Analysis of National Background Check Program (NBCP) Registry Checks

I. Introduction
II. General methodology for this report
III. State Registry Use
IV. The Registry Checking Process in Selected States
V. An In-Depth Look at Registry Results in Selected States
VI. Efficiency in Registry Checking
VII. Summary of Findings
VIII. Recommendations for States
Appendix: Registries Used in the Background Check Process by State

I. Introduction

The Affordable Care Act requires that State background check programs “utilize a search of State-based abuse and neglect registries and databases, including the abuse and neglect registries of another State in the case where a prospective employee previously resided in that State.” The Affordable Care Act also requires States to search “the records of any proceedings in the State that may contain disqualifying information about prospective employees (such as proceedings conducted by State professional licensing and disciplinary boards and State Medicaid Fraud Control Units).”

Accordingly, the NBCP grants also require checks of registry databases. Registry checks are a quick and low-cost or free source of relevant information about an applicant for long term care (LTC) employment. While the NBCP specifies some registry checks—for example, the Office of Inspector General’s (OIG) List of Excluded Individuals and Entities (LEIE)—it defines others more generally, giving States some discretion in their implementation of registry checks.

In a study performed by CNA in 2016, we document the use, efficiency, and effectiveness of registry checks under the NBCP. By presenting the results of this study, we hope to:

• help States understand the value registries provide, and potentially identify enhancements that can increase the efficiency and effectiveness of the NBCP program;
• help the Centers for Medicare & Medicaid Services (CMS) develop data-driven guidance on how States can best implement registry checks; and
• provide registry owners feedback on the usage and value of their data.

II. General methodology for this report

To complete this analysis, we collected two sets of data – one set from all States and a second set from a subset of States selected for in-depth analysis of their registry checking processes and results. From all States, we collected a list of registries used in registry research; these registries are the ones that appear regularly on their background check system’s (BCS’s) research registries page. State Liaisons collected this information directly from the States or verified the information if it was obtained from other sources.

Four States were selected for in-depth analysis. We selected these States because they all have mature programs that use the core BCS developed through the technical assistance (TA) contract; thus, data elements and definitions would be similar across the four States. While having a similar structure for their BCSs, they implemented registry checks in a variety of ways, using different types of registries and different registry checking processes. This variety gave the analysis a broad look at how States conduct registry checks.

We were able to collect data from three of the four selected States. One State was unable to provide the data we requested as State IT staff did not have time available to fulfill the request.

We asked the selected States to provide an analytic data file from their BCS database. A query was created by the TA team that States could run against their databases to produce the data file, allowing States to produce the requested data file with little or no programming effort on their part. The data file produced from the query contained all records from several BCS data tables, including the application, determination, registry, registry determination, employment, and appeal tables. No personally identifiable information was requested or obtained.

To assist the data analysis, a data dictionary was completed that listed all data elements in the data tables obtained from the selected States and gave a definition for each. Data was analyzed at the individual registry check level.

With the help of State program staff, we also developed a profile documenting registry procedures for each participating State.

III. State Registry Use

Data on the registries used for registry research were collected from all 23 States with current NBCP grants plus one graduated State – the State of Illinois.¹ At the time of this analysis, five of these States (Hawaii, Kansas, Maine, Puerto Rico, and Rhode Island) had not yet implemented a BCS with registry checks. The appendix presents the list of registries used by each State.

States with registry checking regularly use on average between five and six registries for registry research. This does not include out-of-State registries that appear on the research registries screen only when an applicant has disclosed a prior out-of-State address. All States conducting registry checks use both federal and in-State registries, but they differ in the number and types of registries checked. The number of registries checked varies from a low of three to a high of 10.

States use the following federal registries:

• Office of Inspector General (OIG) List of Excluded Individuals and Entities (LEIE);
• National Sex Offender Public Website (NSOPW);
• General Services Administration (GSA) System for Award Management (SAM—formerly, the Excluded Parties List System); and
• Social Security Administration (SSA) Death Master File.

All but one of the 18 States with registry checking use federal registries.² Table 1 shows the number of States that use each of the federal registries.

Table 1. Number of States Using Each Federal Registry

For Table 2, in-state registries have been categorized into 14 registry types, and the table shows the number of States that use each registry type for registry research. Almost all States (16 of 18) include the State’s nurse aide registry (NAR) in their list of registries that must be checked. The second most used registry type is State sex offender registries. Note that there is both a federal sex offender website (which provides a search of sex offender data from all States) and individual State sex offender registries. Four NBCP States include mandatory searches of both the federal website and their State’s sex offender registry as part of their registry research. Four States use the State sex offender registry only, six use the federal registry only, and four States use neither the State nor the federal registry (data not shown). One advantage of the State sex offender registries is that often their searches can be automated, while the federal registry must be searched manually.

Table 2. Number of States that Use In-State Registries by Registry Type

Note that one State includes a link to the State’s motor vehicle database on its research registries screen. Checking this database is optional for providers, and the link is provided as a convenience to providers who want to conduct this check.

IV. The Registry Checking Process in Selected States

This section of the report provides a detailed look at the registry checking process in the States that provided data for the in-depth analysis. We gave States a set of questions about how registry checks are conducted and asked the States to provide answers so that we could obtain a complete profile of each State’s registry checking process. Details of the registry checking process in each State are described below.

A. Who checks the registries?

• State 1: Provider users check the registries.
• State 2: Primarily provider users check the registries, but State users will check any registries that providers flag for review. State users will also do the registry checks for any application that is initiated outside of the BCS. About 20 percent of providers in State 2 (mostly small, rural providers or Amish providers) initiate a background check manually.
• State 3: State users conduct all registry checks.

B. Where in the new applicant workflow (NAW) is registry checking completed?

Table 3 shows the answer to this question for the selected States.

Table 3. Steps in the NAW, by State

In both State 1 and State 2, registry research is part of the application process. Registry research is completed before the application is submitted and fingerprints are taken. In State 3, payment must be made and fingerprints taken before the State begins registry research for an applicant.

C. Which registries are checked, and are they checked manually or automatically?

Table 4 shows the answers to this question for each of the selected States.

Table 4. Federal and State Registries Checked, by State and Type of Check (Manual or Automatch)

All three States use the same federal registries—OIG LEIE and NSOPW. The OIG LEIE check is done using an automatch process in States 2 and 3, but is a manual check in State 1. When a registry check is manual, the BCS provides a link to the registry’s website. Providers must click on the link and complete the registry check on the registry’s website. When a registry check is conducted via automatch, the BCS uses already entered applicant demographic information to automatically search the registry. Automatches can be conducted using a stored procedure (that is, the registry database is downloaded into the BCS making the registry a “local” database) or a web service (which calls a registry website in real time, sends applicant information for matching, and returns match information if found). The automatch process works best when an exact match can be made using Social Security Numbers (SSNs). Without SSNs, the match process returns all records that are a possible match based (usually) on last name and first name or last name and date of birth. These “fuzzy” matches can result in many possible matches, especially for applicants with common last names.

D. Are all checks mandatory?

• State 1: Yes
• State 2: Yes
• State 3: Yes, except that State 3 does not conduct registry checks for submitted applications that connect to already eligible determinations. That is, when an individual applies to a second employer and already has an eligible determination, the registry checks are not redone. The registry checks completed for the first application are “reused” for any connecting applications as long as the first application has an eligible determination.

E. What are the drop-down choices for registry results?

• State 1: Users can mark the registry as Cleared or Not Cleared.
• State 2: Users can mark the registry as Cleared, Not Cleared, or Flagged for Review.
• State 3: Users can mark the registry as Cleared, Not Cleared, or Information–Not Disqualifying.

F. Are out-of-state registry checks mandatory when a prior address is listed?

• State 1: Yes. Providers are asked to list applicant’s addresses for the previous seven years.
• State 2: Yes. Providers are asked to list applicant’s addresses for the previous seven years.
• State 3: Yes. No limit is imposed on how far back prior address information can be recorded.

G. Is registry information used in the determination process?

• State 1: No. Applications are closed after a Not Cleared registry result. Applicants with Not Cleared registry results are not fingerprinted and no determination is made for these individuals.
• State 2: No. Applications are closed after a Not Cleared registry result. Applicants with Not Cleared registry results are not fingerprinted and no determination is made for these individuals.
• State 3: Yes. Negative registry findings can result in an “ineligible” fitness determination.

H. Can registry results be appealed?

• State 1: No. Applicants must appeal registry results to the owner of the registry.
• State 2: No. Applicants must appeal registry results to the owner of the registry.
• State 3: Yes. Applicants can request a variance for negative registry findings. The variance is a rehabilitation review. It does not remove the negative findings from registries.

State registry checking processes can change over time. States have added new registries as they have become available. State 3 added the State’s Medicaid exclusion list and Child Protective Services registries. State 1 has removed registries from its list of mandatory registry checks. Initially, State 1’s registry checking process included NAR and professional license database checks from border States. These databases were checked regardless of the applicant’s current or prior addresses. This step proved to be too time consuming, so the border State databases were removed from the default list. Also removed from its list of mandatory registry checks is State 1’s professional licensure database. State 3 is in the process of changing its OIG LEIE registry check from a manual to an automatch search, while the State 1 went the opposite way with the OIG LEIE; it started with an automated search but felt that the search returned too many potential matches that were time consuming to resolve, so it switched to a manual search.

V. An In-Depth Look at Registry Results in Selected States

An analysis of State data was conducted at the individual registry check level. The data allowed us to review how often individual registries are checked and the results of those checks. Table 5 presents some descriptive data from the States’ analytic data files. The time period for the data differs by State as the analysis included as much data as possible from each. The total number of applications varies considerably as well, reflecting both the size of the State and the number of years of data.

Table 5. Average Number of Registry Checks per Application

States 1, 2, and 3 currently check three, four, and 10 registries, respectively, as part of their background check processes. Frequent checking of out-of-state registries in State 1 resulted in an average number of registry checks per application of 4.7. Out-of-state registries are checked much more frequently in State 1 than in State 2 and State 3, which makes sense as State 1 has a metropolitan area that includes parts of neighboring States. Although State 3 checks 10 registries, the average number of checks per application is 7.4. The average number of checks per application is lower than the number of registries checked because no registries are checked for submitted applications that connect to already eligible determinations.

A. Federal registry results

Results from checks of federal registries in our three States are shown in Table 6. The table shows the count of Cleared and Not Cleared results only. For State 3, Cleared results include checks with Cleared and Information—Not Disqualifying results. (The data also include small numbers of pending, flagged for review, and automatch error results.) In each State, checks of federal registries result in a small number of disqualified applicants.

Table 6. Federal Registry Results

Note: Pending, flagged for review, and automatch error results not shown.

B. In-state registry results

Table 7 shows Cleared and Not Cleared results for in-state registries. Registries are categorized by type. The only type of registry that is used in all three States is the State NAR. State 3 checks several types of registries that are more difficult to check, but that result in higher numbers of Not Cleared results including court, offender, Child Protective Services, and name-based criminal history.

Table 7. In-State Registry Results

C. Out-of-state registry results

Table 8 presents results from out-of-state registry checks. Out-of-state registries appear on the research registries screen of the NAW when an out-of-state prior address is listed for an applicant or when a user adds the registry through the “research registries not listed” function. We looked at these registry results by whether or not the registry was from a border (or close) state. In State 1, more checks of out-of-state registries were of border States (88 percent) than non-border States. In State 2 and State 3, 40 and 6 percent of out-of-state registry checks were of border State registries, respectively.

Table 8. Out-of-State Registry Results

We also looked at out-of-state registry results by whether or not the applicant had a prior address listed. Again, State 1 is not like the other States. Seventy-one percent of out-of-state checks were conducted for applicants without a prior address in State 1, whereas just 29 percent of out-of-state checks were conducted for applicants without a prior address in State 2 and 4 percent in State 3. At one point, State 1 required neighboring State registry checks of all their applicants, and this is undoubtedly why so many out-of-state checks were done without a prior address. No Not Cleared registry results were found in State 2 or State 3 for out-of-state registry checks when the applicant had no prior address in the State. But in State 1, two-thirds of the Not Cleared results were obtained when the applicant did not have an out-of-state prior address.

D. Overall Results

Table 9 summarizes the registry check results across all registries and compares the number of individuals with a Not Cleared registry check to the number of Not Eligible determinations made in each State. In State 1, the registry checking process resulted in 69 Not Cleared results for 55 individuals. Applications for these individuals were closed prior to a determination being made. In the same time period, State 1 made a total of 148 Not Eligible determinations based on criminal history.³ Therefore, a total of 203 individuals were prevented from employment in LTC in State 1 due to the background screening process, and registry checking accounted for 27 percent of this total. While a Not Cleared finding on a registry check may happen rarely, registry checks can still add significantly to the number of applicants prevented from being employed in LTC due to disqualifying events in their backgrounds.

Table 9. Overall Registry Results

In State 2, 46 Not Cleared registry results were found for 37 applicants. One of these applicants had a Not Eligible determination, but all others had applications that were closed prior to a determination being made. Applicants denied employment due to a registry check represented about 1.6 percent of all applicants prevented from employment in LTC. In State 3, where registry results are used in the determination process, the 318 individuals with a Not Cleared registry check also had a Not Eligible determination. We could not identify individuals whose Not Eligible determinations were due solely to registry results using the State 3 data, so we don’t know the extent to which registry checks added to the number of individuals who were prevented from being employed in LTC.

VI. Efficiency in Registry Checking

To determine efficiency in registry checking, we analyzed the amount of time it takes to complete registry research based on start and complete dates and times recorded. Figure 1 shows the distribution of applications by the amount of time spent on registry research in State 2. (The curve is similar in State 1.) With 4.1 registries checked per application (three of them being automatched), the median⁴ State 2 user spends about one minute on registry research. When an application is closed due to the registry check results, the median users spends 7.7 minutes on registry research. In State 1, the respective numbers are 2 and 9 minutes. State 1 users check on average 4.7 registries per application, and only one is automated.

Figure 1. Amount of Time Spent on Registry Research in State 2 by Application

State 3 has a much different registry checking process, and the State’s data for registry research started and completed dates and times were incomplete. Many applications were missing data; for those with data, the average amount of time spent on registry research was measured in days.

VII. Summary of Findings

Findings from our analysis of registry use in NBCP States are summarized as follows:

• States have discretion in implementing registry checks, and they use it. The number and type of registries checked vary widely across the States.
• The registries most often used by the States are OIG LEIE, State NAR, and NSOPW.
• Registry checking processes can also vary widely. However, what is similar across the three States whose processes we reviewed is that they all use some combination of State and federal registry checks, checks are mostly mandatory, and their registry research processes are not static—changes are made to streamline the process for users and add new registries as they become available.
• A Not Cleared registry finding is a rare event. With over a million individual registry checks conducted in our three States, there were only about 700 with a Not Cleared result recorded. While a Not Cleared finding is rare, it doesn’t mean that registry checks are not worthwhile. State 1 data show that registry checks can exclude a significant number of individuals from employment in LTC relative to the number with an ineligible determination.
• Of the most widely used registries, State NARs appear to be the most productive for finding registry matches.
• An efficient registry checking process involves trade-offs. More registries added to the process means more time spent. Automating the registry checking process can allow States to increase the number of registries checked without adding significantly to the provider’s level of effort.

VIII. Recommendations for States

We recommend the following for States:

• To increase the efficiency of registry checks, automate the process as much as possible. Using an exact SSN match will cut down on the number of potential matches returned and the time users spend reviewing potential matches.
• By default, all registries in the TA-developed BCS’s registry database that are associated with a State will appear on the registry research screen when the State is included as a prior address. While the registry database generally has just one or two links to registries per State, some States have several registries listed in the database. (West Virginia, for example, has 26 registries listed in the database, including a separate web link for each professional licensure board.) System administrators can select which out-of-state registries appear on the registry research screen when a prior address is entered for an applicant. Using this function can help simplify out-of-state registry checks for your users.
• States may want to consider adding automated registry recheck functionality to their BCS. The registry recheck process allows providers to upload a roster of their current employees and then as often as monthly conducts an automated recheck of selected registries for employees and for applicants whose applications are in process. The registry recheck process works in much the same way as criminal history rap back, identifying any individuals who may have been added to registries after hire.

End Notes

1. The two States not included in this analysis are Maryland, which has ended its grant program without implementing a background check system, and Delaware, whose grant ended in September 2013.
2. Missouri is the only State not using a federal registry as part of its registry research. Missouri is continuing to use its legacy background check system until legislation can be passed to implement an NBCP-compliant program.
3. It’s likely that State 1 has a low rate of Not Eligible determinations for several reasons, including a background check statute that allows for review of criminal histories only for the past seven years. State 1’s data also shows evidence that the percentage of applicants with a criminal history is declining over time, suggesting that people who might potentially be ineligible for employment due to their past history are refraining from applying for employment in LTC positions.
4. Median results are presented because outliers in the data skew the average numbers.

Appendix: Registries Used in the Background Check Process by State