Overall goals for Healthy People (U.S. Department of Health and Human Services, 2010; the Healthy People goals were defined every decade from 2000 to 2010 to 2020; see www. healthypeople.gov) are to help Americans lead healthy and long lives and to reduce health disparities. There are nearly 600 objectives within 39 topic areas to be met by 2020, and many of the objectives in this document relate to behavior-change strategies. For example, for diabetes, many of the goals in Healthy People 2020 refer to lifestyle changes and education to better manage the disease and avoid complications such as cardiovascular disease and death. Goals for Healthy People 2020 include health behavior change such as adherence to chronic illness regimens and preventive behaviors such as screenings to detect risk factors for disease.

The predominant pattern of illness has changed from acute to chronic as science and technology have advanced. With that technology, treatment regimens have become more complex. However, because of changes in managed health care, these complex regimens are implemented with limited or no supervision as the patient and/or family caregivers carry out these prescribed regimens at home. Therefore, practitioners must be concerned with the extent to which patients can implement the treatment plans they design, as well as the evaluation of the patient’s responses.

Patient responsibility for managing chronic conditions has increased, but there is concern about adherence as it relates to medical outcomes and economic costs. For example, an individual who has insulin-dependent diabetes mellitus (IDDM) may have a computerized insulin pump and a blood-testing device. This individual may, at some point, be a candidate for hemodialysis or renal transplantation because of complications. All of these treatment modalities require adherence behaviors to ensure maximal benefit and minimal harm to the patient. According to DiMatteo (2004), the average rate of nonadherence to treatment across all diseases is 24.8%. If this rate was extrapolated to physician visits by individuals with diabetes, as many as 7.6 million visits would result in nonadherence.

The managed care environment has also had an impact on patient burden in chronic illness. Managed care’s influence on health care has been demonstrated by earlier hospital discharges, shortened office visits, and decreasing home health referrals. In addition, recent literature indicates that as many as 46% of healthcare professionals do not prescribe adequate therapy for their patients (Rhiner & von Gunten, 2010; McGlynn et al., 2003). Therefore, patients and family members have had to shoulder more of the responsibility for the treatment regimen, often in isolation. Although disease management programs have been developed by health maintenance organizations (HMOs), few programs have been implemented and critically evaluated to date. Healthcare professionals working within a managed care system often have little time to address the management of chronic illness and adherence to the recommended regimen (McWlliam, 2009; Golin, Smith, & Reif, 2004; Miller, Hill, Kottke, & Ockene, 1997). Other findings suggest that healthcare professionals and agencies can make a difference in medical outcomes by means of integrating multidisciplinary interventions specifically aimed at assisting patients with managing their chronic disease through education, self-management instruction, prevention, and outreach strategies (Barnestine-Fonseca et al., 2011; McWilliam, 2009; Feachem, Sekhri, & White, 2002).

Literally hundreds of studies have examined adherence behavior, but unfortunately that research has not effected significant changes in behavior (Dunbar-Jacob & Schlenk, 2001; McDonald, Garg, & Haynes, 2002). Even in the 1980s, health behavior researchers such as Conrad (1985) asserted that it was reasonable to assume that a patient with chronic illness attempts to selfregulate in order to gain some control over something that is not always controllable. Rosenstock (1988) also noted that healthcare professionals should “encourage people to make informed decisions, but decisions of their own choice” (p. 72). He added that healthcare professionals are not always right, and there is always the potential for untoward side effects from ordered treatment. These assertions have not changed over the years, and, in fact, we have made little progress even in our understanding of adherence behavior. Furthermore, we have not successfully implemented interventions to ensure continuing adherence with chronic regimens.

Chronic illness regimens can be exceedingly complex, and resources to assist individuals with chronic illness are often limited. Therefore, it is important that the healthcare professional understands the variables that affect the ability of the person to adhere to a regimen. To facilitate understanding, this chapter addresses factors that have an impact on adherence behavior. A discussion of the theories and a description of techniques are also presented to provide a context for the behavioral changes that are required in treatment regimens. We will also present information on goals related to Healthy People 2020, as well as evidence-based guidelines and their relationship to adherence behaviors. Finally, interventions to improve adherence are presented, along with case studies to illustrate key strategies.

Greene (2004) addresses the first use of terms related to patients following recommended treatment regimens. He credits medical sociology and further describes the evolution of terms as well as the development of a discipline concerned with health behavior. There were a variety of labels associated with these behaviors, and patients were labeled as “uncooperative, noncompliant, poorly controlled, resistant, devious, incorrigible, irresponsible, and careless” (Greene, 2004, p. 330). Early descriptions of patients having difficulty following treatment regimens reflected societal attitudes related to tuberculosis in patients who were poor and foreign born. Eventually, the terms used to describe behavior included compliance, adherence, and concordance.

In 1974 a group of scientists gathered in Canada to discuss compliance with therapeutic regimens (Greene, 2004). The term compliance was chosen after some careful debate (well described in the Greene essay). Compliance was used as an umbrella term for all behavior related to healthcare recommendations, particularly in light of the shift from acute to chronic disease seen in the last half of the 20th century.

Adherence and nonadherence are generally used as synonyms for the original terms compliance and noncompliance. The term adherence has eventually been adopted instead of the term compliance on the global stage of healthcare delivery. An example of an interesting classic presentation in the meaning and use of these words was presented by Barofsky (1978), who proposed a continuum of self-care with three levels of patient response to healthcare recommendations: compliance, adherence, and therapeutic alliance. In his model, compliance is linked with coercion; adherence, to conformity; and a therapeutic alliance with provider-patient interactions to self-care. Misselbrook (1998) used the term concordance to indicate the partnership between practitioner and patient in achieving health outcomes. Medication adherence has been divided into two main concepts: adherence and persistence. Adherence refers to the intensity of drug use during the duration of therapy, while persistence refers to overall duration (Bosworth, 2010). The literature reveals different schools of thought related to adherence. One school supports the notion that it is impossible to ever have patients completely adhere to medical regimens. A contradictory school of thought suggests that it is possible, through education or other means, to have patients adhere to their regimen requirements. These contrary schools of thought may be dependent on how a health plan was formulated (Dunbar, 1980). If a plan is formulated by a partnership between the patient and the healthcare professional, the possibility of the patient “adhering” to the plan increases. Adherence also implies a biopsychosocial approach as it focuses on actual medication-taking behavior and its measurement, stresses the importance of the relationship between the patient and provider, and addresses the patient’s motivation, health beliefs, and habits (Lehane & McCarthy, 2009). Should the patient be expected to follow a plan created exclusively by the healthcare professional, without input by the patient, then the patient may or may not “comply.” The WHO has adopted adherence as the term of choice and suggests that it is necessary to incorporate the agreement
of the patient with the prescribed treatment plan (McWilliam, 2009; WHO, 2003).

Creer and Levstek (1996) as well as Dunbar-Jacob (1993) questioned the extent to which we “blame the patient” for their adherence behavior. Part of the responsibility, they assert, belongs to healthcare professionals, and there are instances when nonadherence is wise, given the regimen. Trostle (1997) argued that there is too much emphasis placed on the authority physicians have in recommending healthcare regimens. He further asserted that nonadherence is viewed as “nonconformity with medical advice” (p. 116) and suggested that we look broadly at the behaviors that are being engaged in by patients within the context of their illness. He also cautioned that attempts to motivate patients to comply could be considered coercive and manipulative. In any event, healthcare professionals often make decisions about the effectiveness of treatments without knowing whether the patient is actually following the treatment or in agreement with their healthcare professional (McWilliam, 2009; Rand, 2004).

In order to provide efficacious treatment of chronic disease, healthcare professionals face two challenges. First, we must ascertain whether patients are following the regimen, and secondly we must find effective ways of helping patients to overcome barriers in carrying out complex regimens.

The relevance of adherence to the total wellness-illness continuum was first described by Marston, a nurse, in 1970. Marston considered adherence to be self-care behaviors that individuals undertake to promote health, to prevent illness, or to follow recommendations for treatment and rehabilitation in diagnosed illnesses. She is notable in the history of treatment adherence as the first reviewer of literature in the field (Greene, 2004).

It may be helpful, however, to consider adherence as more than self-care behaviors; rather, it is behavior that is often shared, because patients cannot always implement their medical regimens without the participation of others, even though the delineation of responsibilities is not always clear. Sackett and Haynes (1976) outlined three necessary ingredients before labeling a patient as noncompliant: 1) a correct diagnosis must be made, 2) the recommended treatment must be determined to be efficacious, and 3) the patient must be informed and willing. For example, Greenley, Josie, and Drotar (2006) note that there are misunderstandings about the responsibility for asthma treatment regimens in innercity children, and that this misunderstanding often leads to nonadherence. This is especially true when there is a change in the dependence/independence status of the patient, as with the teenager who assumes greater responsibility for management of his or her healthcare regimen or the older adult who now requires more supervision and assistance by family members.

In the classic book Chronic Illness and the Quality of Life, Strauss and colleagues (1984) noted that family members often take on assisting or controlling roles in influencing patients to adhere to medical regimens. Stephens, Rook, Frank, Khan, and Iida (2010) investigated both the negative and positive strategies spouses used to urge patients with type II diabetes to improve dietary adherence. Findings showed that cautioning the patient about the consequences of
eating an inadequate diet was associated with poorer adherence, and encouragement to select healthier food choices was associated with better adherence. A follow-up study of how couples managed chronic disease revealed that coordination and collaboration between the couple were necessary to carry out the work of the medical regimen (Corbin & Strauss, 1985). More recent work in the field of HIV describes the role that family caregivers provide in the complex regimen (Beals, Wight, Aneshensel, Murphy, & Miller-Martinez, 2006). Likewise, family support in adolescents suffering from asthma was positively associated with asthma control and improved quality of life (Rhee, Belyea, & Brasch, 2010). A review by Knafl and Gilliss (2002) concludes that nursing needs to be more involved with family interventions for chronic illness management. Given that shared responsibility exists, it seems reasonable to conclude that adherence-increasing strategies should be directed toward all those involved in the regimen, and that there may be a need for discussing the division of responsibility among family members.

Theoretical frameworks and conceptual models provide direction for healthcare professionals by guiding the assessment and providing structure for the interaction between patient and provider. At this point in time, the emphasis is on translation of theories and models to effective practice interventions. Although an extensive library devoted to adherence behavior exists, few studies support strategies to improve adherence (DiMatteo & Haskard, 2006). Models for understanding individual health behavior can only be useful if they are based on empirical research and can then be used to create effective interventions. This relates to the current mandate for translational research and evidence-based practice. Brief reviews of behavioral models that are currently used are presented here.

Individuals with chronic medical conditions face a variety of stressful life circumstances involving a range of adaptation demands. Individuals with chronic illness must deal with a loss of independence, the threat of disease progression, and the challenge of modifying their behavior to meet the demands of a prescribed regimen. Lifestyle modifications may become necessary and include, but are not limited to, dietary changes, use of medications, and change in physical activity. Adherence to these modifications has substantial implications for treatment success and decreased disease progression.

For the patient with chronic illness, failure to adhere can result in increased disease complications, increased hospitalizations, and greater treatment costs, as well as disruptions in lifestyle, family dynamics, and coping skills. Although ascertaining the true picture of nonadherence in chronic illness is difficult, the consistency with which poor adherence rates are reported indicates that nonadherence is a major problem in health care. Several studies indicate that adherence rates in chronic illness are approximately 50% (Dunbar-Jacob et al., 2000; Haynes, McDonald, Garg, & Montague, 2002; WHO, 2003), with ranges in nonadherence rates estimated to be 20-40% for acute illness, 20-60% for chronic illness, and an incredible 50-80% for preventive regimens (Bosworth, 2010). In the United States, medication adherence to antihypertensive medications is 51% (Graves, 2000). This problem is not limited to the United States. In developing countries, it is estimated that rates of adherence to antihypertensive medications are less than 50% (van der Sande et al., 2000).

Different definitions of adherence have contributed to difficulties in comparing studies of particular disease groups, and make it impossible to generalize to studies highlighting other diseases. Adherence studies are typically disease-specific; that is, the study population is defined by the presence of a specific disease. However, more recent reviews of adherence behaviors in persons with chronic illness indicate that the nature and extent of adherence problems are similar across diseases, across regimens, and across age groups (Vermeire, Hearnshaw, Van Royen, & Denekens, 2001). A review of studies examining medication adherence reported rates as low as 50%, with some differences in rates seen between settings and measurement methods (Dunbar-Jacob et al., 2000).

Medication adherence is one category of research that spans a variety of diseases. Unfortunately there is no gold standard to measure medication adherence, and current evidence suggests the use of several strategies besides disease outcomes to capture treatment adherence to medication (Krapek et al., 2004; Wagner & Rabkin, 2000; Wendel et al., 2001). Furthermore, polypharmacy in chronic disorders adds an additional variable in observing and measuring medication adherence (Vik, Maxwell, & Hogan, 2004). Despite the complex issues contributing to medication adherence, it has been posited that greater health benefits worldwide would be realized with improved adherence to existing treatments than with the development of new medical treatments (Bosworth, 2010).

Electronic monitors have been used to assess medication adherence in many recent studies. Studies of treatment adherence in HIV populations have used both self-report, diaries, and medication event monitors (MEMs). Not specific to HIV populations, electronic monitoring typically provides lower estimates of adherence than self-report data (Wagner & Rabkin, 2000). In a study of individuals with ankylosing spondylitis, only 22% adhered strictly to prescribed medication (de Klerk & van der Linden, 1996). Although adherence rates were not as low among patients with rheumatoid arthritis— consistently below 50% (Elliott, 2008), epilepsy —34% (Cramer, Vachon, Desforges, & Sussman, 1995); major depression—51% upon initiation, 42% upon continuation/follow-up (Akincigil et al., 2007), (Carney, Freedland, Eisen, Rich, & Jaffe, 1995; Demyttenaere, Van Ganse, Gregoire, Gaens, & Mesters, 1998); schizophrenia— less than 55% (Duncan & Rogers, 1998; McCann & Lu, 2009), diabetes mellitus—47% (Mason, Matsayuma, & Jue, 1995); hypertension—30-47% (Mounier-Vehier et al., 1998; Le et al., 1996); tuberculosis—39% (Ailinger, Martyn, Lasus, & Garcia, 2010); ischemic heart disease—38-45% (Carney et al., 1998; Straka, Fish, Benson, & Suh, 1997); and use of inhaled corticosteroids in asthma— 44-72%, with only 13% continuing to fill prescriptions after 1 year (Borrelli, Reikert, Weinstein, & Rathier, 2007). These nonadherence behaviors were nonetheless significantly associated with poor control of symptoms.

Other methods for measuring medication adherence (drug-dosing recall, pill counts, selfreport surveys, and pharmacy refills) have been used and have indicated similar rates of adherence (DiMatteo, 2004; Dunbar-Jacob, Schlenk, & Caruthers, 2002). In a study using subjects with chronic pain, paper diaries and electronic diaries were compared. The electronic diaries offered a time-stamped variation on the paper diary and outperformed the latter with regard to adherence of use by the subject (Stone, Shiffman, Schwartz, Broderick, & Hufford, 2003).

Forty-eight percent of patients with tuberculosis were reported as having defaulted on their recommended medication prescriptions (Pablos-Mendez, Knirsch, Barr, Lerner, & Frieden, 1997). Treatment nonadherence following renal transplantation is associated with loss of the graft. In addition, a history of poor adherence prior to transplantation has also been associated with graft loss (Butler, Roderick, Mullee, Mason, & Peveler, 2004; Marcén &Teruel Briones, 2011). Self-reported medication nonadherence among renal transplant recipients ranged between 13% and 36% (Greenstein & Siegal, 1998; Hilbrands, Hoitsma, & Koene,
1995), whereas heart transplant recipients showed nonadherence rates of up to 37% (Grady et al., 1996). Among liver transplant patients nonadherence to immunosuppressive drugs ranges between 15% and 40% while nonadherence to clinical appointments is in the range of 3-45% (Burra et al., 2011). Although persons with life-threatening disorders may adhere somewhat better than other patients, researchers suggest that even moderate alterations in their treatment have a significant clinical impact (de Geest, Abraham, & Dunbar-Jacob, 1996; Dew et al., 2007).

Several patient characteristics that influence adherence have been examined. These include demographic factors, psychological factors, social support, past health behavior, somatic factors, and health beliefs (Dunbar-Jacob et al., 2002). Ethnicity was addressed in a review by Schlenk and Dunbar-Jacob (1996) and by Joshi (1998), indicating that more research is needed in this area. More recent literature has examined ethnicity as an influence in adherence with diagnostic testing (Cook et al., 2010; Strzelczyk & Dignan, 2002). Strzelczyk and Dignan (2002) reported that African American women were more likely to be nonadherent with mammography screening. Conversely, Cook and colleagues (2010) reported that Latino and African American women were 2 and 1.45 times more likely to receive Pap smear screening, respectively when compared to Caucasians. With respect to retention in clinical trials, African American subjects were more likely to drop out of participation in a rheumatoid arthritis treatment adherence study than Caucasians (Dunbar-Jacob et al., 2004). An interesting study by Taira and colleagues (2007) examined predictors of medication adherence among Asian American subgroups in Hawaii and found that Filipino, Korean, and Hawaiian patients were less likely to adhere than Japanese patients. More research is needed to examine strategies among various ethnic groups to increase adherence behavior.