to ﬂuctuation, is challenging; to evaluate the effect of an
intervention, the number of falls—together with the num-
ber of patients at risk or the fall rate—needs to be
reported for similar study periods.
One study summariz-
ing a systematic review on fall prevention published in
1998 indicated that the usefulness of published evaluations
is limited because of small sample sizes, the research
design used, and study quality.
The current study found
the even more basic problem that data were not described
sufﬁciently to enable effects to be evaluated.
In the few studies that reported data, the pooled inter-
vention effect estimate was not statistically different from
the preintervention status or standard care control group.
Results of meta-analyses summarizing the international lit-
erature vary and report, for example, a statistically signiﬁ-
cant effect for historic control studies but not for
or no consistent results across outcomes
(rate ratio vs number of fallers).
Patient falls are not a
novel problem in hospitals, so to understand the effect of
a new intervention, the comparator status (part of the
intervention context) needs to be known (which fall-reduc-
tion strategies were in place before the tested intervention
or in a concurrent control group). The comparator is an
important determinant of the success (the achieved change)
of the intervention. Unfortunately, fewer than half of the
included studies reported on existing, routine fall preven-
tion approaches present in the comparator group. Recent
publications have emphasized that, to comprehend study
effects, more information is needed on the context in
which interventions take place.
Similarly, details of the
implementation process have been singled out as a crucial
element in patient safety practice evaluations to advance
the science of patient safety,
but information on how a
fall prevention intervention was introduced into clinical
practice in the target organization, for example through
staff education or known continuous quality improvement
strategies, was seldom documented.
Individual study results varied, and there was evidence
of statistical heterogeneity between studies. It was hypoth-
esized that the implementation intensity, intervention com-
plexity, comparator information, and adherence to care
processes were effect modiﬁers for the effectiveness of
interventions to reduce falls, but the large majority of
included studies could not be statistically analyzed. Meta-
regressions showed some evidence of the importance of
adherence levels (data on whether the intervention took
place as intended and implemented care processes were
indeed adhered to) and the intensity of the intervention,
but effects were not consistent across available data.
Adherence strategies are of particular importance for long-
term changes. Initial success might not be maintained
because adherence to introduced care processes fades in
clinical practice, use of the introduced risk assessment tool
may not be sustained, and recommended measures may no
longer be systematically applied. Some barriers encoun-
tered in clinical practice included forgetting to remove
identiﬁcation signs next to call lights after high-risk
patients were discharged and failing to educate new staff
about fall prevention programs.
This systematic review relied on published information.
The amount of reported details, in particular regarding
implementation and adherence strategies, may depend on a
journal’s word limit and preferences. Contacting primary
authors may have provided answers to unresolved ques-
tions, but fall prevention interventions are only as good as
their implementation and adherence strategies, and sufﬁ-
cient data to communicate the nature of the comparator
and its intensity are crucial to understanding study effects.
The Standards for QUality Improvement Reporting Excel-
lence (SQUIRE) criteria provide detailed guidance for how
complex interventions to improve the quality of healthcare
delivery should be reported.
Low statistical power limited these quantitative analy-
ses. The absence of deﬁnitive ﬁndings should therefore not
be interpreted as evidence that implementation strategies,
intervention complexity, and level of adherence are unim-
portant. Until better data are available, readers may bene-
ﬁt from reviewing the successful studies documented in
this review and pursuing approaches that are most com-
patible with their hospital culture and patient populations.
Promising approaches exist, but better reporting of
outcomes and detailed information on intervention compo-
nents and comparison groups, as well as the implementa-
tion strategy and adherence to care processes, need to be
included in published fall prevention evaluations to estab-
lish a strong evidence base for successful interventions to
reduce patient falls in hospitals.
We thank Paul Shekelle (Veterans Affairs Greater Los
Angeles; Evidence-based Practice Center, RAND) and
Rhona L. Imcangco (Agency for Healthcare Research and
Quality (AHRQ)) for pertinent comments, and Tanja Perry
(RAND) and Aneesa Motala (RAND) for administrative
Conﬂict of Interest: The editor in chief has reviewed
the conﬂict of interest checklist provided by the authors
and has determined that the authors have no ﬁnancial or
any other kind of personal conﬂicts with this paper. This
project was funded under Contract HHSA290201000017I
TO #1 from the AHRQ. Additional support was provided
through the U.S. Department of Veterans Affairs, Veterans
Health Administration, Veterans Affairs Health Services
Research and Development (HSR&D) Service through the
VA Greater Los Angeles HSR&D Center of Excellence
(Project VA CD2 08–012–1 and a locally initiated
Author Contributions: Hempel S., Ganz D. A., Saliba
D., and Specter W. D.: designed the study. Shanman R.:
provided the literature searches. Hempel S., Wang Z.,
Ganz D. A., Shier V., and Newberry S.: extracted the data.
Johnsen B.: managed the data. Booth M.: performed the
statistical analyses. All authors contributed to interpreta-
tion of the data. Hempel S., Johnsen B., and Ganz D. A.:
drafted the manuscript. All authors provided critical
revisions to the ﬁnal manuscript.
Sponsor’s Role: The opinions expressed in this docu-
ment are those of the authors and do not reﬂect the ofﬁcial
position of AHRQ or the Department of Veterans Affairs.
AHRQ and the Department of Veterans Affairs had no
role in the design, methods, subject recruitment, data col-
lections, analysis and preparation of this paper; the
expressed views are those of the authors.
JAGS APRIL 2013–VOL. 61, NO. 4 FALL PREVENTION INTERVENTIONS IN HOSPITALS 493