Article Text
Abstract
Introduction Given healthcare resource constraints, voices are being raised to hold patients responsible for their health choices. In parallel, there is a growing trend towards shared decision-making, aiming to empower patients and give them more control over healthcare decisions. More power and control over decisions is usually taken to mean more responsibility for them. The trend of shared decision-making would therefore seem to strengthen the case for invoking individual responsibility in the healthcare priority setting.
Objective To analyse whether the implementation of shared decision-making would strengthen the argument for invoking individual responsibility in the healthcare priority setting using normative analysis.
Results and conclusions Shared decision-making does not constitute an independent argument in favour of employing individual responsibility since these notions rest on different underlying values. However, if a health system employs shared decision-making, individual responsibility may be used to limit resource implications of accommodating patient preferences outside professional standards and goals. If a healthcare system employs individual responsibility, high level dynamic shared decision-making implying a joint deliberation resulting in a decision where both parties are willing to revise initial standpoints may disarm common objections to the applicability of individual responsibility by virtue of making patients more likely to exercise adequate control of their own actions. However, if communication strategies applied in the shared decision-making are misaligned to the patient's initial capacities, arguments against individual responsibility might, on the other hand, gain strength.
- Autonomy
- Decision-making
- Resource Allocation
Statistics from Altmetric.com
Background and problem
Adam and Eve suffer from type 2 diabetes and they are both recommended to observe a rather strict diet, engage in regular physical exercise and follow dosage guidelines regarding insulin. Eve fails to observe the diet, abstains from exercising and is highly variable in meeting the insulin administration recommendations. As a result of this, she suffers from increased symptoms and ends up with renal failure. Adam, instead, adheres to the recommendations but still (due to unknown factors) suffers similar symptoms and renal failure. Do Adam and Eve have equally strong claims to public healthcare resources? According to one notion in medical ethics they do, because they have similar healthcare needs. According to other ethical notions, they do not, as Eve voluntarily caused her healthcare need and therefore is in some sense responsible for it in a way that Adam is not. If both of them cannot have their needs optimally satisfied, this suggests that Eve should be assigned a lower priority in spite of her equal needs, a view that we will refer to as ‘individual responsibility’ (IR).1–3 Theories affording IR may, of course, be debated as such, quite independently of healthcare issues. In this paper, however, we will assume that some such theory is acceptable in order to address arguments criticising the practical applicability of IR within healthcare under non-ideal circumstances. These arguments focus on (1) the difficulty of establishing whether an individual is in fact responsible for a healthcare need; (2) feared consequences of introducing an IR theme into clinical dialogue and practice; and (3) structural risks or unwanted consequences if IR is implemented as policy. In particular, we explore whether these arguments can be ameliorated by the introduction of shared decision-making (SDM).
SDM has emerged in the last couple of decades as part of a trend towards a more patient- or person-centred care,4 ,5 and means that the patient is given room to voice his or her subjective perspective as a basis for a collaborative approach to healthcare decision-making. The main reasons for SDM are that the room for patient autonomy may then be expanded while retaining opportunities to influence the patient's health,6 ,7 and that SDM may result in better adherence to care decisions when these are adapted to the patient's situation.8
SDM presupposes that the patient either is competent and able to partake in a dialogue, make decisions about care in his or her daily life, or that SDM serves to empower patients to attain such abilities within a trusted relationship. This seems to ring well with advocacy of IR, which usually refers to a similar sort of competences in patients.9 Given this characterisation, SDM seems to have the potential of weakening objections to the applicability of IR to healthcare priority setting. Can this intuition be supported when analysing the relationship between SDM and IR more in detail? Can SDM thereby support the introduction of IR in healthcare? These questions depend on the closer understanding of both IR and SDM, but also on analysing more closely the various ways in which they may relate ethically to one another: as facilitators or undermining factors for practical application, and as exponents of underlying ethical ideals in potential conflict or support. This relationship between IR and SDM is of interest to proponents of IR, but also to those who generally oppose IR, as it may then provide them with reasons to resist SDM. However, if SDM would in fact support the introduction of IR in healthcare, it is also of interest to the proponents of SDM as they might find this aspect more or less welcome.
The paper has the following structure. In the sections on ‘Shared decision-making’ and ‘IR: entangling responsibility and priority setting in healthcare’ we elaborate on SDM and IR, respectively. Based on this, we then argue in the section on ‘Combining SDM and IR: what difference does it make?’ that at least some version of SDM and IR indeed seem to strengthen the case for holding some non-adherent patients responsible through some degree of de-prioritisation in the delivery of healthcare services. Other versions may instead have no or the opposite tendency. In the section on ‘Juxtaposing IR and SDM: deeper tensions’ we argue that it is not the case that SDM provides independent support for IR, but rather that IR and SDM rest on potentially conflicting underlying normative ideas. At the same time, IR may be used to limit some unwelcome side effects of SDM.
Shared decision-making (SDM)
We will not repeat the independent arguments for introducing SDM or analyse them in detail, since this has been explored elsewhere.4 ,8 ,10 ,11 Still, we need a clearer picture of what SDM may amount to. Sandman and Munthe have presented a taxonomy of nine generic models of SDM.10 Several of these do not suggest the shift of responsibility discussed in the present setting. In this paper we will therefore focus on the versions that do. We have referred to these as ‘high level dynamics’ SDM. These models all assume that the patient has shared a narrative of his or her situation with the professional, disclosing immediately medically relevant aspects and also more general aspects regarding his general conditions, experiences, values and goals in life, which may affect how a chosen treatment proceeds.10 The patient and professional then collaborate in a decision-making process that may have three outcomes: joint decision; a decision made by the patient without professional support; or a compromise where the professional strategically accepts something less than ideal (from the perspective of the professional) to achieve consensus.
These versions of SDM all accept that professional standards and goals are open to criticism from the patient, that the patient's preferences and views may be discussed and questioned by the professional, arguments can be proposed and accepted or rejected, and both parties may revise their initial standpoints. This ‘high level dynamics’ implies that the patient is given substantially more involvement and power in the decision-making, thereby possibly acquiring the control underlying the thesis that SDM may promote IR.10 Of course, actual attempts at high level dynamics SDM may come closer or further from this generic ideal type. However, even in imperfect cases, there will be considerably increased room for patients to influence the decision on treatment, either by arguments accepted by the professional or by declaring unwillingness to adhere to what the professional suggests, pressing the latter to a compromise. If the patient insists on too risky or unmotivated treatments, however, the professional might withdraw her support and declare to the patient that she cannot condone such an action.
For concrete illustration of this, let us introduce a third patient, Lily. Lily suffers the same symptoms and renal failure as Adam and Eve. Unlike these two, however, her care has involved what we have called high level dynamics SDM. Her GP has taken the opportunity of more advanced dialogue to empower Lily to better understand and appreciate the treatment and the risks she is exposing herself to, and first attempted to reach a consensus on what she perceives as the ideal treatment but failed. Lily has made it clear that she cannot imagine a life without alcohol and sweets, without a social life implying late night activities and sometimes irregular eating hours, that she has difficulty motivating herself to training and that she is a bit sloppy with her insulin intake. Lily's GP then explained that the risks could be better (though not ideally) managed with a reduced alcohol intake, somewhat less irregular eating hours and late nights, a bit more exercise, somewhat more adequate insulin handling and so on, and as a compromise declared herself willing to accept and support Lily in an attempt to accommodate Lily's personal priorities and accomplish such an (imperfect) improvement. However, at the end of the day, Lily decided not to adhere to what was agreed, likely causing her current renal failure.
Individual responsibility (IR): entangling responsibility and priority setting in healthcare
IR rests on two general claims that entangle responsibility and resource allocation:
If a person is in adequate control of her own actions and their consequences, she is also responsible for these (the control claim).
If a person is responsible for an action and its consequences, she has a weaker claim on others to assist her in the continued performance of these actions or the management of these consequences than if she had not been so responsible (the reduced assistance-claim).
If these conditions are met in a healthcare context, such as in our examples with Eve and Lily, this may be taken as a strong argument for IR itself:IR: Assuming that needs for healthcare are equally large and resources scarce, society should allocate fewer healthcare resources to people who are responsible for their own ill health than to people who are not.
As indicated above, we will simply accept IR as a claim of distributive justice at an ideal level for the sake of the argument. As IR is formulated, we have taken the stand that the reduced assistance claim should be based on the notion of ‘option luck’ rather than a morality of desert.2 ,12 According to this so-called luck egalitarian reasoning, the unfairness is not that Eve and Lily receive undeserved benefits, but that this implies that others with equal (or greater) needs (eg Adam) might have to be unfairly prevented from receiving the same benefits (regardless of whether this concerns the whole or parts of the benefit).i 13 ,14
Even if we assume that IR should be applied in healthcare, there are still a number of arguments about its practical applicability, however, and our focus in the following analysis is on these.1 ,9 These arguments can be grouped under four different headings:
Questioning the applicability of the control claim.
People are not able to control their unhealthy behaviour in the way assumed by the control claim and therefore the reduced assistance claim is not applicable;
and/or
There is no plausible way for healthcare professionals to ascertain that patients have fulfilled the control claim and thus they lack practical means to assess when the reduced assistance claim applies. B. Questioning the therapeutic consequences of IR.
IR threatens the trust between patients and professionals.
Questioning the overall benefit of IR.
Behaviour that poses risks to health often confers other benefits to patients;
and/or
IR makes people generally less ready to tend to their health;
and/or
IR means that people will be left to suffer and die from curable or manageable conditions.
Structural arguments against the social efficiency of IR.
It is too costly to ascertain the required control of patients as well as to monitor to what extent it is in place in single cases;
and/or
IR will in practice express social biases and prejudices regarding which risky behaviours are normal or acceptable and which are not, thereby producing skewed, unfairly discriminatory distributive results.
Combining SDM and IR: what difference does it make?
In this section we will be using the contrasting cases of Eve and Lily to test the force of the arguments against IR one by one to assess if the addition of the high level dynamics SDM in the case of Lily makes any difference. Existing ethical discussion of IR typically assumes cases like Eve as a point of departure: a patient arriving as a pristine wreck with dire needs of care but no previous treatment history or any sort of SDM, albeit a long life of health mismanagement assumed to be readily knowable. In contrast, what we address is cases like Lily where there is a treatment history including high level dynamics SDM and where the health problems have come gradually during this process, due to the fact that the patient does not keep to agreed treatment plans. At this point, when Lily has developed renal failure, her GP might not be willing to deny her dialysis but still may be motivated to decrease the time for future consultations or to increase the time between these, becoming less active in re-offering services of a sort that Lily has proved unlikely to use, and so on.ii This might also be an openly addressed topic in the continuing SDM between Lily and her GP.
The applicability of the control claim
This questioning of IR applies when either people are not able to control their unhealthy behaviour in the relevant sense although they have an idea of how to do this, or when they lack such relevant knowledge. It can be debated what sort of control is implied by the control claim and to what extent people generally do have the appropriate control over their health-related behaviour and its consequences (cf. criminal behaviour).15 ,16 In the present context we will simply assume that the appropriate control is not so demanding that it is unattainable for an adult human being in standard circumstances. We may then grant that, probably, many people will not initially be in much control when being struck by ill health or a health threat.9 However, the control can then be increased through support offered by health professionals. Lily has been given extensive help to understand her health problem and the opportunity to present her narrative about the kind of life she wants to live, discussed this together with her GP and appreciated the risks she is facing. Based on this, Lily and her GP have deliberated about what adjustments to her lifestyle and treatment recommendations could be acceptable and Lily has come to terms with some adjustments as necessary to continue with a valued life. In the continued process she has tested these adjustments, realised that she could follow some while others were more difficult, which led to further adjustments of her treatment plan. Although this plan will result in somewhat worse than the best possible outcome with regard to her health, this is due exactly to her own ability to control the treatment decisions on the basis of a developed understanding of her disease in relation to her own life. If she then chooses not to adhere further to what has been decided, plausibly this non-adherence is also within her control. While the poor health from the joint compromise (compared to an unlikely perfect treatment) is the responsibility of both Lily and her GP, the responsibility for the non-adherence to the compromise rests on Lily (assuming nothing else has occurred that undermines her control).iii
This observation leads over to the second way of questioning the applicability of the control claim, whether it is possible to successfully assess the extent to which the outcome of care is a result of the patient's voluntary actions, taken under appropriate control. Since it may be difficult to distinguish how patient choices and other independent factors have been influencing the healthcare needs, using IR in practice requires the acceptance of a certain degree of imprecision and margin of error. This margin may have to be increased further due to ethical restrictions against otherwise effective means for monitoring patient capacities due to respect for patient privacy. At the same time, since this can be explored in an ongoing treatment relationship based on SDM, this argument is somewhat ameliorated compared to the traditional situation.
The therapeutic consequences of IR
The way in which health professionals and patients communicate and interact may in itself have considerable therapeutic effects. In order to actually apply IR, health professionals will need to actively hold patients responsible for what they do. This is often understood as a call for blaming patients, something that might threaten the ‘therapeutic alliance’ wherein it is often thought that professionals should retain the main responsibility themselves in order not to endanger the trust necessary for a working alliance.9
Accepting this possible downside of applying IR, we note that openly holding patients responsible for their health-related behaviour is an age-old and common ingredient in the traditional therapeutic relationship—the health professional telling the patient about the expected outcome of various health-destructive behaviours, mismanagement of treatments, and so on—suggesting that the trust in a working alliance is assumed to cut both ways.9 Adding high level dynamics SDM with its extended consultation, where the patient's perspective, preferences and decision are acknowledged, taken seriously and adapted to, possibly applying various support and empowerment measures, is likely to provide better grounds for trust-building between patient and professional.
In the case of Lily, the GP clearly shows that she is willing to adapt to Lily's perspective or compromise with her own preferred line of treatment to allow Lily some leeway. This signals that the GP trusts Lily to be able to handle her own life as she sees fit (within the constraint that Lily participates in the SDM). In other words, adding SDM introduces a number of ways to strengthen and preserve a mutual trustful care relationship, which may compensate for and perhaps even counteract the assumed threat to trust by openly applying IR. In addition, assuming that IR is applied by the GP, SDM opens the way for Lily to challenge the GP's responsibility and priority setting judgements, thus introducing an element of control on Lily's part also in this respect.
Questioning the overall benefit of IR
The first argument under this heading is the observation that what explains a patient's unhealthy behaviour may be the fact that this person has other values and aspirations in life than just good health. As we have seen, SDM may serve to strengthen such aspects of patients either not adhering to a treatment or (as in the case of Lily before non-adherence) adapting the treatment to secure adherence but with a worse health outcome. However, this does not seem to undermine the applicability of either the control or the reduced assistance claim or present a forward-looking reason against IR. To the extent that the control claim applies, this argument underlines the fact that a patient may find good reasons to voluntarily care less about his or her health than other things, and this is a reason to give the healthcare needs of this person lower priority. For the professional to insist on another priority would amount to disqualifying this person's view of how to value different aspects of his or her own life. Adding SDM to this picture would seem to strengthen the case for the applicability of the control claim, and therefore also the just made point.
The second argument is that the use of IR may make people more reluctant to tend to their health, since open resentment such as blame may result in a feeling of hopelessness and pointlessness: “If you think I'm such a failure, why should I care or think I can improve?” A recent group of studies of adolescent diabetes (type 1) care provides reason to acknowledge this risk, and that SDM may make it worse if performed according to a standard conception that puts the achievement of rational and autonomous decisions at the focus of attention.17–19 The reason for this outcome is that, if the patient is assumed to possess robust decision-making and responsibility-taking capacities not necessarily in place, the focus on how the patient fails to take care of him/herself stimulates a self-image that undermines the applicability of the control claim, and thus also IR. If so, the SDM process has undermined the patient's already vulnerable capacities to adhere to treatments they may very well prefer. Herlitz and colleagues have sketched a markedly different ‘counselling, self-care, adherence approach’ to SDM, focusing on long-term empowerment to develop everyday capabilities of taking responsibility for care-related behaviour.19 Such strategies focus on gradually helping the patient to develop more robust capacities to manage self-care according to plan, and during this development the control claim may not apply and therefore nor does IR. However, at the end of the day, when the patient's self-image and capacities are more robust, this person might after all prefer to prioritise other things in life than health (cf. Lily). Then the reason against IR does not apply, as it is not IR that has made the patient care less about his or her health, but rather a systematic avoidance of applying IR while empowering, supporting and capacitating the patient.
The last argument under this heading is the proposition that IR may mean that people with serious healthcare needs may be left to die or suffer in spite of the presence of effective treatments. This is true, but it should be observed that it is also true of any priority-setting decision in a situation of scarcity, even if made on strictly needs-based grounds. Untreated mild conditions might suddenly develop into serious and even fatal conditions where it is too late to intervene. In other words, the effect held out here seems primarily to depend on scarce resources. IR may stand out as adding insult to injury in this respect if it is applied to more dire needs—for example, if Lily is not only denied more counselling and support but also dialysis or a kidney transplant. Hence, from this perspective it will be important to consider the severity of the healthcare needs to which IR is applied.
A similar reasoning applies to the use of IR based on the luck egalitarian reasoning indicated in the section ‘IR: entangling responsibility and priority setting in healthcare’. If scarcity is decreased, the reasons for setting the priority of attending to the healthcare needs of a non-adherent patient lower than an adherent patient are weakened. Our conclusion is, therefore, that this argument concerns not IR as such, but what level of resource scarcity should be accepted in the healthcare sector. Whether or not SDM is used does not seem to affect this analysis at all.
However, imagine a health system where the application of SDM is systematic: half of patients end up in outcomes like Lily, the other half end up like Adam (pretty much adhering to the original professional recommendations). As a result, there is more scarcity of healthcare resources due to the behaviour in the first group and, because of this, some in the second group will run higher risks of not receiving optimal treatment when struck by the most serious complications, even if they have not taken part in causing this situation. In this situation IR may therefore apply to the extent that the control claim applies (and we assume the reduced assistance claim to hold) and supports the contention that less attention should be given to the non-adherent group (at least when it comes to less serious needs).
Structural arguments against the social efficiency of IR
One claim under this heading is that it is too costly to ascertain the required control of patients, as well as to monitor to what extent it is in place in single cases. This suggestion partly takes us back to the second version of the argument against the control claim. The validity of this claim seems largely dependent on how much certainty and precision is required for the assessment and monitoring, and compromises regarding this may lead to less of a basis for applying IR unless margins of error are accepted. At the same time, we have also noted that the use of high level dynamics SDM seems to involve both a constant empowerment and capacity promotion of patients and a more extensive monitoring of the output of this through interaction and dialogue. This will result in increased costs. Such increased costs may be seen to boost the reason to apply IR, as that would help to balance the costs and the use of resources where they have better effect. However, it may also be the case that SDM on the whole has positive effects that balance the costs.iv The situation may, of course, also be the opposite, but this is basically an empirical question depending on many factors besides the mentioned cost.v 8 ,20
Another claim under this heading is that implementation of IR as an institutional policy is very likely to express social biases and prejudices regarding which risky behaviors are normal or acceptable and which are not, thereby producing skewed and unfairly discriminatory distributive patterns. Health risks resulting from broadly undertaken activities, or activities viewed as part of a ‘normal’ social life, will have less influence on priority setting than others, albeit the resulting healthcare needs may be the same. We acknowledge this risk as quite real, and it remains even if we assume the patient to achieve a better control and capacity through the addition of high level dynamics SDM. However, compared with the implementation of IR with a traditional approach to clinical decision-making, it may seem that such SDM may weaken the force of this claim, since SDM facilitates openly addressing and questioning implicit attitudes about which health risks are normal or more socially acceptable and which are not.21
Juxtaposing IR and SDM: deeper tensions
If IR is used to prioritise healthcare, combining it with high level dynamics SDM fitted to a patient's capacities may significantly decrease the applicability of a wide collection of objections, mainly practical or pragmatic ones at the clinical as well as overarching policy levels. The specific variant of SDM used may make some difference to what extent the increased responsibility for the outcome of care ascribed to patients amounts to a transfer of responsibility from professionals to patients (as noted in the case of Lily when moving from adhering to the compromise to becoming non-adherent). If high level dynamics SDM misaligns communicative approaches to patients' capacities—for example, by assuming a deliberative capacity of the patient not present and thereby causing confusion and insecurity in the patient—this may instead undermine patients' ability to control care decisions and health outcomes (thus strengthening the case against the applicability of IR). SDM variants other than the high level dynamics ones—that is, not involving a joint deliberation where patients can press for revisions of the goals of the care—do not seem to make any difference to whether or not IR is applicable.
None of this, however, means that (well fitted) high level dynamics SDM provides a direct argument for using IR. On the contrary, we will now argue that, despite the fact that SDM may increase the applicability of IR, ethical ideas underlying SDM are in conflict with IR.
We noted that SDM and IR may seem to share a similar ideal that people should be enabled, capacitated and empowered to decide and take responsibility for their own life and the choices in it. On closer inspection, however, the ethical undercurrents of these respective ideas are less in agreement than may initially appear. First, IR does not rest on the contention, often advanced politically, that a better society is a society populated by citizens to which the control claim applies (ie those who have control over their actions and their consequences). IR (as interpreted here) is silent on the issue of what makes for a better life or a better person or a better citizen or society; it only makes an ethical claim about what follows if the control claim happens to apply. Nor does IR imply a necessary focus on reducing public spending; IR says nothing about this, but only states one principle for distributing resources whatever level of spending there is. High level dynamics SDM, in contrast, does seem to embrace as a central idea the notion of the capacitated and empowered patient as a positive ideal to strive for, although this ideal does not harbour the notion of specific ways for patients to take responsibility based on that. This includes the notion of spending the resources necessary to achieve the desired emancipating result, no matter what this implies for health (although there often seems to be a hope that emancipated patients will also choose the healthier courses of action).
At the same time, different SDM variants express slightly different variations on this basic theme, suggesting different attitudes to the exact scope of the emancipation, from the patient deciding on her own to a compromise within professional standards.
If IR is introduced, this picture is changed by the added conversational theme of the patient receiving less attention from healthcare if choosing courses where options accepted by professionals are ignored. This does not automatically follow from the point of view of SDM and its capacitation and emancipation ideal for patients. The only way to find a reason for IR from this standpoint is to combine SDM with a concern for the overall dynamics of healthcare spending and the effects on all (potential) patients in a healthcare system in view of the scarcity of resources.
Hence, the exact version of SDM will influence to what extent SDM and IR could be combined without limiting SDM. When the patient is in full control over the decision and basically defines her healthcare need, the health professional becomes a deliverer of services to a customer on the behest of the latter.20 If so, use of IR undercuts the very idea of SDM, both as it introduces an ingredient of open threat seemingly at odds with the capacitating and empowering aimvi and as its output seems contrary to SDM regarding the issue of what to spend resources on (as IR speaks in favour of withdrawing resources from some activities which SDM strives to promote).
When, on the other hand, SDM implies a compromise between patient and professional, one of the reasons for wanting to limit the patient's influence is distributive effects due to scarce resources. Introducing IR could here serve to constrain how far a professional strategic compromise to adapt to a patient's special situation and preferences can go before a resulting greater need implies a de-prioritisation of the patient. Such limits for risk taking will facilitate a patient's deliberation about his own risk taking, sometimes possibly providing an incentive for the patient to stay within the accepted risk taking levels.
Conclusions
In this paper we have explored the thesis that the practice of SDM may diminish the force of commonly cited arguments against IR. The crucial findings of this analysis are the following.
It is not the case that SDM (and arguments in favour of SDM) constitute an independent argument in favour of employing IR, but rather that these notions rest on quite different underlying values. However, these theoretical tensions do not seem to constitute a problem for combining SDM and IR in practice. If a health system employs SDM, IR may be used to limit how far resources are spent on accommodating patient preferences outside of professional standards and undermining health. If a healthcare system employs IR, what we have called high level dynamics SDM may disarm common objections to the applicability of IR to healthcare priority setting, but only on certain conditions. High level dynamics SDM may pave the way for IR by virtue of empowering patients and making them more likely to exercise adequate control of their own actions. However, if communication strategies applied in the SDM are misaligned to the patient's initial capacities, the result may be the very opposite. Other types of SDM would not seem to affect the applicability of IR at all.
References
Footnotes
Twitter Follow Christian Munthe at @christianmunthe
Contributors All authors have contributed equally to the manuscript. LS wrote the first draft, which has then been rewritten by EG and CM in turn. All authors take full responsibility for the content of the manuscript.
Funding Swedish Research Council, Swedish Research Council for Health, Working Life and Welfare (2014-40).
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
↵i Luck egalitarianism has generated a wide body of research and interested readers are referred to Knight and Stemplowska or Lippert-Rasmussen for further explorations.13 ,14
↵ii This opens up more difficult decisions concerning how far we can accept a patient's condition to deteriorate. In the luck egalitarian literature this is described as the abandonment objection and different authors relate differently to this. For example, Segall suggests an add-on to the luck egalitarian theory according to which we guarantee the patient sufficient healthcare (regardless of IR) to avoid this.3 This is done for other moral reasons than out of fairness or justice. However, it is important to note that this problem is also present in relation to non-adherent patients in general (ie how far should the healthcare professional go to have the patient accept treatment and will s/he have to accept even more dire consequences for the patient?
↵iii If so, since Lily and her GP have an ongoing treatment relationship, this could be exposed and explored in the continuing high level dynamics SDM and motivate further empowering measures to increase Lily's control in these respects.
↵iv This will depend on how the effects of SDM are evaluated not only in cases where the outcome is suboptimal from a health standpoint but across the entire health system.
↵v One such factor is the extent to which SDM will lead to a ‘fragmentation’ of treatment strategies due to an increased plurality resulting from individual adaptations. As has been held out by Sandman et al and Gustavsson and Sandman, such a plurality may undermine the very basis for assessing the cost-effectiveness of treatments, based on controlled studies of larger series of standardised treatment protocols.8 ,20
↵vi Although sometimes a forceful behavioural ‘incentive’, fear mostly incapacitates and disempowers people.