The Delivery of Immunizations and Other Preventive Services in Private Practices

(1)

The

Delivery

of Immunizations

and

Other

Preventive

Services

in Private

Practices

W. Clayton Bordley, MD, MPH; Peter A. Margolis, MD, PhD; and Carole M. Lannon, MD, MPH

ABSTRACT. Objectives. To measure the proportion

of children cared for in private pediatric practices who are fully immunized and have been screened for anemia,

tuberculosis (TB), and lead poisoning by 2 years of age.

Design. Cross-sectional chart review.

Setting. Fifteen private pediatric practices in central North Carolina (11 chosen randomly).

Patients. One thousand thirty-two randomly selected 2-year-old children.

Main Outcome Measures. Proportion of children

im-munized and screened for anemia, TB, and lead

poison-ing by 24 months of age and immunization and screening rates of the practices.

Results. Sixty-one percent of the children were fully immunized at 24 months of age; the rates among practices

varied widely (38% to 82%). Sixty-eight percent of the

children had been screened for anemia, 57% had been screened for TB, and 3% had been screened for lead

poisoning. Physicians overestimated the proportions of

fully immunized children in their practices by an average

of 10% (range, -3% to 17%). The median number of well child visits by 2 years of age was 5 (range, 0 to 14), and only 19% of the entire sample made 8 or more well child

visits, the number recommended by the American

Acad-emy of Pediatrics in the first 18 months oflife. The numbers of well child and non-well child visits were the strongest predictors of complete immunization. Practice

characteris-tics associated with being fully immunized included the

use of preventive services prompting sheets (eg, flow

sheets) in the medical records, not seeing the same

physi-aan for all well child care, and having nurses review

pa-tients’ immunization status during their visits to the office.

Conclusions. Underimmunization and inadequate

screening are significant problems in private pediatric practices in North Carolina. Physicians are unaware of the rates of underimmunization in their offices. Pediatrics 1996;97:467-473; immunization, private

prac-tice, screening, anemia, lead poisoning, child, preschool child, health promotion, immunization schedule, infant,

newborn infant, pediatrics.

ABBREVIATIONS. AAP, American Academy of Pediatrics; TB,

tuberculosis; DiP, diphtheria-tetanus-pertussis; OPV, oral polio vaccine; Hib, Haemophilus influenzae type b; CI, confidence

inter-val; OR, odds ratio.

From the Department of Pediatrics, University of North Carolina at Chapel

FEll.

Preliminary reports of this research were presented at the 34th Annual

Meeting of the Ambulatory Pediatric Association, Seattle, WA, May 1994. Received for publication Mar 3, 1995; accepted May 15, 1995.

Reprint requests to (W.C.B.) Community Pediatrics, CB 7225, Wing C,

Medical School, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599-7225.

BACKGROUND

Data from the National Health Interview Survey

show that nationally only 65% of 2-year-old children

are fully immunized.1 Retrospective reviews of

vac-cination coverage in 21 cities found that only 11 % to

58%

of children were fully immunized by their

sec-ond birthdays.2 Studies of immunization rates have

concentrated on disadvantaged populations7 or

have been population-based surveys.” These studies

have provided useful information for targeting

pub-lic health efforts, but they have not provided

infor-mation about the extent of the problem in specific

provider settings. Although more than 50% of

chil-dren in the United States receive their

immuniza-tions from private physicians,1#{176} little is known about

the effectiveness of immunization delivery in private

physicians’ offices. It is also not known if the

effec-tiveness of individual practices at immunizing

pa-tients is influenced more by practice organizational

characteristics or by the socioeconomic makeup of

the patient populations.

Immunizations are only one component of

com-prehensive well child care. The American Academy

of Pediatrics (AAP) recommends that, in addition to

immunizations, children be screened for

develop-mental disorders, anemia, lead poisoning, and, in

certain populations, tuberculosis (TB) within the

con-text of eight to nine well child visits during the first

2 years of life. Several studies have suggested that

children often do not receive the recommended

num-ber of well child visits.9”2 Children who are

under-immunized often have not received other aspects of

well child care.13 On the other hand, it is not known

if full immunization alone accurately reflects

whether children have received all aspects of

preven-tive care.

The objective of this study was to determine the

proportion of children cared for in private pediatric

practices who are fully immunized and have

re-ceived other recommended preventive services (eg,

screening for anemia, TB, and lead poisoning). We

also examined whether patient or practice

character-istics were associated with full immunization in

pri-vate practices and the relationship between

immuni-zation rates and the receipt of other preventive

services.

METHODS

The study was performed in 15 private pediatric practices in

central North Carolina in 1992 and 1993. Four of the practices were

included because they were part of an ongoing study of access to

care. Eleven additional practices were randomly selected from the

at Viet Nam:AAP Sponsored on September 1, 2020

www.aappublications.org/news

(2)

approximately 75 pediatric practices located within 120 miles of

the University of North Carolina. One of the investigators (W.C.B.)

contacted the senior partner at each practice by phone to invite

participation. No financial incentives were provided, and all of the

practices agreed to participate.

Physicians in the participating practices completed a

question-naire that included questions about their immunization practices.

They were also asked to estimate the proportion of children in

their practices who were fully immunized.

Information about practice characteristics was obtained

through interviews with the office manager at each practice. We

asked about the size of the practice staff, patient volume, payer

mix, and hours of operation. We also asked how often patients see

the same physician for acute and well child care, and whether the

practices used specific systems to identify patients eligible for immunizations.

We used chart abstractions to determine practice immunization

rates. In each practice, we selected a random sample of children

between 24 and 35 months of age from their lists of active patients.

In all 15 practices, lists were generated by their computerized

billing systems. Children who made three or more visits of any

kind to the practices were eligible for inclusion in the study;

therefore, the study included only regular users of the practices.

The charts of 75 children in each practice were abstracted to

provide the practices with estimates of their performance that had

a precision of ± 10%.

Chart abstractions were performed by a research assistant and

a member of each office’s staff. We collected the following

infor-mation from each chart: the date and type of all visits to the

practice, the dates of all immunizations given, the dates that

screening procedures (anemia, TB, and lead) were performed, the

presence in the chart of completed growth charts and preventive services or flow sheets, the child’s date of birth, insurance status

(indicated at the most recent visit), and parents’ occupations and

marital status. Many practices did not record sources of insurance

for patients who had health insurance from carriers for which the

practices did not file claims. Thus, a child whose source of

insur-ance was not recorded may have been uninsured or had some

form of health insurance not processed by the practice. The

insur-ance status of these children was categorized as indeterminate.

Visits were categorized as well child if they were identified in

medical records by descriptive terms such as “WCC (well child

care),” “J-up (checkup),” “well check,” “Early Periodic Screening

and Diagnostic Testing,” and “4 month.” All other visits (eg, visits

for acute illness, follow-up of an acute problem, and follow-up of

a chronic problem) were categorized as non-well child visits.

Visits intended for well child care during which a problem was

found (eg, unsuspected otitis media) were categorized as both.

Information on immunizations received outside the practice

(eg, local health departments) was abstracted if the date, location,

and specific vaccine received were documented in practice charts.

For children found to be underimmunized according to the

records of their practices, we searched the records of local health

departments (n

=

21) to see whether these children received

immunizations not recorded in their practice charts. We added all

immunizations received at health departments to the children’s

records before finally determining if they were fully immunized.

Entries such as “shots at the health department” were not

consid-ered sufficient evidence that specific vaccines had been received.

Immunization Definitions

We used the schedule of immunizations recommended by the

AAP.14 We assessed immunization rates at 3 months (one

diph-theria-tetanus-pertussis [DTPI vaccine and one oral polio vaccine

[OPV]), 6 months (two DiP vaccines and two OPVs), 12 months

(three DiP vaccines and two OPVs), and 24 months (four DiP

vaccines, three OPVs, one measles-mumps-rubella vaccine, and at

least one Haemophilus influenzae type b [Hib] vaccine). These points

of measurement allowed for grace periods of 1, 2, and 6 months,

respectively. The immunization schedule for H influenzae changed

during the 2 years preceding our data collection. We considered

children fully immunized against H influenzae if they received one

or more doses of Hib by 24 months of age. To determine the

impact of including Hib in the definition of full immunization, we calculated immunization rates at 24 months with and without Hib.

Universal immunization against hepatitis B was not

recom-mended by the AAP until February 1992,’ and we did not include

it in our definitions of full immunization.

Any hematocrit or hemoglobin level recorded in a patient’s

medical record was considered evidence of screening for anemia,

even if the test was done in the context of an acute-care visit.

Receipt of either a multipuncture or Mantoux test for TB was

acceptable as evidence of TB screening. Blood lead, free erythro-cyte protoporphyrin, or written evidence that screening questions,

such as those recommended by the Centers for Disease Control

and Prevention,’6 had been asked were acceptable as evidence of

screening for lead poisoning. Screening for anemia, TB, and lead

poisoning was categorized as complete if the procedures were

recorded in a child’s chart before the second birthday.

The study was approved by the Institutional Review Board of

the University of North Carolina School of Medicine.

Statistical Analysis

Data from the practice and health department chart

abstrac-tions were combined to obtain point estimates and 95% confidence

intervals (CIs) of the proportions of children fully immunized at 3,

6, 12, and 24 months of age and screened for anemia, TB, and lead

poisoning by 2 years of age. We also calculated practice level rates

of complete immunization and screening for anemia, TB, and lead

poisoning for each of the 15 practices.

We used stratified analyses to evaluate the relationship

be-tween patient and practice characteristics and full immunization.

We used the x test to compare nominal and categorical variables

and Student’s t test to compare continuous variables. We used

logistic regression to assess the relative importance of patient and

practice characteristics on immunization status. Models were

ex-amined using a child’s immunization status as the dependent

variable. Independent variables included the number of well child

visits and non-well child visits, gender, race, method of payment,

and the presence of preventive services flow sheets in the charts.

These variables were chosen because each was collected at the

individual chart level and was significantly associated with

im-munization status in the bivariate analysis. Collinearity and

inter-action were assessed in the full model using standard techniques.’7

The CIs reported for the logistic analysis were adjusted to account

for the cluster sampling of the data.’8

Reliability of the chart abstraction process was assessed by

reabstraction of 40 charts, a 10% sample of charts in the first four

practices studied. We found very little observer variability. .c

indexes on a number of dichotomous variables ranged from 0.85

to 1.00, indicating excellent strength of agreement.’9

Study Participants

RESULTS

Table 1 summarizes the types of practices included

in the sample, although the diversity of the practices

cannot be fully captured by a table. For example, one

of the large groups (seven physicians) is the only

pediatric practice in the state’s largest and

fifth-poor-est county; two of the larger practices serve affluent,

white, suburban communities; one of the solo

prac-titioners practices in a predominantly

African-Amer-ican, urban neighborhood, and the other practices in

a small, racially mixed town typical of North

Caro-TABLE 1. Description of the Practices (n

=

15)

Characteristic n

Size

Solo practice 2

2 physicians 2

Group practice 11

Setting

Urban 8

Rural 7

Proportion of practice covered by Medicaid

s25% 5

26%-49% 6

(3)

lina. Forty-eight (87%) of the 55 physicians in the

study practices completed the surveys. The median

number of physicians in the 15 practices was 3

(range, I to 8). The physicians in the study saw a

median of 28 (range, 10 to 40) patients per day.

Practices had a median of 95 patients per day in the

summer, 125 in the winter, and 30 newborn infants

each month.

Characteristics of the 1032 patients sampled are

summarized in Table 2. Race was not recorded for

30% of the sample, because several practices either

did not record race or did so inconsistently. Most

children had some form of health insurance, either

from a third-party source or Medicaid. Insurance

status was not recorded for 17% of the patients.

Forty-seven percent of the patients received some or

all of their immunizations at local health depart-ments.

Well Child Visits

The 1032 children we studied made 15 523 visits to

the 15 practices. The median number of total visits

per child by 2 years of age was 11 (range, 3 to 52).

The median number of well child visits by 2 years of

age was 5 (range, 0 to 14). Only 19% of the children

had 8 or more well child visits, the number of visits

recommended by the AAP in the first 18 months of

life.11 Fifty-two percent had 5 or more well child

visits, the number of visits needed to receive all

recommended vaccines during well child visits.

CM!-dren who were fully immunized made more well

child visits than those who were not fully

immu-nized (5.5 vs 3.2 visits; P < .001). However, only 27%

of the children who were fully immunized made 8 or more well child visits.

Immunization Rates

The proportion of children fully immunized (Table

3) was highest at 3 months (79%) and lowest at 24

months (61%). In addition, there was considerable

variation in the immunization rates among the 15

practices. The variation in the proportions of

chil-dren immunized in each practice was greatest at 24

months (Figure). On average, physicians overesti-mated the proportions of their 2-year-old patients

who were fully immunized by 10% (range, -3% to

17%).

Several patient and practice characteristics were

associated with immunization status in our stratified

TABLE L Description of Pati ents (n

=

1032)

Characteristic Study

Population, n (%)

Sex, male 572 (55)

Race

White 380 (37)

African-American 293 (28)

Native American 42(4)

Other 4(<1)

Not recorded 313 (30)

Insurance

Third party 469 (45)

Medicaid 377(37)

Not recorded 186 (18)

analysis (Table 4). The patient characteristics

in-cluded the number of visits a child made to a practice

(for both well child and non-well child care), the

child’s race, and the source of insurance. Only

77%

of

the children who made five or more well child visits,

the number needed to receive all the recommended

vaccines, were fully immunized. In addition, 36% of

the children who were not fully immunized at 24

months had made well child visits after the dates of

their last immunizations and before their second

birthdays.

Practice characteristics associated with full

immu-nization included whether a chart contained a

prompting sheet, not seeing the same physician for

all or most well child visits, and whether nurses

assessed the children’s immunization statuses

dur-ing their visits to the offices. Practice factors that

were not associated with full immunization included the locations of the practices (urban versus rural), the

numbers of physicians in the practices, the ratios of

nurses to physicians, the daily patient volumes of the

practices, and the availability of the practices at night

and on weekends.

Using logistic regression, the number of well child

visits (odds ratio [ORI, 1.97; 95% CI, 1.64-2.36) and

non-well child visits (OR, 1.22; 95% CI, 1.06-1.40)

were the strongest predictors of a child’s

immuniza-tion status. In addition, the presence of a completed

preventive services flow sheet, a practice

character-istic, also remained associated with full

immuniza-tion (OR, 1.58; 95% CI, 1.00-2.52). Race, sex, and

insurance status were not associated with

immuni-zation status in our final model.

Rates of Other Preventive Services

The proportion of children screened for anemia

and the range of practice rates were similar to those

for full immunization at 2 years of age; the rates for

TB and lead screening were lower (Table 3). As with

the immunization rates, there was considerable

van-ation in the practice rates for all the screening

pro-cedures.

Children who were fully immunized by 24 months

were more likely to have been screened for anemia

and TB than children who were not fully immunized

(Table 5). However, among those who were fully

immunized, 22% had not been screened for anemia,

30% had not been screened for TB, and 97% had not

been screened for elevated lead levels.

DISCUSSION

We found that immunization rates for children

cared for in this sample of private practices were low

and no higher than national averages. This finding

challenges the idea that problems with this country’s

immunization delivery system exist only in publicly

funded clinics. Even the practices with the highest

rates did not achieve the national goals of 90% set

forth by the US Public Health Service#{176}and the Chin-ton administration’s Childhood Immunization Initia-tive.21

Few studies have focused specifically on private

practices, but several have suggested that

immuni-zation rates in this setting are uniformly high.

(4)

Service Fully Immunized, % (95% Confidence Interval)

Range of 15 Practice

Immunization

Rates (%)

Immunization rates*

3 mo (1 DTP, I OPV) 79 (77-81) 49-91

6 mo (2 DiPs, 2 OPVs) 75 (72-78) 49-87

12 mo (3 DiPs, 2 OPVs) 77 (74-80) 61-88

24 mo (4 DiPs, 3 OPVs, MMR) 63 (60-66) 48-83

24 mo (4 DiPs, 3 OPVs, MMR, 1 Rib) 61 (58-64) 35-82

Screening rates

Anemia 68 (65-71) 44-81

Tuberculosis 57 (54-60) 14-83

Lead 3 (2-4) 0-7

measIs-mumns-rube1ta:

.

I-Ih Haemophilus influenzae type b.

V

0

N C

E E

90

80

70

60

50-40

3 mo 6 mo 12 mo 24 mo

TABLE 3. Immunization and Screening Rates Measured at Various Ages (n

=

1032)

* DiP, diphtheria-tetanus-pertussis; OPV, oral polio vaccine; MMR,

U I I

Age

Figure. Practice variation: immunization rates of 15 practices by

age.

field et aP and Marks et al reported immunization

rates ranging from 75% to 84%. However, these

stud-ies were performed in the 1970s, before the 1988

through 1990 measles epidemic, and induded only

volunteer practices. More recently, in a study of

missed opportunities in seven primary-care settings

in Rochester, New York, Szilagyi et a!24 found that

75% to 91% of children cared for in three private

practices were fully immunized. In a study of

pa-tients seen in an urban emergency department,

Rodewald et al found that 85% of 274 children

cared for in 46 different practices were

age-appropri-ately immunized; because they evaluated only 6

pa-tients per practice, the authors could not measure

practice rates of full immunization. Hueston et

al’-measured immunization coverage in a birth

certifi-cate survey and found that 66% of children cared for

by private physicians were fully immunized

com-pared with 44% of children cared for in public clinic

settings. The design used in this study also

pre-cluded any examination of individual practice rates

of complete immunization.

There are several possible explanations for the low

levels of immunization we found. Immunization

rates in private practice settings may vary by region,

and some of this variation may be explained by

differences in insurance coverage. Children included

in our study had access to primary care, but many

may have been uninsured or had health insurance

that did not pay for immunizations. Only 4.8% of

children in North Carolina are insured by

group-model health maintenance organizations compared

with 13% nationally.27 In addition, the penetration

of managed care organizations in the state lags

be-hind that in many areas of the country. In contrast,

few children in the Rochester area are uninsured or

have insurance that does not cover immunizations.24

Therefore, the cost of immunizations to parents may

have been a significant barrier to receiving

immuni-zations in the practices we studied.

The high cost of immunizations in physicians’

of-fices prompts many parents to go to local health

departments for immunizations. More than 90% of

physicians in North Carolina refer at least some of

their patients to health departments for

immuniza-tions, and 40% refer all or some of their patients who

are receiving Medicaid.29 In the absence of a reliable

tracking system, the receipt of immunizations at

health departments cannot be verified easily. There

is no way to know what proportion of the parents of

the underimmunized children in our sample

indi-cated to their physicians that they were going to take

their children to health departments but then failed

to do so.

Another explanation for our findings is that

pri-vate practices may not be organized optimally to

deliver immunizations and other preventive

ser-vices. Numerous studies have shown that

opportu-nities to immunize children are frequently

missed5”224’’#{176}, and several have included patients

cared for in private practices.24’ We had intended to

obtain data on missed opportunities, but in pilot

testing our chart abstraction methods, we found it

frequently impossible to determine whether visits

met criteria for appropriate contraindications. This

was usually because of progress notes that either

contained inadequate information or were illegible.

The fact that 20% of the charts we reviewed did not

contain flow sheets of any kind also suggests that

(5)

TABLE 4. Characteristics Associated With Full Immunization

at 24 Months by Stratified Analysis

Variable Fully

Immunized, %

P

Patient characteristics

No. of well child visits

0-1 38 <.001

2-4 49

5-6 65

7 86

No. of non-well child visits

0-3 46 <.001

4-6 63

7-12 67

12 76

Race (n

=

719)*

White 68 <.001

Nonwhite 55

Insurance (n

=

846)*

Third party

Medicaid

67 57

<.001

Indeterminate 47

Practice characteristics Prompting sheet used

Yes 70 .005

No 59

Seeing the same physician

Yes 58 .007

No 67

Who checks immunization status

Physician Nurse

59 65

.06

*n < 1032 because of missing data.

TABLE 5. Proportions of Children Screened for Anemia,

Tuberculosis, or Lead Poisoning Stratified by Immunization Status

Patients Screened for:

Anemia* Tuberculosis* Lead

Yes, % No, % Yes, % No, % Yes, % No, %

Fully immunized

Yes 78 22 70 30 3 97

No 53 47 36 64 2 98

*These differences are statistically significant; P < .001.

have contributed to low rates of immunization; flow

sheets have improved the delivery of preventive

ser-vices in other primary-care sethngs. The failure of

practices to be adequately organized around the

pro-vision of preventive services may be reinforced by

the tendency of physicians to overestimate their

per-formance. The fact that the pediatricians in this study

overestimated the proportions of their patients who

were fully immunized is similar to that in studies of

physicians caring for adults that have demonstrated

significant gaps between physician self-reports

of screening and counseling practices and actual

performance.

The proportions of children who had been

screened for anemia (68%) and TB (57%) were

ap-proximately the same as the proportion who were

fully immunized, suggesting that immunization

sta-his may be a marker for the receipt of other

preven-tive services. However, among children who were

fully immunized, only 27% had made the number of

well child visits recommended by the AAP, and 22%

had not been screened for anemia by 2 years of age.

In a study of children attending a large, multisite

public practice in Denver, Brown et aP2 made similar

observations. Despite the fact that 90% of their

sam-ple were fully immunized by 2 years of age, only 38%

had made the AAP-recommended number of well

child visits. In a national sample, Mustin et a!9 found

that only 42% of white infants and 29% of black

infants received both adequate immunizations and

adequate well child visits by 8 months of age.

Con-sidered together, these data indicate that

immuniza-tion alone does not ensure the receipt of other

corn-ponents of comprehensive primary care.

The proportion of children screened for lead

poi-soning is consistent with data maintained by the

North Carolina Childhood Lead Poisoning

Preven-tion Program. Lead screening was not endorsed by

the state health department until November 1992,

and screening by private physicians did not begin to

increase until 1993 and 1994.

Although we studied only 15 practices in a single

state, several factors may allow our results to be

generalizable to other settings. Aside from

free-standing health maintenance organizations, we

in-cluded the spectrum of private practice types and

settings that exists in other regions of the country.

The proportion of children in North Carolina at the

time of this study who were fully immunized by 2

years of age (59%) and who received immunizations

in the private sector (45%) mirrored national data.10’9

The limitations of this study should be noted.

There is no standard definition for an active patient

in practice settings. By including children from this

study who made only three or more visits, we may

have excluded noncompliant families; this may have

led to an overestimation of immunization rates.

Con-versely, physicians may be unwilling to consider

children who have has made only one or two visits to

their offices active patients in their practices. All the

physicians in the practices we studied accepted the

validity or our eligibility criteria.

Data were collected by chart review. Children

pos-sibly received immunizations that were not recorded

in their practice records. It is unlikely that

immuni-zations administered in the practices were not

re-corded, because federal law requires physicians to

document all aspects of immunization delivery (ie,

consent and lot numbers). We searched records at

local health departments to avoid having to rely on

the practices to record immunizations received at

public clinics. Children possibly went to other local

private providers for immunizations, and this

infor-mation may not have been recorded in their charts.

We did not contact patients directly to see how often

this may have occurred. However, for the seven rural

practices we studied, few, if any, other local

primary-care physicians provided childhood immunizations.

It is also possible that the provision of other

screening procedures was not thoroughly

docu-mented in patients’ records. Tests for anemia are

typically billable and leave a paper trail, often in the

laboratory section of a medical record. TB test results

may not have been reliably recorded in practice

records; this may have resulted in an

underestima-tion of the proportion of children actually screened.

(6)

The quality of the data available in the office charts

precluded us from measuring nonlaboratory

preven-tive services such as risk assessment and counseling

activities. Information on race and insurance status

was missing for 30% and 17% of our sample,

respec-tively. Practices also did not keep accurate data on

parental employment or education level, so the

im-portance of these patient characteristics and

socio-economic status indicators could not be assessed

fully.

The recently adopted Standards for Pediatric

Im-munization Practices call for providers to conduct

semiannual audits to assess immunization coverage

and to review immunization records in the patient

populations they serve? Based on our experience,

private practices will find this mandate difficult to

implement. Practices must begin by defining “the

patient population they serve.” How many visits

must families make before practices think they are

accountable for their preventive care? Practices may

then be confronted by limitations in the data

con-tained in their records. In the absence of a

commu-nity-wide immunization tracking system, obtaining

accurate information on immunizations received

outside of the practice is time consuming. Measuring

the performance of other screening procedures,

es-pecially counseling activities, is even more difficult.

The issue of adequate documentation of preventive

care is becoming more important as a result of the

emphasis the American Board of Pediatrics is placing

on record keeping as part of its newly developed

program for recertification (Program for Renewal of

Certification in Pediatrics, American Board of

Pedi-atrics, January 1994) and the growing attention

man-aged care organizations are placing on measuring

performance and quality of care.

Our findings suggest that there are advantages to

studying practices as a unit of analysis. The

consid-enable variation in the immunization and screening

rates of the 15 practices we studied was not fully

explained by characteristics of the patients who used

. these practices. Other authors have attempted to

identify physician characteristics that are associated

with the immunization rates of their patients.39

How-ever, in the group practices we studied, many

cM!-dren saw more than one physician within the

prac-tices, and it was not possible to determine an

individual physician’s immunization rate. In

addi-tion, physicians share administrative, billing, and

nursing staffs, and systems for handling issues such

as missed appointments are usually similar for all

physicians within a practice. Interventions designed

for implementation in physician practices may be

more effective in improving the delivery of

immuni-zations and other preventives services than those

that target individual physicians. Future studies

should enroll a sufficient number of practices to

as-sess the association between modifiable practice

characteristics and the performance of preventive

services.

Finally, the fact that the number of visits patients

made to the practices was the strongest predictor of

full immunization underscores the importance of

ed-ucating parents about the immunization schedule

and the importance of well child visits to receive

other important preventive services. Parents must

bring their children to their physicians for a sufficient

number of well child visits if we are to expect

phy-sicians in practice to provide comprehensive

preven-tive care. Physicians are being encouraged to seize

every opportunity to immunize children (eg, at acute

and follow-up visits). However, it is possible that,

after having their children receive immunizations in

these situations, some parents may think that

sched-uled checkups are unnecessary. Further research is

needed on the impact of immunizing children at

non-well child visits to see if this practice leads to an

unlinking of immunization from other aspects of

preventive care. Helping parents understand the

im-portance of well child care must be an integral part of

efforts that practices make to improve their provision

of preventive care.

ACKNOWLEDGMENTS

Support for Dr Bordley’s work was provided by the Robert

Wood Johnson Clinical Scholars Program and by National

Re-search Service Award 5 T32 HS00032-05 from the Department of

Health and Human Services.

We thank Carol Porter and Joanne Garrett at the Sheps Center for Health Services Research for their assistance with data man-agement and analysis and Joseph Garfunkel, MD, and Jayne Stu-art, MPH, for their thoughtful editorial comments. We especially

express our gratitude to the staff of the 15 participating practices for allowing us into their offices and assisting us with this project.

REFERENCES

1. Centers for Disease Control and Prevention. Vaccination coverage of 2-year-old children-United States, 1992-1993. MMWR. 1994;43:

282-283

2. Zell ER,Dietz V, StevensonJ, Cochi S, Bruce RH. Low vacdnation levels of US preschool and school-age children. Retrospective assessments of vaccination coverage, 1991-1992. JAMA. 1994;271:833-839

3. Centers for Disease Control and Prevention. Retrospective assessment

of vaccination coverage among school-aged children-selected U. S. cities, 1991. MMWR. 1992;41:103-107

4. Abbotts B, Osborn LM. Immunization status and reasons for

immuni-zation delay among children using public health immunization clinics. Am J Dis Child. 1993;147:965-968

5. Farizo KM, Stehr GPA, Markowitz LE, Patriarca PA. Vaccination levels and missed opportunities for measles vaccination: a record audit in a public pediatric clinic. Pediatrics. 199289:589-592

6. Orenstein WA, Atkinson W, Mason D, Bernier RH. Barriers to vacci-nating preschool children.JHealth Care Poor Underseroed. 1990;1:315-330 7. Centers for Disease Control and Prevention. Early childhood

vaccina-tionlevels among urban children-Connecticut, 1990 and 1991. MMWR. 1992;40:888-891

8. Centers for Disease Control and Prevention. Vaccination coverage of 2-year-old children-United States, third quarter, 1993. MMWR. 1994;

43:556-559

9. Mustin HD, Holt VL, Connell FA. Adequacy of well-child care and immunizations in US infants born in 1988. JAMA. 1994;272:l111-1115

10. National Vaccine Advisory Committee. Special communication-the

measles epidemic: the problems, barriers, and recommendations.

JAMA. 1991;266:1547-1552

11. American Academy of Pediatrics. Policy Reference Guide of the American Academy of Pediatrics. Elk Grove Village, IL: American Academy of Pediatrics; 1995:673

12. Brown J, Melinkovich P. Gitterman B, Ricketts S. Missed opportunities in preventive pediatric health care. Am IDis Child. 1993;147:1081-1084

13. Rodewald LE, Szilagyi PC, Shiuh T, Humiston SC, LeBaron C, Hall CB. Is underimmunization a marker for insufficient utilization of preventive and primary care? Arch Pediatr Adolesc Med. 1995;149:393-397

14. American Academy of Pediatrics. Red Book: Report of the Committee on

Infectious Diseases. 23rd ed. Elk Grove Village, IL: American Academy of

Pediatrics; 1994:23

(7)

16. Centers for Disease Control and Prevention. Preventing Lead Poisoning in

Young Children: A Statement by the Centers for Disease Control. Atlanta:

Centers for Disease Control and Prevention; 1991

17. Hosmer DW, Lemeshow S. Model-bedding strategies and methods for

logistic regression. In: Applied Logistk Regression. New Yoric Wiley-Interscience; 1989. Wiley Series in Probability and Mathematical Statis-tics

18. Stats Corporation. Stata Reference Manual. Release 3.1, 6th ed. College Station, TX: Stata Corporation; 1993:405-414

19. Feinstein AR. Statistical indexes of association. In: Clinical Epidemiology:

The Architecture of Clinical Reseach. Philadelphia: ‘NB Saunders; 1985:

184-186

20. US Department of Health and Human Services. Healthy People 2000: National Health Promotion and Disease Prevention Objectives. Washington, DC: Government Printing Office; 1990. US Dept of Health and Human Services publication PHS 91-50213

21. Robinson CA, Sepe SJ, Un KF. The president’s child immunization initiative-a summary of the problem and the response. Public Health Rep. 1993;108:419-425

22. Starfield B, Seidel H, Gertrude C, Garvin W, SeddonJ. Private pediatric

practice: performance and problems. Pediatrics. 197352:344-350

23. Marks JS, Halpin TJ, Irvin JJ, Johnson DA, Keller JR. Risk factors

associated with failure to receive vaccinations. Pediatrics. 1979;64:

304-309

24. Szilagyi PG, Rodewald LE, Humiston SC, et al. Missed opportunities for childhood vaccinations in office practices and the effect on vaccination

status. Pediatrics. 199391:1-7. Published erratum appears in Pediatrics. 199391:545

25. Rodewald LE, Szilagyi PC, Humiston SC, et al. Is an emergency

de-partment visit a marker for undervacdnation and missed vaccination opportunities among children who have access to primary care? Pedi-atrics. 199391:605-611

26. Hueston WJ, Mainous AG, Palmer C. Delays in childhood

immuniza-tions in public and private settings. Arch Pediatr Adolesc Med. 1994;148: 470-473

27. Bureau of the Census. Statistical Abstract ofthe United States: 1992. 112th ed. Washington, DC: Bureau of the Census; 1992

28. State Center for Health and Environmental Statistics. North Carolina

Health Statistics Pocket Guide. Raleigh, NC: North Carolina Department

of Environment, Health, and Natural Resources; 1992

29. Bordley WC, Freed CL, Garrett JM, Byrd CA, Meriwether R Factors

responsible for immunization referrals to health departments in North

Carolina. Pediatrics. 199494:376-380

30. Centers for Disease Control and Prevention. Impact of missed

oppor-tunities to vaccinate preschool-aged children on vaccination coverage levels-selected US sites, 1991-1992. MMWR. 1994;43:709-718 31. Dietz VJ, Stevenson J, Zell ER, Cochi 5, Hadler 5, Eddins D. Potential

impact on vaccination coverage levels by administering vaccines simul-taneously and reducing dropout rates. Arch Pediatr Adolesc Med. 1994;

148:943-949

32. Gamertsfelder DA, Zimmerman RK, DeSensi EG. Immunization barn-era in a family practice residency dinic. IAm Board Fam Pract. 1994;7:

100-104

33. McConnochie KM. Roghmann KJ. immunization opportunities missed

among urban poor children. Pediatrics. 199289:1019-1026

34. Hutchins SS, Escolan J,MarkOWitZ LE, et al Measles outbreak among

unvaccinated preschool-aged children: opportunities missed by health care providers to administer measles vaccine. Pediatrics. 1989;83:

369-374

35. Gann P, Melville SK, Luckmann R. Characteristics ofprimary care office

systems as predictors of mammography utilization. Ann Intern Med. 1993;118:893-898

36. Lewis CE. Disease prevention and health promotion practices of pri-mary care physicians in the United States. Am JPrey Med. 1988;4(suppl):

9-16

37. Norman EH, Bordley WC, Hertz-Picciotto I, Newton DA. Rural-urban blood lead differences in North Carolina children. Pediatrics. 199494:

59-64

38. Ad Hoc Working Group forthe Development ofStandards for Pediatric

Immunization Practices. Standards for pediatric immunization

prac-tices. JAMA. 1993;269:1817-1822

39. Szilagyi PG, Roghmann KJ, Campbell JR, et al. Immunization practices ofprimary care practitioners and their relation to immunization levels. Arch Pediatr Adolesc Med. 1994;148:158-166

BABIES FOR SALE

Utah already has the highest birth rate in the nation, but Republican leaders here

have come up with a plan that may bring even more children into the Beehive

State: they want to pay unwed, pregnant women to carry their babies to term and

put them up for adoption.

The price for a baby in Utah will be about $3 000, paid to the mother by the state

and reimbursed by the adoptive parents.

Egan T. Take this bribe, please, for values to be received. New York Times. November 12, 1995.

Submitted by Student

(8)

1996;97;467

Pediatrics

W. Clayton Bordley, Peter A. Margolis and Carole M. Lannon