Economic Grand Rounds: Predictors of Missed Appointments Over the Course of Child Mental Health Treatment

This column presents data from 2,903 consecutive referrals (1,727 males and 1,176 females) of patients between three and 17 years of age who were seen for more than one therapy session in our child and adolescent psychiatry clinic between January 2003 and December 2008. The analyses are based on 31,941 total appointments. The median number of appointments per family was 12 (range two to 194). Forty-five percent of the children (N=1,306) were referred by their parents, 20% (N=580) by their primary care physicians, 14% (N=406) by their school, and 11% (N=319) by the county Department of Social Services. The median number of children per family in the sample was two. Mothers and fathers had a similar level of education (M±SD=12.51±1.2 and 12.44±1.3 years, respectively). Thirty-one percent of the children (N=899) lived in a single-parent household. The clinic population is ethnically diverse and included 1,538 Caucasians (53%), 580 African Americans (20%), 493 Latinos (17%), 232 American Indians (8%), and 60 Asian Americans (2%). Based on the Hollingshead Index of socioeconomic status, 16% (N=464) were lower class, 51% (N=1,480) lower middle class, and 20% (N=580) middle class.

The most prevalent child diagnoses were disruptive behavior disorders (for example, attention-deficit hyperactivity disorder, oppositional defiant disorder, and conduct disorder) (N=848, 29%), followed by anxiety disorders (for example, separation anxiety, obsessive-compulsive disorder, and generalized anxiety disorder) (N=754, 26%), mood disorders (for example, major depressive disorder and bipolar disorder) (N=597, 20%), and autism spectrum disorders (for example, autism, and pervasive developmental disorder not otherwise specified) (N=408, 14%). The mean global assessment of functioning score at initial intake for patients in the sample was 49.6±17.6, indicative of serious symptoms and impairment ( 6 ).

For this analysis we defined a missed appointment for a family in ongoing treatment as one that was not kept and not cancelled at least 48 hours in advance. Because children and adolescents who were engaged in therapy longer were more likely to miss an appointment, we calculated a percentage of the total number of appointments attended divided by the total number scheduled. Data indicating which family member attended the appointment were not included in the analyses.

Child and adolescent clinical diagnoses were assigned as part of the clinic intake process and were based on DSM-IV criteria. The disorder that resulted in the most impairment was considered the primary diagnosis; if a child or adolescent patient had comorbid conditions, only the primary diagnosis was considered in the analyses. If the intake diagnosis and discharge diagnosis were not identical, we decided a priori to regard the discharge diagnosis as correct. Parental diagnoses were determined by their response to the following question: "Have you ever been diagnosed with a psychiatric disorder?"

Each sociodemographic and clinical predictor was entered into a simple univariate linear regression model predicting the outcome variable, logit-transformed percentage of missed appointments. Continuous variables such as age and the number of children in the family were entered directly into the linear regression models. However, most of the variables were categorical variables with multiple levels, for which dummy variables were created (0 for the base category and 1 for the additional variable). For all categorical dummy variables, 0 was assigned to the first variable (for example, males, 0, and females, 1). Many categorical variables had more than two levels, necessitating the creation of multiple dummy variables. All remaining conditions had a code of 1 for the variable that described that condition and a code of 0 for the other dummy variables.

The missed appointment rate for the entire sample was 13% (N=4,152 appointments). Thirty-eight percent of the families in our sample never missed an appointment without cancelling at least 48 hours in advance. The mean number of missed appointments among those who missed at least one was 3.7±1.1. No significant difference was found in the mean number of kept appointments for the missed-appointment group (18.6±5.5) and for those who did not miss an appointment (16.9±3.8).

Children whose treatment was covered by private insurance (N=1,742) attended fewer total sessions than children who had public insurance (N=1,161) (16.7±2.8 versus 24.4±3.1; F=5.36, df=1 and 1,829, p=.005, h2=.01). Two variables were both independently associated with attending fewer overall sessions and dropping out of treatment earlier—maternal bipolar disorder ( β =-.124, p=.003) and paternal substance abuse ( β =-.119, p=.005). No other variables, including child diagnosis, gender, or race-ethnicity were associated with total number of treatment sessions.

Multiple factors predicted the logit-transformed percentage of missed appointments. The best predictors of missing more appointments were maternal depression ( β =-.442, p<.001), living more than 30 miles from the clinic ( β =-.258, p=.008), and having a parent who was single (never married, separated, or divorced) ( β =-.297, p=.018). For children an anxiety disorder diagnosis predicted fewer missed appointments ( β =.230, p=.005). These four variables alone explained 58.1% of the variance in the missed appointment ratio (F=41.04, df=3 and 2,895, p<.001). [A table presenting the results of this analysis is provided as an online supplement to this column at ps.psychiatryonline.org .]

We identified a range of factors associated with missed appointments for children in ongoing psychiatric treatment. Most of the predictive variables were related to parent and family characteristics rather than to characteristics of the child. The strongest predictor was a mother's reported history of depression. Although we were not surprised that maternal depression would be related to missed appointments, we were struck by the sheer magnitude of the effect. It was especially impressive because we relied on self-reported history of depression rather than on a standardized assessment of current mood status.

Because of the strength of maternal depression as a predictor, clinicians who treat children should focus on this factor from intake to discharge, if for no other reason than to help ensure treatment adherence. This finding adds to an already substantial body of literature that shows the impact of maternal depression on many central domains: the accuracy of maternal responses on child behavior rating scales ( 7 ); the potential benefit of parent training, cognitive-behavioral approaches, and preventive interventions ( 8, 9 ); and a child's increased risk of future psychopathology ( 10 ). Maternal depression has also been implicated in a low level of positive emotions among children, which may be a mediator to developing depression ( 11 ). On the basis of these findings and the results of our study, the argument could be made that maternal depression should be regarded as a critical factor in the management of child psychiatric disorders. The findings might also justify use of the family treatment approach to pharmacological management whereby parents are treated for their own psychiatric problems alongside their children ( 12 ).

Our findings suggest another way in which an adult's depression can have powerful negative consequences. Depression is second only to coronary heart disease in the functional limitations, role impairment, and economic costs entailed ( 13 ). Fewer than half of adults with current major depression are diagnosed as such by primary care physicians ( 14 ). Therefore, identification of parental depression and referral for treatment should perhaps be among the responsibilities of the child mental health care provider.

In addition to maternal depression, living more than 30 miles from the clinic and having a parent who was single (never married, separated, or divorced) were significant predictors of missing a higher percentage of appointments. Prior research on missed initial appointments has found that single parents were more likely to miss appointments than married or divorced parents ( 15 ); our study extends this finding to missed appointments during ongoing therapy. These data are also consistent with other studies finding that individuals and families who have to travel substantial distances to the clinic are more likely to miss appointments ( 16, 17 ). However, this result should be viewed cautiously until replicated. Similarly, our finding that families with a child in treatment for an anxiety disorder were more likely to keep appointments is worthy of further exploration.

Investigation of potential predictors of missed appointments during ongoing child treatment that were not considered in our analysis is warranted, including seasonal factors, the day of the week, waiting times at the clinic visit, the severity of the child's psychiatric symptoms and functional impairments, and the social support networks available to the family. Studies should also explore the link between appointment adherence and treatment outcome. To the best of our knowledge, that relationship, although it is presumed to be positive, has never been established in child psychiatric populations.

A potential limitation of this study is related to clinic management policies that may be somewhat atypical. Specifically, the overall missed appointment rate of 13% is below national estimates. The relatively low rate may be the result of policies we have instituted over time to increase attendance, including asking families to read and sign an attendance policy at the first visit. We also enforce that policy when it is clinically indicated. Our computerized system allows us to track in real time missed appointment rates by patient, staff, type of service, and program. This capacity allows us to attend to missed appointments both administratively and in our clinical management. Similarly, our use of a 48-hour cancellation requirement may be more stringent than is common, although we are not aware of any standard in the field. Therefore, for whatever reason, the results of this study may have been affected by ongoing administrative efforts to enhance clinic attendance.

Identification of patients who are prone to miss appointments allows clinicians and clinic managers to focus on that group, monitor their adherence, and develop strategies for improving attendance. Data on missed appointments also allow a clinic to develop policies and procedures aimed at monitoring and reducing no-shows. Such strategies might involve both enforcement mechanisms (such as termination policies) as well as interventions aimed at making attendance easier (such as transportation, child care, and extended hours). Reducing rates of missed appointments and enforcing treatment adherence policies can improve service availability without funding additional staff or programming. Given the nationwide shortage of mental health services for children and adolescents, a 10%–20% improvement in missed appointment rates represents a substantial benefit.

Dr. Gordon and Dr. Antshel are affiliated with the Department of Psychiatry and Behavioral Sciences, State University of New York Upstate Medical University, 750 East Adams St., Syracuse, NY 13210 (e-mail: [email protected]). Dr. Lewandowski and Ms. Seigers are with the Department of Psychology, Syracuse University, Syracuse, New York. Steven S. Sharfstein, M.D., Haiden A. Huskamp, Ph.D., and Alison Evans Cuellar, Ph.D., are editors of this column.