Transcript RESOURCES YOU WILL NEED

Lessons Learned
from Psychotropic Medication Monitoring
for Texas Foster Children
A joint project of
Texas Department of Family and Protective Services (DFPS)
Texas Health and Human Services Commission (HHSC)
Texas Department of State Health Services (DSHS)
All opinions are the authors and do not represent an official position or policy
BACKGROUND





Concerns were raised about psychotropic prescriptions given to fostercare children around
2004.
In February 2005, DFPS, DSHS and HHSC released the Psychotropic Medication Utilization
Parameters for Foster Children, which were updated in June 2007 and most recently in
December 2010. The current version can be found at:
http://www.dfps.state.tx.us/Child_Protection/Medical_Services/guide-psychotropic.asp
These Parameters use eight criteria to indicate a need for further review of the child’s
medication regimen. Including such things as:
 Multiple drugs in the same class
 Five or more drugs
 Lack of appropriate diagnoses
 Dosage outside of recommendations
A system was needed to assess the degree to which these parameters were being met. I was
involved in initially setting this system up and continue to work relatively closely with it.
We were very fortunate to have cross agency support and cooperation at all levels to get this
project done.
RESOURCES YOU WILL NEED
Access to Medicaid data.
You will need eligibility, claims/encounters and prescriptions.
Obtain reference files and decode tables to make sense of the
Medicaid data.
Doctors specialty information, drug class descriptions and
related information, diagnoses and procedure code descriptions.
Access to your own programs administrative data if you don’t
have dedicated Medicaid program codes for foster care.
A statistical program such as SAS, SPSS or R and one (or more)
persons who can serve as programmer.
A relatively high initial set up cost in time and effort but a
lower ongoing cost to keep the data flowing once its set up.
One or more doctors and program experts at least on occasion
for consultation.
Some consultation needed at the start and more later for
specific report interpretation.
IMPORTANT QUESTIONS
Who will you keep?
We keep only children age 0-17 for the time they are
actually in foster care and under the state guardianship,
other Medicaid time is excluded.
How often and how long will you look at the data?
We look at quarterly reports, and annual reports. A rolling
six month time frame is probably ideal, but people like firm
time windows with clear cutoffs too.
What will be your review process, medically complex cases
can have surprising medical needs?
Cases outside parameters are reviewed by the HMO
administering the plan.
How will you communicate with your providers and how
often?
Currently
How will you treat all the different varieties of foster care
such as relatives, substitute care, residential treatment
facilities?
At this point we don’t distinguish these on a regular basis.
our primary communication is on a formal
quarterly basis with the HMO and as needed as well.
Approximately every year we review the highest
providers.
Children outside parameters should be flagged at any
point by the HMO’s data systems.
MAJOR DRUG DECISIONS
What drugs will you keep?
We keep most antipsychotics, antidepressants,
anticonvulsants and stimulants as well as
miscellaneous drugs.
Will you exclude drugs based on diagnoses?
We exclude anti-convulsants based on seizure disorder
diagnoses and diphenhydramine (Benadryl) based on
allergy diagnoses.
How many prescriptions and how long?
We look at all, but the focus is on two or more
prescriptions covering 60 days. Since all we have is
claims data, we use this criteria to indicate a drug is
not a test, but a serious treatment.
How is polypharmacy defined?
We use 60 concurrent overlapping days for two drugs
to define the minimum bound of polypharmacy. One
major idea is the availability of drugs to be
metabolized simultaneously. Different prescriptions in
different parts of a month doesn’t seem like much like
polypharmacy it seems like sequential medication.
What will be your parameters for deciding what might
too much, too young?
Currently five or more of any drug, and two or more in
the same class. We use our own formulary classes (but
AHFS or Medi-Span would work just as well).
Looking at only single drugs will create enormous
workloads. Some parameters change based on age.
BASIC STEPS
1. Merge the Medicaid data files with their references.
2. Then more merges and roll ups:
Roll up patients and flag services and diagnoses,
Roll up scripts by chemical names by patients (calculate lengths of scripts),
Roll the chemical names up by class by patients (calculate length and
polypharmacy),
Summarize patients with flags for drug class and chemical name and length of
scripts.
3. Retain merged and labeled files to run further analyses and reports.
CHALLENGES
 Prescriptions and diagnoses are not connected with one another.




You have to pull the diagnostic information from the claims/encounters.
Diagnoses and Drugs have to be matched based on the person id.
Individual drug prescriptions have to be built up into larger groups to cover periods
of time.
There will probably be a time lag in your data probably a minimum of 30 days and
probably more like 90 days plus.
USEFUL REPORTS AND DATA
A big picture overview of patients
A summary of prescriptions for all clients and by age, maybe by
program type or facility.
Individual Patients prescriptions
The individual prescription claims data for clients who are
outside of parameters, which requires going back from the
rolled up flag file to the original claims data.
An overview of doctors
A summary of the number of patients each doctor has with the
number with prescriptions in certain categories: 60 day plus,
class poly pharmacy, multi drug poly pharm, and poly pharm.
A Break down by type of doc: Psych vs Others.
Basic Research/Reference Files
A file that rolls up and merges both prescription information
and diagnoses information by patient (one row per client).
Roll up the start and end dates for all drugs by chemical name
received in a period of time and keep them as a reference for
patients outside of parameters so you can see what drugs, when
and how long for polypharmacy patients.
Helpful troubleshooting file for doctors
Interleave the raw data lines by date of prescriptions, services
with diagnoses (one row per claim).
Update on the Use of Psychotropic Medications in
Texas Foster Children: Fiscal Years 2002-2011
35.0%
Percentage of Children in Foster Care
30.0%
28.0%
29.6%
29.9%
26.4%
25.0%
24.7%
25.0%
21.5%
21.2%
20.5%
20.0%
19.3%
36% Decrease
since 2004
15.0%
10.0%
5.0%
5.0%
3.3%
4.0%
0.7%
1.1%
1.4%
FY2002
FY2003
FY2004
68% Decrease
since 2004
3.4%
2.5%
2.5%
0.9%
0.7%
0.7%
FY2005
FY2006
FY2007
0.0%
Psychotropic Meds 60 days+
2.0%
1.8%
1.8%
1.6%
0.6%
0.5%
0.5%
0.4%
FY2008
FY2009
FY2010
FY2011
Class polypharmacy
Five or more Meds polypharmacy
70% Decrease
since 2004
CLOSING THOUGHTS
Our results indicate that both the parameters (~ 05-07) and managed care
(~ 08-11) have had relatively large effects.
•
•
•
•
The parameters and provider outreach decreased a baseline (~ 02-04) of
either multi-drug or class polypharmacy of slightly over 5% to well below
4%.
The parameters and provider outreach alone decreased either multi-drug or
class polypharmacy from below 4% to slightly over 2%.
Overall both interventions have resulted in well over a 50% reduction in
either type of polypharmacy.
We are hoping for further improvements with our participation in MEDNET
and further refinements of our monitoring system.
Setting up a monitoring system and looking at your data can help you stay
on top of the drugs and diagnoses in your system of care and help
improve quality of care for your patients.