Transcript Enhancing Health Management: Understanding Physician

Enhancing Health Management:
Predicting Physician Utilization of
Integrated Electronic Prescribing
Laurel K Taylor
McGill University
6 June 2007
Orlando, Florida
Messages
o Things can be better in health care
o Technology is a key facilitator for improvement
o Technology uptake extremely variable
o Utilization of technology can be predicted /
modified
Electronic Decision Support Systems
The Potential
o Provide complete information on current drugs for
physicians and pharmacists
o Reduce prescribing and transcription errors
o Improve match between need and therapy
o Enhance compliance
o Improve disease management and patient outcomes
o Increase clinical and cost effectiveness of treatment
Inappropriate Prescriptions
Proportion of patients with at least
one inappropriate prescription
-The Canadian Context-
5-8
120
physicians
16%
100
9+
1
physicians
physicians
3%
27%
80
60
40
3-4
20
0
physicians
30%
1
2
3-4
5-8
9+
Number of prescribing physicians
2 physicians
24%
49% of patients visit 3 or
more physicians
Source: Tamblyn, CMAJ, 1993
Inappropriate Prescriptions
-The Canadian ContextMultiple
Pharmacies
41%
Inappropriate Rx per
1,000 Patient Visits
32
31
30
29
28
27
Only 1
Pharmacy
26
59%
25
One Pharmacy
Many Pharmacies
Source: Tamblyn, CMAJ, 1993
Many patients visit
multiple pharmacies
Electronic Decision Support Systems
The Challenges
o Inconsistent features across applications
o Lack of integration with existing IT systems
o Poor integration with provider work flow
o Lack of systematic and rigorous evaluation
methodology
Electronic Decision Support Systems
The Challenges
Extreme variability in physician
utilization
Unrealized benefits
Primary Care And IT
-The Canadian Context-
o
23% - electronic medical records
o
15% - access to hospital records
o
11% - e-prescribing capabilities
o
8% - electronic test ordering
International Health Policy Survey of Primary Care Physicians in Seven Countries,
The Commonwealth Fund, 2006
Research Objectives
To define and analyze predictors of physician
utilization of electronic prescribing through an
integrated drug and disease management
system.
Research Setting
MOXXI Project
(Medical Office of the XXI Century)
o
61 general practitioners
o
26 practice sites
o
Located in an urban Canadian centre
o
Developed physician and practice characteristics
based on 18 months of data prior to implementation
o
o
o
o
Survey data
Medical services claims database
Medication services claims database
Collected 6 months of electronic prescribing utilization
data subsequent to implementation of an electronic
integrated drug and disease management system.
o
Electronic audit trails
MOXXI
Perceived Benefits of the System
Drug Cost Information
Stop/Change Function
Info on ER visits &
Hospitalization
Drug Monograph
Current Medications List
Refill Compliance Indicator
List of RX prescribed by Others
Drug Interactions
Re-prescribing function
Printed Prescription
1
Not
Beneficial
2
3
4
5
Very
Beneficial
Physician Questionnaire Rating 4 Months Post Implementation (October 2005 – February 2006)
MOXXI
-Patient CharacteristicsParticipating
Not Participating
n (%)
n (%)
7471 (40.2)
29166 (40.2)
11133 (59.8)
40873 (56.4)
1334 (7.2)
23226 (32.0)
30-45
2728 (14.7)
16439 (22.7)
46-60
6045 (32.5)
17140 (23.6)
>60
8497 (45.6)
13233 (18.3)
n (SD)
n (SD)
4.4 (3.7)
2.5 (2.6)
Sex
Male
Female
Age (years)
<30
Average # Visits
MOXXI
-Physician Characteristicsn (%)
Year of Graduation
>1980
31 (51)
1960-1979
30 (49)
Sex
Male
33 (54)
Prior Computer Experience
<5 hours/week
36 (58)
5-15
22 (35)
>15
3 (5)
Physician Typology
Pragmatist
39 (64)
Receptive
9 (15)
Seeker
Traditionalist
10 (16)
3 (5)
MOXXI
-Practice CharacteristicsMean
SD
Range
Number of Unique Patients
1840
877
19-3880
Number of Patient Visits
4193
1703
23-9085
Continuity of Care Index
0.57
0.09
0.22-0.72
Average Medication Use
2.84
0.83
1.27-4.74
MOXXI
Utilization Indicator – e-Rx/visits
Study Period:
1 Oct, 2005 – 3 July, 2006
Include all patients consented
before index date.
Numerator:
Select all patients
included in the
denominator.
Denominator: Select patients that
had an outpatient
visit during the
study period.
Select visit if e-Rx
written (prescription,
not dins) from
MOXXI .
Select patients
consenting to
MOXXI before 1 Oct
2005.
# e-Rx
Select patients
visiting physician
in outpatient
setting
# visits
MOXXI
Results: Full Model R2=.4997
Physician Characteristics
E-Rx Rate
p-value
Sex
Male
Female
33.2
27.5
.2528
ref
Grad Year
1965
1981
1999
36.3
30.8
24.8
.2787
Typology
Seeker
Receptive
Pragmatist
Traditionalist
24.0
37.7
32.5
4.5
.0732
.0042
.0058
Ref
Prior Computer
Experience
< 5 hrs/week
5-15 hrs/week
> 15 hrs/week
27.0
35.8
48.8
.0040
MOXXI
Results: Full Model R2=.4997
Physician/Practice Characteristics
E-Rx
Rate
p-value
Continuity of Care
Lowest Quartile
Second Quartile
Third Quartile
Highest Quartile
33.7
34.3
28.8
25.6
ref
.7063
.4717
.5843
Average Medication Use
(2.9 drugs)
30.6
.0358
Patient Volume
Lowest Quartile
Second Quartile
Third Quartile
Highest Quartile
34.5
29.3
36.1
22.4
ref
.6209
.7881
.3103
MOXXI
Results: Final Model R2=.4633
Physician/Practice Characteristics
Estimates
p-value
.1971
.3310
.2989
Ref
.0404
.0010
.0010
Ref
Prior Computer Experience
.0096
.0018
Medication Use
.0758
.0027
.0740
.0974
Typology
Practice Volume
Seeker
Receptive
Pragmatist
Traditionalist
Third Quartile
MOXXI
- Implications for Practiceo Implementation may require staged approach
o Modular approach to physicians with little or no
computer experience
o Early intervention where necessary
o Deeper understanding of credible evidence for
practice decisions
o Integration into current workflow important
MOXXI
- Implications for Policyo
IT availability insufficient to sustain utilization
o
Need to identify strategies to enhance adoption
and utilization
o
May require availability of customized training
programs
o
Rigorous evaluation of clinical applications for
features, workflow integration assessment
Acknowledgements
Support for this research was provided by:
o
The Commonwealth Fund.
”The views presented here are those of the authors and should not be attributed to
The Commonwealth Fund or its directors, officers, or staff.”
o
Canadian Institutes of Health Research NET Grant
o
Canadian Health Services Research Foundation
Medical Office Of The XXI Century (MOXXI)
-Backup Slides-
*Total Patient Consents = 9052
*Total e-Rx Written = 7990
MOXXI System Overview
Doctor’s Office
eRx
Régie de l’assurance
maladie
MOXXI
Server
Printer
Real-time
adjudication
eRx
Chart
Patient
Pharmacy
Technology Adoption Model
34%
Early
majority
2 ½%
Innovators
13 ½%
Early adaptors
34%
Late
majority
16%
Laggards
Time of adoption innovations
Understanding Predictors of Utilization
The TAM (Technology Acceptance) Model

Developed by Davis in 1989 for predicting user acceptance of
computers
Perceived
Ease of Use
Behavior Intention
Perceived
Usefulness
Computer Usage
Understanding Predictors of Utilization
The Physician Typology Model
Seeker
Evidence
Receptive
Traditionalist
Pragmatist
Experience
Nonconformity
Practicality
Green, Gorenflo and Wyszewianski, 2002
Medical Office Of The XXI Century (MOXXI)
Utilization Indicator – e-Rx/Rx
Study Period:
1 Oct, 2005 – 3 July, 2006
Include all patients consented
before index date.
Numerator:
Select the DINs from
denominator and match to
an eRx during the study
period
Denominator: Select patient if ≥
1 dispensed DIN
*Total Patient
Consents
during
study = 9052
*Total e-Rxperiod
Written = 7990
Select patients
consenting MOXXI
patients before 1
Oct 2005.
# e-Rx
DINs
Select patients with
RAMQ coverage
(75% not gaps)
during study period
Select DINs
# visits
prescribed during
study period