. J Am Geriatr Soc.
2001;49:667– 672.

13

Connecticut Census Data. Available at: www.factfinder.census.gov/

home/saff/main.html. Accessed February 3, 2005.

14

Berwick DM. Disseminating innovations in health care. JAMA.

2003;289:1969 –1975.

15

Berwick DM. A primer on leading the improvement of systems. BMJ.

1996;312:619 – 622.

16

Jamtvedt G, Young JM, Kristoffersen DT, et al. Audit and Feedback:

Effects on Professional Practice and Health Care Outcomes. Cochrane Data-
base of Systematic Reviews. Chichester, United Kingdom: John Wiley &
Sons Ltd; 2004:1.

17

Thomson O’Brien MA, Oxman AD, et al. Educational Outreach Visits:

Effects on Professional Practice and Health Care Outcomes. Cochrane Data-
base of Systematic Reviews. Chichester, United Kingdom: John Wiley &
Sons Ltd; 2004:3.

18

Davis DA, Thomson MA, Oxman AD, Haynes RB. Changing physi-

cian performance. a systematic review of the effect of continuing
medical education strategies. JAMA. 1995;274:700 –705.

19

Soumerai SB, Avorn J. Principles of educational outreach (“aca-

demic detailing”) to improve clinical decision making. JAMA. 1990;
263:549 –556.

20

Rosenberg RN. Translating biomedical research to the bedside: a

national crisis and a call to action. JAMA. 2003;289:1305–1306.

21

Prochaska JO, Velicer WF, Rossi JS, et al. Stages of changes and

decisional balance for twelve problem behaviors. Health Psychol. 1994;
13:39 – 46.

22

Sherrington C, Menz HB. An evaluation of footwear worn at the

time of fall-related hip fracture. Age Aging. 2003;32:310 –314.

23

Menz HB, Lord SR. The contribution of foot problems to mobility

impairment and falls in community-dwelling older people. J Am Geriatr
Soc. 2001;49:1651–1656.

24

Koch M, Gottschalk M, Baker DI, et al. An impairment and disabil-

ity assessment and treatment protocol for community-living elderly
persons. Phys Ther. 1994;74:286 –298.

25

Feinstein AR. Principles of Medical Statistics. Washington, DC: Chap-

man & Hall/CRC; 2002:417.

26

Jekel JF, Katz DL, Elmore JG. Epidemiology, Biostatistics, and Preventive

Medicine. 2nd ed. Philadelphia, Pa: WB Saunders Co; 2001:177.

27

Bowker AH. A test for symmetry in contingency tables. J Am Stat

Assoc. 1948;43:572–74.

28

Headrick LA, Speroff T, Pelcanos HI, Cebul RD. Efforts to improve

compliance with the National Cholesterol Education Program guide-
lines: results of a randomized controlled trial. Arch Intern Med. 1992;
152:2490 –2496.

29

Browner WS, Baron RB, Solkowitz S, et al. Physician management of

hypercholesterolemia: a randomized trial of continuing medial educa-
tion. West J Med. 1994;161:572–578.

30

Skeff KM, Stratos GA, Bergen MR. Evaluation of a medical faculty

development program. Eval Health Prof. 1992;15:350 –366.

31

Levinson W, Gordon G, Skeff KM. Retrospective versus actual

pre-course self-assessments. Eval Health Prof. 1990;13:445– 452.

32

Streiner DL, Norman GR. Health Measurement Scales: A Practical Guide

to Their Development and Use. New York, NY: Oxford University Press;
1995:72–75.

Physical Therapy . Volume 85 . Number 5 . May 2005

Brown et al . 403