Increasing Identification of Psychosocial Problems

To the Editor. Drs Kelleher et al1 tackled a critical and neglected topic in their research on detection of psychosocial problems in primary care. Although the apparent increase in identification rates is encouraging, the study's research protocol undoubtedly influenced detection rates. Physicians in both 1979 and 1996 were given a list of diagnostic codes to consider at each visit. The list essentially functioned as a standardized reminder to consider psychosocial development at each visit. In the real world, only about 25% of providers use standardized tools for developmental or behavioral/emotional screening.2Had the study included a control group, it is likely that far fewer children would have been identified. One chart review study showed that physicians failed to recognize psychosocial problems in 83% of children.3 Using similar methods, Palfrey et al4 found that clinicians did not detect developmental disabilities in more than 70% of patients. Parents have the ability to offer highly accurate indicators of developmental and behavioral/emotional status, but in the absence of a standardized tool for eliciting the parents' concerns, only 30% to 40% discuss their worries spontaneously.5–6 In general, the use of structured forms is associated with significantly higher levels of both recorded and observed performance and even with overdocumentation.7Thus, primary care providers should not interpret results from the Kelleher et al as an indicator that all is well in the early identification of psychosocial problems in primary care. Rather, the study should be viewed as strong support for the use of validated tools in early detection. Clinicians interested in learning about accurate screens may wish to visit the American Academy of Pediatrics' Section on Developmental and Behavioral Pediatrics website:http://www.dbpeds.org/articles/dbtestingIn Reply. We thank Dr. Glascoe for her interest in our study1and respond to each of her points below:1. “Although the apparent increase in identification rates is encouraging, the study's research protocol undoubtedly influenced detection rates.”Dr Glascoe suggests that the use of a checklist for clinicians may have resulted in the greatly increased rate in psychosocial problems identification from 1979 to 1997. However, the same format and categories were used in both periods and were, therefore, unlikely to account for the large change observed.2. “One chart review study showed that physicians failed to recognize psychosocial problems in 83% of children.”Dr Glascoe cites 2 chart review studies about underidentification by primary care clinicians. However, we have evidence that even without structured forms or reminders, clinicians underreport diagnosed psychosocial problems in medical records because of concerns about insurability, reimbursement, and stigma.2 Thus, chart review studies are unlikely to capture physician recognition for psychosocial problems.3. “In general, the use of structured forms is associated with significantly higher levels of both recorded and observed performance.”We agree. However, even with a specific question asking about the presence of a psychosocial problem, <60% of children symptomatic by parent report were identified by the primary care clinicians. At the same time, a large number of children not reported as symptomatic by parents were identified by primary care clinicians, suggesting that clinicians also have concerns not raised by parents. These findings, which we report elsewhere,3 deserve additional investigation.4. “The study should be viewed as strong support for the use of validated tools in early detection.”Like Dr Glascoe, we are distressed at the large number of children with behavioral and emotional problems currently unrecognized and untreated. However, we disagree with her conclusion that structured screening tools will help, especially in the absence of more effective intervention for school-aged children with emotional and behavioral disorders. The US Preventive Services Task Force4 has outlined criteria for screening in primary care that includes easily implemented screening and scoring, evidence that earlier intervention is effective, evidence that services of proven effectiveness are available in the community. Primary care practice for psychosocial problems is far from meeting these simple criteria. In fact, studies among adults5 and children6 of screening implementation for psychosocial problems have failed to show improvements in practice or outcomes for persons with emotional or behavioral disorders.We hope that someday specific, standardized screening instruments will be implemented, but preferably only in combination with innovative interventions to improve outcomes.