Pages et al developed a model that can help identify a psychiatric patient who is more likely to leave against medical advice (AMA). These can help identify a patient for whom early discharge planning for outpatient treatment is recommended. The authors are from the University of Washington in Seattle.


Predictors of leaving AMA:

(1) significant substance abuse problem (large quantity and/or large number of substances)

(2) >= 2 previous inpatient psychiatric hospitalizations

(3) non-White ethnicity

(4) absence of functional impairment secondary to physical illness

(5) male gender

(6) none to mild suicidality on admission


total number of predictors =

= SUM(number of predictors present)



• minimum number of predictors: 0

• maximum number of predictors: 6

• The presence of >=4 predictors was associated with a higher risk for leaving AMA.

• The presence of <= 1 predictors was associated with a lower risk of leaving AMA.


Other factors associated with leaving AMA:

(1) tend to be uncooperative

(2) show significant denial at discharge

(3) more likely to live alone

(4) tend to be unemployed


Patients who leave AMA have worse outcomes, more severe symptoms at discharge and a higher rate of rehospitalization than those who do not.


To read more or access our algorithms and calculators, please log in or register.