Subscription is FREE for qualified healthcare professionals in the US.
6 Articles in Volume 1, Issue #4
Breaking Down the Barriers of Pain: Part 4
Facing Reimbursement Challenges
Getting Back on Track
Taking the Hurt Out of Pain
The Mind-body Connection
The Neural Plasticity Model of Fibromyalgia Theory, Assessment, and Treatment: Part 2

Facing Reimbursement Challenges

Delivering solid, validated evidence of treatment can be achieved by utilizing an outcomes measurement system.

Today’s health care professional, already faced with reimbursement challenges, is under continuing pressure to demonstrate and document the effectiveness of treatment. No discipline is under more pressure than that of chronic pain management. Payors, ever aware of the need to control costs, are rapidly eroding traditional practices and will continue to do so. The emphasis by regulators, public and private accreditation bodies, employers, and patient groups on the improvement of quality of care, along with reduction of associated cost, challenges the practitioner to execute valid models of therapeutic care and to furnish measures of the results of that care. The recent focus on patient satisfaction and health-related quality of life has further increased the demand for documentation of results. The clinician who cannot deliver solid, validated evidence of successful treatment will find it increasingly difficult to survive. Further, those seeking to define treatment protocols must have evidence of efficiency to support clinical applications.

Given the widely differing requirements of an outcomes measurement system suited for general clinical use, the challenge to those embarking on the creation or selection of such a system is to address a variety of audiences and factors with a single research design.

According to Slater,1 outcomes research is any research that attempts to link either structure or process, or both, to the outcomes of medical care at the community, system, institution, or patient level. Slater refers to the framework offered by Donabedian for describing the elements of care that may be considered when assessing quality.2

The concept of structure, the settings in which medical care takes place and the organization of which it is a product, encompasses all of the environment as well as the participants in the delivery of care + human, physical, and financial resources.2 System characteristics include the number and geographical disposition of hospitals and facilities, the availability and training of professionals, and the means of payment for medical services. The scope of examination may be national, regional, institutional, or a combination that includes specific diagnostic, delivery and/or reimbursement groups.

Among provider characteristics that influence the quality of care are availability, specialty training, and preferences, as well as job satisfaction and workload. The consideration of provider characteristics other than availability and discipline is rarely seen in outcomes measurement instruments currently available.

Patient factors to be considered concerning structure include diagnosis or condition, severity, comorbidity, and health knowledge and habits. Patient self-reports and provider’s assessments comprise the elements of this level.

Process, specifically the technical aspect, is the area of measurement that has traditionally received the most attention in the discussion of quality in medical care.

As part of the approval by the FDA for the marketing of both medical devices and pharmaceuticals, clinical trials still are the reason for much of the outcomes research carried on today. Clinical trials traditionally have not considered the interpersonal aspects of the provider/patient relationship, and of course do not seek to emulate normal treatment settings and their accompanying logistical restrictions. Results of clinical trials are a definition of efficacy, whereas the results of clinical processes are an indication of effectiveness.

Process measures are frequently used in evaluating performance. They are generally easier to use, and are more understandable by both clinicians and patients. They are the basis for developing programs of continuous quality improvement (CQI) that seek to enhance care by reducing variation in treatment and cost of care.

Along with procedures and tests, there is an interpersonal aspect to process that affects both the provider’s and the patient’s impressions of quality. While those impressions cannot be said to reflect on the actual quality of care from a medical standpoint, they contribute to therapeutic success in that they impact compliance and the patient’s attention to treatment as well as feelings of satisfaction with care. In the evaluation of the process of care, the interpersonal aspect is generally overlooked, or perhaps more precisely, avoided, primarily due to the difficulty in expressing and measuring the quality of the relationship and the participants’ abilities in interpersonal communication.

The final area of investigation is the obvious one, the outcome of treatment in terms of changes in patient status. The measurement of change implies a baseline assessment followed by at least one post-treatment measure and possibly interim measures as well. Areas of inquiry include clinical endpoints, health-related quality of life, and patient satisfaction with care.


The reasons for adoption of a medical outcomes measurement system may be classified as follows:

  • Regulatory/statutory requirements
  • Accreditation requirements
  • Quality control and improvement
  • Development of protocols/evidence-based practice
  • Reimbursement and plan participation requirements
  • Marketing

The current climate in which medical decisions are made is subject to turbulence in a number of areas, as suggested by the range of justifications for measuring outcomes detailed above. As may be seen, a variety of motivations exists for measuring and documenting medical treatment outcomes. Depending on the primary motivating factor, the weight of emphasis on the elements defined above will vary, but a comprehensive outcomes measurement system will consider most, if not all, of them.

Accrediting bodies for health care facilities impose their own requirements for outcomes measures. The Joint Commission on Accreditation of Healthcare Organizations (JCAHO) instituted its ORXY initiative for incorporating outcomes and performance measures into the accreditation process in 1973. The Commission on Accreditation of Rehabilitation Facilities (CARF) introduced similar requirements at about the same time. Since 1992, the American Academy of Pain Management (AAPM) has required outcomes measures, using the National Pain Data Bank (NPDB), as a part of a comprehensive chronic pain management program.4


The NPDB is an outcome measurement system designed and sponsored by the American Academy of Pain Management (AAPM). It is intended for general use in measuring the outcome of treatment for any chronic pain condition. Information describing the specific condition treated is collected using ICD codes. The specific treatments utilized, using a combination of CPT codes, modalities, disciplines seen, and medications, are collected as well. Outcomes are measured in all of the dimensions commonly found in comprehensive systems. As such, the NPDB is a hybrid of general and dimension-specific systems, with the additional ability to collect information about the specific condition and treatment used, reimbursement mechanisms, and cost. Data is collected at three points in time: immediately prior to the start of treatment (Intake), at the conclusion of treatment (Time 2), and one year following the second administration (Follow up.) The timing of administration therefore allows both proximal and distal evaluation of treatment in the measured dimensions.

Data collection began in 1992 using paper-based electronically scanned submission. In 1997, computer-based administration was added. Participating programs receive a report on a calendar-quarter schedule providing summary information about the results of their programs performance, as well as a comparison to similarly organized programs. To date, approximately 13,000 patient records are included in the database.

For more information on the NPDB visit the Academy’s website at npdb1.html. Or see the article titled, Raising the Bar, by B. Eliot Cole, MD, MPA, in the January/February 2001 issue of Practical Pain Management.

Internal efficiency and validation of effectiveness of treatment are important to providers for financial and professional satisfaction reasons. The recent emphasis on quality management and improvement, while motivated initially by cost-control concerns of payors, has also grown into recognition of the utility of measures for process improvement and its accompanying advantages. Additionally, providers take more responsibility for an outcome measure that is not externally imposed on them, enhancing the process and the quality of information collected.5

The development of practice guidelines and outcomes measures compliments each other in a reciprocal relationship that promises to encourage the proliferation of both in the future. Guidelines serve to provide access to the latest advances in clinical practice, to further cost containment, to decrease unnecessary procedures, reduce medical liability, and provide an informed basis for selecting among treatment options.6 By doing so, the standardization of procedures can enhance the quality of outcomes research, while outcomes measures contribute to providing support for accepting guidelines as improvement to care. Further, the guideline creation process can identify additional areas in need of research.

Probably the most significant factor in the development of outcomes measures has been the increase in the cost of medical care and the response to that increase by purchasers in the form of managed care. The impact has been felt in two different ways + first, the demand for information by third-party payors for the purpose of decision-making and reimbursement, and second in the use of that information to select providers for participation and procedures for inclusion in benefits. Payors have had a “data advantage” in that they have ready access to claims databases which encompass patient populations far larger than the provider normally has access to, but are subject to inaccuracies as far as diagnosis and risk-adjustment are concerned. When faced with an adverse inclusion or reimbursement decision, the provider has access to quality information about patients and the outcome of treatment is in a better position to prevail when requesting reconsideration. Similarly, when responding to a request for a proposal, the provider that is measuring the results of care has an informed basis for evaluating prospects for success, as well as evidencing a commitment to cost-effectiveness.

Sources of Data

Outcomes studies rely on three major sources of data + claims databases, medical records, and surveys. No single source of data is superior to the others, as each has its advantages and disadvantages. As with the level of observation, the source or sources of data will depend on the nature of the research, available financial resources, and the timeline for the study.

Claims databases are often the least expensive and most timely approach to process measurement. Information available through diagnostic and procedure codes includes cost of care, utilization of resources, frequency of specific procedures, etc. The cost of customized data extraction and reporting may be a barrier for smaller organizations without dedicated technical staffing, however.

The medical record can also be a source of valuable information on outcomes. Motheral points out the concern for whether the measure is one of the quality of care or, rather, the quality of the record.7 Keeping this caveat in mind, the medical record can be useful in evaluating care from both a process and outcomes standpoint. It can, however, be a labor-intensive, and therefore costly, approach.

When change in status is measured, there will usually be a need to draw information from providers, patients, or both. The inclusion of the sociological and humanistic factors will require collection from the participants in the process. An important consideration in this area is the burden on the source in the collection of data. Where the design requires use of patient or provider responses, the use of short, feasible, reliable instruments is desirable. When measures are used to compare changes in status over time, adequate planning for follow-up administrations will improve response rates.

Types of Instruments

Outcomes measurement tools fall into three categories + general health measures, disease or condition specific measures, and functional status measures. Generic measures are usually of most interest to policy and decision-makers, whereas condition-specific and functional status measures are of more interest to patients and providers.

General, or generic, measures are best suited to establishing the extent of a condition or disability in a large population or patient group. Areas of measure may include physical, social, and emotional well-being. Because of the general nature of the questions, when the data source is patient self-report, these tools have the advantage of being fairly easy to administer and score, and are well-suited for investigations of the effect of structure. This type of measure is not well suited to the measure of change in specific conditions, however, nor is it especially useful in measuring the effect of specific processes.

For example, the Medical Outcomes Study Short Form 36 (SF-36) developed by John Ware at the Rand Corporation, has an extensive body of published research to document its validity and reliability when used as a generic measure of health status. The 36 items on the instrument gather a patient’s self-report in the areas of limitations in physical, usual, and social activities due to health and emotional problems, as well as bodily pain and current health perceptions. It also is available in a shorter, 12-item version.

Dimension-oriented instruments attempt to assess the status in specific domains, such as mental well-being or functional status. These instruments may be used in a variety of groups to measure a particular status, and with repeated administration to document change in a dimension over time or with intervention. Examples of this type of instrument include the Beck Depression Inventory (BDI), the Behavioral Assessment of Pain (BAP), and the McGill Pain Inventory (MPI).

If the area of concern is a specific condition or treatment, the instrument will require information of a more detailed nature. A search of the literature may reveal that a reliable, validated instrument already exists that will suffice for the measure desired. The use of standardized instruments also allows for comparison with other studies should that be part of the design. Often it is necessary to utilize multiple instruments in a study to achieve coverage of the necessary dimensions.

When deciding whether to construct a new set of outcomes measures using available internally collected data or to use an existing system or instrument, the construction of the tool must be adequately considered. Regardless of the intended user, data collection efforts should employ instruments having demonstrated reliability and validity. This approach will increase the likelihood of dependable results, in terms of information. Internal users will benefit when making quality-improvement decisions, and external users will have increased confidence in the results of measures. If comparability with the results obtained by others is desirable, it is important to be assured that both sets of data are based on reliable and valid results. Standardized instruments are the only vehicles to obtaining comparability, and the researcher needs to have a complete understanding for the basis of the information examined.

One of the problems facing chronic pain treatment providers in measuring outcomes is the lack of such standardized tools8 resulting in difficulty in comparing results among providers and between treatments. Additionally, the range of dimensions often demanded by users can require the use of a number of different instruments to provide adequate coverage.

Barriers to Successful Implementation

A reality of outcomes measures is that the implementation and execution of studies is fraught with peril. Because of the wide array of players and users, and the complexities of data collection and reporting, careful planning is a pre-requisite to a successful effort. Without adequate preparation, a variety of problems pose a threat to the collection of usable information.

Participant Resistance

Resistance to the additional work of measuring outcomes is encountered in all parties involved + providers, staff, and patients. Providers are often reluctant to commit the resources necessary to an activity that is not directly reimbursable. Staff members, likely already working to capacity, may also see outcomes research as a low-priority drain on their time. Patients, already subject to requests for a considerable amount of information in order to provide the necessary history and reimbursement information, will also possibly question the need for more. The presenting condition will possibly be an obstacle in gathering data from some.

There is no easy answer to the problem of resistance. Education of providers and staff, as well as patients as appropriate, about the importance and uses of outcomes research, will help to some extent. The more each participant feels included in the process, the more likely cooperation will ensue.

Providers, especially those in private practice, need to be aware of the potential uses for measures of their own outcomes. Often a fear exists that, in the hands of third parties, it will be cost-reduction rather than quality improvement that will influence decision-making. For the provider with a good understanding of measurement and use of data, the answer may be more and better information about the results of treatment.

Additional information provided by outcomes measures can also enhance the treatment experience for all involved, provided that the results collected are made available in a timely fashion. Providers can benefit from additional patient assessment information for use in diagnosis and treatment planning. Patients can be involved in review of reports while still in the office, and the results may be incorporated into treatment and discharge procedures.

Lack of Coordination with Existing Systems

Much of the data necessary for determination of outcomes, particularly the sociological and humanistic domains, does not exist in information systems found in medical applications today. This is a situation that will gradually improve, as users demand features that address outcomes, and developers of medical information systems become familiar with the requirements. Until such time, ancillary systems such as the NPDB offer automation, but usually at the expense of some duplication of data collection. Existing databases can provide some information, but suffer from the same shortcomings in domain information, the added expense of mining the data, and the limitations on distribution resulting from both confidentiality laws and logistical obstacles.


The state of outcomes research today does not allow for a determination of cost-effectiveness. In fact, it is extremely difficult to use outcomes data to evaluate treatment from a standpoint of cost and benefit; taking the analysis one step further to examine the utility of the measurement process itself is even more difficult, if it is possible at all. There is no precise measure of the benefits of care, such as quality of life, and so no way to place a dollar value on them.

In all practice settings, the financial burden of developing measures, as well as ensuring that they are psychometrically valid and reliable, is considerable. Adding to the upfront cost is the creation of a mechanism for collecting data from the target population, and finally actual collection and analysis.

The cost of ensuring and attesting to psychometric reliability and validity can also be substantial. This cost must be taken into consideration when deciding whether to use existing instruments or to develop new studies internally. If sufficient expertise exists within an organization, the option may be viable. If it does not exist, and outside consultation is required, the organization must weigh the added cost against the cost, if any, of acquiring a ready-made instrument.

The current emphasis on outcomes measures in healthcare will only intensify as the rapid evolution of the structure of the delivery system continues. The interests of all stakeholders involved will benefit from the results. Over the past two decades, the dimensions measured by outcomes studies have expanded, as has the utility of tools to accomplish the task. Today a comprehensive examination of outcomes includes measures of functional status, disability assessment, emotional health, social interaction, general health perception, quality of life, and patient satisfaction. These additions to the traditional objective measures can be quantified in a meaningful way, and instruments are available for gathering data about them.

The major challenge facing providers is to organize, interpret, and report information on the results of treatment, both in terms of cost-effectiveness and efficiency, and to be in the position to compare results with other providers and treatments. This challenge is now permanently in place, and providers should begin preparation to meet it.

Last updated on: May 16, 2011
close X