Chapter 9 Screening

Screening

After reading this chapter you will be able to:

  • Appreciate the role that physicians can play in promoting health and preventing diseases at the individual and community level (MCC 78-3)
  • Be able to describe the health impact of community-level interventions to promote health and prevent disease (78-3)
  • Apply the principles of screening and be able to evaluate the utility of a proposed screening intervention, including being able to discuss the potential for lead-time bias and length-prevalence bias (78-3)
  • Explain the goals of screening and also the limitations of screening;
  • Apply the criteria for judging whether or not to screen and so be able to evaluate the utility of a proposed screening intervention
  • Identify ethical issues with the restricting of individual freedoms and rights for the benefit of the population as a whole (78-3)

Linking these topics to the Medical Council exam objectives, especially section 78-3.

Note: The colored boxes contain optional additional information; click on the box open it and to close it again.
Words in CAPITALS are defined in the Glossary

To screen or not to screen

Paul Richards consults Dr. Rao because his back has been causing him pain recently and he wonders if he needs an X-ray. In passing, he heard about the “prostate test” on the radio and wants to know if he should get one. After taking a history and examining his back, Dr. Rao advises Paul to continue his normal activities and, if possible, to get more exercise. He tells Paul that he does not find any indication for imaging. He checks the clock and wonders if he has time to discuss prostate screening or if he should ask Paul to arrange another appointment…

Goal of Screening

Screening aims to detect disease in its early stages – before it becomes clinically manifest. As such, it is generally classified as secondary prevention (Chapter 4). In formal terms, screening is the application of a test or procedure to asymptomatic people in order to classify them as being at higher, or at average risk of developing symptomatic disease. This illustrates Bayesian thinking, which concerns how far the addition of new information (here, from the screening test) alters the clinician’s judgment of whether treatment is required. People found by screening to be at higher risk undergo further diagnostic investigation to identify those who have the disease. These are then treated to prevent the effects of the disease. Note that the difference between a screening and a diagnostic test is subtle; it can be defined in terms of difference in the risk of the disease for which the person is being investigated. In screening, the person is asymptomatic and at apparently average population risk. In diagnosis, the person has symptoms or signs indicating a higher risk of the disease being tested for.

The logic underlying screening assumes that the earlier a disease is diagnosed the better it will respond to treatment, improving survival and quality of life for patients. However, the decision of whether or not to screen must be considered carefully. Wilson and Jungner commented, “the central idea of early disease detection and treatment is essentially simple. However, the path to its successful achievement (on the one hand, bringing to treatment those with previously undetected disease, and, on the other, avoiding harm to those persons not in need of treatment) is far from simple though sometimes it may appear deceptively easy.”1, p164  First, the screening test may not be perfectly accurate: there may be false positives, or cases may be missed (see Figure 9.1); SENSITIVITY and SPECIFICITY were explained in Chapter 6. Second, not all people with early signs will develop clinical disease; this is discussed below. As screening represents an intervention on people who may not have the disease being screened for, some will not directly benefit except, perhaps, by being reassured. Therefore, the clinician has a responsibility to ensure that the risk of harm is very small. Growing awareness of the limits of screening has resulted in the elimination of several tests that were formerly routine (see Nerd’s corner box) and greater rigour in evaluating proposed screening activities.

Screening failures

Here is a list of once widespread screening tests that have been discarded because they do not meet the criteria for a useful screening test.

  • Urine dipsticks for renal disease – very poor specificity.
  • Regular electrocardiograms – poor predictor of cardiac events, and harm can arise from investigations of positive screens.
  • Homocysteine for heart disease – poor predictor of disease.
  • CA 125 for ovarian cancer – not specific and no effect on mortality.
  • Regular chest x-rays for lung cancer – not sensitive or specific.
  • Breast self-examination – no impact on mortality.
  • Self-examination for testicular cancer – positive predictive value unknown, low incidence, and good response to treatment.
  • Infrared temperature sensors at airports for SARS – temperature results unreliable as they are influenced by many other factors.
Figure 9.1 Outcomes of screening

Clinicians most commonly screen patients during routine office visits, based on guidelines for which tests are appropriate for the patient’s age and sex; examples were given in Chapter 4. Screening can also be focused on patients at high risk due to their previous medical history, taking the form of more active case finding. Screening may also be applied by public health agencies to segments of the whole population: mass screening (see the Definitions box).

Screening terminology

Routine screening is applied to patients according to their age and sex, based on evidence that the screening test accurately identifies disease and its application leads to interventions that do more good than harm. The evidence supporting each test is reviewed by the Canadian Task Force on Preventive Health Care. 2

Responding to preventive guidelines, clinicians may systematically screen their practice population, for instance by implementing an office reminder system (Chapter 8), to ensure that all their practice’s adult patients have regular blood pressure checks.

Extending this, multiphasic screening refers to administering multiple screening tests in a single checkup session.

By contrast with routine screening, clinicians may also screen opportunistically using any consultation as an opportunity to administer appropriate screening tests.

Case-finding also aims to diagnose pathological processes in the asymptomatic stage, but the population targeted is usually defined by risk factors other than demographic, for example checking cholesterol levels in family members of a patient with familial hypercholesterolemia, or checking for retinopathy in patients with diabetes mellitus. Of course, case-finding should also be carried out only when the expected benefits outweigh the risks.

Note that the term case-finding is also used:

  1. In infectious disease control when diligent efforts are made to locate and treat persons who have had close or intimate contact with a known case. This is also known as contact tracing,
  2. In epidemic control to mean seeking people who have been exposed to a risk or to potentially harmful factors and who have the disease. 3

Mass or universal screening is aimed at a demographically defined population. Examples would be the provincial newborn screening programs or the Canadian breast cancer screening program.

Criteria for Introducing a Screening Test

Given screening’s superficial attractiveness but underlying potential for wasteful, even harmful, use of resources, in 1968 the WHO published criteria for judging screening programs developed by Wilson and Jungner.1 These were updated through a consensus review in 2018.4 The criteria consider the disease characteristics and the organizational context of the screening. An abbreviated summary is given in the table below.

Consolidated criteria for introducing a screening test

  1. In relation to the disease:
  • The condition being sought should be an important health problem.
  • The natural history of the condition should be adequately understood, and there should be a recognizable latent or early symptomatic stage
  • The target population for the screening test should be clearly defined, and must be reachable.
  1. In relation to the test:
  • Screening test performance should have adequate sensitivity, specificity, and predictive value. The test should be acceptable to the population and safe.
  • The test results should be clearly interpretable, with clear cut-off scores to identify participants who should receive diagnostic testing.
  • There should be an agreed course of action for participants with positive screening scores, covering diagnostic procedures and follow-up care. This should be available and acceptable to those affected. The effects of false positive and false negative results should be minimal.
  1. In relation to the health care system:
  • There should be adequate infrastructure (financial, human resources, facilities, equipment) to allow timely access to the program.
  • All parts of the screening program should be coordinated and integrated into the broader health care system.
  • All parts of the program should be clinically, socially and ethically acceptable to the participants, to clinicians and society; informed consent must be provided, and the autonomy of participants should be protected.
  • The costs and benefits of screening should be clearly defined and acceptable and supported by high-quality evidence.
  • An economic evaluation of the program should be conducted to assess its full costs as an approach to managing the condition.
  • The screening program should have clear goals that guide program planning and evaluation, with continuing evaluation to ensure the program meets its goals.

These criteria are considered by agencies such as the Canadian Task Force on Preventive Health Care in formulating their clinical screening guidelines.

More Detail on the Criteria

The condition should be an important health problem

“Important” refers to the population burden of morbidity and mortality of the disease, its severity, including complications, and to its population incidence. Screening for the common cold would not make sense as it does little harm to the sufferer. Screening for a rare disease also runs into problems: large numbers of people will undergo the potential harms of screening in order to find a single case. The POSITIVE PREDICTIVE VALUE of a screening test is lower with rare diseases so more people who screen positive and  undergo diagnostic testing (with its attendant stress, risks and costs) will eventually be found not to have the disease after all.

To improve the program’s efficiency, screening programs may target the sections of the population most at risk to improve the positive predictive value. For instance, breast cancer screening is aimed at women from 50 to 75 years old even though breast cancer can occur in both sexes and much earlier.

When screening is not worthwhile

Chinese women in Hong Kong have a much lower incidence of breast cancer than women in western countries; therefore, a mammography program there would currently bring little benefit.5 However, because Hong Kong lifestyles are changing, screening may become beneficial in the future.6

The Canadian Task Force recommendations for cervical cancer screening do not currently start until age 25, since below that age cancer is extremely rare, and the minor changes related to Human Papillomavirus that can generate abnormal test results commonly regress spontaneously.

The natural history of the condition, including development from latent to manifest disease, should be adequately understood

Chapter 1 presented a simplified idea of natural history as though disease progresses steadily. But this is not always the case and diseases may regress because of the patient’s natural defences. And some conditions may be indolent and will never progress enough to harm the patient. A challenge in prostate screening lies in distinguishing rapidly progressing cancers that need immediate treatment from slowly progressing ones that will likely never need treatment (see the Illustration box). These variations may distort screening results in a way termed ‘length bias’.

Screening and overdiagnosis

About two thirds of cases of mild cervical dysplasia regress to normal, whereas progression from mild to severe dysplasia or worse occurs at a rate of about 1% per year.7 Slow progression may mean that a patient could die from other causes before succumbing to the cervical cancer detected by the screening.

The old adage that “more men die with prostate cancer than die of prostate cancer” is true. The prostate specific antigen (PSA) test may only be about 40% sensitive but can still be useful in detecting early cases of prostate cancer if a trend in scores over time is used.8 However, autopsy series find that approximately 30%–40% of men over the age of fifty have prostate cancer—a proportion that rises with age—yet only about 3% of men die from it. The best way to manage early prostate cancer is still uncertain.

Studies of breast cancer screening have shown that up to 30% more breast cancers are diagnosed in the screened group than in the unscreened group, probably due to the detection of early cancers that would never otherwise have become apparent.9

Length bias

Screening programs should be evaluated, but variability in the natural history of a disease complicates this evaluation. Slowly progressing variants of a disease remain in a pre-symptomatic (but screen-detectable) stage for longer than rapidly progressing variants of the same disease. Because of their longer duration, intermittent screening is more likely to detect a disproportionate number of slow-progressing variants, and these have longer survival after the eventual appearance of symptoms. If the program is then evaluated based on survival time in this biased sample a falsely positive conclusion will be reached: see Figure 9.2.

Figure 9.2: Length bias. In this population there are equal numbers of slow (blue) and rapidly-progressing (red) cases. But the screening test will identify five slower cases and only two rapidly progressing ones. Calculating mean survival from these seven cases will give an impression of longer average survival than occurs in the population.
Figure 9.2: Length bias. In this population there are equal numbers of slow (blue)
and rapidly-progressing (red) cases. But the screening test will identify five slower cases
and only two rapidly progressing ones. Calculating mean survival from these seven cases
will give an impression of longer average survival than occurs in the population.

There should be a recognizable latent or early symptomatic stage

If screening is to detect disease in asymptomatic people then there must be an adequate time between the stage when it is detectable by screening and the stage when it causes symptoms. Successful early identification of rapidly progressing disease would require frequent screening which could be unacceptable for the health system and the target population. Diseases with long latent periods are more amenable to early detection.

Pre-disease

The term “pre-disease” has been used to describe the early stage or latent phase when a disease is amenable to detection by screening.10 An example would be nuclear dysplasia that may eventually become a cancer. In principle, intervening in the pre-disease phase should arrest the development of disease. Removing a pre-cancerous lesion should prevent the onset of cancer, although many pre-cancers regress without any intervention.

Other examples of pre-disease are “pre-diabetes” and “pre-hypertension”. At the time of writing, diabetes is diagnosed at a fasting blood glucose of 7 mmol/l or more (or > 11.1 mmol/l two hours post prandial or HbA1c of 8.5%).11 This level is generally below the level at which symptoms (polydipsia, polyuria and weight loss) occur and far below the level for diabetic ketoacidosis. It is chosen because population-level studies show that it is the threshold for an increased risk of cardiovascular events and microvascular disease.12 So, in practice, we are treating diabetes in order to prevent cardiovascular and microvascular disease; we do not affect the progress of the diabetes itself. There is little than can be done to delay or reduce the onset of type 1 diabetes. In type 2 diabetes, some authors point out that treating prediabetes may delay the clinical onset by 3 to 4 years.13 However, not all untreated pre-diabetics progress to diabetes. Nonetheless, impaired glucose tolerance should be the trigger to modifying lifestyles to reduce weight and increase physical activity, and even to using pharmacotherapy.11 Hence, for some in this debate, pre-diabetes and diabetes merit much the same treatment.

Essential hypertension is another condition in which defining levels on a continuous scale as thresholds between pre-disease and disease may lead to over treatment. Hypertension manifests itself only by its effect on end organs, increasing the risk of renal damage, heart failure and stroke. We try to prevent end organ damage using lifestyle counselling and antihypertensives to reduce the blood pressure. Patients on antihypertensive medications are likely to experience side effects, so may feel worse with treatment than without it.

Diagnosing pre-diabetes and pre-hypertension may lead to treating risk factors for these conditions, which are themselves risk factors for microvascular, cardiovascular, renal and cerebrovascular disease, and yet there is little evidence for improved eventual outcomes for those diseases.10

Lead-time bias

Long latent periods may also create an error in evaluating screening known as lead-time bias. This arises from the delay between when disease is detectable by screening and when it produces symptoms and would be diagnosed without screening. Screening leads to an earlier diagnosis, as shown in Figure 9.3, so the patient lives longer with the diagnosis, but (if treatment does not prolong survival) may not live longer overall. Hence, evaluating a screening program in terms of survival time after diagnosis may give a falsely positive impression of success. Controlled trials of screening are the only way of eliminating lead-time bias.

Figure 9.3: Lead time bias
Figure 9.3: Lead time bias

A final problem in evaluating screening programs is that people who readily adopt new, supposedly healthy behaviours tend to be healthier than those who do not. Observational studies that compare people who chose to get screened with others who did not, generally have a greater proportion of this healthier group in the screened group; they therefore tend to show that people who opted for screening have better outcomes than those who did not. For instance, women who have Pap smears are less likely to die from cervical cancer. However, they also tend to be better informed and wealthier and, therefore, to have a lower base-line risk of dying from cervical cancer than those who don’t get Pap smears. Even after controlling for stage at diagnosis, women in higher socio-economic groups have a better prognosis than those in lower socio-economic groups. It is difficult to determine how much the favourable outcome in women who get screened is due to the test and treatment, and how much is due to their other favourable health determinants.

There should be an accepted treatment for patients with recognized disease

If nothing can be done to reduce the impact of a disease, screening increases morbidity by lengthening the time the patient is ill. The asymptomatic person becomes a patient, assumes the “sick role” and is subject to a longer period of being stressed by a serious diagnosis. Furthermore, an effective treatment is available, it must also be acceptable to the patient. For example, it is important to discuss the implications of transitioning a patient with Type 2 diabetes to pharmacological management which may either represent a major loss of dignity for them, or (conversely) be welcomed as freeing them from the need for limiting their diet.

We linked screening to secondary prevention, but when there is an effective treatment, screening can sometimes be an initiator of primary prevention. For instance, while treatment of HIV will not always eliminate the infection, screening followed by counselling and treatment may reduce HIV transmission to others by modifying the patient’s behaviour and by reducing their infectivity by reducing the viral load.

The test

There should be a suitable test or examination

In principle, a screening test should be simple, quick, inexpensive, reliable and valid. In reality not all tests measure up. For instance, sigmoidoscopy and colonoscopy for colon cancer, mammography for breast cancer and cytology for cervical cancer require skilled professionals and sophisticated equipment, so can be costly. Although blood pressure is ostensibly simple to record, a number of factors can distort its measurement, so most patients require more than one instance of elevated blood pressure to be diagnosed with hypertension (see Chapter 6). Prostate specific antigen (PSA) is a marker of prostate cancer, but it can be elevated in benign prostatic disease. Furthermore, it cannot distinguish between cancers that will eventually kill and those that progress too slowly to cause problems. A valid test is one that tells us what we want to know. In effect, although PSA is simple, quick, and reliable, it is not a valid indicator of cancer that requires treatment.

The test should be acceptable to the population

Screening acceptability is largely a practical matter: is it going to be painful, inconvenient, or embarrassing? For example, there have been significant increases in gonorrhea screening after the development of a urine test (compared to the previous swab). Although promoting screening can increase participation, there are ethical considerations. Patient autonomy may be compromised by the way that information is given. For instance, most information about breast cancer screening stresses its population benefit rather than the benefit to the individual. Recalling the ways of summarizing risks from Chapter 5, relative statements such as “Mammography screening will reduce the risk of death from breast cancer by 30%” are commonly used. This sounds impressive but may not help an individual patient who would benefit more from an absolute figure, such as: “Mammography screening will reduce your risk of death from breast cancer from 4% to 3%.” As Rose noted (Chapter 8), although the benefit for a population as a whole may be considerable, the benefit for the individual can be small.

Sending patients letters or other reminders improves screening uptake. However, if patients perceive the reminders as having more authority than they do they may abdicate their right to informed consent.14 Even careful clinicians can have difficulty ensuring that patients have a realistic view of the risks and benefits of screening so that they can give properly informed consent (see Chapter 10 on communicating risks). Finally, deprived, vulnerable people are generally at greater risk of disease. They are also less likely to respond to promotion of screening than those with greater health information who are at lower risk. Screening activities that do not ensure equitable access can increase health inequities.

The health care system

There should be an agreed policy on whom to treat as patients

Few indicators of pre-clinical disease are dichotomous and differentiate sharply between those who will develop the condition and those who will not. Indeed progress in medical imaging and screening techniques means that disease can be identified increasingly early. Unfortunately, the natural history of these very early lesions is poorly understood. For instance, in cancer screening, it can be very difficult to distinguish between inflammation and early neoplasia. Although some findings on imaging or cytology can be definitively categorized as neoplastic or as benign, others are intermediate. Even experienced readers may have difficulty interpreting these. For interval-scale measures such as blood sugar, cholesterol, or blood pressure a cutting-point must be set to define a threshold for intervention, as described in Chapter 6. If the cut-off point is set too far towards the high risk end of the scale, the screening is likely to be ineffective because it will miss many patients who could potentially have benefitted from treatment. If the point is too far towards the low risk end of the scale, the costs of investigating and reassuring large numbers of patients at low risk will outweigh the benefits of evaluating and treating high-risk patients.

Because it is often hard to detect disease in its early stages and bearing in mind the potentially damaging consequences of an erroneous test result, only tests with excellent sensitivity and specificity should be used for screening.

Facilities for diagnosis and treatment should be available

Evidently, if waiting lists are long or if facilities for follow up and treatment are not conveniently accessible, the screening program should be reconsidered. Harm occurs when asymptomatic people are labelled as possibly having a disease but further investigation cannot be carried out. Harm also occurs when they are diagnosed and no treatment offered. In developing the Canadian breast cancer screening system, provinces had to ensure that diagnostic facilities were prepared to deal with the added workload produced by systematic screening. Primary care providers should consider the impact of their screening programs on the local health system as well as the availability of specialist services.

The cost of case-finding (screening and the resulting diagnosis) and the treatment of the found cases should be economically balanced in relation to possible expenditure on medical care as a whole.

Economic costs

The cost of a single test may be low but the cost of screening a whole population can be considerable. Patient management systems are needed for call, recall and follow up of results; systems are needed to ensure that screening equipment is maintained and that screening staff are adequately trained, and patient safety systems should be in place to reduce the risk of medical error. Systematic evaluation of screening processes and outcomes should also be in place.

The cost of investigating people with screen-detected abnormalities, many of which are false positives, and the cost of treating the true positive tests have also to be considered.  One comparison of the costs of breast cancer treatment with and without screening found that screening reduced the cost of treating breast cancer by 21%. At the same time, screening increased overall health care costs, including those not related to breast cancer, by nearly 6%.15  Clinicians should recognize that spending on a procedure that has not been shown to be beneficial reduces the resources available for effective procedures.


Other costs – Beneficence and non-maleficence

While Wilson and Jungner mention only economic costs, the ethical clinician needs to balance all likely costs against all likely benefits.  Apart from the cost to the healthcare system, screening incurs physical, mental, and financial costs to patients and their families, as summarized in Table 9.1.

The clinician who proposes interventions to healthy patients, particularly if they have not asked for them, has an added responsibility to ensure the benefit is likely to outweigh the harm. Screening often has little or no impact on an individual’s risk from the disease being screened for, while the test and follow-up procedures carry an inherent, although generally minimal, risk of harm.

Patients also ask for screening tests, such as total body scans, an annual physical examination or a PSA test, for which the evidence does not support their use in most populations. The physician must balance his duty not to harm against his duty in relation to patient autonomy of choice. There is no ready answer to this dilemma: physicians should educate their patients and negotiate with them bearing these concerns in mind.

Table 9.1: The physical and emotional costs of screening

Costs due to: Examples
The test itself Colonoscopy carries a small risk of bowel perforation; mammography exposes the breast to radiation, albeit minimal.
Procedures engendered by false positive results Diagnostic and treatment procedures carry risk. In the case of treatment for putative cancers, these can be severe. About 10% of women who undergo breast cancer screening for the first time and about 6% of women getting subsequent screening undergo further testing only to find that the screen result was falsely positive.16
Unnecessary anxiety after false positive results Most patients experience anxiety on being told of a positive screening result. Some patients take considerable time to get over the anxiety. As many as 5% of Pap tests in young women are false positives.
False reassurance after negative results After having a negative screen, whether a true or a false negative, some patients ignore subsequent symptoms of the disease.
Imposition of difficult choices Men over 65 found on screening to have an asymptomatic abdominal aortic aneurysm have a choice of either undergoing a very risky operation immediately or face a 70% risk of death within the next five years.
Prolongation of the period of illness As their disease is detected earlier, patients who are screened live longer with the diagnosis. In some cases, the consequent anxiety and changes to lifestyle reduce the quality of life.
Labelling If the disease is viewed negatively, the patient runs the risk of being viewed, or viewing himself, prejudicially. Consequences are both social and psychological. They can also be financial if the patient loses his job or can’t find insurance after being diagnosed.

 Screening should be a continuing process and not a “once and for all” project.

The screening interval (time between screens) should be based on the natural history of the disease – how long between the onset of screen detectable disease and irreversible damage – and the sensitivity of the test – at what stage can the test detect the disease. Too long an interval means that many cases will become symptomatic in the intervals between screens, so the benefit of early detection is lost. Too short an interval can reduce the PPV as well as increasing the number of tests required to detect a case. Annual screening is often recommended, not from evidence, but for convenience. For instance, some health professionals still recommend annual Pap smears even though the best outcomes for cervical screening are from Finland and the Netherlands, where the screening interval is five years but the program reaches a very high proportion of the population. The common practice of screening compliant women every year ties up resources and turns attention away from groups most at risk from invasive cervical cancer and who are least likely to be screened, such as women living in rural areas, Aboriginal women, recent immigrants, older women, and those with lower incomes. This is an example of the inverse care law.

The inverse care law

An unfortunate inverse relationship between need and demand was termed “the inverse care law” by Julian Tudor Hart: those at highest need are the least likely to receive care, and particularly preventive care.17

An unintended consequence of interventions is often that their initial uptake occurs among groups such as opinion leaders, informed people and the “worried well” so that their health improves, widening the disparities in health between them and other, less informed people. If subsequent population health efforts succeed in increasing uptake by specifically targeting the less informed people, the gap can narrow again. However, innovations are always being proposed, so the health gap produced by delayed uptake of innovations persists. A Public Health Agency of Canada report presented a diagram illustrating how delayed adoption of innovations can generate socioeconomic disparities in health (see page 9, Figure 3 of the report).

Paul’s prostate

Having started to discuss prostate cancer screening, Dr. Rao realized that Paul has deep-seated fears and false beliefs on the subject. Dr. Rao asked Paul to drop into Nurse Jennings’ office to get fact sheets on prostate cancer and prostate cancer screening. He also asked Paul to make another appointment when Dr. Rao would have more time to continue the discussion.

WHO synthesis of screening criteria18

Since the publication of the Wilson and Jungner criteria in 1968, some formal, organized screening programs have been introduced, and others have been considered. Opinions on screening criteria have also progressed; the criteria now include the notion of screening programs, informed choice for patients, equity in application of the program, and evidence of costs and benefits, and of effectiveness of the program.

  1. The screening program should respond to a recognized need.
  2. The objectives of screening should be defined at the outset.
  3. There should be a defined target population.
  4. There should be scientific evidence of screening program effectiveness.
  5. The program should integrate education, testing, clinical services, and program management.
  6. There should be quality assurance, with mechanisms to minimize potential risks of screening.
  7. The program should ensure informed choice, confidentiality, and respect for autonomy.
  8. The program should promote equity and access to screening for the entire target population.
  9. Program evaluation should be planned from the outset.
  10. The overall benefits of screening should outweigh the harm

Self-test questions

1. One of your colleagues has just read an article that shows that screening for orange disease using the marmalade test can prolong survival of patients with the disease. What information do you need before deciding whether or not to support this conclusion?

You need to know what kind of evidence is presented in the article. An observational cohort study that showed that patients who were screened had better results that those who were not is subject to bias due to self-selection; patients who choose to have a test generally do better than those who don’t because they tend to also have other more positive health determinants. A case series that showed that patients presenting after screening had a longer survival time than those presenting when the disease is clinically apparent is subject to lead-time and length biases; screening picks up cases earlier on in the natural history of disease with the result that, as the time of diagnosis is advanced, the period during which the patient is known to have the disease is prolonged. This makes it appear that survival has been prolonged.  Furthermore, screening favours the diagnosis of indolent forms of disease, so cases picked up by screening are likely to be those which progress slowly.Because of such problems the only acceptable evidence for screening comes from a randomised controlled trial. High level evidence in screening is particularly important because screening always prolongs the period of illness and often brings psychological, social and financial problems to people who would otherwise have enjoyed apparently good general health for a longer period.

If you are satisfied that the article describes a valid controlled trial which shows that screening results in a reasonably large clinical benefit, you then need to check a few other factors before forming an opinion. The main one concerns how much work has already been done on the subject and is if the article agrees with it. It is generally rash to accept findings that disagree with those of other valid studies. It is also rash to act on the basis of a single trial.

You would also like to know background information about orange disease, about its frequency, its likely outcomes and the effectiveness of treatment options. Is the natural history known, and is early treatment more effective than treating at a later stage? Your assessment includes consideration of both length and quality of life, as well as the risks of treatment. Secondly, is a screening test possible – is there a preclinical phase that lends itself to early diagnosis?

Once you are satisfied that the evidence from a number of sources is solid, then you need to look at issues in implementation. (See the WHO criteria) You would like to know about costs of screening and of treatment for the patient and for the health care system, and whether or not you have at your disposal the means of carrying out the test, diagnosing patients who are found to be at high risk of orange disease and the means of treating those confirmed to have orange disease. You would like to know about the performance characteristics of the serum marmalade test: its positive and negative predictive values in the general population. You would also like to know whether or not patients find it acceptable.

Reflection Questions

1.   Describe the breast cancer screening program in your province.
2.   What are your provincial guidelines for prostate cancer screening?
3.   At what level of glycemia or blood sugar do your professors recommend action should be taken in healthy patients?

Misc

References

  1. Wilson JMG, Jungner G. Principles and practice of screening for disease. Geneva: World Health Organization, Public Health Papers No. 34; 1968.
  2. Canadian Task Force on Preventive Health Care. About the CTFPHC 2016 [Available from: http://canadiantaskforce.ca/about-us/about-the-ctfphc/.]
  3. Porta M, editor. A dictionary of epidemiology. New York (NY): Oxford University Press; 2008.
  4. Dobrow MJ, Hagens V, Chafe R. Consolidated principles for screening based on a systematic review and consensus process. CMAJ. 2018;190(14):E422-E9.
  5. Leung GM, Lam TH, Thach TQ, Hedley AJ. Will screening mammography in the East do more harm than good? American Journal of Public Health. 2002;92(11):1841-6.
  6. Leung GM, Thach TQ, Chan E, Foo W, Meng O, Fielding R, et al. Short-term, medium-term, long-term, and lifetime risks of developing and dying of breast carcinoma in a Westernized Chinese population: Evidence from Hong Kong between 1976 and 2000. Cancer. 2005;103(3):501-8.
  7. Holowaty P, Miller AB, Rohan T, To T. Natural history of dysplasia of the uterine cervix. Journal of the National Cancer Institute. 1999;91(3):252-8.
  8. Thompson IM, Ankerst DP. Prostate-specific antigen in the early detection of prostate cancer. Canadian Medical Association Journal. 2007;176(13):1853-58.
  9. Morrison AS, Brisson J, Khalid N. Breast cancer incidence and mortality in the breast cancer detection demonstration project. Journal of the National Cancer Institute. 1988;80(19):1540-7.
  10. Viera AJ. Predisease: when does it make sense? Epidemiol Rev. 2011;33:122-34.
  11. Canadian Diabetes Association. Clinical practice guidelines 2013 [Available from: http://guidelines.diabetes.ca/?_ga=1.41276011.839397301.1478810223.]
  12. World Health Organization. Definition and diagnosis of diabetes mellitus and intermediate hyperglycaemia 2006 [cited 2016 November]. Available from: http://www.who.int/diabetes/publications/diagnosis_diabetes2006/en/.]
  13. Yudkin JS, Montori MV. The epidemic of pre-diabetes: the medicine and the politics. BMJ. 2014;349:g4485.
  14. Osterlie W, Solbjor M, Skolbekken JA, Hofvind S, Saetnan AR, Forsmo S. Challenges of informed choice in organised screening. Journal of medical ethics. 2008;34(9):e5.
  15. Johnston K. Modelling the future costs of breast screening. European Journal of Cancer. 2001;37(14):1752-8.
  16. Health Canada. Organized Breast Cancer Screening Programs in Canada, 1999 and 2000 Report. Minister of Public Works and Government Services Canada; 2003. p. 60.
  17. Hart JT. The inverse care law. Lancet. 1971;1(7696):405-12.
  18. Andermann A, Blancquaert I, Beauchamp S, Dery V. Revisiting Wilson and Jungner in the genomic age: a review of screening criteria over the past 40 years. Bulletin of the World Health Organization. 2008;86(4):317-9.

Français (French)