For patients with poorly controlled asthma, medical treatment guidelines advise clinical intervention and stepping up therapy [48]. But it is also recommended that, before doing so, the possibility of medication nonadherence should be considered, and, if prese nt, addressed. If the patient is determined to be adherent, then evaluation of the patient’s regimen and other factors that may be affecting control is in order. Because a change in regimen may lead to subsequent adherence issues, the use of strategies that lead to better adherence is clearly warranted. Patients with disease that is not well controlled and who are nonadherent to their regimen are clearly the primary target population for an adherence intervention (as illustrated by the lower right quadrant in Fig. 6.1), including attention to the regimen and the patient’s goals and preferences. However, in the same sense that a decision to step up therapy should not be undertaken without considering whether nonadherence to the current regimen may be compromising its effectiveness, an adherence intervention is arguably inappropriate in patients with poorly controlled disease unless it is known, or can be assured by the intervention process itself, that the regimen to which the patient is expected to adhere is consistent with accepted medical guidelines and appropriate for that patient. Attempting to motivate patients to adhere to an inadequate or potentially inadequate treatment regimen, using strategies that are isolated from the patients’ medical care is unlikely to be successful in improving adherence, especially when that care is discontinuous or inadequate in some other respect. Ignoring the possibility of deficiencies in patients’ regimen (as is the case in adherence interventions that do not involve personal contact with the patient) does not make these problems go away and may be one reason why such interventions have not demonstrated much success in terms of clinical improvement [109].

In research to evaluate adherence interventions, the vast majority of studies have not selected patients whose disease was poorly controlled. Instead, studies have used other eligibility criteria, such as patients who were prescribed an ICS or other asthma controller medication within some preceding time period or patients who had asthma of some particular severity level (mild-moderate, for example, in one study, or moderate-severe in another). This is understandable in studies carried out prior to the clarification of the distinction between asthma severity (applicable to treatment-naïve patients) and asthma control (applied to treated patients), and prior to the development of straightforward instruments and procedures to assess control. However, even more recent adherence intervention studies have not always defined their target population clearly or chosen their eligibility criteria appropriately for that target. The implicit assumption may have been that this is unnecessary because patients’ adherence will be poor, as it so often is. However, not focusing on the appropriate target population has resulted, in some studies, in the recruitment of patients with relatively high adherence at enrollment and/or relatively well-controlled disease (FEV₁ ~80 %; low symptom level), and may have decreased the likelihood that any clinical benefits would be detected. While, from a study design perspective, the adequacy of the treatment regimen might be addressed during a run-in period and prior to randomization and delivery of an intervention of interest, we would suggest that it is more appropriate to integrate reexamination of the patient’s regimen in relation to their level of control into the adherence improvement strategy itself.

There is a further potential problem. In carrying out research on interventions that use strategies that place intervention personnel in direct contact with the participants but that do not directly address the patient’s medication regimen, the possibility exists that the interventionists may become aware of problems in a patient’s regimen or medical care, and this can put them in an uncomfortable conflict between their responsibility not to deviate from the research protocol and their responsibility to the individual patient.

The fact that both the shared decision-making and clinician decision-making care management interventions in the study by Wilson et al. [21] were associated with clinically and statistically significant improvement in clinical outcomes relative to usual care may not only reflect the effect of intervention strategies such as patient education, effective communication to the patient of the individual patients’ level of asthma control and the possibility of improvement, written asthma management and action plans, but also the fact that a diagnosis of asthma was confirmed as an eligibility criterion, that the participants had poorly controlled disease, and that the possibility that their current treatment and their adherence to it were inadequate were both addressed [21]. It is quite possible that greater clinical improvement would have been observed in other trials that reported clinical outcomes had those samples consisted of patients with poorly controlled disease and had the nature of the regimen, as well as adherence, been addressed.