Pediatricians and public health researchers presented their most recent results from the Children's Sentinel Nutrition Assessment Program (C-SNAP) in San Francisco at the Pediatric Academic Societies Annual Meeting, April 29-May 3.
Topics included:
Dr. Hans Kersten, Clinical Director of The GROW Clinic at St. Christopher's Hospital, also presented two posters with his Drexel colleagues.
Does Safety Net Program Participation Decrease Impact of Food Insecurity on Developmental Risk Among Low Income Children of Color?
Deborah A. Frank, Ruth Rose-Jacobs, Maureen Black, Patrick Casey, Diana Cutts, Alan F. Meyers, Mariana Chilton, Suzette Levenson, Nicole Neault, Timothy C. Heeren. Boston U Schls of Medicne and Public Hlth, Boston, MA; U of Maryland, Baltimore, MD; U of Arkansas, Little Rock, AR; Hennepin Cty Med Ctr, Minneapolis, MN; Drexel U Schl of Public Hlth, Philadelphia, PA.
BACKGROUND: We have previously shown that safety net programs decrease but do not eliminate the association of household food insecurity (FI) with poor health and growth of young children of color, who in national surveys show higher rates not only of FI but of developmental lag than white children.
OBJECTIVE: To explore the association between FI and developmental risk in African American (AA) and Latino/Hispanic (LH) children and test whether safety net programs attenuate that relationship.
DESIGN/METHODS: From 7/04 to 6/05 the Children
s Sentinel Nutrition Assessment Program interviewed 1320 AA and 487 LH caregivers of children 4-36months without private insurance in 5 pediatric sites (AK, MA, MD,MN,PA), documenting demographics, the US Food Security Instrument, and the Parents Evaluation of Developmental Status (PEDS) (Glascoe, 1998) which classifies children with 
2 significant parental concerns as 
at developmental risk. Childrens weight for age z score was calculated.
RESULTS: Developmental risk was prevalent in both groups (LH, 12%, AA 14%), as was FI (LH 37%, AA 19%). Base models controlling for childs age, gender, low birthweight, current weight for age z score, and mothers education, age, and place of birth (US vs not US) found FI was associated with increased developmental risk among LH (AOR 2.23 95% CI 1.09,4.54, p=.03) and AA children (AOR 1.57, 95% CI 1.02, 2.40, p=.04). Within each group, analyses controlling one at a time for family participation in Food Stamps, WIC, Temporary Assistance to Needy Families (welfare), Housing Subsidies, and Low Income Home Energy Assistance found no attenuation of this association.
CONCLUSIONS: The association between household food insecurity and young childrens developmental risk occurs in both AA and LH families. Safety net programs which partially protect health and growth of children of color in FI families, are not sufficient to decrease this association in this small sample, but larger studies are needed.
Stunting Is Associated with Poor Health Among Low-Income Children
Maureen M. Black, Deborah A. Frank, Suzette Levenson, Patrick Casey, Mariana Chilton, Diana Cutts, Alan Meyers, Timothy Heeren, John Cook, Nicole Neault, Carol Berkowitz. Pediatrics, Univerity of Maryland, Baltimore, MD; Pediatris, Boston Medical Center, Boston, MA; Boston Univ. School of Public Health, Boston, MA; Univ. of Arkansas for Medical Sciences, Little Rock, AK; Drexel Univ. School of Public Health, Philadelphia, PA; Hennepin County Medical Center, Minneapolis, MN; Harbor-UCLA Medical Centr, Los Angeles, CA.
BACKGROUND: Stunting, length-for-age < -2 z-scores (< 2.28 percentile), early in life is a sign of chronic undernutrition and predicts subsequent cognitive and academic deficits.
OBJECTIVE: To evaluate the prevalence and correlates of stunting among low-income children < 36 months of age.
DESIGN/METHODS: From 2000 - 2005, the Childrens Sentinel Nutritional Assessment Program gathered weight and length on 14,562 children ages 1-36 months at 6 pediatric clinic/emergency departments (AK, CA, MA, MD, MN, PA). Weight-for-age and length-for-age z-scores were calculated using CDC norms. Caregivers were interviewed regarding demographics, hospitalizations, and perceptions of their childs health.
RESULTS: Overall, 6% (877/14,562) of the children were stunted, 2.64 times the national rate of 2.28%. Stunted children were more likely to be low birth weight (52% vs. 10%, p < .0001), shorter gestational age (35.6 vs. 39.1 weeks, p < .0001), male (58% vs. 53%, p = .004), not breast fed (65% vs. 44%, p < .0001), live in households with unemployed mothers (65% vs. 53%, p < .001) and use public insurance (83% vs. 79%, p = .001). In multivariate logistic regression, adjusting for site, gender, race, birth weight, gestational age, breast feeding history, maternal place of birth, maternal employment, and insurance, stunted children were one-third more likely than non-stunted children to have been hospitalized in the past, OR = 1.36, 95% CI = 1.14, 1.62, p = .0007; 57% more likely to be hospitalized the day of the assessment, OR = 1.57, 95% CI = 1.18, 2.10, p = .002; and 65% more likely to be reported as being in fair or poor health, OR = 1.65, 95% CI = 1.35, 2.04, p < .0001.
CONCLUSIONS: Rates of stunting are higher among low-income children than national norms and stunting is associated with indicators of poor health. Clinicians should track young childrens length as a health indicator.
Household Food Insecurity and Risk for Childrens Developmental Problems
Ruth Rose-Jacobs, Maureen Black, Patrick Casey, Mariana Chilton, John Cook, Diana Cutts, Timothy Heeren, Suzette Levenson, Alan Meyers, Nicole Neault, Deborah A. Frank. Boston U, Boston, MA; U of Maryland, Baltimore, MD; U of Arkansas, Little Rock, AK; Drexel U, Philadelphia, PA; Hennepin County Med. Ctr, Minneapolis, MN.
BACKGROUND: While infant malnutrition has been linked to developmental problems, no previous study has evaluated the relationship of household food insecurity (FI), limited or uncertain availability of adequate or safe foods to development among children under age 3.
OBJECTIVE: To evaluate the relationship between household FI and risk for developmental problems in children 4-36 months of age, after control for potential confounding variables.
DESIGN/METHODS: Between July 2004 and June 2005 the Childrens Sentinel Nutritional Assessment Program interviewed in English, Spanish, and Somali 2,478 caregivers from low income households with children ages 4-36 months at five pediatric clinic/emergency department sites (AK, MA, MD, MN, PA). Interviews included demographics, the US Food Security Scale, and a 10 question caregiver reported developmental screening instrument, the Parents Evaluations of Developmental Status (PEDS) (Glascoe, 1998). The PEDS meets the American Academy of Pediatrics standards for developmental screening and classifies children with 2 significant concerns on the PEDS to be 
at developmental risk
. Children also were weighed and weight-for-age z-score calculated.
RESULTS: Overall, 21% (N=508) of the children lived in food insecure households and 14% (N=341) were developmentally at-risk on the PEDS. In logistic analyses after controlling for interview site, child (gender, age, birth weight, weight-for-age z score) and caregiver (age, US born, depression, education) factors, FI households were two thirds more likely than food-secure households to have children at risk for developmental concerns (AOR=1.63, 95% CI=1.18, 2.24, p=.003) . The relationship between FI and the PEDS was maintained even when families reporting FI with hunger (N=149) were removed from the analysis (AOR=1.69, 95% CI=1.18-2.43, p=.004).
CONCLUSIONS: After controlling for established developmental correlates 4-36 month old children from low income households with even the mildest levels of FI are more likely than those from food secure households to be at developmental risk. Public policies which ameliorate household FI also may improve early child development/later school readiness.
Children of Immigrants: Impact of Food Stamps on Health
Mariana Chilton, Diana Cutts, Carol Berkowitz, Maureen Black, Patrick Casey, John Cook, Suzette Levenson, Alan Meyers, Timothy Heeren, Nicole Neault, Luz Neira, Deborah A. Frank. Drexel Univ. School of Public Health; Hennepin County Medical Center; Harbor-UCLA Medical Center; Boston Medical Center; Univ. of Maryland School of Medicine; Univ. of Arkansas for Medical Sciences; Boston Univ. School of Public Health; St. Marys Center for Maternal and Child Care.
BACKGROUND: 1/5 children < age 6 years in the US has immigrant parents. Though 93% of children of immigrants are US citizens, Food Stamps are not as readily accessible for them as for children of citizens. Implications of differential access for health of these children are unknown.
OBJECTIVE: To investigate associations between Food Stamp participation and family wellbeing among immigrant households.
DESIGN/METHODS: Since 1998, the Childrens Sentinel Nutrition Assessment Program tracked growth, health, and access to public assistance of children < age 3 in emergency departments and health clinics in 7 urban areas. Citizen children of immigrants comprise nearly 40% of the sample (6,973/18,002). Caregivers were interviewed in English, Spanish and Somali on demographics, public assistance, food insecurity (USDA Food Insecurity Scale), child health, hospitalization history, and maternal depression.
RESULTS: Most of the immigrant sample was Mexican, Somali, Central American, and Caribbean. For immigrant households, household food insecurity rates were twice the rate found in citizen households (36% vs 16%, p< .0001). Child food insecurity for citizen children of immigrants was 4 times the rate found among children of non-immigrants (24% vs 6%, p<.0001). Food stamp participation rates differed between immigrant and non-immigrant households (23% vs 48%, p<.0001). Immigrant households receiving food stamps showed slightly higher rates of good/excellent child health (78% vs 75%, p<.0001) and decreased rates of maternal depression (26% vs 30%, p<.0001) when compared to non-participating immigrant households.
CONCLUSIONS: Immigrant households with citizen children experience more food insecurity than non-immigrant households. Although immigrant families infrequently access food stamps, participation in the Food Stamp program is associated with positive health outcomes for the family. Measures to decrease access of immigrant parents to food stamps may increase rates of poor health among our youngest New Americans.
Direct Observation of Competencies (DOCs) in Ambulatory Settings
Hans B. Kersten, Nancy D. Spector, Shareen Kelly, Robert L. Bonner, Harry Salvati, Pierre P. Chanoine, Paul S. Matz, Shoshana Melman, Laura E. Smals, Dan Taylor, Cynthia Warren. Department of General Pediatrics, Drexel University College of Medicine, St. Christophers Hospital for Children, Philadelphia, PA.
BACKGROUND: Direct observation (DO) of resident/patient encounters by faculty and documentation of resident competence in history taking and physical examination skills is now required by the RRC. Many barriers exist for the effective implementation of DO in ambulatory clinics, including lack of validated evaluation tools and faculty time to observe the encounters and provide feedback.
OBJECTIVE: To assess the usefulness of a teaching module series with accompanying evaluation tools designed to evaluate resident history taking and physical examination skills in ambulatory care settings.
DESIGN/METHODS: Modules were created to teach history and physical exam skills. Each module was designed to teach and evaluate a portion of the patient history or physical examination (e.g. HPI or abdominal exam). The evaluation tools were developed for use at point of care by faculty and residents. The 5-part DOC process includes 1) teaching the module to the resident 2) observation of that portion of the patient encounter and 3) confirmation of the findings by the attending 4) completion of the evaluation forms by the attending/resident 5) oral feedback by the attending. Completed evaluations are included in the residents evaluation portfolio. DOCs were introduced to the faculty who precept in two ambulatory clinics at St. Christophers Hospital for Children. Residents and faculty were surveyed after DOCs were introduced to assess their usefulness for faculty and residents.
RESULTS: DOCs were introduced in the ambulatory clinics in November 2005. To date, faculty have found DOCs easy to use, took little time to use and guided their feedback to residents. Residents felt that faculty feedback was specific and helpful in improving their skills, but felt the exercise could be disruptive to patient encounters.
CONCLUSIONS: DOCs are easy to use and provide helpful information for attendings to give meaningful feedback on patient encounters. Residents felt that the feedback was relevent and concise, but could be disruptive to patient flow. DOCs are a useful tool to document resident competency achievement and could be used in other busy clinical venues (e.g. ED, inpatient), and for other types of trainees (e.g. students, fellows).Measuring Competency Through EBM Evaluation
E. Douglas Thompson, Hans B. Kersten, Tahniat S. Syed. Department of Pediatrics, Drexel University College of Medicine 
St. Christophers Hospital for Children, Philadelphia, PA.
BACKGROUND: Evaluation of evidence-based medicine (EBM) projects provides documentation of the Residency Review Committees (RRC) requirement that pediatric residencies must evaluate the competence of residents in performing an evidence-based exercise
based on predetermined criteria. These evaluations can also provide documentation for resident competencies. Limited data is available on methods to evaluate EBM skills.
OBJECTIVE: To assess the validity and reliability of an evaluation tool for use by preceptors of EBM projects.
DESIGN/METHODS: During 3 clinical rotations, residents and medical students completed EBM projects requiring them to frame a clinical question and identify a journal article to answer this question through the use of computerized literature searches (IT) and basic appraisal skills. The authors designed a tool to evaluate 3 content areas of the EBM project. Two existing standardized EBM tests provided a frame of reference. To measure validity, the tool was presented at the 2005 Association of Pediatric Program Directors (APPD) meeting and qualitative feedback from the participants was used to adapt the tool. To measure reliability, the authors independently evaluated anonymous EBM projects. Inter-rater reliability was assessed by measuring the Pearson correlation coefficients between the 3 reviewer pairs.
RESULTS: Participants from the APPD workshop report the tool is easy to use and is strong in measuring the residents abilities in clinical question formation and IT. Evaluation of critical appraisal skills is basic, but consistent with the authors intent. 44 projects were scored. The correlation coefficients between reviewer A and B (AB) for clinical question formation, IT and basic critical appraisal skills are 0.778*, 0.457* and 0.587*, respectively (*p<0.01). The corresponding coefficients are 0.727*, 0.187 and 0.606* between evaluators B and C (BC), 0.806*, 0.122 and 0.502* between evaluators A and C (AC), and 0.649*, 0.830* and 0.635* for the total project scores for AB, BC and AC.
CONCLUSIONS: The evaluation tool demonstrates content validity as well as reliability in the areas of framing a clinical question, basic appraisal skills and the overall project score. The tool is an easy-to-use method that can be used to document competency of EBM skills of residents.