Improving Child and Parent Mental Health in Primary Care: A Cluster-Randomized Trial of Communication Skills Training
OBJECTIVE. We examined child and parent outcomes of training providers to engage families efficiently and to reduce common symptoms of a range of mental health problems and disorders.
METHODS. Training involved three 1-hour discussions structured around video examples of family/provider communication skills, each followed by practice with standardized patients and self-evaluation. Skills targeted eliciting parent and child concerns, partnering with families, and increasing expectations that treatment would be helpful. We tested the training with providers at 13 sites in rural New York, urban Maryland, and Washington, DC. Children (5–16 years of age) making routine visits were enrolled if they screened “possible” or “probable” for mental disorders with the Strengths and Difficulties Questionnaire or if their provider said they were likely to have an emotional or behavioral problem. Children and their parents were then monitored for 6 months, to assess changes in parent-rated symptoms and impairment and parent symptoms.
RESULTS. Fifty-eight providers (31 trained and 27 control) and 418 children (248 patients of trained providers and 170 patients of control providers) participated. Among the children, 72% were in the possible or probable categories. Approximately one half (54%) were white, 30% black, 12% Latino, and 4% other ethnicities. Eighty-eight percent (367 children) completed follow-up monitoring. At 6 months, minority children cared for by trained providers had greater reduction in impairment (−0.91 points) than did those cared for by control providers but no greater reduction in symptoms. Seeing a trained provider did not have an impact on symptoms or impairment among white children. Parents of children cared for by trained providers experienced greater reduction in symptoms (−1.7 points) than did those cared for by control providers.
CONCLUSION. Brief provider communication training had a positive impact on parent mental health symptoms and reduced minority children's impairment across a range of problems.
In both developed and developing countries, 12% to 20% of children have mental health conditions severe enough to warrant treatment.1 Many are seen in primary care settings but receive limited help.2,3 Efforts to improve the management of mental health conditions in primary care settings have centered on helping practices implement evidence-based protocols for specific disorders.4–6 These steps alone, however, may not be sufficient.
First, in community and primary care surveys, the prevalence of children who do not meet the criteria for a disorder but have clinically significant impairment (referred to as mental health “problems” in the Diagnostic and Statistical Manual for Primary Care)7 ranges from equal to to twice the prevalence of children with disorders.8–10 Identifying and helping these children may prevent progression to a disorder. Second, currently available primary care mental health resources may be underused. In one study, only 41% of parents with psychosocial concerns about their children chose to disclose them at pediatric visits.11 In another study, only 61% of mental health referrals made from primary care practices were completed.12 These failures of communication and engagement affect minority children disproportionately. Their parents report less-satisfactory interactions with primary care providers and less-positive expectations of mental health treatment.13–15 Finally, disorder-specific treatment may or may not address cooccurring parent distress, a critical factor both for engagement of families in treatment and for child mental health outcomes.16,17
Therefore, in addition to identifying and treating specific mental disorders, primary care providers need skills that are broadly applicable to eliciting parent and child mental health concerns, partnering with families to find acceptable forms of treatment (potentially including treatment by the provider, additional diagnostic steps, or referral), and increasing the expectation that treatment will be helpful. Strong evidence from studies of mental health care supports the ability of these sorts of interventions to improve outcomes.18 We developed a set of skills that address these needs, and we tested the impact of teaching primary care providers to use the skills during routine visits. Our primary hypothesis was that visits with trained providers would favorably affect children's mental health symptoms and impairments in carrying out daily social, home, and school functions. We also hypothesized that the training's emphasis on engagement could result in greater effects on symptoms and functioning for racial/ethnic minority children. In addition, because pediatricians spend considerable time interacting with parents19 and because parent mental health status and child mental health status are closely related,20 we hypothesized that visits with trained providers would reduce parents' mental health symptoms, albeit to a greater extent among parents whose children were less severely impaired.
This cluster-randomized trial began with recruitment of practice sites. Within sites, providers were assigned randomly to receive training or to be control subjects. Next, children making routine visits were screened, to identify those who possibly or probably had a mental disorder. Identified children and their accompanying parent were monitored for 6 months, to compare changes in symptoms and impairment (children) and distress (parents) among those cared for by trained or control providers. The study was approved by the institutional review boards of the Johns Hopkins Bloomberg School of Public Health and the study sites.
The training contained communication skills that built on providers' existing knowledge of child behavior and development. Skills were drawn from patient-centered care,21 family therapy,22 solution-focused cognitive therapy,23,24 motivational enhancement,25,26 and family engagement (Table 1).27
Training methods mixed structure with opportunities for active learning and locally relevant content.28 Training was delivered in 3 cycles, spaced 3 weeks apart.29 Each cycle began with a 60-minute, small-group discussion, held on-site during work hours and led by a child psychiatrist. The discussion used a computerized slide set containing background material, written and video examples of skills, and programmed pauses for interaction. Audiotapes of the discussions monitored the psychiatrist's fidelity to planned content.
Immediately after each discussion, providers practiced skills in a 10-minute, standardized, patient visit. To reduce stress, providers were told the challenges the standardized patients would present and were encouraged to consult their notes, if desired. Videotapes of these visits were returned to the providers for guided self-assessment.30 The total time for training was ∼4 hours (copies of the training materials are available from the corresponding author). Control providers were aware of their status. They received brief feedback from 2 standardized patient visits and a written manual that paralleled the psychiatrist-led training.
Sites were recruited by approaching directors of practices typifying the range of payers, locations, and structures seen in pediatric primary care in the United States.31,32 Thirteen of 16 directors approached agreed to participate. The 3 who declined cited lack of time, imminent staff turnover, or vacancy in the lead clinician position. Participating rural sites (n = 7) included a solo pediatric practice, a hospital-based, multimember, pediatric practice, 4 freestanding multispecialty offices, and a practice staffed by 2 family nurse practitioners. Urban sites (n = 6) included 3 community medical centers (2 multispecialty and 1 pediatric), a group private pediatric practice, a hospital-based family practice, and a community center serving Latino immigrants. No site had formal arrangements with psychiatrists or psychologists, and all cared for patients with a mixture of insurance types. Sites received no compensation, but providers received continuing education credits.
Provider Participants and Randomization
At each site, we attempted to recruit all full-time providers routinely involved in pediatric primary care. Before random assignment, providers reported on their training and length of time at the site. They also completed the 14-item Physician Belief Scale,33,34 which measures attitudes toward treating psychosocial problems. Higher scores reflect more-negative attitudes and are correlated inversely with patients' disclosure of psychosocial information.34,35 Within sites, providers were assigned randomly by placing identifiers in a randomly ordered list and tossing a coin to determine the first assignment. Subsequent providers were given alternating status. Imbalance was resolved in favor of training, to ensure that there were ≥2 participants in each provider training session.
Provider training was staggered across sites. Immediately after training was completed at any given site, patients at the site were recruited for evaluation. Recruiting thus took place between December 2002 and August 2005 but lasted only 6 to 8 weeks for any given site. As families came for a child's care (health maintenance, acute problem, or follow-up visit), they were approached and assessed for initial eligibility (age of 5–16 years, visiting a participating provider, and not acutely ill). Written consent was obtained at this point; parents and children were told that providers were participating in training for the care of children's behavioral and mood problems and help was sought to evaluate its impact. They were not told their provider's training status. The recruiter at the Latino center was a native Spanish speaker. Families received $15 for screening and $15 for follow-up evaluation.
Next, parents completed the Strengths and Difficulties Questionnaire (SDQ), a brief survey of child emotional and behavioral symptoms and associated functional impairment that has been validated in several languages.36,37 For children 5 to 10 years of age, parents consented to completion of a parallel SDQ by the child's teacher; youths 11 to 16 years of age completed a parallel report of their own. After the visit, parent/teacher and parent/youth reports were combined by using an algorithm that classified the child as unlikely to have, possibly having, or probably having ≥1 disorder of attention, conduct, or emotions.38 Providers were not given these results. Children were eligible if they scored as possible or probable or if they were rated by their primary care provider as being likely to have an emotional or behavioral problem at the time of the index study visit.
Parents reported on their child's use of mental health services in the previous 6 months by using questions from the Methods for the Epidemiology of Child and Adolescent Mental Disorders study.39 Parents supplied sociodemographic data regarding themselves and their children, including age, gender, race/ethnicity, education, and insurance.
Data Collection and Scoring
Research assistants who were masked with respect to provider training collected all parent and child data. Outcomes were change scores, computed by subtracting the 6-month follow-up values of measures from the values at enrollment.
Child Symptoms and Functional Impairment
We computed total symptom (range: 0–40) and impairment (range: 0–10) scores from the parent-rated SDQ (in contrast to the use of the SDQ for study eligibility, which combined symptom and impairment ratings from 2 observers to yield a categorical likelihood of disorder). In the US National Health Interview Survey, parent-rated SDQ symptom scores of ≥16 identified the most symptomatic 9% of children; parent-rated SDQ total impairment scores of ≥2 identified the most impaired 8%.40
Parent Mental Health Symptoms
We assessed parent symptoms with the 28-item General Health Questionnaire (GHQ).41 Possible GHQ scores range from 0 to 28, with scores of ≥5 being predictive of psychiatric disorders. The GHQ has been used internationally and is available in both English and Spanish. Its items cover anxiety, depression, social dysfunction, and somatic concerns often associated with changes in mood.42
We conducted patient-level, intent-to-treat analyses, maintaining participants in the intervention group of the provider they had seen at the time of enrollment. Because this was a group-level randomized trial with a relatively small number of groups, we controlled for baseline differences that might affect outcomes or function as confounders.43 To test the effect of the intervention, we estimated multivariate linear regression models with before/after change scores as the dependent variable, intervention status as the independent variable, and baseline symptom or function scores plus potential confounders and factors associated with nonresponse as covariates. Hypothesized differential effects of the intervention according to ethnicity and child mental health status were explored by inserting appropriate interaction terms in the regression models. We used the Stata 8 command “xtgee” to adjust confidence limits for clustering of children and parents within providers (Stata, College Station, TX). We calculated effect sizes to compare the impact of training with other child mental health interventions. Effect sizes reflect the ratio of change attributed to an intervention to the size of random variation in the outcome.44 This ratio then becomes a way of comparing changes attributable to interventions with similar goals but different outcome measures.45 The estimated size of the sample at follow-up evaluations, based on Monte Carlo simulations and a range of possible intracluster and intercluster correlations, was ∼350 to detect effect sizes ranging from 0.2 to 0.4 by using 2-tailed tests with α = .05 and power = .80.46
Of 69 providers approached for participation, 58 (84%) agreed; 27 became control providers and 31 were trained. Two control providers left after patients were enrolled but are included in the evaluation. All trained providers completed the study. To increase minority patient enrollment, 3 control providers subsequently received full training and were assessed with a second, unique cohort of children 1 year after follow-up evaluation of their first cohort ended. These 3 providers were selected on the basis of their practice site (availability of previously unrecruited, minority patients), before assessment of study outcomes.
Providers had practiced an average of 15 years since obtaining their degree (range: 2–40 years; median: 13 years); 85% had been at their site for >1 year, and 60% were female. Most were trained in pediatrics (60%) or family practice (38%). Most (83%) were physicians, 16% were nurse practitioners, and 1 was a physician's assistant. Approximately one fourth (21%) had additional training in child behavior, and 7% had training in development or counseling. Table 2 compares baseline characteristics of trained and control providers.
Child and Parent Participants
Recruiters approached 875 families to determine eligibility. Of these, 28 patients were not patients of participating providers, 507 were patients of trained providers, and 340 were patients of control providers (Fig 1); 39 (5%) declined, were ineligible, or agreed but did not complete screening. Of 808 with completed SDQs, 337 (42%) were rated possible or probable and 130 were rated unlikely but were nominated by their provider as having an emotional or behavioral problem. Of the 467 children eligible for follow-up monitoring, 49 parents (10%) declined further participation or could not be contacted after the index visit, leaving 418 patients in the follow-up group (248 having seen trained providers and 170 having seen control providers). Of the 418 patients, 46% were female. The mean age was 10.4 years. Just under one half were members of minority racial or ethnic groups (30% black, 12% Latino, and 4% other or declined to self-identify). Approximately one half (47%) were from rural sites. A similar proportion (46%) had private insurance. One half (49%) had received some form of counseling outside primary care in the previous 6 months. On the basis of SDQ results, 39% possibly or probably had a disorder related to hyperactivity, 51% to conduct, and 31% to emotions.
Accompanying parents/guardians ranged from 16 to 71 years of age (median: 36 years; mean: 37.3 years). Almost all (91%) were female, and most (74%) had graduated high school. By parent report, just under one half (49%) of index visits were for health maintenance, 26% for acute medical concerns, 11% for emotional or behavioral problems, and 14% for other medical follow-up care.
Table 3 shows baseline characteristics of children and parents seeing trained and control providers. Children seeing trained providers were more likely to be female and black; they were slightly older and were less likely to be from rural sites.
Of the child characteristics that differed between trained and control providers, race/ethnicity, residence, and age had the characteristics of potential confounders for study outcomes. Across intervention and control groups, minority race/ethnicity was associated with an increase in parent symptoms during the study period, relative to the change among white parents (mean difference: 1.8 GHQ points; 95% confidence interval [CI]: 0.88–2.8; P < .001). Relative to urban residence, rural residence was associated with greater decreases in parent symptoms (mean difference: −2.1 GHQ points; 95% CI: −3.1 to −1.2 points; P < .001) and child symptoms (mean difference: −3.1 points on the SDQ symptom scale; 95% CI: −4.8 to −1.5 points; P < .001). Older children tended to have greater decreases in symptom scores (mean difference: −0.18 points on the SDQ symptom scale per year of age; 95% CI: −0.36 to −0.01 points per year of age; P = .036).
Child and Parent Follow-up Monitoring
Of the 418 children in the evaluation, 367 (88%) were monitored for 6 months. There was not a significant difference in follow-up rates between children seen by trained providers (89%) and those seen by control providers (86%) (Fig 1). Children who were retained in the follow-up group were more likely to have been seen by family physicians (25%), compared with children who were lost to follow-up monitoring (8%; P = .002); retained children were more likely to have been seen at a rural site (47%) than were those lost to monitoring (35%; P = .065). Parents of retained children were on average older than parents of those lost to monitoring (37.3 vs 34.4 years; P = .01) and were more likely to be female (91% vs 87%; P < .01). Neither parents nor children lost to follow-up monitoring differed significantly in baseline clinical measures from those retained in the study.
Child and Parent Outcomes
Tables 4 and 5 show unadjusted changes in children's symptoms and impairment and in parents' symptoms, stratified according to race/ethnicity. Examining the direction of these changes illustrates the different impact of the training for minority and white families. For all children, mean symptom scores tended to decrease (improve) during the study period, regardless of provider training (Table 4). This was also true for mean impairment scores for white children. In contrast, mean impairment scores for black and Latino children seeing control providers tended to increase (worsen), whereas impairment scores for minority children seeing trained providers decreased.
A similar diminishing of race/ethnic disparity was seen for parent symptoms (Table 5). Symptom scores among white, non-Latino parents tended to decrease regardless of provider training. Symptom scores among black and Latino parents tended to increase, but to a smaller extent among parents seeing trained providers.
Multivariate analyses controlling for baseline mental health status and for confounders (child age, ethnicity, and rural/urban residence) further explain the unadjusted results for the 3 study outcomes. Regarding child symptoms, seeing a trained provider did not have an impact on change during the study, and there was no statistical evidence that seeing a trained provider had different effects on minority versus white children. Regarding child impairment, as suggested by the unadjusted results, there was a statistically significant interaction between minority race/ethnicity and provider training. Minority children (black and Latino combined) seeing trained providers had a mean decrease in impairment of 0.91 points more (95% CI: −1.8 to −0.01 points; P = .049), compared with white children. To examine further how race/ethnicity modified the relationship of training and impairment, we conducted separate regression analyses for black and Latino children. For black children, seeing a trained provider was associated with a significantly greater decrease in impairment compared with children seeing control providers (mean difference: −1.1 points; 95% CI: −0.24 to −2.0 points; P = .012; effect size: 1.1/2.5 = 0.44). For Latino children, there was a trend toward greater reduction among those seeing a trained provider (treatment/control difference in impairment changes for Latino children: −0.40; 95% CI: −1.5 to 0.66; P = .46).
Regarding parent symptoms, seeing a trained provider was associated with a significantly greater decrease, compared with seeing a control provider (on average, −1.7 GHQ points; 95% CI: −3.2 to −0.11 points; P = .035; effect size: 1.7/5.4 = 0.31). Although it was suggested by the unadjusted results, there was no statistical evidence that seeing a trained provider had different effects on minority versus white parents. However, the impact on parents of seeing a trained provider did vary with their child's mental health status. When the analysis was performed separately for children with different SDQ ratings at enrollment, seeing a trained provider was associated with a significantly greater decrease in parent symptoms for children rated as possibly having a mental disorder (mean difference for parents seeing trained versus control providers: −1.6 GHQ points; 95% CI: −2.9 to −0.25; P = .019) but not for children rated as probable (mean difference: −0.70 GHQ points; 95% CI: −2.6 to 1.2; P = .47) or unlikely (mean difference: 0.79 GHQ points; 95% CI: −0.95 to 2.5; P = .38).
Brief primary care provider training reduced mental health symptoms among parents of children with emotional and behavioral disturbances and led to reduction in impairment among minority children. We think that this is the first study to show that broadly applicable communication skills training can affect parent and child outcomes in pediatric care.
Previous work in adult primary care showed that teaching doctors emotion-handling skills could reduce patients' emotional distress.47 The skills taught in our training address both eliciting emotions and addressing negative ones with specific techniques. Their impact was greatest for parents of children rated possible for a disorder with the SDQ, similar to the Diagnostic and Statistical Manual for Primary Care “problem” category. Parents of more severely ill children (those likely to have a disorder) probably need more-intensive support and specific treatment for their child. Reduced distress among parents of children in the possible category, however, could help prevent progression to the disorder level.
The impact of training on parent distress and child functioning was of similar magnitude as that reported for more-complex interventions. A meta-analysis of the impact on maternal psychosocial health of parenting programs for children with behavioral problems found average effect sizes of 0.3 for parents' depressive symptoms and 0.5 for parents' anxiety and stress.48 A trial of primary care treatment for depressed teens found effect sizes of 0.23 for overall mental health and 0.24 for depression.6 The effect sizes we observed may have particular public health significance because they represent changes in a heterogeneous population of parents and children, rather than changes among individuals with a specific disorder.
The training program demonstrated potential for dissemination among community-based providers. Despite initial apprehension about practicing with standardized patients, >80% of those approached took part once the lead clinician at a site agreed, and none assigned randomly to training dropped out. Several factors in the training design contributed to engagement; short sessions built into the workday reduced opportunity costs and allowed trainees to try new skills immediately. Training materials balanced a consistent core of information with attention to site-specific interests33; training content was delivered in multiple formats so that participants could access it in ways that most suited their learning styles and time constraints.
Training might have differentially benefited minority children by helping providers address minority parents' low expectations of mental health treatment. A randomized trial of quality improvement for primary care treatment of adult depression also found greater benefit for minority patients. The investigators of that study hypothesized that their program, although not focused directly on communication skills, increased providers' ability to elicit and to tailor their responses to patient preferences and concerns.49,50
Our study has several limitations. Cluster randomization and our approach to increasing minority enrollment resulted in treatment and control groups that were unbalanced with regard to factors associated with our primary outcomes, and adjustment was required to create unbiased models.43 Control providers received feedback and written training materials. To the extent that these interventions influenced provider behavior, our estimates of the full training impact may be conservative. Our heterogeneous population of providers and families demonstrated variation in response to the training but was not large enough for us to explore the many possible combinations of parent, child, and provider characteristics likely to influence outcomes.51 A stronger test of clinical utility would use diagnostic instruments and longer follow-up periods, but the SDQ and GHQ allowed us to simulate the impact of training on a primary care practice that screened routinely for emotional and behavioral problems with commonly used instruments.52
The SDQ is a readily used and powerful instrument. To establish eligibility for the study, we used a scoring method that combined symptom and impairment ratings from 2 observers to yield a trichotomous likelihood (unlikely, possibly, or probably) of having ≥1 mental disorder. This ensured that we would include children representing the full range of severity seen in practice and not just those likely to have a disorder. We then used the SDQ parent symptom and impairment scales separately (as in the US National Health Interview Survey47) to explore these 2 distinct clinical outcomes, both of which are associated independently with mental health service utilization.9 We cannot rule out the possibility that improvement in impairment among minority children was an artifact of decreased parent symptoms.53 We also cannot speak to the longer-term impact of provider training on children who seem unlikely to have a mental disorder, on the basis of both SDQ and provider ratings. Providers overly dwelling on mental health could prove aversive to families without these concerns. We did, however, gather reports of satisfaction with the index visit from all parents of screened children. Among parents of children with no identifiable risk of having a mental health problem, there was no statistically significant difference between those seeing trained providers and those seeing control providers in willingness to discuss mental health or in the parent feeling that she had accomplished her goals for the visit. Those seeing trained providers were more likely to report having been asked for their opinion in developing treatment plans.
Brief provider training in broadly focused communication skills improved parent symptoms and reduced minority children's impairment associated with mental health problems. Although much work remains to replicate these outcomes, to increase their degree and breadth, and to study their duration, our results support the potential to develop a readily learned communication skill set to expand help for children's mental health problems beyond treatment for specific disorders.
The study was funded by grant RO1MH62469 from the National Institute of Mental Health. Dr Brown's work was supported by predoctoral National Research Service Award F31MH75531.
We are grateful to the families and clinicians who took part in the study. Dr Barry Lewis helped refine the training methods and Drs Scott Krugman, Gaurav Kumar, and Steve Shearer provided guidance about training family physicians. Alex Suchman helped organize data collection at the outset of the work. Anonymous reviewers provided many helpful comments about analysis and presentation of the data.
- Accepted July 23, 2007.
- Address correspondence to Larry Wissow, MD, MPH, Johns Hopkins School of Public Health, 624 N. Broadway, Room 703, Baltimore, MD 21205. E-mail:
The authors have indicated they have no financial relationships relevant to this article to disclose.
This work was presented in part at the annual meeting of the Pediatric Academic Societies; April 29–May 3, 2006; San Francisco, CA.
Dr Brown's current affiliation is Mathematica Policy Research, Washington, DC.
Dr Luo's current affiliation is Department of Biostatistics, University of Minnesota School of Public Health, Minneapolis, MN.
- ↵Costello EJ, Egger H, Angold A. 10-year research update review: the epidemiology of child and adolescent psychiatric disorders, part I: methods and public health burden. J Am Acad Child Adolesc Psychiatry.2005;44(10) :972– 986
- ↵Kelleher KJ, McInerny TK, Gardner WP, Childs GE, Wasserman RC. Increasing identification of psychosocial problems: 1979–1996. Pediatrics.2000;105 (6):1313– 1321
- ↵Foy JM, Earls MF. A process for developing community consensus regarding the diagnosis and management of attention-deficit/hyperactivity disorder. Pediatrics.2005;115 (1). Available at: www.pediatrics.org/cgi/content/full/115/1/e97
- Campo JV, Shafer S, Strohm J, et al. Pediatric behavioral health in primary care: a collaborative approach. J Am Psychiatr Nurses Assoc.2005;11 (5):276– 282
- ↵Wolraich ML, Felice ME, Drotar D, eds. The classification of child and adolescent mental diagnosis in primary care. In: Diagnostic and Statistical Manual for Primary Care (DSM-PC), Child and Adolescent Version. Elk Grove Village, IL: American Academy of Pediatrics; 1996:11
- ↵Costello EJ, Shugart MA. Above and below the threshold: severity of psychiatric symptoms and functional impairment in a pediatric sample. Pediatrics.1992;90 (3):359– 368
- ↵Briggs-Gowan MJ, Owens PL, Schwab-Stone ME, Leventhal JM, Leaf PJ, Horwitz SM. Persistence of psychiatric disorders in pediatric settings. J Am Acad Child Adolesc Psychiatry.2003;42(11) :1360– 1369
- ↵Minkovitz CS, Strobino D, Scharfstein D, et al. Maternal depressive symptoms and children's receipt of health care in the first 3 years of life. Pediatrics.2005;115 (2):306– 314
- ↵Weissman MM, Pilowsky DJ, Wickramaratne PJ, et al. Remissions in maternal depression and child psychopathology: a STAR*D-child report. JAMA.2006;295(12) :1389– 1398
- ↵Stewart M, Brown JB, Weston WW, McWhinney IR, McWilliam CL, Freeman TR, eds. Patient-Centered Medicine: Transforming the Clinical Method. Thousand Oaks, CA: Sage Publications; 1995:216–228
- ↵Allmond BW Jr, Tanner JL, Gofman HF. The Family Is the Patient: Using Family Interviews in Children's Medical Care. 2nd ed. Baltimore, MD: Williams & Wilkins; 1999
- ↵Walter J, Peller J. Becoming Solution-Focused in Brief Therapy. New York, NY: Brunner/Mazel; 1992
- ↵Miller WR, Rollnick S. Motivational Interviewing: Preparing People to Change Addictive Behavior. New York, NY: Guilford Press; 1991
- ↵Rubenstein LV, Mittman BS, Yano EM, Mulrow CD. From understanding health care provider behavior to improving health care: the QUERI framework for quality improvement: Quality Enhancement Research Initiative. Med Care.2000;38(6 suppl 1) :I129– I141
- ↵Woodwell DA, Cherry DK. National Ambulatory Medical Care Survey: 2002 summary. Adv Data.2004;(346) :1– 44
- ↵Ashworth CD, Williamson P, Montano D. A scale to measure physician beliefs about psychosocial aspects of patient care. Soc Sci Med.1984;19(11) :1235– 1238
- ↵McLennan JD, Jansen-McWilliams L, Comer DM, Gardner WP, Kelleher KJ. The Physician Belief Scale and psychosocial problems in children: a report from the Pediatric Research in Office Settings and the Ambulatory Sentinel Practice Network. J Dev Behav Pediatr.1999;20 (1):24– 30
- ↵Leaf PJ, Alegria M, Cohen P, et al. Mental health service use in the community and schools: results from the four-community MECA study: Methods for the Epidemiology of Child and Adolescent Mental Disorders study. J Am Acad Child Adolesc Psychiatry.1996;35 (7):889– 897
- ↵von Korff M, Üstün TB. Methods of the WHO collaborative study on psychological problems in general health care. In: Üstün TB, Sartorious N, eds. Mental Illness in General Health Care: An International Study. Chichester, England: Wiley; 1995:19–38
- ↵Murray DM. Design and Analysis of Group-Randomized Trials. New York, NY: Oxford University Press; 1998
- ↵Cohen J. Statistical Power Analysis for the Behavioral Sciences. 2nd ed. Hillsdale, NJ: Erlbaum; 1988
- ↵Slymen DJ, Hovell MF. Cluster versus individual randomization in adolescent tobacco and alcohol studies: illustrations for design decisions. Int J Epidemiol.1997;26 (4):765– 771
- ↵Roter DL, Hall JA, Kern DE, Barker LR, Cole KA, Roca RP. Improving physicians' interviewing skills and reducing patients' emotional distress: a randomized clinical trial. Arch Intern Med.1995;155(17) :1877– 1884
- ↵Barlow J, Coren E, Stewart-Brown S. Meta-analysis of the effectiveness of parenting programmes in improving maternal psychosocial health. Br J Gen Pract.2002;52(476) :223– 233
- Copyright © 2008 by the American Academy of Pediatrics