Transcript Insert Session Title - AMDA Foundation

Certified Nursing Assistants’ Perceptions of
Nursing Home Patient Safety Culture:
Is There a Relationship to Clinical or Workforce
Outcomes?
Alice F. Bonner, PhD, RN
Graduate School of Nursing
University of Massachusetts, Worcester
March 7, 2008
Faculty Disclosures:
Dr. Bonner has disclosed that she has no relevant
financial relationships.
Learning Objectives
By the end of the session, participants will
be able to:
• Discuss the importance of patient safety culture
in long term care
• List at least two clinical outcomes and one
workforce outcome relevant to the study of
patient safety culture
• Consider future research opportunities related to
the study of patient safety culture in long term
care
Acknowledgement
This study was supported by an American
Medical Director’s Association
Foundation/Pfizer Quality Improvement
Award
Consultants and Coauthors
Nicholas Castle, PhD
Associate Professor, Graduate School of Public
Health
Aiju Men, MS
Analyst
Steven Handler, MD, MS
Assistant Professor, School of Medicine
University of Pittsburgh
Pittsburgh, Pennsylvania
Background and Significance
• The significance of medical error
– 2000 Institute of Medicine (IOM) report, To Err is
Human
• Definition of patient safety culture
• Development of instruments to measure PSC
– Several hospital studies
– At least six nursing home studies to date
Original 12 Domains
Hospital Survey of Patient Safety Culture
(HSOPSC)
•
•
•
•
•
•
•
•
•
•
•
•
Overall Perceptions
Frequency of events reported
Management expectations and actions
Organizational learning
Teamwork within units
Communication openness
Feedback and communication about errors
Non-punitive response to error
Staffing
Management support for resident safety (attitudes)
Teamwork across units
Handoffs and transitions
Selecting the Topic:
CNAs’ Perceptions of Nursing Home
Patient Safety Culture
• Certified nursing assistants (CNAs): the
heart of the interdisciplinary team (IDT)
– CNAs provide 80-90% of the direct care in
nursing homes
– CNAs are on the frontlines and are often the
first line of defense against accidents or injury
The Research Question
• Is there a relationship between CNAs’ values,
perceptions and attitudes about patient safety and
clinical outcomes (rates of falls, daily restraint use, and
pressure ulcers)?
• Is there a relationship between CNAs’ values,
perceptions and attitudes about patient safety and a
workforce outcome (nursing staff turnover)?
Significance: to date, no study has been published
that examines the relationship between patient
safety culture (PSC) scores and actual clinical or
workforce outcomes in nursing homes
Core Organizational
Structure
Conceptual Framework
CONCEPTUAL FRAMEWORK
OF CNA PATIENT SAFETY
CULTURE IN NURSING
HOMES
Leadership
--valuing safety
--philosophy
--in touch with
frontline workers
Organizational Structure
--fair and just culture
--policies and procedures
--learning culture
--safety systems
Facility Level Outcomes
Unit Level Process Domains
Communication
Management
Attitudes
Staffing
Handoffs
Teamwork
across units
Circle of
Safety
Climate*
Management
Expectations
Overall
Perceptions
Teamwork
Feedback
Organizational
Learning
PSC: Safety Attitudes*,
Perceptions*, Behaviors
Workforce Outcomes
--job satisfaction
--turnover*
Clinical Outcomes
--accidents/injuries
--physical restraints*
--pressure ulcers*
--falls*
*Denotes concept or relationship examined in this study.
Adapted from Stone, P. et al. (2005). Organizational Climate of Staff Working Conditions and
Safety - An Integrative Model.
Specific Aim 1
Hypothesis
Nursing homes with higher CNA PSC total
scores and domain subscores will have lower
rates of falls, daily restraint use, and
pressure ulcers, and lower staff turnover
(RN/LPN/CNA)
Specific Aim 2
Hypothesis
Nursing homes with higher CNA total PSC
scores and domain subscores will have
higher staffing levels (RN/LPN/CNA) and
lower turnover (RN/LPN/CNA)
Specific Aim 3
Hypothesis
CNAs with more total years of education (in
addition to CNA training), more years of
experience and longer tenure in the
nursing home will have higher average
PSC scores than less educated, less
experienced CNAs
Specific Aim 4
Hypothesis
Nursing homes with higher CNA PSC scores
will be located in non-rural counties, have
higher bed occupancy, have lower bed
size, have higher private pay occupancy
and will have either not-for-profit status
or will be members of a chain
Parent Study: Methods
• Castle, N.G. (2006). Nurse Aides’ ratings of the resident
safety culture in nursing homes. International Journal for
Quality in Health Care, 18(5), 370-376.
– Sample
• 5 randomly selected states
• 10% random sample (240 homes)
• 72 nursing homes (30% response rate)
• 1579 CNAs (55% response rate)
– Procedures
– Human subjects
– Measures
– Data collection
Dissertation Study: Methods
• Secondary data analysis
– Hospital Survey on Patient Safety Culture (HSOPSC)
– Matched with data from the Minimum Data Set
(MDS), Online Survey Certification and Reporting
(OSCAR) System, and Area Resource File (ARF)
• Procedures
– Approval obtained from both the University of
Massachusetts and University of Pittsburgh IRB
(exempt status)
– Power analysis
• Data analysis
– Exploratory factor analysis
– Poisson, linear and multinomial logistic regression
– Generalized Estimating Equations (GEE)
Selected Outcome Variables
•
•
•
•
Falls
Use of physical restraints
Pressure ulcers
Nursing staff turnover
– Based on work by Teigland, Capezuti, Rubenstein, Berlowitz,
Schnelle, Sullivan-Marx, Strumpf, Castle, Engberg, Mor, Morris,
Harrington, Rantz, Scott-Cawiezell and others
Risk Adjustment
Based on available data
• Falls
– cognitive impairment, Alzheimer’s disease, behaviors, ADL
status, average number of medications, facility characteristics
• Daily use of physical restraints
– cognitive impairment, behaviors, number of medications, ADL
status, Alzheimer’s disease, facility characteristics
• Pressure ulcers
– cognitive impairment, Alzheimer’s disease, behaviors, ADL
status, average number of medications, facility characteristics.
Sample facilities had very small numbers of low risk residents,
therefore categories were combined
• Turnover
– facility characteristics such as staffing, profit or chain, bed size,
county unemployment rate, facility occupancy
RESULTS
Demographic Data
for CNAs
•
•
•
•
•
•
82.7% were Caucasian
91.9% of CNAs had a high school degree
98.1% were female
Average age was 30 years
Average tenure in the facility was 4.6 years
Average tenure as a CNA was 7.8 years
Demographic Data
for Nursing Homes
•
•
•
•
•
•
•
82.4% were non-rural
58% were non-profit
37.8% were chain members
Average facility bed size was 102.5
Average facility fall rate was 12.3%
Average facility pressure ulcer rate was 8.2%
Average facility restraint rate was 6.7%
Demographic Data
for Nursing Homes
•
•
•
•
•
•
Average CNA turnover rate was 33.6%
Average LPN turnover rate was 28%
Average RN turnover rate was 24%
Average CNA staffing 29 FTE/100 residents
Average LPN staffing 9 FTE/100 residents
Average RN staffing 11 FTE/100 residents
(includes administrative RN staff)
CNA PSC and Falls Rates
Poisson Regression (N=74 facilities)
• A higher average CNA total PSC score
was associated with a higher rate of falls
(B=.015; p=.000). In addition, a higher rate
of falls was associated with:
– fewer beds (B=-.001; p=.028)
– higher cognitive performance scale (CPS) scores (more
cognitively impaired residents) (B=.182; p=.003)
– lower activities of daily living (ADL) scores (less functionally
dependent residents) (B=-.182; p=.006)
– higher rate of Alzheimer’s disease in the facility (B=.011; p=.017)
– lower proportion of Medicare residents in the facility (B=-.013;
p=.000).
CNA PSC and Restraint Rates
Multinomial Logistic Regression (N=74)
• Facilities with higher average CNA total PSC
scores were more likely to report moderate
restraint use, whereas facilities with lower
average CNA total PSC scores were more likely
to report high restraint use (B=.172; p=.017).
In addition:
– Facilities reporting moderate restraints had more medications
per resident than those reporting high restraints (B=.895; p=.023)
– Facilities reporting moderate restraints had slightly lower ADL
scores (more functionally independent) than facilities reporting
high restraints (B=-.003; p=.028)
CNA PSC and Pressure Ulcer Rates
Poisson Regression (N=74 facilities)
• Average CNA total PSC scores did not have a
statistically significant association with pressure
ulcer rates (B=-0.001; p=0.807)
CNA PSC and CNA Turnover
Linear Regression (N=74)
• Higher CNA PSC scores were associated with
lower CNA turnover (B=-.052; p=.030). In
addition:
• Lower CNA turnover was associated with not for profit
status (B=1.446; p=.001)
• Lower CNA turnover was associated with higher facility
occupancy (B=-.188; p=.000)
• Adjusted R square for this model was .639.
GEE Model for Staffing, Turnover and Total PSC (N=1761)
Variable
B
SE
P-Value
CNA staffing
-0.555
0.353
0.116
LPN staffing
-0.938
0.295
0.002**
RN staffing
0.540
0.447
0.227
CNA turnover
-2.390
1.094
0.029**
LPN turnover
-0.173
1.386
0.901
RN turnover
0.842
1.373
0.540
Dependent Variable: Total CNA Patient Safety Culture Score
**p<.05
GEE Model for CNA Demographic Characteristics (N=1761)
Variable
B
SE
P-Value
4.868
3.254
0.135
Gender
Female
Male
0a
Education
High School Degree
2.040
1.763
0.247
Associates Degree
-3.167
3.062
0.301
0.039
0.025**
Bachelors Degree or Higher
Age
Dependent Variable: Total CNA Patient Safety Culture Score
a. Set to zero because this parameter is redundant.
**p<.05
0a
-0.087
Combined GEE Model for Facility level and CNA Demographic
Characteristics (N=1761)
Variable
Not for profit
Profit
Non-chain member
Chain member
Non-rural location
Rural location
High School Degree
Associate Degree
Bachelors Degree or higher
CNA turnover
Age
Tenure as a CNA
Tenure in the facility
County unemployment rate
Number of nursing homes in the county
Average facility occupancy
Average facility private pay occupancy
Bed size
LPN staffing
B
SE
P-Value
-1.771
0a
-1.524
0a
-1.537
0a
1.963
-3.576
0a
-1.337
-0.049
-0.058
0.009
1.297
-0.052
-0.010
-0.006
-0.020
-0.801
1.755
0.313
1.645
0.354
2.684
0.567
1.769
2.950
0.267
0.225
0.515
0.068
0.100
0.135
0.838
0.091
0.217
0.162
0.014
0.211
0.009**
0.469
0.559
0.946
0.122
0.571
0.964
0.970
0.134
0.000**
Dependent Variable: Total CNA Patient Safety Culture Score
a Set to zero because this parameter is redundant.
**p<.01
DISCUSSION
CNA PSC and Falls Rates
Poisson Regression
• A higher average CNA total PSC score was
associated with a higher rate of falls
(B=.015; p=.000). Why?
– Ascertainment bias in MDS measure
• Comprehensive fall prevention programs include
more effective reporting and documentation
strategies
– Resident selection (higher percentage of
dementia, more functional independence)
– Unmeasured confounding factors
CNA PSC and Restraint Rates
Multinomial Logistic Regression
• Facilities with higher average CNA total PSC
scores were more likely to report moderate
restraint use, whereas facilities with lower
average CNA total PSC scores were more likely
to report high restraint use (B=.172; p=.017)
– Facilities with a less developed PSC are more likely to
use restraints
– In homes with zero or very low restraints, CNAs may
perceive this as a lack of safety for residents
CNA PSC and Pressure Ulcer Rates
Poisson Regression
• Average CNA total PSC scores did not have a
statistically significant association with pressure
ulcer rates (B=-0.001; p=0.807). Why not?
– Are pressure ulcers a quality or a safety
measure?
– Do some staff consider pressure ulcers
inevitable?
– Failure to detect differences (study only
powered to detect moderate to large effect)
GEE Model for Staffing, Turnover and Total PSC (N=1761)
Variable
B
SE
P-Value
CNA staffing
-0.555
0.353
0.116
LPN staffing
-0.938
0.295
0.002**
RN staffing
0.540
0.447
0.227
CNA turnover
-2.390
1.094
0.029**
LPN turnover
-0.173
1.386
0.901
RN turnover
0.842
1.373
0.540
Dependent Variable: Total CNA Patient Safety Culture Score
**p<.05
GEE Model for CNA Demographic Characteristics (N=1761)
Variable
B
SE
P-Value
4.868
3.254
0.135
Gender
Female
Male
0a
Education
High School Degree
2.040
1.763
0.247
Associates Degree
-3.167
3.062
0.301
0.039
0.025**
Bachelors Degree or Higher
Age
Dependent Variable: Total CNA Patient Safety Culture Score
a. Set to zero because this parameter is redundant.
**p<.05
0a
-0.087
Combined GEE Model for Facility level and CNA Demographic
Characteristics (N=1761)
Variable
Not for profit
Profit
Non-chain member
Chain member
Non-rural location
Rural location
High School Degree
Associate Degree
Bachelors Degree or higher
CNA turnover
Age
Tenure as a CNA
Tenure in the facility
County unemployment rate
Number of nursing homes in the county
Average facility occupancy
Average facility private pay occupancy
Bed size
LPN staffing
B
SE
P-Value
-1.771
0a
-1.524
0a
-1.537
0a
1.963
-3.576
0a
-1.337
-0.049
-0.058
0.009
1.297
-0.052
-0.010
-0.006
-0.020
-0.801
1.755
0.313
1.645
0.354
2.684
0.567
1.769
2.950
0.267
0.225
0.515
0.068
0.100
0.135
0.838
0.091
0.217
0.162
0.014
0.211
0.009**
0.469
0.559
0.946
0.122
0.571
0.964
0.970
0.134
0.000**
Dependent Variable: Total CNA Patient Safety Culture Score
a Set to zero because this parameter is redundant.
**p<.01
Discussion
• CNA PSC scores were associated with some (falls,
restraints), but not all (pressure ulcers) clinical outcomes
• Few PSC subscore associations were noted in the data,
possibly related to sample size. Further work on
subscales/domains is needed
• Associations were noted between CNA PSC scores, CNA
turnover and LPN staffing, suggesting the importance of
staff mix and nursing staff models
Discussion
• Factor Analysis
– The factor structure was similar, but some differences
were noted, suggesting that further work on adapting
the HSOPSC to nursing homes and CNA populations
may improve the validity of the instrument
Strengths and Limitations
• Strengths
– National, randomly selected sample
– HSOPSC has previously reported reliability, validity
– Focus on CNAs, critical to NH PSC
• Limitations
– Reliability and validity of MDS, OSCAR data
– Different clinical indicators and variables for risk
adjustment may need to be examined
– Sample size may have had limited ability to detect
some significant results
Policy and Practice
Implications
• Interventions that improve CNA PSC may influence
clinical outcomes, such as falls and restraint use
• Efforts to reduce CNA turnover should include a
focus on nursing management and staff models,
and enhancing PSC (Advancing Excellence goals
include reducing nursing staff turnover)
• The QIOs’ 9th SOW includes recommendations to
include the nursing home culture survey in
comprehensive resident safety programs
Implications and Future
Research
• Future studies should consider fall-related injuries
as well as falls rates, and other clinical indicators
• Future studies should build on our knowledge of
CNA PSC and include other members of the IDT
• Review of the literature in nursing homes in press
and targeted for April issue of Annals of Long
Term Care (co-authors Castle, Perera, Handler)
• Original Study submitted to JAMDA and currently
under review
Conclusions
• Measuring patient safety culture may be
helpful in working with your IDT on specific
areas such as communication, teamwork,
handoffs
• Interventions that improve CNA PSC may
influence clinical outcomes, such as falls and
restraint use
• Revised NH Culture survey should be
available on the AHRQ website after May
2008