Erroneous Payments In Childcare Centers Study (EPICCS)

The summary below presents key findings that are reliable and rigorous, but relate to broad trends instead of specific estimates. For transparency and scientific integrity, we have published the study data on AgData Commons (these data are accurate and may be used by independent researchers). We will use the lessons learned during EPICCS to plan upcoming research to evaluate payment errors in the child nutrition programs.

Key Findings

The payment error patterns found in EPICCS resembled those found in the Access, Participation, Eligibility and Certification (APEC) study series. The most notable difference between the certification processes used in the Child and Adult Care Food Program (CACFP) and those used in the school meal programs examined by APEC is that CACFP does not directly certify children based on their household’s participation in other programs. Direct certification is more likely to result in correct certification than household applications.

In the EPICCS study sample:

• Certification error was highest for children certified for reduced-price meals and lowest for children certified for free meals. Certification error was about as likely to result in overpayment as underpayment.
• Household reporting error was a more common source of certification error than administrative error.
• Household reporting error led to overpayment more often than underpayment. The most common source of household reporting error was incorrect reporting of household income. Households may make reporting errors for a variety of reasons. Another common source of household reporting error was incorrect reporting of household size. This led to underpayment more often than overpayment, because households tended not to report everyone living in the home.
• Administrative error was more common for children certified for paid meals than for children certified for free or reduced-price meals.
• A significant number of child care centers used center-level meal counting methods instead of child-level meal counting methods. As expected, payment error was lower in these child care centers (because certification errors for specific children offset one another at the level of the center).
• The aggregation error rate was very small. Most child care centers claimed the correct number of meals for reimbursement, while a small number of child care centers had relatively high error rates. As expected, error decreased with higher levels of aggregation. In part, this is because aggregation errors are net errors, but it also indicates that sponsors do a good job auditing and correcting their centers’ meal counts.
• When child care centers didn’t meet USDA meal pattern requirements, it was more often because they served the wrong portion sizes than because they had incorrect or missing meal components. The meal component that was incorrect most often was whole milk instead of low-fat or non-fat milk.
• Breakfast meals were less likely to meet USDA meal pattern requirements than lunch meals. Meals served in child care centers with 3- to 5-year-old children, mixed ages eating together, and family-style service (as opposed to cafeteria-style service) were less likely to meet USDA meal pattern requirements than other child care centers.

Why did we do this study?

The Payment Integrity Information Act of 2019 requires federal agencies to give Congress information about payment errors for federal programs, like CACFP. CACFP operates in several types of outlets, which serve nutritious meals and snacks to the children and adults in their care. Each month, we reimburse CACFP outlets for those meals and snacks. Payment errors happen when outlets are reimbursed too much (overpayments) or too little (underpayments).

It's difficult to estimate payment errors for CACFP. Program rules differ for each outlet type so payment errors must be estimated differently for each outlet type. We planned EPICCS to estimate payment errors in CACFP child care centers.

How did we do this study?

EPICCS included a nationally representative sample of sponsored, independent, and Head Start child care centers serving 2- to 5-year-old children, and the children’s households. EPICCS did not estimate payment errors for other CACFP outlets (Table 1).

Table 1. CACFP outlet and meal types included in and excluded from the Erroneous Payments in Childcare Centers study (EPICCS).

	Included in EPICCS	Excluded from EPICCS
Outlet Type	Child care centers serving 2- to 5-year-olds: Sponsored centers Independent centers Head Start centers	Infant care centers Family day care homes At-risk afterschool centers Outside school hours centers Adult care centers Emergency and homeless shelters
Meal Type	Breakfast Lunch	Snacks Supper

The amount of money FNS reimburses child care centers for meals depends on two main factors (Table 2):

• Meal type: Child care centers that participate in CACFP may serve breakfast, lunch, supper, and snacks. Snacks have the lowest reimbursement rate, followed by breakfast, followed by lunch and supper (which have the same reimbursement rate).
• Children’s certification for free, reduced-price, or paid meals: Children are eligible for free, reduced-price, or paid meals based on their household income. (Children are eligible for free meals when their household has an income at or below 130% of federal income poverty guidelines. Children are eligible for reduced-price meals when their household has an income between 131% and 185% of federal income poverty guidelines. Children at Head Start centers are automatically eligible for free meals.) Child care centers or their sponsoring organizations certify children as eligible for free, reduced-price, or paid meals based on information households report on an application.

Reimbursement rates are adjusted every year for inflation. EPICCS examined payment errors made in 2017, which was the year the study collected data from child care centers.

Table 2. The amount of money we reimbursed for child care centers for CACFP breakfasts and lunches in 2017.

Meal Type	Reimbursement for a Meal Served to a Child Certified for Free Meals	Reimbursement for a Meal Served to a Child Certified for Reduced-Price Meals	Reimbursement for a Meal Served to a Child Certified for Paid Meals
Breakfast	$1.71	$1.41	$0.29
Lunch	$3.16	$2.76	$0.30

In CACFP, payment errors are made due to certification error and aggregation error. These two sources of payment errors are described below.

A certification error is when a child’s eligibility and certification differ. EPICCS measured certification error in a sample of 5,710 children by comparing two kinds of information:

1. The children’s certification: EPICCS asked child care centers or their sponsoring organizations which children they certified as eligible for free meals, reduced-price meals, and paid meals (Figure 1, Part A). The child care centers and sponsoring organizations certified children using information on household applications (Figure 1, Part B).
2. The children’s eligibility: EPICCS collected information from households to verify each child’s eligibility for free, reduced-price, or paid meals (Figure 1, Part C).

A certification error can happen when a household reports inaccurate information on their application (household reporting error) and when a center or sponsor certifies a child inaccurately using the household application (administrative error).

Aggregation errors happen when sponsor and state meal counts differ. Each month, child care centers add up the number of free, reduced-price, and paid meals they serve. Sponsors add up the meal counts from their centers, and states add up the meal counts from their sponsors to send to FNS for reimbursement (Figure 2). Aggregation errors are net errors rather than gross errors because meal counts that are too high and too low offset each other at each step.

EPICCS measured aggregation error by comparing the meal counts that 131 sponsors sent to their state agencies with the meal counts that state agencies submitted to FNS. EPICCS used sponsor-level meal counts instead of center-level meal counts because sponsors audit their centers’ meal counts and make corrections.

The study team calculated payment error in the study sample using the certification error and aggregation error measurements. They also calculated the payment error rate (the percent of FNS reimbursements that were overpayments and underpayments) in the study sample.

In addition to measuring payment error, EPICCS measured the percent of CACFP meals that didn’t meet USDA meal pattern requirements, which is a standard of service error. EPICCS observed the breakfast and lunch meals that 405 child care centers served and noted when the centers served the wrong portion sizes, served meals that were missing required components (for example, milk), or served components didn’t meet USDA standards (for example, whole milk instead of low-fat or non-fat milk).

Since EPICCS collected data from a sample of CACPF child care centers, the study team needed to scale up the findings to produce national estimates. The study team’s weighting method used national data that states report to FNS about child care.

After EPICCS data collection and analysis were completed, the Study of Nutrition and Activity in Childcare Settings in USDA’s CACFP (SNACS) was published. SNACS examined similar CACFP-participating child care providers to EPICCS, and revealed new information that was critical for estimating certification error. EPICCS originally assumed that child care centers count the number of meals FNS must reimburse at the free, reduced-price, and paid meal rates at the level of the child. Thanks to the SNACS data, we learned this was an incorrect assumption. Some child care centers count the number of meals for reimbursement at the level of the center. Child-level meal counts and center-level meal counts have different effects on payment errors:

• Child-level meal counts: These centers use the certification of each child to count the number of meals FNS must reimburse at the free and reduced-price and paid rates. In these child care centers, every certification error causes an overpayment or underpayment; in other words, payment errors in these child care centers are gross errors (Table 3).

Table 3. Payment errors are gross errors in child care centers with child-level meal counts.

Child	Child's Certification	Child's Eligibility	Payment Error for Lunch?
1	Reduced-price meals (reimbursement rate: $2.76)	Reduced-price meals (reimbursement rate: $2.76)	No FNS correctly reimburses the child care center $2.76 for this child’s meal.
2	Reduced-price meals (reimbursement rate: $2.76)	Free meals (reimbursement rate: $3.16)	Yes FNS underpays the child care center $0.40 for this child’s meal.
3	Free meals (reimbursement rate: $3.16)	Reduced-price meals (reimbursement rate: $2.76)	Yes FNS overpays the child care center $0.40 for this child’s meal.
Total payment error: $0.80

• Center-level meal counts: These centers use the proportion of children who are certified for free, reduced-price, and paid meals to claim the number of meals FNS should reimburse at each rate. In these child care centers, certification errors for specific children can offset one another at the level of the center; in other words, payment errors in these child care centers are net errors (Table 4).

Table 4. Payment errors are net errors in child care centers with center-level meal counts.

Child	Child's Certification	Child's Eligibility	Payment Error for Lunch?
1	Reduced-price meals (reimbursement rate: $2.76)	Reduced-price meals (reimbursement rate: $2.76)	No FNS correctly reimburses the child care center $8.68 for 2 reduced-price meals and 1 free meal.
2	Reduced-price meals (reimbursement rate: $2.76)	Free meals (reimbursement rate: $3.16)
3	Free meals (reimbursement rate: $3.16)	Reduced-price meals (reimbursement rate: $2.76)
Total payment error: $0.00

In 2023 (after the COVID-19 pandemic), the EPICCS Adjustment study went back and asked EPICCS child care centers whether they used child-level or center-level meal counts in 2017. The study team used this information to recalculate payment error due to certification error. As expected, the adjusted payment error rate was lower than the original payment error rate.

After completing the EPICCS Adjustment study, we found a weighting error that impacted the EPICCS and the EPICCS Adjustment study findings. The study team incorrectly included some outlets that EPICCS excluded in the weighting calculations. As a result, the national estimates in EPICCS and the EPICCS Adjustment study are incorrect.

This study is in the FNS Study & Evaluation Plan for fiscal year 2014, where it is called Improper Payments in CACFP Centers.