Clinical trials have identified interventions that reduce adverse outcomes such as falls in
nursing home (NH) residents but attempts to translate these into practice quality improvement
(QI) techniques have not been successful. Using a complexity science framework, our previous
RO1 showed that low connection, information flow, and cognitive diversity among NH staff
explains quality of care for complex problems such as falls. Our pilot of "Connect," a
multi-component intervention that encourages staff to engage in network-building and use
simple strategies to make new connections with others, enhance information flow, and use
cognitive diversity, suggests that staff can improve the density and quality of their
interactions. This 5-year study uses a prospective, cluster-randomized, outcome assessment
blinded design, with NHs (n=16) randomized to either Connect and a falls QI program (Connect
+ Falls) or QI alone (Falls). About 800 residents and 576 staff will participate. Specific
aims are to, in nursing homes: 1) Compare the impact of the Connect intervention plus a falls
reduction QI intervention (Connect+Falls) to a falls reduction QI intervention (Falls) on
fall risk reduction indicators (orthostatic blood pressure, sensory impairment, footwear
appropriateness, gait; assistive device; toileting needs, environment, and psychotropic
medication); 2) Compare the impact of Connect+Falls to Falls alone on fall rates and
injurious falls, and determine whether these are mediated by the change in fall risk
reduction indicators; 3) Compare the impact of Connect+Falls to Falls alone on complexity
science measures (communication, participation in decision making, local interactions, safety
climate, staff perceptions of quality) and determine whether these mediate the impact on fall
risk reduction indicators and fall rates and injurious falls. Cross-sectional observations of
complexity science measures are taken at baseline, at 3 months, at 6 months, and at 9 months.
Resident fall risk reduction indicators, fall rates, and injurious falls are measured for the
6 months prior to the first intervention and the 6 months after the final intervention is
completed. Analysis will use a 3-level mixed model to account for the complex nesting of
patients and staff within nursing homes, and to control for covariates associated with fall
risk, including baseline facility fall rates and staff turnover rates.
Although clinical trials have identified interventions that reduce adverse outcomes such as
falls in nursing home (NH) residents, attempts to translate those interventions into practice
using current standard of care quality improvement (QI) programs[1, 2] have not led to
expected improvements.[3, 4] Barriers encountered in previous studies point directly to a
need for effective nursing management practices (NMPs).[1, 3, 5] Many studies now show that
relationship-oriented NMPs such as open communication, participation in decision-making,
teamwork, and leadership result in better resident outcomes.[5-10] Our recent case-studies
described how NMPs work in day-to-day practice, and identified new NMPs associated with
better NH care. We found that staff at all levels engaged in these NMPs, albeit erratically,
suggesting that NHs have substantial untapped capacity to provide better resident
care.[11-15] Thus a new intervention that fosters systematic use of NMPs may provide a
foundation for more effective QI programs.
QI programs are the current standard for improving resident outcomes for common and costly
conditions such as falls, pressure ulcers, pain, and depression. Such geriatric syndromes are
inherently multifactorial, requiring modification of multiple risk factors to improve
outcomes.[16, 17] Clinical trials using study staff to implement multiple risk factor
reduction have improved resident outcomes,[18-20] but QI programs teaching existing NH staff
to implement multiple risk factor reduction have not shown significant effects.[1, 21-24] One
proposed reason for this failure is that QI programs seek to change individual clinician
behavior but fail to account for the interactive dynamics of care. We propose that CONNECT,
an intervention to foster systematic use of NMPs, will enhance the effectiveness of a Falls
QI program in NHs by strengthening the one-on-one staff interactions that are necessary for
clinical problem-solving about geriatric syndromes.
We have developed the Connect intervention based on complexity science and empirical
research to target these local interactions among staff in a new approach to facilitating
organizational learning. Connect is a multicomponent intervention that includes: 1) helping
staff learn new strategies to improve the effectiveness of day-to-day interactions; 2)
helping staff identify important relationships and encouraging interaction at the point of
care; and 3) mentoring to reinforce and sustain newly acquired interaction behaviors.
Complexity science and empirical research suggest that interaction patterns determine
information flow, ease of knowledge transfer, and capacity to monitor behaviors and outcomes
in health care settings. [10, 26-28] Thus, Connect has the potential to improve resident
outcomes when combined with QI programs for clinical problems such as falls. Falls is an
excellent outcome for this initial test of Connect because: 1) there is ample evidence that
multifactorial risk factor reduction interventions effectively reduce fall rates in NHs; 2)
accepted practice guidelines and fall prevention programs exist;[29-32] and 3) falls is an
important clinical problem in its own right.
The specific aims of this longitudinal, two arm, randomized intervention study are:
Aim 1: Compare the impact of the Connect intervention plus a falls reduction QI intervention
(Connect+Falls) to the falls reduction QI intervention alone (Falls) on fall-related process
measures in nursing home residents.
Aim 2 (exploratory): Compare the impact of Connect+Falls to Falls alone on fall-related
outcome measures in nursing home residents, and determine whether these are mediated by the
change in fall-related process measures.
Aim 3 (exploratory): Compare the impact of Connect+Falls to Falls alone on staff interaction
measures as reported by NH staff, and determine whether these mediate the impact on
fall-related process measures and fall-related outcome measures.
With its focus on improving local interaction, Connect is an innovative new approach
targeting the learning environment to maximize NH staff's ability to adopt content learned in
a Falls QI program and integrate it into knowledge and action. Our pilot work shows Connect
to be feasible, acceptable and appropriate. Connect results from empirical findings; local
interaction behaviors already exist in NHs, albeit to a limited extent and not in a way that
effectively enables the staff to adopt evidence-based current practice for falls prevention
inherent in the Falls approach. We are confident that in most NHs the capacity exists to
develop and focus these behaviors using existing staff and resources and, therefore, the
Connect intervention has the potential to enhance the effectiveness of Falls by promoting its
adoption. Also, being a system intervention, Connect can be applied in future projects to
examine the adoption of evidence-based practices for a wide variety of clinical problems such
as pressure ulcers, pain, and depression. This study offers a unique opportunity to
understand the circumstances in which such proven interventions (e.g., Falls) are likely to
be translated into practice. Our future work will build on this study to establish correlates
of the sustainability of the intervention in NHs and examine transferability to other
clinical problems and other health care settings. The results of this research will be of
interest to NH leadership and policy makers, particularly in light of ongoing state and
national initiatives to improve care in NHs.
- Eligible residents will be long-term care residents at least 65 years of age who have
resided in the NH at least 6 months and are likely to survive at least 6 months.
Residents must be potentially at risk for falls, which we define as ambulatory or
transfer-independent as recorded on the Minimum Data Set.