Data Quality Initiatives



The Necessity for a Standards-Based System to Enhance Data Quality

In an effort to improve data quality, SPARCS is committed to implementing national standards. One way data quality can be realized is through administrative simplification. This can only be achieved when duplication and data redundancies are eliminated at the source, which are the provider information systems.

Background

When the SPARCS system was implemented in the late 1970s, patient discharge data was submitted using proprietary formats. In 1992 New York State did not request an extension of the waiver to use state billing formats because Medicare indicated this extension would not be granted. Therefore New York State providers started billing their Medicare claims using the national standard. And rather than maintaining two separate systems, proprietary formats were eliminated. Two reporting formats would cause inconsistencies in the quality of hospital reporting. Therefore, in the interest of data quality, the process was begun to align the SPARCS system with the national billing standard.

Migration to State and National Standards

The Universal Data Set for Institutional Providers (UDS/IP) Task Force was established by the Department of Health in collaboration with the Hospital Association of New York State (HANYS). Participants include payers, provider groups, and vendors.

The initial mission of the UDS/IP Task Force was to identify UB-92 migration issues and to establish a process for change. A process was established for standardizing data needs in New York State for all payers and government entities including public health information systems. The task force exerted political pressure for payers and providers to adhere to the standard state extensions. Payers now had a forum to get data needs satisfied in a uniform way using a consensus process. Providers had an opportunity to educate payers on the data collection implications and thus had a voice in the process.

The establishment of state standards was an important first step in migrating to national standards. Ground rules for establishing these standards were:

  • One Field - One Purpose - One Use
  • Processors agreement to ignore data irrelevant for their own purposes rather than reject the claim
  • Identification of state fields for New York needs when no national field could be found.

The task force identified national groups that were key participants in the national standardization process and focused its energy toward learning how these organizations functioned. This was supported by Department of Health management. Information gained at these national meetings was beneficial, providing valuable knowledge to be brought back to the task force.

Because of the national perspective gained, once the HIPAA legislation was enacted, the UDS/IP Task Force was in a leadership position to disseminate information throughout New York State.

The UDS/IP Task Force helped to identify SPARCS data elements that were no longer useful. These elements have been eliminated and other elements have been refined. This has contributed to SPARCS data quality.

Conclusion

If the systems that collect information for public health research are isolated from mainstream data collection in provider systems, then national administrative simplification will be difficult. The burden of non-standardization will fall on providers, and data quality will most likely suffer. The enhancement of data quality occurs when separate entities agree to unify similar needs. SPARCS is committed to a standards-based system using administrative simplification as a way of achieving this goal.

Enhancing Quality Through Communications

SPARCS prides itself on working cooperatively with the health care industry nationwide. To do so, we communicate through electronic technology, personal contacts, and mailings.

Electronic Communications

We communicate SPARCS information through our Internet Web site www.health.state.ny.us/statistics/sparcs/

  • The most current specifications about SPARCS are available here to assist facilities in submitting their data as well as to allow others nationwide to learn about our system.

  • SPARCS Bulletins are produced on an 'as-needed' basis, posted on our Web site, and sent e-mail to those who have HPN Data Upload Access and others who request to be included in our e-mail distribution list. These publications outline the newest specification changes and other useful information.

  • Audit and Submission History Reports are updated monthly to assist facilities in determining the status of their data submissions.

  • Annual Reports are produced annually summarizing SPARCS data into 19 standardized tables.

  • Data Request Forms are available for facilities to request their Hospital's Own Data (HOD).

The Health Provider Network (HPN) provides a secure method of transferring data:

  • Facilities can submit their data electronically to SPARCS, using the Data File Upload found on the SPARCS Web site under Health Provider Network (HPN).

  • Copies of previous edit reports and error records can be sent to facilities.

Personal Contacts

  • Site visits: SPARCS administrative staff make field visits to facilities to provide support and answer any submission questions. In addition, these visits provide the opportunity to determine the relevance of or problems with the submission of current data elements.

  • Electronic tracking system: This system enables SPARCS staff to coordinate information about calls received from and made to facilities.

  • Meetings: Participation at national and state meetings provides an opportunity to share concerns and information. It also enables state needs to be discussed and negotiated. SPARCS staff have addressed both state and national associations at various meetings across the country.

Mailings

Annual Reports

In the fall of each year, the previous year's data is summarized into statistical tables, annotated, and published as part of the SPARCS Annual Report. These reports are mailed to the Chief Executive Officer at each hospital, libraries, Department staff, and other interested parties. These reports are also available on our Internet site at:
www.health.state.ny.us/statistics/sparcs/annual.htm.

Reference Lists

  1. The SPARCS Hospital Profile lists all hospital and collector contacts who are involved with data submissions.

  2. The Department of Health's secured Web site, the Health Provider Network (HPN), lists individuals who are authorized to send and receive data via the Internet.

Procedures to Support Data Quantity

SPARCS regulations establish a specific time frame for all Article 28 hospitals to meet in submitting acute care inpatient discharges and hospital-based and free-standing ambulatory surgery data.

Monthly Procedures

Following each hospital's data submission, an edit report is generated that identifies any format or content errors that occurred with the data during processing. This is described in more detail in the Operations Guide.

  • A format error in any particular record will cause the whole submission to be rejected.

  • A content error will fail only that particular record. A failed submission or failed record is treated as if the record was never submitted.

  • The edit report provides information the facility needs in order to make any necessary corrections in their data submission and return the corrected data to SPARCS. The corrected data is then resubmitted and is included in the facility's cumulative data base.

An Audit and Submission History Report is generated following each monthly update. This is described in more detail in the Operations Guide. This report shows every facility's cumulative monthly data submissions for each year. This report is posted each month on the SPARCS Web site.

  • Every 30 days an automated delinquent letter is mailed to each facility with a 120-day lag in data submissions.

  • The delinquency letter is discontinued for the previous year after the July update. A new delinquency letter cycle for each new year begins with the August update.

Annual Activities

The annual review of submissions is based upon the assumption that each facility will have approximately the same number of submissions each year or will provide an explanation for any discrepancy. The following process will determine whether the shortfall is attributable to reductions in patient census or data submission shortfalls.

Since the December data submission records are due by the end of March, SPARCS staff begin to contact facilities in April that have a 50% variance between the previous data year and the year just completed.

  1. The purpose of this preliminary review is to identify and resolve any large scale systemic submission problems.

  2. It is important to note that only 95% of the data is due at this time.

SPARCS regulations require that 100% of the data is due 180 days following the end of the year. In July SPARCS staff again contact each facility that continues to show a 50% decline in submissions between the current year and the previous year.

The purpose of this contact is to identify the extent of the shortfall (if any) and to establish a timetable for completing the data submissions.

  1. Facilities are informed that their complete data is crucial for SPARCS Annual Reports and other health care researchers.

  2. Facilities are also advised that in cases where there are data shortfalls, these shortfalls are identified together with the reasons and are published in the Annual Report so that the data is not misinterpreted.

  3. If data is not submitted according to regulation, facilities are subject a reimbursement rate penalty.

Reports

Customized tables and analytical reports, data submission timetables, and procedures may be provided to assist a facility in completing its data requirements. These are described in more detail in the Operations Guide. The reports below can be used to complement the existing file maintenance reports and data monitoring tools available to hospitals:

  • Edit Report

  • Audit Report

  • Exception File Reports

  • History Reports

  • Hospitals Own Data Tape

  • Customized Reports

  • Log File Summary Reports

The following is available to assist facilities in completing data submission and correction requirements.

  • SPARCS5 - DOS-based software used for submitting and correcting data

  • Download data to a diskette and mail it to the Department for processing

  • Health Provider Network (HPN) - Data files created by SPARCS software can be transmitted to the Department using the SPARCS Data File Upload application on the HPN. The HPN data file upload process provides an efficient and secure data transmission option by using Internet Secured Socket Level (SSL) encryption technology.

Efforts to Enhance Data Quality

SPARCS is highly sensitive to the need for reliable data. Consequently, questions about data quality are looked upon as positive steps to improve the data. The more data is used and scrutinized, the better it will become.

Data Quality Analysis

SPARCS data is reviewed as a result of questions from data users:

  • Groups within the Department of Health have historically been the heaviest users of SPARCS data. The Bureau of Biometrics, Hospital Rate Setting, Quality Assurance, and others have been instrumental in enhancing data quality by exposing irregular data reporting patterns.

  • Increased data use by individual hospitals, independent contractors, and private vendors have provided increased review of SPARCS data.

SPARCS data is proactively reviewed by the staff. SPARCS verifies the data with other data files that share a common event. Related data fields from two data sources are compared for consistency. For example, SPARCS data was matched and compared with:

  • the Vital Statistics Birth Registry

  • the Vital Statistics Death Registry

  • a file of newborn intensive care unit data for a few hospitals

Verification of SPARCS data is accomplished by examining irregular patterns over time. A code can be valid for an individual record. However, that code is not necessarily valid for all of the data records for a hospital. For example, a disposition code indicating the patient was discharged to home is acceptable. But if all patient records from a hospital indicate a discharge to home, that would warrant examination of that data item more closely.

Process to Resolve the Data Problem

The problem is first verified to make sure that the data user understands the data item and to confirm that the data used is current. The extent and cause of the problem is identified. This is achieved by asking questions such as:

  • Does it happen for one or more facilities?

  • Does it happen across discharge years?

  • Did it happen for just one submission?

  • Can it be determined when the problem first started?

Remedial action is taken in order to:

  • Correct the data, if possible. Data correction is done jointly with the hospital. If it is not possible to correct the data, a summary report of the scope of the problem is produced and published on QADATA-L, a department data quality Listserv maintained by the Bureau of Biometrics.

  • Prevent recurrence of the problem. This is accomplished by contacting the facilities involved, generating specialized reports for them, and helping them to be aware of and understand the problem.

The mission of the SPARCS Unit is to assist facilities in achieving a data profile that comprehensively reflects the health care services provided at each hospital. To continue our goal of enhancing SPARCS data, we encourage all comments and suggestions.