Transcript here - Angelfire

Evidence-Based Practices in
Medications:
Psychiatric Services and Clinical
Knowledge Enhancement System
(PSYCKES)
4/7/2005
Edith Kealey, MSW, MA
Molly Finnerty, MD, Director
Bureau of Adult Services and Evaluation Research
Outline




Overview
 Motivation, Timeline, & Data Flow/Challenges
PSYCKES Demo
 Types and examples of content
Preliminary Results
 From OMH databases, evaluation, and usage logs
Future Plans
 Development
 Analysis Opportunities
2
Motivation


Purpose / Rationale
 Missing or inaccurate historical data can lead to duplicative
care, deviation from guidelines, and longer patient stays
 Raw data not enough – need intelligent decision support
 Administrative data is available and can be harnessed for
clinical decision support
Goals / Intended Outcomes
 Give MDs and supervisors access to data
 Support guideline-driven, cost-conscious QI at the state,
facility, ward, and patient levels
 Track change over time of fiscal and clinical quality
indicators at state and local levels
3
Development Timeline
Year
PSYCKES Activities
1998
Development of printed patient-level report summarizing
all pharmacy data since 1989 (6 hospitals)
1999
Field testing; physicians request a) timeline graph and b)
caseload summary
2001-2002
Printed report for all adult, non-forensic hospitals
2002-2003
Move to Web-based format, adding timeline graph,
caseload summaries, and management reports with drilldown capabilities. Beta testing in one hospital with 5 MDs
2003-2004
Addition of Infobuttons. Run expanded to include all
hospitals. Commissioner commits to statewide
implementation.
Dec. 2003April 2005
Version 1 implementation in all adult facilities.
4
PSYCKES Data Flow
1. NKI IRDB
2. OMH
MEDSoln &
DMHIS
3. State MCD
Database
4. Additional
OMH & State
Databases:
MHARS, etc.
OMH Atomic Pharmacy Data Warehouse
PSYCKES web-based application
1. State Inpatient
EBP Initiatives
2. Community Initiatives 3. State Outpatient
(MCD population)
Initiatives
5
Data Quality Challenges

Who



What



Drugs – misspelled drugs; new / systemic ones
Diagnosis – only admitting, not updated
When



Patient – multiple conflicting IDs
Attending – irregular updates; resident/limited permit MDs
not entered
Dates – gaps / overlaps
Frequency – multiple, hard-to-interpret values
Where

Ward – conflicting codes; poor linkage to Patient
6
PSYCKES Content & Demo
Performance Measures
 EBP/CPG based quality indicators
 Fiscal measures
 Data quality measures
 Medication reports
 Attending level reports
 Patient level reports
 Demo

7
EBP-based Performance Measures

Measures developed and incorporated into PSYCKES
for state inpatients at facility and clinician level:
 Antipsychotic polypharmacy
 Psychotropic polypharmacy
 Combinations of antipsychotics
 Dosing within recommended range
 Duration within recommended range
 Clozapine eligibility
 Dosing frequency > QD
8
Fiscal Measures

Measures developed and incorporated into
PSYCKES
 Cost savings by dose
 Cost savings by antipsychotic monotherapy
 Cost savings by form/frequency (to date
developed for risperidone only)
9
Data Quality Measures

Measures developed for state inpatients
incorporated into PSYCKES
 Missing date of discharge (false inpatient
status)
 Use of unknown frequencies
 Inappropriate frequency / dose
 Dosing gaps
 Concurrent admissions
10
Attending Caseload Summary


Patient Summaries
 Good dose / duration?
 Antipsychotic Hx
Guideline Measures
 % doses high/low
 % durations high
 % polypharmacy
 % long stays not on
clozapine
 Calculated on trial
basis
11
Patient Prescribing Summary



Demographics
Medication History
 Classification
 Admission
 Medication Trial
Details
 Start / Stop
 Max Dose (Trial)
 Dose Timeline
 Discharge Dose
12
Medication History Timelines






Used to determine which
regimens were effective in
the past
Dark-gray background
shows outpatient status
Medication history
grouped by drug class
12- or single-year views
Log scale – actual dose
Normalized scale –
whether dose in range
13
Demo for MDs – Click Clinical Reports to …
Fictitious PC
14
… Select self from list of MDs to …
Fictitious PC
Fictitious PC
15
… View own caseload. See that sole patient on 3
antipsychotics for 2 years, but Clozapine never
tried. Click patient name for more details …
Fictitious PC
Fictitious PC
16
…See all drug trials, with their durations, doses,
and effectiveness. Click Timeline to see that …
Fictitious PC
Fictitious PC
Fictitious PC
Fictitious PC
Fictitious PC
Fictitious PC
Fictitious PC
17
…Higher dose Risperidone worked at another
hospital in 1995, so switch to Risperidone or
Discharged on
Clozapine
Fictitious PC
Risperidone 16 mg qd
Fictitious PC
Maximum of
Risperidone 8 mg qd
18
Demo for Clinical Supervisors – Click
Antipsychotic Regimens by Patient to see …
19
… Patients on most complex drug regimens. Click
on Patient name to review prescribing history.
Fictitious PC
Fictitious PC
20
Click Concurrent Antipsychotics by Attending to
see …
21
… Distribution of polypharmacy by MD. Click on
a physician to review caseload and patient-level
data.
22
Demo for Operations – Fiscal reports show
potential cost savings from improved
guideline compliance, such as …
23
…Maximizing reasonable mono-therapy
24
…Reducing very high doses
25
Easy Access to Knowledge Resources:
Infobuttons and Web Sites
26
Current PSYCKES Limitations




Contains only psychotropic medications
Two months out of date
Tracks only current inpatients
Many data quality and completeness issues have been
challenging to resolve
27
Implementation Status
(as of March 31, 2005)
•
•
•
•
650+ registered users statewide
Trainings held at 17 adult facilities
• Interactive, hands-on format
• # sessions based on number of staff to be trained
• Follow-up technical support by phone and e-mail
Trainings at remaining adult facilities to be completed
by April 2005
Implementation planning initiated at 4 interested
children’s facilities
28
PSYCKES Usefulness Data




Rated the single most useful source of information about
patients’ medication histories (8.8 on a 10-point scale)
High average usefulness scores (6 or more on a 7-point scale) on
all dimensions of a standardized scale, including “useful in job”
(6.6), “accomplish tasks more quickly” (6.4), and “improve job
performance” (6.4)
Physicians without access to PSYCKES correctly identified only
24.6% of their patients’ medication trials, but improved to 76.9%
with PSYCKES
Physicians using PSYCKES recorded a 59.8% decrease in the
time needed to assemble a medication history
29
PSYCKES Usability Data


High average usability scores (6 or more on a 7-point
scale) on all dimensions of a standardized scale,
including “easy to become skillful” (6.2),”easy to use”
(6.1), and “clear and understandable interaction” (6.0)
Discomfort with computers is not a barrier to
navigating PSYCKES
 Average post-training test scores for those who rated
themselves below average on a “Computer Comfort
Scale” similar to the average of all respondents (89%
vs. 88%)
30
QI Performance over Time
(as of 12/31/04, adult facilities only)
Quality Indicator
1/31/04
12/31/04
>= 2 antipsychotics*
47.9%
44.6%
>= 5 psychotropics
16%
16%
>1000mg/day CPZ eq
44%
42%
>1year trial length
41%
43%
> QD dosing
39%
38%
*statistically significant decrease (p<.05)
31
PSYCKES Use by User Role
(Facility Users)
Type of user
Registered
users
Active users* (%
of reg. users)
Avg. hours
used/mo
(SD)
Avg. hits/ mo
(SD)
Administration
80
53 (66%)
2.0 (3.0)
15.2 (19.7)
Attending
Psychiatrist
138
101 (73%)
2.5 (3.7)
15.4 (16.2)
Supervising
Psychiatrist
39
31 (80%)
2.8 (4.4)
16.3 (20.4)
Other Clinical
212
47 (22%)
2.3 (5.3)
13.1 (25.2)
TOTAL
469
232 (50%)
2.4 (4.0)
15 (19.6)
*“Active user” means someone who used PSYCKES outside of a training session.
Data is as of 1/13/05.
32
PSYCKES Use by Attending
Psychiatrists
Variation across facilities
 % of active users ranges from 31% to
100%
 Avg. hours used per month ranges from
0.3 to 4.2
 Avg. hits per month ranges from 3.1 to
22.9
 Variation within facilities

33
Reasons for PSYCKES Use by Attending
Psychiatrists

Medication History Review
To assess dose and duration of past trials
 To check what medications have not been tried
 To narrow scope of chart review (e.g., date trial
ended used to find reasons for discontinuation)
 To check patients’ histories when transferred to new
ward
 To confirm patient statements re: medication history
 To supplement thinned/culled chart

34
Reasons for PSYCKES Use by Attending
Psychiatrists (cont’d)

Supervision
To prepare for case review
 To brief supervisor during case presentation


Documentation
To prepare court documents
 To prepare application for clozapine
 To prepare discharge note


Information

To obtain information about medications, guidelines
35
Reasons for PSYCKES Use by
Supervisors
 To
review patient hx prior to/during
case conference
 To identify candidates for case review
 To identify candidates for clozapine (in
conjunction with facility records)
 Aggregate data used as basis for group
discussion of practice style
36
Reasons for PSYCKES Use by
Administrators
 Tool
for dialogue around QI
 Tool for identifying data quality issues
 JCAHO
 Strategic Plan
37
Future Activities:
New Data (Versions 3 / 4)



Medicaid Data
 Outpatient Pharmacy & Services to fill-in gaps
 Detect continuity of care; compliance
Lab Values
 Blood levels for drugs
Clinical Outcomes
 Adverse Events (NIMRS) – falls, seclusion
 Symptom Scores – BPRS/TMAP, QOL?
 Vital Signs / Side Effects – obesity, BP, …
38
Future Activities:
New Report Development



Fiscal Reports
 Identify cost drivers, with drill-through to clinician,
patient, and drug levels
Clinical Reports
 Custom views for Adult and Child MH Diagnoses
 Reports for co-occurring systemic disorders
Progress / Trending Reports
 Fiscal and Clinical trends over time
 Statistical Analyses of change
39
Future Activities:
Analyses


Implementation Evaluation Study
 Impact of use on clinical, fiscal, and patient
outcomes
Exploratory Analyses
 Complex relationships between medications,
treatments, and outcomes
 Hypothesized and emerging trends (e.g. using
Clementine data mining tool)
 Develop new business rules (e.g. knowledge bases
and KNAVE)
40