As we go beyond the year 2000, health care professionals, regardless of their area of practice, will be required (like it or not) to participate in evolving areas of medical practice that have not been part of the traditional medicine of earlier decades: outcomes assessment, patient quality-of-life (QOL) issues, statistical methods such as decision analysis and meta-analysis, and new areas of patient interest such as alternative medicine, to name a few.1 A voluminous literature already exists concerning QOL assessment in head and neck cancer (HNC) patients, but it is too embryonic to have fostered distinct controversies at this point. This chapter presents a brief overview of where we have been, where we are, and where we are going on issues related to QOL in HNC patients. The goals of this presentation are to make health care professionals aware of (1) new issues that must be dealt with; (2) how espousing, rather than resisting, these new trends can optimize relationships with patients and foster compliance with any desired treatment regimen; (3) current deficiencies in the existing literature toward the goal of correcting them expeditiously; and (4) what can be done now in daily practice to enhance the QOL of HNC patients.
What Is QOL and Why Should We Study It?
QOL refers to a spectrum of factors that, in aggregate, comprise “well-being,” a concept that is currently being defined.2–13
Global well-being includes physical, functional, emotional, and social domains as influenced by a person’s beliefs, expectations, and perceptions, and as influenced by the patient’s education and knowledge about expected outcomes of therapy. Such perceptions may change during the course of therapy and in conjunction with the patient’s disease/health status, cancer-free or not. The person under discussion is the patient–not the health care professional. An individual’s past experience, personality, and ethnic background lead to different expectations regarding health and varying ability to cope with limitations and disability. Two people with the same health status may have different QOL related to different perceptions of health and satisfaction with life. QOL outcomes are often impacted by functional factors involved in cancer treatment. For example, two patients who are “cured” of larynx cancer may experience disparate levels of satisfaction if one has a persistent tracheotomy and/or feeding tube, and the other does not. But QOL is a multidimensional construct in which functional ability is only one component. Outcomes assessment has shown that the severity of physiologic or functional measures often correlates minimally with global QOL, which correlates more accurately with patient satisfaction and is manifested as increased compliance with treatment regimens.
Interestingly, traditional cancer outcome measures (e.g., response rates, sites of failure, disease control, overall survival, cause of death, treatment toxicities) do not always correlate with patient benefit, especially in advanced cancer patients, which makes assessing QOL essential. The ultimate goal of adding such assessments is that the combination of symptoms, function, QOL, cost, and cancer treatment endpoints will lead to better definition of treatment efficacy to facilitate the choice of treatment.9 Interestingly, in a completion of the circle, there is evidence that optimizing global QOL may be a significant predictor of survival time and that improving the QOL of cancer patients may result not only in better quality-adjusted survival, but in longer survival as well.14
QOL issues differ when considering curative intent versus palliation. In the former approach, aggressive attempts to control locoregional cancer may be justified despite attendant toxicities that negatively impact on QOL. In the latter situation (incurable metastatic disease), such an aggressive stance is less appropriate and symptom control—especially pain control—becomes paramount. QOL is also important in the context of adjuvant treatment, which, by definition, is given in patients without symptoms and without objectively evident cancer in whom the main criteria of effectiveness are relapse-free survival and long-term survival.
Except for the inclusion of constitutional B symptoms in lymphoma staging, cancer staging systems have included only factors related to the tumor, and not to the patient. The complexity of the tumor–host interaction in HNC patients has been reemphasized in the recent literature15 and the effect of having left out the patient half of the equation in the past may well negate the putative incremental improvements in cancer control claimed with some of the newer treatment strategies. For example, two recent articles demonstrate that the variable of alcohol consumption (none versus heavy) is associated with a 40 to 50% difference in survival,16, 17 which is much more than the less than 10% differences that can currently be related to variations in cancer treatment for HNC, and that reversing alcoholism correlates with improved survival over time.16 Certainly, in view of these results, this usually neglected factor, often considered impossible to assess accurately in unreliable HNC patients, deserves renewed attention.
The most important goal of QOL assessment is to have a cured and intact patient who can assume as near-normal function as possible for the duration of his or her lifetime. An important practical application of studying QOL is identifying patients and/or families who will have, or are having, problems adapting to their illness and its treatment. The identification of these problems at their onset can lead to expeditious implementation of psychosocial treatment, behavior modification, and instruction in coping mechanisms so that both patient and spouse/family can make a more rapid transition to psychological stability.
One gets the feeling from the existing HNC QOL literature that there should be no morbidity associated with any type of treatment for advanced HNC. In fact, by its very nature, HNC is a mutilating disease. Treatment of any type intrudes on an individual’s physical, emotional, and social well-being, in contradistinction to the age-old perception that the morbidity of HNC treatment relates largely to the surgical component. The achievement of complete normalcy in the treated HNC patient is an unrealistic expectation. Persistent pain, dysphagia, altered voice, and the need for tracheotomy and/or feeding tubes can accompany initial treatment with chemo-and radio-therapy or altered fractionation radiation or traditional radiation with curative intent, as well as with surgical regimens. Among the QOL studies emerging from the University of Chicago, one showed that the 40% of patients who had pain before aggressive chemo- and radiotherapy also continued to experience pain after treatment,18 demonstrating that an objective tumor response does not always correlate with freedom from distress. Another study showed that in a group of patients free of disease at 1 year after organ preservation treatment (although biased toward a favorable result due to selecting cancer-free survivors), that the most important performance impairment was the inability to eat a normal, solid food diet, with 50% of the patients able to eat only soft foods or liquids.19 Obviously, when cancer affects the complex and compact anatomy of the head and neck, the important functions of speech, swallowing, breathing, and appearance will inevitably be affected. It is unrealistic to use the patient’s premorbid condition as the baseline for comparison of functional and QOL results. It is more realistic to focus on the degree of compromise of structure, function, and ensuing psychosocial morbidity to be expected with HNC of various stages and sites, and currently such information is not available. A large standardized sample study including oncologic, functional, and QOL outcomes in the HNC population would help establish such norms.
QOL issues relate not only to patients with recurrent/persistent cancer, but also to cancer survivors. “Successful” cancer treatment (often prematurely defined) has resulted in a class of “walking worried,”20–23 such as treated prostate cancer patients who regularly focus on their prostate-specific antigen (PSA) levels or breast cancer survivors who participate in ongoing hormone or chemotherapy trials. Ongoing anxiety over maintenance of the cancer-free status can become an important health problem. Patients may look and feel physically well, but they are not well emotionally. Survivorship issues for HNC patients relate to long-term use of Salagen24 or participation in chemoprevention trials, with daily medications being an ongoing reminder of potential disease and treatment, as well as ongoing monitoring for second primary tumors. Many studies now assess the cost/benefit of follow-up regimens for treated cancer patients, and an important emerging issue is whether the omnipresent reminders of such surveillance outweigh the potential benefit to well-being of less rigorous follow-up. Such questions raise obvious ethical concerns but are inherent in the modern era of cost containment in medical practice. This leads to an additional reason for health care providers to get involved in QOL and other types of outcomes research. Many other groups (medical consortia, HMOs, federal agencies, insurers) are engaged in synthesizing outcome and economic data toward the development of clinical practice guidelines. Some of the tools of modern medical practice include the development of evidence-based medicine (the 10 to 15% of the medical literature that is considered scientifically valid, based on a randomized, controlled, prospective format), cost-containment research, and outcomes research. It is inevitable that in the not too distant future, health care providers may have to justify their treatment plan to health care payors based on outcomes data and patient satisfaction responses. The “sciences” of practice guideline development and implementation are both in their infancy and too embryonic to serve as a basis for treatment decisions in the individual patient, but there is also no question that investigating health care and documenting health care outcomes is now an industry in its own right in the United States. This fact cannot be ignored. A new class of entrepreneurs has emerged—private companies who are vendors of outcomes instruments. Many flashy systems are available, but the descriptive data collected often have limited usefulness. The push for further development in QOL and related areas now appears to come largely from third-party payors. Unless the medical community actively participates in research that will lead to valid information based on deep medical expertise, outcomes analysts will soon be dictating what success is based on more superficial analysis. In particular, a specialist should not sacrifice involvement in the post-treatment aspects of patient management in an era when surveillance of treated cancer patients is being directed away from them and toward family physicians. Currently, about 40 articles in the QOL literature (often published in journals such as Medical Decision Making , Quality of Life Research Journal, and Journal of the Medical Information Association ) document the cost-effectiveness and importance of ongoing specialist intervention.
QOL Assessment Instruments
QOL issues, heretofore considered too soft and subjective for specialists to deal with are, in fact, objectively quantifiable and do not merely represent the absence of disease. QOL is generally measured by structured questionnaires that can be scored and quantified to generate hard data. Most widely used QOL assessments include measures of physical, psychosocial, economic, and global well-being. There is an explosion in the QOL literature, with more than 30 generic tools available (to assess global QOL and health considerations) and more than 300 disease-specific tools available, including 35 to 45 for cancer patients. The otolaryngologist–head and neck surgeon who ventures into the QOL literature published in journals of psychology, epidemiology, health science, psychosomatic medicine, and sociology may well be confused by the statistical methodology and jargon encountered, such as, “Construct validity was examined in this study through factor analysis using principal components and orthogonal rotation, which revealed that the average scores were function and network properties which did load onto three factors.”25
Different methodologies assess different health outcomes. For example, QOL scales (self-assessment questionnaires) measure the patient’s perceived quality of health retrospectively, whereas utility scales measure projected health state. An example of the latter is the time/trade-off method that identifies the number of years of perfect health the interviewee (often not a patient) is willing to exchange for 5 to 10 years in each health state. Products of outcomes research often apply linear analogue scoring or category rating not only to physical status, but also to psychological and social well-being by assessing variables such as pain, mobility, capacity for sexual relationships, and symptom-free days. Additional complexity arises when cost-effectiveness of treatment alternatives is included.
In general, articles in the QOL literature show that whatever instrument is under evaluation, is reliable (yields consistent values), valid (targets what it claims to measure), responsive (detects changes over time), sensitive (reflects true changes in individual patients), and practical to administer. Instruments that fail to document these parameters probably do not get published, although it would be interesting to see some scales that have not fulfilled these criteria; possibly much could be learned from them. Often lacking are articles that demonstrate the reproducibility of a given instrument; there appears to be more interest among authors in generating a slightly different instrument that is better than any of the others rather than in validating someone else’s. All instruments vary slightly in methodology (e.g., self-response vs proxy, telephone vs mail vs interview) and psychometrically in what is included in the questionnaire. Outcomes research has demonstrated that patients generally like to fill out questionnaires (satisfied patients generally respond more than dissatisfied ones who manifest their dissatisfaction in other ways), but there is concern over how the information gathered will be used; to discriminate against patients in a punitive way? There is also significant concern on the part of providers who are often reluctant to participate in such studies; will patient satisfaction results be used to punish (or reward) providers (“report cards”)? The theoretical desire to include all relevant outcomes and costs is daunting, and there are already too many tools to choose from. It is more realistic to select a QOL product that meets most of the needs of the patient population under consideration and to use it in a prospective, longitudinal manner in large numbers of patients.
QOL studies must be performed prospectively and longitudinally over time, as a cancer patient’s perception of health often changes, and if QOL is assessed at only one point or over a short interval, a distorted picture will emerge. Recall bias is a well-known factor, and responses often differ when they are asked near to or far after an event. An individual’s satisfaction with QOL early after treatment often differs significantly from that perceived years later, especially if side effects become more troublesome or if the cancer recurs. One example from the prostate cancer literature shows that with longer follow-up, persistent incontinence may dampen early enthusiasm over having had a successful operation.26 An analogy in the area of HNC might be that patients who require delayed or have long-term tracheotomies and difficulty swallowing may become depressed, whereas they were initially happy to have their larynx preserved. These conditions may not develop for years after nonsurgical organ preservation treatments, as a result of progressive soft tissue fibrosis in the head and neck, and may not be recognized if cancer follow-up terminates at 5 years, the traditional cure endpoint. Additional surveillance of HNC patients for second primaries should continue indefinitely with yearly chest radiographs that can detect early, operable second primary lung cancers.
Thus, measurement of QOL is dynamic, rather than static, and requires periodic reassessment. Short-term results cannot be extrapolated to reflect long-term outcome. It is likely that prospective, long-term, longitudinal studies, which are lacking in the HNC literature, may better elucidate the patient’s true experience in the future. Similarly, isolated observations in long-term survivors are biased if they are meant to represent the population at large, because patients who have had less favorable cancer-control results and outcomes have already expired or otherwise been selected out.27
How often QOL outcomes should be measured depends on the condition and the treatment given. Assessment before treatment and after each component of multimodality treatment, if clearly definable, as well as longer-term assessment, might be reasonable for the HNC population. For example, in patients who undergo surgery followed by radiation or chemoradiotherapy, assessment before treatment, after surgery/before radiotherapy, after radiotherapy, 6 months after the end of treatment, and 1, 3, and 5 years after treatment seems logical. Even longer assessment, at 10 years, for example, is probably indicated to assess functional changes over time caused by progressive tissue fibrosis.
The fact that outcomes and QOL research are not yet exact sciences is demonstrated by highlighting some of the problems in the existing literature. Table 19–1 lists some general problems with questionnaire studies. For example, depending on how questions are phrased, it is easy to get the desired response; a fact that pollsters are well aware of. If identical probabilities are proposed to subjects but are asked in two different ways (chance of survival vs chance of death), subjects usually make different decisions, even though the probabilities are identical.28, 29 Table 19-2 lists areas of apparent consensus in the QOL literature.
HNC QOL Perspective: Where Have We Been and Where Are We Now?
Before about 1985, the major focus of solid tumor surgical oncology was to cure the patient and/or increase the length of survival. Any degree of organ resection with its accompanying physical and functional “mutilation” was acceptable, as the alternative was death. Treatment planning was based on tumor-related factors. The patient’s nonphysical status was not a priority. The literature largely consisted of retrospective cancer treatment reports of oncologic results. During the 1960s, some articles appeared on rehabilitating surgically imposed deficits, and attention focused on certain postoperative conditions, such as the “oral cripple.” The literature on QOL for HNC patients was largely nonexistent, with the exception of one contribution, which was subsequently widely misquoted.30 A holistic approach to the patient was considered the domain of general medical practitioners. For the most part, this approach failed to enter the realm of specialty training, where the focus is on technical advances. Topics such as communicating with patients,31 psychosocial considerations in patient management, and disease prevention/behavior modification,32 were largely absent from medical school curricula.
|
QOL, quality of life, HNC, head and neck cancer.
|
QOL, quality of life; HNC, head and neck cancer; HNSCC, head and neck squamous cell carcinoma.
aDirksen SR. Theoretical modeling to predict subjective well-being. West J Nurs Res 1990;12:629.
Starting in the 1980s, changes in the focus of surgical oncology were motivated by patient demands for less mutilating treatment options (e.g., lumpectomy vs radical mastectomy). Over time, this led to a function-preserving emphasis (in selected patients) such as sphincter-preserving operations for gastrointestinal and genitourinary cancers, limb-preserving sarcoma operations, and nerve-sparing prostatectomy. This functional approach to solid tumors has led to a minimally invasive strategy in general surgical oncology, as exemplified by the increased use of laparoscopic techniques for cancer resection. However, such a modus operandi has largely failed to penetrate HNC surgery, with the exception of more widespread use of selective neck dissection and transoral laser excision of laryngeal tumors by some practitioners. In HNC surgical oncology, the popularity of free-flap reconstruction during the 1990s led to the philosophy that there is no longer a need to “scrimp and save” because large organ territories can be resected and adequately reconstructed. The functional validity of these reconstructive techniques has largely been assumed, rather than documented in the literature, however. Although an organ preservation strategy has not developed in the area of HNC surgical oncology for advanced tumors, this is the focus of strategies using various combinations of chemo- and radiotherapy as the initial treatment in patients with advanced (and now less advanced) HNC, in an effort to avoid initial surgery altogether—presumably a response to widespread aversion to the traditional radical, mutilating operations for advanced HNC. The organ preservation strategies of the past 10 to 15 years, synonymous with using nonsurgical combinations of chemo- and radiotherapy in the initial treatment of HNC patients, have become (surprisingly, based on the lack of evidence-based support in their favor), a community standard of care in many parts of the United States.33 The multimodality management of HNC remains controversial in many aspects, which are beyond the scope of this chapter, but which have been reviewed in the recent literature.15 For patients who present with advanced cancer that is difficult to cure, maintaining QOL assumes greater importance.
Consistent with this new emphasis in head and neck oncology, QOL studies are now appearing in addition to the omnipresent retrospective cancer treatment reports in the HNC literature. The “state of the art” of HNC QOL is currently to report functional outcomes related to various treatments (only a small component of global QOL/well-being), as well as validating a variety of QOL instruments that assess the general health as well as the disease-specific status of these patients.
There appears to be competition between academic institutions in generating better QOL instruments for HNC patients, and there are a plethora of these.18, 19, 34–38, 67 Concomitantly, there appears to be little interest in verifying the validity of someone else’s instrument. Most published reports are still retrospective in nature, resulting in a large number of studies on small numbers of patients, further complicated by the many stages and natural histories of HNC in different sites. However, there now appears to be a realization that the proper format for QOL data collection is prospective. There is a need to select one good instrument and to apply it uniformly to the HNC population in the form of national prospective studies, in order to produce the type of data that are necessary to generate valid practice guidelines.
In the context of documenting function and relating it to various treatments, the tendency is to generate very sophisticated technical and multivariate analysis such as objective measures of vocal function and to use these data to demonstrate the advantage of one treatment over another. By now, the reader of this chapter should be able to deduce that such high-tech quantitative assessments of function may have relatively little relationship to a patient’s perceived QOL, although these measures may have intrinsic interest from a research viewpoint.
A question that arises from recent publications is what to do with anti-intuitive or contradictory results. For example, one study from the University of Chicago19