How Can We Get
the Best Medication History?
Stephen Shalansky, Pharm.D., FCSHP Pharmacy Department, St. Paul’s Hospital
Faculty of Pharmaceutical Sciences, UBC
How Are We Doing Now?
Completeness of Medication Histories Documented by Various Health Care Professionals
Before-Admission Medications
Health Care Number of Number Proportion Number NOT Professional Med Histories of Meds Documented Documented
Reviewed
___________________________________________________________________
Physicians
Staff 29 5.5 ± 3.6 70 ± 24 1.8 ± 2.2
Residents 22 5.6 ± 2.8 69 ± 29 2.1 ± 2.0
Nurses 63 5.0 ± 3.1 74 ± 31 1.6 ± 2.2
Paramedics 21 5.0 ± 2.6 63 ± 27 2.2 ± 1.9
Pharmacists
With PharmaNet 18 7.0 ± 2.8 76 ± 21 1.6 ± 1.4
Without PharmNet 16 5.5 ± 3.0 77 ± 30 1.4 ± 1.6 ___________________________________________________________________
Lo, Shalansky, Menezes. Can J Hosp Pharm 2004;57:32-38
Class 1: unlikely to cause patient discomfort or clinical deterioration
Class 2: potential to to cause moderate patient discomfort or clinical deterioration Class 3: potential to to cause severe patient discomfort or clinical deterioration
39% of discrepancies had the potential to cause moderate to severe discomfort or clinical deterioration
54% of patients had an unintended discrepancy
No significant predictors of unintended discrepancies were identified
Is PharmaNet the Answer?
Canada Health Infoway Inc. is leading a
national program towards the implementation of :
–Electronic health records
–Electronic drug profile and prescribing systems
The Accuracy of a
Prescription Claims Database
for Determining Medication Profiles of Heart Failure Patients
Stephen Shalansky, Janice Yim, Linda Jang, Andrew Ignaszewski, Leon Jung, Catherine Clark
Pharmacy Department and Division of Cardiology, St. Paul’s Hospital UBC Faculties of Medicine and Pharmaceutical Sciences
Burrard Pharmacy, Vancouver
UNIVERSITY OF BRITISH COLUMBIA
PharmaNet
• Records all prescription transactions
processed by community Pharmacies in British Columbia
• Not limited to specific health plans
• Excluded:
– HIV meds
– meds administered in hospital
• Can be accessed by:
• community pharmacists
• Hospital pharmacist
• Emergency room physicians
• Ministry and researchers can apply for data
PharmaNet Data
Hospital access:
– Generic drug name and strength – Dosage form
– Manufacturer
– Date and quantity dispensed – Dose and directions
– Prescribing physicians
Also available:
– Cost/billing, drug interactions, allergies, medical conditions, clinical notes, pharmacy
demographics, physician demographics, access
info., etc.
Use of PharmaNet Data
Clinical, research, policy applications:
– Current medication profiles/medication history – Adherence
– Therapeutic duplication – Drug interactions
– Major source of medication use data for epidemiology research
– Ministry-initiated drug use evaluations
Objectives
To quantify agreement between the PharmaNet database and:
a) A thorough medication history interview
– Current medication profile:
• the number of outpatient prescription medications currently consumed (Primary)
• the types and doses of outpatient prescription medications currently consumed (Secondary)
b) Medication Event Monitoring System (MEMS)
– medication adherence over a 6 month period (Pilot study).
Results
Demographics of 194 Patients Interviewed
Male 149 (77%)
Age, median (range) 65 (27-94) Recruitment Site
Heart Function Clinic 162 (83%) Pre-Heart Transplant Clinic 19 (10%) Community Pharmacy 13 (7%) CHF class
Class I 41 (21%)
Class II 66 (34%)
Class III 19 (10%)
Class IV 1 (1%)
Unknown 67 (34%)
Prescription meds (mean ± SD) 7.5 ± 2.9
Total number meds reviewed 1457
Med Profile Discrepancies
Between PharmaNet and Interview
194 Medication Profiles; 1457 meds
All discrepancies
# of patients with a discrepancy 138 (71.1%)
# of discrepancies 353
Discrepancies in # of active meds
# of patients 109 (56.2%)
# of discrepancies 230
Discrepancies in type of meds
# of patients 115 (59.3%)
# of discrepancies 268
Discrepancies in doses of meds
a# of patients 64 (33.0%)
# of discrepancies 85
a for medications listed in both PNet and Interview
Bland-Altman Agreement Plot
Number of Medications Currently Consumed:
PharmaNet versus Interview
Difference:
PNet - Interview
Number of Prescription Meds Currently Consumed
Mean
difference:
-1.0 meds per patient
Limits of agreement:
-4.7 to 2.8
n = 194
Medication Class Involved in 353 Discrepancies
Diuretics 76 (21.6%)
Beta-blocker 37 (10.5%)
ACE inhibitor 29 (8.2%)
Supplements 26 (7.4%)
Oral anti-diabetics 23 (6.5%)
Other heart medications 48 (13.6%)
Others 114 (32.3%)
Reasons for Discrepancies
Number and Type of Medications
n = 1457 meds
Med listed on PNet but incorrectly appeared inactive 205 (14.1%)
Overdue for refill by < 7 days 72 (4.9%) Overdue for refill by >7 but < 30 days 69 (4.7%) Overdue for refill by >30 days 64 (4.4%) Med listed on PNet but incorrectly appeared active 40 (2.7%) Med did not appear anywhere on PNet 23 (1.6%)
Note: dosing discrepancies excluded: all medications involved in dosing discrepancies appeared on PharmaNet
Reasons for Discrepancies
Number and Type of Medications
Reason for Discrepancy N = 268 Discrepancies
Overdue for refill; pt stated will refill soon 40 (15.0%)
Dosing regimen recently changed 39 (14.6)
Unknown 37 (13.9%)
Medication stopped 31 (11.6%)
…reason unknown 16
…because of SE’s 7
…because of med switching 4
…because symptoms resolved 2
…because of other reasons 2
Supply left from previous refill 30 (11.2%)
Hospital provided medication during hospital stay 21 (7.9%) Medication recently restarted / old Rx 18 (6.7%)
Patient stated non-compliance 14 (5.2%)
Sample medication 11 (4.1%)
Taken only when needed 10 (3.7%)
Other known reasons 17 (6.3%)
Reasons for Discrepancies
Dosing
N = 85 Discrepancies
Side Effects 11 (12.9%)
Medication not working 9 (10.6%) Reason unknown to patient 8 (9.4%) Other known reasons 33 (38.8%)
Not recorded 24 (28.2%)
Adherence: PharmaNet versus MEMS Bland-Altman Agreement Plot
40.00 60.00 80.00 100.00 120.00
-25.00 0.00 25.00 50.00 75.00 100.00 125.00
% Adherence (average of PNet and MEMS) Discrepancy
(PNet – MEMS)
Mean discrepancy 6.8% +/- 23.7%
Limits of agreement:
-40.6% to 54.2%
N = 45
Adherence: PharmaNet versus MEMS Bland-Altman Agreement Plot
40.00 60.00 80.00 100.00 120.00
-25.00 0.00 25.00 50.00 75.00 100.00 125.00
% Adherence (average of PNet and MEMS) Discrepancy
(PNet – MEMS)
MEMS misadventures in red (N=5)
Mean discrepancy 6.8% +/- 23.7%
Limits of agreement:
-40.6% to 54.2%
N = 45
Adherence: PharmaNet versus MEMS Bland-Altman Agreement Plot
40.00 60.00 80.00 100.00 120.00
-25.00 0.00 25.00 50.00 75.00 100.00 125.00
% Adherence (average of PNet and MEMS) Discrepancy
(PNet – MEMS)
Mean discrepancy 0.7% +/- 11.9%
Limits of agreement:
-23.1% to 25.5 N = 40
Conclusions
• 70% of PharmaNet profiles contained some inaccurate or misleading information
regarding current medication consumption
• Most medications currently consumed appeared somewhere on the PharmaNet profile
• It was often difficult to determine which were still active
• At least some doses were inaccurate for a third of patients interviewed
Conclusions
• Adherence calculated using refill data from PharmaNet may inaccurate:
• Limits of agreement with MEMS: -23.1% to 25.5
• 5/45 (11%) of patients did not use MEMS appropriately
• Gold Standard for measuring adherence?
A Couple More Points on Adherence
An accurate list of what was prescribed or filled is not necessarily an accurate list of what is being consumed Initiating medications that the patient was not previously
taking can be dangerous
Predictors of Nonadherence
Depression
Cognitive impairment Asymptomatic disease
Inadequate discharge planning/follow-up Side effects
Lack of belief in benefit Lack of insight in illness
Poor provider-patient relationship Barriers to care
Missed appointments Complexity of treatment Cost
Osterberg L. N Engl J Med 2005;353:487-97
Predictors of Nonadherence
Non-Adherent Adherent P value
(N=49) (N=328)
Age 62 ± 10 61 ± 11 0.34
Male 41 (84%) 239 (73%) 0.12
Number of Rx meds 4.0 ± 2.8 6.0 ± 3.4 <0.001
Adverse effects 18 (37%) 151 (46%) 0.22
Use of adherence aids 18 (37%) 180 (55%) 0.02 Alternative medications 49 (20%) 85 (26%) 0.38
Attends clinic 22 (45%) 144 (44%) 0.93
Shalansky SJ, et al. Ann Pharmacother 2002;36:1532-9
Predictors of Nonadherence
Non-Adherent Adherent P value
(N=49) (N=328)
Age 62 ± 10 61 ± 11 0.34
Male 41 (84%) 239 (73%) 0.12
Number of Rx meds 4.0 ± 2.8 6.0 ± 3.4 <0.001
Adverse effects 18 (37%) 151 (46%) 0.22
Use of adherence aids 18 (37%) 180 (55%) 0.02 Alternative medications 49 (20%) 85 (26%) 0.38
Attends clinic 22 (45%) 144 (44%) 0.93
Shalansky SJ, et al. Ann Pharmacother 2002;36:1532-9
More Meds = Better Adherence?
N = 367
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
1 to 3 4 to 6 7 to 9 10 or more
Number of Regularly Scheduled Prescription Medications
Adjusted Odd Ratio
0%
2%
4%
6%
8%
10%
12%
14%
16%
18%
20%
Proportion Non-Adherent
Adjusted OR for non-adherence Proportion who were non-adherent (unadjusted) N = 111
N = 137
N = 70
N = 49
Shalansky SJ, et al. Ann Pharmacother 2002;36:1532-9
“Have you changed your daily routine to accommodate your medication schedule?”
Morisky Score
Non-Adherent Adherent P value (N=49) (N=328)
Do you ever forget to
take your medicines? 29 (59%) 138 (42%) 0.03 Are you careless at time
about taking your medicine? 13 (27%) 72 (22%) 0.47 When you feel better,
do you sometimes stop
taking your medicine? 6 (12%) 7 (2%) <0.001 Sometimes if you feel worse
when you take your medicine,
do you stop taking it? 12 (25%) 27 (8%) <0.001
Shalansky SJ, et al. Ann Pharmacother 2002;36:1532-9
Morisky Score
Non-Adherent Adherent P value (N=49) (N=328)
Do you ever forget to
take your medicines? 29 (59%) 138 (42%) 0.03 Are you careless at time
about taking your medicine? 13 (27%) 72 (22%) 0.47 When you feel better,
do you sometimes stop
taking your medicine? 6 (12%) 7 (2%) <0.001
Sometimes if you feel worse when you take your medicine,
do you stop taking it? 12 (25%) 27 (8%) <0.001
Shalansky SJ, et al. Ann Pharmacother 2002;36:1532-9
There’s still a place for medication
history interviews!
Recommendations
Questions for Medication History Interviews
Ask about all medications:
– Prescription
– Over-the-counter
– Anything from a herbalist or health store?
– Herbs
– Vitamins or supplements – Teas
– Traditions remedies (from other countries?)
Name, dosage form, dose, schedule, last dose taken (be specific about prn medications)
Have you recently started any medications, or had the dose changed?
– Side effects/non-adherence more likely
http://www.saferhealthcarenow.ca/en/pdf/GSK%20Reconciling%20Medications%20FINAL%20rev6Nov2005.pdf
Recommendations
Questions for Medication History Interviews
• Balance open-ended questions with yes/no questions
• Nonbiased questions
• No leading questions
• Vague responses may indicate non-adherence
• Avoid medical jargon
http://www.saferhealthcarenow.ca/en/pdf/GSK%20Reconciling%20Medications%20FINAL%20rev6Nov2005.pdf
Recommendations
Questions for Medication History Interviews
• Encourage questions from patient
• Encourage bringing meds and use of medication wallet card
– Give out wallet cards
• Prompt regarding non-pill dosage forms and prns
– Creams, drops, inhalers, spray, samples
• Allergies: ask about symptoms
• Use multiple sources of information:
– Medication labels – Family
– Community pharmacy – Family physician
http://www.saferhealthcarenow.ca/en/pdf/GSK%20Reconciling%20Medications%20FINAL%20rev6Nov2005.pdf
Recommendations
Questions for Medication History Interviews
If you have access to prescription refill records:
“Did the doctor change the dose or stop any of your medications recently?”
– Many discrepancies identified in our study resulted from physician-initiated changes or discontinuation
“Have you changed the dose or stopped any of your medications recently?”
– Many of the identified discrepancies were the result of patient-initiated deviations from the written prescription
Recommendations
Questions for Medication History Interviews
If you have access to prescription refill records:
“Do you have trouble taking or remembering to take any of your medications?”
– A substantial number of patients admitted non-adherence with little prompting
“Have any of the medications been causing side effects?”
– Experiencing side effects is one of the common reasons for stopping a medication or changing the dosing of a medication
If you’re pressed for time, focus on:
•Medications with a relatively high incidence of side effects:
•e.g. beta blockers
•Medications that are doses based on symptoms:
•e.g. diuretics
Recommendations
Questions for Medication History Interviews
If you have access to prescription refill records:
Your PharmaNet profile indicates that you may have run out of some medications. Are you still taking any of these?
– Overdue for refills were the most common reason identified for discrepancies
Have you spent any days in the hospital for the past year?
– Hospital providing medications during admission accounted for 8% of the discrepancies in this cohort
Recommendations
Questions for Medication History Interviews
Adherence:
Review the prescription refill records if possible.
Morisky Questions:
“When you feel better, do you sometimes stop taking your medicine?”
“Sometimes if you feel worse when you take your medicine, do you stop taking it?”
Health Beliefs:
“Have you changed your daily routine to accommodate your medication schedule?”
Don’t assume patients taking few medications are adherent, or patients taking a large number of medications are non-adherent
Ask the Right Questions!
How much time do you spend using your exercise treadmill each week?