Lesson

Transcript

Let's revisit the two different types of data and data sources. Recall that primary data sources contain information specifically entered into the medical record by the healthcare professional providing the care or services; this could be a physician, nurse, technician, or others. It is when this data or information is taken from the primary source, the medical record, and used in registries (cancer, trauma, birth defect, implant, transplant, immunizations, etc.), indexes (master population or master patient index [MPI], diseases, physician, operation, or procedures indexes, etc.), or databases (CDC, National Practitioner Data Bank, Health Integrity and Protection Data Bank [HIPDB], clinical trials, vital statistics, etc.) that it can be considered a secondary data source. In addition, data can be classified as either patient identifiable (which relates to a specific patient), or aggregate data (which is data that cannot be related back to a specific patient and is de-identifiable). In addition, the information contained in the various components of the medical record can be aggregate data.

Aggregate data have been collected, or collated, from various sources and can answer the general questions of who, what, when, where, why, and how care was provided for research or other institutional purposes. Aggregate data can help organizations assess the effectiveness of healthcare delivered through quality and financial reviews, both internally and externally. An example of aggregate data might be statistics on average length of stay (ALOS) for patients who were discharged with a specific DRG (diagnosis-related group). Think about how easy this task is with a database rather than manually examining paper medical records.

Healthcare and HIT professionals must understand where their data come from, both from a data entry (storage) and a database abstraction (retrieval) perspective. Healthcare data populate the many documents used in healthcare today and are used for healthcare assessment. If you are managing data and you have missing or corrupt data, your applications and documents will be corrupt and could impact the quality of care being delivered.

There are six key categories of data that we will discuss. The chart below is not meant for memorization, but rather as a resource. Remember that any professional is only as good as his or her resources.

HIM Data Resource Chart

Key 1: Primary Health Data and Information Media: First, let's define primary healthcare data. These are field items entered into a database that are generated from primary data sources: patient registration information forms, clinical provider documentation and reports, patient financial data, referral data, and so forth. Primary data today is often incorporated with nonpatient specific data, such as clinical guidelines (best practice in medical care), clinical pathways (roadmaps or care plans), and outcome measurements (evaluation of care), as well as patient survey data for quality assurance in patient care and satisfaction monitoring. Primary healthcare—raw data—is a reflection of the individual patient record, but this data is often reported cumulatively for a variety of analytical reasons, including marketing, disease management, patient safety action plans, scientific clinic trials, medical research, and of course, financial reports.

Key 2: Secondary Data Sources: These data sources are most valuable in internal benchmarking and developing external industry standards. For example, the Medicare Provider Analysis and Review (MEDPAR) is one of many claims databases. The MEDPAR file contains 100% of the data reported for all hospital inpatient services. All individual patient identification information has been removed so that the providers can't identify the patients. That said, however, the information may be used to identify all other database field elements reported to CMS, including the hospital where the services were performed.

Key 3: Healthcare Quality Data Sets: These are used to measure health insurance and health system quality of care. They are not mandated for the most part. If the provider or facility is credentialed via a national body, however, they may be mandated. For example, the National Committee on Quality Assurance (NCQA) Healthcare Effectiveness Data and Information Set (HEDIS) provides guidelines for "effectiveness of care measures," which includes such things as adult and child BMI assessment; childhood immunization status; breast, cervical, and colon cancer screenings; and much more. According to NCQA (which credentials payers and providers and is responsible for HEDIS guidelines), 90% of all insurance companies use these guidelines. Most large integrated delivery systems and physicians utilize quality measures in a variety of ways as well, but again, this is particularly true if the provider organization is certified, or if it is in a provider-owned HMO environment.

Key 4: Healthcare Mandated Requirements: Some health-information-mandated requirements are applicable to all providers, such as CMS documentation requirements and Retention and Destruction of Health Information.

Some mandates, however, are by choice. Being a JCAHO- or AAAHC-certified medical facility is a voluntary credential. Therefore, if the facility is certified under one of these national bodies, it must follow the rules and regulations of that credentialing organization. The caveat is that most of the regulations are required by CMS, the respective state, and other payers. It is in the best interests of the facility to have a private national credential in order to cover its bases proactively with CMS and its payers. It is also better to be told by JCAHO that you have a facility deficiency and be able to create a plan for improvement (PI) rather than be told by CMS via your state agency that your facility is deficient, which may be associated with fines, expensive PIs, or facility closure.

Key 5: Healthcare Information Standards: Why do we need so many HIM standards? The standards mandated in HIM today are needed to assist us in developing organization-wide health record documentation guidelines to ensure continuity and quality in patient care and business functions. Providers of care, regardless of type, are required to maintain organizational compliance with regulations and standards as we have seen with medical documentation guidelines and reporting. Most importantly, standards allow us to interpret, communicate, and apply current laws, accreditation, licensure, and certification standards related to health information initiatives at the national, state, local, and facility levels. Essentially, the standards—such as HL-7 and UMLS—allow all providers to be on the same page in regard to health languages, standards, and requirements.

Key 6: Healthcare Taxonomies and Clinical Vocabularies: Every industry has its own taxonomy and vocabulary; healthcare is no different. In healthcare, however, it is critically important to have a common, comprehensive clinical vocabulary. Without Systematized Nomenclature of Medicine–Clinical Terms (SNOMED-CT), two different physicians could stage the same cancer in a patient two different ways—one giving the patient less than a year to live and the other giving him or her five to ten years. This may seem extreme, but standardized medical nomenclature allows us to be accurate and consistent in the language of medicine. The healthcare medical informatics community, providers, medical vendors, and all users of healthcare systems can communicate because we have synchronized vocabularies, national standardized medical data code sets, and electronic data information (EDI) code sets. All of these things help us to maintain HIM processes, policies, and procedures that ensure accuracy and integrity of health data.

Quality in Data

There are seven key characteristics of quality data: (1) relevancy, (2) completeness, (3) accuracy, (4) precision, (5) currency, (6) timeliness, and (7) granularity.

• Data relevancy—collecting meaningful and useful data. It would be a waste of time to collect data that was not useful, but it is done all the time.
• Data completeness—collecting all the necessary data. It would be irrational to collect only part of the data needed, but it happens.
• Data accuracy—collecting accurate data. It would be negligent to enter data that is inaccurate.
• Data precision—collecting the appropriate level of data detail. It will be ineffective if databases are not precise in the level or the detail required.
• Data currency—collecting current, up-to-date data. Data is useless if it is outdated.
• Data timeliness—having access to the data when needed. Why have data if you can't get at it when you need it?
• Data granularity—designing the data fields, so that the data values cannot be divided any further. For example, the "date" is broken down to month, day, and year. Granularity helps with data reporting and more specific data reporting contributes to quality improvement activities.

See Data Quality Management Model Domains and Characteristics for further definitions and explanations of data characteristics.

How does one know if one's organization has quality data? The organization must establish a data quality policy and program that incorporates the five key characteristics of data quality. There are five key steps that teams may follow to ensure that their data quality is optimized. Let's briefly look at each one separately.

Key 1: Data Definitions: First, the data quality requirements must be established. Each field item in the database must be clearly labeled and defined. Does the field "Name" denote patient name, insurer's name, or guarantor's name? Therefore, it is more appropriate to have the labels and definitions be, for example, Name = patient, IName = insured, GName = guarantor, and so forth. Positioning in the databases is also important, as well as defining and documenting which data elements are in which database; for example, is GName in the patient demographic database, or in the financial database?

Key 2: Data Measures and Monitoring: Developing measuring and tracking systems to monitor data quality is important for problem solving database issues. Some systems have audit functions built in so that the DBA (database administrator) and health information management professionals know exactly by whom, what, when, and where data was entered. Most systems today have some type of built-in event monitoring, which is intended to identify crucial events based on clinical, coding, or financial rules as data is entered into the database. If data is entered incorrectly, or if it is not appropriate, error alerts are generated. For example, if a medical clerk uses unnecessary punctuation when registering patients, an alarm goes up, because this may create electronic data interchange problems. If a physician orders a medication that is not compatible with the patient’s current medications, an alert goes up. If a nurse fails to enter key fields in a patient health record, an error report is generated. If a medical coder uses a code that is not deemed medically necessary for a given procedure, an alert goes up. If the insurance clerk fails to bill a group of claims, a report of unbilled claims is generated. All of these help to create a clean, functional, and well-performing database. It also helps with improvements in patient safety, system communications, and the financial bottom line of the organization.

Key 3: Data Analysis: There are several ways to analyze data and the results of your data monitoring. You start by defining what is in your database and then you redefine it and benchmark it against itself. For example, if you see that the registration staff is dropping 30% of the insurance payer information, creating laborious research for the A/R patient services staff, then, as a manager, you may have the DBA, or financial analysts, benchmark the payer information each week until this is no longer a problem. In addition, if you ask the DBA to provide you with a registration user report, he or she may be able to link the missing data directly to the input clerk and thereby identify the root cause of the problem. You may then address the problem by doing a performance review on the clerk(s) who is failing to enter payer information.

Key 4: Data Improvement: To improve data quality, one has to improve data processes. For example, if on analysis, we find that we are unable to enter a financial code for secondary insurance, which causes our A/R (accounts receivable) reports to be skewed from incorrect classifications on patient balances, it may be that there is no field in the data entry registration screen to do so. The way to improve the data, therefore, is to add a field to the screen and do the necessary programming to reclassify accounts with balances due from a secondary insurance. Often this type of improvement is done by the software vendors, but it is coordinated by the HIMS (Health Information Management Services) manager, A/R manager, or the DBA (Database Administrator).

Key 5: Data Controls: So we've monitored, analyzed, and improved our data and database, but how do we keep it accurate? First, we have to implement policies that allow for data controls; for example, implementing an appropriate training system from the point of hire to additional training for every employee who changes jobs. In our policy manual, we need to monitor and correct action for all of the issues we have uncovered in the past and anticipate potential issues and corrective action for problems that could occur in the future. For example, the DBA noticed that data labels changed in the last computer upgrade performed by the vendor. They noted this because they had thoroughly documented all field labels and definitions in their database manuals. To prevent this from happening again, they now can instruct the vendor, providing it with field documentation so that, at the next upgrade, the integrity of the data labels and definitions is not compromised. Are you beginning to see how standards help organizations and HIT professionals to ensure the quality of data used to assess a healthcare organization's clinical and business operations?

HIM professionals are often involved in a number of compliance activities, other than medical record documentation. They may engage themselves in the areas of clinical documentation improvement (CDI), compliance with Meaningful Use requirements, peer review, medical staff appointments and privileges, and overall data reporting to the federal and state governments, as well as accrediting bodies. 

Clinical documentation improvement (CDI) is an activity that focuses on assuring complete and detailed documentation of clinical data for diagnostic and procedural coding, research, patient safety, and quality scorecards. AHIMA CDI Toolkit (2010, 6) identifies the following goals for CDI programs.

• Identify and clarify missing, conflicting, or nonspecific physician documentation related to diagnoses and procedures
• Support accurate diagnostic and procedural coding, DRG assignment, severity of illness, and expected risk of mortality, leading to appropriate reimbursement
• Promote health record completion during the patient's course of care
• Improve communication between physicians and other members of the healthcare team
• Provide education
• Improve documentation to reflect quality and outcome scores
• Improve coder's clinical knowledge

Meaningful Use (MU) is a government program initiated by the Centers for Medicare and Medicaid Services that provides financial incentives to hospitals and providers who use electronic health records in a meaningful way. There are specific requirements that define what is considered "meaningful", such as using computerized physician order entry to enter orders for medications or tests, using e-Prescription functions, providing patients with a summary of their visit at the end of the encounter, and so forth. These requirements need to be measured and met at a certain level, such as 50% or higher. CMS has stages and timelines for the implementation and those details are outlined on the CMS website: http://www.cms.gov/Regulations-and-Guidance/Legislation/EHRIncentivePrograms/Meaningful_Use.html.The ultimate goals of this program are to increase the exchange of health information, the communication with providers and patients, the coordination or care, and the overall quality of care.

Peer review activities are conducted by Peer Review Organizations (PROs), now called Quality Improvement Organizations (QIOs), with the goal of promoting quality, effectiveness, and efficiency of health services for Medicare beneficiaries. At the basis of all these activities are monitoring, evaluation, and communication with providers and beneficiaries. HIM professionals can support this process by collecting and reporting the data necessary, such as cardiovascular data for patients with acute myocardial infarctions and heart failure. HIM professionals can also contribute by analyzing the data, identifying trends and issues, and communicating them to the physicians and the clinical team.

Medical staff appointments and privileges are often referred to as credentialing. Credentialing is done during the initial appointment and reappointment, which are usually done periodically. Some hospitals establish review schedules, such as every two years for each physician on the medical staff. The review includes data collection, organization, and reporting, specifically pertaining to physician's education, background, experience, quality profile, patient volume, mortality rate, average length of stay, incidents and medical errors, and so forth. One important part of of this task is obtaining information from the National Practitioner Data Bank (NPDB).

Data reporting is a very important aspect of the health information department. Reportable data include various hospital inpatient and outpatient clinical and service usage data, cancer data through the cancer registry, births, deaths, and so forth. HIM professionals who work in these areas have the opportunity to collect, analyze, and report data. In addition, they often have to communicate with the receiving party, check on data transmission issues, and handle error reports, which are errors that may occur because of data formatting, system design issues, or simply errors originated from data collection.

The practice exercise below demonstrates a deeper understanding of concepts presented in this week's lecture.