PHM Glossary: B


A set of critical observations or data used for comparison or a control.


All studies of interventions (such as population health management) have certain measurable outcomes. Studies (see study design) are constructed to assess the value of the intervention by evaluating outcomes.

The outcomes measured in an intervention study must be compared with those measured in a group or situation in which the intervention was not applied. In a pre-post study, the values of these outcomes in the pre-group or situation constitute the baseline values. For example, a chronic care management program for people with heart failure should affect outcomes such as heart failure hospitalization rate, quality of life, and dollars spent.

The values of outcomes are measured in the group or individual under study but also in some reference situations where the intervention didn’t occur (or wasn’t offered). In a study with an external control group, one can measure and compare the value of outcomes for that group; however, in the more common pre-post type of chronic care management study, the comparison group is a group meeting selection criteria but before the intervention was offered (the group may be the intervened-on people before they received the intervention). The values of the outcomes of interest in this pre-group are known as baseline values. In other words, baseline refers to the before values of the outcomes, as opposed to the after (intervention) values.

The term baseline refers to the value of a measurement, while the baseline period is the time frame in which a baseline value is measured.

Examples of types of baseline values include:

  • The last value before exposure to the intervention;
  • The average or median of values during a specific time frame ending with exposure to the intervention; and
  • The trend of the value (as in a time series) during a specific time frame ending with the intervention.

The baseline values may be reported for an individual (e.g., before and after the intervention) or for a group.


Issues relating to fair measurement of baseline and postintervention values are discussed in more detail in the Disease Management Program Evaluation Guide(Care Continuum Alliance, 2004). Briefly, good practice relating to measuring baseline values includes the following:

Clearly defining what is being measured and determining exactly what constitutes a value (e.g., the units of measurement, or a binary value);

If measuring for a group, clearly defining the numerator and denominator of what is being measured (for example, as a percent who met the criterion or in a mean or median numerical value); and

Specifying whether the baseline is measured at a point in time or over a time period.


Group baseline values are dollars per member per month of the baseline. In determining whether a population health management program saved money during 2003, one needs to know the cost of a similarly chosen group in a previous period. In such a pre-post study, it is critical to select an appropriate baseline so that the question “did the population health management program result in savings?” can be answered.

The program decides to use costs in a group selected using the same criteria (e.g., those with heart failure) as the baseline. The numerator is all covered charges for members of the denominator in 2002, and the denominator is the number of member-months for these members with heart failure. In an actual savings measurement, the baseline would probably need to be adjusted for numerous factors that can affect medical expenditures in 2003 (such as change in case mix, financial trend, and benefit design changes.

A diabetes chronic care management program asks all new members to have a low density lipoprotein (LDL) cholesterol test if they have not had one within three months of enrollment. Enrollee Bob Jones asks his doctor for the test, and the result is 150 mg/dL. When Mr. Jones completes the program, his LDL cholesterol is 120 mg/dL. The first value is his baseline value.

The baseline LDL levels of all new enrollees could be averaged and compared with the LDL values for all graduates to determine whether the program had an effect on participants’ LDL cholesterol. In an actual measurement, the change in LDL cholesterol in a larger population of diabetics would have to be compared with that of program enrollees over a similar time frame to ensure that the chronic care management program caused the improvement.

See actuarial adjustments.


Baseline. (n.d.). In Merriam-Webster’s medical dictionary. Retrieved from

Duncan, I. (2004). An actuarial methodology for evaluating disease management outcomes. Unpublished.

Fetterolf, D., Wennberg, D. and DeVries, A. (2004). Estimating return on investment in disease management programs. Disease Management, 7(1), 5-23.

Johns Hopkins/American Healthways. (2003). Standard outcome metrics and evaluation methodology for disease management programs. Disease Management, 6(3), 121-138.

SOA Home Page. Society of Actuaries. Retrieved from

Behavior Modification, Transtheoretical, and the Prochaska Model

Behavior modification is the dynamic process of supporting members in accomplishing tangible beneficial lifestyle changes from a state in which their own behaviors are contributing factors to potentially negative outcomes. Various methods may be applied or be operative to effect this process, but all depend on a group’s or individual’s decision and motivation to respond to supplied information or interventions. Disease management programs usually embrace an informational and empowerment approach to encourage these changes.

However, direct positive or negative reinforcement strategies may also be applied. Joint Commission on Accreditation of Healthcare Organizations defines behavior modification as: “The targeted outcome of an organized patient education program wherein patients successfully integrate the theory and skills necessary to manage their disease(s) or condition(s).”


This model approaches the temporal and motivational dimensions of behavioral modification by defining five distinct stages that reflect the readiness (commitment) and motivation of an individual to effect change.

The stages of change are pre-contemplation [person not contemplating change and may be totally unaware of the need to make the change], contemplation [person is aware and is contemplating making the change, e.g. concerned about the health impacts of smoking], preparation [person is making the preparations needed to make the change, e.g. setting a date to quit, enrolling in a smoking cessation program, speaking with the family physician], action [person is taking action to actively make the lifestyle change a part of her or his life] and maintenance [person has successfully made the change and is able to maintain the changed lifestyle]. [Prochaska et. al., 1994]

A sixth stage, relapse, is occasionally operative and is described as a resumption of old behaviors.

The transtheoretical model of behavior change was the result of Prochaska’s comparative analysis of 18 major theories of psychotherapy and behavior change which applies specific stage-matched interventions that take into consideration the advantages and disadvantages associated with the behavior as well as the self-efficacy of the individual.

Behavior modification techniques and stages of change theory are typically incorporated into the health coach function of a disease management intervention. Health coaches support program participants in identifying behaviors (e.g., smoking, lack of compliance to treatment) that require change to meet their life goals (e.g., better health, improved mobility, reduced risk of illness or death). Health coaches then work with participants to establish a commitment to change through the establishment of change goals.

Throughout this process, the health coach assesses the participants’ readiness to change and tailors their support to enhance their readiness or, if ready, work with the participant directly on the behavior change. In some programs, readiness to change assessment is incorporated into a survey process, serving as a triage mechanism, to focus limited resources on individuals ready for change and theoretically more likely to be affected.


DiClemente, C., et al. (1991). The process of smoking cessation: An analysis of precontemplation, contemplation, and preparation stages of change. Journal of Consulting and Clinical Psychology, 59(2), 295-304.

Levesque, D., et al. (Fall 2001). Assessing Medicare beneficiaries’ readiness to make informed health plan choices. Health Care Financing Review, 23, 87-104.

Parker, M., Pederson, D. and Bergmark, R. E. (Fall/Winter 2003). Healthy lifestyle coaching, motivational interviewing, and stages of change: Outcomes on what does and does not work. ISM-USA Newsletter, 7–13.

Prochaska, J., et al. (January 1994). Stages of change and decisional balance for 12 problem behaviors. Health Psychology, 13(1), 39-46.

Prochaska, J., DiClemente, C., and Norcross, J. (1992). In search of how people change: Applications to addictive behaviors. American Psychologist, 47(9), 1102-1114.

Sarkin, J., et al. (2001). Applying the Transtheoretical Model to regular moderate exercise in an overweight population: Validation of a stage of change measure. Preventive Medicine, 33, 462-469.


An external measure used to compare against the internal value of a measure of interest, often in a normative (or best practices) manner.

See also benchmark studies under study design.

Beneficiary – Medicare or Medicaid

Beneficiary is the designation the Centers for Medicare and Medicaid Services applies to an individual who has health care insurance through the Medicare or Medicaid program. Beneficiary is synonymous with the term member as it is commonly used by private insurers and disease management organization.


Glossary Centers for Medicare and Medicaid Services. Department of Health and Human Services. Retrieved from


Bias is a systematic difference between the outcomes measured in a study of a sample of the entire population study and the true results. Bias leads to an overestimation or an underestimation of the strength of the association between the dependent and independent variables.

It is important that possible sources of bias be understood and identified so that they can be minimized in study design, controlled for in measurement and revealed in publication.


There are two general types of bias.

  1. Selection bias occurs when a sample of patients chosen for study in an intervention is selected by a nonrandom process or is not representative of the entire population for measurement. For example, in population health management measurement, selection bias may affect interpretation of an observed outcome if those who elected to participate are compared with those who did not elect to participate but who had the same condition. The people who volunteer for a program are more motivated to respond to the program’s recommendations and may have more ability to improve than the nonvolunteers, because the participants are more compliant, better self-managed members. Selection bias exists because the volunteers represent a different risk profile than the nonvolunteers.
  2. Observational bias occurs when measurements (observations) are inaccurately performed on the exposure or outcome.
    • Observer bias occurs when an observer consistently over- or underreports a particular variable. For example, a nurse consistently and incorrectly records blood pressure using a sphygnomometer.
    • Recall bias occurs when individuals with a particular adverse health outcome consistently remember and report their previous exposure experience differently than those in the sample with less-adverse outcomes. For example, an individual never hospitalized is likely to remember the date and time of her first hospitalization, even if it was more than 6 months previous. An individual hospitalized six times last year would likely have difficulty remembering the dates of the first hospitalization and might even have a problem remembering the number of admissions.
    • Interviewer bias is any systematic difference in the solicitation of responses or interpretation of information given by the participant. One example would be if one individual routinely accepted Abb levels reported from a participant of about 7 while another consistently did not accept any reported level that was not received from a physician. Differences in the results of the population would be affected based on this bias.
    • Follow-up bias occurs when subjects are lost to follow-up. This problem occurs because individuals who were eligible for chronic care management and who are part of an intervention group terminate coverage or change health plans, while corresponding members of a comparison group do not terminate or change coverage.
    • Misclassification bias occurs when subjects are erroneously categorized. For example, if healthy individuals enroll in a hypertension program and were incorrectly measured as hypertensive, the results of the intervention would likely be biased toward no difference between those hypertensives treated and those who were not.
    • Refusal bias is also known as nonresponse bias and results when patients who meet selection criteria for a disease management program refuse to participate (either actively or passively).
    • Publication bias is the bias of peer-reviewed journals to publish only those papers that report positive or topical results.

Confounding is an important concept related to bias.

A few tools are available with which to estimate the bias and standard error of a statistic of interest. One such technique, “bootstrapping,” is discussed in Jain and Shapiro (see Bibliography). Finally, it is important to note that some biases are not technically measurement biases but impact our ability to clearly understand effectiveness of analyses and magnitude of results. Research motivations or reviewer perspectives, for example, may skew research findings.


Hennekens, C., and Buring, J. (1987). Epidemiology in medicine. Boston: Little, Brown & Company.

Jain, L., and Shapiro, A., eds. (2003). Intelligent and other computational techniques in insurance. Hackensack, NJ: World Scientific.

Mausner, J., and Kramer, S. (1985). Epidemiology, an introductory text (2nd ed.). Philadelphia: Elsevier (1985).

Body mass index (BMI)

BM is defined as the weight in kilograms divided by the square of the height in meters (kg/m2). A BMI of 18.5 to 24.9 kg/2 is defined as normal, BMI of more than 25 kg/m2 as overweight, and BMI of more than 30 kg/m2 as obese. These markers provide common benchmarks for assessment, but the risks of disease in all populations can increase progressively from lower BMI levels.


Simple BMI calculations may be misleading. According to the existing definition and calculation of BMI, anyone with a BMI more than 25 would be classified as overweight, whether their body is composed of fat or muscle. Athletes, for example, may be considered to be overweight even though they may have very little visceral fat. BMI is an imperfect indicator of risk of disease. People with the same BMI but different amounts of visceral fat face different risks of disease. Furthermore, weight is only one among many risk factors.

BMI is calculated the same for adults and child but is interpreted differently for children. The Centers for Disease Control and Prevention notes that “for children ages 2-20 years, BMI is plotted on a growth chart specific for age and gender.”

Although some authors use categories such as “moderately overweight” for those with BMI of 25-30, the NHLBI does not designate overweight with such qualifiers. The extreme obesity classification (BMI>40) is a commonly used cut-off for determining qualification for bariatric surgery. However, BMI cut-offs for obesity vary around the world.


Care Continuum Alliance. (2008). Obesity toolkit. Washington, DC: Care Continuum Alliance.

Centers for Disease Control and Prevention. Health Weight – it’s not a diet, it’s a lifestyle! Retrieved from

Division of Nutrition and Physical Activity, National Center for Chronic Disease Prevention and Health Promotion. Retrieved from

World Health Organization. Obesity and Overweight. Retrieved from

Breach notification rules

The requirement to notify individuals of breaches of their protected health information.

Pursuant to section 13402 of the Health Information Technology for Economic and Clinical Health (HITECH) Act, the Department of Health and Human Services (HHS) created rules for breach notifications in the healthcare setting. In addition, HITECH expanded the regulatory and enforcement authority of the Federal Trade Commission (FTC) to include personal health record vendors.  


American Recovery and Reinvestment Act of 2009, § 13407, Pub. L. no. 111-5, 123 Stat. 115 (2009).