Disease Screening
Another postulate of wellness care is that the early detection of diseases, particularly cancer, will result in more effective and inexpensive treatment. The annual physical exam is an expression of this principle. Unfortunately, the annual physical has been shown to be almost worthless for most of the population.21 In place of the annual physical, there are a number of specific screening tests, mostly for cancer, that are now utilized. Certainly in the medical realm, cancer screening is considered to be an important part of any wellness program. It is axiomatic that the earlier cancer is diagnosed, the greater the chances are of curing the disease or at least of extending the survival time. The premise behind this is illustrated in Figure 1.
| A | B | C |
Undetectable Period | Detectable By Screening | Symptomatic Period |
Treatable Period | Untreatable Period | |
Figure 1. Cancer timeline if screening is going to be effective.
The top row of this figure illustrates the progression of a cancer in terms of its detectability. At some point A, a normal cell(s) undergoes malignant degeneration and the body's own defenses are unable to arrest its growth. It is at this point microscopic in size and undetectable by any means. At some point B, the malignancy, either by reaching a particular size or by causing physiologic changes, becomes detectable through a specific screening or diagnostic procedure (mammography, hemoccult, pap smear, prostate specific antigen, etc.). At some point C, the malignancy finally becomes manifest symptomatically (rectal bleeding, difficulty urinating) or through routine diagnostic exam procedures (breast palpation).
The bottom row of the illustration shows the progression of a cancer in terms of its treatability. At point A, when the tumor is small and localized, the cancer is certainly treatable if only one knew of its existence and location. The tumor remains treatable for a period of time until it metastasizes or becomes so large or invasive as to become untreatable, point C in this figure. As suggested by this illustration, there is a period, B-C, during which a cancer is detectable by those specific screening procedures, and not detectable otherwise, and during which the cancer changes from being treatable to untreatable. To the extent that these two parallel timelines reflect reality, cancer screening procedures are useful.
This illustration assumes a number of things. It assumes that there is some sort of treatment (surgery, chemotherapy, radiation, guided imagery -- something) that offers a reasonable expectation for benefiting the patient: it not a cure, at least prolonged survival. It assumes that the risks and side effects of the treatment do not outweigh the benefits. It also assumes that the screening procedure is reasonably reliable in terms of its specificity and sensitivity. (Sensitivity is a measure of a test's ability to detect all cases of a condition or disease. If a test is not sensitive it produces false negative results. Specificity is a measure of a test's ability to detect only cases of a condition or disease. A test that is not specific produces false positive results.)
When specific cancer screening programs are analyzed in these terms, all too often they fail to meet these criteria. Unfortunately, effective cancer treatments remain all too elusive. For the most common cancers, treatment efficacy ranges from not-at-all (lung) to variable (breast, colon, prostate, cervical), to readily treatable (basal and squamus cell carcinomas of the skin). It is precisely those cancers in the variable range for which screening procedures have been developed. Any generalization regarding the effectiveness of treatment for this group of cancers will necessarily be inexact, but in many cases, Figure 2 more closely reflects reality than does Figure 1. In this case, early detection offers no benefit.
| A | B | C |
Undetectable Period | Detectable By Screening | Symptomatic Period |
Treatable Period | Untreatable Period | |
Figure 2. Cancer timeline for ineffective screening program.
Perhaps an even greater problem lies with the screening tests themselves. All these cancer screening tests, to varying degrees, are very inexact in terms of their specificity and sensitivity. A false negative test result, one that fails to detect an existing cancer, is obviously of no value and may even be harmful by creating a false sense of security. A factor that is often overlooked when assessing the net value of screening tests is the false positive result, a test that identifies a cancer that is not there. At first glance this would not seem to be a terribly big problem; the screening test is positive, this is followed up by a more sophisticated test which delivers the good news -- there is no cancer -- and no harm is done.
A more accurate characterization of a false positive test is something like this: The false positive is followed by a second more invasive, more expensive, and perhaps more painful diagnostic test. Biopsy or colonoscopic exams are examples of these. Or the false positive test may lead directly to treatment which is of course unnecessary and perhaps harmful. Assuming that a biopsy is performed which then corrects the original false positive finding, the patient has to endure days or even weeks of uncertainty between these two tests, during which time the thought "I've got cancer," is a constant preoccupation. This last point in particular is often not figured into the analysis, but in fact every year hundreds of thousands of people are subjected to this anxiety because of false positive tests.
Table 1 lists the possible outcomes of a cancer screening procedure. Outcomes numbers 1-6 are clearly of no benefit to the patient, may be harmful to the patient, and add to the cost of health care. (Some might argue that outcome #1 benefits the patients. Patients receive peace of mind from knowing they do not have cancer. On the other hand, there is always a period of uncertainty from the time the test is administered to the time the results are known, a period during which the question, "Do I or don't I have cancer?" is likely to prey on one's mind. The peace of mind benefit can hardly justify the use of these tests.) Outcome #7 may or may not benefit the patient. It depends upon the relative amount of the benefit versus the side effects (7a=no net benefit; 7b=net benefit). Outcome #8 is unequivocally of benefit to the patient. The question becomes, how may 1-7a (no benefit) outcomes are there relative to the 7b and 8 (benefit) outcomes and how much does it cost to achieve the 7b and 8 outcomes?
1 | The test gives a correct negative result. |
2 | The test gives a false negative result and the cancer is missed. |
3 | The test gives a false positive result and the patient undergoes additional diagnostic tests at additional expense, discomfort, and causing significant emtional distress to the patient. |
4 | The test gives a false positive result and leads to an unnecessary therapeutic procedure at additional expense, discomfort, emotional distress, and perhaps even injury to the patient. |
5 | The test accurately detects the cancer, but the treatment does not alter the course of the disease. |
6 | The test accurately detects the cancer, but the treatment does not alter the course of the disease, and leaves the patient worse off because of disfigurement (mastectomy), side effects (nausea, hair loss) or some other functional loss (incontinence, impotence). |
7 | The test accurately detects the cancer, the treatment does alter the course of the disease, but at a cost of functional loss, side effects, or disfigurement. |
8 | The test accurately detects the cancer, treatment does alter the course of the disease and there is minimal functional loss or disfigurement. |
Table 1. Possible Cancer Screening Outcomes
Thus a screening procedure is of value if it is: a) sensitive; b) specific; c) the subsequent treatment is safe and effective. When particular screening procedures are viewed in this light, the results are often disappointing. Three tests in particular (hemoccult, mammography, and prostate specific antigen) have come under criticism because of their poor reliability and the uncertainty of treatment outcomes.
One analysis of the mammography asked what would be the likely results of screening 1,000 women for 10 years.22 It was calculated that:
- 300 women would receive accurate negative result during this decade. These women received no benefit and no harm from the screening. There would be a financial cost however (outcome #1).
- Seven women would receive false negative reports. That is, the mammography would fail to detect existing tumors (outcome #2).
- Almost 700 women would receive some sort of false positive results that would require additional diagnostic procedures. Some might even be improperly treated for breast cancer they did not have. All these women received a net harm from screening mammography (outcome #3, #4). There is also a financial cost associated with screening these women for one decade.
- Fifteen tumors would be accurately detected. Of these 15, an uncertain number would benefit form this early detection (outcome #7b, #8) and some would not (outcome #5, #6, #7a).
Even assuming an accurate detection of an existing breast tumor, it is not always clear that mammography benefits the patient. Breast cancer treatment has progressed little over the last decades and early detection of the cancer does not often translate into successful treatment of the cancer. In other words, the progression of breast cancer from treatable to untreatable more closely resembles Figure 2 than Figure 1. A much publicized Canadian study found no reduction in breast cancer mortality resulting from mammograms for women under 50.23 This finding has been confirmed elsewhere and there is now a consensus that women in this age group receive no benefit from mammograms.24,25 The National Cancer Institute has recently changed its policy and now recommends against screening mammography for women under 50.26 There is also no evidence of benefit for women over 65 years of age. This leaves women from ages 50-65 who may benefit. Mammography does seem to produce a reduction in breast cancer mortality in this age group, but it is a small reduction. This small reduction in mortality balanced against the extremely high false positive rate, has lead some to conclude that the net benefits of mammography, even in the 50-65 age group, is nil.27,28 There is considerable debate and disagreement over this conclusion, and no doubt future studies will clarify the issue, but it is clear that at best mammography offers only very marginal protection against breast cancer mortality at a relatively high cost.
The detection of fecal blood to identify colorectal cancer has been regarded as a model for cancer screening. It is cheap, safe, and the test can be self-administered by the patient. Unfortunately, the poor sensitivity and specificity of the test make it almost worthless. In one of the largest prospective studies, (n=12,312) patients over 50 years of age who were relatives of colorectal cancer patients, i.e., a group at relatively high risk, were screened for fecal blood and followed for a three-year period.29 The study showed a sensitivity of 26 percent (74 percent of cancers were not detected) and a positive predictive value of 5-8 percent (92-95 percent of test-positive patients did not have cancer). Both sensitivity and specificity would be expected to be even poorer in subjects who are not at a high risk. Not surprisingly, several controlled studies have failed to demonstrate a reduction in cancer mortality using fecal blood screening.30,31
Recent improvement in the detection of prostate cancer, primarily through the detection of prostate specific antigen (PSA), have led to the paradoxical situation where there is concern that prostate cancer is now being detected too early. Because prostate cancer is a slow-growing cancer and because it occurs in the elderly population, it is not a significant factor in reducing longevity. Relatively few men with prostate cancer, only 8.5 percent, die from prostate cancer.32 They usually outlive their cancer and die of other causes. Many older men die with prostate cancer, rather than of prostate cancer. In many of these men the prostate cancer is asymptomatic at the time of death. Because treating prostate cancer can have serious unwanted side effects such as impotence and incontinence, the decision is often made not to treat the cancer in elderly men. By detecting the cancer earlier, younger men are now being treated and many of them are suffering from those side effects. Many men, if not treated, would not have died of prostate cancer at all. Yet the quality of their remaining years has been seriously diminished by the side effects. In light of this, the recommended treatment for some prostate cancer patients is no treatment at all.33 What is the sense in screening for a condition for which no treatment is recommended?
Even as benign and inexpensive a screening procedure as the visual inspection of asymptomatic adolescents for the presence of idiopathic scoliosis has been found to be of limited or no value. In this case, the sensitivity and specificity of the screening tests are actually fairly good -- from 74-100 percent and 78-91 percent respectively. The problem is that there is no evidence that early detection results in better clinical outcomes. In fact, early detection may result in harm to the patient through excessive radiation exposure, financial cost, and diminished self-esteem.34 The US Preventive Services Task Force thus recommends against the use of routine scoliosis screening.35 Chiropractors might quarrel with this conclusion, arguing that early intervention with spinal manipulation will produce better outcomes. Unfortunately there are no data to support this argument.
These problems, the uncertainty of treatment outcomes and the poor reliability of the screening procedures, suggest caution in the use of screening procedures as a part of wellness care.
Preventive Intervention
There are a number of services offered by the medical community that are designed to directly prevent disease. Immunization is the prime example. Dentists have probably been the most successful of all health care practitioners at preventing the diseases they treat. Fluoride treatments, dental sealants, and cleaning services have dramatically reduced the incidence of tooth decay. Does chiropractic have an equivalent service to offer?
I suspect that by the time some chiropractors have read this far they are scratching their heads and muttering to themselves, "What's all this talk about cholesterol and mammograms? We're chiropractors and we keep people well through spinal adjustments." In other words, what many chiropractors mean when they talk about wellness care or prevention is not diet, or exercise, or lifestyle, but chiropractic itself. Chiropractic adjustments are viewed as a form of preventive health care. It is asserted by some that asymptomatic, disease-free people will benefit from regular preventive or maintenance adjustments. With the spine in optimal alignment and the nervous system's function thereby normalized, our resistance to disease is presumably enhanced.
The problem of course is that chiropractic-adjustments-as-prevention is an hypothesis. There is no evidence that regular adjustments prevent anything. To some in the profession the logic of adjustments-as-prevention is unassailable. To question this is to question the very premise of chiropractic. To others in the profession, most outside the profession, and certainly public health professionals, this logic is not compelling in the absence of any supporting data.
Adjustments-as-prevention is a testable hypothesis. One might compare the populations of chiropractors and their families to a comparable group that does not receive regular chiropractic adjustments, say, podiatrists and their families. If there is a health-promoting, disease-preventing effect of adjustments, there should be a measurable difference between these two groups in morbidity and mortality. If chiropractors can demonstrate some generalized health-promoting effect of regular adjustments, I suspect the profession's difficulties, political and otherwise, would vanish. But until such time as there are data to support the adjustments-as-prevention hypothesis, it is inappropriate for the profession to promote this idea or use it as a model for health.
Criteria for Wellness Care
Given the preceding discussion, I propose that if chiropractors, or any other health providers for that matter, are to institute a program of wellness care, all the following criteria need to be met:
- Risk factors or disease causing agents must be accurately identified and shown to be causally related to a particular condition. Cigarette smoking satisfies this criteria, but for most other factors there is considerable doubt and ambiguity. The experience with cholesterol and heart disease suggests that accurately identifying causally related risk factors is not easy. In the absence of compelling experimental evidence demonstrating causality it is possible to do more harm than good with prevention programs. Notes one critic of screening programs: "Iatrogenic harm to both individuals and populations may be caused by premature translation of hypotheses (concerning risk factors) into action."36
- The magnitude of the risk factor or behavior must be significant. Again, smoking certainly meets this standard. In all other cases the risk factors are substantially less significant. Additionally, for most diseases, such as coronary artery disease, there are many risk factors (some of which are not modifiable, like genetics) and altering any one factor is not likely to have a substantial effect on the incidence of the disease. For other diseases, such as prostate cancer, lowering the incidence of the disease has little effect on longevity because the condition primarily affects the elderly.
- The intervention must be effective in modifying the risk factor. Any intervention that relies on modifying a patient's behavior, particularly in the area of weight loss, smoking cessation, or diet and exercise modification, is suspect until it can demonstrate effectiveness. Non-practitioner based strategies such as the cultural changes described above are more likely to be effective.
- The risk/benefit ratio of the intervention must be acceptable. For many medication-based interventions, particularly for cholesterol and blood pressure lowering programs, there are associated risks and side effects that must be balanced against the benefits. In many cases the risks appear to outweigh the benefits. For cancer screening programs, the effects on patients of false positive test results must be factored into the net benefit of the procedure.
- There must be some rationale for the intervention to be provided by health professionals rather than by others. I take it as given that if a physician (DC or MD) provides a service it will cost more money than if someone else provides that service. That extra cost can be justified if the physicians' training or expertise adds something to that service. It's often difficult to make that case. Take for example health risk assessment evaluations. A number of companies have developed questionnaires that evaluate a person's lifestyle by asking questions about their eating, drinking, driving, smoking, and exercising habits and then calculate how many years are added or subtracted from their chronological age by their behavior. Whatever misgivings one might have about the validity of such a calculation, it is a useful way of illustrating to people the consequences of their behavior. These health risk assessment questionnaires are most often administered in the workplace by non-physicians. Typically after scoring the questionnaires, individual and group counseling sessions are held during which appropriate recommendations are made. Chiropractors certainly could, and perhaps some already do, use these instruments, but I wonder if it can be done as effectively or economically as it is currently done by others? And we are still entitled to ask whoever is administering the instrument, "Does this change behavior?"
- The intervention must be cost effective. It is taken as an article of faith by many that preventive health care will save money. The logic seems sound -- it's cheaper to prevent or at least identify earlier disease processes than it is to treat the full blown condition. Preventive health care is offered as one of the solutions to the problem of escalating health care costs. Unfortunately it doesn't work that way. Most studies done on preventive health care services indicate that they add to the net cost of health care.37 Those services may or may not be worth the extra cost, but it is an extra cost. Some preventive services clearly are cost effective. Prenatal care and well baby care, including immunizations, are relatively cheap and yield measurable improvements in outcomes. Other types of preventive services, such as cancer screening, do not perform nearly as well on this standard.
- The intervention must be effort effective (on the part of the patient). It costs little or nothing, in terms of dollars and cents, to modify behavior. There is a cost to the patient however in terms of the effort and sacrifice and the patient is entitled to ask if the potential benefits are justified. The data on cholesterol suggest that for most people a great deal of effort and sacrifice (a lifelong program of cholesterol reduction) will yield very small benefits (optimistically, days or weeks added to lifespan). It is difficult to justify such a program given this effort/benefit ratio, and patient compliance is likely to be correspondingly and appropriately poor. The same calculation should be made for any lifestyle-modification program.
- There must be an economic framework within which to provide the intervention, i.e., someone has to be willing to pay for the intervention. If wellness care is to constitute anything more than a footnote to chiropractic practice, there must be some economic incentives to provide this care. Someone, either patients themselves or their parties, have to be convinced to pay for these services. This is not entirely under the control of the providers, but chiropractors or any other practitioners interested in providing this care must be able to make the case that these services are worth paying for. It remains to be seen whether that case can be persuasively made.
References
28. Roberts M. Breast screening: time to rethink? BMJ, 1989;299:1153-55.
29. Ahlquist D, Wieand S, Moertel C, McGill D, Loprinzi C, et al. Accuracy of fecal occult blood screening for colorectal neoplasia. JAMA, 1993;269:1262-67.
30. Selby J, Friedman G, Quesenberry C, et al. A case-controlled study of screening sigmoidoscopy from colorectal cancer. N Eng J Med, 1992;326:653-657.
31. Newcomb P, Norfleet R, Storer B, et al. Screening sigmoidoscopy and colorectal cancer mortality. J Natl Cancer Inst, 1992;84:1572-1575.
32. Johansson J, Adami H, Anderson S, Bergstrom R, Holmberg L, Krusemo U. High 10-year survival rate in patients with early, untreated prostatic cancer. JAMA, 1992;267;2191-96.
33. Fleming C, Wasson J, Albertsen P, Barry M, Wennberg J. A decision analysis of alternative treatment strategies for clinically localized prostate cancer. JAMA, 1993;269:2650-2658.
34. US Preventive Services Task Force. Screening for adolescent idiopathic scoliosis. JAMA, 1993;269:2667-2672.
35. US Preventive Services Task Force. Screening for adolescent idiopathic scoliosis. Policy statement. JAMA, 1993;269:2664-2666.
36. McCormick J, Skrabanek P. Coronary heart disease is not preventable by population interventions. Lancet, 1988;Oct.8:839.
37. Russell L. The role of prevention if health reform. N Eng J Med, 1993;329:352-354.
Craig Nelson, DC,
Associate Professor
Wolf-Harris, Clinical Research Center
Northwestern College of Chiropractic