Below are answers to commonly asked questions organized into six categories.
Q. What rule governs how MHDO releases data to the public?
A. MHDO Rule Chapter 120 governs the MHDO’s release of data to the public. You can find a copy of that rule on our Rules and Statutes page.
Q. How do I request data from MHDO?
A. You can find out how to request data on our Request Data or Reports page.
Q. What data can I request?
A. The Data Availability page lists the available data sets. It also shows the organization types that supply the data, the information contained in the data, and the time periods the data represent.
Q. What is the comment period and who can submit a comment?
A. The comment period is the period of time a data request is publicly posted for comment. The comment period allows the data providers and the public to review the data request and submit concerns about the data release to the MHDO. MHDO posts new data requests to its website, on its Current Data Request page, on the first business day of every week; MHDO also sends an electronic notification to the data providers who submit the data and other interested parties, notifying them of new data requests. The comment period is 30 business days after the request first appears on the MHDO website.
Q. How much does it cost to access data from MHDO?
A. MHDO’s data access fees can be found on our Pricing Information page.
Q. What is the difference between the APCD (All Payer Claims Data) data and the Hospital encounter data?
A. Payers submit claims data to the MHDO and hospitals submit hospital encounter data for all encounters inpatient and outpatient as well as for their provider based clinics.
The APCD contains health care claims paid for Maine residents by insurance companies licensed in Maine. It also includes Medicare claims paid for Maine residents and MaineCare (Maine Medicaid) claims. The submissions include files with member eligibility, medical claims, pharmacy claims, and/or dental claims information. These data contain information on the amounts that insurers paid for services. Please see the Data Availability page to see what types of entities submit data for each source and what data are available.
Hospital encounter data are submitted by hospitals and provider-based clinics in Maine. These submissions contain data for every service provided to each patient, regardless of how they were paid (commercial insurer, public, uninsured). However, these data do not include financial information on amounts billed or paid. Please see the Data Availability page to see what types of entities submit data for each source and what data are available.
Q. Does the MHDO data (APCD and hospital encounter) include patients who are uninsured and/or who paid for their care, or only data from payers?
A. Hospital encounter data does include patients who are uninsured but does not include payment information. The Maine APCD does not contain data on patients who are uninsured.
Q. What are the main differences between Level I and Level II data sets?
A. MHDO's Level I data sets are considered de-identified data which means the data elements do not directly or indirectly identify an individual patient and for which there is no reasonable basis to believe that data can be used to identify an individual patient.
MHDO’s Level II data sets are considered a limited data set which includes limited identifiable information specified in the HIPAA regulations. Level II data sets may only be used in ways that maintain individual anonymity.
Q. I am interested in purchasing a Level II data set. Do I need to purchase both the Level I and Level II data sets to obtain all the releasable data fields available?
A. No, the Level II data set contains the data elements in Level I, so there’s no need to purchase both.
Q. Is it possible to obtain patient level and provider level data? Is this identified or de-identified data?
A. In the Level II data sets for both APCD and Hospital encounter data, de-identified patient data are available and each patient has a MHDO assigned unique identifier. Identifiable provider data is available for both the APCD and hospital encounter data. The MHDO data dictionary is available and provides details on the specific data elements. Some highlights of the provider data include facility name, facility code, geographic information, service provider state, data processing center code and a national provider identifier where appropriate. For the Hospital encounter data, a unique hospital or clinic is provided for each encounter with the name of the facility, geographic location, and de-identified ordering and performing provider.
Q. What format are the data sets released in?
A. MHDO Data sets are released in fixed-width text (.txt) files. These can be imported on a variety of platforms and are widely recognized by data management and statistical analysis tools. Data users typically use a database like SQL server or a statistical analysis package like SAS.
Q. Is there a way to link the APCD and Hospital encounter data sets together?
A. No, not at this time as there are no direct data elements released in the data that would allow the linkage between the All Payer Claims Database (APCD) and the Hospital encounter data. The MHDO has discussed the feasibility and utility of linking these data sets internally and making available as a combined data set. At this time there is limited utility in developing one patient index across the different data streams (claims and hospital data). Instead MHDO is focusing our resources on making sure that we have the data elements needed to identify unique individuals in the data streams so that we can accurately link records internally and release de-identified records.
Q. Once I’ve been approved to receive MHDO data, how will the data be sent?
A. There are two Options for how Authorized MHDO Data Users can access and work with MHDO datasets
| Specifications |
Direct Data Access in Data Enclave
Analyst (Standard Subscription Fee) |
Direct Data Access in Data Enclave
Custom Project (DE Estimated Fee) |
|---|---|---|
| Data Access & Formats | Access database tables/views in Vertica HP (read only) | On request, requiring Feasibility Assessment and Estimating by NORC |
| Storage for data files, reports, spreadsheets, documents |
50 GB personal H Drive:
200 GB shared storage available for the research team |
Custom Database Storage + Custom File Storage as agreed with NORC |
| Database Software | Vertica HP database with Dbeaver SQL Client Tool | Vertica HP database with Dbeaver SQL Client Tool + Other Data Management Tools if supported by NORC Data Enclave |
| Data Analysis Software |
SAS or STATA, R Studio, Python
Access to Tableau Desktop with provision of external license |
SAS or STATA, R Studio, Python + Other Data Analysis Software, if supported by DE
For Additional Software, customer needs to provide license; software should be Citrix compatible |
| Presentation Software | Microsoft Office Productivity Suite - Excel, PPT, Word, etc. | Microsoft Office Productivity Suite - Excel, PPT, Word, etc. |
| Download Files/Reports | 5 SDC-reviewed export requests per month (Aggregated Data files, Reports) | Custom Number of Downloads and Size, as agreed with NORC |
| Upload 3rd Party Data | 50 GB for Ancillary Data Files | Custom Number of Uploads, as agreed with NORC |
| Support Tier Description | 2 day response service level agreement (SLA) for technical issues, defined as: account management, application launch failure, and database connectivity errors. Guaranteed issue resolution. | 1 day response service level agreement (SLA) for technical issues, defined as: account management, application launch failure, and database connectivity errors. Guaranteed issue resolution. |
| Training | Self-guided onboarding to Data Enclave/MHDO Data via FAQs and short web-based video. 1 hour initial environment and data access training | Self-guided onboarding to Data Enclave/MHDO Data via FAQs and short web-based video. 1 hour initial environment and data access training |
| Data Objects Permissions | Read only access to data views for all available years in Vertica (low memory and temp table limits) | Creation of Temporary tables in Vertica is allowed. Tables/Views can be migrated to PROD (persistent) by NORC IT only |
|
Database New Objects
development (Tables/Views/Summary Tables) |
Creation of Temporary tables in Vertica is allowed | Creation of Temporary tables in Vertica is allowed |
| Request Process | Request and Approval by MHDO/and NORC Data Enclave Database Managers | Project Request and Approval to MHDO/scope & feasibility assessments by NORC |
| NORC Fee Structure |
Hosting Fee Per User/Per Year = $5,200
Additional File Storage = $400/100G |
Hosting Fee Per User/Per Year = $5,200
1 TB Vertica Storage ~ $20K / 1 TB 1 TB File Storage ~ $1,000 / 1 TB Virtual Machine (VM) server 24 CPU / 128 GRAM / 500 HDD with Windows 2012 R2 ~ $10,000 / 12 month Virtual Machine (VM) server 12 CPU / 64 GRAM / 500 HDD with Windows 2012 R2 ~ $5,000 / 12 month Managed Services Fee (OS patches, Vulner PT Monitoring, Backups, etc.) ~ $150 / machine / month Total Fees TBD based on final design |
| Note: MHDO Data access fees apply as described in MHDO Rule Chapter 50 | ||
*Process for the download option: When the data are ready to be released, you’ll receive an e-mail from MHDO's data vendor, NORC with the links to the data release files to be downloaded. We use the file sharing platform Accellion to ensure a secure file transfer. If you do not have an Accellion account, you’ll need to register for one.
The following instructions will be included in the e-mail:
To begin downloading the files you’ll need to click on one of the links. Because
this is a secure download you will be asked to enter the address that received the
e-mail. If you do not already have a download account, you will receive a second
verification e-mail. Click on the link in that e-mail, choose a password, and you
will then be able to authenticate and download the files.
Q. I received an email to download MHDO data, but I need one of my colleagues to download the data. Can I forward them the email?
A. No, the data transmittal email cannot be forwarded to others. Only original recipients are authorized to download the files. Please contact MHDO to modify the Data Custodian of your request if needed.
Q. How long will I have to download the data?
A. You have 90 days after receipt of the e-mail with the instructions to download the data. If you need to access the data after this time you will need to e-mail the MHDO at Webcontact.MHDO@maine.gov and include your Data Request Number in the subject line. The e-mail should explain why you need extended access to download the data files.
Q. What is the average time it takes to download one of MHDO’s larger files?
A. There are a number of factors that can influence the download speed but typically the largest files can take anywhere from several hours to a day to download.
Q. What do I do if I have questions regarding the data?
A. Send your questions in an e-mail to MHDO at Webcontact.MHDO@maine.gov with your Data Request Number in the subject line of the message.
Q. How should we utilize the external cause of injury and Present on Admission Indicators (POA) variables? Which diagnosis codes do they refer to?
A. For the ICD-10 fields:
Q. Is there a Present on Admission Indicator (POA) for the single ICD-9 ECODE?
A. No, there is no Present on Admission Indicator for any of the ICD-9 fields in the MHDO APCD data.
Q. Is there a Present on Admission Indicator (POA) for MC200_PRINDGNS (the principle ICD-10 diagnosis)?
A. Yes, the Principal Diagnosis field (MC200_PRINDGNS) also has a corresponding Present on Admission indicator (MC201_POA).
Q. Why are there a significant number of HCPCS codes in PRNPRCDRCD field in 2014 and 2015 when it is supposed to be an ICD-10 procedure code field? Should HCPCS codes be removed from the PRNPRCDRCD field when using this data?
A. The ICD-10 fields were added to the Rule that governs the submission of APCD
data Rule Chapter 243 a year before the official cutover on 10/1/2015. This was
done so that data submitters could perform their implementation and testing. The
MHDO did not perform validation on the ICD-10 fields until the cutover. This field
has always been an ICD-10 field, however claims incurred before the 10/1/2015 cutover
to ICD-10 should have populated the MC058 ICD-9 procedure code field.
Any non-ICD-10 values in an ICD-10 field (or non-ICD-9 values in ICD-9 fields) should
be considered invalid. Even if it is a valid HCPCS code, it was populated by the
payer in error and should not be relied upon.
Q. Are the ICD-9 procedure codes indeed header level variables?
A. Yes, ICD-based procedure fields, required to be filled out only on inpatient claims, are claim - level fields which would generally be included in a claim-header file. In other words, we would expect the values in these ICD-9 or ICD-10 procedure fields to be the same across claim lines. The claims data submitted by payers to the MHDO is claim line-level data; we do not receive separate claim-level information. If submissions included cases of claims where ICD-9 field values varied across claim lines, that would be a data quality issue. The MHDO does not currently create a claim-header file; this aggregation would need to be done by the data user. The MHDO is planning to create claim header analytic files and make them available for request in early 2019.
Q. Can a single claim have multiple ICD-9 procedure codes?
A. Inpatient claims may have a single value for principal procedure (MC058) recorded at the claim level. There are no fields in the layout for the documentation of other ICD-9 Procedure codes in the rule that governs the APCD data submission, MHDO Rule Chapter 243 which can be found on our Statutes and Rules page: https://mhdo.maine.gov/rules.htm
Q. How do I identify the principle procedure if a claim has multiple ICD-9 procedure codes?
A. This type of scenario would not be expected as per CMS claim reporting requirements. The value present in field MC058 should be consistent across claim lines of the same claim ID, and it represents the principal procedure. If variation is observed, those instances are most likely data quality issues.
Q. What is the field locator that indicates the admit source?
A. The field locator that indicates the admit source is IP2007_PointOrigin. Be sure to refer to NUBC's current list of values for Point of Origin and not the discontinued Source of Admission. More information is available via the MHDO Data Dictionary here: https://mhdo.maine.gov/mhdo-data-dictionary
Q. Does the Maine APCD contain social determinants of health information directly linked to patients?
A. With the change in coding systems that occurred in October 2015, from the International
Classification of Diseases, Ninth Revision (ICD-9) to the Tenth Revision (ICD-10),
new ICD code values were introduced that can be used to track “potential health
hazards related to socioeconomic and psychosocial circumstances.” These types of
codes are in the Z55-Z65 range and can be reported by physicians and other providers
through the principal and other diagnosis fields on health insurance claims. For
more information about this subset of ICD-10 codes and its subcategories, reporting
practices and efforts to improve reporting, please consult the following American
Hospital Association resource at:
https://www.aha.org/system/files/2018-04/value-initiative-icd-10-code-social-determinants-of-health.pdf.
Aside from this, the Maine APCD does not include other social determinants of health
information. In the future, MHDO may be looking to enhance the data released to
the public by making available additional data about patient’s location of residence
from external, non-APCD data sources (for example, information on socioeconomic,
educational, health and other factors that describe a patient’s neighborhood or
census tract).
We have confirmed that a small share (1-2%) of the Maine APCD claims for health
services received between October 2015 and June 2018 do have such environmental
factors codes present. When present, it is most likely to be on a Medicaid (or MaineCare)
claim—as opposed to commercial or Medicare claim—and more likely to be from a relatively
small subset of Maine healthcare providers (25-100 providers) that either possibly
serve a larger share of population for which reporting on these factors is more
relevant or may have had substantial training on verifying and reporting, if applicable,
these factors for their patients. Given the observed share of claims with at least
one Z55-Z65 code present, it would be safe to assume that such environmental factors
are underreported in the Maine APCD.
Q. How are the ICD-10 codes validated on intake?
A. These codes are submitted to the Maine APCD on the ICD-10 Principal Diagnosis and Other Diagnosis fields on medical claims, validated against the standard list of ICD-10 codes, and made available to data users as part of the standard release data elements for Maine APCD Medical Claims. Original code values submitted to APCD are not edited; if the validation checks discover that there are issues in these fields, MHDO works with the submitter to clarify the issue and, in some cases, resubmit the data after addressing it. Detailed information on validations applicable to the ICD-10 Principal Diagnosis and Other Diagnosis fields on medical claims is available in this spreadsheet.
Q. The data we receive and the entity relationship diagrams (ERDs) refer to a ‘submitter’ field which is not in the MHDO Data Dictionary. Can you tell us what will be populated in this field?
A. The ‘submitter’ field is referring to the XX001_PAYER fields (ME001_PAYER, MC001_PAYER, PE001_PAYER, PC001_PAYER, etc.). These fields contain the MHDO-assigned identifier of payer submitting claims data. The XX002_NPLAN fields (ME002_NPLAN, MC001_NPLAN, PE001_NPLAN, PC001_NPLAN, etc.) contain the MHDO-assigned code of the insurer / underwriter in the case of premiums-based coverage, or of the administrator in the case of self-funded coverage. More information for each field is available in the MHDO Data Dictionary.
Q. Where can I go to see which data elements are included in the APCD?
A. The MHDO data dictionary is available on our Data Dictionary page.
Q. Is the MHDO data dictionary the same across all payers (e.g., commercial, MaineCare and Medicare)?
A. Yes, the data dictionary applies to all claims records regardless of payer.
Q. Does your medical claims database include both professional and facility claims?
A. Yes, both professional and facility claims are included in the Maine APCD.
Q. Do I need to request the Practitioner Identifiable data elements if I am looking to identify a hospital as a provider in the APCD data?
A. No, you do not need to receive the Practitioner Identifiable data elements to identify a hospital. A facility table is provided in both the Level I and Level II data releases. The table allows data users to link the hospital identifier in the data and determine the claim lines associated with each hospital.
Q. Does the APCD contain denied claims?
A. No, the MHDO claims data submission, Rule Chapter 243, specifically excludes denied claims from all medical, pharmacy, and dental claims file submissions. When a claim contains both approved and denied service lines, only the approved service lines are included as part of the health care claims data set submittal. You will typically see reversals in the data with a claim status code of 22 and negative values in the data are adjustments to previous paid claims. Generally, payers fully reverse a claim line before issuing a replacement. However, we know that there is some variance in this practice. For instance, some Payers will only issue a partial adjustment (meaning that the sum of the entire claim still gives you the right bottom line figure).
Q. What is the difference between the APCD claims datasets and the eligibility data sets?
A. Claims data sets include:
Eligibility data sets include:
For more information, please see our Data Dictionary page.
Q. Can eligibility data be linked to the claims data?
A. The eligibility data can be linked to the claims data; in fact, one of the pre-release QC steps that we perform is checking the match rate between the claims data. You can find the match report from the most recent Release Report on the worksheet "MC Match to Eligibility".
However, the linkage of claims to eligibility can be complex because a payer may submit multiple eligibility records for an individual for a given month, each for a different product type. The eligibility associated with a claim may also come in under a different payer code than the claim.
The fields DC902_IDN, MC902_IDN, ME902_IDN, and PC902_IDN are unique row identifiers within their respective tables. They aren't used for inter-table linkages. Claims data can be linked to corresponding eligibility records by MHDO Member ID and incurred date. The following example SQL shows how the MC data can be linked to the ME data (the table names are our standard level II MC and ME release tables):
SELECT *
FROM vwRelease_MC_Level2_Base a left join
vwRelease_ME_Level2_Base b
ON a.MC911_MHDO_MEMBERID = b.ME910_MHDO_MEMBERID
and YEAR(a.MC059_FDATE) = b.ME004_YEAR
and MONTH(a.MC059_FDATE) = b.ME005_MONTH
;
This will return MC claim records and any eligibility records the member has for the month and year that the service was incurred. Please note that, depending on your analysis, you may want to restrict the linkage further by payer code, coverage type or insurance product type. Individuals may have more than one eligibility row for a given month and year; depending on your analysis, you may need to take this into account to prevent duplicating claim lines.
For instance, to restrict an analysis to only medical claims that have a corresponding ASO or ASW eligibility record from the same payer, you could do a query like the following:
SELECT *
FROM vwRelease_MC_Level2_Base a
WHERE EXISTS
(
SELECT NULL
FROM vwRelease_ME_Level2_Base b
WHERE a.MC911_MHDO_MEMBERID = b.ME910_MHDO_MEMBERID
and YEAR(a.MC059_FDATE) = b.ME004_YEAR
and MONTH(a.MC059_FDATE) = b.ME005_MONTH
and a.MC002_PAYER = b.ME002_PAYER
AND b.ME029_COVERAGE in ('ASO','ASW')
)
;
This query will not duplicate medical claim lines even if there are more than 1 qualifying eligibility records available. It will only include medical claim lines that have at least one eligibility record that meets the match criteria. The bottom line is that linkage is possible, but it isn't as simple as linking on a single field. How the linkage is done depends on the specific analysis you are doing.
Q. Will the information provided include carriers that cover employer groups who reside in Maine?
A. The Maine APCD includes all payors as defined by MHDO with greater than $2,000,000 of adjusted premiums or claims processed per calendar year. Please see the Data Availability page to determine if a specific payer would be included in the requested dataset. . Note: As a result of the Gobeille v. Liberty Mutual decision (March 1, 2016) submission of claims data for self-funded ERISA plans is no longer a State requirement. However, we are accepting these data submissions on a voluntary basis.
Q. Does the All Payer Claims Database (APCD) include Medicare Advantage?
A. Yes, the MHDO receives Medicare Part C claims data also referred to as Medicare Advantage Plans. There is a flag in the claims that identifies Medicare plans administered through commercial payers. We have separated these from Medicare plans administered by the government. Therefore, you can determine which claims were Medicare Advantage claims.
Q. Does the APCD Pharmacy Claims include all payers (i.e., MaineCare, Medicare, and private insurance)?
A. The APCD includes pharmacy claims data from the private insurance companies and from Medicare Part D plans (outpatient Prescription Drug Insurance). Part D is provided only through private insurance companies that have contracts with the government.
Q. In the APCD, is there any kind of identifier that we can use to follow an individual across insurers and over time? Are encoded beneficiary identifiers the same across all parts of the data, and would it be possible to see claims of the same individual when they switch from a commercial plan into Medicare/Medicare HMO?
A. Every claim and eligibility record is assigned a MHDO de-identified member number to allow the tracking of de-identified individuals across time and between payers; however, due to variations in how payers supply eligibility information, sometimes the same individual may be assigned more than one MHDO de-identified member number. In general, the Medicare data includes eligibility data that allows us to assign a de-identified member number consistently The Medicare Advantage data are not always submitted with consistent member eligibility data. This is also an issue with pharmacy claims, where a number of plans only provide a contract number that cannot be tied to an individual in the medical claims. For Medicare Part A and B claims data administered by the government, we receive only medical claims. For Medicare Advantage plans, we receive both pharmacy and medical claims.
Q. How can I link data across files? Is there a Member ID to use?
A. The level II data elements include an MHDO assigned member ID. However, this is only useful for tying records together within the APCD. So, for instance, if you requested both pharmacy and dental claim information, this member ID would allow members to be tied together both between and within the datasets. The actual member identifiers sent by the payers (PC007_SUBSSN Subscriber Social Security Number, PC009_SEQNO Member Suffix or Sequence Number, and PC010_MEMSSN Member Identification Code) are not releasable under MHDO Rule Chapter 120. The MHDO replacement value for a member’s SSN is releasable as a level II data element, but that is an integer substituted value that won’t be able to be linked to non-APCD data.
Q. Which field in the pharmacy claims can be used to determine dosage?
A. The MHDO receives information on the quantity of medication dispensed through the PC033_QTY Quantity Dispensed field. The NDC code of the drug being dispensed is available in field PC026_NDC Drug Code. The NDC code can be looked up on the FDA’s website to determine the active ingredient strength. So, if a claim has a PC033 value of 90 and a PC026 value of 6818051701, we can determine that the member has received 90 tablets of 40mg Lisinopril. (https://www.accessdata.fda.gov/scripts/cder/ndc/index.cfm - search for NDC Code 68180-517-01). Note that the FDA site requires the addition of hyphens to the NDC code following one of three schemes. For this reason, it is often quicker to simply use Google (e.g., “NDC 68180051701”) to get the information via a third party site such as ndclist.com. Alternately, the FDA makes a list of NDC information available at https://www.fda.gov/drugs/drug-approvals-and-databases/national-drug-code-directory; the product table provides active ingredient information.
Q. What geographical identifiers are available in the APCD data? From the data dictionary, it seems that for both plan members and providers, city name, state/province, and ZIP code are available. Could you confirm this?
A. The geographical data elements available in the APCD data are as follows for both providers and de-identified individuals:
Geographical information related to city, state and zip has to be specifically requested and justified during the data request process.
Q. Using the claims information, I need to be able to track the exact health insurance plan for each submitted claim. What are the plan identifiers in your data?
A. The APCD data includes a field which indicates the payer and product type (HMO versus PPO, etc.) but not the specific plan identifier. The claims file has the claim filing indicator code and the eligibility file has the insurance type code which can be referenced in ASCx12.
Q. How are medical claims versioned in the data?
A. According to Rule Chapter 243, the original claim will have a version number of 0, with the next version being assigned a 1, and each subsequent version being incremented by 1 for that service line. There are some payers that are versioning claims differently from what is described in Rule Chapter 243. We have asked these payers for their methodologies and will post them and update this FAQ by July 2017.
Q. Is there a way to identify prescriptions that were filled through mail-order?
A. There is no mail-order indicator, however the pharmacy NPI is released which allows for the pharmacy to be identified.
Q. Can the same member identifiers in past files, be used with future files?
A. The member IDs we assign are consistent across time so that data users don’t
need a full data refresh with each data release- when we send a new quarter of data,
the user should be able to append it to the previous MHDO data that you received.
For every quarterly release, we include information on the member match to eligibility,
which represents the percentage of claims that have a matching eligibility record
for the member. In a recent release, which included all of calendar year 2016 for
commercial data, Q2-Q4 2016 for MaineCare (Medicaid) data and Q3-Q4 2015 for Medicare
data, we had high overall match rates – 97.9% for medical claims, 96.2% for dental
claims and 98.7% for pharmacy claims. MHDO is looking into the stability of the
member identifiers across time and will follow up with users in 2017.
Note: Every claim and eligibility record is assigned a MHDO de-identified member
number to allow the tracking of de-identified individuals across time and between
payers; Note: due to variations in how payers supply eligibility information, sometimes
the same individual may be assigned more than one MHDO de-identified member number.
Q. How does the MHDO define runout?
A. Runout claims data is data with an incurred date within the data request period that is paid after the data request period ends. MHDO’s default runout period is 6 months. Example: if you request MHDO claims data through Q4 of 2016, you will receive data through Q2 2017 (the 6-month runout period will include records incurred in 2016 that were paid during Q2 of 2017).
Q. Where do I find the insurance type or product code that indicates the type of insurance coverage the individual has?
A. Refer to MHDO Rule Chapter 243 for the source information for these fields. Depending on the file type (medical, pharmacy, dental, eligibility), this information is released in the following fields.
Q. How do I interpret the Medicare product codes?
A. Below is a list of ME912_MHDO_PRODUCT codes that are non-standard and specific to Medicare.
Q. How is Medicare Part C and D coded?
A. This information can be found in the MHDO_PRODUCT fields in the release files.
In the eligibility files, these are the codes:
In the claims files, these are the codes:
Q. Does the MHDO get date of death for Medicare patients from CMS and if so, is it included in any of the MHDO releases?
A. Date of Death is not a releasable field and not included in our data releases.
Q. Where can I go to see what data elements are included in the Hospital data sets?
A. The MHDO data dictionary is available and provides details on the specific data elements included.
Q. Are there sample data layouts available?
A. The Excel documents "Outpatient Level II Base Layout" and “Inpatient Level II Base Layout” provide a detailed description of the data layout. The MHDO hospital data releases includes ICD-10 data elements. In general, records with discharge dates before 10/1/2015 used ICD-9 data elements while those discharged on or after this date used the new ICD-10 data elements. Note: ICD-9 and ICD-10 data elements are stored in different fields. For instance, the ICD-9 principal diagnosis is stored in the field OP7004_PrincipalDiagnosisCode while the ICD-10 version is stored in OP7104_PrincipalDiagnosis.
Whenever possible, data elements are prefixed with the data element name from the input layout specified in Rule Chapter 241, Uniform Reporting System for Hospital Inpatient Data Sets and Hospital Outpatient Data Sets, which provides additional details on the derivation of each element in Appendix B-2.
Value add fields that are not directly mapped from the input layout are prefixed with identifiers that begin with 4 alphabetic characters, to allow easy differentiation. For example, the patient age, which is calculated using date of birth and the admission/start of care date, is found under the field name OPMVA21_AGE and IPMVA21_AGE.
Effective with the implementation of MHDO’s Rule Chapter 120 –Release of Data to the Public, by default, data releases will include patient county and age. Patient city, Zip code and date of birth will require special justification for inclusion.
Q. Do you have data dictionaries for the hospital encounter outpatient and inpatient data sets?
A. Yes, the MHDO data dictionary is available on our Data Dictionary page.
Q. When each quarter of CY17 hospital inpatient data is released, will it only be that quarters data or are you also releasing the most recent 12 months of data ending in that most recent quarter?
A. The quarterly Inpatient Hospital files are meant to represent encounters that happened at a facility during that quarter, and only in that quarter. For example, the Q1 2017 Inpatient encounter file will include data for the months January- March 2017. The annual 2017 Hospital Inpatient Encounter file would be a compilation of the four quarterly files.
Q. What diagnosis fields are available in the current Rule Chapter 241 layout for the Hospital data?
A. In the current version of Rule Chapter 241, the diagnosis fields that are available are Version 040 which has one principal diagnosis, one admitting diagnosis code, two external injury codes, and eight other diagnosis codes (ICD-9); the versions 050 and 060 layout has one principal diagnosis code, one admitting diagnosis code, two external injury codes, and eight other diagnosis codes (ICD-10). Note: ICD-9 and ICD-10 data elements are stored in different fields. For instance, the ICD-9 principal diagnosis is stored in the field IP7004_PrincipalDiagnosisCode while the ICD-10 version is stored in IP7104_PrincipalDiagnosis. Whenever possible, data elements are prefixed with the data element name from the input layout specified in Rule Chapter 241, Uniform Reporting System for Hospital Inpatient Data Sets and Hospital Outpatient Data Sets, which provides additional details on the derivation of each element in Appendix B-2.
Q. What is the MHDO-assigned Medical Record Number (MRN)?
A. The MHDO-assigned Medical Record Number (MRN) is an obfuscated and transformed version of the MRN that is submitted by facilities to uniquely identify patients. Data users are reminded that the MHDO-assigned Medical Record Number generally cannot be used to track individuals between facilities; the same MRN may be used at different facilities to represent different individuals. Also, even within the same facility, an individual may not retain the same MRN across time; when hospitals merge or when they transition to new data systems, new MRNs may be assigned. The MHDO has no control over the MRN assignment policies within facilities. The MHDO is developing data elements that will allow an individual to be more reliably tracked both across time within a given facility and between facilities. MHDO plans consider a Rule Change to add patient SSN, patient name and patient street address to Chapter 241 Uniform Reporting System for Hospital Inpatient Data Sets and Hospital Outpatient Data Sets in 2017-2018 with an effective date of 2019.
Q. Is this to be implemented for the age at admission or the age at discharge?
A. Age at admission. In the MHDO Hospital Inpatient data, the age field IPMVA21_AGE is calculated based on the admission / start care date (IP2011) and patient’s date of birth (IP2005).
Q. If a patient is 90 years old, then the age on the data file is 90 and the date of birth is allowed to be non-missing? Or set to missing?
A. The rule of top coding applies to patients ages 90 or older. A person that is 90 years old at admission will have the age field IPMVA21_AGE set to value ‘90’ and the date of birth field IP2005_PatientDOB set to missing.
Q. If a patient is older than 90 years old (91, 92, etc.), then the age on the data file is set to 90 and the date of birth is set to missing?
A. The rule of top coding applies to patients ages 90 or older. A person that is 91 years old or older at admission will have the age field IPMVA21_AGE set to value ‘90’ and the date of birth field IP2005_PatientDOB set to missing.
Q. In the MHDO Hospital data, does Emergency Department data include both Inpatient and Outpatient data sets?
A. Yes, Emergency Department (ED) data is a subset of the Hospital for both the Inpatient and Outpatient data sets. The ED data follows the same format as the Outpatient Hospital data. It is processed with the Inpatient and Outpatient data and is available for the same time periods as these datasets.
Effective with the 2015 hospital data there is an ED Flag (IPMVA25_EDFLAG & OPMVA25_EDFLAG) in both the Inpatient and Outpatient data. This flag is set based on the presence of ED-related revenue codes present for each encounter in the Inpatient data and either ED-related revenue codes or ED-related CPT codes in the Outpatient data.
The methodology that we use to identify ED records is any inpatient or outpatient visit that has a revenue code of 0450, 0452, 0456, or 0459 are considered ED visits. In addition, any outpatient visits that include CPT code 99281-99285 are considered ED visits. Note: CMS guidelines call for these CPT codes to always be assigned to 045x revenue codes, however, it allows hospitals to associate these codes with other revenue codes if they had historically done so (https://www.cms.gov/Regulations-and- Guidance/Guidance/Transmittals/downloads/r167cp.pdf). Thus, inclusion of the CPT codes prevents overlooking ED visits due to historical billing practices at facilities or for non-CMS payers.
Update 3/1/2018: It came to our attention that we may be undercounting ED visits at Critical Access Hospital because certain payers are being split out and we cannot use revenue codes to identify ED-related visits. MHDO/HSRI will discuss with CAH’s the best way to identify all ED visits regardless of the payer for future releases. We will be updating our methodology once we find out how we can identify these ED visits.
Q. What Diagnosis Related Group (DRG) information is available in the Inpatient data.
A. MHDO assigns DRGs using the 3M Grouper software. Currently, two different sets of DRG codes and Major Diagnostic Categories (MDC) codes are created: one based on the All Patients Refined Diagnosis Related Groups (APR-DRG) and the other on the Medicare Severity-Diagnosis Related Groups (MS-DRG). The MDHO had previously also distributed two older versions of DRGs (AP-DRGS and CMS-DRGs) which have since been depreciated by their maintainers.
DRG information available in the Inpatient Encounter Data:
| Field Name | Description | Data Type | Length |
|---|---|---|---|
| IPMG20_MSDRG | MS-DRG code | Varchar | 3 |
| IPMG21_MSMDC | MS-MDC code | Varchar | 2 |
| IPMG22_MSVer | MS-DRG verion number | Integer | 19 |
| IPMG23_APRDR | APR-DRG code | Varchar | 3 |
| IPMG24_APRMDC | APR-MDC code | Varchar | 2 |
| IPMG25_APRVER | APR version number | Integer | 19 |
Both the APR-DRG and the MS-DRG are revised annually. The DRG version used is available in the data and will be determined by the discharge date on the encounter record following the table below.
DRG Version Table:
| Calendar Year | DRG Version | Date Range |
|---|---|---|
| 2013 | v30 | 1/1/2013 - 9/30/2013 |
| v31 | 10/1/2013-12/31/2013 | |
| 2014 | v31 | 1/1/2014 - 9/30/2014 |
| v32 | 10/1/2014-12/31/2014 | |
| 2015 | v32 | 1/1/2015 - 9/30/2015 |
| v33 | 10/1/2015-12/31/2015 | |
| 2016 | v33 | 1/1/2016 - 9/30/2016 |
| v34 | 10/1/2016-12/31/2016 | |
| 2017 | v34 | 1/1/2017 - 9/30/2017 |
| v35 | 10/1/2017-12/31/2017 | |
| 2018 | v35 | 1/1/2018 - 9/30/2018 |
| v36 | 10/1/2018 - 12/31/2018 |
Note each version starts in the 4th quarter of one year and extends through the end of the 3rd quarter of the subsequent year. This means that in each year's hospital encounter data, we will have two different versions of the DRGs—one for Q1-Q3 and the subsequent version for Q4.
Q. What do the MHDO Assigned Payer Category Codes mean?
A. In 2015, we updated the MHDO assigned payer category codes based on conversations with the data user group. The codes ‘12’ Medicare Advantage and ‘00’ Unknown were added, while ‘05’ Blue Cross has been removed. Blue Cross is now recoded as ‘06’ Commercial Carriers. In addition to the category codes, we also include the National Association of Insurance Commissioners (NAIC) Payer Code and Payer Name as received on the encounter records. Payer Names are released if the name does not disclose an individual - over 90% of payer names are released.
As part of the Payer Category Code assignment, MHDO codes payers with 10 or more encounters in the data warehouse at the time of the data processing. For encounters through December 2017, Payer Names below this frequency threshold have a blank Payer Category Code value. Starting with the 2018 data, Payer Names below this frequency threshold have a Payer Category Code value of ‘99’, meaning ‘undetermined’. The payer name frequency is reassessed periodically and values that pass the “10 or more encounters” threshold—whether previously assigned a value of ‘99’ or left null in the Payer Category Code—will be included in the assignment process.
| MHDO Assigned Payer Category Code | Category Name | Payer Name Examples |
|---|---|---|
| 01 | MEDICARE | MEDICARE, MEDICARE A B, MEDICARE PART A IP |
| 02 | MEDICAID | MAINECARE, MEDICAID - OUT OF STATE |
| 04 | TRICARE/USVA | CHAMPVA, TRICARE, VA TOGUS |
| 06 | COMMERCIAL CARRIERS | AETNA HMO, ANTHEM BCBS, UNITED HEALTH |
| 07 | CHARITY/UNCOMPENSATED CARE | DISCOUNTED CARE, FREE CARE, UNCOMP CARE |
| 08 | SELF PAY | SELF PAY NO INSURANCE, S/P SELF PAY |
| 09 | WORKERS COMPENSATION | BATH IRON WORKS WC, LIBERTY MUTUAL WC |
| 11 | OTHER | RISK MANAGEMENT, HOSPICE |
| 12 | MEDICARE ADVANTAGE | AETNA MEDICARE HMO, HUMANA MEDICARE |
| 00 | UNKNOWN | MH NET, RM DEPT, MISC PAYER MISC ADDRESS |
| 99 | UNDETERMINED |
Q. What does the data element IPMVA23_RVHITS represent in the MHDO Inpatient hospital data?
A. The data element IPMVA23_RVHITS represents a count of the unique ancillary revenue code (IP60) values that are reported on inpatient encounter records.
Important notes:
IPMVA23_RVHITS is not calculated as the maximum sequence number on the ancillary revenue (IP60) rows that belong to distinct inpatient encounter records (meaning, the total number of rows among rows with the same encounter identifier, IP60AD02_MasterIDN, in the IP60 dataset).
The ancillary revenue (IP60) dataset is structured so that one distinct IDN could have multiple rows and on each row there can be up to three positions populated with a revenue code.
To calculate the count one would need to look both across rows (in as many rows as shown by the IP6002 Sequence Number field) and across columns (three revenue code fields: IP6004, IP6006, IP6008) in the IP60 dataset.
Example: If we have a distinct inpatient encounter with 3 rows in the IP60 dataset (3 rows of data sharing the same IP60AD02_MasterIDN value), there would be a total of 9 “cells” or revenue code positions that could have revenue code values. If only 7 out of the total of 9 “cells” that can hold revenue code values are populated, and they have 7 distinct values, the calculated IPMVA23_RVHITS would be ‘7’.
For situations where a revenue code value appears more than once for a particular IDN, it is counted a single time. As such, in the aforementioned example, if a revenue code appeared twice or more out of the 7 populated values, IPMVA23_RVHITS would be less than 7.
Q. How are geocodes assigned?
A. If your data request includes the release of ZIP codes, you will also receive a geocode. The MHDO assigns a geocode when the city, state, and ZIP code match the entries that appear on the canonical list of geocode values (a data table of Maine geocodes provided by the Maine Office of GIS (MEGIS) is the canonical list used for the assignment and is included in this release). The MHDO will not impute geocodes based on incomplete or conflicting city, state, or ZIP code information. In order to improve our ability to assign geocodes, we are working with those facilities that have not provided consistent city, state, and ZIP code information. Since the beginning of 2016 we are able to assign a geocode for over 99% of the encounters.
Q. How are Hospital Service Areas (HSA) assigned?
A. If your data request includes the release of ZIP codes, you will also receive HSA assignments. These assignments are created by the Dartmouth Institute for Health Policy and Clinical Practice and made publicly available on their website. In historical hospital data releases for data through December 2020, two different assignment methodologies were available to data users. One (OPML23_OriginalSA or IPML23_Original_SA) was developed in 2004 based on the Dartmouth Institute methodology and the other (OPML25_MMC_HSA or IPML25_MMC_HSA) was developed in 2016 based on a methodology developed by the MaineHealth/Maine Medical Center Planning Department, which was also based on the Dartmouth Institute methodology with a few modifications.
If a data user is going to publicly release in the form of a report, presentation or on a website, the HSA assignments contained in the MHDO inpatient and outpatient hospital data (beginning with 2020 Q4 hospital inpatient data and 2020 Q1-Q4 hospital outpatient data) the following citation must be made:
"The data used to determine the HSA assignments was obtained from the Dartmouth Atlas Data website, which was funded by the Robert Wood Johnson Foundation, The Dartmouth Clinical and Translational Science Institute, under award number UL1TR001086 from the National Center for Advancing Translational Sciences (NCATS) of the National Institutes of Health (NIH), and in part, by the National Institute of Aging, under award number U01 AG046830."
Q. What is the Location of Service in the outpatient data set?
A. The starting with the 2016 data, each release includes a new coding scheme that allows the field OP4005_LocationofService to be categorized as either a 1= Hospital Outpatient encounter or a 2 = Other locations, such as clinics, labs or physician practice. This new field appears on the base record in OP4005A_LOS_Category. The support table vwSupport_LOS_Codes contains descriptions of the LOS provided by the hospitals. The 2018 data is being submitted under new requirements that the LOS only be populated for physicians/physician practices. When OP4006 is populated with a subset of Place of Service codes (11, 17, 20, 22, 49, 50, 71, 72), the LOS field OP4005 must be populated and the full name of the physician/physician practice along with the other information must be listed in the LOS crosswalk.
Q. What does it mean if the LOS is blank?
A. If the LOS field is blank for a given record, this usually indicates that the location of service is at the main facility indicated by the OP0102_SubmitterEIN field rather than an associated clinic or other sub-facility. As of the submission of 2018 data, the LOS can be blank when Place of Service (OP4006) is not populated with codes 11, 17, 20, 22, 49, 50, 71, 72
Beginning with the 1st Quarter of 2018 Inpatient data and 1st Quarter of 2017 Outpatient data, the value add field for patient’s age (IPMVA21 and OPMVA21) has a null value if the calculated age is above 110 years old, or if the admission/start of care date or statement covers period from date (IP2011 or OP2012, respectively) represents a value later than the discharge/statement covers period through date (IP2013 or OP2013, respectively).
Q. What does the Length of Stay calculation represent in the Inpatient and Outpatient data?
A. The Length of Stay field is calculated for both Inpatient and Outpatient data, however it reflects two different types of duration concepts. In Inpatient data, the Length of Stay field (IPMVA20) reflects the time period of hospitalization (the actual duration of the inpatient stay), whereas in the Outpatient data (OPMVA20) it pertains to the time period between the earliest services on the encounter record and the most recent services on the encounter record, so it does not represent a continuous “outpatient stay”.
Q. What should be submitted in Record Type 73 (OP7301 – OP7327) if there was not an external cause of injury? Should we submit empty records?
A. OP73 codes are only expected when there was an external cause of injury. If there is an S or T code (Injury or Poisoning) in OP7104, the first Ecode field is enforced. Only include records when there are values to report; do not send empty records.
Q. When should Record Type 74 (OP7401 – OP7427) records be populated?
A. The OP74 records will only be populated if there are additional (“other”) diagnoses that need to be provided in addition to the principal DX code. Do not send empty records.
Q. How do I use the Sequence Number fields such as IP7402?
A. The field IP7402_SequenceNumber derives from Chapter 241 field IP7402 and can have a value of 1 or 2, meaning that the facility can submit up to two of these records per inpatient encounter. Each IP74 record has 12 “slots” for ICD-10 CM “Other Diagnosis Information”. This means that the facility can submit up to 24 distinct “Other” DX codes for an encounter: the first 12 would appear on the record with an IP7402_SequenceNumber value of 1 and the last 12 would appear on the record with an IP7402_SequenceNumber value of 2.
Q. How is an outpatient encounter defined?
A. The volume count provided with the released files represents a count of distinct IDNs (unique encounter identifier) in the MHDO Hospital Outpatient Datasets. Typically, outpatient records are distinct healthcare encounters in the hospital outpatient setting: a single appointment or visit with the hospital or affiliated clinic where all the services received occur on the same date. The vast majority of outpatient records fall in this group. There are a subset of the outpatient records that cover multiple outpatient services performed on different dates with the same facility, for the same person. These represent a small share (2-4%) of all hospital outpatient records which is consistent across calendar years, though it may fluctuate across facilities; in general, it represents a small volume of IDNs. Some of the types of services that are part of these outpatient encounters are: labs and diagnostic procedures that are performed on separate dates for the same encounter, multiple physical therapy or behavioral health appointments that fold under a single rehabilitation / therapy service, other similar diagnostic or therapeutic services.
Q. What do the values in OP2007 PointofOrigin represent?
A. The source agency for the standard Point of Origin code list is the National Uniform Billing Committee (NUBC http://www.nubc.org/subscriber/index.dhtml). NUBC provides documentation on code lists and regular updates via paid licenses. While the MHDO Data Warehouse Contractor holds a license for the Point of Origin and other NUBC-issues code lists for uses within the Data Warehouse, redistribution is not permitted. As a result, these and other similar code lists and value descriptions cannot be made available to data users through the MHDO Data Dictionaries or through direct communication.
For quick reference, there may be external entities that make available Point of Origin code lists that are closely comparable or identical to the current standard list and can be found through internet searches. Additionally, we would recommend searching within the documentation published by CMS for public information around current standard lists, including guidance around code uses and their updates. For example, users can refer to the notifications published on the CMS’ Medicare Learning Network pages: https://www.cms.gov/Outreach-and-Education/Medicare-Learning-Network-MLN/MLNGenInfo/Index.html. On this website there are some Point of Origin updates from 2010 and 2014 (links below). As of May 2019, the CMS Point of Origin notification from 2014 listed below appears to be the most recent notification related to the Point of Origin code list.
Q. What is the easiest way to distinguish between facility records and professional records in the hospital outpatient data file?
A. Facility or institutional claims (UB-04) are identified by the population of any Revenue Center Code fields (OP6104, OP6111, OP6118) and the population of Type of Bill (OP4004). Professional claims (CMS-1500) are identified when none of the Revenue Center Code fields (OP6104, OP6111, OP6118) fields are populated and the any of the HCPCS Code fields (OP6105, OP6112, OP6119) are populated.
Q. We only use UB-04 data for acute care hospitals and specialty/rehab. Type of Bill (OP4004) being populated indicates that the record is from UB-04. Almost all the data we received does have Type of Bill populated, but it is null for a portion of records, primarily from one facility (this facility is a Clinic, and it does have other records that do have Type of Bill populated). The dataset does not currently have the Place of Service (OP4006) variable. Do you have a way to identify if the null Type of Bill records are submitted in UB-04 or CMS-1500?
A. Since the facility is a clinic, we would assume it would have come in on a CMS-1500,
and that would be why Type of Bill isn’t populated. Without accounting for Type
of Bill and Place of Service, it is difficult to be certain, that is part of why
Place of Service was added as a field and will be available as part of the 2018
outpatient data release.
The fields that are most telling and could be used to get an approximation of determining
when a field is CMS-1500 vs UB04 for Outpatient would be the following:
Note that the MHDO Hospital data submission do allow providers to populate most data fields if they have the data populated in their system, even if it would not typically be included in the claim form they use, so Start of Care Date being populated does not definitively indicate that an encounter was UB04 whereas a blank does indicate either CMS 1500 or that there was an anomaly in a UB04 encounter’s data (rare, but may have a few instances).
Q. Should I see a procedure code on all Inpatient admissions?
A. No, the service lines where procedures codes are missing are what we would expect to see – for example medical sub-service lines (medical cardiology, medical orthopedics and spine, etc.).
Q. Does the Hospital data include uninsured patients’ diagnoses? In other words, would the uninsured patients’ data be the same as all other patients’ data except that there would be no payment information?
A. Yes, services rendered to the uninsured are reported in the Hospital Inpatient and Outpatient data. Patients with no insurance appear with “self-pay” indicated as the primary “payer.” Note: The hospital encounter data releases exclude all financial data.
Q. Why do some hospitals have high rates of self-pay for pay2 and pay3 fields?
A. There are two hospitals that have confirmed that their system defaults to self-pay in the payer fields after all validated insurance carriers are included
Q. How do you identify Inpatient Rehabilitation Discharges since the conversion to ICD-10?
A. Since the switch to ICD-10 from ICD-9 there is no longer a way to identify rehabilitation discharges in the MHDO hospital inpatient data. Before the transition to ICD 10, hospital data coders were instructed to use a “V” code for any inpatient rehabilitation patient discharge which made the identification that a patient was an inpatient rehabilitation patient. With the introduction of ICD-10s, this concept was not carried forward as the National Center for Health Statics (NCHS) feels that rehabilitation is a procedure, not a diagnosis.
Beginning in the 4th Quarter of 2015 and all of 2016 data, inpatient rehabilitation encounters are not all being grouped into DRGs 945 and 946, instead they are being grouped into other DRGs. NCHS has suggested that looking at revenue codes is the way to identify rehabilitation cases. Below is a list of revenue codes that may be useful for the identification of rehabilitation discharges. MHDO is working with data users and coding experts to determine if there are other ways to code these discharges in the existing data or if we would need additional information submitted from the hospitals.
Q. Is there a way to calculate how many procedures were in an encounter?
A. It sounds like you are looking to compute a measure of utilization. The response varies depending on what particular characteristic of the outpatient procedures you are looking to analyze. Here are some scenarios:
Starting with the 2017 Hospital Outpatient release, MHDO makes the Hospital Outpatient procedure detail (OP61) in a new, restructured format, which may be best to use in any of the scenarios mentioned above. From the respective release notes: “This release contains a new file: OP61_Narrow. This file contains the same detail information as the existing OP61 file. However, it has been restructured. Rather than each row having three sets of fields documenting revenue codes, HCPCS codes, etc., the OP61_Narrow format only has a single set of fields. So, if an encounter has 2 records with all fields populated on the OP61 table with OP6102_SequenceNumbers 1-2, these would appear as 6 records on OP61_Narrow with OP6102_SequenceNumbers 1-6. The MHDO will continue to supply the OP61 information in both formats for reasons of backwards compatibility.”
Q. Why is there only one operating provider attached to an encounter with multiple procedures?
A. The data submission rule, Chapter 241, only gives the hospital the option to provide one operating or attending provider in the layout (record 80). Unfortunately, we cannot tease out which particular procedure the attending or operating was associated with.