Table 4.

Warfarin management and monitoring

Study	Study objective, (intervention/exposure and outcomes)	Study design, data source	Study population, study setting, time period	Results	Quality assessment, funding Source
Part A: Quality of oral anticoagulant prescribing and monitoring
Aspinall et al. (2010) [15,30]	To describe the quality of warfarin prescribing and monitoring in VA nursing homes and to assess factors associated with maintaining a therapeutic INR	Design: retrospective cohort study Data source: medical record review	Population: all veterans (160) who received warfarin Setting: 5 VA nursing homes (US) Time period: Jan 1, 2008 - June 30, 2008	INRs were in therapeutic range for 55% of the 10,380 total person-days of warfarin. In a 4-week period, patients had an average of 5.2 (SD = 2.7) INRs obtained. 99% of the INR tests were repeated within 4 weeks of the previous result. 49% of patients had INRs in the target range for ≥ 50% of their person-days. Achieving this outcome was more likely in patients with prevalent warfarin use than with new use (adjusted OR = 2.86; 95% CI = 1.06-7.72). Patients with a history of a stroke (adjusted OR = 50.38; 95% CI = 50.18-0.80) were less likely to have therapeutic INRs for > 50% of their days. Approximately 89% of the patients at baseline were receiving ≥ 1 medication that potentially interacts with warfarin. The most frequently prescribed interacting drugs at baseline were omeprazole (51% of patients), simvastatin (45%), aspirin (34%), citalopram (18%), and levothyroxine (13%). During the study period, 46% of patients were prescribed a medication with the potential to interact with warfarin	Quality assessment for observational studies: 1) Unbiased selection of the cohort? Yes 2) Selection minimizes baseline differences in prognostic factors? Yes 3) Sample size calculated/5% difference? No 4) Adequate description of the cohort? Yes 5) Validated method for ascertaining exposure? Yes 6) Validated method for ascertaining clinical outcomes? Yes 7) Outcome assessment blind to exposure? Yes 8) Adequate follow-up period? Partial; length of follow-up not clearly stated in Methodology 9) Completeness of follow-up? Yes 10) Analysis controls for confounding? Yes 11) Analytic methods appropriate? Yes Funding: various VA centers and US government agencies
Gurwitz et al. (2007) [25] (repeated from table 1)				The percentages of time in the < 2, 2-3, and > 3 INR ranges were 36.5%, 49.6%, and 13.9%, respectively
Gurwitz et al. (1997) [8] (repeated from table 1)				Of 122 warfarin users with adequate INR data, warfarin therapy was monitored at least every 2 weeks in 52% of the subjects, every 2-4 weeks in 32% of the subjects, and less frequently than every 4 weeks in only 16% of the subjects. On average, 117 NVAF residents with available INR data were maintained in the therapeutic range 39.6% of the time, in the subtherapeutic range 44.8% of the time, and in the supratherapeutic range 15.6% of the time; < 23 subjects (20%) were in the therapeutic range ≥ 60% of the time
Karki et al. (2003) [31]	To evaluate the warfarin management patterns in an academic nursing home and evaluate what pre-determined factors are associated with variability in the INR	Design: case control study Data source: medical chart review	Population: 37 residents receiving warfarin therapy for > 3 consecutive months in a calendar year Setting: 566-bed academic medical center nursing home (US) Time period: 12-month period; dates not specified	For patients who had INR values exceeding the therapeutic range there was no significant difference between "easy" management (INR fluctuations of 0.5-0.99 and outside therapeutic range ≤ 10% of time, n = 18) and "difficult" management (with INR fluctuations > 0.99 and outside therapeutic range > 10% of time, n = 19) in all factors examined. The "difficult management" group received more medications known to interact with warfarin than the "easy" management. These medications may have caused the INR to increase above the normal range (P=0.003), as well as produced large (P=0.001) or small fluctuations (P=0.0007) in the INR. 54% of residents on warfarin therapy initiated a potential warfarin-interacting drug. Of all interacting medications, 55% were antibiotics and 28% were analgesics	Quality assessment for observational studies: 1) Unbiased selection of the cohort? Partial; limited to residents of a single nursing home 2) Selection minimizes baseline differences in prognostic factors? Yes 3) Sample size calculated/5% difference? No; very low power 4) Adequate description of the cohort? No; only limited description of nursing home and resident characteristics reported 5) Validated method for ascertaining exposure? No; assignment of subjects to "easy management" and "difficult management" cohorts not validated 6) Validated method for ascertaining clinical outcomes? Yes 7) Outcome assessment blind to exposure? Yes 8) Adequate follow-up period? Yes 9) Completeness of follow-up? Yes 10) Analysis controls for confounding? Partial; evaluated association of age, gender, and number of illnesses with cohort 11) Analytic methods appropriate? Partial; only limited univariate analyses conducted Funding: not stated
Lackner et al. (1995)[9] (repeated from table 1)				The INR was within the recommended range for NVAF over a 6-month period 37% of the time and recommended PT, 52% of the time. An equal percentage of warfarin dose changes occurred in response to a PT ratio outside the recommended range as occurred with an INR outside the recommended range
McCormick et al. (2001) [11] (repeated from table 1)				In the 42% of AF patients who were receiving warfarin therapy, the therapeutic range of INR values was maintained only 51% of the time, was below the therapeutic range 36% of the time, and was above the therapeutic range 13% of the time
Verhovsek et al. (2008) [22]	To determine how effectively warfarin was administered to a cohort of residents in LTC facilities by measuring TTR, to identify the proportion of residents prescribed warfarin-interacting drugs and to ascertain factors associated with poor INR control	Design: Retrospective cohort study Data source: medical chart review	Population: 105 LTC residents receiving warfarin therapy Setting: 5 LTC facilities in Hamilton, Ontario (Canada) Time period: 12 months of data for each resident between October 2004 and April 2005	3065 INR values were available. Residents were within, below, and above the therapeutic range 54%, 35% and 11% of the time, respectively. 79% of residents were prescribed ≥ 1 warfarin-interacting medication during the period in review. The 5 most common drugs were acetaminophen (40% of residents), citalopram (25%), acetylsalicylic acid (16%), diltiazem (11%), and simvastatin (10%). Residents receiving interacting medications spent less TTR (53.0% vs 58.2%, OR = 0.93; 95% CI = 0.88-0.97, P=0.002). Adequacy of anticoagulation varied significantly between physicians (TTR range 45.9-63.9%)	Quality assessment for observational studies: 1) Unbiased selection of the cohort? Yes 2) Selection minimizes baseline differences in prognostic factors? Yes 3) Sample size calculated/5% difference? No; low power for both residents and physicians studied 4) Adequate description of the cohort? Yes 5) Validated method for ascertaining exposure? Yes 6) Validated method for ascertaining clinical outcomes? Yes 7) Outcome assessment blind to exposure? Yes 8) Adequate follow-up period? Partial; some residents may have had < 6 month follow-up of INR 9) Completeness of follow-up? Yes 10) Analysis controls for confounding? No; association of INR outcomes with subject attributes not analyzed 11) Analytic methods appropriate? No; analysis of residents limited to simple one-way tabulations Funding: CIHR and the Regional Medical Associates
Part B: Medication management interventions
Allen et al. (2000) [23]	To evaluate the effectiveness of nurse practitioner management of anticoagulation using a protocol. Outcomes were frequency of blood draws as well as frequency and percentage of INRs that were out of range	Design: retrospective cohort study Data source: nurse practitioner maintained records	Population: 47 patients on long-term anticoagulation therapy Setting: 9 area nursing homes (US) Time period: 6 months beginning June 1997	Average number of venipuncture ranged from 0.7 -2.7 per month. Reasons for out-of-range INRs were identified 35% of the time. Percentage out of range was 15%	Quality assessment for observational studies: 1) Unbiased selection of the cohort? Cannot be determined; convenience sample of 47 residents 2) Selection minimizes baseline differences in prognostic factors? Yes 3) Sample size calculated/5% difference? No 4) Adequate description of the cohort? No; resident characteristics not described 5) Validated method for ascertaining exposure? Yes 6) Validated method for ascertaining clinical outcomes? Yes 7) Outcome assessment blind to exposure? No 8) Adequate follow-up period? Cannot be determined; post-intervention follow-up period not described 9) Completeness of follow-up? Cannot be determined 10) Analysis controls for confounding? No; no evaluation of association of resident attributes with INR outcomes 11) Analytic methods appropriate? No; Statistical error not reported Funding: not specified
Papaioannou et al. (2010) [24] (repeated from Table 1)				Overall, TTR increased during the MEDeINR phase (65-69%), but was significantly increased for only 1 facility (62-71%, P< 0.05). The percentage of time in supratherapeutic range decreased from 14% to 11%, P=0.08); there was little change for the subtherapeutic range (21% to 20%, P=0.66). Overall, the average number of INR tests per 30 days decreased from 4.2 to 3.1 (P< 0.0001) per resident post-implementation. Feedback from LTC clinicians and staff indicated that the program decreased workload, improved confidence in management and decisions, and was generally easy to use

CI, confidence interval; CIHR, Canadian Institute of Health Research; INR, international normalized ratio; LTC, long-term care; NVAF, nonvalvular atrial fibrillation; OR, odds ratio; PT, prothrombin time; SD, standard deviation; TTR, time in therapeutic range; VA, Veteran's Administration