Constructing a Systematic Review for Argument-Based Clinical Ethics Literature: The Example of Concealed Medications

Abstract

The clinical ethics literature is striking for the absence of an important genre of scholarship that is common to the literature of clinical medicine: systematic reviews. As a consequence, the field of clinical ethics lacks the internal, corrective effect of review articles that are designed to reduce potential bias. This article inaugurates a new section of the annual “Clinical Ethics” issue of the Journal of Medicine and Philosophy on systematic reviews. Using recently articulated standards for argument-based normative ethics, we provide a systematic review of the literature on concealed medication for the management of psychiatric disorders. Four steps are completed:

1. identify a focused question;

2. conduct a literature search using key terms relevant to the focused question;

3. assess the adequacy of the argument-based methods of the papers identified; and

4. identify conclusions drawn in each paper and whether they apply to the focused question.

We identified seven papers and provide an assessment of them. While none of the papers fully meet the standards of argument-based ethics, they did provide rationales for the use of concealed medications, with the important requirement such a practice be accountable in explicit organizational policy to prevent abuse of patients with mental illness or dementia.

Keywords:

• systematic review
• argument-based ethics
• concealed medications
• mental illness
• dementia

I. INTRODUCTION

The literature of clinical ethics incorporates both evidence-based or empirical ethics as well as argument-based or normative ethics investigation of a wide range of clinical topics. The methods of empirical ethics are well-understood and include accepted methods for both qualitative and quantitative research (Sugarman & Sulmasy, 2001). A set of standards for argument-based ethics has recently been identified (McCullough, Coverdale, & Chervenak, 2004). It is remarkable that a literature meant, at least in part, for clinicians, to influence their clinical judgment, decision making, and behavior, lacks a genre of scholarship that is common to the literature of clinical medicine: systematic reviews. As a result, the literature of clinical ethics lacks the corrective effect of review articles that are designed to reduce the potential for bias in the conclusions drawn in both the empirical and normative clinical ethics literature. This article inaugurates a new section of the annual clinical ethics issue of the Journal of Medicine and Philosophy on systematic reviews and we invite readers to submit their own to the Journal's peer-review process.

Systematic reviews of the empirical ethics literature should be conducted following already accepted standards for reviews of both qualitative and quantitative literature in clinical medicine (Drummond, Richardson, O'Brien, Levine, & Heyland, 1997; Giacomini & Cook, 2000a, 2000b; Guyatt, Sackett, & Cook, 1993, 1994; O'Brien, Heyland, Richardson, Levine, & Drummond, 1997; Oxman, Cook, & Guyatt, 1994; Wilson, Hayward, Tunis, & Bass, 1995a, 1995b). To date, there has been no explicit identification of an appropriate method for constructing a systematic review of the normative clinical ethics literature. The purpose of this paper is to identify what should constitute a systematic review of the normative clinical ethics literature and to demonstrate its application to one clinical ethics topic: the use of concealed medicine to treat non-adherent patients with mental illness or dementia.

II. A METHOD FOR SYSTEMATIC REVIEWS OF THE NORMATIVE CLINICAL ETHICS LITERATURE

The general method for conducting a systematic literature review is well-understood. The first step is to identify a focused question that arises from a clinical scenario, an organizational management challenge, or a public health problem. A focused question specifies the target group of interest, identifies a specific intervention or organizational/public policy issue, and identifies relevant outcomes for the review. The second step is to conduct a literature search, using key terms relevant to answering the focused question. These key terms and the databases searched should be clearly identified. Inclusion and exclusion criteria for the types of papers searched should be clearly specified. The purposes of this step are to ensure a comprehensive attempt to identify literature relevant to the focused question and to allow readers to appraise the adequacy of the search. The third step is to assess the adequacy of the methods of the papers selected on the basis of the literature search. The fourth step is to identify the conclusions drawn in each paper and whether to apply these to the focused question. This step is crucial because the focus of evidence-based medicine is to improve clinical practice, organizational leadership, and health policy in accordance with the best information available in the literature.

An appropriate method for conducting a systematic review of the normative ethics literature adopts the first two steps unchanged and adapts the third and fourth steps in response to the distinctive method of argument-based ethics. The third step assesses the adequacy of the ethical analysis and argument of each individual paper selected against the standards of argument-based ethics. These are matters of judgments, just as they are in systematic reviews of the empirical literature. These judgments should be explained. The fourth step identifies the conclusions and their clinical applicability, inasmuch as the focus of clinical ethics is on improving clinical practice (McCullough, Coverdale, & Chervenak, 2004).

III. AN EXAMPLE OF A SYSTEMATIC REVIEW OF THE NORMATIVE CLINICAL ETHICS LITERATURE

To provide an example of a systematic review of the normative clinical ethics literature, we chose the topic of concealment of medication in the drink or food of the mentally ill or demented. This topic was generated by discussions that arose in the clinical care and teaching of one of us (JHC) in one of Baylor College of Medicine's primary affiliated hospitals. Our interest was prompted by the case of a hospitalized patient with a major mental disorder whose behavior was disruptive on the unit and who was impaired in his ability to appreciate the need to take his medications and the consequences for him and the other patients in the unit of not taking his medications. Providing medication to this patient, who was repeatedly refusing it, by means of concealment in his food, was proposed by a member of the clinical team, as an alternative to forcibly holding the patient down and injecting medication. (Texas law allows for such forcible medication with a court order.) This team member argued that concealment of medication appeared to be less violative of the patient than forcible medication and therefore should be seriously considered.

A. Focused Question

In response to this case we formulated the following focused question: In patients with mental disorders (schizophrenia, dementia), is use of concealed medications in food or drink, rather than prescribing medications in the usual way or forcibly administering them, ethically justifiable? There are four parts to this focused question, as is the case in focused questions in evidence-based medicine. The first part concerns the patient population, which is patients with mental disorders or dementia. The second part concerns the intervention, which is the use of concealed medications in food or drink. The third part is the comparison of this intervention with the alternative of prescribing in the standard fashion, i.e., openly with the patient's consent, or forcibly provided that a court order can obtained. Forcible medication involves the physical restraint of the patient and then injection intramuscularly, without the patient consent and against the patient's objections. The fourth part concerns the outcome of interest for the question, which here concerns whether there is an ethical justification for concealed medication.

B. Literature Search

The terms selected for the literature are a function of the focused question, because the goal is to identify as many publications as possible that might be relevant for answering the focused question. For our focused question, we selected combinations of “concealed” or “covert” or “surreptitious” medications, with “ethics,” with “schizophrenia,” with “dementia,” and with “delirium” and/or “psychiatry” and searched the PUBMED database. A “grey literature” was performed non-selectively by examining the titles of chapters in major anthologies on ethics in psychiatry. We also searched in the reference lists and notes of the articles chosen for this systematic review.

C. Methods

The inclusion criteria were any article (scholarly paper, opinion, editorial) or chapter in an anthology that addressed the focused question, i.e., raised and addressed ethical issues pertaining to the focused question (Ahern & van Tosh, 2005; Griffith & Bell, 1996; Honkanen, 2001; Stroup, Swartz, & Appelbaum, 2002; Treloar, Philpot, & Beats, 2001; Welsh & Deahl, 2002; Whitty & Devitt, 2005). We excluded articles that provided empirical studies, e.g., surveys of attitudes of healthcare professionals toward concealed medication (Srinivasan & Thara, 2002; Treloar, Beats, & Philpot, 2000). We also excluded letters, because this form of publication usually does not allow for extended ethical analysis and argument, which was true of the letters we identified (Chua, Choy, & Wong, 2001; Noroian, 2005).

As can be seen from Table 1, for each publication chosen, we identified the patient population to which the article or chapter was addressed. As can also be seen from Table 1, based on recently published standards for argument-based ethics, we used a simple scoring system for each article in the following domains: whether the publication stated a focused question; literature search (present and clearly defined); quality of the ethical analysis and argument; clear statement of conclusion; and clear statement of the clinical application of the ethical analysis and argument and their conclusion. We assigned an overall score and identified the position taken in the publication on our focused question. We scored in each domain as follows: 0 = absence of the domain; 1 = entirely present; and a fraction of ½ to indicate partial fulfillment of a domain.

TABLE 1 Critical Analysis of Papers on Concealed Medications

	Patient Population	Focused Question	Literature Search	Ethical Analysis and Argument	Conclusions	Clinical Application	Overall Score	Position Taken
Griffith & Bell, 1996	History of possible temporal lobe epilepsyand current elevated mood	1	0	0	1	1	3.0	Concealed medication in the case considered was not unethical
Treloar, Philpot, & Beats, 2001	Not clearly specified	0	0	½	1	1	2.5	Permissible with safeguards to prevent abuse
Honkanen, 2001	Dementia	0	0	0	1	1	2	“Underground” mediation not ethically justifiable; medication may be justifiable with surrogate consent
Stroup, Swartz, & Appelbaum, 2002	Schizophrenia	½	0	1	1	1	3.5	Not usually justifiable; consider advance directives and other approaches
Welsh & Deahl, 2002	Not clearly specified	1	0	½	1	1	3.5	Best-interest judgments are not just clinical but also societal and legal
Whitty & Devitt, 2005	Severe mental illness	1	0	½	1	1	3.5	In absence of a single rule applied to all cases, adopt a multidisciplinary approach
Ahern & van Tosh, 2005	Not clearly specified	0	0	0	1	1	2.0	Never ethically permissible

D. Results

We found seven publications that met our inclusion criteria. Four are articles and three are opinions or editorials.

Griffith and Bell (1996) commented on a particular case in which a nurse was suspended for obeying a physician's order to administer concealed tranquilizing medication to a patient with possible frontal lobe epilepsy who refused work-up and treatment and was described as “cognitively intact and physically strong” (Kellett, 1996, p. 1249). This commentary's strengths included a focused question, conclusary statements, and clinical application to the case of concealed medication. However, there was no literature search or ethical analysis and argument. As a result, the opinions often lack supporting justification and therefore count as “mere opinion,” which is not acceptable in argument-based ethics.

Treloar, Philpot, and Beats' (2001) article states a conclusion and its clinical application. On the other hand, there was no clearly focused question, nor a literature search. There was a partial ethical analysis and argument to the effect that, when patients lack the capacity to make decisions, the legal duty to care should guide clinicians in using concealed medication, provided that they take steps in organizational policy and practice to prevent abuse. They also remark that “[a]dvance directives may perhaps help provide some indication as to a patient's views on covert medication” (Treloar, Philpot, & Beats, 2001, p. 63), but they do not identify this remark for what it is, viz., a proposal for a preventive ethics approach to concealed medication in which the patient will have consented in advance to this therapeutic modality. They conclude that “medication may be concealed in foodstuffs in extreme circumstances” (Treloar, Philpot, & Beats, 2001, pp. 63–64).

Honkanen's (2001) very brief paper does take a position that has clinical implications and does define the patient population. However, it does not provide either an ethical analysis or argument, despite the fact that it uses the technical language of medical ethics. This paper provides an example of what should not be done in medical ethics, asserting a conclusion without a supporting ethical analysis or argument.

The next three papers received the highest score, 3.5 of 5, in our evaluation. Stroup, Swartz, and Appelbaum (2002) provide a more comprehensive review of the literature, although they do not formally identify their search strategy. Their paper was prompted by an empirical study that indicated that hiding medicines in food was not uncommon in a clinic in India for patients with schizophrenia. Of all the papers we identified, this paper provides the most thorough ethical and legal analysis of the issues involved in concealed medications, including the consideration of “Ulysses' contracts” or a kind of advance directive that mentally ill patients, when they have decision-making capacity, could use to authorize treatment against their resistance later when they relapse or go off their medications and lose their decision-making capacity. Stroup, Swartz, and Appelbaum (2002) take a broad approach, considering the roles of the patient, the patient's family, psychiatrists, and the service system in providing care for patients with schizophrenia. They concluded that psychiatrists should not routinely direct family members to conceal medications. Instead, advance directives and other approaches, including psychoeducation, should be considered.

Welsh and Deahl (2002) aim to “identify relevant legal and ethical perspectives” (Welsh & Deahl, 2002, p. 123) on “covert medication.” Their patient population was not clearly defined, nor was a formal search strategy. Their paper asks a number of relevant questions without follow-through, thus resulting in a paper that lacks the substance of discursive argument. Their paper makes the important point that judgments about the best interests of patients are not just clinical but also societal and legal.

Whitty and Devitt (2005) define their patient population as those with severe mental illness. They do not define a search strategy. Their ethical analysis appeals mainly to the consequences, both advantageous and disadvantageous, of “surreptitious prescribing.” Strengths of their analysis are that both ethical and legal consequences were identified and that there was an emphasis on the risk of breach of trust in the physician-patient relationship. While they do not take a definitive position, they do suggest that the final decision is likely to be multidisciplinary, involving all healthcare professionals contributing to the patient's care, an important strategy for managing uncertainty.

In a one-page editorial, Ahern and van Tosh (2005) take a strong preliminary position that “surreptitious prescribing” is “coercive and forced treatment at its most sinister” (Ahern & van Tosh, 2005, p. 383). Their editorial refers only to another article in the same issue of Psychiatric Services, but there are no other references, no focused question, no ethical analysis, and no argument. Their position that concealed medication is never ethically permissible is a forcefully presented but unargued opinion.

IV. DISCUSSION

Our findings should be understood in the context of the following limitations. First, we did not systematically review the “grey” literature — publications other than those in the peer-reviewed literature. In clinical ethics, and in bioethics and philosophy of medicine more generally, original scholarship often appears in the form of invited chapters in anthologies. This is unlike chapters in basic and clinical science textbooks, which are not expected to be original scholarship but rather summaries and consolidations of existing literature. We did hand-search anthologies in ethics in psychiatry, but we may have missed some.

Second, we did not score the articles independently but decided on scores by consensus. Therefore, we could not assess inter-rater reliability. Other raters might have reached different judgments and therefore scores.

Third, we approached the review topic on the assumption that the literature would be opposed to this practice but that a case could be made for a controversial clinical intervention, concealed medication. We took care to prevent these assumptions from influencing our assessments of the publications that we identified. We took our methodologic instruction in this matter from the qualitative research literature that urges investigators to declare their preconceived ideas before the analysis begins.

Fourth, our scoring system is based on the domains identified in our previous paper on argument-based ethics (McCullough, Coverdale, & Chervenak, 2004). We are confident that these domains comprehensively identify the domains of argument-based ethics. We therefore recommend its use for the preparation of systematic reviews for the clinical ethics literature.

There are four noteworthy findings of this systematic review. First, even though three papers scored rather well, none of the papers fully meet the standards of argument-based ethics. They therefore do not provide a reliable point of departure for addressing our focused question. Second, there is no stable nomenclature used in the titles and texts of these publications to characterize the practice of concealed medication. Third, the words chosen are freighted with the potential for bias, e.g., “surreptitious,” “insidious and deceitful practice” that causes “irreversible damage” in the absence of empirical evidence that would support such a judgment. Fourth, none of the publications provided a formal search strategy or literature review based on such a strategy, making it difficult for the reader to reach a judgment about whether the literature cited in the publications omits other publications that might be relevant to the focused question of our systematic review.

Despite these problems, our review of these publications did help us to become aware of the range of ethical, legal, organizational, and societal issues that are prompted by our focused question. Keeping these in mind in the clinical setting is an important antidote to bias in favor of a particular position on concealed medications. To this end, we found helpful the emphasis on setting up a process of accountability, involving, for example, multidisciplinary teams and explicit organizational policy to prevent abuse of patients with mental illness or dementia. In others, authors who supported and authors who opposed concealed medication were thinking in terms of preventive ethics, i.e., the need to anticipate ethical challenges in a controversial area and to responsibly manage those challenges through organizational policies and practice.

While the positions taken in these publications vary, the weight of positions across articles is that a strategy of concealed medications should not be used in the clinical setting or be authorized by physicians to be used in community settings for the various medical and psychiatric decisions prescribed. This creates a modest burden of proof for a position in favor of such a strategy. That burden of proof would be steeper if those opposed had fully met the standards of argument-based ethics. Since those in favor did not do so either, we were left only with some help in addressing the ethical issues.

In our judgment, these ethical issues concern mainly whether concealed medication violates patient autonomy and undermines trust in the physician-patient relationship. Patients with advanced dementia or inadequately managed major mental illness often lack the capacity to make decisions. They may not be able to pay attention, to absorb, retain, and recall information, to reason from present events to future possible circumstances, to appreciate relevant clinical information, and to assess that information in terms of their values and beliefs, although they may usually be able to say “no” or to physically resist medication. In such cases, to assert that respect for the patient's autonomy creates an ineluctable constraint on what otherwise would be behavior that is deceitful misunderstands the implications of this ethical principle. This has implications for the trust-argument, because patients with significantly impaired decisional autonomy lack the cognitive apparatus to appreciate a trusting relationship in the first place. In other words, an autonomy-based objection to concealed medication, which was the most frequent objection (but not argument) made, does not succeed.

Troelar, Philpot, and Beats (2001) appreciate this logic of respect for autonomy and thus rightly emphasize that the main consideration is a duty to care, which they couch mainly in legal terms, but is, more importantly, an appeal to the well-known ethical principle of beneficence. This ethical principle requires clinicians to provide interventions that involve the greater balance of clinical goods over harms for patients. The alternatives to concealed medication are non-treatment and forcible treatment. Non-treatment violates the principle of beneficence and therefore professional integrity, and is therefore ruled out. Forcible treatment risks physician and psychological injury that could be serious, long-lasting, and irreversible on a magnitude perhaps greater than these sequelae in the case of concealed medication. Forcible medication involves biopsychosocial harms, where concealed medications involve mainly psychosocial harms, provided that dosing and efficacy are well established. That is, there is a beneficence-based case to be made for concealed medication.

At this point concern in publications about abuse and preventing abuse through a system of organizational accountability becomes very pertinent. A thoroughgoing beneficence-based case for concealed medication must take seriously the potential psychosocial harms of a practice of concealed medication and seek to prevent them. A system of prior review and justification, accompanied by rigorous quality enhancement, may well achieve this important beneficence-based goal.

In summary, we conclude that the answer to our focused question is the following. For patients with seriously impaired decisional autonomy, who physically resist medications and clinically deteriorate as a result, concealed medication directed by an organizational policy of accountability is ethically justified. This is an example of a topic normative clinical ethics for which an argument can be developed in the absence of an already well-developed argument identified as a result of a systematic literature review.

V. CONCLUSION

The clinical ethics literature, and the bioethics and philosophy of medicine more generally, lack the disciplined approach afforded by conducting systematic reviews. Such reviews improve the quality of subsequent clinical decision making and scholarship. There are annotated bibliographies on many topics and the very useful “Scope Notes” series in the Kennedy Institute of Ethics Journal, but these should not be categorized as systematic reviews. Systematic reviews play a major role in the basic and clinical sciences, and they should play a similar role for clinical ethics.