Understanding the Influences on the Association between Nurse Staffing and Preventable Patient Complications

1
Understanding the Influences on the Association
between Nurse Staffing and Preventable Patient
Complications

• Deborah Dang, PhD, RN
• 2007 Interdisciplinary Research Interest Group
• on Nursing Issues
• Academy Health
• June 2, 2007

2
Problem and Significance

• Nurse staffing may be necessary, but not
  sufficient, to prevent adverse patient events
• Few studies have examined characteristics of the
  work environment at the unit level that may
  affect staffing
• Nurse staffing and characteristics of the work
  environment are modifiable features in hospitals

3
Purpose

• Examine the impact of nurse staffing and
• potential confounders on preventable
• adverse events at the unit level over a
• seven year period.

4
Research Questions

• Is unit-level nurse staffing associated with the
  failure to rescue, falls, and medication errors
  between 1998 and 2004?
• Do unit characteristics confound the relationship
  between nurse staffing and adverse events between
  1998 and 2004?

5
Staffing and Strength of Patient Outcomes
Evidence
Central line-associated BSIs
Decubitus ulcer
Falls
Failure to rescue
Medication errors
Mortality
Pneumonia, hospital acquired
Pneumonia, vent-associated
Post-op PE/DVT
Post-op respiratory failure
UTI, catheter-associated
consistent evidence mixed evidence
6
Conceptual Model
System
Outcome
Intervention
Client
Quality Health Outcomes Model (Mitchell,
Freketich, Jennings, 1998)
7
Study Model

• System
• Unit
• Nurse Staffing Characteristics
• Total hours ? Agency proportion
• RN hours ? Orientee proportion
• RN Proportion ? RN education
• RN shortfall ? RN experience
• ? RN turnover
• ? Patient turnover

• Outcomes
• Adverse Events
• Failure to rescue
• Falls
• Medication errors

8
Research Design

• Design
• Descriptive correlational using secondary data
• Setting
• 945 bed, Magnet-designated, not-for-profit, urban
  academic medical center
• Sample
• Convenience sample of 31 adult and pediatric
  inpatient units
• Unit of analysis
• Care-giving unit

9
Type of Patients and Level of Care

• Adult units 25 (84)
• Acute care 16 (52)
• IMC 8 (26)
• ICU 7 (23)
• gt 100 due to rounding

10
File Development

• Design and construct the database and merge
  procedures
• Restructure database
• Develop procedures for handling
• Changes in cost centers
• Inconsistent reporting periods
• Handling missing data
• Construct 3 separate files for analyses
• Failure to rescue (N 28)
• Falls (n560)
• Medication errors (n 341)

11
Analyses

• Hospital-level Failure to Rescue
• Descriptive
• Bivariate zero-order and first-order
  correlations
• Unit-level Falls and Medication errors
• Poisson regression with a robust variance
  estimator
• Adjust for type of unit
• Account for change in AE rates over time by
  including a quadratic function of time

12
Sample Benchmarks
Total HPPD Total HPPD Total HPPD RN Proportion () RN Proportion () RN Proportion ()
Acute IMC ICU Acute IMC ICU
Current study (31 units) 8.7 12.7 21.5 81.5 83.7 87.5
Dunton (1751 units) 7.6 9.1 15.9 63.6 69.7 89.2
Blegen (39 units) 8.6 18.0 69.0 90.0
Median
13
Unit Characteristics Overall Mean
Agency proportion, 10.06
Orientation proportion, 5.04
RN education, BSN 77.66
RN experience, years 7.77
RN turnover, 9.32
Patient turnover, ADT 9.12
admissions, discharges, transfers
14
Rate of Failure to Rescue
15
Correlation FTR and Staffing
Zero-order Correlation R2
Total hours -.795 .63
RN hours -.797 .64
RN proportion -.085 .01
RN shortfall .678 .46
16
First-order Correlation FTR and Staffing Removing First-order Correlation FTR and Staffing Removing
Zero-order Correlation Agency Proportion RN Turnover
Total hours (R2) -.795 (.63) -.756 (.57) -.762 (.58)
RN hours (R2) -.797 (.64) -.757 (.57) -.767 (.59)
RN shortfall (R2) .678 (.46) -.679 (.46) .654 (.43)
p lt .05, p lt .01
17
Rate of Falls by Type of Unit (n 560)
18
Relative Rate of Falls for each 10 Increase in
Staffing
Unadjusted and Adjusted Relative Rates
19
Rate of Medication Errors by Type of Unit (n
341)
20
Relative Rate of Medication Errors for each 10
increase in Staffing
Unadjusted and Adjusted Relative Rates
21
Summary of Findings

• Staffing effect found for all AEs
• RN shortfall effect found for FTR and falls
• Unit characteristics had little to no influence
  on association between staffing and AEs
• Unclear explanation for findings in unexpected
  direction

22
Limitations

• Measurement error
• Risk adjustment for falls and medication errors
• Other unmeasured factors
• Generalizability

23
Implications for Practice

• Longitudinal data presented opportunity to
  explore multiple predictors at the unit level
• Investment needed by hospitals to collect and
  monitor unit-level data
• Unique conceptualization and measurement of
  staffing

24
Future Research

• Unit-level risk adjustments methods
• Large scale unit-level studies
• Attribution of failure to rescue to unit-level
• Impact of organizational factors on staffing and
  patient outcomes

25

• Questions