“BEST PRACTICES RESEARCH” Jim Mold, M.D., M.P.H. The University of Oklahoma Department of Family and Preventive Medicine The Oklahoma Physicians Resource/Research Network (OKPRN)
Jun 19, 2015
“BEST PRACTICES RESEARCH”
Jim Mold, M.D., M.P.H. The University of Oklahoma
Department of Family and Preventive Medicine
The Oklahoma Physicians Resource/Research Network (OKPRN)
OBJECTIVES• Teach you about “best practices
research” as a concept
• Teach you the method
• Give you enough examples that you feel that you could do it
• Share some of our discoveries
Process of Care Questions What is the best way to maximize
pneumococcal vaccination rates? What is the best way to handle laboratory
test results? What is the best way to manage
prescription refills? What is the best way to manage diabetic
patients What is the best way to maximize
colorectal cancer screening rates?
Traditional Approach Measure current performance
Choose or construct a theoretical model or examine barriers, facilitators, opportunities, and threats
Design an approach that ought to work
Test the approach in an RCT or before/after
“Best Practices Research” Identify the steps/components involved in
the process under investigation Define “best” for each step in terms of
values (e.g. accuracy, efficiency) Identify existing “best” methods for each
step by finding exemplars and examining their methods
Combine best approaches into a unified best method
(Test combined method in an RCT)
Advantages Draws upon the wisdom and
experience of clinicians/end user
Efficiently gets to an answer
The answer is likely to be practical, feasible, acceptable, and effective
Disadvantages Don’t learn much about why something
works
Perhaps no one has figured out a particular step
Solutions identified may be unique to a practice
Parts may not fit together well
First Effort – Pneumovax Funded by Merck Vaccine Division Steps not identified in this case Literature review summarized and
shared with participants Financial incentive and opportunity
to improve Highest baseline rate Most improvement
Increasing Pneumococcal Vaccination Rates
Practice Performance Audits
Clinician
A B C D E F
Initial Rate
12% 15% 18% 33% 35% 67%
Final Rate
15% 15% 22% 36% 36% 86%
Pneumococcal Immunization Physician must believe in it Nurse authorized to give it (standing
orders) Physician must conduct regular
oversight/review Immunization clinics in Fall;
pneumovax linked to flu shots
Increasing Delivery of Preventive Services
Preventive Services Reminder System Nurse-operated PDAs linked to
decision support and registry
Printed summary of services due and done for review by physician
Preventive Services Reminder System2
PDA/Clinic
Appointment Database
Prev. Serv. Database
IntoleranceRecord
AllergyRecord
RiskRecord
Service
NotServiceReason
PastServiceRecord
NotImmReason
PastImmRecord
Vaccination
PatientInsurance
PatientAppointment
PIF Enterprise Server
(Patient Info)
Reference TablesReference TablesReference TablesReference Tables
Pendragon Internet Forms (PIF) PIF Entp. Server
Java Server(Recommendations)
Patient Database
Central Server/OUHSC FMC
PSRS is a comprehensive electronic tool designed to improve documentation and enhance delivery of primary and secondary preventive services. This system includes 3 integrated subsystems: a Palm Operating System -based PDA (Palm, Handspring and Sony) & PC running Widows 98/2000/XP Operating System, connected to the PDA and a Central Server System (Enterprise Server and Recommendation JAVA Server).
Efficacy of the PDA Version
2-3 year olds Controls PSRS p-value
DTaP#4: 53% 86% 0.001
HepB#3: 61% 93% 0.0005
Pneumo: 27% 38% NS
MMR#1: 61% 93% 0.0005
Efficacy of the PDA Version
Adult Diabetics Controls PSRS p-value
Smoking status: 70% 93% 0.02
Smoking counselling: 13% 78% 0.0004
Pneumovax: 33% 78% 0.0003
Management of Laboratory Test Results
Identification of model/steps in the process Literature review
Focus groups
Listserv discussions
Delphi process
Management of Laboratory Test Results
1. Track tests sent out until the results come back
2. Notify patients of test results3. Document patient notification4. Track patients with abnormal
results to be sure that they follow-up
Defining “best” Identify values/Set standards
All steps: Accuracy (90%)
All steps: Cost (<$5 per patient)
Step 2: (Patient notification) patient satisfaction (>90% satisfied)
Physician surveys, blanket audits
Selective practice audits
Identifying potentially effective methods for each step
Lab Tests 11 practitioners satisfied with their
method for at least one step
2 different methods identified for each step
Audits of practices; time/motion studies
Patient reports regarding time/satisfaction
Combining Methods for Steps into a Combined Best Method
Choose best methods for individual steps
Try to put them together into a process that makes sense
Lab Tests Log and nurse tracking (dual system)
Physician writes note to patient on lab results sheet
Lab results sheet dated; copy mailed to patient generic explanation of tests for chemistry panels
??Tickler file system??
Lab Tests – Time/Cost $5.17 per set of lab tests for steps 1-
3 Almost half of the cost is physician
time Methods that rely on nurse/patient
call backs are more expensive
LabMan Laboratory Test Results Management
Appl. 5
Screen 1. Shows list of labs due “today” and list of follow-ups due “today”. This list can be populated by lab type (“Populate Labs Due By Type”) and by patient name. “Follow-ups Due” generates a list for due follow-ups only. “*” indicates that a lab is due, while “~” indicates that a follow-up is due on an abnormal lab result. Due lab and follow-up lists can be printed via an infrared printer port from the PDA for review, or documentation. Screen 2-3. Lab result data detail. Lab can be selected from a drop-down menu (“Lab Due”), and the lab result, lab order date and return time can be entered as well. Default return time can be customized for each lab. “Lab Due Date” is calculated by the PDA automatically. If labs come back, the user can check the lab off (“Check, if lab came back”). At this point the user can keep, or delete the lab and, if abnormal, can schedule a follow-up for the lab. Labs can also be deleted manually (“Delete Lab”). Screen 4. The user can schedule a follow-up for abnormal labs, by selecting the lab type and entering a short message that indicates the nature of the follow-up. A pre-formulated quick-entry drop-down menu assists the user in entering the free text information (“Message: Select”). A default follow-up time can be entered and the PDA calculates the due date on the follow-up. Checking off the follow-up is similar to that of the lab result cheek-off. Follow-ups can be deleted manually by the “Delete” button. Screen 5. Patient demographics (name, DOB, phone number) and individual patient lab profiles can be managed on this screen. Labs can be added quickly to the particular patient’s profile by the “Add Lab” button.
“Management of Laboratory Test Results In Family Practice” Mold, et al. J Fam Pract. 2000 Aug;49(8):709-15.
Management of Prescription Refills
Steps/Components Patient access to the system Clinical decision-making Notification of pharmacy Notification of patient Documentation
Values Patient access
Patient satisfaction Efficiency
Clinical decision-making Accuracy Efficiency
Values Notification of pharmacy
Accuracy Efficiency
Notification of patient Patient satisfaction Efficiency
Documentation Accuracy Efficiency
Practice Audits
3 4 5
A X X
B X X X
C X X
D X X
E X X X X X
Prescription Refills – Patient Access Patient calls pharmacy to request
refills 75% adherence rate Mean time required by patient: 2.5
minutes Satisfaction high
Prescription Refills – Decision-making Accuracy reduced by using written
protocol 4-10% error rate vs. 0-2% Too many special
circumstances/variability M.A. or L.P.N. more efficient and as
accurate as physician
Prescription Refills – Communication with Pharmacy/Patients
Pharmacy faxes prescription to office for authorization
Authorization faxed back to pharmacy Multiple methods less efficient (hot
line/phone/fax) Pharmacy communicates with patients Denials/reasons/instructions written on fax to
pharmacy If likely to be controversial, patient phoned by
nurse Mean patient satisfaction: 3.5-4.1/5
Prescription Refills – Documentation EMR increases efficiency
substantially $400-500 cost saving per MD per year
Error rates 0-3% no clear advantage for EMR Transfer of information (double entry)
should be avoided if possible
Maximizing Quality of Care for Diabetic Patients
Diabetes Quality Improvement Project Indicators
A1c Q 1 year Lipid panel Q 1 year UA for protein Q 1 year Eye exam Q 1 year Foot exam Q 1 year Flu shot Q 1 year Pneumococcal vaccine ever
Diabetes Care Oklahoma Foundation for Medical Quality
audits Exemplars (90% adherence) identified for
two or more items 5 exemplars covered all items with overlap
Phone interviews – transcripts Analysis of transcripts identified 6
principles
Diabetes Care See all diabetic patients every 3 months
for diabetes care Label diabetic charts with sticker Protocol for office staff Registry of diabetic patients Work with one or two eye doctors who are
faithful about sending reports Chart documentation sheet/flow sheet
PDAs and PEAs We developed a PDA-based
reminder/registry/flow sheet generator
Practice Enhancement Assistants work with 8 practices over an extended period of time to help them implement practice improvements
Diabetes Patient Tracker Enterprise MS Access and SQL Database
Versions
3
Diabetes Patient Tracker Enterprise is one of OKPRN’s most utilized PDA solution at this point. With it’s quick menu options and colorful icon-coded multiple object screens, this application is even more flexible, effective and user friendly. Individual patient report function and printable flowcharts along with enhanced electronic chart audit function can provide the ultimate solution for diabetes patient tracking, electronic documentation and clinical decision support system. Syncs up to a central MS Access, or an SQL database
Diabetes Care QI Studies Initial study of best practices plus practice
enhancement assistant completed 1/2003 30 clinicians in 10 practices I explained/we discussed the 6 principles
High rate of acceptance of six principles Wide acceptance of a PDA-based diabetic
registry Dramatic improvements in adherence to
guidelines
Quality of Care Indicators A1c: 87% - 96% p=0.0003 UA protein: 53% - 64% p=0.05 Lipid Panel: 69% - 80% p=0.02 Foot Exam: 71% - 82% p=0.004 Retinal Exam: 48% - 59% p=0.04 Pneumovax: 42% - 61% p=0.0006 ACEI for BP: 72% - 86% p=0.03 ACEI for prot: 53% - 64% p=0.05
Next Steps AHRQ grant to study feedback vs. FB
+ best practices vs. FP+BP+ practice enhancement assistant
Focus on the most effective ways to reduce BP, LDL, and A1c
Colorectal Cancer Screening
Steps: At risk patients identified Screening offered Screening completed Patients who screen positive receive
needed follow-up/further testing
NCI – Colorectal Cancer – Critiques “The ‘best practices’ approach has
enormous potential.” “To ‘systematically tap into the
wisdom of practicing physicians’ seems to have a lot more to recommend it than the usual top-down, theory driven efforts to improve practice.”
More Questions What is the best way to manage no-shows
and late cancellations What is the best way to deal with
pharmaceutical representatives What is the best way to manage drug
samples What is the best way to handle patient
phone calls What is the best way to deal with
residents who are performing poorly
Management of No-Shows Student summer project Survey of all FP residency program clinic
managers In depth interviews with exemplars• Steps
• Reduction in umber/% of no-shows• Management of no-shows when they occur
(e.g. number of patients seen/half-day)
What are your challenges?
Would the methods described be a potentially useful way to address them?
What are the implications of this approach for practice-based research networks?
How could such an approach be used to improve primary care practice nationally?
OBJECTIVES• Teach you about “best practices
research” as a concept
• Teach you the method
• Give you enough examples that you feel that you could do it
• Share some of our discoveries
Questions/Reference
Mold JW and Gregory ME Best practices research. Family Medicine 2003, 35 (2): 131-134