HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Bailit, H.L. Search for Related Content PubMed PubMed Citation Articles by Bailit, H.L.

Health Services Research

Professor Emeritus and Director of the Health Policy and Primary and Primary Care Research Center, School of Medicine, University of Connecticut, 263 Farmington Avenue, Farmington, CT 06030; bailit{at}nso1.uchc.edu

	Abstract

TOP Abstract Introduction Data Sources Data Barriers Demonstration Project Conclusions References

 
The major barriers to the collection of primary population-based dental services data are: (1) Dentists do not use standard record systems; (2) few dentists use electronic records; and (3) it is costly to abstract paper dental records. The value of secondary data from paid insurance claims is limited, because dentists code only services delivered and not diagnoses, and it is difficult to obtain and merge claims from multiple insurance carriers. In a national demonstration project on the impact of community-based dental education programs on the care provided to underserved populations, we have developed a simplified dental visit encounter system. Senior students and residents from 15 dental schools (approximately 200 to 300 community delivery sites) will use computers or scannable paper forms to collect basic patient demographic and service data on several hundred thousand patient visits. Within the next 10 years, more dentists will use electronic records. To be of value to researchers, these data need to be collected according to a standardized record format and to be available regionally from public or private insurers.

KEY WORDS: Informatics • dental delivery • health services research

	Introduction

TOP Abstract Introduction Data Sources Data Barriers Demonstration Project Conclusions References

 
Health services research (HSR) is a relatively new field within biomedical research. The predecessor organization to the current Federal organization that funds HSR, the Agency for Healthcare Research and Quality (AHRQ), was established in the 1970s. This new research agency was not part of the National Institutes of Health (NIH) and was viewed as addressing issues outside of the traditional NIH mission. With a budget of about $300 million annually, HSR remains a small but important part of the medical research community.

According to AHRQ: "HSR examines how people get access to health care, how much care costs, and what happens to patients as a result of this care. The main goals of health services research are to identify the most effective ways to organize, manage, finance, and deliver high-quality care; reduce medical errors; and improve patient safety" (Health Services Research Definition, 2002).

Clearly, the focus of HSR is on the delivery of health services, but the definition is so broad that many NIH-funded projects have HSR components. Within HSR, there are several informally recognized sub-disciplines representing the different health professions, including dentistry. Nationally, some 20 to 30 investigators have a special research interest in the delivery of dental services. HSR is an interest section within the International and American Associations for Dental Research and is part of the Behavioral Sciences & Health Services Research Group. Most dental HSR programs are located at campuses that have schools of dentistry and public health. The American Dental Association (ADA) and the American Dental Education Association (ADEA) also have active HSR groups.

The purpose of this paper is to discuss the challenges researchers face in accessing information on the dental delivery system, since this is the primary interface between dental HSR and informatics. The paper starts with a discussion of the data used in HSR. Then the technical, organizational, and social barriers to accessing these data are examined. The paper ends with a description of a large demonstration project that collects data on the services provided during several hundred thousand patients’ visits from 200 to 300 different delivery sites.

	Data Sources

TOP Abstract Introduction Data Sources Data Barriers Demonstration Project Conclusions References

 
Secondary data collection
Because of the substantial costs associated with collecting primary data on dental services, charges, oral health status, etc., most dental HSR projects use secondary datasets. One important source of data comes from Federal agencies that do health population surveys. These national surveys are done periodically and focus on oral health status obtained from clinical examinations and on utilization, expenditures, perceived oral health status, and patient satisfaction obtained from interviews and supplemented with some clinical record audits to assess data validity (Massey et al., 1989; Harper et al., 1991; Miller, 1997). The Government makes these datasets available to the research community.

Another major source of secondary data is paid dental insurance claims from Medicaid or from employer-sponsored dental insurance plans. Approximately 40 million people, mainly children and single mothers, are covered by Medicaid (US Department of Health and Human Services, 2000). Typically, Medicaid-eligible populations have very low annual utilization rates, ranging from 20 to 40%, depending on the state (Eklund et al., 2003). Some 105 million people have private dental insurance, where employers contribute to the premium (Bailit, 1999). For privately insured populations, annual utilization rates of 50 to 70% are common. For reasons that will be discussed later, it is difficult for researchers to obtain paid claim data on publicly and privately insured populations. For the 50% of Americans who are not covered by public or private insurance plans, there are no secondary data sources available on their use of dental services. Further, data from public or private dental insurance plans cannot be used to generalize to the entire population.

A few state health departments undertake oral examinations and/or surveys of state or local populations. These are seldom done on random samples of the population and are not carried out on a periodic basis. For these reasons, these datasets have limited usefulness and are seldom available to researchers.

The American Dental Association (ADA) conducts surveys of practitioners every two years and special surveys as needed (American Dental Association, 2002a). Working with the American Dental Education Association (ADEA), the ADA also conducts annual surveys on dental school educational programs and finances and the career plans of dental graduates (American Dental Association, 2000b). These surveys are carefully designed and implemented and represent a valuable source of data that are used by internal staff to prepare reports and papers. These datasets are available to researchers under special circumstances.

Many dental schools and a few large group practices have computerized data available on the care provided to patients. These datasets are used mainly by internal management staff but under special circumstances are also available to researchers if patient confidentiality and other Federally regulated privacy policies are adhered to.

The National Association of Dental Plans, the trade organization for (some) dental insurance companies, contracts annually with Interstudy Publications (St. Paul, MN, USA), a survey research firm specializing in managed care, to survey insurance and managed care companies with dental products (National Association of Dental Plans, 1998). The results of this survey are published annually. Interstudy will sell the data files to researchers.

Finally, some data come from dental items in medical surveys carried out by many different government and private organizations. Examples include surveys undertaken by the Bureau of Labor Statistics, Employee Benefit Research Institute, and benefit consulting firms. The results of these surveys are published, but the data files are not available to researchers.

Primary data collection
The collection of oral health status data is very expensive, because of the costs associated with identifying and scheduling people for examinations and with employing dentists and support staff to do clinical examinations. Likewise, the abstraction of dental records is also time-consuming and expensive (see below). The few studies that collect primary oral health status data on large populations are usually funded by Federal agencies and involve the assessment of dental treatments rather than classic health services research.

In contrast, many studies collect primary population data from patient surveys. The development, testing, distribution, and analysis of data obtained from paper or Web-based survey instruments and interviews are less expensive compared with clinical examinations. Thus, a fair amount is known about subjects such as patient and provider satisfaction and the impact of oral diseases on disability and social functioning (Reisine, 1984; Slade et al., 1996; Locker et al., 2000).

	Data Barriers

TOP Abstract Introduction Data Sources Data Barriers Demonstration Project Conclusions References

 
This section describes the technical, organizational, and social barriers to collecting HSR data.

Technical
A major limitation in collecting service data on patients receiving dental care is the lack of a uniform dental record. Thus, practitioners use different record formats, and even those using the same format may not collect the same data on similar types of patients (e.g., age, diagnosis). Dental professional organizations (e.g., ADA, American Academy of Periodontics) have not defined a common dataset that should be available on all patients or recommended a specific clinical record format. Consequently, it is very difficult to obtain a standard set of data on patients from multiple practices.

A related issue is the cost of abstracting information from dental records. Abstracting records is sometimes done, but a great deal of attention needs to be paid to define data elements that are available in most dental records and to train staff to abstract the records reliably. An equally difficult task is selecting a random sample of records from a practice, since the records are stored in many different ways.

Electronic dental records have great promise for increasing access to the information in dental records, but relatively few dentists now use electronic dental records (Schleyer et al., 2003). Further, the electronic records suffer from many of the same problems noted for paper records. There is no uniform dataset on patients, and several companies sell dental record software and use different record formats and technical specifications.

Another barrier is the lack of diagnosis codes. Most dental records and all insurance claims list the treatments provided to patients, but the treatments received cannot be linked to a specific problem or diagnosis. The ADA has developed diagnosis codes and has urged dentists to use them. So far, this effort has had little impact on the behavior of practitioners. Even if the dentists use the codes, the variations among dentists in diagnostic decision-making may limit their research value.

A final problem is the nomenclature used to describe dental diseases and treatments. For example, stating that the patient has periodontal disease means little, since this is a very broad term and does not describe the location, severity, or type of periodontal disease. Further, there is considerable subjectivity in how different dentists use this term. This means that periodontal disease may mean one thing in one practice but something different in another practice. Of course, this problem is not unique to dentistry, but it is a problem for researchers trying to use record data to understand why some patients are receiving certain services.

Organizational
There are also organizational barriers to collecting data for HSR studies. Perhaps the major barrier is the structure of the dental delivery system. Most care (70.3%, 2000) is provided in privately owned solo practices (American Dental Association, 2002a). There are few large group practices or hospital dental programs. This is a barrier because the average practice has relatively few patients, and it is costly to collect data from these small delivery units.

Another serious organizational barrier is the fragmented private insurance industry. At the local market level, few insurers have more than 10% of the market share. As a result, any one dentist will have relatively few patients from any one insurer, making it impossible to obtain a stable estimate of the practice patterns in that dental office. The same problem exists in trying to estimate the services provided to people living in a community. The population will be insured by multiple firms, and perhaps half the people will not have insurance.

One solution to this problem is to combine data from different insurers. This is almost impossible to do, because of the difficulty in gaining access to paid claim files from multiple firms. Equally difficult are the technical problems in combining claim sets from different insurers, because of differences in coding conventions and data organization. This is in spite of the fact that all dentists and private insurers use the uniform dental claim form developed by the ADA and the insurance industry.

Because of the fragmentation of the dental insurance industry, intermediary data processing organizations, called claims clearinghouses, have formed to accept claims for many different insurers. These firms then aggregate the claims by insurer and send them in batches to the carriers. The problem with claims data from clearinghouses is that the claims are not linked to the benefit plan that patients are eligible to receive. As a result, it is impossible to know the reasons for differences in service patterns among patients. Some of the differences are due to different benefit designs and some to differences in dentist and patient treatment preferences. Because of this problem, few researchers use data from claims clearinghouse firms.

Social
Another set of barriers to accessing data on the services provided to patients is gaining permission from dentists to access the data. The great majority of dentists are in solo practice and have had limited, if any, experience in clinical or health services research. Consequently, it takes time and resources to convince dentists and their staffs to cooperate in research efforts and to train them to collect data needed for research. Unlike medicine, where there are many practice-based research networks, few dentists participate in research networks and have experience interfacing with research organizations.

A more recent social barrier is the demands on researchers made by the Health Information Portability and Accountability Act (HIPAA). The implementation of this 1996 legislation requires researchers to have explicit permission from patients to access any patient-related information available in dental offices, hospitals, etc. This becomes a big issue when linking different datasets such as paid claims and census data. In these circumstances, the patient’s name or some other identifier is needed to link datasets.

	Demonstration Project

TOP Abstract Introduction Data Sources Data Barriers Demonstration Project Conclusions References

 
The Robert Wood Johnson Foundation (RWJF), The California Endowment (TCE), and The W.K. Kellogg Foundation have come together to support a large demonstration project, entitled "Pipeline, Profession, and Practice: Community-based Dental Education" (http://dentalpipeline.columbia.edu). The primary goals of the project are:

1. to double the number of underrepresented minority and low-income students enrolled in participating dental schools;
   
2. to have senior students and general and pediatric dentistry residents spend an average of 60 days in patient-centered community clinics and practices treating underserved patients; and
   
3. to provide senior students and residents with appropriate didactic courses and clinical experiences to prepare them for treating disadvantaged patients in community clinics and practices.
   

Based on a national competition, 15 dental schools were selected to participate in the "Pipeline" program. Each school received support for at least four years. The schools started their second year of program implementation in the fall of 2004. The National Program Office, the unit responsible for directing the program, is based at Columbia University. A research group from the School of Public Health, University of California at Los Angeles (UCLA), is responsible for evaluating the project.

A critical issue for the project was obtaining a uniform dataset for senior students and residents providing care in community clinics and practices. The 15 schools are expected to have contracts with at least 20 community sites by the end of the project. This means that services will be delivered in 300 different practice sites.

As expected, there was no uniformity in the data systems used by the different schools or the community clinics and practices. Further, there were differences among schools in student requirements for computers. About half of the schools required students to purchase a laptop computer.

It is important to stress that the community practice sites have to be patient-centered to qualify for the project. This means that the primary mission of the sites is the delivery of dental services to patients. Education cannot interfere with the achievement of patient care objectives. This also means that the information system used to manage and evaluate the "Pipeline" program cannot interfere with efficient patient care operations.

The UCLA evaluators were assigned primary responsibility for developing the new information system. The National Program Office supported this effort and encouraged the grantees to participate in the design and testing of the system. A workgroup consisting of school representatives, the evaluators, and the National Program Office was established. The group met twice and had many conference calls to complete the task within the allotted 12 months.

The group decided to design a simple system based on the collection of a minimal dataset on paper forms or computer. The design for the computer program also included the capacity for schools to add data fields to customize the system for their special needs. The system was built with the Microsoft Access database program and was designed for installation on a laptop or a personal data assistant.

The minimal dataset is shown in the Fig. Basically, the data collected include information on the provider, facility, patient, payment source, the type of visit and services provided, and their state of completion. Because of HIPAA and Internal Review Board requirements, patient and provider names or other identifiers are not available to the evaluators or the National Program Office. The system uses select menus to collect data and does not require any free text entry. This should make it easy for students and residents to complete the form after each patient encounter. The major limitation of this system is its inability to track the care provided to individual patients over time.

View larger version (30K): [in this window] [in a new window]   Fig. — Minimal dataset used in the "Pipeline, Profession, and Practice: Community-based Dental Education" Demonstration Project to collect clinical data.

Schools using paper collection forms are responsible for sending an electronic file of the data to the evaluator, who will then make it available to the National Program Office. For schools that are collecting encounter data on a laptop computer, students and residents will send the clinical data to the schools using a modem from the field sites. The schools will aggregate the data and send it to the evaluator monthly. The development of a Web site for data collection and transmission is under consideration.

The evaluation team will provide schools with normative data on the types and numbers of patients and the services students and residents provide in different community sites (e.g., hospital clinics, community health centers, private practices). In addition, the evaluators will make standard and customized reports available to the schools on the performance of their programs.

	Conclusions

TOP Abstract Introduction Data Sources Data Barriers Demonstration Project Conclusions References

 
Dental health services research is an established research discipline and is becoming more important to health policy-makers as the nation deals with the challenges of reducing income and ethnic disparities in access to care and oral health. Obtaining data on the dental delivery system is a major problem for investigators. There are many barriers to collecting population-level data from multiple delivery sites and patient groups. These barriers are related primarily to the organization and structure of the dental delivery system and will not be easily solved. The developing field of dental informatics has considerable potential for reducing some of these barriers and for highlighting others that need to be addressed by organized dentistry, payers, government regulators, etc.

The "Pipeline" demonstration program has developed a simple system for collecting encounter data on thousands of patients receiving care in hundreds of delivery sites. This is an opportunity for dental health services and informatics researchers to gain invaluable experience on the organization and operation of a large-scale, patient-level data collection system.

	Acknowledgments

 
I appreciate the efforts of my colleagues who have reviewed this paper and made helpful suggestions, including Drs. Kathryn Atchison and Richard Manski. This paper was supported, in part, from a research grant, "Pipeline, Profession, and Practice: Community-based Dental Education", funded by The Robert Wood Johnson Foundation and The California Endowment.

	Footnotes

 
Publication supported by Software of Excellence (Auckland, NZ)

	References

TOP Abstract Introduction Data Sources Data Barriers Demonstration Project Conclusions References

 
American Dental Association (2002a). The 2000 survey of dental practice. Chicago, IL: ADA.

American Dental Association (2002b). 2000/2001 survey of predoctoral dental education. Finances, Vol. 5. Chicago, IL: ADA.

Bailit HL (1999). Dental insurance, managed care and traditional practice. J Am Dent Assoc 130:1721–1727.[Abstract/Free Full Text]

Eklund S, Pittman J, Clark S (2003). Michigan Medicaid’s Healthy Kids Dental Program: an assessment of the first 12 months. J Am Dent Assoc 134:1509–1515.[Abstract/Free Full Text]

Harper T, Berlin M, DiGaetano R, Walsh D, Ingles J (1991). National Medical Expenditure Survey: household survey final methodology report. Deliverable No. 1.163. Rockville, MD: Westat, Inc.

Health Services Research Definition (2002). Agency for Healthcare Research and Quality. http://www.academyhealth.org/hsrproj/definitionofhsr.htm

Locker D, Clarke M, Payne B (2000). Self-perceived oral health status, psychological well-being, and life satisfaction in an older adult population. J Dent Res 79:970–975.[Abstract/Free Full Text]

Massey J, Moore T, Parsons V, Tadros W (1989). Design and estimation for the National Health Interview Survey, 1985–1994. National Center for Health Statistics 2(110). DHHS Publication No. (PHS) 89-1384. Hyattsville, MD: National Center for Health Statistics.

Miller HW (1997). Plan and operation of the National Health and Nutrition Examination Survey, 1971–73. Vital Health Stat. 1 (10a and 10b). Hyattsville, MD: National Center for Health Statistics.

National Association of Dental Plans (1998). Industry profile: dental HMO/PPO. Minneapolis, MN: Interstudy.

Pipeline, Profession, and Practice: Community-based Dental Education. http://dentalpipeline.columbia.edu

Reisine S (1984). Social, psychological and economic impact of oral health conditions, diseases, and treatments. In: Social sciences and dentistry. Vol. II. Cohen L, Bryant P, editors. London, UK: Quintessence Publishing Co.

Schleyer T, Speller H, Battling WC, Corby P (2003). The technologically well-equipped dental office. J Am Dent Assoc 134:30–41.[Abstract/Free Full Text]

Slade G, Spencer A, Locker D, Hunt R, Strauss R, Beck J (1996). Variations in the social impact of oral conditions among older adults in South Australia, Ontario, and North Carolina. J Dent Res 75:1439–1450.[Abstract/Free Full Text]

US Department of Health and Human Services (2000). Oral Health in America: a report of the Surgeon General. Rockville, MD: National Institute of Dental and Craniofacial Research.

This Article Abstract Full Text (PDF) Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Bailit, H.L. Search for Related Content PubMed PubMed Citation Articles by Bailit, H.L.