Title: Working Environment and Adverse Patient Events in Home Healthcare: Issues and Findings
1Working Environment and Adverse Patient Events in
Home Healthcare Issues and Findings
- Penny H. Feldman, Ph.D. - Principal
InvestigatorRobert J. Rosati, Ph.D. -
Investigator - Supported by AHRQ Grant R01 HS11962
- Visiting Nurse Service of New York
2Collaborators
John Bridges, Ph.D. - Investigator Christopher
Murtaugh, Ph.D. - Investigator Timothy Peng,
Ph.D. - Research Associate Doreen Wang, M.A. -
Programmer Lori King, B.A. - Research
Assistant Visiting Nurse Service of New York
3Study Aims
- To examine the variability of adverse events and
organizational team characteristics across 86
care teams within one agency - To examine the relationships between adverse
patient events and - Patient characteristics
- Care team objective characteristics
- Care team subjective work environment
4Conceptual Model
- Organizational/Structural Characteristics
- Communication Processes
- Governance
- Information Technology
- Leadership
- Values
- Style
- Strategy
Core Structural Domains
Process Domains
- Quality Emphasis
- Patient Centeredness
- Safety
- Innovation
- Outcome Measurement
- Evidence-based Practice
- Supervision
- Style
- Employee Recognition
- Work Design
- Staffing
- Resources/Training
- Rewards
- Autonomy
- Group Behavior
- Collaboration
- Consensus
- Healthcare Worker Outcomes
- Satisfaction
- Burnout
- Intention to Leave
- Organizational Identification
- Patient Outcomes
- Patient Satisfaction
- Patient Safety
- Health and Function
Outcomes
5Current Study
Measures of care team objective
characteristics and subjective work environment
- Quality Emphasis
- Safety
- Patient Care
- Supervision
- Style
- Employee Recognition
- Work Design
- Staffing
- Resources/Training
- Rewards
- Autonomy
- Group Behavior
- Collaboration
- Consensus
- Patient Outcomes
- Patient Safety
Adverse Events
6Study Context
- Largest non-profit home health care organization
in the United States - Average daily census is 25,000 patients
- Diverse and complex patient population with
post-acute and long-term care needs - Disproportionately high numbers of
African-American and Hispanic patients - Decentralized workforce providing care in
patients homes - Patients are seen by over 100 different care
teams (all programs) across seven counties
7Data Sources
- Outcomes Assessment Information Set (OASIS)
- Standardized patient assessment data
- Patients assessed at the beginning and end of
each episode of care - Employee survey
- Patient level service utilization data
- Human Resources employee data
- Other administrative data
8Employee Survey
- Mailed self-administered survey to acute and
congregate care clinicians and front-line
managers - 46 response rate (1,027 surveys)
- 169 items compiled from work environment
literature - Approximately 30 minutes to complete
- The survey measured 15 constructs
- Teamwork (0.83)
- Leadership (0.90)
- Top-Down Communication (0.84)
- Perceived Quality (0.76)
- Safety Culture
- General (0.81)
- Leadership (0.84)
- Communication (0.91)
- Manageable Workload (0.74)
- Job Demands (0.86)
- Autonomy (0.88)
- Equity (0.85)
- Access to Resources (0.68)
- Satisfaction (0.85)
- Staff/Patient Communication (0.72)
- Physician Relations (0.94)
Shown with team-level reliability coefficients
(Chronbachs alpha)
9Study Design
- We used a 2-stage regression model to analyze the
relationship between the dependent variablethe
risk-adjusted incidence of any CMS adverse
eventand a set of independent objective and
subjective variables (e.g., aspects of the team
structure, staff beliefs about their team, etc.) - The final results are based on how the dependent
variable was related to each of the independent
variables while controlling for (or taking into
account) all the other variables included in the
analysis
10Study Design (continued)
Patient Episode Data
Employee Survey Data
11CMS Adverse Events
- The 13 CMS Home Healthcare defined adverse events
serve as markers for potential problems in home
care because of their negative nature and
relatively low frequency - Whether or not individual patient situations
resulted from inadequate care provision can only
be determined by investigating the actual care
provided to specific patients - An episode was considered to have an adverse
event if one or more of the adverse events
occurred - For study purposes, we counted events at
discharge/transfer and at the end of each 60-day
episode
1213 CMS Adverse Events
- Emergent Care (EC) Events
- EC for injury caused by fall or accident at home
- EC for wound infection or deteriorating wound
status - EC for improper medication administration or
medication side effects - EC for hypo/hyperglycemia
Significant Decline in Health or
Function Development of urinary tract
infection Increase in number of pressure
ulcers Substantial decline in 3 or more
ADLs Substantial decline in management of
medications
Significant Unmet Care Need Unmet need for wound
care or medication assistance Unmet need for
toileting assistance Unmet need for management of
behavioral problem
Serious Unexpected Event Unexpected nursing home
admission Unexpected death
13Characteristics of Patient Episodes
- The data set contained 51,560 episodes of care (6
months) - The populations average age was 71 years
- The majority of the group was female (67)
- 40 lived alone 30 had no primary caregiver
- The most common clinical diagnoses were diabetes
(14), hypertension (7), and congestive heart
failure (5) - The ethnicity of the population was
14Characteristics of Survey Respondents
- The data set contained 951 survey respondents
- Their average age was 45 years
- They had an average of 9.3 years of experience in
the home care setting - Their average tenure at the agency was 7.5 years
- 81 had a bachelors (or higher) degree
15Adverse Events Across Teams
Variation in rates of adverse events across care
teams
of teams
Episodes with any one CMS event over a 6-month
time period (unadjusted mean approx. 7)
16Results Volume and Distribution of Work
- ? Volume R2.010
- ? Concentration of visits among nurses R2.021
- ? Number of weekend admissions R2.126
- lower rates of team-attributable adverse events
The degree to which the teams visits are evenly
distributed across many nurses or concentrated
among a few nurses is measured using a
Herfindahl-Hirschman Index of team
visits Risk-adjusted based on a methodology
developed by the health economist on our team
Note The partial R-squared is displayed next to
each variable
Overall Model R-squared .345
17ResultsWorkforce Characteristics
- lower rates of team-attributable adverse events
Note The partial R-squared is displayed next to
each variable
Overall Model R-squared .345
18ResultsPerceived Organizational Climate
- ? Equity R2. 011
- ? Teamwork R2.019
- ? Supervisor Safety Culture R2. 022
- ? ManageableWorkload R2.046
- lower rates of team-attributable adverse events
Overall Model R-squared .345 including all
eight variables
May be due in part to association of increased
awareness of safety with increased reporting
(identified in health care literature)
Possibly due to complacency or burnout
factors
Note The partial R-squared is displayed next to
each variable
Overall Model R-squared .345
19Summary of Results
- After controlling for patient case-mix severity,
this study found that work design, safety
culture, teamwork, and workforce characteristics
all affected team-attributable adverse event
rates. - Specifically, adverse event rates were lower for
teams that had
- higher volume
- higher concentration of visits
- fewer weekend admissions
- more experienced clinicians
- greater perceived equity
- greater perceived teamwork
- lower perceived supervisorsafety culture
- less workload manageability
Possibly due to a combination of safety
awareness, complacency, and burnout factors
20Summary of Results
Overall Model R-squared .345
21Implications for Patient Safety
- Findings are consistent with other patient safety
studies that - Link higher volume and greater experience to
safer care - Show higher rates of reported adverse events
where safety culture is emphasized - Show that managerial span of control and teamwork
make a difference - Association between better patient safety
outcomes and less manageable workload - Raises questions about possible burnout among
high performing teams or complacency among
poorer performing teams - Suggests the need for further investigation
22Next Steps
- Complete chart reviews to analyze the
relationship between adverse events and specific
types of patient care errors for two adverse
events - Emergent care for wound infection or
deteriorating wound status - Emergent care for hypo/hyperglycemia
- Prepare study results for publication
- Pursue additional research
- Assess the impact of quality/safety on team
climate (reverse causation) - Design and evaluate interventions to improve
teamwork and reverse adverse events - Develop a diagnostic tool to predict the
likelihood of future adverse events
23Contact Information
Robert J. Rosati, Ph.D. Center For Home Care
Policy and Research Visiting Nurse Service of New
York 5 Penn Plaza, 11th Floor New York, NY
10001 (212) 609-5776 Robert.Rosati_at_vnsny.org