| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
Correspondence: Address correspondence to Barry D. Lebowitz, PhD, National Institute of Mental Health, 6001 Executive Blvd., Rm. 7160, MSC 9635, Bethesda, MD 20892. E-mail: blebowit{at}mail.nih.gov
 |
Abstract |
|---|
 TOP Abstract References |
|---|
Key Words: Clinical trials • Alzheimer's disease • Depression • Methodology
What a story this report tells: the capacity of contemporary health care to address issues of chronic disease, the multiple actions of a dozen different medications, a daughter closely involved in her mother's care, an assisted living facility developed specifically to care for people with Alzheimer's disease. And this only scratches the surface of the issues that are raised by this woman at the end of her life.
I begin with this "case report" for many reasons, but principally as a reminder that health policy, and the research that informs that policy and provides the evidence base for its development, must address the issues that are raised by this woman's needs. Stanley J. (Steve) Brody used to say that health policy was easy if someone thinks that all it involves is moving boxes and arrows on a piece of paper. The trouble is that health care is not just boxes and arrows. Health care is the woman in this report; it is people and families, the clinicians who care for and about them, and the administrators and policymakers who create the context in which that care is provided. Our challenge in research is providing the evidence that can guide, or at least inform, the development of those policies.
The Gerontological Society of America (GSA) recognizes that challenge. Our mission in the GSA is to promote the conduct of multi- and interdisciplinary research in aging by expanding the quantity of and improving the quality of gerontological research and by increasing its funding resources and to disseminate gerontological research knowledge to researchers, to practitioners, and to decision and opinion makers (http://www.geron.org/mission.htm). This dual mission has fundamentally shaped the nature and direction of the work that I am honored to present here. And by doing so, I hope, in a small way, to pay respect to the legacy of M. Powell Lawton. For over 30 years, Powell was my mentor, teacher, colleague, and friend. In a very real sense, his influence shaped the direction of my entire career and the careers of many others of my generation.
Since our challenge is to provide evidence that can be used to guide or inform policy development, we must look at what we study, how we study it, and what we do with the results. In particular, there are four aspects of the issue of evidence that I intend to explore. By doing so, I hope to identify some considerable strengths in our field but also to specify areas and issues in need of attention and development. I will emphasize four broad areas:
Targets of Significance
As part of a volume that, in retrospect, serves as a festschrift to Lawton (Lebowitz, 2000b), I linked the emerging "public health" model of treatment research to many of the themes that had been articulated in Lawton's work. I argued that mainstream mental health research was moving away from short-term symptom management studies in highly selected populations. Research in young and midlife adult populations was beginning to address problems of long-term functional outcomes in heterogeneous and representative populations. I concluded that the models that Lawton had developed in gerontology and geriatrics were now being exported across the life course. In particular, attention to comorbidity, function, disability, and quality of life in settings of general health care, social service, and long-term care was now the standard against which new approaches were being developed in research with early and midlife populations. Looking back, many will recall the discussions in the 1970s and 1980s about the need to "gerontologize" training and practice in health and social services. How far we have traveled from trying to get attention to our issues to, now, providing leadership in conceptual and methodologic development.
In the description of the public health approach to treatment research, I used the example of Alzheimer's disease to illustrate the point. For reasons having to do mainly with academic politics and strategies to optimize reimbursement for care, the prevailing sense was that Alzheimer's disease should be seen as strictly a cognitive disorder. Any other symptoms or dysfunctions were regarded as secondary and incidental to the core cognitive deficit. My main point in the article involved arguing the somewhat unusual case that we needed to get back to Alzheimer's original conceptualization, specifically that this disease was multidimensional, with dysfunction not only of cognition but also of mood, thinking, and behavior.
Much has happened since then to establish this multidimensional principle as the standard in the field. This view has guided a new generation of treatment development studies. And once again, gerontology and geriatrics have led the way—this time, for a whole new approach to the targets of treatment. The issue can be posed as follows: What is it that we treat when we develop treatments for brain disorders?
First and foremost, we are dealing with a very complex system, and the complexity seems to be increasing as our knowledge increases. The scale of the central nervous system goes from 102 (the number of neurotransmitters) to 1012 (the number of synapses). In addition, data from imaging and from neuropsychology show that the pathology associated with brain disorders is widely distributed through the central nervous system. Genetic studies have shown that many, if not most, brain disorders are polygenetic with multiple biochemical pathways that may contribute to the expression of a single disease. As an interesting link to some of Lawton's early work on environmental press (Lawton & Nahemow, 1973), it has been estimated that environmental factors may contribute as much as half the variance in disease expression. Moreover, the joint action of environment and genetic factors may be necessary for disease incidence (Caspi et al., 2003; Mossner et al., 2001). Coupled with a system of diagnostic classification that is often ambiguous and incomplete, our challenge in treatment development is formidable indeed (Kennedy, Farrer, Andreasen, Mayeux, & St. George-Hyslop, 2003; Merikangas & Risch, 2003).
Although we like to think that we treat diseases, for the time being at least, it is symptoms, and occasionally syndromes, that are treated. In other disease areas (e.g., rheumatoid arthritis), syndromal targets are well accepted. Not so in the central nervous system, however, where concerns with "pseudospecificity" have limited target identification (Laughren, 2001). With Alzheimer's disease, research in neuropathology, brain imaging, and neuropsychology all provided the basis for what developed as the first instance of regulatory and policy recognition that development of syndromal indications for treatment of the central nervous system is appropriate and necessary.
In the United States, the certification of biological treatment approaches is regulated by the Food and Drug Administration (FDA). The FDA has identified a number of conditions that must be satisfied in order for a treatment to be approvable (Laughren, 2001), that is, even before assessing the safety and efficacy of the proposed treatment. Two different approaches are possible.
One approach takes the route of demonstrating that the proposed treatment is specific to the clinical entity under consideration but not "pseudospecific." That is, there cannot be a claim of uniqueness for a particular clinical entity in the absence of data supporting such a claim. There must be general acceptance of the treatment target in the expert community, identification of a reasonably homogeneous patient population, and, ideally, some understanding of etio/pathophysiologic mechanisms. Alzheimer's disease satisfies each of these criteria (Lebowitz, 2000b), thus paving the way for development and validation of specific, operationally definable diagnostic criteria. Along with the general categories of Alzheimer's disease and mild cognitive impairment, specific diagnostic criteria have been developed for three syndromes in Alzheimer's disease: psychosis (Jeste & Finkel, 2000), depression (Olin et al., 2002), and sleep disturbance (Yesavage et al., 2003).
A second approach to gaining approval for a syndromal treatment indication is to demonstrate that the syndrome is similar across a wide variety of illnesses. Pain and fever are two such generalized examples; agitation seems to fall into this category as well. By establishing the principle that approval of syndromal indications is achievable in Alzheimer's disease, basic and translational research in the field has established a new context for treatment development studies (De Deyn et al., 1999; Katz, Jeste, Mintzer, Clyde, Napolitano, & Becher, 1999; Lyketsos et al., 2003; Schneider, Katz, Park, Napolitano, Martinez, & Azen, 2003; Street et al., 2000).
It is certainly the case that not all conditions would fulfill syndromal indication criteria. In the absence of evidence of a direct pathophysiologic link, it would be appropriate to regard the joint occurrence of common conditions as reflecting comorbidity (e.g., depression following myocardial infarction, as in Glassman et al. [2002]). In other cases, the condition might be regarded as treatment emergent (e.g., the depression associated with interferon treatment, as in Musselman et al. [2001]).
In addition to the obvious clinical benefit of having treatment options available for a range of severely disabling conditions, a major contribution of this syndromic approach is that research in geriatrics has once again paved the way for developments in the rest of the life course. The regulatory approvals of the atypical antipsychotic clozapine as a treatment for suicidality in schizophrenia and of injectable antipsychotics for the treatment of agitation in schizophrenia or mania are derived directly from the concept of syndromal indications in Alzheimer's disease. A renewal of interest in the enhancement of cognitive function in adults with schizophrenia and in the treatment of depression in Parkinson's disease has been stimulated by the acceptance of syndrome-based indications in Alzheimer's disease. The public health implications of these developments are profound and important.
Appropriate Methodologies
In order for our data to contribute to the policy process, the research upon which it is based must meet the highest standards of methodologic excellence. My work has been focused on therapeutics and clinical trials, so let us examine the use of clinical trials data to guide what has come to be called "evidence-based practice." In other words, how good are the treatment approaches we use?
The evidence from clinical trials does not provide a definitive answer to that question. Why? It could be that our treatments are not very good. Or it could mean that we do not study them correctly.
Most controlled trials of antidepressant medications fail to demonstrate superiority of active treatment over placebo both in midlife (Khan, Khan, Walens, Kolts, & Giller, 2003) and in late life (e.g., Roose, 2002). And of those trials that do identify positive treatment effects, the size of that effect is so small (Schneider et al., 2003; Tollefson, Bosomworth, Heiligenstein, Potvin, & Holman, 1995; Williams et al., 2000) or the intervention so complex or unfeasible (Bruce et al., 2004; Unutzer et al., 2002) that any reasonable person would avoid these treatments whenever possible. Large placebo effects have been observed in trials of new agents to treat depression or the psychotic symptoms of Alzheimer's disease. At the same time, paradoxically, we know that millions of people are treated for these disorders worldwide and that many of them derive substantial benefit from treatment. How do we account for the difference? How is it that treatments work very differently in actual practice than they do in controlled clinical trials?
I believe that a major source of the difference lies in the fact that the methods we use to study new approaches to treatment in human clinical trials were set, for the most part, in the 1940s and 1950s with the first modern randomized controlled trial, the Medical Research Council trial of streptomycin (1948). Having frozen our methodology and not taken advantage of developments in contemporary statistical science has had a downside, however. Nearly everything we do in clinical trials minimizes treatment response, enhances placebo response, and, in the case of combination treatment studies, distorts the value of certain active comparators. There are many factors that contribute to this. I will focus on three of them: patient selection, study operations, and statistical analysis.
Patient Selection
Participant selection often is limited (Anand, Hartman, Sohn, Danyluk, & Graham, 2003) by censoring the severity distribution through entry criteria where exclusions often require elimination of patients who are suicidal or psychotic or have a variety of common comorbidities such as physical illness, substance abuse, or personality disorder. Since these are typically associated with severity, elimination leads to preferential recruitment of less severely ill participants (Schneider, Olin, Lyness, & Chui, 1997) or otherwise nonrepresentative patient samples. This is not restricted to geriatrics but is generally true in all registration-oriented studies in mental disorders (Licht, 2002; Robinson, Woerner, Pollack, & Lerner, 1996; Zimmerman, Mattia, & Posternack, 2002). Similarly, many studies have long no-treatment periods for observation and diagnostic purposes. Clinicians are not likely to withhold treatment from severely ill patients in whom immediate treatment is seen as an absolute necessity. Thus, only those patients who can tolerate a long no-treatment period are recruited. These are likely to be less severely ill (De Deyn et al., 1999; Katz et al., 1999; Street et al., 2000). It has long been observed that the lower the severity of illness, the higher the likelihood of placebo response.
If advertising is used as the primary source of recruitment, a group of high-likelihood placebo responders, the "worried well," are often the ones recruited. If, on the other hand, recruitment is largely from active clinic or practice populations, there is likely to be preferential treatment of nonresponders under the justification that it is inappropriate to "rock the boat" for patients who are doing well by switching to a new treatment.
In conclusion, the patients recruited for typical studies are either very likely to respond to placebo or to respond to nothing at all—this is hardly an ideal laboratory for the testing of a new approach to treatment.
Study Operations
There seems to be a general notion that treatment studies are "easy," an impression that can be validated at any elementary school science fair. Methodology is seen as straightforward: participant recruitment, exposure of some to one treatment option and the others to a different one, and follow-up. Even if studies were as easy as this to design—and they are not—they are very difficult to actually carry out (see, e.g., Friedman, Furberg, & DeMets, 1998). Many studies fail not because they are poorly conceived but because they are poorly organized and run.
In addition to overall problems of operations, there are several specific issues that compromise study outcomes. For example, in large, multiple-site studies, recruitment is often set as a horserace with enrollment concluded when the overall study sample is achieved. Payment to sites is largely on a per-patient basis. Therefore, as the study recruitment goal appears closer, there is a financial incentive to enroll participants immediately, that is, while enrollment is still possible. Over time, that could lead to a drift in criteria: Mild symptoms are seen as more serious, and borderline eligibility gets to be considered acceptable. This erosion of severity promotes recruitment of those with milder severity who, as described above, are more likely to respond to placebo.
In combination treatment studies, that is, those that include both a pharmacologic and a nonpharmacologic approach, operational concerns may minimize differences between pharmacologic and psychosocial treatments. Typically, the approach to pharmacotherapy is rigid with fixed dosing and short duration—an approach uncharacteristic of clinical practice. Clinicians in the pharmacotherapy arm must behave robotically and in a manner unlike the way they generally work. Training and supervision are minimal. Those doing the psychosocial treatment, on the other hand, are typically highly selected, well paid, well trained, and closely supervised. We therefore commonly encounter trials in which (expertly done) psychotherapy beats (poorly done) pharmacotherapy. In these studies, however, it is very difficult to separate the operational aspects from the true clinical effect of the different treatment approaches (Thompson, Coon, Gallagher-Thompson, Sommer, & Koin, 2001).
Statistical Analysis
The preferred approaches to statistical analysis of trials data, such as "intent to treat" and "last observation carried forward," may reward placebo response. Newer approaches such as mixed-effects modeling and survival models may provide crisper alternatives for the identification of treatment effects. And, of course, statisticians continually remind us that effect size estimation, not statistical significance, should be the criterion applied to all trials.
Clinical trials often fail because we feel constrained to follow the classic approaches to clinical trials methodology. New science and the new treatments derived from it should be assessed with a methodology that is appropriate and built upon the best of our current knowledge. As we develop new treatment approaches for major disorders (see, e.g., Holden, 2003), there is a pressing need to re-engineer the standard approaches to clinical trials in the mental disorders (Califf, 2003; Katz et al., 2002). This does not necessarily imply full endorsement of a neo-Bayesian approach to research design (e.g., Lilford, 2003) but does mean that we should take into account a variety of patient, clinician, and environmental factors that could inform design decisions (e.g., Lavori et al., 2001; Sachs, Thase, Otto, Bauer, Miklowitz, & Wisniewski, 2003).
More than ever before, we now have opportunity to revolutionize treatment discovery for the mental illnesses. With advances in cognitive and behavioral science, molecular biology, genomics, proteomics, bioinformatics, and automation, we may have the tools to move to a whole new approach to treatment development (Collins, 1999; McKusick, 2001; Subramanian, Adams, Venter, & Broder, 2001). By using genomics and proteomics, it will be possible to redefine phenotypes (Merikangas & Risch, 2003) by creating mechanism-based classifications of clinical patient populations to use in the development of therapeutic or prophylactic interventions. All this will move us from an approach based on serendipity to one based on rationality (Peltonen & McKusick, 2001). The therapeutics of the future will be expected to identify specific targets with minimal side effects in genetically defined clinical populations (Collins & Guttmacher, 2001), with, as Lawton would have put it, appropriate attention to the social and cultural factors important in late life.
We also need to remember that discovery and development are the beginning and midpoint of treatment development—not the end. Traditional models have limited generalizability and restricted outcome measures and leave substantial amounts of nonresponse, residual symptomatology, and associated disability (Lebowitz, 2000a). New pragmatic or practical trials, based on approaches first articulated by Peto and colleagues (Yusef, Collins, & Peto, 1984) and refined by Tunis, Stryer, and Clancy (2003), are expanding our vision with respect to treatment assessment in our field (Norquist, Lebowitz, & Hyman, 1999). At the same time, the risk of muddying interpretation owing to possible confounding factors and varying treatment exposures needs to be evaluated and taken into account when drawing conclusions.
Instruments and Measures
A highlight of Lawton's contribution to our field was the development and use of new measures: approaches to functional assessment, instrumental activities of daily living, quality of life, daily affect. We continue to have great need for better instruments to use as outcome measures in clinical trials. In psychiatry trials in general, and in depression trials in particular, there is a tradition of using too many measures, most of which are rating scales or self-report forms of dubious validity and notorious unreliability.
Measures are selected because they have worked in the past and are therefore uninformed by contemporary science and knowledge of etiology or pathophysiology. This is important because treatments may be having effects that our standard measures are incapable of capturing. The standard outcome measure used in depression trials, the Hamilton Rating Scale for Depression (Ham-D), was developed nearly a half-century ago to assess the impact of the then-new class of antidepressants, the tricyclics, on severely depressed inpatients (Hamilton, 1960). Some of the problematic aspects of use of the Ham-D in its multiple formats include the following: (a) It does not map well onto the full set of diagnostic criteria needed for depression, (b) it contains items that measure multiple constructs, and (c) it has vague anchor points (IsHak, Burt, & Sederer, 2002; Yonkers & Samson, 2000).
Other areas of health take other approaches that we might adapt. In the 2000 festschrift (Lebowitz, 2000b), I identified several dimensions that held out the possibility of a new look at outcomes: institutionalization, morbidity, mortality, disability, resource use, quality of life, family burden. I cited a few examples of these in that 2000 article. There is a need to go beyond the discrete measure, however, to develop composite, event-based measures for use in mental health. This type of measurement was developed for use in cardiovascular trials (major adverse cardiac events [MACE]). There has been one example used in schizophrenia: a composite measure of suicidality in people with schizophrenia that was used in the trial of clozapine (Meltzer et al., 2003). We need to have more. I have proposed, for example, that we develop a measure of major adverse depression-related events for use in depression trials using suicide attempts, hospitalization, needs for rescue treatment, and significant exacerbation of symptoms.
Conclusions That Make a Difference
Targets, methods, and instruments—all partial answers to a single underlying question: What special considerations are necessary in the design of a geriatric protocol? That is, how do we establish a general understanding of what would constitute a geriatric protocol to assess the safety and efficacy of new treatments (Arean, Cook, Gallagher-Thompson, Hegel, Schielberg, & Schulz, 2003; Lebowitz & Harris, 2000; Norquist et al., 1999)? A proper geriatric protocol is not simply an established "adult" protocol with the age range extended into late life. In studies using drugs, the dosing and duration of treatment must reflect the potential effects of age-related changes in pharmacokinetics and drug metabolism and the effects of comorbid conditions and their treatments. In studies using nonpharmacologic approaches, the type of treatment selected for study must reflect the social, cultural, and clinical realities of late life. In all cases, restrictive exclusion criteria should be minimized in order to optimize generalizability, and both symptomatic and functional outcomes should be assessed.
There is more than a purely scientific context for this set of issues. In his 1986 Donald P. Kent Memorial Lecture, Steve Brody (1987) wisely observed that public policy emerges and is guided by "the felt necessity of the time." One of the felt necessities of our time is the need to address the economics of pharmaceuticals for older people. As I write this, a prescription drug benefit in Medicare is an issue of major prominence in policy development. There seems to be general agreement regarding public funding of many drugs for many older people. It seems to be a question of "when" and "how"—no longer a question of "whether." It seems logical that at some point in this process of policy development, there will be questions raised about the adequacy of the evidence upon which to base drug prescribing for participants in the Medicare program. This evidence needs to come from the kind of studies that I have been describing.
What are the incentives to produce this evidence? There have been two approaches to improving evidence of the safety and efficacy of drugs used by older people. It is beyond the scope of this article to describe these in detail, and I will use very broad brushstrokes.
The earliest of these, developed in 1994 by the International Conference on Harmonization (an international body representing drug regulatory authorities in North America, the European Union, and Japan; FDA, 2003b), establishes the principle that drugs should be studied in all appropriate age groups and be reasonably representative of potential consumers. This principle has never been fully implemented.
A second approach, developed in 2001 by the FDA in the United States, establishes a "Geriatric Use" section of drug labeling with information coming from any source, including spontaneous reporting, not restricted to controlled clinical trials. The variability in the information available under this very open policy has limited the impact of this guidance.
One highly successful approach to generating specific evidence regarding treatments is from the area of pediatrics. Since 1997 in the United States, there has been an incentive, in the form of an additional 6 months of patent protection, for those drugs that are specifically studied in children. These studies must be requested by the FDA, which, as of October 2003, has requested nearly 300 studies and has approved patent extension for 84 drugs (FDA, 2003a). Concerns with the selection of drugs and the overall cost impact have been raised in the context of this incentives-based approach, but no one can dispute the public health value of the program. Indeed, renewal of the legislation authorizing this program was approved unanimously by the U.S. Senate in October 2001. Many of us believe that this type of approach needs to be examined as a way to increase the base of our knowledge in geriatrics.
"Older persons need a dream, not only a memory." Rabbi Abraham Joshua Heschel taught us this at the 1961 White House Conference on Aging. Arthur S. Flemming inspired us with those same words 34 years later, at the 1995 White House Conference (Flemming, 1995). Much remains to be done to fulfill this charge. This spirit has guided the line of work that Lawton established, that those of my generation have been pursuing, and that I have had the honor to present in the M. Powell Lawton Award Lecture.
|
Footnotes |
|---|
1 National Institute of Mental Health, Bethesda, MD.
Decision Editor: Linda S. Noelker, PhD
Received for publication November 18, 2003. Accepted for publication March 11, 2004.
|
References |
|---|
|
TOPAbstractReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  M. A. Espeland, T. M. Gill, J. Guralnik, M. E. Miller, R. Fielding, A. B. Newman, M. Pahor, and for the Lifestyle Interventions and Independence f Designing Clinical Trials of Interventions for Mobility Disability: Results From the Lifestyle Interventions and Independence for Elders Pilot (LIFE-P) Trial J. Gerontol. A Biol. Sci. Med. Sci., November 1, 2007; 62(11): 1237 - 1243. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Pinquart, P. R. Duberstein, and J. M. Lyness Treatments for Later-Life Depressive Conditions: A Meta-Analytic Comparison of Pharmacotherapy and Psychotherapy Am J Psychiatry, September 1, 2006; 163(9): 1493 - 1501. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. B. Mitrani, D. J. Feaster, B. E. McCabe, S. J. Czaja, and J. Szapocznik Adapting the Structural Family Systems Rating to Assess the Patterns of Interaction in Families of Dementia Caregivers Gerontologist, August 1, 2005; 45(4): 445 - 455. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
