Economic Grand Rounds: The Cost of Collaboration: Predictors of Hours Spent in Collateral Contacts

Although it has become an article of faith that the delivery of effective mental health care for youths requires a seamless integration of services across the continuum of care, little attention has been paid to the actual cost of those collaborative efforts. Collateral contacts, while at the heart of wraparound care, are time consuming and often nonreimbursable. When clinicians visit schools, talk with case workers, or phone a pediatrician, they do so as part of proper case management, not because it accrues to their clinical hours or the agency's bottom line.

We are unaware of any existing published data on the average amount of time that pediatric clinicians spend on collateral activities (such as meetings or phone calls) or the typical ebb and flow of those activities over the course of treatment. Without data, it is challenging for administrators to judge what constitutes a reasonable caseload or how to charge for services in contracts that stipulate coordination of efforts. The absence of valid information precludes evidence-based decision making for how child mental health services should be reimbursed. The literature is also devoid of information regarding which factors predict collateral time. If those predictors could be identified, administrators could make more empirically based decisions regarding case assignments.

This column presents data from a large number of patients regarding the amount of time that clinicians spent in clinical activities outside the context of the billable service. We also wanted to determine whether the collateral ratio (collateral time to billable time) can be predicted by demographic variables, child diagnosis, parental psychopathology or family history of mental disorders, and staff variables.

Collateral contacts

The sample consisted of 1,639 patients (956 males, or 58%, and 683 females, or 42%) between the ages of three years and 17 years who were seen for a total of 22,127 appointments (mean±SD=13.5±28.8 appointments) in a child and adolescent psychiatry clinic based within an academic medical center's department of psychiatry. Data were amassed on consecutive referrals over a six-year period spanning January 2002 to January 2008. We analyzed data only from families that attended more than the intake session and for whom the clinician had recorded collateral activities. This group represented 57% of all referrals to the clinic over this period (1,639 of 2,875 referrals). The study was approved by the university's institutional review board.

The study sample included 852 (52%) Caucasians, 329 (20%) African Americans, 280 (17%) Latinos, 131 (8%) American Indians (8%), and 47 (3%) Asian Americans. On the basis of the Hollingshead Index of socioeconomic status, 855 families (52%) in our sample were predominantly lower-middle class, 327 (20%) were middle class, and 267 (16%) were lower class. The mean number of years of education was comparable for mothers (12.51±1.23 years) and fathers (12.44±1.31 years). A total of 524 children in our sample (32%) lived in a single-parent household.

The most prevalent child diagnoses were disruptive behavior disorders (for example, attention-deficit hyperactivity disorder, oppositional defiant disorder, and conduct disorder) (N=442, 27%), anxiety disorders (for example, separation anxiety disorder, obsessive-compulsive disorder, and generalized anxiety disorder) (N=394, 24%), mood disorders (for example, major depressive disorder and bipolar disorder) (N=311, 19%), and autism spectrum disorders (for example, autism and pervasive developmental disorder, not otherwise specified) (N=213, 13%). The mean Global Assessment of Functioning (GAF) score for child patients at the initial intake was 49.8±17.3, indicative of serious symptoms and impairment ( 1 ). (Possible scores on the GAF range from 1 to 100, with higher scores indicating better functioning.)

Collateral activities were defined as any case-related clinical effort that occurred outside of the standard therapy visit. These included such activities as phone calls, school visits, meetings, and paperwork, but they did not include supervision. We intentionally made the definition broad to capture the entirety of nonbillable efforts that contributed to treatment. Because of the wide range of treatment length, we computed a "collateral ratio" that represented the total time of collateral activities divided by the total time of appointments. Higher values indicate a higher ratio of collateral activities to time in direct patient contact.

Child patients were assigned DSM-IV diagnoses from intake to discharge. We applied the following set of a priori decision-making rules: In the case of comorbid conditions, we considered only the primary diagnosis for the purposes of this study. If the intake diagnosis and discharge diagnosis differed, we analyzed only the discharge diagnosis. We entered a parental diagnosis based entirely on a response to the questions, "Have you ever been diagnosed with a psychiatric disorder" and "If so, what diagnoses did you receive?" Parents were not administered a formal diagnostic interview as part of their child's clinic intake.

Data on collateral activities were collected through a computerized clinic management system. As a part of routine procedures, clinicians entered the type of activity (for example, phone call or meeting) and the duration of that activity in increments of five, ten, 15, 30, and 45 minutes up to three hours. The clinic tracking was developed in a way that made it easy for clinicians to enter the information as they composed the collateral notes required by Office of Mental Health documentation requirements.

Findings

The mean collateral ratio for our sample was .33±.20. For every 60 minutes of direct patient contact, 19.7±15.0 minutes of collateral activities were performed by the clinician. Of the 18,503 collateral activities that were performed, the most common collateral activities were paperwork (N=7,031, 38%), phone calls (N=6,661, 36%), and meetings with external contacts at schools, social services, medical offices, and the juvenile justice system (N=2,590, 14%). Meetings accounted for 35% of all time spent on collateral activities, averaging 72.4±15.1 minutes per meeting, followed by phone calls (23%, with a mean of 15.9±16.2 minutes per phone call), and paperwork (21%, with a mean of 14.4±20.1 minutes per time doing paperwork).

Collateral ratios were higher during the first and last third of treatment (F=6.11, df=2 and 1,637, p=.007). Collateral ratios were higher for African-American and Latino families than for families of other races and ethnicities (F=4.96, df=4 and 1,634, p=.001). Children whose parents were married required fewer collateral activities than children whose parents were divorced, separated, or never married (F=19.18, df=2 and 1,636, p<.001). All other demographic variables, including age, gender, and payment source (public versus private), were nonsignificant.

Children with a history of psychiatric hospitalization (N=263) and suicidal ideation or attempt (N=574) required more collateral activities than those without (F=11.71, df=1 and 1,637, p<.001, and F=10.46, df=1 and 1,637, p<.001, respectively). Children who were misusing or abusing substances (N=410) required more collateral activities than other patients (F=14.74, df=1 and 1,637, p<.001), as did those with a history of maltreatment or trauma (N=546) (F=23.21, df=1 and 1,637, p<.001). No other child clinical variables, including child diagnosis, were associated with the collateral ratio. According to post-hoc analyses, a maternal diagnosis of depression or anxiety was associated with higher collateral ratios (F=35.50, df=5 and 1,655, p<.001).

As for service delivery variables, the collateral ratio was higher for children who were given prescriptions for medication than for those who were not (F=6.29, df=1 and 1,637, p=.012). Trainees in psychology and psychiatry and clinicians at the M.S.W. level spent more nonbillable time with their patients than Ph.D.s and M.D.s (F=7.25, df=3 and 1,635, p<.001). Finally, correlations between initial clinician ratings of impairment and case acuity were all nonsignificant, with Spearman rho correlation coefficients ranging from -.1 to .1. However, the higher the number of missed appointments, the higher the collateral ratio ( ρ =.463, p=.001).

Several variables significantly predicted a logit-transformed collateral ratio. The best predictors, accounting for 62.3% of the collateral ratio (F=34.77, df=18 and 1,604, p<.001) were having parents who were not married ( β =.314, p<.001), a mother with depression ( β =.328, p<.001) or anxiety ( β =.333, p<.001), a child patient with a history of substance misuse or abuse ( β =.214, p=.008), and a child patient with a history of maltreatment ( β =.382, p<.009). Maternal psychopathology, marital status, and childhood history of substance abuse collectively accounted for 44.7% of the variance (F=23.78, df=3 and 1,629, p<.001).

Discussion

On average, for every 60 minutes of direct patient contact, a child clinician spent approximately 20 minutes on collateral activities, such as meetings, phone calls, and paperwork. That collateral ratio of 3:1 varies somewhat depending on the phase of treatment; clinicians spent the most collateral time at the beginning of the case when, presumably, case management is especially important. An uptick also occurred at the end of the case when clinicians collaborated around discharge planning. Although variation around that mean of 20 minutes was relatively small, those additional minutes accrue to meaningful durations over months of treatment.

Although the 3:1 ratio is a rough approximation (and one that likely underestimates the average collateral effort), it nonetheless stands as a stark reality against which to consider current clinic business models and reimbursement methods and rates. According to these data, a clinician with 30 appointments per week will spend at least another ten hours on the phone, in case-related meetings, or completing paperwork. Because clinicians commonly have responsibilities for attending clinic meetings, supervising other employees, being supervised, and managing administrative duties, a 40-hour week will not suffice unless staff either cut the time that they pursue collaborative efforts or spend in direct patient contact. Because clinicians are generally evaluated and reimbursed on the basis of what they bill out for patient care, it is likely that collateral activities will be the first to suffer. Furthermore, systems that fail to account realistically for collateral demands will inevitably encounter significant problems with staff retention and burnout. Expecting clinicians to provide good care in a context that does not allow for sufficient time to pursue the activities required for good care ultimately promotes stress and discouragement.

The data on the extent to which child and adolescent mental health cases demand additional, nonbillable efforts for each appointment are especially sobering against the current push for more integrated, intensive case management services that are "wrapped around" families ( 2 ). By definition, the focus on continuum of care engenders the expectation that clinicians become integrally involved in nonbillable collaboration. It would seem that these models could include only participation from the mental health component of care if clinicians were reimbursed for that involvement. Otherwise, a philosophical commitment to the importance of wraparound services will fall prey to financial and logistical realities.

The most powerful predictors of the collateral ratio were related to parent and family variables. Cases that placed the heaviest collateral demands involved mothers with a history of anxiety or depression and families in which the parents were not married. Perhaps mothers with anxiety problems are more likely to call clinicians because they seek reassurance or because they are acting on what others might regard as less urgent events.

The child variables that predicted collateral overhead (a history of being abused and presenting problems that included substance use) may have emerged as significant because both factors are associated with a youngster's being involved in the child protective or juvenile justice systems. Such cases inevitably require ongoing contact with case workers, probation officers, lawyers, and social workers.

The lack of predictive power of some factors was interesting. For example, our data indicated that, contrary to widely held belief, families on Medicaid were no more demanding of collateral contacts than those who paid through insurance. It was also intriguing that the child-centered variables that were significantly predictive of a higher collateral ratio were not the child's diagnosis, age, or gender, but rather a history of trauma or substance use. What we found most sobering is that clinicians were unable to predict accurately at intake how demanding a case would ultimately be in terms of collateral efforts. According to our data, clinicians should be careful not to jump to conclusions on the basis of initial impressions about how time consuming a case will be to manage.

As for study limitations, the most prominent were related to the aforementioned tendency for clinicians to be somewhat lax in logging all of their collateral activities. Therefore, the 3:1 ratio may underrepresent the number of billable to nonbillable hours clinicians spent on cases. Also, because all paperwork in this clinic is completed on the computer in templates, our clinicians likely spent far less time than those in most facilities on this aspect of collateral activity. The collateral ratio and its predictors may vary depending on the clinic philosophy regarding the importance of pursuing collateral activities, expectations for caseload size, and referral patterns. For example, the clinic that provided these data is situated in a department of psychiatry within a tertiary medical center that may attract more complicated cases. However, because of the unusually large sample size and the sheer number of appointments considered, these results at least provide a crude estimate of the additional time that clinicians put in for appointments.

Conclusions

Perhaps these findings can serve as the initial step in a systematic effort to establish the parameters that determine collateral activities. They might raise awareness among agency administrators, policy makers, and payers that collateral overhead in case management is a significant factor to consider when planning mental health services for children and adolescents. They may also indicate a need for managers to develop innovative ways of helping clinicians be more efficient in how they complete paperwork and interact with case-related sources.

Acknowledgments and disclosures

The authors report no competing interests.