Using Prospective MDS Assessment Data and Clinical Information to Improve Quality of Care in Nursing

1

• Using Prospective MDS Assessment Data and
  Clinical Information to Improve Quality of Care
  in Nursing Homes
• American Health Quality Association 2003
  Technical Conference
• Orlando, FL February 2003
• Christie Teigland, Principal Investigator
• Colene Byrne, Project Manager Outcomes Research
  Analyst

2
Prospective Use of MDS Data

• We hope to achieve a shift in focus
• ?from using MDS data for investigating adverse
  outcomes in a facility after they occur
• ?to a preventive focus centering on safety of
  individual residents before the adverse event
  occurs.

3
If information about individual
residents risks for adverse outcomes can be
gotten into the hands of staff that work with the
residents, they would then be able to adjust
interventions individually based on risk factors,
and even focus on prevention of adverse resident
outcomes.This initiative could transform
nursing home care by utilizing existing data
sources and known risk relationships to actually
shape nursing home care in a positive way.
Nancy Watson, PhD, RN, Director of the Center for
Clinical Research on Aging, University of
Rochester Medical Center
4
American Medical Directors Association Caring for
the Ages Newsletter (April 2002)

• We shouldnt be looking at falls only because
  the surveyors ask about them or its part of CMS
  QIs
• Medical Directors and Administrators would be
  wise to develop and implement a comprehensive,
  facility-wide process for determining causes and
  assessing risks of falls.
• Otherwise, caregivers may miss important
  diagnostic clues, thus bypassing opportunities to
  correct modifiable risk factors and avert
  subsequent falls.

5
The basic guidelines for screening, management,
and treatment to reduce elder falls and injuries
are known, but there is still work to do.
6
Patient Safety Projects 2001 Research Awards

• 94 grants Only 7 in long term care!
• Congress Health systems and providers should
  utilize all available and appropriate
  technologies to reduce the probability of future
  medical errors.

7
Project Background

• Nursing staff are plagued with too muchdata and
  too little information.
• Staff need easy access to the right information
  at the right time if it is to be used
  effectively to improve the outcomes of care in
  nursing homes and other long term care settings.
  
• Giving clinicians information regarding the
  likelihood of a preventable adverse outcome prior
  to the outcome actually occurring, and
  identifying the resident- specific risks
  involved, will greatly reduce the occurrences of
  these events.

8
Project Background

• It has been demonstrated that many healthcare
  errors and adverse outcomes are due to peoples
  limitations as data processors.
• Avoiding practice errors and preventing adverse
  outcomes requires committing more time to
  processing patient data, but medical and nursing
  staff are simply too busy to consistently analyze
  and detect the multitudinous conditions specified
  by the numerous protocols and standards of care.

9
Project Background

• This study has developed web-based on-demand
  reports
• alerting staff to those residents at greatest
  risk for an adverse outcome
• resident-specific risk profiles specifying
  addressable risk factors so preventive actions
  can be taken.
• These reports are used in care plan development
  and staff education and replace numerous ad hoc
  risk assessment tools and manual forms, thus
  saving significant staff time and ensuring all
  risk factors are identified and addressed.

10
Measurable and Sustainable Improvements in
Quality of Care

• Although a few falls have a single cause, the
  majority result from interactions between
  long-term or short-term predisposing factors..
  Preventing Falls in Elderly Persons, New England
  Journal of Medicine, Jan, 2003
• Medicare is spending billions to treat
  preventable injuriescost of 1,272 per
  incidentinterventions are not widely
  disseminated.Nov-Dec issue of Health Affairs

11
3 Preventable Adverse Outcomes

• Falls
• Pressure Ulcers
• Urinary Tract Infections (UTIs)

12
Predicting Falls Using MDS
Risk Factors in MDS from Previous Research,
RAPS, IP,Literature
Previous Fall
Cog Decline
Wandering
Meds
Linked Historical Data
Restraints
Age
Mobile, Needs Assist
Cane/Walker
New and Improved Statistical Models
Improved Resident- Specific Risk Profiles
13
New and Improved Indices Predictive of Risk of
Falling
Predicts Probability of Falling by ADL Index Score
ADL Score 6 thru 11
ADL Score 5, 12, 13
ADL Score 4, 14, 15
14
Onset of Cognitive Impairment

• The measure of cognitive impairment used in the
  CMS QIs uses only 2 MDS items B4 decision
  making and B2a short term memory problem.
• Definition misses a large number of cognitively
  impaired residents at all levels.
• Only 3 QIs are currently risk adjusted for
  cognitive status.

15
Comparison of CPS to Current QI Definition of
Cognitive Impairment
CPS
CPS Codes 0Intact 1Borderline 2Mild
3Moderate 4Moderately Severe 5Severe 6Very
Severe
QI Categories 0 No Impairment 1 Cognitively
Impaired 2 Severely Cog. Impaired
NYS MDS Database 2000 Annual, Significant
Change, Significant Correction Prior Full
16
Prevalence of Fallsby Cognitive Performance Level

• Rate of falls increases with level of cognitive
  impairment.
• QI Prevalence of falls should be risk adjusted
  using CPS scale.

17
Falls Risk Factor Information from MDS Used to
Predict -- Provided in Resident Risk Profiles
18
Examples -Analyzing MDS Data Identifies Highest
Risk Categories
Probability of Falling Degree of
Assistance Independent
17.7 Supervised or Total Dependence 24.0 He
lp to Transfer or Bathe, or did not bathe 37.7
Bladder Training Program None
30.0 Yes 43.5 Level of Continence
Total Incontinence
21.6 Continent 29.5 Somewhat
Incontinent 40.0
Underscores importance of accurate coding to risk
assessment!
19
Falls Risk Prediction Model
26 of residents in these 3 risk groups
9 of residents in these 3 risk groups
74 of All Falls (in next quarter) Were for
Residents in our 3 Highest Risk Categories
Accurately predicted 75 of Falls in Very High
Risk New Residents
Overall, false positives and negatives less than
10.
20
Types of Risks Reports Currently Available
Highlight the one you want and click.
21
(No Transcript)
22
Just click on any resident name to see the
residents risk profile report.
23
Each risk profile is unique to resident!
24
Residents Assigned Same Risk Level Can Have Very
Different Set of Risk Factors
25
CMS QI Prevalence of Stage 1-4 Pressure
Ulcers
Many more residents are at highrisk than
identified by QI definition.
26
CMS QM Residents with Pressure Sores Two
different QMs (1) without FAP and (2) with FAP
27
New CMS QMs most recent quarter!
28
Many Resident Characteristics Increase Risk of
Pressure Ulcers and Are NOT Used for
Risk-Adjustment in QI or QM
Linked Historical Data
These are used in our statistical models to
predict risk of pressure ulcers
29
New and Improved Predictive Model Explored Risk
of Pressure Ulcers
New Mobility Index Predicts Better than
Individual MDS Items
30
NYAHSA AHRQ Patient Safety Project Research
Findings Stage 1-4 Pressure Ulcers

Risk Factors
ODDS RATIO   Previous Pressure
Ulcer (past 8 quarters) 6.00 Indwelling
Catheter 2.39 Admitted w/in last 6
months 2.11 Dehydrated 2.09 Limited
Physical Functioning 2.05 Recent Hospital
Stay 2.05 Weight Loss 2.00 Medications
Affecting Mood 1.93
31
Other Risk Factors
ODDS RATIO UTI 1.91 Deteriorating Health
Status 1.89 Edema 1.64 ER Visit (one or
more) 1.59 Hip Fracture 1.55 Pain 1.53 Incon
tinence of Bowel 1.41

• All of these risk factors put residents at far
  greater risk for development of a pressure sore.
• Many are addressable or modifiable.
• Identification and care planning for these risk
  factors can reduce the occurrences of this
  adverse outcome.

32
(No Transcript)
33
(No Transcript)
34
Different set of risk factors!
35
Risk Levels for All 3 Outcomes
36
Click on any column to SORT!
37
Project Evaluation

• Monitor utilization of risk reports.
• Investigate how at risk reports changes care
  plans/practices and QI processes.
• Evaluate accuracy of at risk reports (predicted
  vs. actual falls, pressure sores, UTIs).
• Determine whether information reduces occurrences
  of adverse events (using time series models to
  compare to control group).
• Develop lessons learnedand key success factors
  in using informatics in long term care.

38
Growing Publicity Regarding Quality of Care in
Long Term Care Gives Greater Importance to
Providing the Tools Needed toEvaluate,Ensure,
and Improve Quality of Carefor Our Nations
Frail Elderly