Adolescent immunization practices: a national survey of US physicians.

BACKGROUND: Adolescent immunization rates remain low. Hence, a better understanding of the factors that influence adolescent immunization is needed. OBJECTIVE: To assess the adolescent immunization practices of US physicians. DESIGN AND SETTING: A 24-item survey mailed in 1997 to a national sample of 1480 pediatricians and family physicians living in the United States, randomly selected from the American Medical Association's Master List of Physicians. PARTICIPANTS: Of 1110 physicians (75%) who responded, 761 met inclusion criteria. OUTCOME MEASURES: Immunization practices and policies, use of tracking and recall, opinions about school-based immunizations, and reasons for not providing particular immunizations to eligible adolescents. RESULTS: Seventy-nine percent of physicians reported using protocols for adolescent immunization, and 82% recommended hepatitis B immunization for all eligible adolescents. Those who did not routinely immunize adolescents often cited insufficient insurance coverage for immunizations. While 42% of physicians reported that they review the immunization status of adolescent patients at acute illness visits, only 24% immunized eligible adolescents during such visits. Twenty-one percent used immunization tracking and recall systems. Though 84% preferred that immunizations be administered at their practice, 71% of physicians considered schools, and 63% considered teen clinics to be acceptable alternative adolescent immunization sites. However, many had concerns about continuity of care for adolescents receiving immunizations in school. CONCLUSIONS: Most physicians supported adolescent immunization efforts. Barriers preventing adolescent immunization included financial barriers, record scattering, lack of tracking and recall, and missed opportunities. School-based immunization programs were acceptable to most physicians, despite concerns about continuity of care. Further research is needed to determine whether interventions that have successfully increased infant immunization rates are also effective for adolescents.

Childhood vaccine risk/benefit communication in private practice office settings: a national survey.

UNLABELLED: Communication about childhood vaccine risks and benefits has been legally required in pediatric health care for over a decade. However, little is known about the actual practice of vaccine risk/benefit communication. OBJECTIVES: This study was conducted to identify current practices of childhood vaccine risk/benefit communication in private physician office settings nationally. Specifically, we wanted to determine what written materials were given, by whom, and when; what information providers thought parents wanted/needed to know, the content of nurse and doctor discussion with parents, and the time spent on discussion. We also wanted to quantify barriers to vaccine risk/benefit discussion and to prioritize materials and dissemination methods preferred as solutions to these barriers. METHODS: We conducted 32 focus groups in 6 cities, and then administered a 27-question cross-sectional mailed survey from March to September 1998, to a random national sample of physicians and their office nurses who immunize children in private practices. Eligible survey respondents were active fellows of the American Academy of Pediatrics or American Academy of Family Physicians in private practice who immunized children and a nurse from each physician's office. After 3 mailings, the response rate was 71%. RESULTS: Sixty-nine percent of pediatricians and 72% of family physicians self-reported their offices gave parents the Centers for Disease Control and Prevention Vaccine Information Statement, while 62% and 58%, respectively, gave it with every dose. In ~70% of immunization visits, physicians and nurses reported initiating discussion of the following: common side effects, when to call the clinic and the immunization schedule. However, physicians reported rarely initiating discussion regarding contraindications (<50%) and the National Vaccine Injury Compensation Program (<10%). Lack of time was considered the greatest barrier to vaccine risk/benefit communication. Nurses reported spending significantly more time discussing vaccines with parents than pediatricians or family physicians (mean: 3.89 vs 9.20 and 3.08 minutes, respectively). Both physicians and nurses indicated an additional 60 to 90 seconds was needed to optimally discuss immunization with parents under current conditions. Stratified analysis indicated nurses played a vital role in immunization delivery and risk/benefit communication. To improve vaccine risk/benefit communication, 80% of all providers recommended a preimmunization booklet for parents and approximately one half recommended a screening sheet for contraindications and poster for immunization reference. The learning method most highly endorsed by all providers was practical materials (80%). Other desirable learning methods varied significantly by provider type. CONCLUSIONS: There was a mismatch between the legal mandate for Vaccine Information Statement distribution and the actual practice in private office settings. The majority of providers reported discussing some aspect of vaccine communication but 40% indicated that they did not mention risks. Legal and professional guidelines for appropriate content and delivery of vaccine communication need to be clarified and to be made easily accessible for busy private practitioners. Efforts to improve risk/benefit communication in private practice should take into consideration the limited time available in an office well-infant visit and should be aimed at both the nurse and physician.

Impact of vaccine financing on vaccinations delivered by health department clinics.

OBJECTIVES: This study measured the number of childhood vaccinations delivered at health department clinics (HDCs) before and after changes in vaccine financing in 1994, and it assessed the impact of changes in financing on HDC operations. METHODS: We measured the number of vaccination doses administered annually at all 57 HDCs in New York State between 1991 and 1996, before and after the financing changes. Interviews of HDC personnel assessed the impact of financing changes. A secondary study measured trends in Pennsylvania and California. RESULTS: HDC vaccinations for preschool children in New York State declined slightly prior to the financing changes (6%-8% between 1991 and 1993) but declined markedly thereafter (53%-56% between 1993 and 1996). According to nearly two thirds of New York State's HDCs, the primary cause for this decline was the vaccine-financing changes. HDC vaccinations for preschool children in Pennsylvania declined by 12% between 1991 and 1993 and by 56% between 1993 and 1997. HDC vaccinations for polio-containing vaccines in California declined by 31% between 1993 and 1997. CONCLUSIONS: Substantially fewer vaccinations have been administered at HDCs since changes in vaccine financing, thereby keeping preschool children in their primary care medical homes.

Decline in physician referrals to health department clinics for immunizations: the role of vaccine financing.

BACKGROUND: Physicians frequently refer children to health department clinics (HDCs) for immunizations because of high out-of-pocket costs to parents and poor reimbursement for providers. Referrals for immunizations can lead to scattered care. In 1994, two vaccine financing reforms began in New York State that reduced patient costs and improved provider reimbursement: the Vaccines for Children Program (VFC, mostly for those on Medicaid and uninsured) and a law requiring indemnity insurers to cover childhood immunizations and preventive services. OBJECTIVE: To measure reported changes in physician referrals to HDCs for immunizations before and after the vaccine financing reforms. DESIGN: In 1993, a self-administered survey measured immunization referral practices of primary care physicians. In 1997, we resurveyed respondents of the 1993 survey to evaluate changes in referrals. SETTING/ PARTICIPANTS: Three hundred twenty-eight eligible New York State primary care physicians (65% pediatricians and 35% family physicians) who responded to the 1997 follow-up immunization survey (response rate of 82%). RESULTS: The proportion of physicians reporting that they referred some or all children out for immunizations decreased from 51% in 1993 to 18% in 1997 (p<0.001). In 1997, physicians were more likely to refer if they were family physicians (28% vs. 13%,p<0.01), or did not obtain VFC vaccines (29% vs. 13%,p<0.001). According to physicians who referred in 1993, decreased referrals in 1997 were due to the new insurance laws (noted by 61%), VFC (60%), Child Health Plus (a statewide insurance program for poor children, 28%), growth in commercial managed care (23%), Medicaid managed care (19%), and higher Medicaid reimbursement for immunizations that is due to VFC (18%). For physicians noting a decline in referrals, the magnitude of the decline was substantial-60% fewer referrals for VFC-eligible patients and 50% fewer for patients eligible under the new insurance law. CONCLUSIONS: Vaccine financing reforms decreased the proportion of physicians who referred children to HDCs for immunizations, and may have reduced scattering of pediatric care.

How much time is spent on well-child care and vaccinations?

CONTEXT: Because well-child care represents the most important prevention opportunity in the health care system, a growing number of activities and indicators have been proposed for it. OBJECTIVE: To measure the time spent in the various components of well-child care. DESIGN: Time-and-motion study. SETTING: Five private pediatric practices and 2 public providers in Rochester, NY. PARTICIPANTS: One hundred sixty-four children younger than 2 years. MAIN OUTCOME MEASURE: Duration of family's encounters with the primary care provider (physician or nurse practitioner), nurse, and other personnel. RESULTS: The median encounter times and their component parts in minutes were: (1) primary care provider, 16.3 (physical examination, 4.9; vaccination discussion, 1.9; discussion of other health issues, 9.5; vaccination administration, 0); (2) nurse, 5.6 (physical examination, 3.5; vaccination discussion, 0; other health discussion, 0; vaccine administration, 1.6); and (3) other personnel, 0 for all categories. Public provider setting, African American race of the child, and administration of 4 vaccinations were significantly associated with an increase (3-4 minutes) in the duration of the primary care provider encounter. Only 8 (5%) of families read vaccine information materials. CONCLUSIONS: Depending on whether a child makes the usual 3 or recommended 6 number of well-child visits, the total time of well-child care is 45 to 90 minutes during the first year of life and declines to less than 30 minutes per year thereafter as the number of recommended visits diminish. Because high-risk children make half as many well-child care visits as other children, a 3 to 4 minute increase in encounter time is insufficient to provide them with the same level of care as other children.

A randomized study of tracking with outreach and provider prompting to improve immunization coverage and primary care.

OBJECTIVE: To compare and measure the effects and cost-effectiveness of two interventions designed to raise immunization rates. SETTINGS: Nine primary care sites serving impoverished and middle-class children. SUBJECTS: Complete birth cohorts (ages 0 to 12 months; n = 3015) from these sites. INTERVENTIONS: Two 18-month duration interventions: 1) tracking with outreach [tracking/outreach] to bring underimmunized children to their primary care provider office, and 2) a primary care provider office policy change to identify and reduce missed immunization opportunities (prompting). DESIGN: Randomized, controlled trial, randomizing within sites using a two-by-two factorial design. Subjects were allocated to one of four study groups: control, prompting only, tracking/outreach only, and combined prompting with tracking/outreach. Outcomes were obtained by blinded chart abstraction. MEASURES: Immunization status for age; number of days of delay in immunization; primary care utilization; and rates of screening for occult disease. RESULTS: Out of 3015 subjects, 274 subjects (9%) transferred out of the participating sites or had incomplete charts and were excluded. The 2741 (91%) remaining subjects were assessed. At baseline, study groups did not differ in age, gender, insurance type, or immunization status. Of the remaining subjects, 63% received Medicaid. Final series-complete immunization coverage levels were: control, 74%; prompting-only, 76%; tracking/outreach-only 95%; and combined tracking/outreach with prompting, 95%. Analysis of variance showed that: 1) tracking/outreach increased immunization rates 20 percentage points; 2) tracking/outreach decreased mean immunization delay 63 days; 3) tracking/outreach increased mean health supervision visits 0.44 visits per child; 4) tracking/outreach increased mean anemia screening 0.17 screenings per child and mean lead screenings 0.12 screenings per child; 5) impact of tracking/outreach was greatest for uninsured and impoverished patients; and 6) the prompting intervention had no impact on the studied outcomes, and its failure was caused by inconsistent use of prompts and failure to vaccinate ill children when prompted. Using tracking/outreach, the cost per additional child fully immunized was $474. Each $1000 spent on the tracking/outreach intervention resulted in: 2.1 additional fully vaccinated children and 668 fewer child-days of delayed immunization; 4.6 additional health supervision visits and 5.9 additional other visits to the primary care provider; and 1.8 additional anemia screenings and 1.3 additional lead screenings. CONCLUSIONS: Outreach directed toward children not up-to-date on immunizations improves not only immunization status, but also health supervision visit attendance and screening rates. The cost per additional child immunized was high, but should be interpreted in view of the spillover benefits that accompanied improved immunization. Effective means to improve coverage by reducing missed immunization opportunities still need to be identified.

Effect of 2 urban emergency department immunization programs on childhood immunization rates.

BACKGROUND: Emergency departments (EDs) are recommended as sites for immunizing children. However, there is little information about the effect of ED immunization programs on immunization rates. OBJECTIVES: To assess the ability of 2 ED immunization programs to vaccinate children and to measure the effect of the programs on immunization rates after the ED visit and 6 months later. DESIGN: A prospective cohort study. Emergency department patients were screened for immunization status, and vaccinations were offered to patients who either were documented to be eligible or were eligible by age and had no documented records. A systematic, sequential sample of those accepting vaccinations (study patients) was compared with a systematic, sequential sample of those not vaccinated (control subjects). Telephone interviews and medical record reviews were performed 6 months after the ED visit to verify dates of immunizations. Results were weighted to reflect the sampling frames of patients screened by the 2 programs. SETTING: Two EDs in New York City (in Manhattan and the Bronx) and the surrounding primary care offices. PATIENTS: Children (aged 0-6 years) screened for immunization status by the ED immunization program during a 10-week period; these included 210 children from the Manhattan ED (106 vaccinated in the ED) and 274 children from the Bronx ED (129 vaccinated in the ED). INTERVENTION: Emergency department immunizations. MAIN OUTCOME MEASURES: Proportion of patients (vaccinated, not vaccinated, and ED population) up-to-date for immunizations (1) at the time of the ED visit, (2) 1 day later, and (3) 6 months later. RESULTS: Two thirds of the patients in each ED had Medicaid, and one tenth were uninsured. At the time of the ED visit, 20% of the vaccinated children in each ED were actually up-to-date and were unnecessarily vaccinated; 74% (Manhattan ED) and 72% (Bronx ED) of the not vaccinated children were up-to-date (the remainder were later determined to have been eligible for vaccinations). One day after the ED visit, and 6 months later, the immunization rates of the vaccinated and not vaccinated children were similar. The results of the weighted analysis were as follows: for the entire ED population screened for immunization status, compared with up-to-date rates at the time of the ED visit, rates 1 day later were 11% (Manhattan ED) and 8% (Bronx ED) higher in each ED (P < .05); and rates 6 months later were the same in the Manhattan ED and 10% lower in the Bronx ED (P < .01). Eighteen percent of all children screened for immunization status were vaccinated; 10 to 15 children were screened and 2 to 4 children were vaccinated per 8-hour ED shift. CONCLUSIONS: This ED immunization program temporarily improved the immunization rates of the ED population, but substantial personnel time was required to achieve these small gains. Urban ED immunization programs are unlikely to be cost-effective.

Effect of emergency department immunizations on immunization rates and subsequent primary care visits.

BACKGROUND: The Standards for Pediatric Immunization Practices recommend the routine use of emergency department (ED) encounters for screening the immunization status of children and, if indicated, immunizing them. OBJECTIVE: To test the hypothesis that ED immunizations will improve immunization rates without decreasing subsequent primary care visits. DESIGN: A randomized controlled trial of 2 interventions. Children (aged 6-36 months) (n = 1835) were enrolled in the study in the ED; informed consent was obtained from their parents. They were randomized into 1 of 3 groups: (1) the control group (n = 614), in which no intervention was undertaken; (2) the letter group (n = 610), in which a letter to the primary care physician was written indicating the child's estimated likelihood of being underimmunized; and (3) the ED vaccination group (n = 611), in which, based on a decision rule, those likely to be underimmunized were offered immunizations in the ED. After randomization, parents were interviewed in the ED using a decision rule to estimate the likelihood of the child being underimmunized. One year after enrollment in the study, the medical records of the children at their primary care sites were reviewed to determine the immunization status of the children and primary care use patterns. SETTING: An urban ED and 54 primary care sites in Monroe County, New York. RESULTS: The mean age of the participants was 17.9 months. Medical record review-verified underimmunization rates at the time of the ED visit were 33%, 31%, and 28% for the control, letter, and ED vaccination groups, respectively. The demographic characteristics and baseline immunization rates were not different among study groups. According to the decision rule, 248 children (41%) in the ED vaccination group were likely to be underimmunized. Parents of these 248 children were offered immunizations for their children; 117 (47%) accepted, and their children were immunized (with 230 separate immunizations). One month after the ED visits, the underimmunization rates of the study groups were 31%, 28% (P = .40 compared with the control group), and 23% (P = .002). One year later, these rates were 28%, 25% (P = .20), and 25% (P = .20). No clinically meaningful differences were present at either of these times. One year after the ED visit, no differences in the rates of primary care use were found among groups. CONCLUSIONS: This study provides evidence that the immunization of children in this ED was ineffective at raising their immunization rates; primary care attendance was also unaltered. Major obstacles were as follows: (1) an inability to ascertain accurately the immunization status in the ED and (2) a high rate of parental refusal to accept immunizations in the ED. The standards should be modified to de-emphasize the ED as a routine immunization site for children with access to primary care. Efforts and resources should be directed toward strengthening the primary care system and tracking immunization status.

Reducing missed opportunities for immunizations. Easier said than done.

BACKGROUND: Missed opportunities for immunizations are associated with underimmunization of preschool-age children. Practice policies limiting immunizations to scheduled preventive visits and guidelines requiring legal guardians to sign consent forms for vaccinations are 2 factors contributing to missed opportunities. However, methods to change these policies have not been sufficiently evaluated. OBJECTIVE: To measure the effectiveness of (1) changing practice policies to incorporate the new national standard to screen and vaccinate eligible children at all office visits and (2) eliminating legal guardian signature requirements. DESIGN: A randomized controlled trial of 2 interventions: (1) changing practice policy and routine to have office nurses screen for immunization status at all visits, attach immunization reminder cards to medical charts for eligible patients, and have providers vaccinate eligible children ("no missed opportunities" intervention) and (2) changing practice guidelines to allow vaccinations without a legal guardian's signature. The first intervention was performed at both sites; the second only at the neighborhood health center (NHC). SETTING: A Pediatric Continuity Clinic in a teaching hospital (hereafter referred to as Clinic), and an NHC. PATIENTS: Enrolled in the trial were 1005 Clinic patients and 983 NHC patients, 0 to 2 years of age. MAIN OUTCOME MEASURES: Missed opportunity rates, immunization rates, and rates of preventive services. RESULTS: Eliminating the requirement for a legal guardian's signature had no effect on any of the outcome measures. The no missed opportunities intervention was partially effective. Study patients had slightly fewer missed opportunities than control patients at each site: (0.60 vs 0.90 per patient per year at the Clinic, P = .01; 1.1 vs 1.3 per patient per year at the NHC, P = .02). For study group patients, immunization reminder cards were attached to medical charts in only one third of vaccine-eligible visits; when attached, they markedly increased vaccination by providers (odds ratio for vaccinating at a visit was 6.9 comparing visits when immunization reminder cards were attached vs not attached). However, at the end of the study, immunization rates were similar for study and control groups at each site. The number of undervaccinated days was slightly lower for the no missed opportunities study group at the Clinic than for the control group (56 days vs 77 days, P < .001), but they were similar for both groups at the NHC. There were no differences in rates of preventive visits or screening tests between study and control groups. CONCLUSIONS: The interventions evaluated to reduce missed opportunities did not increase immunization rates. The key problem was failure to screen for immunization status at all visits. More effective interventions will be needed to overcome barriers within busy primary care practices to substantially reduce missed opportunities.

Physician and parent opinions. Are children becoming pincushions from immunizations?

OBJECTIVE: To determine parent and physician opinions regarding the administration of multiple childhood immunizations by injection. DESIGN: Confidential mailed survey to physicians and residents; interview of parents during office visits for immunizations. PARTICIPANTS: Physicians and parents from Rochester, NY. RESULTS: The survey included 215 practicing physicians and 74 residents; response rate was 82%. Of the 197 parents interviewed, 93% were mothers, 68% were white; the mean (+/- SD) age was 25.8 +/- 5.2 years, with 12.8 +/- 1.8 years of education; 59% had private insurance, and 35% had Medicaid coverage. Of the parents, 31% had strong concerns about their child receiving a single injection; an additional 10% (total, 41% vs 31%; chi 2 = 4.05, P = .04) had the same concerns about their child receiving three injections. More practicing physicians than parents had strong concerns about children 7 months old or younger receiving three injections (60% vs 41%; chi 2 = 7.71, P < or = .01). Physician concern increased further when physicians were asked about administration of four injections (80% vs 60%; chi 2 = 18.77, P < .001). Of the parents, 64% preferred one rather than two visits to have three injections administered, if their physician recommended it; 58% still preferred one visit even if four injections were needed. CONCLUSIONS: Physicians have more concerns than parents about the administration of multiple injections at a single visit. Pain for the child was the main concern of all respondents. While most physicians have strong concerns about administering three or more injections at one visit, most parents prefer this practice. Continued education and reassurance of parents and physicians is needed to address concerns about children becoming "pincushions" from immunizations.

Is underimmunization a marker for insufficient utilization of preventive and primary care?

OBJECTIVE: To test the hypothesis that the underimmunization of young children is a marker for the lack of preventive and acute primary care. SETTING: Primary care center serving an impoverished population (90% Medicaid). DESIGN: Historical cohort study (N = 1178) of children aged 12 to 30 months that determined each child's immunization status, anemia, tuberculosis, and lead screening status; and office utilization history. Screening delay was defined as missing a recommended screening by more than 3 months past the standard screening age. RESULTS: Thirty-four percent of the population were underimmunized at 12 months of age. Compared with fully immunized children, these children were at greater risk for screening delay: anemia, 38% vs 5% (risk ratio [RR], 7.5; 95% confidence interval [CI], 5.4 to 10.4); tuberculosis, 76% vs 44% (RR, 1.7; CI, 1.6 to 1.9); and lead, 69% vs 33% (RR, 2.1; CI, 1.9 to 2.4). These RRs increased with greater immunization delay. Compared with fully immunized children, the underimmunized group made 47% fewer preventive health visits (2.5 vs 4.7 visits per infant per year, P < .001) and 43% fewer illness visits (2.5 vs 4.4, P < .001) and had 50% more missed appointments (2.1 vs 1.4, P < .001). Logistic regression, predicting anemia screening delay at 12 months of age, showed that underimmunization had an effect independent of utilization, with an odds ratio of 7.7 (CI, 5.2 to 12.0). CONCLUSION: Underimmunization was a powerful, independent marker for inadequate health supervision in this population. IMPLICATIONS: The current emphasis on immunizations has the benefit of targeting children at risk of lack of preventive and acute care. Improving immunization rates may have the potential to improve other aspects of primary care if immunization provision is not uncoupled from primary care.

Immunization practices of pediatricians and family physicians in the United States.

OBJECTIVE: To assess current practices and attitudes among pediatricians and family physicians across the United States regarding immunizations. DESIGN: Survey of a random sample of pediatricians and family physicians. SUBJECTS: Fellows of the American Academy of Pediatrics (N = 746) and American Academy of Family Medicine (N = 429). SURVEY TOPICS: General immunization practices (eg, types of visits during which vaccinations are provided, mechanisms to identify undervaccinated children); and opinions about perceived barriers to immunizations, acceptance of alternative sites for immunizations, and possible immunization requirements for Medicaid and The Special Supplemental Food Program for Women, Infants, and Children (WIC). RESULTS: Pediatricians and family physicians (combined) reported the following: immunizing children during acute illness visits (28%), follow-up visits (90%), and chronic illness visits (77%); using computer or reminder files to identify undervaccinated children (13%); and simultaneously administering four vaccines (diphtheria-tetanus-pertussis, oral poliovaccine, measles, mumps, and rubella and Haemophilus influenzae type b) to an eligible 18-month-old child (66%). Physicians perceived the following as barriers to immunizations: missed preventive visits (40%), vaccine costs (24%), lack of insurance coverage (24%), inability to track undervaccinated patients (22%), incomplete immunization records (12%), and missed vaccination opportunities (12%). Physicians agreed with offering vaccinations during hospitalizations (51%) or emergency department visits (30%), and with immunization requirements for continued eligibility for Medicaid (66%) or WIC (64%). Pediatricians were more likely to vaccinate during chronic illness and follow-up visits, and were more likely to use systems to track undervaccinated children (P < .05); however, most immunization practices and attitudes of pediatricians and family physicians were similar. Physicians who graduated from medical school more recently and those in high-risk urban practices were more likely to vaccinate during acute illness visits, provide simultaneous vaccinations, and favor vaccinations in hospital settings. CONCLUSIONS: Vaccination rates might be improved by closer adherence to current immunization guidelines regarding vaccinations during all encounters and simultaneous vaccinations, by developing systems to identify undervaccinated children, and by reducing patient costs for vaccinations. Current immunization practices fall short of the immunization guidelines; changes in individual practice styles will be required to conform with these standards.

Intent to immunize among pediatric and family medicine residents.

OBJECTIVE: To determine whether contraindications to immunization are inappropriately broadened for children with a fever or a neurologic condition. PARTICIPANTS: Pediatric and family medicine residents (N = 52 and 23, respectively) at the University of Rochester (NY). DESIGN: Cross-sectional survey. Residents rated how likely they would be to administer a diphtheria-tetanus-pertussis or measles-mumps-rubella vaccine in 17 clinical scenarios according to a rating scale ranging from 1 (never) to 5 (always). For all scenarios, the immunization was recommended by the American Academy of Pediatrics or the Immunization Practices Advisory Committee. RESULTS: In only five and three of 17 scenarios would 90% or more of the pediatric residents and family medicine residents, respectively, have administered an immunization. For diphtheria-tetanus-pertussis vaccine, pediatric residents reported a lower likelihood of vaccinating a 2-month-old child with a low fever (temperature, 38.1 degrees C) than an afebrile child (mean score, 3.0 vs 4.7; P < .01). A 2-year-old child with idiopathic epilepsy, a 2-month-old child with intraventricular hemorrhage, and a 2-month-old child who had a parent with a seizure disorder each had a lower reported likelihood to be vaccinated than a same-aged child without a neurologic condition (2.8 vs 4.5; 4.1 vs 4.7; and 4.3 vs 4.7, respectively; each P < .01). For measles-mumps-rubella, pediatric residents reported a lower likelihood of vaccinating a 15-month-old child with a low fever than an afebrile child (4.2 vs 4.9; P < .01). A child with a progressive neurologic disease had a lower reported likelihood to be vaccinated than a child without a neurologic condition (3.5 vs 4.9; P < .01). CONCLUSIONS: Residents reported a lower likelihood of immunizing children with a fever or neurologic condition. Such practice styles may contribute to underimmunization. Residents need to be educated regarding which medical conditions contraindicate an immunization.

ROC curves for classification trees.

A common problem in medical diagnosis is to combine information from several tests or patient characteristics into a decision rule to distinguish diseased from healthy patients. Among the statistical procedures proposed to solve this problem, recursive partitioning is appealing for the easily-used and intuitive nature of the rules it produces. The rules have the form of classification trees, in which each node of the tree represents a simple question about one of the predictor variables, and the branch taken depends on the answer. The authors consider the role of misclassification costs in developing classification trees. By varying the ratio of costs assigned to false negatives and false positives, a series of classification trees are generated, each optimal for some range of cost ratios, and each with a different sensitivity and specificity. The set of sensitivity-specificity combinations define a curve that can be used like an ROC curve.