 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
ARTICLE |
Departments ofa Pediatrics
b Epidemiology and Preventive Medicine, University of Maryland School of Medicine, Baltimore, Maryland
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
METHODS. A total of 216 mothers (because 96% of caregivers were mothers, we use this term) bringing in children <6 years of age for child health supervision completed a parent screening questionnaire in a primary care clinic. The parent screening questionnaire, a brief screen for psychosocial problems developed for the study, includes 2 questions on depressive symptoms. Mothers then completed the computerized study protocol within 2 months. This included the parent screening questionnaire as well as the Beck Depression Inventory II. Different combinations of the depression questions were evaluated against Beck Depression Inventory II clinical cutoff values.
RESULTS. Twelve percent of the mothers met the Beck Depression Inventory II clinical cutoff value for at least moderate depressive symptoms. There was moderate stability of the screening questions. When a positive response to either or both of the 2 questions was considered, the sensitivity was 74%, the specificity was 80%, the positive predictive value was 36%, and the negative predictive value was 95%.
CONCLUSIONS. Maternal depressive symptoms are prevalent. A very brief screen can identify reasonably those who could benefit from additional evaluation and possible treatment. This should benefit mothers, families, and children.
Key Words: child maltreatment • risk factors • depression • prevention • screening • primary care
Abbreviations: PSQ—parent screening questionnaire • BDI—Beck Depression Inventory
Depression in adults is a highly prevalent problem. Point prevalence rates of depression or depressive symptoms in women of parenting age range from 12% to 48%.1–4 As many as 23% of mothers bringing their children to a pediatric primary care clinic screened positive for depression.5
The harmful impact of maternal depression on children's health, development, and behavior has been amply demonstrated.6–10 Higher rates of low birth weight, poor growth, behavior and sleep problems, somatic complaints, learning difficulties, noninflicted injuries, and affective illness have been found for children of depressed mothers, compared with children whose mothers were not depressed.10 Field7 described several physiologic findings for infants of depressed mothers, including elevated norepinephrine and cortisol levels, lower vagal tone, and neurologic delays. McLennan and Kotelchuck11 found that mothers who reported high levels of depressive symptoms were less likely to implement preventive practices, such as using car seats and covering electrical outlets. Maternal depressive symptoms have also been associated with increased risk of infant hospitalization, use of corporal punishment, and lower likelihood of having a smoke alarm and using the supine sleep position.12 Minkovitz et al13 found that maternal depressive symptoms were associated with increased use of acute care, including emergency department visits, and decreased receipt of preventive care, including immunizations. Maternal depression has also been linked to diminished parenting expectations, confidence, and skills,7,9 as well as to child maltreatment.14,15
Treatment of depression can be quite effective. For example, pharmacotherapy and psychotherapy are recommended for moderate to severe depression. Some studies have found medications alone to help patients with major depressive disorder significantly.16 Cognitive, behavioral, and interpersonal approaches have been found to be as effective as medication in treating mild to moderate depression,17–19 and psychotherapy alone is recommended. Recently, treatment of depressed mothers was shown to benefit their children.20 Another study found that mothers treated successfully for depression reported fewer behavior problems in their children.21
Despite the high prevalence of depression and the effectiveness of treatment, few mothers are treated for depression,22 partly because of their lack of contact with health or mental health care services.23 Frequent visits with a pediatrician offer an opportunity to address this barrier. Olson et al5 reported that 40% of women who screened positive for depression in pediatric clinics accepted referrals for additional intervention.
On the basis of the aforementioned evidence, pediatricians have been encouraged to include screening for depression in primary care settings.24,25 In addition, mothers seem interested in pediatricians doing so.23 Nevertheless, pediatricians have been found to identify depression rarely.26 Few have implemented a systematic approach to screening for maternal depression.22
The US Preventive Services Task Force has recommended routine screening of all adults for depression, with the use of 2 questions pertaining to mood and anhedonia (ie, loss of pleasure).27 Whooley et al28 found that 2 questions performed as well as longer screening measures, such as the Beck Depression Inventory (BDI)29 and the Center for Epidemiological Studies-Depression scale. In addition, both interview and paper-and-pencil approaches have been validated,30,31 although Olson et al5 found a much higher rate of positive results with paper screening (22.9% vs 7.6%).
Despite considerable evidence supporting screening of mothers for depression in pediatric primary care, we are aware of only a single study that examined the validity of brief screening tools in such settings. This study compared a 3-item version with the 8-item Rand measure from which it was derived.2 The objective of the present study was to test the stability, sensitivity, specificity, and positive and negative predictive values of a 2-item screen for parental depression in a pediatric primary care clinic. This was performed as part of a larger project that trained residents to address prevalent psychosocial problems, implemented a parent screening questionnaire (PSQ) for the targeted problems in their continuity clinics, and included a social worker in efforts to meet identified needs.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
Five hundred seven parents were approached to participate; some left the clinic before a research assistant could speak with them. Of those approached, 382 (75%) agreed to participate, and 308 (81% of those recruited) completed the initial study protocol (Fig 1). The remaining 74 of those recruited did not keep the scheduled appointments, and a few actively withdrew. Of the 308, 92 did not complete the PSQ in the clinic within a 2-month period or had incomplete data. There were thus 92 who needed to be excluded, leaving 216 for the analyses.
|
The present article is based on the sample of 216 caregivers in the intervention group who completed the PSQ in the primary care clinic and a computerized version in the study protocol within 2 months (Table 1). Most of the caregivers were single mothers who were unemployed. They averaged 25 years of age and had varying levels of education. The children averaged 11.8 months of age. Approximately one half were male, and most were black. Families had an average of 2.3 children and 2.2 adults in the home, and most were receiving Medical Assistance.
|
In addition, parents were approached and asked to participate in research to validate the PSQ. If they were interested, then informed consent procedures were implemented, as approved by our university's human subjects review committee. Parents were asked to return within 2 weeks to complete an audio-computerized self-interview of the study protocol that included the PSQ. The median time before this was completed was 8 days. The computerized interview also had more comprehensive, standardized measures of the targeted problems, such as the BDI II, to examine the validity of the screen. Subjects were paid $60 for completing the study protocol.
Measures
PSQ
The PSQ was developed as a tool for pediatric practices to screen parents briefly for major psychosocial problems and risk factors for child maltreatment (eg, maternal depression, substance abuse in the family, or intimate partner violence). We reviewed the literature to prioritize risk factors that were reasonably amenable to intervention. Next, validated screening measures for several risk factors were reviewed and incorporated. The 2-item depression screen developed by Whooley et al28 was found to have very good psychometric properties. However, this measure required minor modifications in the wording to be suitable for a low-income, less-educated population. Consequently, the words "bothered by" were removed from the questions, "During the past month, have you often been bothered by feeling down, depressed, or hopeless?" and "During the past month, have you often been bothered by little interest or pleasure in doing things?" An advisory committee of community pediatricians guided the development of the PSQ and recommended keeping the PSQ brief, with a yes/no response set. The PSQ was designed to be easy for pediatricians to assess by readily noting the yes responses. The PSQ was then pilot tested with 
40 parents in neighboring pediatric clinics. Parents' input was used to reduce the number of items and to clarify those that were unclear. Because the PSQ solicits sensitive information, it was important to frame this in a supportive context. Therefore, the screen started with an introduction that conveyed an empathetic tone, concern about safety, and a willingness to help (Appendix).
|
BDI II
The BDI II is a widely used measure for screening for depression in adults.29 It has high internal consistency (ranging from 0.86 to 0.93) and high test-retest reliability. It has excellent concurrent validity with other self-report measures and clinical ratings of depression. There is also substantial evidence of its construct validity. Its reliability, validity, and factor structure are similar across diverse populations. In addition, it has been found to have adequate positive and negative predictive values.32,33 Although the BDI II does not provide a definitive diagnosis of depression, it does identify the presence and severity of symptoms consistent with DSM-IV criteria. Ideally we would have conducted clinical interviews to establish a diagnosis of depression, but resources for the study precluded this option. The BDI II thus served as a reasonable compromise to validate a very brief, 2-item screen.
The BDI II is composed of 21 questions, each with scores ranging from 0 to 3, reflecting the frequency with which the problem occurs. The total BDI II score (range: 0–63) is the sum of the scores for the 21 questions. A total BDI II score of >19 has been found to correspond to moderate or severe clinical depression. Missing responses for any of the questions make it difficult to ascertain an accurate total BDI II score. To address this problem, any parent with a total BDI II score of >19, regardless of missing responses, was considered at risk for moderate depression. In addition, if the total BDI II score and number of missing responses were in the following combinations, then parents were categorized as not clinically depressed: score of 
16 and 1 missing item (a score of 3 for the missing item could yield a total score of 19, just short of the clinical cutoff point), score of 13 and 2 missing items, score of 10 and 3 missing items, score of 7 and 4 missing items, score of 4 and 5 missing items, or score of 1 and 6 missing items. Twenty participants were excluded because missing items could have placed them over the clinical cutoff point; therefore, the final sample size was reduced to 196.
Data Analysis
Stability
To determine the stability of the depression screen, we compared the same 2 depression questions on the clinic PSQ and the computerized PSQ. Because a caregiver could have completed >1 clinic PSQ, the last one within 2 months before the computerized interview was chosen for analyses. The 
value described by Cohen34 was used as the measure of reliability. Values range from –1 to 1; values of <0.21 indicate no to slight agreement, values of 0.21 to 0.6 indicate fair to moderate agreement, and values of >0.6 indicate substantial to perfect agreement.
Validity
To determine the validity of the clinic depression screen, we compared it against the BDI II by examining the sensitivity, specificity, positive predictive values, and negative predictive values. We compared each screening item, and the 2 items in different combinations, with the BDI II assessment (depressed/not depressed) by using the cutoff point of >19, representing moderate to severe depression.
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
values of 0.62, 0.45, 0.57, and 0.50 were observed for questions A, B, A or B (either one yes), and A and B (both yes), respectively.
Validity
Table 3 presents results pertaining to the validity of the depression screen items, compared with the BDI II. The sensitivity of the screen was maximized when a positive screen was defined as a yes response to either or both of the depression questions (A or B). Use of this definition also maximized the negative predictive value; a respondent with a negative screen had a 95% likelihood of not being depressed.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
High sensitivity is one goal of a screening questionnaire, to ensure that few persons with the condition are missed. The 74% sensitivity obtained in this study is satisfactory, although not as high as we hoped. Nevertheless, screening for depression helps identify many parents in need of help, albeit not all, and seems preferable to not screening at all. It is noteworthy that depression is masked frequently and may not be apparent in brief encounters.20,25 Another consideration is that those who do not disclose a problem may not be at a stage where they are ready to address it.35 Therefore, although high sensitivity is the goal, it may be that those who are missed are least likely to engage in treatment. It is possible that a seed may be sown simply through questioning and parents may appreciate pediatricians' interest in the parents' own health. Over time, parents may learn to recognize pediatricians as a resource and may confide more in them. It seems that pediatricians generally have good rapport with parents, and many parents are comfortable discussing their problems with pediatricians.23 This should be helpful in facilitating the receipt of services parents may need. Indeed, parents may be grateful for the attention shown to them and may be more satisfied with their children's pediatric care.
There is also a concern regarding false-positive results, unduly "labeling" a parent as depressed. Indeed, the 36% positive predictive value with use of a positive response to either question indicates that only one third of those who screened positive had moderate or worse depressive symptoms. It is critical to distinguish between a screen and a diagnosis. Pediatricians and parents need to appreciate the difference; the screen only identifies a possible problem. Additional evaluation is clearly warranted. Pediatricians can be taught to assess parental depression briefly and to clarify whether a mental health evaluation seems indicated. Those working closely with social workers or other mental health professionals may be able to assess the situation readily in the practice setting. Alternatively, a referral to a community mental health resource may be needed. In these ways, concerns regarding possible erroneous labeling can be minimized.
In order to comprehensively assess a screening test’s performance, one needs to examine its sensitivity, specificity, and positive and negative predictive values. There are inevitable compromises depending where one sets the threshold for "positive." When positive was defined as a yes to either screening question, the rather high specificity indicates many true-negatives and few false-positives. The 74% sensitivity indicates that three quarters of those with depressive symptoms were detected (true-positives), but one quarter were missed (false-negatives). The 95% negative predictive value adds that most of those who screened negative were identified correctly as not being depressed.
It is likely that, for some parents, inquiries into their own health in a pediatric clinic may be surprising and may seem intrusive. In the course of this study, however, we were not aware of any objections to the screening questions. Responses were voluntary and parents could choose, for a variety of reasons, not to disclose their problems. A study found that mothers were more likely to discuss depressive symptoms with family members and friends than with a pediatrician, because of mistrust and fear of being judged, although they were receptive to doing so if they thought that the pediatrician "knew them well."36
In addition to identifying possible depression, the screen may be valuable in signaling to parents that their own health is of interest to pediatricians. The focus in pediatric practice has long been on the child, with relatively scant attention to problems such as maternal depression.22 Attention to the "New Morbidity,"37 the Bright Futures program,24 and the American Academy of Pediatrics38 have drawn attention to the importance of considering the influence of children's social environment on their health.
The challenge remains how to advance pediatric practice to take advantage of the opportunities offered by well-child visits to identify and to begin to address parental depression. The time allocated for routine visits is limited,39 and pediatricians have much to cover. Addressing problems such as parental depression requires some time. Pediatricians can be trained to do this efficiently, with a few key questions. With practice, pediatricians likely can become competent and comfortable addressing such problems. Our training included brief specific information to access community resources, as well as parent handouts. A larger issue is raised, however. If the time available for these visits remains so limited, then there is a need to reconsider how best to take advantage of the opportunity and to prioritize areas in which pediatricians can be most helpful and can make an important difference in children's health and well-being. This requires rethinking some aspects of pediatric primary care that have become ritualized but may have little value.
In addition to training and commitment, pediatricians need valid practical tools. The depression screen in this study was included in a brief PSQ aimed at identifying several psychosocial problems. The use of such questionnaires has been found to be useful and efficient in pediatric practice.2,5 The questionnaires can be filled out by parents while they are waiting for their child to be seen. The pediatrician can peruse the questionnaire quickly and focus on problem items.
Several questions remain for future research. What are the best methods for screening? One study found that more mothers reported depressive symptoms on a paper-based screen, compared with the same questions in a scripted interview (22.9% vs 5.7%).5 How can sensitive new questions be introduced into the well-child visit? We began the PSQ by explaining that we were addressing problems facing many families and that we wished to help and to make sure that children were safe, building on the concern for children's safety that has long been part of well-child care. Accordingly, the first questions were about poison control and smoke alarms.
One key question related to screening is what threshold to set. On one hand, one wishes to identify those likely to benefit. On the other hand, casting the net too widely results in many false-positive results and associated problems. We decided to set the criterion for validating the screen at moderate to severe depression. It seems reasonable that these likely would be good candidates to identify for additional evaluation and possible treatment. When we examined the sensitivity by using a lower threshold that included mild depression (score of >13), the sensitivity was only 62%.
What are the best questions to ask? The questions in the present study worked well and are similar to those recommended by the US Preventive Services Task Force, but additional research is needed, and it is likely that wording may vary slightly to be culturally sensitive for different populations. What are the best times to ask these questions? We planned to do so only periodically, but for logistic reasons it was easiest to have parents of preschool-aged children complete the PSQ at each well-child visit. There were no complaints, but it may be preferable to ask these questions less frequently or to intersperse these questions with others addressing different issues. Most importantly, there is a need to study whether screening leads to parents in need receiving appropriate services and experiencing improvement in their depression. We are examining this issue currently.
There are signs that pediatricians may be ready to embrace some changes to practice. Olson et al5 found that 4 community pediatric practices were able to incorporate screening for depression in well-child visits. In a survey of 508 community pediatricians, 57% felt responsible for recognizing maternal depression and 21% were willing to modify their approach to identification.22 The survey also found that inadequate training (65%) and time to assess the problem (70%) were major barriers. Training is clearly needed to help pediatricians address parental depression.26 Still, Olson et al22 found that 43% of pediatricians did not feel responsible for identifying maternal depression; more needs to be known about their resistance, and education is needed to demonstrate the importance of this problem.
Our study has several limitations. It was conducted in an urban, low-income, mostly black population, necessitating caution about applicability to other groups. Even within our clinic, we found that those who participated were more likely to report depressive symptoms than those not in the study. The sensitivity and specificity should be similar in other settings. The predictive values, however, may vary in other populations.
The PSQ was first administered in a primary care clinic and then repeated with the study protocol in a self-administered computerized interview. The variation in methods could contribute to differing responses and precludes strict examination of test-retest reliability. Despite this, the screening questions proved to be quite stable over time.
Ideally, the screen would have been validated against a standard diagnostic measure such as a comprehensive clinical interview. We lacked the resources to do so, relying instead on the BDI II, a well-standardized, lengthier screen for depression. It should be noted that this measure has performed well in comparison with clinical interviews.28 In addition, Field et al40 found that depressive symptoms, in contrast to a diagnosis of depression, were more frequently related to problematic mother-child interactions.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Address correspondence to Howard Dubowitz, MD, MS, Department of Pediatrics, University of Maryland School of Medicine, 520 W Lombard St, Baltimore, MD 21201. E-mail: hdubowitz{at}peds.umaryland.edu
The authors have indicated they have no financial relationships relevant to this article to disclose.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  J. Stevens, K. J. Kelleher, W. Gardner, D. Chisolm, J. McGeehan, K. Pajer, and L. Buchanan Trial of Computerized Screening for Adolescent Behavioral Concerns Pediatrics, June 1, 2008; 121(6): 1099 - 1105. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Dubowitz, L. Prescott, S. Feigelman, W. Lane, and J. Kim Screening for Intimate Partner Violence in a Pediatric Primary Care Clinic Pediatrics, January 1, 2008; 121(1): e85 - e91. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Dubowitz, W. Lane, and S. Feigelman Implementing Maternal Depression Screening: In Reply Pediatrics, August 1, 2007; 120(2): 449 - 450. [Full Text] [PDF]
|
|
|
|
|
|
K. J. Kemper, K. Kelleher, and A. L. Olson Implementing Maternal Depression Screening Pediatrics, August 1, 2007; 120(2): 448 - 449. [Full Text] [PDF]
|
|
|
|
 |
|
 Simultaneous Screening for Parental Depression and Psychosocial Risk Journal Watch Pediatrics and Adolescent Medicine, May 16, 2007; 2007(516): 3 - 3. [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
||
|
||
Home | My Pediatrics | Journal Information | Current Issue | Past Issues | Subscriptions & Services | Contact Us Click here for faster international access © 2007 American Academy of Pediatrics. All rights reserved. |
||
|
||
