BACKGROUND/RATIONALE:
Atrial fibrillation (AF) is a major cause of preventable stroke; anticoagulation with warfarin decreases stroke risk, but warfarin's narrow therapeutic window necessitates careful monitoring of labs (INR). Clinicians may have concerns that patients with mental health conditions (MHC) will not be able to take warfarin safely. It is not known how often patients with MHC receive anticoagulation therapy, nor how they fare on it.

OBJECTIVE(S):
Primary objectives were to characterize warfarin eligibility and to determine whether processes of care (receipt of warfarin, INR lab monitoring, Time in Therapeutic Range of INR [TTR]) and outcomes (strokes, hemorrhages, death) differ for patients with and without MHC.

METHODS:
For VA patients with AF (based on VA National Patient Care Database [NPCD] and Medicare administrative data), in FY04 we identified processes of care from Decision Support System, supplemented with Medicare data. We identified outcomes from VA's Vital Status file and from NPCD and Medicare data. Patients had one of 15 MHCs (derived from AHRQ's Clinical Classifications Software) if they had an ICD9 code for MHC prior to FY04 and a confirmatory ICD9 code during the study period.

FINDINGS/RESULTS:
In our main cohort, 22,261 had MHC and 103,442 had no MHC. Key findings were as follows:

(1) Most AF patients with MHC and without MHC (97% of each group) were warfarin-eligible based upon a CHADS2 stroke risk score of 1 or more. More of those with MHC than without MHC had high stroke risk: 83% versus 82% had a CHADS2 score of 2 or more, and 62% versus 57% had a score of 3 or more. However, risk factors for hemorrhage were also more common in patients with MHC: for example, 1.2% versus 0.5% had a history of intracranial hemorrhage, 27% versus 20% had a history of other hemorrhage, 17% versus 3% had a history of dementia, and 2% versus 1% had a history of cirrhosis.

(2) In unadjusted analyses, patients with MHC were less likely than those with no MHC to receive warfarin (48% versus 52%, respectively). This pattern remained even after adjusting for age, sex, race/ethnicity, CHADS2 stroke risk score, and bleeding risk: AOR 0.80. Among AF patients treated with warfarin, those with MHC had more regular INR monitoring (AOR 1.24 for monthly monitoring in patients with versus without MHC, adjusting for age, sex, race/ethnicity and physical comorbidity index). However, those with MHC spent less time in the therapeutic range of INR: median TTR was 0.61 versus 0.68, respectively. Furthermore, fewer patients with MHC (54%) than without MHC (62%) had a TTR above the desired threshold of 0.58; this finding persisted in adjusted analyses (AOR 0.78 after adjusting for sociodemographics and physical comorbidity).

(3) Among both warfarin-treated patients and among patients not treated with warfarin, those with MHC had worse outcomes (strokes, major hemorrhages, death) than did those without MHC. These differences persisted after adjustment for sociodemographic characteristics and comorbidity. Among those treated with warfarin, 20% of those with MHC versus 16% of those without MHC had at least one adverse outcome (adjusted OR 1.27). Among those who did not receive warfarin, adjusted OR was 1.46 for the composite outcome.

(4) The association between MHC and processes/outcomes of AF care varied by specific MHC type. Outcomes tended to be worst in those with alcohol use disorders or psychosis. For example, among warfarin-treated patients, for the composite outcome (stroke, hemorrhage or death), adjusted OR was depression 1.24, PTSD 1.32, other anxiety 1.07 (NS), psychotic disorders 1.51 and alcohol use disorders 1.40, respectively.

IMPACT:
Clinicians should take MHC into account when developing a treatment plan for patients with MHC, since MHC may confer excess risk for adverse outcomes in anticoagulated patients. However, the risks versus benefits in an individual patient need to be carefully considered, and presence of MHC should not be considered an absolute contraindication to anticoagulation. The subset of warfarin-treated MHC patients who have alcohol use disorders or psychotic disorders may represent a group at particularly high risk of out-of-range INR values and adverse outcomes; if clinicians decide to treat them with warfarin, it is possible that adjunct measures, such as AF case management, may enhance safety of anticoagulation care. This study is part of an emerging body of evidence suggesting that processes and outcomes of medical care can be worse in patients with mental illness.

---

External Links for this Project

---

PUBLICATIONS:

Journal Articles

1. Walker GA, Heidenreich PA, Phibbs CS, Go AS, Chiu VY, Schmitt SK, Ananth L, Frayne SM. Mental illness and warfarin use in atrial fibrillation. The American journal of managed care. 2011 Sep 1; 17(9):617-24. [view]
2. Turakhia MP. Quality of stroke prevention care in atrial fibrillation: many moving targets. Circulation. Cardiovascular quality and outcomes. 2011 Jan 1; 4(1):5-8. [view]
3. Yoon J, Scott JY, Phibbs CS, Wagner TH. Recent trends in Veterans Affairs chronic condition spending. Population health management. 2011 Dec 1; 14(6):293-8. [view]

Journal Other

1. Turakhia MP, Heidenreich PA. Using restriction to minimize bias in observational studies. JAMA : the journal of the American Medical Association. 2010 Dec 1; 304(21):2359; author reply 2359-60. [view]

VA Cyberseminars

1. Frayne S, Lipson L, Phibbs C, Yano EM. Improving Care for Women Veterans: How to Move from Observational to Interventional and Implementation Research. [Cyberseminar]. 2011 Nov 2. [view]

Conference Presentations

1. Frayne SM, Turakhia M, Moos RH, Friedman SA, Schmitt SK, Berlowitz D, Phibbs CS. Anticoagulation outcomes in atrial fibrillation: Impact of mental illness. Poster session presented at: Society of General Internal Medicine Annual Meeting; 2011 May 6; Phoenix, AZ. [view]
2. Frayne SM, Turakhia M, Moos RH, Friedman SA, Schmitt SK, Berlowitz D, Phibbs CS. Anticoagulation outcomes in atrial fibrillation: Impact of mental illness. Paper presented at: AcademyHealth Annual Research Meeting; 2011 Jun 13; Seattle, WA. [view]
3. Frayne SM, Turakhia M, Schmitt S, Friedman SA, Chiu V, Moos RH, Heidenreich PA, Berlowitz D, Phibbs CS. Atrial Fibrillation Management in Patients with Mental Illness: Are they Anticoagulated? Are They Monitored? Paper presented at: VA HSR&D Field-Based Mental Health and Substance Use Disorders Meeting; 2010 Apr 29; Little Rock, AR. [view]
4. Frayne SM, Turakhia M, Schmitt SK, Friedman SA, Chiu VY, Moos RH, Heidenreich PA, Berlowitz D, Phibbs CS. Atrial Fibrillation Management in Patients with Mental Illness: Are they Anticoagulated? Are They Monitored? Poster session presented at: VA HSR&D National Meeting; 2011 Feb 25; National Harbor, MD. [view]
5. Turakhia M, Phibbs CS, Schmitt SK, Heidenreich PA, Frayne SM. Facility Variation in Quality of Anticoagulation is Directly Related to Facility INR Test Frequency: A National VA Study. Poster session presented at: Heart Rhythm Society Annual Scientific Sessions; 2010 May 14; Denver, CO. [view]
6. Turakhia M, Phibbs CS, Moos RH, Friedman SA, Heidenreich PA, Frayne SM. Facility volume is not associated with increased quality of anticoagulation care for patients with atrial fibrillation. Poster session presented at: American Heart Association Quality of Care and Outcomes Research in Cardiovascular Disease and Stroke Annual Scientific Sessions; 2010 May 21; Washington, DC. [view]
7. Turakhia M, Phibbs CS, Schmitt S, Moos RH, Heidenreich PA, Le S, Frayne SM. Systematic Variation of Facility Quality of Anticoagulation for Atrial Fibrillation: A National VA Study. Paper presented at: VA HSR&D Career Development Annual Meeting; 2010 Feb 25; San Francisco, CA. [view]
8. Chiu VY, Turakhia M, Phibbs CS, Moos RH, Schmitt SK, Le SY, Friedman SA, Frayne SM. Time to Next INR as an Anticoagulation Quality Measure in Atrial Fibrillation. Poster session presented at: Society of General Internal Medicine Annual Meeting; 2010 Apr 28; Minneapolis, MN. [view]

DRA: Health Systems, Mental, Cognitive and Behavioral Disorders
DRE: none
Keywords: Patient outcomes, Stroke, VA/non-VA comparisons
MeSH Terms: none