CMS-R-305 Validating Encounter Data

OMB Approval No.
0938-0786

VALIDATING ENCOUNTER DATA

A protocol for use in Conducting Medicaid External Quality
Review Activities

Department of Health and Human Services
Centers for Medicare & Medicaid Services

Final Protocol
Version 1.0
May 1, 2002
According to the Paperwork Reduction Act of 1995, no persons are required to respond to a collection of information unless it displays a valid
OMB control number. The valid OMB control number for this information collection is 0938-0786. The time required to complete this
information collection is estimated to average 1,591 hours per response for all activities, including the time to review instructions, search existing
data resources, gather the data needed, and complete and review the information collection. If you have comments concerning the accuracy of the
time estimate(s) or suggestions for improving this form, please write to: CMS, 7500 Security Boulevard, Attn: PRA Reports Clearance Officer,
Baltimore, Maryland 21244-1850.
Form CMS-R-305

VALIDATING ENCOUNTER DATA
I. PURPOSE OF THE PROTOCOL
Encounter data (i.e., data on the distinct health care services provided to each Medicaid managed
care enrollee) can be a useful source of information for States, as well as managed care
organizations (MCOs) and Prepaid Health Plans (PIHPs). Encounter data can be used to assess
and improve quality, as well as monitor program integrity and determine capitation payment rates.
However, in order for encounter data to effectively serve these purposes, it must be valid; i.e.,
complete and accurate. At present, completeness and accuracy of encounter data vary across
States, MCOs, and PIHPs. This protocol specifies processes for assessing the completeness and
accuracy of encounter data submitted by MCOs and PIHPs to the State. It also can assist in the
improvement of the processes associated with the collection and submission of encounter data to
State Medicaid agencies.

II. ORIGIN OF THE PROTOCOL
This protocol was developed from documents in both the public and private sectors, as well as
interviews with personnel from three State Medicaid agencies (Alabama, Arizona and Oregon)
experienced in the collection of encounter data. The documents reviewed included: 1) A Guide
for States to Assist in the Collection and Analysis of Managed Care Data - second edition 1
(draft); 2) The MEDSTAT Group (MEDSTAT)’s Final Design Report for Verification of
Encounter Data (part of the evaluation of the Medicare Choices Demonstration); and 3) the
National Committee for Quality Assurance (NCQA)’s 1999 HEDIS7 publication: Volume 5,
HEDIS Compliance AuditTM Standards and Guidelines.
Beginning in 1995, the Centers for Medicare and Medicaid Services (CMS) (formerly the Health
Care Financing Administration (CMS)) and MEDSTAT began developing a series of tools to help
State Medicaid agencies collect, validate and use encounter data for managed care program
management and oversight. The tools and approaches developed for this contract were further
refined and narrowed as part of the pseudo-claims (encounter data) validation project CMS
commissioned for the Medicare Choices Demonstration. For that project, CMS specifically
requested the development and application of a statistically reliable encounter data validation
process using medical records as the reference information. MEDSTAT has used similar
approaches for validating Medicaid managed care encounter data in a number of States, although
these approaches have varied depending on the sophistication of the MCO/PIHP and State
information systems, the amount of encounter data collected, and each State=s approach to
improving the quality of encounter data. However, in all of these States, MEDSTAT has used an
approach composed of three core activities:
(1)
(2)

1

Assessment of the MCO or PIHP’s information system (IS)
Analysis of MCO or PIHP electronic encounter data for accuracy and
completeness, including analysis of data reasonableness

The final version of this document is available online at: (www.hcfa.gov/medicaid/enguide2.htm)

Page 1

(3)

Review of medical records for additional confirmation of findings.

As part of the validation process, MEDSTAT has developed data collection instruments designed
to evaluate, troubleshoot and facilitate improvement of encounter data and the information
systems from which encounter data are produced. These instruments are intended to be completed
by the MCO/PIHP. The information obtained through these instruments is to be confirmed
through face-to-face interviews of MCO/PIHP staff by staff conducting the assessment of the
MCO/PIHP IS. The tools also are designed to be sensitive to the burden placed on an MCO/PIHP
by information collection, while maintaining the integrity of the information collected.
The NCQA HEDIS Compliance Audit tool was designed for auditing encounter data when
encounter data are used to calculate certain performance measures. It also includes an IS
assessment analogous to the first of the three elements of the MEDSTAT validation process. The
NCQA tool is designed to gather information about MCO/PIHP’s IS capabilities across all
payors, rather than concentrating specifically on Medicaid or Medicare. The MEDSTAT tools
were developed for Medicaid and Medicare only, although MEDSTAT also has companion
approaches designed for commercial payors.
All the validation processes reviewed address the collection of information about the
MCO/PIHP’s IS capabilities as a first step. All also include medical record review as a
component of the validation method. Each document provides a method to calculate a
statistically valid sample size for the medical record review. Interviews with State Medicaid
agency personnel found that their protocols also include medical record review.
The NCQA and MEDSTAT tools also are similar in that they: 1) contain pre-onsite visit
questionnaires to be completed by MCO/PIHP staff, and site visit interview forms to be
completed by staff performing the IS assessment; 2) require MCOs/PIHPs to provide information
on the level of specificity of the diagnosis and procedure coding systems used; 3) devote
considerable attention to medical record review; and 4) explore the issue of provider contracts
and physician compensation. Provider contracts and compensation are significant elements in
understanding the flow of encounter data from the providers to the MCOs/PIHPs, which
influences the timeliness and completeness of encounter data.

III. PROTOCOL OVERVIEW
This protocol is based almost entirely on the guide for States developed by MEDSTAT for
validation of encounter data. The elements contained in MEDSTAT’s document are consistent
with the other documents reviewed.
This protocol also makes the following assumptions:
1.

For the purposes of this protocol, an encounter refers to the electronic record of a service
provided to an MCO/PIHP enrollee by both institutional and practitioner providers
(regardless of how the provider was paid) when the service would traditionally be a
billable service under fee-for-service (FFS) reimbursement systems. Encounter data
provides substantially the same type of information that is found on a claim form (e.g.,
UB-92 or CMS 1500), but not necessarily in the same format.
Page 2

2.

The State will further specify an operational definition of an “encounter” and the types of
encounters (e.g., physician, hospital, dental, vision, laboratory etc) for which encounter
data are to be provided. The State will also specify the information (data fields) to be
submitted for each type of encounter.

3.

Encounter data can be considered “complete” when they can be used to describe the
majority of services that have been provided to Medicaid beneficiaries who are enrollees
of a MCO/PIHP. 2

4.

Development of accurate and complete encounter data is an iterative process. Because
encounter data are an outgrowth of MCO/PIHP IS and data policies, it is often not
possible for MCOs and PIHPs to overcome all limitations in their IS and data policies in
one year. As a result, in the first year that a State requires the submission of encounter
data from its MCOs and PIHPs, the data may be significantly incomplete and contain
errors. Improving the completeness and accuracy will take place through continuous
quality improvement (CQI) processes implemented year after year. Because of this,
States will need to develop a “phased-in” approach for using standards for encounter data
accuracy and completeness. “Phased-in” standards acknowledge the start-up issues
affecting both MCO/PIHPs and State Medicaid information systems receiving the
encounter data.

5.

The State will establish standards for encounter data accuracy and completeness.

6.

States will specify objective standards to which encounter data submitted by their
Medicaid MCOs/PIHPs will be compared. These standards can be national, regional or
State standards, as discussed in ACTIVITY 3 of this protocol.

The protocol consists of five sequential activities:
(1)
(2)
(3)
(4)
(5)

Review of State requirements for collection and submission of encounter data
Review of each MCO/PIHP’s capability to produce accurate and complete
encounter data
Analysis of MCO/PIHP electronic encounter data for accuracy and completeness;
Review of medical records, as appropriate, for additional confirmation of findings
Submission of findings.

IV. PROTOCOL ACTIVITIES
2

A State may decide to use other sources of information, such as an immunization registry, to substantiate
or complete the information on the provision of services to Medicaid beneficiaries when the encounter data format
does not easily or accurately capture the required information.

Page 3

ACTIVITY 1:

Review State requirements for encounter data collection and
submission.

Prior to performing encounter data validation, the External Quality Review Organization
(EQRO) 3 needs to be familiar with the State’s requirements for the collection, processing and
submission of encounter data by MCOs and PIHPs to the State. Some State requirements may be
unique to a particular State program. States need to provide the EQRO with: 1) the State=s
requirements for collection and submission of encounter data by MCOs/PIHPs. (These typically
are found as specifications in the contracts between the State and the MCO/PIHP.) 2) the data
submission format specified by the State for MCO/PIHP use, 3) the State”s data dictionary, 4) an
explanation of the information flow from the MCO/PIHP to the State, 5) State standards for
encounter data completeness and accuracy, 6) the time frames for data submission, 7) any
historical problems experienced in this process; and 8) any other information relevant to
encounter data validation.
The EQRO should also obtain from the State a listing of the types of encounter data to be
validated. For each type of encounter data (e.g., office visit, inpatient, laboratory, et al.) to be
validated, the State should specify the rates of missing, surplus, or erroneous encounters (as
defined below) that it will find acceptable. The Acceptable Error Rates Specification Form
below (or a similar form) can be used to summarize these specifications for each of the different
types of encounter data.
The State also should specify acceptable rates of accuracy and completeness for each data field
submitted for each encounter type. Attachment 1 contains MEDSTAT’s recommendations for
eventual accuracy and completeness standards for typical data fields. The EQRO will need to
tailor this chart or generate a form or forms similar to Attachment 1 to identify accuracy and
completeness standards specified by the State for all data fields the State requires for the
different types of encounters. The standards should be more lenient in the early years of
collecting encounter data, and more stringent as MCO/PIHP IS capabilities improve over time.
Acceptable Error Rates Specification Form
Type of Encounter
Office Visit (excludes dental and mental
health / substance abuse visits)

Office Visit - mental health / substance

Error Type

Acceptable Error Rate

Missing

<

%

Surplus

<

%

Erroneous

<

%

Missing

<

%

3

It is recognized that a State may choose an organization other than an EQRO as defined in Federal
regulation to perform encounter data validation. However, for convenience, in this protocol we use the term
“external quality review organization” (EQRO) to refer to any organization conducting validation of encounter data.

Page 4

abuse

Office Visit - dental

Inpatient admission - (excludes mental
health / substance abuse visits)

Inpatient admission - mental health /
substance abuse

Other types of encounters as specified
by the State; e.g., laboratory, pharmacy,
physical therapy.

Surplus

<

%

Erroneous

<

%

Missing

<

%

Surplus

<

%

Erroneous

<

%

Missing

<

%

Surplus

<

%

Erroneous

<

%

Missing

<

%

Surplus

<

%

Erroneous

<

%

Missing

<

%

Surplus

<

%

Erroneous

<

%

The EQRO should add as many
additional rows to this chart as needed
to incorporate all types of encounters
specified by the State.

Definitions:
Missing - encounters that occurred but are not represented by an electronic record.
Surplus - encounters which are represented by an electronic record, but either did not occur or
duplicated other records.
Erroneous - encounters that occurred and are represented by an electronic record, but contain
incorrect data elements.
Acceptable Error Rate - the maximum percentage of missing, surplus, or erroneous records that
the State is willing to consider acceptable.

ACTIVITY 2:

Review each MCO/PIHP’s capability to produce accurate and
complete encounter data.

It is not feasible to review all encounters that beneficiaries have with MCO/PIHP providers to
assess whether they are completely and accurately recorded. Therefore, efficiently assessing
Page 5

encounter data completeness and accuracy involves: 1) determining if the MCO/PIHP has
structured its information system in a way that is likely to capture complete and accurate
encounter data; and then, 2) examining more closely the data produced by the IS system to detect
patterns that can indicate its completeness and accuracy. Activity 2 addresses the first of these
two; the second is addressed in Activity 3.
Activity 2 attempts to answer the question: ΑTo what degree is an MCO/PIHP’s information
system likely to produce complete and accurate information on all encounters between Medicaid
enrollees and their providers (both institutions and practitioners). Reviewing the capability of
each MCO/PIHP to do so is accomplished through two activities:
1.

Reviewing a standardized assessment of each MCO/PIHP’s IS capabilities; and

2.

Interviewing personnel at each MCO/PIHP to augment information obtained
through the standardized assessment.

Step 1: Review or conduct a standardized assessment of each MCO/PIHP’s IS capabilities.
A standardized assessment (i.e., an assessment that does not vary by the individual performing
the review or by the questions being asked) is necessary to promote reliable assessments of
information systems. A standardized tool and approach for conducting such a review is found in
Appendix Z (Information System Capabilities Assessment (ISCA) for Managed Care
Organizations and Prepaid Health Plans).
An MCO/PIHP may already have undergone such an assessment of its IS. For example,
assessment of IS is conducted when validating performance measures and performing
accreditation reviews. The EQRO needs to determine if the MCO/PIHP whose encounter data
are being validated has already undergone such a review, and if so, if the review findings are
current. If a recent IS assessment has been conducted, the EQRO should receive a copy of the
findings, review the results of the prior assessment, and seek more recent information where
necessary. If the MCO/PIHP has not recently undergone an assessment, one will need to be
conducted as part of encounter data validation consistent with the process described in Appendix
Z.
Whether the EQRO reviews the results of an earlier IS assessment or conducts its own
assessment, the content included in Appendix Z should be addressed. This content and the reasons
for its significance include the following:
-

General Information
1.
2.

Managed Care Model Type: Encounter data completeness and accuracy are likely
to be better in staff model than non-staff model MCOs/PIHPs.
Year of incorporation: Encounter data accuracy and completeness is likely to be
better in more mature MCOs/PIHPs.

Page 6

3.

-

Information Systems Capabilities
1.

-

2.
3.

4.
5.

2.

Processing issues: Points in the process where errors are particularly likely to
occur should be identified.
Edit checks: MCOs/PIHPs should have an established, standard set of edits which
verify field content and consistency.

Enrollment Issues
1.
2.

-

Forms used: If the MCO/PIHP is not using standard claims forms (e.g., UB 92 or
CMS 1500) or encounter forms which are similar to the CMS 1500 or the UB 92,
the forms used should be reviewed to ensure that they capture key data elements.
Submission methods: Processing paper forms is more prone to error than direct
electronic data submission.
Required data fields (data elements): Standard measures of plan performance
typically require the availability of data on: patient date of birth/age, sex, place of
service, diagnoses, procedures, dates of service, revenue codes, and provider
specialty.
Number of diagnosis and procedure codes retained in data fields: Data fields
should allow a minimum of two diagnoses and two procedure codes to be retained.
Coding schemes: Knowledge of the coding schemes used by the MCO/PIHP is
necessary to verify the accuracy of their use.

Claims/Encounter Processing
1.

-

System descriptions: This provides an indication of the MCO/PIHP’s overall level
of data management sophistication.

Data Acquisition Capabilities
1.

-

Member enrollment: A larger enrollment may indicate that the MCO/PIHP has
more experience working with encounter data, but may also offer more
opportunity for errors.

Type and frequency of updates: Infrequent or inaccurate updates will result in
invalid encounter data.
Use of unique identifiers: Without reliable identification of enrollees and
providers, encounter data validity is not possible.

Vendor/Contractor Data
1.
2.

Data submission policies: Consistently applied policies that leave no room for
variations in interpretation increase the accuracy of the data submitted.
Contract requirements: Data are more likely to be complete when they are
required as a condition of payment than when they are optional.

Page 7

-

Provider Contracting Arrangements
1.

2.

Compensation arrangements: Salaried providers will submit data on a timely
basis if data submission is a parameter in their contract with the MCO/PIHP. Feefor-service (FFS) providers have the greatest incentive to submit accurate and
complete data, since their payment depends on it.
Contract requirements: Data are more likely to be complete when they are
required for payment.

After reviewing each MCO/PIHP’s IS Capabilities Assessment, the EQRO staff will record their
analytic findings on a standard form such as the Information Systems Capabilities Assessment for
Managed Care Organizations and Prepaid Health Plans - Reviewer Worksheet and Interview
Guide (Reviewer Worksheet and Interview Guide) found in Appendix Z. A form such as the
Reviewer Worksheet and Interview Guide serves to document the findings of the EQRO staff
when reviewing the IS Capabilities Assessment for each MCO, and to identify those issues to be
addressed in Step 2, the follow-up interview with MCO/PIHP personnel.
Step 2: Interview personnel at each MCO/PIHP to augment information obtained through
the standardized IS Capabilities Assessment.
Whether the EQRO is reviewing a previous assessment of MCO/PIHP IS capabilities, or whether
it has asked the MCO/PIHP to complete a new Information Systems Capabilities Assessment such
as that in Appendix Z, the written descriptions of IS capabilities submitted by the MCO/PIHP
must be reviewed and supplemented by conversations with the MCO/PIHP staff to clarify or
gather more detailed information regarding the MCO/PIHP’s IS capabilities. The EQRO staff will
interview appropriate MCO/PIHP staff using a standard interview protocol such as the
MCO/PIHP IS Capabilities Assessment - Reviewer Worksheet and Interview Guide found in
Appendix Z. However, all information submitted by the MCO/PIHP on its Information Systems
Capabilities Assessment might not need further clarification. In addition, not all questions in the
Interview Guide may need to be discussed with MCO/PIHP staff with respect to encounter data
validation. This is because assessment of information systems can be conducted for different
purposes; e.g., validation of performance measures and determining compliance with MCO/PIHP
structure and operational standards. Further, some questions in the Reviewer Worksheet and
Interview Guide will need to be reworded for newly formed MCOs/PIHPs. However, the
following areas, at a minimum, should be fully described either in the MCO/PIHP’s written
documentation of its IS or through subsequent follow-up discussions between the MCO/PIHP and
the EQRO.
Information Systems: Data Processing and Procedures
1.
2.
3.

Data Base Management System (DBMS) Type
Programming language
Updating the program to meet changes in State requirements.

Claims/Encounter Processing

Page 8

1.
2.
3.

Overview of the processing of encounter data submissions
Completeness of the data submitted
Policies/procedures for audits and edits.

Claims/Encounter System Demonstration
1.
2.
3.
4.

Processes for merges and/or transfer of data
Processes for encounter data handling, logging and processes for adjudication
Audits performed to assure the quality and accuracy of the information and the
timeliness of processing
Maintenance and updating of provider data.

Enrollment Data
1.
2.
3.

Verification of claims/encounter data
Frequency of information updates.
Management of enrollment/disenrollment information

Based on this review of the MCO’s/PIHP’s IS capabilities, the EQRO should note for each
encounter type listed in the Acceptable Error Rates Specification Form (described previously)
whether or not there are concerns about certain types of encounter data, and note it in the fourth
column in the chart below. This should trigger further investigation.

Page 9

Acceptable Error Rates Specifications and Identified Areas of Concern Form
Encounter Type

Error Type

Acceptable Error Rate

Missing

<

%

Surplus

<

%

Erroneous

<

%

Missing

<

%

Surplus

<

%

Erroneous

<

%

Missing

<

%

Surplus

<

%

Erroneous

<

%

Missing

<

%

Surplus

<

%

Erroneous

<

%

Missing

<

%

Surplus

<

%

Erroneous

<

%

Other types of encounters as
specified by the State; e.g.,
laboratory, pharmacy, physical
therapy.

Missing

<

%

Surplus

<

%

Erroneous

<

%

The EQRO should add as many
additional rows to this chart as
needed to incorporate all types
of encounters specified by the
State.

Missing

<

%

Surplus

<

%

Erroneous

<

%

Office Visit - (excludes dental
and mental health / substance
abuse visits)

Office Visit - mental health /
substance abuse

Office Visit - dental

Inpatient admission - (excludes
mental health / substance abuse
visits)

Inpatient admission - mental
health / substance abuse

ACTIVITY 3:

Area of Concern
(Yes / No)

Analyze electronic encounter data for completeness and accuracy.

This activity represents the core of the process that the EQRO will use to test the validity
(completeness and accuracy) of the encounter data. Once the steps in this activity have been
completed, the EQRO and the State will have an excellent assessment of whether the data can be

Page 10

used for analysis. If the EQRO is unsure of the quality of the encounter data at the completion of
Activity 3, then it should not proceed to the medical record review activity (Activity 4). Rather,
it should either review the steps of this activity and identify areas where information did not
satisfy the EQRO or it should seek additional assistance to determine why there is uncertainty
about the quality of the encounter data. If the steps in Activity 3 are completed thoroughly and
accurately, there should be little doubt about the quality and usefulness of the submitted
encounter data.
In this activity, the EQRO undertakes analysis of each MCO/PIHP’s encounter data through four
steps. Information obtained from these four steps and the previously conducted IS Capabilities
Assessment and the Structured Interview, should yield four classes of information available for
each MCO/PIHP:
1. General magnitude of missing encounter data. Evidence of whether the MCO/PIHP
has been unable to submit any encounter data and reasons for failures, such as the inability
to process the encounter data without edits.
2. Types of potentially missing encounter data. MCOs/PIHPs which have sub-capitated
or sub-contractor relationships with providers often experience difficulty in receiving
information from those providers. Knowledge of the MCO/PIHP’s contractual
relationships with providers will help identify specific areas to investigate for missing
services.
3. Overall data quality issues. Identification of data quality problems such as inability to
process or retain certain fields on the encounter data record. Some MCOs/PIHPs may not
currently have room in their systems to maintain all the information which is expected to
be submitted.
4. MCO/PIHP data issues. Problems with how the files are compiled and submitted to
the States.
Step 1:

Analyze information from the IS Capabilities Assessment and the followup structured interview with MCO/PIHP staff.

At the completion of Activity 2, the EQRO will have an excellent description of the MCO/PIHP’s
IS and should know what to expect when the MCO/PIHP’s data files are investigated. The
information from Activity 2 is incorporated into a plan for testing the quality of the data. This plan
specifies the areas that will be tested in the data (the areas of investigation) and the expected
results. Having such a plan ensures that parts of the data quality review are not overlooked. For
example, it is expected that MCOs that pay a substantial portion of their primary care providers on
a capitated basis will have a lower encounter data submission rate than MCOs that pay their
primary care providers predominantly on a FFS basis. As a part of a data quality test plan, provider
groups would need to be identified by their type of payment and rates of outpatient visits per
eligible beneficiary calculated. The rates of outpatient visits would then be compared to test the
assumption that capitated providers have a lower rate of encounter submission than FFS providers.
If this assumption is not supported by the data, and the outpatient visit rates differ from the

Page 11

benchmark, then other questions need to be answered about the MCO/PIHP’s processing of FFS
claims. This is an example of how the information from Activity 2 is incorporated into Activity 3,
Step 1. Other questions and issues can be addressed in the same way.
Using information provided by the State in Activity 1, and information obtained from the
MCO/PIHP through the IS Capabilities Assessment and the Structured Interview, the EQRO staff
will develop a data quality test plan that will:
-

adjust their own error detection programming specifications to reflect those used by the
MCO/PIHP
adjust their own report specifications to match those used by the MCO/PIHP
create ad hoc data investigation specifications, based on the information gleaned from the IS
Capabilities Assessment and the interview
create ad hoc report writing specifications
compile notes to assist in data interpretation.

Step 2:

Inspect the MCO/PIHP’s Encounter Data files

Step 2 and Step 3 of this Activity are closely inter-twined. To make the steps clearer, they have
been broken into two parts because having two steps more accurately reflects the way a standard
data quality review process would occur. When data are reviewed for accuracy and completeness,
they are subjected to a macro and micro analysis. These steps are described separately to prevent
the EQRO from rushing forward to generate a large number of reports and analyses before the
basic integrity of the data have been verified. Step 2 represents the macro analysis. Step 3
represents the micro.
Step 2 describes a basic integrity check of the data files. It answers the questions: Are there data?
Do they generally fit with expectations? Are they of sufficient basic quality to proceed with more
complex analyses? In general, all of the analysis that is required in Step 2 should be highly
automated and generated as a standard data review process. The analysis required in Step 2 can be
separately performed on each of the different encounter data files (e.g., hospital, dental,
ambulatory, etc.) for each of the data fields in those files, while the analysis described in Step 3
requires that encounter, eligibility, and provider data be linked together. The EQRO will obtain the
encounter data to analyze either by accessing the State’s information system or by receiving from
the State an encounter data extract from the States= data system that replicates the data that the
State has. In the latter case, the EQRO will access the data using its own analytical processes. The
EQRO will inspect the files and perform the following activities:
1.

Assure that the enrollment information that the State transfers to the MCO is
accurately incorporated into the MCO information system and is being reported
back to the State correctly. In many cases, MCO information systems do not use
the State’s Medicaid identifier as the means for tracking enrollees. When the
encounter data are reported to the State, the Medicaid ID must be reattached to the
data file. In this step, the EQRO will verify that the Medicaid IDs are being
reported correctly. As an additional, but not required, step, the EQRO could
compare the encounter data file to a State eligibility file and check for accuracy of

Page 12

the IDs, and other eligibility information; (e.g., age, sex and eligibility category).
This step is optional since in the majority of situations, the encounter data file that
the EQRO is reviewing has been edited in the State system - where eligibility
checking is one of the core elements. This step should be considered in those
situations where the encounter data have not been edited in the State system. 4 In
addition, the EQRO will determine whether there are encounter data for the
majority of beneficiaries, rather than a large volume of data but only a few IDs.
The primary focus of this step is the verification of the correct eligibility numbers.
In other parts of this activity, analysis will be done to ensure that the scope and
volume of services are consistent with the eligibles.
2.

Apply general edit and consistency checks, such as verifying that critical fields
contain non-missing values and that values are consistent across fields; e.g.,
pregnancy and related diagnoses and procedures are for individuals whose sex is
coded as female.

3.

Inspect the data fields for general validity (i.e., information for each critical field is
within required ranges, and the volume of data is consistent with the MCO/PIHP’s
enrollment).

4.

Capture more detailed validity information from the encounter data fields used for
reporting purposes.

The analysis includes a review of each data element and a general review of the volume of data
by type or place of service. This review concentrates on two areas: field validation and
completeness:
1. Field validation
-

Percent present: required data fields are present on the file and have information in that field
Percent valid: data in the field are of the requested type; i.e., numeric fields have numbers,
character fields have characters, etc.
Percent valid values: In those fields, the values are the expected values; e.g., are there valid
ICD-9 codes in the diagnosis field, not just random numbers? This review requires comparing
the specific field to sources of information showing valid values. 5
Field validation will require the use of State standards. Examples of eventual State
standards are found in Attachment 1.
4

In a few States, Medicaid programs have chosen to by-pass the MMIS and have MCOs submit data directly to
an outside source. In these cases the eligibility checking may not be as intense as that done within the MMIS. In those
cases the EQRO might choose to add this additional eligibility review.
5

This is a place where the information from the IS Assessment becomes critical. One of the things that will
be asked is which version of the ICD-9 codes is used by the MCO. Often the State and the MCO maintain different
versions of these files resulting in values being considered invalid when they really are valid.

Page 13

2. Data completeness
-

Distribution by service type: determine whether there are data distributed as expected in the
large data types: institutional, provider, pharmacy, dental, etc.
Across time: determine the data volumes by month. Are all data types seen in all months? Is
the data volume consistent across the months?

Step 3:

Generate and Review Analytic Reports

Using simple statistical procedures such as measures of central tendency, univariate descriptive
statistics and bivariate distributions, the EQRO will analyze the data to obtain a “data validity”
overview of each MCO/PIHP’s encounter data. This process will analyze and interpret data on: 1)
submitted fields, 2) volume/consistency of encounter data, and 3) utilization rates.
Analyzing and Interpreting Data in Submitted Fields
There are three questions to be addressed in a field-specific review:
1.

Is there information in the field, and is that information of the type
requested? Each field will have a definition which will include data type (alpha,
numeric, mixed) and size. The fields must be checked to determine whether the
information is of the correct type and size. For example, if ICD-9 diagnosis codes
have been requested, the field should have 5 digits. If CPT-4 codes are requested,
the field should have 4 digits. If the State’s Medicaid beneficiary ID is requested,
the field should contain the correct number of letters and digits.

2.

Are the values valid? When compared to an external standard, are the values in
the field valid for that standard? For instance, if ICD-9 diagnosis codes have been
requested, are the values in the diagnosis field valid ICD-9 diagnosis codes? A
field could have 5 digits in it but those digits might not represent valid codes.

Page 14

Findings for questions 1 and 2 could be recorded on a standard form such as that below:
Sample Form for Recording Evaluation of Submitted Fields
Required Field

Information
present
#

%

Correct type
of information
#

%

Correct size of
information
#

%

Presence of
valid value ?
#

%

Enrollee ID
Plan ID
Provider ID
Principal Diagnosis
Procedure Code
Date of Service
Units of Service
Others (continue
adding fields as
appropriate). . . . .

3.

Are the values reasonable? A frequency distribution of the values needs to be
developed and then compared to an external standard to determine whether the
values make sense for the submitted population. For instance, if one of the required
fields is “place-of-service,” there should be a reasonable distribution between
inpatient hospital, outpatient hospital, emergency room and physician office. In this
data review, the values in the fields could have passed the test in steps 1 and 2, but
failed at this stage. A plan could submit data with a single valid value in a field;
these data would pass steps 1 and 2 but fail step 3.

Analyzing and Interpreting Volume / Consistency of Encounter Data
This type of evaluation provides basic statistics on the encounter data. It describes, among other
things, the number of Medicaid enrollees, the number of encounters, and counts and totals for
various demographic subgroups, diagnoses and types of services. The EQRO should run
frequency distributions on specific fields as well as on the variables created explicitly for data
validation reporting purposes. The EQRO may also run distributions on subsets of variables and
observations where the result indicates potential data validity concerns. For instance, a subset of
rates of outpatient services by provider zip code might highlight missing zip codes in the edit file
which results in the rejection of all encounters with that zip code. This initially looks like an

Page 15

overall low rate of services but by looking at a subset of a specific field and checking for
reasonableness, a different problem is detected. The EQRO also should generate univariate
statistics (e.g., means, medians and modes) as appropriate on continuous and discrete data fields.
The output produced for these reports should be checked for both reasonableness and to detect
specific problems, such as entire categories of data missing from the regular data submissions.
The EQRO should also analyze encounter data for other volume/consistency dimensions. These
dimensions can include time, provider type, type of service, and demographic groupings, but
States may have additional dimensions (aid category) which also need to be included. This
information allows the EQRO to look for trends such as the following:
-

Time - This analysis would examine encounter data both by service date and by processing
date to check consistency. MCOs/PIHPs often have problems processing encounter data and
in many cases these claims are processed sporadically. When such a situation is present, it can
often be an indication of other problems within the MCO/ PIHP’s information system. After
establishing the length of time between service dates and processing dates by MCO/PIHP, the
EQRO could compare these with existing benchmarks for data submission and processing.

-

Provider - Encounter data validation can verify the presence of encounter data for all
provider types and determine if there are significant fluctuations in patient visits per time
period. In addition, information collected during the capabilities assessment will be used to
identify missing encounter data for specific provider types. The distribution of encounter data
will be compared by provider type with the benchmark information described above.

-

Service Type - The EQRO should verify whether ancillary services (e.g., labs, x-rays,
therapy, etc.) are evenly represented as visits. The clinical connection between the use of
services is being evaluated. If a Medicaid beneficiary is receiving x-ray or lab tests, one
would expect to see an office visit (or perhaps a hospital admission) in the same time frame.
Other areas where reasonableness between the encounter data should be tested include: 1)
relationship of outpatient visits to number of prescriptions, 2) relationship of primary to
specialty care visits, and 3) outpatient services associated with inpatient admissions.
Examination of the data in this way will reveal whether there are missing encounters.

-

Age- and Sex-Appropriate Diagnoses and Services - The EQRO will determine if the
diagnoses and services reflect expected care by age and sex. As an example, one would expect
sex-specific diagnoses (such as endometriosis or undescended testes) and procedures (such as
deliveries or hysterectomies) would have the patient’s sex coded correctly. Conversely one
would expect that men and women would not have procedures coded that cannot be
performed on a person of that sex.

As part of the review, the EQRO will find it helpful to display the data quality findings
graphically. It is nearly always true that in these situations, “a picture is worth a thousand words.”
These graphs will be useful internally for identifying issues, and externally for conveying the
results of the data quality review. An example of the type of chart that might be generated is a
frequency distribution of the number of encounters per Medicaid enrollee. This chart shows how
many enrollees had zero, one, two, three, etc. encounters. This chart can also be replicated by

Page 16

different age, sex, and race groupings, to see whether there are differences along these
dimensions. These charts should be generated for all MCOs/PIHPs in the aggregate, and for each
MCO separately so that issues with a specific plan can be identified.
Analyzing and Interpreting Utilization Data
The EQRO also should routinely compile and review, on a periodic (e.g., monthly, bi-monthly,
quarterly or other periodicity as directed by the State) basis, statistics displaying information on
utilization rates overall and by specific diagnosis, procedure, service and provider types when
appropriate. These reports initially should be generated both for each MCO/PIHP and on the
entire encounter data set for all MCOs/PIHPs together to account for problems associated with
small numbers of encounters for certain MCOs/PIHPs. During the program start-up phase, many
MCOs/PIHPs may have very few encounters for some diagnoses and services, which would make
their rates statistically imprecise.
One method for estimating the completeness of the encounter data for each MCO/PIHP is to
benchmark the utilization rates for a given MCO/PIHP against utilization statistics from other
sources. These benchmarks would be incorporated into the MCO/PIHP-specific utilization
reports. Utilization rates could be broken-out by patient demographics, diagnosis, type of service
and type of provider.
Step 4: Compare findings to State-identified standards
The EQRO will next compare the encounter data submitted by each MCO/PIHP to standards and
benchmarks that are identified by the State. These standards can be obtained from a number of
different sources, including: aggregate encounter data from all Medicaid MCOs/PIHPs in that
State or other States which are considered to be comparable, historical FFS Medicaid data in that
State, or data from a State’s Primary Care Case Management (PCCM) program. The State may
also look to: commercial managed care plans, national standards, or other benchmarks. The State
will need to identify such standards and document them using a form such as that found in
Attachment 2: Table of Benchmark Utilization Rates. States will also need to specify acceptable
variation from these standards. Both the standards and acceptable variations from the standards
can be made more stringent over time.
For instance, when comparing encounter data utilization rates to a Medicaid FFS benchmark, one
might expect to see a drop in emergency room utilization per member month under managed care.
However, large, downward swings in other types of utilization (e.g., >30 percent drops in
ambulatory care) may indicate incomplete encounter data rather than a change in provider
practice patterns. The EQRO should test their assumptions about changes in utilization under
Medicaid managed care with the MCOs/PIHPs and with the State, as a further test of the
completeness of the encounter data.
The results of ACTIVITY 3 are used to form the basis for a long-term monitoring strategy for
assessing the quality of the encounter data. As the data improve over time due to monitoring and
Page 17

validation, the EQRO will be able to design targeted validation strategies by using analytic testing
on the encounter data files to identify problem areas requiring medical record review validation,
thus allowing the conservation of resources by avoiding unfocused Αbroadside≅ medical record
review.

ACTIVITY 4:

Review of medical records for confirmation of findings of analysis of
encounter data

Medical record review can provide additional verification of the information obtained from the
preceding analysis of electronic encounter data for accuracy and completeness (Activity 3).
However, medical record validation is a complex and resource-intensive process. While it can
easily be used to validate specific areas of concern, it is not an efficient method for performing a
more general validation of encounter data. As stated at the beginning of Activity 3, if the EQRO
is unsure of the quality of the encounter data at the completion of Activity 3, then it should not
proceed to medical record review. Rather, it should either review the steps in Activity 3 and
identify areas where information did not satisfy the EQRO or it should seek additional assistance
to determine why there is uncertainty about the quality of the encounter data. If the State is in the
initial phase of collecting encounter data from its MCOs/PIHPs, there may be a time lag of as
much as three years before the encounter data are of sufficient completeness and accuracy to
warrant the investment in medical record review
Further, medical record review should not be used to validate information that is collected at
another source and considered to be more accurate. For example, information collected as part of
eligibility determination is considered the primary source for demographic information such as
patient age, sex, and race. This eligibility data should be used as the source of demographic
information when validating encounter files.
In this protocol, the following assumptions are made with respect to medical record review:
-

Medical record review for encounter data validation is being performed independently of
medical record review for evaluation of performance measures or other purposes, however
these reviews could be coordinated in an effort to reduce the burden on the MCOs/PIHPs.

-

When medical record review should begin, and how frequently it should be performed, will be
decided by the State.

-

One medical record review is determined by the State to be appropriate, the EQRO will draw
a sample of medical records for validation on a regular and periodic basis specified by the
State.

Validating encounter data using medical records must be approached as if it were a research
question. There needs to be clear hypotheses, well defined populations, and stated error
tolerances. A rigorous research design ensures that the results of this resource intensive effort will
be meaningful and useful.
The approach to medical record validation depends on the questions/hypotheses to be addressed.
Depending on the stated hypotheses, one would begin with a sample of encounters, enrollees, or
Page 18

both. The table below illustrates how two different questions are answered by using different
sampling universes.
Examples of questions to be
addressed
Is the information found on
the encounters accurate when
compared to the medical
record?
Are there electronic
encounters for all the services
that were provided to
enrollees?

sampling universe
Encounters

Enrollees

Use sample of encounters

Use sample of enrollees

Another possible question may be: “Are the encounters fully coded for all diagnoses?” Often
when providers submit encounters, they only provide the primary diagnosis code, since that is
usually sufficient for payment. Another question might be “Are the procedures fully coded?”
Providers might record on the encounter form only those procedures that historically have
affected reimbursement. In such cases, additional procedures performed can only be discovered
by reviewing the medical record. In this type of medical record validation, a sample of encounters
is selected, and the medical records for those encounters are reviewed.
When one wants to determine whether all encounters have been received for services that were
delivered, then a sample of enrollees is selected, and their medical records are reviewed against
encounters. In this case, the EQRO would sample MCO/PIHP enrollees rather than encounters for
a specific time period. Services recorded in the medical record would be matched with those
found on the encounters to assess the completeness and accuracy of encounter data. To reduce the
number of medical records that must be located for each enrollee, the EQRO should consider
limiting medical record review to a specific type of encounter, such as inpatient admissions or
physician office visits.
Where medical record review is performed is a complicated question and each solution has its
own strengths and weaknesses. In all cases, the medical record must be located, the reviewers
must be trained and experienced, the confidentiality of patient records must be maintained, and
costs must be minimized. Balancing these four requirements is difficult and results in varied
solutions that fit with the specifics of each particular situation. In many cases, reviewers decide to
look at medical records away from the site of health care delivery and request that providers copy
charts and send them to a central location. This approach places a large amount of the
responsibility for locating the charts on the plans and providers, but also increases the number of
charts that are located. Providers should be compensated for the direct costs of copying and
submitting the files.
Page 19

Step 1: Sampling for medical record review
When medical record review is to be undertaken, the EQRO may initially choose to conduct “trial”
reviews of the encounter data using medical records previously obtained for other purposes, such as
focused studies of clinical care topics. This trial review would provide an opportunity for the
EQRO to acquire experience in using medical records specific to individual MCO/PIHPs for
encounter data validation, as well as highlighting community practice patterns with regard to
developing encounters from medical records, without adding unnecessary administrative burden to
the MCOs/PIHPs.
Once medical record review is to be conducted on a more routine basis, the size of the sample
must be determined for each MCO/PIHP. The size of the sample of medical records that must be
drawn in order to make statistically valid inferences regarding the validity of the encounter data
depends on a number of factors, including:
-

The minimum error rate the State wants to be able to detect
The frequency with which the State wants to perform the review
The subsets of encounter data the State intends to validate.

Because of these and other factors; e.g. the size of an MCO/PIHP’s enrollment and previous
validation results, it may be statistically appropriate to determine different sample sizes for each
MCO/PIHP. However, it may be operationally more efficient (and also statistically valid) to
specify the same sample size for all MCOs/PIHPs. The EQRO will need to determine the sample
size for each MCO/PIHP either as directed by the State, or, if the State allows the sample size to be
determined by the EQRO, in consultation with a qualified statistician. Whether one sample size or
multiple sample sizes will be used will be determined by the State, in consultation with the EQRO.
Once the sample size is determined, the EQRO will select this number of enrollees from each
MCO/PIHP for medical record review in order to calculate fault rates for each MCO/PIHP. The
sampling must be performed using methodologically sound techniques that defend against
sampling bias. The fault rate is the ratio of missing and erroneous records to the total number of
encounters that took place during the time period being examined. A fault rate can be calculated
for each encounter type.
It is anticipated that fault rates initially will be at least 30 percent. However, each MCO/PIHP’s
targeted fault rate should be below 5 percent (f<0.05) for each time period examined, and if
possible, demonstrate a decrease over time. Consequently, for each time period examined, the
EQRO will test the hypothesis that there is no difference between the services recorded in the
medical record and those found in the encounter data (i.e., the “null” hypothesis). The ability of
this test to reject the null hypothesis (Ho) when it should be rejected will depend on the size of the
encounter sample, the size of the total population of true encounters, and the MCO/PIHP’s actual
fault rate. The EQRO should set sample sizes sufficient to estimate the fault rate for each type of
encounter within each MCO/PIHP, with equal precision for each time period to be studied.
Step 2: Review medical records and record findings on a standardized worksheet.

Page 20

Fields to be validated through medical record review should include a few socio-demographic
fields (e.g. date of birth, sex) that may be needed to identify the correct beneficiary. Typically,
when medical records are selected for review, the provider is given the patient’s name, age, and
sex, the provide’rs name, and the target dates of service. This information helps the provider find
the correct medical record. When the comparison is done between the encounter data and the
medical record, this information is included in the data elements that are reviewed to further
ensure that the correct record was actually reviewed. The demographic information on the
medical record is not considered the definitive source of demographic information. It is included
here to support the medical record review process. Other fields on the encounter data should be
assessed so that it is possible to create measures of access and quality of care. These data
elements include: date of service and the clinical codes (such as ICD-9-CM, HCPCS and CPT
codes) that define diagnoses, procedures, and other services.
For diagnoses, the medical record review staff should review codes based on the diagnoses stated
by the provider in the medical record, (not based on diagnoses indicated by their own judgement).
Because of this, medical record review staff should be experienced clinical coding validators.
Clinical coding validators should have substantial clinical background, including anatomy and
physiology, pathology, microbiology, pharmacology, and disease process. They use this clinical
understanding, combined with their knowledge of appropriate coding guidelines, to assign the
right codes to a record. When a medical record lacks sufficient documentation to select the most
specific code(s), clinical coding validators may consult with each other and other health care
professionals to answer clinical questions.
Sources of coding errors, described by the Institute of Medicine in a study of the reliability of
coded data in the National Hospital Discharge Survey included:
1)
2)
3)
4)
5)

Incorrect selection or sequencing of principal diagnosis codes
Incorrect selection of other individual codes
Coding diagnoses or procedures not documented in the medical record
Errors caused by mistakes in entering the data into a database
Failure to review the entire medical record.

Building on these ideas, the EQRO should develop an error categorization scheme to identify
areas of incompleteness and inaccuracy. Codes assigned by the clinical coding validators will
have three components, as separate data elements. Data can be reported for any single error
component, or by any combination of these codes. These error components are:
Level: Identifies whether an encounter is present in the database. The presence or absence
of an encounter determines the strategy followed by the reviewer to complete the review.
Type: Describes if codes or other data are correctly or incorrectly present or absent.
Source: Assigns the most likely reason for the type of error found. While this may be a
subjective determination, it will be helpful when giving feedback to MCOs/PIHPs so that
they can target their data quality improvement processes.

Page 21

Certain errors may be designated as Αcritical≅ for the purposes of an audit. These can be defined
differently at varying points in time, because issues that are critical in encounter data validation
change over time. For example, in the early rounds of validation, a State may wish to focus on
diagnosis and procedure codes, and not on physician specialty or place-of-service. These fields
are important, but are of little value if the MCOs/PIHPs are not able to produce accurate clinical
coding. Once accurate diagnosis and procedure coding is taking place, knowing accurate specialty
and place of service greatly increases the value of the data. Having “tiers of errors” (e.g., critical,
serious, moderate, etc.) allows the State to move ahead with using encounter data that are not
totally complete and accurate. This is an important goal. In future years, another tier of more
refined error types will be added to the critical error types. Assigning the label of Αcritical≅ to the
errors will be determined by the State.
The findings of each medical record review should be documented on a standardized form, such
as the Medical Record Review Findings Tool for Encounter Data Validation found as Attachment
3. It is also important to provide documentation guidelines to the staff performing the medical
record reviews. These guidelines should describe exactly how to document the findings of the
medical record review. The guidelines also should be closely linked to the reporting requirements
and the data elements chosen for validation, so they should be written after the error classification
scheme is set. Written policies for interpretation of the documentation guidelines should include
the following:
-

Directions for reviewing medical records
Instructions on what to do when faced with conflicting documentation
Instructions on what to do when no code can be readily assigned
Use of optional codes
Definitions of what constitutes errors, and how to document them
List and location of approved reference materials (coding manuals, medical textbooks, etc.)
Whom to consult for additional assistance.

ACTIVITY 5:

Submission of findings

After the performance of Activities 1- 4, the EQRO will create data tables that display summary
statistics for the information obtained from these activities for each MCO/PIHP. Summarizing the
information in tables makes it easier to evaluate, and highlights issues with respect to the
accuracy and completeness of encounter data. A narrative will accompany these tables which will
highlight MCO/PIHP-specific issues.
In addition, a State at its discretion, may direct its EQRO to undertake technical assistance to its
MCOs/PIHPs to improve the accuracy and completeness of encounter data.

END OF PROTOCOL
Page 22

Attachment 1
EXAMPLES OF RECOMMENDED DATA QUALITY STANDARDS FOR EVALUATION OF
SUBMITTED ENCOUNTER DATA FIELDS (Physician and Other Provider)
Data Element

Expectation

Validity Criteria

Enrollee ID

Should be valid ID as found in the State’s
eligibility file. Can use State’s ID unless
State also accepts SSN.

100% valid

Enrollee Name

Should be captured in such a way that
makes separating pieces of name easy.
There may be some confidentiality issues
that make this difficult to obtain. If
collectable, expect data to be present and of
good quality

85% present. Lengths should vary and
there should be at least some last names
>8 digits and some first names < 8
digits. This will validate that fields have
not been truncated. Also verify that a
high percentage have at least a middle
initial.

Enrollee Date of
Birth

Should not be missing and should be a valid
date.

< 2% missing or invalid

MCO/PIHP ID

Critical Data Element

100% valid

Provider ID

Should be an enrolled provider listed in
provider enrollment file.

95% valid

Attending Provider
ID

Should be an enrolled provider listed in
provider enrollment file (also accept the
MD license number if listed in provider
enrollment file).

> 85% match with provider file using
either provider ID or MD license
number

Provider Location

Minimal requirement is county code, with
zip code being strongly advised.

> 95% with valid county code
> 95% with valid zip code (if available)

Place of Service

Should be routinely coded, especially for
physicians

> 95% valid for physicians
> 80% valid across all providers

Specialty Code

Coded mostly on physician and other
practitioner, optional on other types of
providers

Expect > 80% non-missing and valid on
physician or other applicable provider
type claims (e.g. other practitioners)

Principal Diagnosis

Well coded except by ancillary type
providers

> 90% non-missing and valid codes
(using ICD-9-CM lookup tables) for
practitioner providers (not including
transportation, lab and other ancillary
providers)

Other Diagnosis

This is not expected to be coded on all
claims even with applicable provider types,
but should be coded with a fairly high
frequency.

90% valid when present

Page 23

Date of Service

Dates should be evenly distributed across
time

If looking at a full year of data, 5-7% of
the records should be distributed across
each month.

Unit of Service
(Quantity)

The number should be routinely coded

98% non-zero
< 70% should be one if CPT code in
range 99200-99215, 99241-99291

Procedure Code

This is a critical data element and should
always be coded.

99% present (not zero, blank, 8- or 9filled). 100% should be valid, Stateapproved codes. There should be a wide
range of procedures with the same
frequency as previously encountered.

Procedure Code
Modifier

This is important to pick up to separate out
surgical procedures/anesthesia/asst.
Surgeon. It is not applicable for all
procedure codes

> 20% non-missing. Expect a variety of
modifiers both numeric (CPT) and
Alpha (HCPCS). The more common
codes which should appear with at least
a minimal frequency are: 47
(anesthesia) and 80 (asst. surgeon).

EPSDT Indicator
(All States might not
have this.)

If this field is used, the beneficiary should
be < 21 years of age and the provider
should be certified to administer EPSDT
screens.

95% enrollees < 21;
85% providers certified as EPSDT

Patient Discharge
Status Code
(Hospital)

Should be valid codes for inpatient claims
with the most common code to be
“Discharged to Home.” For outpatient
claims it can be coded as “not applicable.”

For inpatient claims, expect >90%
“Discharged to Home.” Expect 1-5% in
all other values (except “not applicable”
or “unknown”).

If facility uses UB92 claim form, should
Revenue Code
always be present.
(Hospital)
Source: The MEDSTAT Group

100% valid.

Page 24

Attachment 2
TABLE OF BENCHMARK UTILIZATION RATES
(for services incurred between XX/200x and YY/200x)
Measure

Value from
MA FFS or
PCCM

Value from
Comparable
State or States

Other
Comparison
Value

Inpatient Discharges
Inpatient LOS
Overall
By high volume DRGs
By eligibility category/patient cohort
Ambulatory Surgeries
Total # surgeries
By high volume CPT codes or by ambulatory
surgery categories
Total # surgeries/1,000 enrollees
By high volume CPT codes or by ambulatory
surgery categories
Number of Providers
Primary care physicians
Specialists
Other (e.g., mental health providers)
Number of Enrollees
Total #
By eligibility category
By age/sex categories
Number of Users (i.e., enrollees who used
services)
Total #
By eligibility category
By age/sex categories
Visits
Total #
#/enrollee
#/user
by visit categories (e.g., well child, well adult,
ob/gyn, mental health, substance abuse, etc.)
Other Services (e.g., prescription drug)
Total #
#/enrollee
#/user
by service category

Page 25

Attachment 3

Sample Medical Record Review Findings Tool
for Encounter Data Validation
Patient ID Number:

Medical Record Number:

_

Patient Name:

Completion Date:

_

Provider Name:

EQRO Reviewer:
Coder Reviewer:

_

Attending Physician Name:
Visit Dates:

Begin Date:

End Date:

_

Required Review: (Check one)
[

] Office Visit - (excludes dental and mental health / substance abuse visits)

[

] Office Visit - mental health / substance abuse

[

] Office Visit - dental

[

] Inpatient admission - (excludes mental health / substance abuse visits)

[

] Inpatient admission - mental health / substance abuse

[

] Other types of encounters as specified by the State; e.g., laboratory, pharmacy, physical
therapy. Specify:_______________________

Diagnosis Codes and Descriptions
Diagnosis Code

Match

No Match

Diagnosis Description

a.
b.
c.
d.
If the diagnoses in the record do not match the billed information, write the correct diagnosis
description(s) on the lines provided.

Page 26

Procedure Codes and Description
Procedure Code

Match

No Match

Procedure Description

a.
b.
c.
d.
e.
f.
If the procedures in the record do not match the billed information, write the correct procedure
description(s) in the spaces provided.

Revenue Center Codes and Descriptions
Revenue
Center

Revenue Center
Descriptions

Match

No

Correct Revenue

Match

Center Description

Correct Code

Codes
a.
b.
c.
d.
e.
If Revenue Centers in the record do not match the billed information, write the Correct Revenue Center
Description(s) in the spaces provided.

NOTE: The EQRO should tailor and add to this form to address all data
fields under review.
Page 27

END OF PROTOCOL ATTACHMENTS

Page 28