Home > DARE Reviews > Evaluating the impact of study-level...

PubMed Health. A service of the National Library of Medicine, National Institutes of Health.

Database of Abstracts of Reviews of Effects (DARE): Quality-assessed Reviews [Internet]. York (UK): Centre for Reviews and Dissemination (UK); 1995-.

Database of Abstracts of Reviews of Effects (DARE): Quality-assessed Reviews [Internet].

Evaluating the impact of study-level factors on warfarin control in US-based primary studies: a meta-analysis

DA Cios, WL Baker, SD Sander, OJ Phung, and CI Coleman.

Review published: 2009.

Link to full article: [Journal publisher]

CRD summary

The authors concluded that patients spent 57% of their time within the therapeutic range while on warfarin treatment. This conclusion was based on heterogeneous studies of unknown quality that may contain publication bias and so cannot be considered reliable.

Authors' objectives

To evaluate the effect of study level factors on patients treated with warfarin in USA.


Studies were included from a previous systematic review (see Other Publications of Related Interest) and MEDLINE, EMBASE, CINAHL, Web of Science and The Cochrane Library were searched from January 2005 to February 2008. Search terms were reported. Bibliographies of retrieved articles were handsearched for additional material. Only studies published in English were considered.

Study selection

North American studies that enrolled 25 or more patients, that contained at least one warfarin dosing group and that monitored International Normalized Ratio (INR) values for at least three weeks were eligible for inclusion in the review. Eligible studies used a patient-time approach to obtain serial INRs for each patient, an interpolation method to estimate anticoagulation status and reported data on the proportion of time patients who received warfarin had INR values in traditional therapeutic ranges. Studies were excluded if they measured serial INRs after administration of vitamin K.

The primary outcome measured was proportion of time within INR therapeutic range. Included studies were retrospective or prospective cohorts and ranged in size from 14 to 7,445 participants. Equal numbers of studies were set in anticoagulation clinics or within community practice. Studies used predominantly linear interpolation. Follow-up ranged from 1.4 to 24,179 patient years (median 222). Time spent within INR therapeutic range was 0.6 to 13,177.6 patient years (median 124.3).

Two reviewers independently selected studies for inclusion in the review.

Assessment of study quality

The authors did not state that they assessed validity.

Data extraction

Incidence densities (proportion of time that the group spent within INR range multiplied by observation time, divided by the total observation time) and 95% confidence intervals (95% CI) were calculated. If studies reported INR control rates for the same patient at different time periods, only the longest duration was included.

Three categories of interpolation methods were used: linear (INR values assumed to change linearly); halving (the first half of the time between successive INRs was assigned the previous value and the second half of the time was assigned the subsequent INR value); and other (not halving or linear).

Data was independently extracted by two reviewers. Discrepancies were resolved by a third reviewer.

Methods of synthesis

A random-effects meta-analysis was used to synthesize weighted-averaged time in therapeutic INR range and 95% CIs.

Subgroup and meta-regression analyses (random-effect and fixed-effect) were conducted to determine the influence of study setting, year of publication, interpolation method, study design and presence of self management. A post hoc analysis was performed on Canadian studies to improve the statistical power of the analysis; it was assumed these studies had similar practices and controls.

Heterogeneity was assessed using the I2 statistic. Publication bias was assessed using funnel plots and Egger's weighted-regression statistics.

Results of the review

Twenty-four studies were included in the review (n=26,979).

Overall, patients spent 57% of their time within the therapeutic range (95% CI 55% to 59%) while on warfarin treatment. Heterogeneity was found between studies. Publication bias was not detected.

Metaregression found that patients who received treatment within community care spent significantly less time within the therapeutic range (-13%, 95% CI -18.1% to -7.9%) than patients who attended anticoagulation clinics. Patients in prospective studies spent more time in the therapeutic range (7.3%, 95% CI 1.5% to 13.1%) than patients in retrospective studies.

Subgroup analysis indicated that interpolation method, study year and self management were not found to significantly effect time in the therapeutic range.

When studies from Canada (34 studies including six RCTs) were included in the analysis, the results were similar to the base-case analysis. Patients in older studies spent less time within the therapeutic range (-8%, 95% CI -14.1% to -2%) than newer studies. Studies that used linear interpolation methods resulted in more time in the therapeutic range (7.2%, 95% CI 1.2% to 13.2%).

Authors' conclusions

Patients maintained INR values within the therapeutic range 57% of the time. Use of anticoagulation clinics was superior to use of community care. Patients treated with anticoagulation clinics had INR values within the therapeutic range less than two thirds of the time.

CRD commentary

This review addressed a clear research question supported by clear inclusion criteria. The authors searched several relevant databases. Only English-language publications were included, which introduced a risk of language bias. There was no apparent search for unpublished material, so relevant trials may have been missed and publication bias could not be ruled out. The included studies were not assessed for quality or validity; therefore, it was difficult to know how much to weight the studies. Adequate steps were taken throughout the review process to minimise errors and bias. Data were analysed using appropriate statistical techniques, but clinical heterogeneity was not investigated (warfarin dose, patient age, duration of therapy before study). Rationales for subgroup analyses were clear, but subgroup analyses was limited by the number of eligible studies considered. The authors conclusions reflected the results, but were based on heterogeneous studies of unknown quality. Therefore, we cannot consider the conclusions reliable.

Implications of the review for practice and research

Practice: The authors stated that anticoagulation services provided superior INR control to standard community care.

Research: The authors stated that further research was required to improve the length of time that patients received warfarin within the therapeutic limits..


Not stated.

Bibliographic details

Cios DA, Baker WL, Sander SD, Phung OJ, Coleman CI. Evaluating the impact of study-level factors on warfarin control in US-based primary studies: a meta-analysis. American Journal of Health-System Pharmacy 2009; 66(10): 916-925. [PubMed: 19420310]

Other publications of related interest

Van Walraven C, Jennings A, Oake N et al. Effect of study setting on anticoagulation control: a systematic review and metaregression. Chest 2006; 129:1155-1166.

Indexing Status

Subject indexing assigned by NLM


Ambulatory Care Facilities; Anticoagulants /administration & dosage /adverse effects /therapeutic use; Community Health Services; Humans; International Normalized Ratio; Prospective Studies; Randomized Controlled Trials as Topic /methods; Retrospective Studies; Treatment Outcome; United States; Warfarin /administration & dosage /adverse effects /therapeutic use



Database entry date


Record Status

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.

CRD has determined that this article meets the DARE scientific quality criteria for a systematic review.

Copyright © 2014 University of York.

PMID: 19420310