By A. Umbrak. Ripon College.
Logistic analysis This is a more general approach to measuring outcomes than using frequency tables moduretic 50 mg with mastercard blood pressure normal limit. Logistic regression estimates the probability of an outcome based on one or more risk factors cheap moduretic 50mg visa prehypertension readings. Results of logistic regression analysis are often reported as the odds ratio, relative risk, or hazard ratio. For one independent variable of interval-type data and relative risk, this method calculates how much of an increase in the risk of the outcome occurs for each incremental increase in the exposure to the risk fac- tor. An example of this would answer the question “how much additional risk of 364 Essential Evidence-Based Medicine stroke will occur for each increase of 10 mm Hg in systolic blood pressure? For multiple variables, is there some combination of risk factors that will bet- ter predict an outcome than one risk factor alone? The identiﬁcation of signiﬁcant risk factors can be done using multiple regressions or stepwise regression analyses as we discussed in Chapter 29 on clinical prediction rules. Survival analysis In the real world the ultimate outcome is often not known and could be dead as opposed to “so far, so good” or not dead yet. It would be difﬁcult to justify waiting until all patients in a study die so that survival in two treatment or risk groups can be compared. Besides, another common problem with comparing survival between groups occurs in trying to determine what to do with patients who are doing ﬁne but die of an incident unrelated to their medical problem such as death in a motor-vehicle accident of a patent who had a bypass graft 15 years earlier. This will alter the information used in the analysis of time to occlusion with two different types of bypasses. Finally, how should the study handle the patient who simply moves away and is lost to follow-up? The data con- sist of a time interval and a dichotomous variable indicating status, either failure (dead, graft occluded, etc. In the latter case, the patient may still be alive, have died but not from the disease of interest, or been alive when last seen but could not be located again. Early diagnosis may automatically confer longer survival if the time of diagnosis is the start time. This is also called lead-time bias, as discussed in Chapter 28, and is a common problem with screening tests. Censoring bias occurs when one of the treatment groups is more likely to be censored than the other. A survival analysis initially assumes that any patient censoring is independent of the outcome. Survival curves The distribution of survival times is most often displayed as a survivor function, also called a survival curve. It is important to note that “surviving” may indicate things other Survival analysis and studies of prognosis 365 9 x 9 x 8 O 8 O 7 x 7 x 6 6 5 x 5 x 4 4 3 O 3 O 2 x 2 x 1 x 1 x 1970 1975 1977 1980 t=0 t = 5 years Fig. Patient 1 lived longer than everyone except patient 4, although it appears that patient 1 didn’t live so long, since their previous survival (pre-1975) does not count in the analysis. We don’t know how long patient 4 will live since he or she is still alive at the end of the observation period and their data are censored at t = 5 years. Two other patients (3 and 8) are lost to follow-up, and their data are censored early (o). These curves can be deceptive since the number of individuals represented by the curve decreases as time increases. It is key that a statistical analysis is applied at several times to the results of the curves. The actuarial-life-table method measures the length of time from the moment the patient is entered into the study until failure occurs. The product-limit method is a graphic representation of the actuarial-life-table method and is also known as the Kaplan–Meier method. The analysis looks at the period of time, the month or year since the subject entered the study, in which the outcome of interest occurred. There are several tests of equality of these survivor functions or curves that are commonly performed. The Cox proportional-hazard model uses interval data as the inde- pendent variable determining how much the odds of survival are altered by each unit of change in the independent variable. This answers the question of how much the risk of stroke is increased with each increase of 10 mm Hg in mean arterial blood pressure. Further discussion of survival curves and outcome anal- ysis is beyond the scope of this book.
Interestingly order 50mg moduretic otc prehypertension kidney disease, the per- spective of the analysis can be the patient moduretic 50 mg lowest price blood pressure diastolic low, the payor, or society as a whole. Overall, patient or societal outcomes ultimately determine the usefulness of a test as a screening tool. Bertrand Russell (1872–1970): The Philosophy of Logical Atomism, 1924 Learning objectives In this chapter you will learn: r the characteristics and deﬁnitions of normal and abnormal diagnostic test results r how to deﬁne, calculate, and interpret likelihood ratios r the process by which diagnostic decisions are modiﬁed in medicine and the use of likelihood ratios to choose the most appropriate test for a given purpose r how to deﬁne, calculate, and use sensitivity and speciﬁcity r how sensitivity and speciﬁcity relate to positive and negative likelihood ratios r the process by which sensitivity and speciﬁcity can be used to make diag- nostic decisions in medicine and how to choose the most appropriate test for a given purpose In this chapter, we will be talking about the utility of a diagnostic test. This is a mathematical expression of the ability of a test to ﬁnd persons with disease or exclude persons without disease. These are the likelihood ratios and the prevalence of disease in the target population. Additional test characteristics that will be introduced are the sensi- tivity and speciﬁcity. These factors will tell the user how useful the test will be in the clinical setting. Using a test without knowing these characteristics will result in problems that include missing correct diagnoses, over-ordering tests, increas- ing health-care costs, reducing trust in physicians, and increasing discomfort 249 250 Essential Evidence-Based Medicine and side effects for the patient. Once one understands these properties of diag- nostic tests, one will be able to determine when to best order them. The indications for ordering a diagnostic test can be distilled into two simple rules. They are: (1) When the characteristics of that test give it validity in the clinical setting. Will that result help in correctly identifying a diseased patient from one without disease? What will a positive or negative test result tell me about this patient that I don’t already know and that I need to know? If the test that is being considered does not fall into one of these categories, it should not be done! Diagnostic tests are a way of obtaining information that provides a basis for revis- ing disease probabilities. When a patient presents with a clinical problem, one ﬁrst creates a differential diagnosis. One attempts to reduce the number of dis- eases on this list by ordering diagnostic tests. Ideally, each test will either rule in or rule out one or more of the diseases on the differential diagnosis list. Diseases which are common, have serious sequelae such as death or disability, or can be easily treated are usually the ones which must initially be ruled in or out. We rule in disease when a positive test for that disease increases the probability of disease, making its presence so likely that we would treat the patient for that disease. This should also make all the other diseases on the differential diagnosis list so unlikely that we would no longer consider them as possible explanations for the patient’s complaints. We rule out disease when a negative test for that dis- ease reduces the probability of that disease, making it so unlikely that we would no longer look for evidence that our patient had that disease. After setting up a list of possible diseases, we can assign a pretest probabil- ity to each disease on the differential. This is the estimated likelihood of disease in the particular patient before any testing is done. As we discussed earlier, it is based on the history and physical examination as well as on the prevalence of the disease in the population. This is the estimated likelihood of the disease in a patient after testing is done. A positive test tends to rule in the disease while a negative test tends to rule out the disease. However, the test can be an item of history, part of the physical examination, a laboratory test, a diagnostic x-ray, or any other diagnostic maneuver. Mathematically, the pretest probability of the disease is modiﬁed by the appli- cation of a diagnostic test to yield a post-test probability of the disease. Likelihood ratios are stable characteristics of a diagnostic test and give the strength of that test. The likelihood ratio can be used to revise dis- ease probabilities using a form of Bayes’ theorem (Fig. Before fully looking at likelihood ratios, it is useful to look at the deﬁnitions of normality in diagnostic tests.
Movement of the spine is restricted in all planes and a limitation of chest expansion may occur 50mg moduretic mastercard arteria spinalis anterior. Acute anterior uveitis moduretic 50 mg on-line hypertension first aid, aortic regurgitation and (spondyloarthropathies) apical lung ﬁbrosis are known extra-articular features. Ankylosing spondylitis Deﬁnition Ankylosing spondylitis is a chronic inﬂammatory arthri- tis predominantly affecting the axial skeleton, causing pain and progressive stiffness. Chapter 8: Seronegative arthritides (spondyloarthropathies) 363 Complications Age Spinal fractures may occur with minimal trauma due to Peak incidence age: 30–50 years. Pathophysiology r Patients should be encouraged to remain active, avoid Synovitis is histologically the same as that of rheumatoid prolonged bed rest and avoid lumbar supports. Phys- arthritis, although bone resorption is sometimes promi- iotherapy involvement is important. Itislikelythatboththeskinlesionsandthearthritis r Pain and morning stiffness are treated with non- are immunologically mediated. Fivepatternsofarthritis osteotomy may be helpful in patients with severe cur- are seen: vature. There is a wide range of severity: In over 85% there is 3 Symmetrical rheumatoid-like polyarthritis. Psoriatic arthritis Investigations Deﬁnition r Blood tests may show raised inﬂammatory markers, Achronic inﬂammatory arthritis occurring with psori- anaemia of chronic disease and presence of autoanti- asis. Other features include 1% of population have psoriasis of which 5% will get periostitis, bone resorption, sacroiliitis and spondyli- arthritis. Second line agents include methotrex- Typically there is an abrupt onset of asymmetrical lower ate and ciclosporin. Achilles ten- have been shown to be effective in reducing the pro- dinitis and plantar fasciitis may also occur. Surgical intervention may have been preceded by a clinical urethritis, prostatitis, prove necessary. Prognosis It is not clear whether any medical intervention has Investigations disease-modifying potential. X-rays are initially normal but may show erosions and features Reactive arthritis similar to ankylosing spondylitis. Deﬁnition Management Acute or chronic synovitis that occurs less than 6 weeks Although unlikely to affect the course of arthritis, an- following infections with various organisms, including tibiotics are given for ongoing urethritis. Ophthalmol- Chlamydia, Yersinia, Salmonella, Shigella and Campy- ogy referral is essential for uveitis and the arthritis is lobacter species. Reiter’s syndrome is a form of reactive usually managed with nonsteroidal anti-inﬂammatory arthritis with the triad of arthritis, uveitis, and urethritis. The few patients who develop a chronic arthritis are treated as for rheumatoid arthritis. Deﬁnition An enteropathic arthritis, sacroiliitis, ankylosing Sex spondylitis or rarely hypertrophic osteoarthritis in as- M > F sociation with ulcerative colitis or Crohn’s disease. Sex 1:1 Pathophysiology In early synovitis there is intense hyperaemia with in- Aetiology ﬂammatory inﬁltration. The arthritis is said to be ster- The aetiology is unknown but the synovitis may occur in ile as bacteria cannot be cultured from joints; however, response to bacterial antigens that have passed through Chapter 8: Connective tissue disorders 365 thedamagedgutwall. Enteropathicarthritisisaseroneg- into a number of chromosomal loci in relation ative non-erosive synovitis. Intra-articular creased cytotoxic T-cell reactions, increased helper steroid injections may be of value. Connective tissue disorders It is thought that these defects may trigger a cascade of events resulting in the production of autoantibod- Systemic lupus erythematosus ies. Prevalence Pathophysiology 40 per 100,000 in United Kingdom, wide geographic The mechanism by which the aetiological factors inter- variation (1:250 American black women). Clinical features Sex Systemic lupus erythematosus is a multisystem disor- 9F : 1M der affecting skin, joints, kidneys, lungs, nervous system, mucous membranes and other organs. Systemic symptoms include general malaise, Aetiology fever(sometimeshighandswinging)anddepression(see r Genetics: Up to 60% concordance in monozygotic Fig.
Surveillance approaches Passive or ‘scanning’ disease surveillance: this involves examination of only clinically affected individuals cheap moduretic 50 mg with amex blood pressure chart sleeping, with no special effort being made to ‘seek out’ infected or diseased cases moduretic 50mg discount hypertension journals ranking. This may involve the routine gathering of information on disease incidents from the general public, medical or veterinary professionals and laboratories dealing with routine cases. Passive surveillance may lead to significant under-reporting of diseases and should, therefore, be supplemented by active disease surveillance particularly for important animal diseases. Active disease surveillance: this involves proactive examination of individuals to actively seek out infection or disease, and targeted searching for evidence of disease in populations. Programmes may be broad-scale to capture any significant disease occurrences, targeted against specific high-threat diseases (e. International trade may also guide surveillance schemes to establish national and regional disease status, especially where it relates to public health and economic initiatives. For livestock diseases which are spread by the movement of infected animals, areas where animals are moving should be targeted for surveillance (e. The speed of information flow between different components of the disease surveillance system (immediate or routine). The rapidity of response required: immediate investigation of disease incidence or routine and regular analysis of data with subsequent adjustments to control activities when required. For a disease surveillance strategy to act as an early warning system, reporting, decision-making and response must be rapid. However, for endemic diseases, it may be more appropriate to evaluate the routine data collected to adjust or target control activities. National surveillance systems should include an integral approach and accommodate all needs. It may beIt may be possible to link and integrate severalpossible to link and integrate several different surveillance systems. The following functions may supporttions may support surveillance systems: setting of standards (e. The key components of aThe key components of a surveillance and monitoring system. The following tasks are recommended for improving animal disease surveillance:The following tasks are recommended for improving animal disease surveillance: 1. Identify key stakeholders and organisations relevant toIdentify key stakeholders and organisations relevant to the site state or local veterinarian or animal health officer (will most likely be lead person inveterinarian or animal health officer (will most likely be lead person inveterinarian or animal health officer (will most likely be lead person in regional surveillance effort)surveillance effort) public health contact veterinary diagnostic laboratoriesic laboratories. Identify relevant animal diseases for the siteIdentify relevant animal diseases for the site notifiable animal diseasesdiseases wildlife animal diseases zoonoses. Familiarisation with country responses with reference to potential disease outbreaks at the site. Establish standardised report forms for disease surveillance including definitions such as “confirmed” and “suspected”. Identify and collaborate with ongoing animal disease surveillance efforts at other wetland sites and government Ministries or Departments e. Identify efficient and effective communication channels with the relevant health authorities and laboratories and other wetland stakeholders and include opportunities for feedback. Prioritising diseases for surveillance The following factors should be considered when determining which diseases to prioritise for surveillance: Whether the disease is of public health or agricultural importance. Whether the disease is a specific target of a local, regional, national or international control programme. Whether the information to be collected will lead to significant successful human/animal health action. Communicable Disease Management Protocol Manual: Communicable disease surveillance. Climate change and the expansion of animal and zoonotic diseases: What is the Agency’s contribution? Wild birds and avian influenza: an introduction to applied field research and disease sampling techniques. Planning an integrated disease surveillance and response system: a matrix of skills and activities. Concepts for risk based surveillance in the field of veterinary medicine and veterinary public health: review of current approaches. Animal disease surveillance: a framework for supporting disease detection in public health. Identifying a departure from ‘usual’, ‘natural’ or ‘expected’ levels of mortality or morbidity can be complex and measures need to be put in place to help this process.