Taking the Convention’s relational notion of disability* seriously prompts to focus an intervention’s context in more diligent ways than traditional RCT designs allow for. Disability* resulting from reciprocal effects between both an impairment and a disabling context, the latter is not only considered a disturbance but also to actively contribute en- or disabling a person. A stressful environment, for instance, might simply be seen as a confounding variable when intervening with stenocardia (10)—or it might be perceived to be an active component itself reducing an impaired person’s chances for participation. Both cases highly differ in how they acknowledge contextual influences, the former reducing context to a set of avoidable and, at best, clearly circumscribable variables, the latter rendering it to an active and complex ingredient in its own right.

Thus, research designs in accordance with the Convention’s relational notion of disability* should employ context-sensitive strategies: testing an intervention against various contexts has been advocated to identify its active ingredients (11). Randomizing groups of subjects instead of individuals (12) or evaluating an intervention’s effects in relation to various fields (13) equally serve to systematically investigate the extent to which contexts may enable or constrain change. And, finally, ethnographic methods may be used to identify contextual active ingredients, as further elaborated below.

If context counts as constrain or resource that dis-/enables change, research designs, further, should dedicate more attention to these transformative potentials. This seems to be particularly reasonable in the case of chronic conditions—when cure of impairments can often not be hoped for. Contextual constraints or resources may involve norms, moral judgments, meta-level policies, or regulations. They might be of technical or social nature, potentially residing in an intervention’s material environment or in an organization’s social architecture. To specify them helps to fill the usual “black box” between environmental conditions and the efficacy of an intervention and thereby to “ecologically model” approaches of care (14).

Yet, context is polyvalent. Various expertise and knowledge thus are required to understand a context’s transformative effects. A good example is the interdisciplinary research collaborations that have been stimulated by the after war need for extensive rehabilitation and that have been labeled “Universal Design” later on in history: these collaborations joined various techniques and instruments to develop, evaluate, and promote accessible design mainly for persons with physical disabilities (15). This is one of only few examples for systematic interdisciplinary inquiry in the field of chronic illness. Usually the disciplinary ethos not only in medical sciences is highly compartmentalized and there is a shortage of adequate models for how to initiate and pursue interdisciplinary collaboration (16).

And yet, what determines a disability* most: a person’s impairments or his or her surrounding? This question asks to decide either for a medical or a social model of disability*. In contrast, the CRPD takes a step forward in locating disability* in between, i.e., as an intermediate function of interrelated elements of both individual and contextual conditions. Accordingly, analyzing the effects of an intervention should focus on relations, i.e., the analysis of the complex interplay between both a person and his or her surrounding, without methodically granting privilege to one of each. For instance, when measuring the effects of some assistant technology for persons with diabetes (insulin injection or pump, blood sugar meter, etc.), it might be difficult to divide out the complex and mutual interrelations between person and technology, requiring methods that allow for analyzing their effects together (17).

It is also in this respect that research designs are underdeveloped: relational forms of analyses are rare to inexistent in the field of medicine. Most notably, there is a lack of theories how conceptualize change models relationally. Exceptions are ethnographic methods, and in particular the longstanding method of participant observation that allows for reciprocally analyzing the mutual effects between an intervention and a particular situation (9, 18). Further, practice theories, originating in various social sciences (19, 20), have proven to be useful, also in researching disabilities* (21, 22), for understanding interrelations between both an impaired person and his or her environment.

In the everyday treatment of chronic diseases, complex interventions are frequently used (23): these forms of care produce change in rather non-linear ways; they usually involve multi-composite arrays of various and mostly highly interlinked, active ingredients. To give an example various psychosocial therapies for psychotic disorders can be perceived to be complex interventions: the critical ingredients of assertive community treatments have been analyzed (24), the active components of peer-support delineated (25). In both examples, change models have been developed that point to often highly complex, reciprocal relationships between interventions, context, and outcomes.

Thus, interventions for chronic diseases often impact due to a non-linear dynamics of manifold interacting components. Accordingly, and in reference to the venerable model “web of causation” (26), a web of effects has to be conceptualized: research designs in accordance with the Convention’s relational notion of disability* should attempt to analyze a highly interrelated range of primary, secondary, process-related, intermediary etc outcomes. They should be well equipped to bring forth complexity, instead of generating reduced findings. This, above all, requires systematic process evaluation (27): it has to be clear, how, why, and where an intervention may produce change. This requires the systematic integration of qualitative methods, an argument that will be further elaborated upon in the next section.

As shown by the concepts and results of Personalized Medicine approaches, individuals are (biologically and socially) diverse (28): they may, in response to different contexts, react differently to an intervention, as demonstrated for various cancer treatments (29). Or persons may employ an intervention for various goods, making use of it in personal ways: for instance, Beta blockers are used by patients with hypertension not only to control their blood pressure but also for sedative effects. In the same way, a range of ethnographies illustrate how persons with chronic diseases creatively act upon a particular intervention: engaged in sustained efforts, they calibrate it until it (usually temporarily) fits—not only to the persons’ own prospects but also to the context around them (19).

Thus, an intervention might have various effects: it may be received differently or be used for different purposes. As a result, it might be inappropriate to come up with short term proceedings or interventionist ready-mades. Instead, evaluating an interventions effects involves a primarily investigative approach, exploring the multiple and often highly individual ways in which it fits to a persons condition or has been tailored to his or her situation (30). Merely employing patient related outcome measures does not suffice in this context. Instead, qualitative methods should be integrated as they allow for analyzing the full range of (experienced, potentially variable, and context-dependent) effects that might be overlooked, when focusing only on a fixed set of (quantitatively measurable) outcomes.