Transcript W-A-1445-1545 Population Based Healthcare

Patient-centered Medical Home (PCMH)/
Medical Home Port (MHP) Evaluation
(with emphasis on chronic conditions, particularly diabetes)
June 2012
Eric Christensen
Overview
• Context
– MHS/Navy Medicine moving rapidly to PCMH/MHP models
– Literature shows the potential of PCMHs, but success is not automatic
– “78% of health spending is devoted to people with chronic conditions”*
• Purpose
– Assess the impact of the Bethesda PCMH on access, quality, and cost
– Assess whether the MHP model is effective for different patient types
and in different settings
• Outcome
– Resource allocation should consider the effectiveness of MHP model
– Which populations to target first with MHP
* Anderson and Horvath, “The Growing Burden of Chronic Disease in America,” Public Health Reports, 2004, 119(3):263-270
2
Topics/outline
• Results on access and quality
– HEDIS metrics
– Patient satisfaction survey
• Results on use and costs
– Overall
– Chronic versus non-chronic patients
– By chronic condition
• Clinical Practice Guidelines (CPGs) adherence—diabetes
• Note
– Some slides use NNMC and others WRNMMC as some of the work was
completed before the Bethesda-Walter Reed merger
– Results in this brief are for the WRNMMC internal medicine MHP
3
WRNMMC HEDIS scores (pre- and post-implementation)
Pre-period
Post-period
Favorable
(unfavorable)
difference
HbA1c test
84.5%
91.8%
7.3%
HbA1c > 9.0
25.9%
17.6%
8.3%
LDL screening
78.2%
88.6%
10.5%
LDL-C value < 100 mg/dl
53.3%
66.1%
12.9%
Asthmatics appropriately prescribed
93.9%
96.1%
2.2%
Pap smear test
80.6%
87.4%
6.8%
Mammography screening
75.3%
82.8%
7.5%
Colorectal cancer screening
60.7%
68.5%
7.8%
HEDIS measure
Note: The values for the pre-period are monthly averages for January-May 2008 compared to February-December 2009 for the post-period. The
transition period was from June 2008 through January 2009.
4
PCMH impact on access and patient satisfaction
0
Satisfied with healthcare at clinic
20
40
73.4
69.9
Access composite (getting care quickly)
74.3
64.6
Able to see PCM when needed
75.0
63.9
Satisfied with ease of scheduling appts
75.8
70.1
Provider communication composite
89.2
86.5
PCM listens carefully Often to Always
90.0
87.1
PCM explains things Often to Always
91.8
88.2
* Patient Activation Measure -13 is
copyrighted by Insignia Health, LLC
Walter Reed PCMH
100
88.3
83.7
Customer service (office staff)
Patient activation composite (level 4)*
80
82.0
78.4
High rating of PCM
Trust in PCM composite (PCAS)
60
75.6
73.6
68.4
61.9
Comparison site
5
How to increase satisfaction and PCM rating?
• Those reporting high levels of access and provider communication
report high satisfaction and PCM rating
– Access (OR: 2.1; CI: 1.4-3.2)
– Provider communication (OR: 1.9; CI: 1.2-3.0)
• Implies that increasing access and provider communication will
increase satisfaction and PCM rating
– But, how to do this?
• Drivers of access
– Ease of scheduling appointments (OR: 4.6; CI: 3.0-7.0)
– Ability to get appt for routine care when needed (OR: 4.4; CI: 3.2-6.2)
– Ability to get appt for urgent care when needed (OR: 3.7; CI: 2.5-5.5)
• Drivers of provider communication
– PCM listens carefully (OR: 13.5; CI: 6.4-28.4)
– PCM provides complete and accurate info (OR: 12.9; CI: 5.2-31.9)
6
Use and cost analysis
• Conducted retrospective data analysis (FY07-10)
– Transition period from June 2008 to January 2009
• Used differences-in-differences approach
• Used two-step process for analyzing health care use and costs
– Step 1: binary regression for user or non-user of a particular service
– Step 2: OLS regression for amount of services for users only
• Used NMC Portsmouth, NMC San Diego, and NHP Pensacola
internal medicine clinics as comparison sites
• Focused on chronic conditions
– Diabetes, hypertension, hyperlipidemia, chronic obstructive pulmonary
disease (COPD), coronary artery disease (CAD), and mental health
7
PCMH impact on use – all enrollees
30%
25%
Primary care use has increased. Presumably this has resulted in reductions
elsewhere. For example, the probability of ER use and inpatient
admissions are down, but average length of stay has increased (likely due
to better management of less severe cases).
20%
15%
10%
5%
0%
-5%
-10%
Inpatient
admissions
Inpatient days
Change in probability of use
ER visits
Specialty care Primary care
encounters
encounters
Change in amount of use for users
Change in average use
8
PCMH impact on use – chronic patients
30%
25%
The results for chronic patients are similar to the results across all
patients. The differences are that the increase in primary care is
greater and specialty care use is down for chronic patients relative
to the results across all patient.
20%
15%
10%
5%
0%
-5%
-10%
Inpatient
admissions
Inpatient days
Change in probability of use
ER visits
Specialty care Primary care
encounters
encounters
Change in amount of use for users
Change in average use
9
PCMH impact on use – non-chronic patients
30%
25%
20%
15%
Most of what we observe across all patients is
driven by or is a reflection of the experiences of
chronic patients as there is almost no statistical
impact on the use metrics for non-chronic patients.
The exception is specialty care. Non-chronic
patients have a higher probability of using specialty
care with the PCMH model than without the
PCMH model.
10%
5%
0%
-5%
-10%
Inpatient
admissions
Inpatient days
Change in probability of use
ER visits
Specialty care Primary care
encounters
encounters
Change in amount of use for users
Change in average use
10
PCMH impact on cost – all enrollees
Pharmacy costs
Ancillary costs
PMPQ costs
0%
-2%
-4%
-6%
-8%
-10%
-12%
-14%
-16%
-18%
-20%
For pharmacy and ancillary costs, both the probability of use and the
amount of use for users decreased substantially. For PMPQ, the decrease
is a result of a change in the amount of use for users and not a change in
the probability of use of the Military Health System overall.
Change in probability of use
Change in amount of use for users
Change in average use
11
PCMH impact on cost – chronic patients
Pharmacy costs
Ancillary costs
PMPQ costs
0%
-2%
-4%
-6%
-8%
-10%
-12%
-14%
-16%
-18%
-20%
The results for chronic patients are similar to the results
across all patients, but of a greater magnitude.
Change in probability of use
Change in amount of use for users
Change in average use
12
PCMH impact on cost – non-chronic patients
Pharmacy costs
Ancillary costs
PMPQ costs
0%
-2%
-4%
-6%
-8%
-10%
-12%
-14%
-16%
-18%
The results for non-chronic patients are generally of a
lesser magnitude than for chronic patients.
-20%
Change in probability of use
Change in amount of use for users
Change in average use
13
Cost impacts associated with chronic enrollees
Chronic
Nonchronic
PMPY without PCMH
$3,136
$750
PMPY with PCMH
$2,803
$697
Change in dollars
-$333
-$53
-10.6%
-7.1%
Total
Change
attributable
to chronic
enrollees
Estimated costs per enrollee
Change in percentages
Average PMPY change by percent of enrollees with chronic conditions
40%
-$165
80.7%
50%
-$193
86.2%
60%
-$221
90.4%
14
WRNMMC PCMH impact by condition
Diabetes
Hypertension
Hyperlipidemia
COPD
CAD
Inpatient admissions
-10.8%
Inpatient days
ER visits
Mental
health
20.2%
19.0%
-3.6%
-0.5%
36.0%
-13.5%
Specialty care
3.4%
Primary care
40.3%
32.0%
32.1%
46.3%
49.3%
24.8%
Pharmacy
-17.0%
-16.1%
-17.0%
-10.3%
NA*
-1.4%
Ancillary
-16.2%
-19.1%
-15.2%
-24.0%
-24.1%
-14.1%
PMPQ
-10.5%
-11.1%
-10.0%
-10.1%
1,595
7,098
7,207
960
WRNMMC enrollees
-8.2%
659
2,426
*Model would not converge.
15
Diabetes CPG metrics
• Screening/exam
–
–
–
–
HbA1c exam (at least yearly)
Eye exam (every two years)
Lipid (LDL-C) screening (yearly)
Nephropathy (yearly)
• Level/control
– HbA1c control (> 9.0%)
– Lipid control (LDL-C < 100 mg/dL)
– Blood pressure control
16
Diabetes CPG adherence rates—HbA1c exam
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
Comparison sites
WRNMMC nonPCMH
2007
2010
WRNMMC
PCMH
Medicare benchmark
17
Diabetes CPG adherence rates—eye care
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
Comparison sites
WRNMMC nonPCMH
2007
2010
WRNMMC
PCMH
Medicare benchmark
18
Diabetes CPG adherence rates—lipid control
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
Comparison sites
WRNMMC nonPCMH
2007
2010
WRNMMC
PCMH
Medicare benchmark
19
Diabetes CPG adherence rates—nephropathy
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
Comparison sites
WRNMMC nonPCMH
2007
2010
WRNMMC
PCMH
Medicare benchmark
20
Are changes in CPG adherence rates significant?
• Controlling for demographic differences and other chronic
conditions, PCMH patients are more likely to have yearly nephrology
and lipid control panels than patients at control sites
Procedure
Odds ratio for
post-PCMH
Significance
Nephrology
1.26
0.025
Lipid control
1.20
0.045
• HbA1c results are not meaningful because of the change in coding
practice at WRNMMC in 2010
• Eye exams are recommended every two years, but with a 1-year
comparison (2007 to 2010), there is a significant decline in patients
receiving eye exams
21
Staff survey – continuity and coordination of care
•
How often do you feel that you can exercise autonomy as opposed to
having to utilize a standard procedure?
– The question was only asked of providers
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
Never or rarely
Sometimes or often
Comparison clinics
Very often or always
NNMC PCMH
22
Enrollment status of diabetics (Navy catchment areas)
23
Chronic disease burden
•
“78% of health spending is devoted to people with chronic conditions.
Quality medical care for people with chronic conditions requires a new
orientation toward prevention of chronic disease and provision of ongoing
care and care management to maintain health status and functioning.”
– Health spending attributable to people with chronic conditions
 1 or more conditions: 88% for prescriptions, 72% for physician visits, 76% for inpatient
 2 or more conditions: 67% for prescriptions, 48% for physician visits, 56% for inpatient
•
Source: Anderson and Horvath, “The Growing Burden of Chronic Disease in
America,” Public Health Reports, 2004, 119(3):263-270
24
Pharmacy Variation
Utilization and variation for maintenance and scheduled drugs
Nevin Aragam, CNA Analysis and Solutions
Outline
•
We investigate Navy pharmacy utilization and variation for FY 2011
to identify and understand patterns among our population of
maintenance and scheduled pharmaceutical users




Methods
All drugs
Maintenance drugs
Scheduled drugs
26
FOR OFFICIAL USE ONLY
Methods
•
•
•
•
•
FY 2011 PDTS via MDR
Used catchment area of record to identify the eligible populations
surrounding navy catchment regions
Identified all eligible beneficiaries in FY 2011 using DEERS and the
demographic information from the most recent FM the beneficiary
appeared
Maintenance drugs are identified with the MDR PDTS field
Maintenance Drug = ‘Y’
Scheduled drugs and identified using the MDR PDTS field DEA
Class = 1, 2, 3, 4, or 5*
* note, there were no drugs identified as having DEA class 1 in PDTS
27
FOR OFFICIAL USE ONLY
All pharmacy age distribution
Age distribution for all pharmacy users
28
FOR OFFICIAL USE ONLY
All pharmacy
utilization
60,000 30 days supplies
implies the average 80
year old takes about 5
medications daily
30 day supplies per 1,000 Navy region beneficiaries by age
29
FOR OFFICIAL USE ONLY
Popular fill source
Source System
Direct
Retail
Mail order
Other
Total
AD
93.6%
5.3%
0.3%
0.8%
100.0%
ADFM
61.3%
37.4%
0.7%
0.6%
100.0%
RET
48.4%
40.6%
6.4%
4.5%
100.0%
RETDEP/
Other
40.1%
50.9%
5.3%
3.6%
100.0%
Total
53.8%
39.2%
4.1%
3.0%
100.0%
Source: MHS Data Repository (MDR PDTS table FY2011).
• AD beneficiaries fill almost exclusively at MTF
pharmacies
• Active duty family members and retirees fill mostly at
MTFs and a sizable proportion at retail pharmacies
• Retirees and their dependents fill fairly evenly across
MTFs and retail pharmacies
30
FOR OFFICIAL USE ONLY
Types of maintenance drugs
•
Top 5 Maintenance drugs (15% of all maintenance drugs
prescribed)
1. IBUPROFEN
2. SIMVASTATIN
3. LISINOPRIL
4. NEXIUM
5. LIPITOR
31
FOR OFFICIAL USE ONLY
Maintenance drug age
distribution
Age distribution for maintenance pharmaceutical users
32
FOR OFFICIAL USE ONLY
Maintenance drug
utilization
50,000 30 days supplies
implies the average 80
year old takes about 4
maintenance medications
daily
30 day supplies of maintenance drugs per 1,000 Navy region beneficiaries by age
33
FOR OFFICIAL USE ONLY
Maintenance drug regional
variation
High:
NH Pensacola
(13,307 30 day
supplies)
Low:
NCA MSMA (9,460
30 day supplies)
High/Low: 1.41
(Beneficiaries in NH
Pensacola use 1.4
times as many
maintenance drugs
as those in the NCA
MSMA)
Regional variation for maintenance drug prescriptions per 1,000 beneficiaries
34
FOR OFFICIAL USE ONLY
Scheduled drugs
Schedule
Abuse
potential
Accepted
medical
use
Likelihood of
psychological or
physical
Dependence
Examples
I
High
No
No accepted safety
for use
Heroin, LSD, marijuana
II
High
Yes
High
Amphetamine (Adderall®),
methamphetamine
(Desoxyn®), cocaine
III
Moderate
Yes
Moderate
Vicodin®, Tylenol with
codeine®, ketamine
IV
Low
Yes
Limited
Alprazolam (Xanax®),
diazepam (Valium®)
V
Low
Yes
Limited
Robitussin AC®,Phenergan with
Codeine®
35
FOR OFFICIAL USE ONLY
Types of scheduled drugs
•
•
Top 5 Scheduled drugs (55% of all schedules drugs prescribed)
1. HYDROCODONE-ACETAMINOPHEN
2. ZOLPIDEM TARTRATE
3. OXYCODONE-ACETAMINOPHEN
4. ALPRAZOLAM
5. DIAZEPAM
Scheduled drugs can also be maintenance drugs:
1. CLONAZEPAM
2. CONCERTA
3. LYRICA
4. ADDERALL XR
5. VYVANSE
36
FOR OFFICIAL USE ONLY
Scheduled drug age distribution
Age distribution for scheduled pharmaceutical users
37
FOR OFFICIAL USE ONLY
Scheduled drug
utilization
1,900 30 days supplies
implies the average 80
year old takes about 1.5
scheduled drug
medications daily
30 day supplies of scheduled drugs per 1,000 Navy region beneficiaries by age
38
FOR OFFICIAL USE ONLY
Scheduled drug regional variation
High:
NH Pensacola
(1,073 30 day
supplies)
Low:
NCA MSMA
(567 30 day
supplies)
High/Low: 1.89
Regional variation for scheduled drug prescriptions per 1,000
(Beneficiaries in NH
Pensacola use
nearly twice as
many scheduled
drugs as those in
beneficiaries
the NCA MSMA)
39
FOR OFFICIAL USE ONLY
Questions?
• Nevin Aragam
• [email protected]
40
FOR OFFICIAL USE ONLY
Appendix
•
•
•
•
•
•
•
•
•
•
•
Navy region definitions by catchment area DMIS ID:
0067 0066 0037 0123 - NCA MSMA
0124 0120 – Tide Water MSMA
0024 0029 - San Diego MSMA
0125 126 127 - Puget Sound MSMA
0104 - NH Beaufort
0091 - NH Camp Lejeune
0039 - NH Jacksonville
0028 - NH Lemoore
0038 - NH Pensacola
0030 - NH Twentynine Palms
41
FOR OFFICIAL USE ONLY