Improving the health of infants on medicaid by collocating Special Supplemental Nutrition Clinics with managed care provider sites

Objectives. This study tested whether collocation of Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) clinics at managed care provider sites improved health care for infants enrolled in Medicaid and WIC.

Methods. Weights and immunization rates were studied for the 1997 birth cohort of African American infants enrolled in WIC and Medicaid in Detroit, Mich. Infants using traditional WIC clinics and health services were compared with those enrolled under Medicaid in 2 managed care organizations (MCOs), of whom about half obtained WIC services at MCO provider sites.

Results. Compared with other infants, those who used collocated WIC sites either were closer to their age-appropriate weight or had higher immunization rates when recertified by WIC after their first birthday. Specific benefits (weight gain or immunizations) varied according to the priorities at the collocated sites operated by the 2 MCOs.

Conclusions. Collocation of WIC clinics at MCO sites can improve health care of low-income infants. However specific procedures for cooperation between WIC staff and other MCO staff are required to achieve this benefit, (Am J Public Health. 2002;92:399-403)

The Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) provides nutrition education and supplemental foods to low-income families across the United States. WIC increases rates of early prenatal care(n1) and reduces the frequency of low-birthweight infants.(n2-n6) Health care cost savings for pregnant women exceed the costs of their WIC benefits.(n7) Nourishment and early growth of infants or preschool children may be improved by WIC.(n8-n10)

WIC can refer clients to other programs.(n11-n14) Cross-referrals with Medicaid for services and sharing of records are specifically encouraged.(n15,n16) However, health care trader Medicaid depends increasingly on private managed care organizations (MCOs). Separation of managed care sites and public-sector support programs, including WIC, could decrease overall service delivery.(n17,n18)

WIC clients normally come for food vouchers every 2 to 3 months. Therefore, we wanted to test the hypothesis that provision of WIC services at managed care sites could improve health care. From a survey of state WIC directors in 1996 through 1997(n19) (also K. Bell, C. Hogue, A. Kendal, unpublished data, 1996-1997), we determined that such an evaluation might be done in Detroit, Mich, with a quasi-experimental approach. Identifying ways that the health of low-income persons might be improved in Detroit is important because Detroit has poor overall pregnancy outcomes and infant immunization rates.(n20,n21)

We studied the 1997 birth cohort of African American infants in Detroit, who were enrolled in both Medicaid and WIC. They constituted about 90% of the 6548 WIC infants receiving Medicaid. Only those infants enrolled in WIC by 22 weeks and reenrolling at 12 to 14 months without either apparent change in source of WIC or health care services or birth of a sibling were included. These limitations and other data editing (for obvious clerical errors) reduced the database to 4648 infants. We divided the infants into 5 groups for analysis (Table 1), according to their sources of health care (MCO A, MCO B, or neither) or WIC services (collocated at MCO A or MCO B or public health department sites). Clients of MCO A or MCO B who used their MCO's collocated WIC sites (groups 1 and 2) were considered the intervention groups. Groups 3 and 4 were clients of MCO A or B who obtained WIC services at Detroit Health Department clinics. These served as controls for groups 1 and 2. Group 5 constituted an additional control group of infants not enrolled in MCO A or B.

Mothers in groups 1 and 2 were more likely than those in the other groups to be welfare recipients (receiving food stamps or cash payments) and less likely to have only 1 child in the household enrolled in WIC (Table 1). Mother's ages and breastfeeding rates (< 10% to 13 weeks for all groups) were not, however, different (data not shown).

Detroit Health Department sites. Although mothers visited a WIC clinic every 3 months for food vouchers, WIC infants were required to be brought in, by appointment, only at about 6 and 12 months for growth monitoring and immunization assessment. Immunization data were recorded into a WIC data system that reported which vaccines were due. To be considered up to date by WIC, a 12- to 14-month-old child would need to have completed the primary series for diphtheria, pertussis, tetanus, polio, hepatitis B, Haemophilus influenzae B, measles, mumps, and rubella. Only 1 clinic (included in group 5) limited WIC vouchers when infants' immunizations were not up to date as an incentive to raise vaccination levels.(n11)

Collocated sites. Requirements for WIC clinics operated through MCOs A and B were the same as those at Detroit Health Department WIC clinics, with the exception that MCO B issued food vouchers for 2-month periods and did not conduct immunization assessments dining WIC visits.

Additional differences between collocated sites were as follows:

• MCO A operated 4 large clinics at which WIC services were available daily and had a network of contracting physicians not offering WIC. When possible, WIC staff collocated at the large MCO clinics sent infants needing immunizations to the MCO's pediatric clinics before completing the visit and records.

• MCO B used only contracted clinics and physicians to provide health care and employed a WIC team that had a regular schedule of visits to about half of these sites. (We could not separately analyze the MCO B clients whose health care provider was not one visited by the WIC team.) The WIC team members attached a special form to the client's WIC file when they believed that the client was at special risk. Files were reviewed by the MCO's nutrition program supervisor, who either made an appointment to see high-risk clients or coordinated with another MCO employee responsible for arranging extra medical or social services.

WIC-Medicaid-MCO data linkage. At 6-month intervals, state information system staff compared Medicaid identification numbers, social security numbers, and names and dates of birth for persons included in the research database with those for persons enrolled in Medicaid during similar periods. This process resulted in validation of Medicaid identification numbers in WIC records and addition of some missing ones. Fewer than 1% of the Medicaid identification numbers in the WIC records contained clerical errors (e.g., transposed or missing digits). Medicaid identification numbers of MCO A and B clients provided every 6 months were linked to the database. The clients' WIC clinic identification numbers were then used to determine whether those clinics were MCO-collocated sites. The reports prepared only aggregated data with no individual client identification were contained.

In regression analysis, the main independent variable was the client's categorization into the 5 groups shown in Table 1. This "MCO variable" was included regardless of its statistical significance. The effect of the MCO variable on outcomes of "midyear evaluation," "first-year weight gain," and "immunization status" was examined, controlling for welfare status, having other children in the household enrolled in WIC ("family size"), and other variables as described in the "Results" section.

The backward elimination method was used to determine the final regression model, with progressive elimination of whichever variable (other than the MCO variable) had the highest P value greater than .05 until each final model contained only the MCO variable and any other variable significant at P<.05.

Outcomes were adjusted for the effect of significant variables. Multiple linear regression analysis was performed for the continuous variable "first-year weight gain," and significance of differences in weight gains among MCO groups was determined by least squares means comparisons. Logistic regression analysis was used for the discontinuous outcomes, and the significance of the effect of the MCO group on the outcome was determined by calculating the 95% confidence intervals (CIs) on the odds ratio (OR) for that outcome's occurrence in each of the possible pairs of MCO groups.

In the summer of 1999, we surveyed mothers of 1- to 2-year-old children at WIC clinics in Detroit, including collocated sites. The subset of questions analyzed had elicited individual response rates of greater than 75% from the 842 clients self-reporting that they were both African American and receiving Medicaid (like the mothers of the infants included in this study).

Although most infants were enrolled in WIC in the first 6 weeks after birth, considerably fewer of these infants appeared for their midyear evaluation at 22 to 40 weeks than appeared for WIC reenrollment after their first birthday (Figure 1). About 75% of the WIC infants receiving Medicaid reenrolled in WIC at 12 to 14 months, with similar rates in all 5 study groups (range=710/o-77%, P> .05). However, longitudinal analysis showed that whereas midyear evaluation visits occurred for 75% and 91% of the infants in groups 1 and 2, respectively (collocated WIC sites), only 56% to 60% of those in the other groups had these visits (Table 2). Rates were adjusted in regression analysis for the variables "welfare status" and "family size." Differences in the rates of midyear evaluation visits between the 2 collocated groups (groups 1 and 2) and between either of these collocated groups and the other groups (groups 3, 4, or 5) were all statistically significant, and group 2 retained its rank as having the most midyear evaluations performed (i.e., group 2 compared with individual groups 3, 4, or 5, 0R=7.04, 8.13; 95% CI=4.52, 12.5).

Mean birthweights were very similar among the 5 groups (3141-3170 g, P>.05), as were mean ages at which recertification for WIC occurred (13.3-13.7 months, P>.05). However, weights and weight gains by the time of recertification for WIC at ages 12 to 14 months were highest for group 2 (MCO B collocated infants; Table 2). This finding could not be accounted for statistically in regression analysis by the variables "age when reenrolled in WIC," "welfare status," or "breastfeeding to 12 weeks." Weight gains were adjusted for statistically significant confounding variables, which were "family size," "birthweight," and "midyear evaluation by WIC." After adjustment (Table 2), MCO B collocated infants were found to have gamed about 400 to 450 g more weight than did those in the other groups (P<.001 in least squares means analysis).

For the overall population of infants enrolled in WIC, having a midyear evaluation by WIC increased the rate of up-to-date immunizations at 12 to 14 months (OR=2.56; 95% CI= 2.04, 3.21) (data for combined groups not shown). About 90% of the immunization assessments recorded used infant immunization records as the data source (data not shown).

For all individual groups except group 2, immunization assessments were recorded by WIC for 62% to 73% of the 12- to 14-month-old infants (Table 2). Group 1 infants had the highest rate of up-to-date immunizations of the 5 groups (61% vs 39%--46%; Table 2). This difference was statistically significant when group 1 was compared with groups 4 and 5 (ORs= 1.50 and 1.65; 95% CIs from 1.08 to 2.13). The difference between MCO A groups 1 and 2 was close to significant (OR= 1.27; 95% CI=0.91, 1.75).

When the mothers of the 1- to 2-year-old African American children receiving Medicaid were surveyed in mid-1999, 50% of those at MCO A collocated sites reported obtaining immunizations for their child while visiting a WIC clinic, compared with fewer than one fifth of the WIC clients at the other sites (Table 2). Of the clients at MCO A collocated sites, 91% liked receiving assessment and delivery of immunizations in conjunction with WIC visits, compared with 59% to 71% of the clients at other locations. The differences for MCO A collocated sites compared with all other sites were significant (P<.005).

This study tested the hypothesis that providing WIC at managed care sites (collocation) improves the health care of WIC clients receiving Medicaid. The 3 outcomes measured (WIC midyear visits for evaluation of infants' growth, infants' weight gain at 1 year, and immunizations at 1 year) showed that instances in which better results were found always corresponded to collocated sites. We attempted to assess the factors responsible. Higher rates of WIC evaluations at collocated sites might result from using motivational strategies or having adequate staff and space to schedule more evaluations, factors that could not be reliably assessed. However, possible reasons for other outcome differences were found when we compared processes at the sites.

Mean weights of infants in the 4 study groups were 550 g below 10.25 kg, the 50th-percentile weight for 13.5-month-old infants,(n22) the average age of infants being recertified for WIC. The mean weight was only 150 g below the age-appropriate weight for infants in the group in which more than 90% were evaluated by WIC at about 6 months and whose WIC staff cooperated with MCO staff to ensure that extra visits with nutritionists or other referrals were arranged for clients at greatest risk. This suggests that active efforts by WIC staff to evaluate infants' growth at about 6 months, and to pay attention to special nutritional or other needs identified at that time, may result in healthier 1-year-old infants.

Observed immunization rates are subject to several caveats. The immunization schedule is complex, and errors may occur during immunization assessments.(n23) Although the WIC clinics we studied used a computer module for immunization assessment, and immunization cards were used more than 90% of the time in the assessments, higher rates might be found from reviews of medical records. Also, the immunization rates we calculated were for infants aged 12 to 14 months. They are not comparable to estimates for children older than 2 years, such as those from the Centers for Disease Control and Prevention's National Immunization Survey.

Despite these caveats, findings that used WIC records appeared to be logical. First, infants in Detroit had more immunizations if they were evaluated at about 6 months by WIC, whether at collocated MCO sites or at traditional WIC clinics. This finding supports other evidence that raising the number of health-related visits increases implementation of preventive measures and confirms the value of involving WIC in health care beyond nutrition.(n11,n24-n26)

Second, the highest immunization rates were for the group of infants at the collocated WIC clinics of one MCO at which special efforts were made to improve access to immunizations from pediatric health care staff on site at the time of the WIC evaluations. This is consistent with our previous report (which did not include findings from WIC sites collocated with MCOs) that WIC clients who used Detroit Health Department sites had more immunizations if the Detroit Health Department provided immunizations on site at times when WIC evaluations were scheduled.(n27)

Both MCO A and MCO B missed opportunities to improve the health of their clients. MCO B failed to design ways to deliver immunizations at visits to collocated MCO B sites, and MCO A failed to identify other types of support, including extra nutritional counseling or intervention, for high-risk clients identified at its collocated sites. Taking advantage of opportunities created by collocation requires the combination of investment in resources and in management time by both the WIC program and the health care programs (including Medicaid contractors or health departments).

The budget for the federally funded WIC programs is based on the number of clients served rather than on the outcomes achieved, and resources may be scarce for nonnutrition services such as immunization assessment. In the absence of changes in WIC funding by the US government, state and local health departments or the private-sector MCOs contracting with Medicaid could probably improve their clients' health care by providing WIC clinics with nurse practitioners or other health professionals who can deliver preventive services to clients.(n28) Such use of non-physician practitioners by MCOs may be criticized as encouraging mothers to forgo regular physician visits for their infants, with possibly negative overall consequences. That risk is somewhat offset by probability, as noted by Szilagyi et al., that the Vaccines for Children Program may have met one of its goals of encouraging more low-income mothers to use a "medical home" for preventive services such as child immunizations? Nevertheless, as noted in that report,(n29) local situations differ, and Detroit may be a location in which the "medical home" concept has yet to succeed, as judged by its lack of increased immunization rates in the past several years.

The current results provide a rationale for health departments and MCOs to cooperate in service delivery in order to improve client health and satisfaction. Linking preventive medical services (for which appointments often are not kept) with the supply of infant food (which is highly sought after by low-income mothers) could provide lifelong benefits for many infants by improving not only immunization rates(n30) but also the use of other important preventive health services such as early and periodic screening, diagnosis, and treatment.(n31)

Requests for reprints should be sent to Alan P Kendal, PhD, Emory University, Rollins School of Public Health, 1518 Clifton Rd, Atlanta, GA 30322 (e-mail: apkenda@ sph.emory.edu).

This article was accepted July 5, 2001.

A.P. Kendal was principal investigator and oversaw analyses. A. Peterson assisted with study design and organized provision and assembly of data by the state of Michigan. C. Manning and F. Xu provided oversight of data management and performed statistical analyses. L.J. Neville organized surveys of WIC clients and provision of information on clinic management, C. Hogue participated in study design and evaluation. All authors were substantially involved in the design and performance of the study and in preparing analyses and interpretations.

Financial support was provided by the National Immunization Program, Centers for Disease Control and Prevention, through a Cooperative Agreement with the Association of Schools of Public Health and by a contract between the Michigan Department of Community Health WIC Program and the Michigan Public Health Institute.

We wish to thank the following people for their cooperation and support: Dave Sachau, WIC Information Specialist, Michigan Department of Community Health WIC Program, for providing WIC data and for organizing linkages of WIC records with Medicaid beneficiary records in Michigan; Dr T. Hershel Gardin and Dr Mark Kashishian of The Wellness Plan, Janet Hunter of the Detroit Urban League, Kathy Smith, RD, of The Omni-Care Health Plan, and Nancy Erickson of the Detroit WIC program, as well as all their colleagues, who helped obtain data and information; and Rebecca Zhang of the Rollins School of Public Health of Emory University for providing additional statistical help.

GRAPH: FIGURE 1--Detroit Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) visits (excluding those for only food vouchers) by infants younger than 60 weeks elegible for WIC in the first quarter of 1997.