Article Text

Download PDFPDF

A capacity-based approach for addressing ancillary care needs: implications for research in resource limited settings
  1. Patricia L Bright1,
  2. Robert M Nelson2
  1. 1Food and Drug Administration, Silver Spring, Maryland, USA
  2. 2Food and Drug Administration, Office of Pediatric Therapeutics, Silver Spring, Maryland, USA
  1. Correspondence to Dr Patricia L Bright, Food and Drug Administration, 10903 New Hampshire Ave, WO51-6272, Silver Spring, MD 20993-0002, USA; patricia.bright{at}fda.hhs.gov

Abstract

A paediatric clinical trial conducted in a developing country is likely to encounter conditions or illnesses in participants unrelated to the study. Since local healthcare resources may be inadequate to meet these needs, research clinicians may face the dilemma of deciding when to provide ancillary care and to what extent. The authors propose a model for identifying ancillary care obligations that draws on assessments of urgency, the capacity of the local healthcare infrastructure and the capacity of the research infrastructure. The model lends itself to a decision tree that can be adapted to the local context and resources so as to provide procedural guidance. This approach can help in planning and establishing organisational policies that govern the provision of ancillary care.

  • Research ethics
  • children
  • research on special populations

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Introduction

Ancillary care obligations are ‘positive obligations to provide care that participants need, but that is required neither to successfully answer the researcher's question nor to avoid or mitigate harm resulting from participation in the research.’1 A clinical trial enrolling infants in a developing country is likely to encounter illnesses unrelated to the study intervention such as gastroenteritis, malaria, sickle cell disease and cardiac anomalies. Some conditions may be incidental: unexpected findings of potential clinical significance discovered through study procedures (eg, an abnormal imaging scan). Others may be endemic in the population with onset during study implementation, such as a new malaria infection. This paper will not address findings of unknown clinical significance or adverse events (AEs) related to the study intervention. Instead, we will focus on conditions unrelated to the study intervention that warrant medical follow-up as judged by a trained research clinician. These include unexpected or reasonably anticipated conditions/illnesses that pre-exist or have onset during study implementation and would likely improve with medical intervention.

Although the focus of our analysis is on children so as to draw on our professional experience, ancillary care concerns are as legitimate for adults as for paediatric participants. We use the term ‘research clinician’ to refer to a clinical member of a research team. The research clinician may or may not serve as a principle investigator on a research protocol. Typically, such a research clinician will be implementing research protocols within a healthcare or research facility.

Background

In the developing country setting, treatment infrastructure may be compromised due to few healthcare providers, lack of medicine, poor diagnostic equipment and so on. Research clinicians may face a dilemma in deciding when to provide ancillary care and to what extent. Although the practice of some clinicians in facility-based research institutions may be to provide ancillary care,2 this has clear cost implications. Chou et al found that in a perinatal HIV prevention study in Uganda, 89% of infant and 95% of maternal AEs were unrelated to the study drug. The cost of evaluating and responding to AEs was driven by the underlying health status of the study population rather than by the research question or the study product.3 Not all investigators and clinical staff have sufficient funding to meet all participant health needs. Furthermore, some advocate referring out care that is beyond that specified in the protocol (or following stabilisation for an emergency).4 Therefore, deciding when and to what extent to provide ancillary care can pose a challenge to research clinicians in resource limited environments.

Conventional ethical views of the research clinicians' obligations pull in opposite directions and serve the paediatric researcher in resource limited settings poorly. One end of the spectrum prohibits medical activity outside the researcher–subject relationship, while the other prioritises care to the point of possibly undermining the research.5 Intermediary models have tried to strike a balance between extending medical care and working within resource constraints.5–9 This paper: (1) summarises prior views on the research clinician's role and their application to ancillary care, (2) proposes a model that lends itself to a decision tree and (3) demonstrates confluence of the model with prior theory.

The fiduciary view

The fiduciary view maintains that research clinicians, as fiduciaries, are obliged to act faithfully and in the best interests of research participants.10 Such a fiduciary relationship would require the clinician to uphold the participant's health as the priority. Applying this view, ancillary care is obligated since the medical needs of the individual are the highest priority. However, implementing such a policy has shortcomings. The research clinician in a developing country may not have the budget or expertise to address all ancillary care needs, particularly in a high risk setting where participants have limited healthcare options. Meeting such needs would draw resources away from the research objectives.1 It may also remove participants from any local healthcare system encouraging dependency on the research facility and resulting in a competing mechanism for healthcare provision.

The non-exploitation approach

In their ethical analysis of placebo-controlled trials and clinical equipoise, Miller and Brody explore the obligations of the physician–investigator. They maintain that the investigator does not have the ethical duty to provide optimal medical care, but instead has a duty to avoid the exploitation of research participants while seeking answers to clinically relevant scientific questions. They note that through implementing a randomised clinical trial, the research clinician has already diverged away from offering individualised care instead offering a randomised treatment where the clinician may not know which individuals receive either treatment.7 ,8 The non-exploitation approach restricts the research clinician's role to the implementation of the protocol and the protection of the study participants; ancillary care would not be obligated. Although it is a simple and scientifically sound approach, ignoring ancillary care needs altogether where there are inadequate alternatives to meet those needs might conflict with a sense of compassion on the part of the research clinician. Also, some participants may not understand why a clinician would refuse to administer needed medical care leading to a loss of trust and undermining the success of the research.

The PE model

The partial entrustment (PE) model, proposed by Richardson and Belsky, is an intermediary between the fiduciary and non-exploitation approaches. An essential distinction from the fiduciary view is that the PE model is transaction based. The authors assert that clinical researchers have limited entrustment responsibilities for the participants' health based on the permissions granted by the participants, their resulting vulnerability to the researcher, and the research clinician's duties of compassion, engagement and gratitude.5 An obligation is generated if the medical need is within the ‘scope’ of the research and shows sufficient ‘strength’ to justify ancillary care. Scope depends on permissions obtained during the consent process required to carry out the protocol generally covering illnesses/conditions related to the disease under study and clinically significant diagnosis identified through the medical information collected for the study.5 To evaluate the extent of obligation, the model evaluates the strength of the duty of care, invoking concepts of the participants' vulnerability, participants' uncompensated risks or burdens, and depth of engagement between the participant and researcher. Richardson further lists these factors as vulnerability, dependence, engagement, gratitude and cost.11

When rapid individual patient care decisions need to be made, however, utilising the concepts of ‘engagement’ and ‘gratitude’ to assess protocol driven ancillary care obligations may be a challenge for the research clinician. The authors state that ‘the depth of the relationship between a researcher and a participant will vary from study to study because different protocols demand interactions of varying intensity, duration and longevity. Researchers have a stronger moral responsibility to engage with a full range of participants' needs when the relationship is deeper.’6 Administering ancillary care in accordance with the engagement level elicited through a protocol may be difficult to operationalise, particularly for a research clinician working across multiple protocols.

Ancillary care could potentially be evaluated as part of a compensation approach to all participants based on protocol requirements if reviewed and approved by an Institutional Review Board. However, even if compensation is based on overall protocol demands, participants are likely to require differing levels of ancillary care within the same protocol. Therefore, a compensation package for research participation that encompassed ancillary care would differ in value across participants, making the degree of compensation unequal and posing a challenge in assessing for fairness.

The contextual approach

An alternative approach is offered by Resnik. He emphasises weighing contextual factors and applies this to different types of research (observational studies, clinical trials and case reports). He advocates that ‘context’ should drive the research clinician's role, citing seven factors for consideration: the research design, benefits and risks, the patient's/subject's reasonable expectations, the patient's/subject's vulnerability, the patient's/subject's motivations, the investigator's ability to benefit the subjects and the investigator's previous relationship with the subject. According to the contextual approach, when the clinical role dominates, the investigator has significant ethical obligations to provide the subject with medical benefits and when the scientific role dominates, such obligations are minimal and ‘there are many variations in between.’9 In his examples, those on one end of the continuum (observational and phase I studies) are more likely to invoke a scientific role for the investigator than those on the other end (phase IV and clinical case reports). However, he emphasises the need to evaluate all contextual factors, noting that a phase I trial of sick participants may lead to different conclusions than a phase I study of healthy participants.

The approach is similar to the PE model in that it advocates for protocol-specific assessments of duties, but introduces variability based on participant differences. However, with an emphasis on ‘weighing’ the seven contextual traits, the approach is heuristic and allows wide room for interpretation. Since the model does not drive an obvious conclusion, it is harder to make the claim of an obligation and, therefore, institutions may understand their obligations differently.

Evaluating the magnitude of obligation incurred across seven factors prior to administering care would be cumbersome for the research clinician and, therefore, does not lend itself readily to bedside implementation. We assert that while such models may help articulate the ethical values at stake, a research clinician needs more pragmatic ancillary care guidance.

Proposal for addressing ancillary care needs

How might an organisation or investigator standardise ancillary care provision and generate an ethically humane policy for their institution? We propose a capacity-based model that tries to avoid the pitfalls of the other approaches.

The capacity-based model

The capacity-based model is comprised of three components: urgency, local capacity and internal research capacity.

Urgency

The more urgent the ancillary care need, the greater the research clinician's obligation to administer or facilitate access to care. The extent of urgency triggers an obligation, but is tempered by capacity (discussed below). The following questions help in assessing urgency: would the condition/illness likely cause serious and/or severe irreversible disability or incapacity without the intervention? Is the condition life threatening? Is the condition/illness causing severe pain? Does the medical need require prompt care?

Local capacity

If adequate quality care is readily accessible outside the research facility, the ancillary care should be referred out (unless the care involves little beyond the cursory examination). However, if there are no sufficient external local resources for care, if access is questionable or if care is needed urgently, the investigator should assess the internal research capacity to provide the care. The investigators will need to know the following: Are there alternative healthcare resources accessible to the participant with the capacity to deliver the needed care? Can local care be obtained within a reasonable time frame? In short, how dependent is the participant on the researcher for the care?

Research team capacity

If there is insufficient local capacity, but the research facility has the capacity, the ancillary care may be provided by the research clinician, contingent on the assessment of urgency. Elective procedures do not warrant an obligation whereas urgent medical care should be provided. When local care is available and there is no internal research capacity for the care and when the need is urgent, research staff may try to stabilise the participant, expedite a referral, and possibly facilitate external treatment by making personal contact or providing transport. However, if there is insufficient local and insufficient internal research capacity, ancillary care cannot responsibly be attempted nor reasonably expected (although the research clinician should investigate alternatives mechanisms, if any, to meet the urgent medical need). Assessing the research team capacity requires answering the following: Does the team have the resources to provide the ancillary care without compromising the study objectives (factoring in the prevalence of the condition/illness, its associated cost and the workforce required for treatment)? Does the clinician have the skills and the tools to provide the care? Does the study have an appropriate setting to deliver safe care?

Capacity-based decision tree

We propose a decision tree utilising the capacity-based model (figure 1). In the study of AE procedure costs from the perinatal HIV clinical trial in Uganda, the majority of infant AEs were unrelated to study product or study procedures and were mild, such as an acute upper respiratory tract infection (URTI); however, some were severe and life threatening,3 such as a malaria infection. To help illustrate the capacity-based decision tree, we will utilise the examples of an enrolled infant with an URTI and an infant with severe malaria infection who are being examined for a study visit by a research clinician who is also trained in the detection and treatment of URTI and malaria.

Figure 1

A decision tree to identify ancillary care obligations using the capacity-based model.

Treatment of a non-urgent mild URTI in an infant would not be obligated according to the capacity-based decision tree. Although not obligated, if the research facility has the capacity to provide the care and the care involved minimal intervention beyond a cursory exam, administering such care would be at the discretion of the organisation with the extent of care outlined in standard operating procedures (SOPs).

For the urgent malaria infection in the infant, if there is no local capacity to render the care within an acceptable time frame and there is capacity within the research facility, provision of such care by the research clinician would be obligated. If there is local capacity to provide the care and no capacity at the research facility, the obligation would be to stabilise the infant and facilitate a prompt referral. If both the local healthcare system and the research facility have the capacity, there would be an obligation to make sure care is rendered to the infant, but whether this involved stabilisation and prompt referral or direct provision of the care would be determined by institutional policy and delineated in the SOPs. If neither the local healthcare system nor the research facility had the capacity, ancillary care would not obligated, but the research clinician would be encouraged to investigate alternative mechanisms (if available) by which the infant could receive the care.

An organisational approach to ancillary care

Triage theory has long grappled with transforming policy governing scarce resources into specific allocation decisions. The approach typically involves a triage officer using an established algorithm or set of criteria to determine treatment priority.12 We suggest that those organisations implementing research protocols also use an algorithm or SOPs for ancillary care decisions for their institutions. This approach would allow for differences in allocation of care based on participant need and available resources, but would not entail different levels of ancillary care obligation for participants experiencing the same need, participants receiving care from different research clinicians within the same research facility or enrolled in different protocols. An organisational policy that delineates ancillary care provision to participants and draws on the capacity-based decision tree would reduce the burden on the research clinician by transferring those decisions to organisational leaders a priori. In addition to enhancing implementation logistics, this approach would help make ancillary care provision more equitable across research clinicians within the organisation. Organisational leaders can adapt the capacity-based decision tree in figure 1 to their local context to translate the decision tree into clear procedural guidance.

Implementing the capacity-based model requires planning. Researcher facility supervisors should investigate the local disease burden and retain some flexibility to respond to unanticipated conditions/illnesses that might arise. The local healthcare capacity should be evaluated: identifying links for referrals and the strengths, weaknesses and quality of accessible care to study participants. In budgeting, researchers should make adequate provision for ancillary care needs, factoring in the use of local capacity when available and adequate. The institution should establish policies and procedures regarding ancillary care that include when to refer locally, when to offer care internally and to provide guidance on assessing urgency. Although the organisational mission may define the population served and the protocol may identify the population for enrolment, policies should further delineate those eligible for ancillary care consideration. The clinician may encounter a participant's sibling or child with an urgent medical need or an urgent need among those screened for participation, but not enrolled. Clear institutional guidance will help the research clinician to navigate such ancillary care boundaries. A description of ancillary care policies should be available to the Institutional Review Board or Research Ethic Committees for review.

Confluence with other approaches

Merritt and colleagues examined ancillary care obligations in community-based public health intervention (CBHI) research, noting that such research tends to be large, takes place within the community and focuses on disease prevention.13 They propose two questions: ‘What are the candidate needs for duty of rescue?’ and ‘For which candidate needs might the study team leadership appropriately bear a duty of rescue?’ The first involves evaluating the severity or urgency of the need and its susceptibility to remediation by individual action. The second involves discerning whether the leadership has the expertise, can apply that expertise without undue cost, whether others can meet the need and competing obligations that preclude action. All of these considerations are factored into the capacity-based model. Thus, the capacity-based model provides a pragmatic framework for the research investigator that encompasses the CBHI context.

As part of a discussion of ‘a community of rights’, Brownsword suggests a procedural four stage test to evaluate for positive rights and responsibilities: ‘1) Is A in a position to assist B? 2) Does A have the capability to assist B in any material respect? 3) Even though A is in a position to assist B and has the relevant capability, would the burden of responsibility on A be unreasonable relative to A's own essential interests? 4) Even though A is in a position to assist B, has the relevant capability, and the imposition on responsibility on A would not be unreasonable (relative to A's essential interests), would B be taking unfair advantage of A if A were required to assist B?’14 Although not rights based (with a focus on identifying the researcher's obligation), the capacity-based model is compatible with Brownsword's four stage test. Questions 1–3 are addressed through evaluating for capacity. The model also only entails an obligation if there is sufficient capacity and urgency. These prerequisites help to curb unreasonable burdens placed on the research infrastructure (helping to guard against ‘B’ taking advantage of ‘A’ as in Brownsword's fourth question). The policy generated from the capacity-based model would establish a ‘duty’, although the justification for the capacity-based model is not ‘duty based’.

The capacity-based model complements guidance offered by the 2006 Georgetown University Workshop on Ancillary Care which generated a list titled ‘The Four P's’: positive duty, planning, partnership and practical provision. This guidance states that researchers/sponsors in developing countries have some positive ancillary care obligation owing to: due concern for welfare, duty of rescue, justice and entrustment.1 It recommends that researchers/sponsors take into account the unpredictable nature of ancillary care needs in their planning. Additionally, the guidance advocates for partnerships within the host community so as to avoid disrupting local health structures and to have representation from the community and population of participants. Through encouraging practical provisions, the guidance seeks to enhance the researcher's/sponsors' ability to address ancillary care needs of participants.1 The guidance also emphasises the need for ethic committees to assess ancillary care obligations and facilitates such scrutiny by providing questions that encourage such an analysis.

The capacity-based model also has some similarities to the partial-entrustment approach although it entails a wider view of scope and narrower view of strength. The scope of the capacity-based model is any urgent condition in a study participant unrelated to the study intervention that is likely to improve with medical care as judged by a trained research clinician. The strength hinges on an evaluation of local capacity and internal research capacity to provide the care (which encompasses vulnerability, dependence and cost, but excludes engagement and gratitude). The capacity-based model also offers the additional facet of providing a framework for utilisation of the local healthcare infrastructure.

Justification

If local health infrastructure has the capacity to provide quality, accessible care, it should serve as a first resource so as not to remove the participant from the local system. This approach borrows from theory on sustainable development: ‘meet[ing] the needs and aspirations of the present generation without compromising the ability of future generations to meet their needs.’15 Encouraging local medical care helps maintain the participants' connection to such care and would minimise the concern of providing duplicative services. Furthermore, if the researchers/sponsors attempt to meet all ancillary care needs, this would strain study ‘budgets and monopolize the scarce time of trained personnel.’1

But where access to local quality care is insufficient and a research facility has the capacity, and especially where the need is urgent, research clinicians should still step in to provide requisite ancillary care. By reviewing the participant's confidential medical information and supervising aspects of the participant's health, the research clinician has assumed the role of a treating clinician triggering the need for a competent response.11 The capacity-based model also embraces justice: persons equal in whatever respects that are relevant should be treated equally. ‘Virtually all accounts of justice in healthcare hold that delivery programs and services designed to assist persons of a certain class…should be made available to all members of that class. To deny benefits to some when others in the same class receive benefits is unjust.’16 The capacity-based model treats study participants as equals for ancillary care consideration based on clinical need, regardless of differences between protocols or of differences in discretion across research clinicians. The capacity-based model lends itself to decision rules which help guard against ‘the unfairness of decisions made arbitrarily or on the basis of personal prejudice.’17 Research clinicians attending to study participants ought to provide care based on policy, driven by medical need and available resources.

Limitations

The capacity-based model does not address healthcare access beyond the study duration, although utilising local health infrastructure as a first resource facilitates reliance on those resources after study discontinuation. Also, the capacity-based model is vague on medical needs of lower urgency leaving room for the organisational leadership to craft ancillary care policies for less urgent needs based on their context and resources. This presents the challenge that ‘internal capacity’ is affected by budget decisions. If applied incorrectly, this gives authority to those who control the budget to define parameters that drive internal capacity rather than having an ethical obligation drive the budget and human resource allocation. The capacity-based model also does not negate concerns related to exploitation through the provision of ancillary care services, but may lessen the potential through utilising the local healthcare infrastructure as the first resource.

Conclusion

Since clinical trials enrolling infants in developing countries are likely to encounter illnesses unrelated to the study intervention and the cost of AE procedures is closely tied to responding to these unrelated events, investigators who plan to work with such populations would be prudent to consider ancillary care issues before engaging in research with that community and to define such parameters for care in their organisational planning, procedures and policies. In contrast to the theoretic models currently proposed as intermediaries between the fiduciary and non-exploitation approaches to ancillary care obligations, simple, pragmatic guidance is needed that can lend itself readily to crafting institutional policies that navigate ancillary care boundaries. The capacity-based model that we propose can help paediatric researchers work within the resources of their research projects, formalise the administration of ancillary medical care, and minimise the possibility and perception of prejudicial allocation of care.

References

Footnotes

  • The opinions expressed in this paper are those of the authors and do not necessarily represent the policies of the Food and Drug Administration or the Department of Health and Human Services.

  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Linked Articles