Monitoring of Patients on Second-Generation Antipsychotics: A National Survey of Child Psychiatrists

The use of second-generation antipsychotics among children and adolescents has been rapidly expanding. This trend has major public health implications because these medications can cause significant side effects, which include weight gain, hyperglycemia, hyperlipidemia, and tardive dyskinesia. Youths are especially vulnerable to these side effects (1). Youths treated with second-generation antipsychotics gain, on average, between 2.1 and 8.0 pounds in a 12-week period, have up to three or four times higher risk of developing type II diabetes, and have an increased risk of developing elevated cholesterol (1).

In 2004, the American Psychiatric Association and the American Diabetes Association published guidelines on screening and monitoring of patients on second-generation antipsychotics. In 2011, the American Academy of Child and Adolescent Psychiatry added its endorsement. The guidelines specify routine monitoring of fasting blood glucose, fasting lipid profiles, body mass index (BMI), and abnormal involuntary movements (AIMs) prior to and periodically after initiating second-generation antipsychotics. Despite dissemination of the guidelines, prescribers—including child psychiatrists—have not adhered to them, according to studies of U.S. prescribing practices for children and adolescents in years 2000 through 2011 (2).

The underlying reasons for nonadherence to monitoring guidelines, especially among child psychiatrists, have not been clear. Experts in health care improvement have emphasized the need to understand the problem, identify groups that should be targeted by efforts to improve adherence, and clarify barriers to implementation of adherence with guidelines. We identified three previous surveys of child psychiatrists that assessed monitoring of second-generation antipsychotics (3–5). In a U.S. survey of 334 child psychiatrists (published after the current survey was conducted), 53% of the respondents reported that patients were noncompliant with blood tests, but the study did not explore other barriers to monitoring (4). In a survey of 126 child psychiatrists in Australia, 55% reported that patient noncompliance with recommended tests was a barrier (5). In a survey of 339 Canadian child psychiatrists and developmental pediatricians, most (65%−75%) endorsed monitoring fasting glucose and lipids of children prescribed second-generation antipsychotics, but the reported intervals between actual monitoring varied greatly (3).

To further understand monitoring of second-generation antipsychotics among children, we conducted a nationwide survey of child psychiatrists regarding adherence with the guidelines for metabolic monitoring among children who are prescribed second-generation antipsychotics. The survey assessed self-reported monitoring, facilitators and barriers to monitoring, and the steps taken by child psychiatrists in adherence to monitoring guidelines.

Methods

We sent the survey to 4,144 working e-mail addresses of child psychiatrists throughout the United States who were registered members of the American Academy of Child and Adolescent Psychiatry, excluding psychiatrists in training. We e-mailed the survey in three waves between December 2012 and February 2013 and incentivized respondents to complete the survey with a chance to win an iPad. The Committee for the Protection of Human Subjects at the Geisel School of Medicine at Dartmouth approved the study. Because the responders completed the survey anonymously and without providing identifying information, the committee waived the need for signed informed consent.

We based the survey on the four steps of the “awareness-to-adherence” model (6), which has been used to evaluate adherence to other practice guidelines. The model addresses knowledge-related, attitudinal, and behavioral factors known to affect guideline adherence. It also identifies methods to improve physicians’ adherence by identifying a path to adherence. To conform with guidelines, physicians must sequentially go through the following steps: be aware of the guidelines (knowledge), agree with them (attitude), decide to follow them at the appropriate times or at the appropriate intervals (adoption), and actually follow (adherence) them for the majority of patients (behavior) (6).

The awareness-to-adherence model survey was developed by Pathman and associates (6) to study dissemination of vaccination recommendations. We adapted the survey by substituting the second-generation antipsychotic monitoring guidelines in place of the vaccination guidelines. The survey takes five to ten minutes to complete. Sections include psychiatrists’ demographic and practice characteristics, their awareness and agreement with monitoring, self-reported monitoring, and self-reported barriers to monitoring.

Our team of three board-certified child and adolescent psychiatrists (including JLM and WBD) adapted the survey questions, pilot-tested the survey for clarity and ease of completion with 18 child and adolescent psychiatry prescribers, and revised the survey on the basis of the pilot. [Details about the steps in the awareness-to-adherence model, the questions developed by Pathman and associates (6), the potential responses for various variables, and how responses were dichotomized in our analyses are available as an online supplement to this report.]

We used descriptive statistics for survey results and multiple linear regressions to identify correlates of each monitoring outcome. We used SPSS, version 19.0, to conduct all statistical analyses, setting the p value at .01 to control for multiple tests.

Results

Approximately one-third (N=1,314, 32%) of the child psychiatrists who were invited to participate completed the survey. Respondents’ geographic locations were consistent with the geographic locations of AACAP members (Northeast, 36%; Southeast, 18%; Midwest, 18%; Northwest, 10%; and Southwest, 17%) (7). The two most common practice settings of respondents were solo outpatient practice and community mental health. [A table in the online supplement shows characteristics of the survey respondents, including gender, time in practice, and geographic location.]

Almost all respondents surveyed were aware of the guidelines for monitoring of children and adolescents on second-generation antipsychotics. Awareness was highest for BMI and lowest for AIMS monitoring (N=1,286, 98%, and N=1,257, 96%, respectively). Fewer respondents agreed with the recommendations, with agreement rates ranging from 69% (N=902) for fasting lipids to 80% (N=1,050) for glucose, 89% (N=1,170) for AIMS, and 91% (N=1,198) for BMI monitoring. A minority of physicians had adopted the monitoring practices recommended in the guidelines, with the lowest rate of adoption for AIMs monitoring (N=114, 9%) and the highest rate of adoption for BMI monitoring (N=362, 28%). Similarly small minorities reported adherence to the guidelines; adherence rates ranged from 9% (N=114) for AIMs monitoring to 25% (N=332) for BMI monitoring. [A figure in the online supplement provides a sequential analysis of the percentage of child psychiatrists reporting awareness of, agreement with, adoption of, and adherence to second-generation antipsychotic monitoring guidelines.]

All of the child psychiatrists who agreed with and adopted the guidelines for glucose monitoring (N=205) and AIMs (N=109) also reported adherence to monitoring at the guideline-prescribed intervals. Most of the child psychiatrists who agreed with and adopted the guidelines for monitoring lipids and BMI also reported adherence to monitoring at the guideline-prescribed intervals (N=133 of 183, 73%, and N=311 of 335, 93%, respectively).

Table 1 shows the independent associations of various factors with monitoring over a one-year period of fasting glucose, fasting lipids, BMI, and AIMs after initiation of a second-generation antipsychotic. Child psychiatrists who endorsed being aware of the guidelines or who were in academic practice had greater monitoring levels of all four variables. Physician’s length of time in practice was inversely related to monitoring levels of glucose, lipids, and BMI. Use of an electronic medical record was associated with greater monitoring of BMI only. Child psychiatrists who received reminders to monitor children using second-generation antipsychotics had greater levels of monitoring AIMs but not other side effects. Child psychiatrists who agreed that the guidelines were easy to keep up with or were important for patient care had greater levels of monitoring glucose and lipids.

A large majority of physicians (N=1,077, 82%) reported that parents who forget to obtain laboratory tests were a barrier to obtaining fasting glucose and lipids. Other commonly cited barriers were parental resistance (N=683, 52%) and children’s refusal to obtain the tests (N=828, 63%). A minority of physicians reported forgetting to order laboratory tests as a barrier to monitoring (N=355, 27%).

Discussion

A majority of child psychiatrists who responded to our survey reported that they were aware of the guidelines for second-generation antipsychotic monitoring, but a minority of child psychiatrists had adopted the guidelines, and even fewer reported full adherence to the guidelines. The low rate of monitoring glucose, lipids, BMI, and AIMs was consistent with the low rates of monitoring found in retrospective claims data for children prescribed second-generation antipsychotics (2).

The application of the awareness-to-adherence model to this survey suggests that awareness of the second-generation antipsychotic monitoring guidelines and agreement with the guidelines are not generally obstacles to monitoring. Thus making physicians more aware of the guidelines by further dissemination of and education about the importance of guidelines is unlikely to be helpful, given that a majority of child psychiatrists already agree with the guidelines (8). The low rate of monitoring, despite high levels of awareness of and agreement with the need to monitor, suggests that obstacles interfere with adopting these guidelines (6,8). Further efforts should focus on improving physician adoption of the guidelines. A few examples of obstacles to adoption may be lack of physician knowledge regarding the recommended monitoring intervals, the absence of reminders indicating when a child is due for monitoring, the lack of a consistent and easily accessible recording mechanism for the monitored parameters (3), and parent and child knowledge and attitudes. Similar to previous reports (4,5), this study found that a large majority of child psychiatrists reported that parents who forget to obtain laboratory tests, children’s refusal to be tested, and parental resistance are barriers to monitoring.

Potential interventions to improve adoption include education with audit and feedback directed at specific providers who have not adopted recommended monitoring practices, monitoring reminder systems, and family education to improve knowledge and attitudes about monitoring (9). Our survey showed that physicians who have adopted the guidelines and monitor patients consistent with the guidelines are much more likely to adhere to them.

Child psychiatrists in academic practice were more likely to monitor all the recommended antipsychotic side effects. Those who had completed training recently were more likely to monitor lab values and BMI, which may reflect different views of managing psychotropic medications based on training era. Similarly, in another small survey sample of child psychiatrists, those practicing for longer periods were less likely to monitor fasting glucose and lipids (10). Thus interventions to improve monitoring should target physicians in nonacademic practices and physicians who have been out of training longer.

Several potential limitations of our study warrant mention. First, our data came from a self-selected subsample of child and adolescent psychiatrists who are part of a national organization. The results, therefore, may not generalize to all 8,300 child psychiatrists in the United States. Nevertheless, the participants’ characteristics were roughly distributed across gender, years in practice, practice location, and practice type, and their characteristics were similar to published characteristics of American child psychiatrists and the characteristics of respondents in other published reports. For example, respondents’ gender was comparable to gender reported in other surveys of AACAP members and child psychiatrists in the United States and comparable to that in the American Association of Medical Colleges’ Physician Specialty Data report (11). Additionally, respondents’ time in practice and practice settings were also comparable to those reported in other surveys of U.S. child psychiatrists (4,12).

Second, selection bias is a limitation of survey research. We obtained a survey response rate of 32%; although this rate is similar to typical response rates by mental health providers in other survey studies (25%−40%) (13), it is possible that responders were overall less (or more) inclined to monitor use of second-generation antipsychotics. Past studies have shown that physician self-report may either over- or underestimate their actual practice compared with audit (8). Third, our study focused on the perspectives of child psychiatrists, but the perspectives of parents and children influence physician adherence to practice guidelines. Future studies should survey parents and children to explore the supports that they need to comply with monitoring and gain an understanding of their barriers to monitoring. Finally, our study explored some potential facilitators and barriers to monitoring, however, others may be important, such as availability of on-site lab monitoring within an integrated behavioral health center.

Despite its limitations, this is the largest study that has evaluated second-generation antipsychotic monitoring behaviors among child psychiatry prescribers in the United States. Additionally, it is the largest survey of second-generation antipsychotic monitoring behaviors among all prescribers. Thus these findings warrant serious consideration in quality improvement efforts within practice settings in the United States.

Conclusions

This survey suggests that American child psychiatrists are generally aware of and agree with the need to monitor potential side effects among children treated with second-generation antipsychotics, but they have not adopted monitoring practices consistent with recommended practice guidelines. Child psychiatrists who completed training less recently and who are not working in academic settings may be least likely to adopt monitoring guidelines. Families may not understand the importance of or be able to participate in monitoring, for reasons that are not yet clear. The field needs interventions to improve adherence with second-generation antipsychotic monitoring guidelines for psychiatrists and families.