By H. Jaffar. University of Tennessee, Chattanooga.
There are two situations for which this must be modiﬁed: two-tailed testing and multiple variables 162.5 mg avalide free shipping. It is important to note that the researcher must hypothesize either an increase or a decrease in the effect generic avalide 162.5 mg fast delivery, not just a differ- ence order 162.5 mg avalide free shipping. This means that the normal distribution of one result is only likely to over- lap the normal distribution of the other result on one side or in one direction. The null hypothesis, H0 is that either there is no difference or drug A is worse than drug B. This states that we are only interested in drug A if it is better and we have good a-priori reason to think that it really is better. It removes from direct experimentation the possibility that drug A may actually be worse that drug B. The use of a one- tailed test can only be justiﬁed if previous research demonstrated that drug A actually appears to be better and certainly is no worse than drug B. When doing a two-tailed test, there is no a-priori assumption about the direction of the result. This can mean that drug A is either better or worse, but not equivalent to drug B. The null hypothesis H0 states that there is no difference between the two drugs or that they are equivalent. This means that we will accept a Type I error one in 10 trials with a one-tailed test rather than one in 20 with a two-tailed test. Conceptually this means that for a total probability of a randomly occurring error of 0. Multiple outcomes The probability of making a Type I error is α for each outcome being measured. If two variables are measured, the probability of a Type I error or a false positive result is α for each variable. The probability that at least one of these two vari- ables is a false positive is one minus the probability that neither of them is a false positive. The probability that neither is a false positive is the probability that the ﬁrst variable is not a false positive (1 – α) and that the second variable is not a false positive (1 – α). This makes the probability that neither variable is a false positive (1 – α) × (1 – α), or (1 – α)2. The probability that at least one of the two is falsely positive then becomes 1 – (1 – α)2. Therefore, the probability that one positive and incorrect outcome will occur only by chance if n variables are tested is 1 – (1 − α)n. Data dredging, mining, or snooping is a technique by which the researcher looks at multiple variables in the hope that at least one will show statistical signiﬁcance. This result is then emphasized as the most important positive result in the study. Suspect this when there are many variables being tested, but only a few of them show statistical signiﬁcance. For one variable, the probability that this association occurred by chance only is 0. The probability that at least one of the 20 variables tested will be positively associated with the disease by chance alone is 1 minus the probability of no association. Therefore, there is a 64% likelihood of coming up with one association that is falsely positive and occurred only by chance. If there are two values that show an association, one cannot know if both occurred by chance alone or if one result is truly statistically signiﬁcant. Then the question becomes which result is the signiﬁcant value and which result is a false positive. This is the previous α divided by n, the number of variables being compared, not the sample size. The Bon- ferroni correction is used when the variables being tested are independent of each other and there are only 10 or fewer variables being measured. This cor- rection is not a true assumption in most cases and other means of estimating α must be used. The variables that came up statistically signiﬁcant will then be measured in another study using only those variables and a new sample called the validation set to see if this relationship still holds. This allows the researcher to ﬁnd a statistically signiﬁcant relation- ship that exists only by chance and claim it as the reason for the study. This tech- nique is only legitimate if the variable that comes up statistically signiﬁcant in the derivation set can then become the explicit hypothesis of a validation set. This gives 124 Essential Evidence-Based Medicine Table 11.
We will discuss this further when talking about quantifying patient val- ues in Chapter 30 avalide 162.5 mg with mastercard. Recommendations about providing the evidence The most important recommendation is to avoid overwhelming the patient with too much information purchase avalide 162.5mg online. The key to avoiding this pitfall is to repeatedly check with the patient before and during delivery of the information to ﬁnd out how much she understands generic avalide 162.5mg online. Using verbal terms such as “usually” instead of numbers is less precise, and may give unintended meaning to the information. When numbers are used as part of the discussion present them in natural frequencies rather than percents. To avoid the framing bias, results should be presented in both positive and neg- ative terms. For our example patient who is interested in aspirin to prevent heart attacks and strokes, it may be most practical to use multiple modalities for pre- senting information including verbal and pictorial presentations, presenting the evidence in this way: “In a large study of women like you who took aspirin for 10 years, there was no difference in number of heart attacks between patients who took aspirin and those who didn’t. In that study, 1 out of 1000 women experienced excessive bleeding from the aspirin. If one has a strong belief that one option is the best for the patient, state that with an explicit discussion of the evidence and how the Communicating evidence to patients 207 option best ﬁts with the patient’s values. When the evidence is less than robust from weak study designs or because there are no known studies available, you cannot give strong evidence-based recommendations and must mitigate this by presenting options. When the evidence is stronger, present a recommendation and explain how that recommendation may meet the patient’s goals. In all cases, the physician has to be careful about differentiating evidence-based recommendations from those generated from personal experiences or biases regarding treatment. For our patient interested in aspirin for prevention of strokes and heart attacks, we might say: “While I understand it has been hard to lose weight and reduce your cholesterol, taking an aspirin won’t help you prevent heart attacks and is only very minimally helpful in preventing strokes. Another important part of this step is to allow the patient time to ask questions. When the physician and the patient are both in agreement that the information has been successfully transmitted and all questions have been answered, then a good decision can be made. Albert Camus (1913–1960) Learning objectives In this chapter you will learn: r the basic concepts of qualitative research r process for critical appraisal of qualitative research r goals and limitations of qualitative research While the evidence-based medicine movement has espoused the critical appraisal and clinical application of controlled trials and observational studies to guide medical decision making, much of medicine and health care revolves around issues and complexities not ideally suited to quantitative research. Qual- itative research is a ﬁeld dedicated to characterizing and illuminating the knowl- edge, attitudes, and behaviors of individuals in the context of health care and clinical medicine. Whereas quantitative research is interested in testing hypothe- ses and estimating effect sizes with precision, qualitative research attempts to describe the breadth of issues surrounding a problem or issue, frequently yield- ing questions and generating hypotheses to be tested. Qualitative research in medicine frequently draws on expertise from anthropology, psychology, and sociology, ﬁelds steeped in a tradition of careful observation of human behavior. Unfortunately, some in medicine have an attitude that qualitative research is not particularly worthwhile for informing patient care. But, you will see that qual- itative studies can be powerful tools to expose psychosocial issues in medicine and as hypothesis-generating studies about personal preferences of patients and health-care workers. Researchers then apply one or more analytic approaches to sift through the available data to identify the main themes and the range of emotions, concerns, or approaches. In the medical literature, in-depth interviews with individuals such as patients or health-care providers and focus-group interviews and discus- sions among patients with a particular condition are the most common study designs encountered. Observations of clinical behavior and analyses of nar- ratives found in medical documents (e. Qualitative research is an appropriate approach to answering research questions about the social, attitudinal, behavioral, and emotional dimensions of health care. When the spectrum of perspectives needs to be known for the develop- ment of interventions such as educational programs or technological implemen- tations, qualitative research can characterize the barriers to and facilitators of change toward the desired practice. This can be the initial research to deter- mine the barriers to adoption of new research results in general practice. Although qualitative research studies have more methodological latitude to accommodate the wide range of data used for analysis, readers of qualitative research reports can nevertheless expect to ﬁnd a clear statement of the study objectives, an account of how subjects were selected to participate and the ratio- nale behind that selection process, a description of the data elements and how they were collected, and an explanation of the analytic approach. Readers of qualitative studies should be able to critically appraise all of these components of the research methods. Designing an intervention to improve the management of Helicobacter pylori infection. The authors’ analysis revealed insights about deﬁnitions, prevalence, process, and content of secrets in primary care. The researchers transcribed the videotaped discussions and reviewed both the videotapes and the transcriptions, coding content related to the speciﬁc types of screening discussed, messages conveyed, and time spent. This objective is often framed as a research question and is the alternative or research hypothesis for the study. Unlike quantitative research studies, where the study objective is generally very speciﬁc and outcome-based, the objective or research question in qualitative studies frequently has a non- speciﬁc or general ﬂavor. In fact, it is one of the strengths of qualitative research that the speciﬁc details surrounding the study objective often emerge through the data collection and the analytic processes can actually change the direction Critical appraisal of qualitative research studies 211 of the research.
Older age and male sex are powerful determinants of risk generic avalide 162.5mg overnight delivery; consequently cheap avalide 162.5 mg visa, it has been argued that the use of the risk stratiﬁcation approach will favour treatment of elderly people and men generic 162.5mg avalide with visa, at the expense of younger people with several risk factors and women. However, while younger people gain more life years if they have a non-fatal event, older people are a lot more likely to die from an event. When discounting is taken into consideration, the quality adjusted life years gained by preventing events in young people are very similar to those gained in old people (Table 3) (50). Concern about the metabolic syndrome, characterized by central obesity, elevated blood pressure, dyslipidaemia, and insulin resistance (51, 52), has raised the question of whether identifying people with this syndrome should be a priority. There is, as yet, insufﬁcient evidence to justify using metabolic syndrome as an additional risk prediction tool (63, 64). People with metabolic syndrome would, in any case, beneﬁt from weight reduction, higher levels of activity (65–71), lowering of blood pressure, avoidance of drugs that tend to cause hyperglycaemia (72–75), lowering of choles- terol with a statin (76–80), and reduction of hyperglycaemia with metformin. There is insufﬁcient evidence from randomized trials to support more speciﬁc management of dyslipidaemias (81). In summary, the great strength of the risk scoring approach is that it provides a rational means of making decisions about intervening in a targeted way, thereby making best use of resources available to reduce cardiovascular risk. Alternative approaches focused on single risk factors, or concepts such as pre-hypertension or pre-diabetes, have been popular in the past, often because they represented the interests of speciﬁc groups in the medical profession and professional societ- ies. Such an approach, however, leads to a very large segment of the population being labelled as high risk, most of them incorrectly. If health care resources were allocated to such false-positive individuals, a large number of truly high-risk individuals would remain without medical attention. Risk scoring moves the focus of treatment from the management of individual risk factors to the best means of reducing an individual’s overall risk of disease. It enables the intensity of interven- tions to be matched to the degree of total risk (Figure 2). Further research is required to validate existing subregional risk prediction charts for individual populations at national and local levels, and to conﬁrm that the use of risk stratiﬁcation methods in low- and middle-income countries results in beneﬁts for both patients and the health care system. These charts are intended to allow the introduction of the total risk stratiﬁcation approach for management of cardiovascular disease, particularly where cohort data and resources are not readily available for development of population-speciﬁc charts. The charts have been generated from the best available data, using a modelling approach (Annex 5), with age, sex, smoking, blood pressure, blood cholesterol, and presence of diabetes as clinical entry points for overall manage- ment of cardiovascular risk. Some studies have suggested that diabetic patients have a high cardiovascular risk, similar to that of patients with established cardiovascular disease, and so do not need to be risk-assessed. In addition, in people with diabetes, there is no gender difference in the risk of coronary heart disease and stroke (82). Therefore, separate charts have been developed for assessment of cardiovascular risk in patients with type 2 diabetes. In many low-resource settings, there are no facilities for cholesterol assay, although it is often feasible to check urine sugar as a surrogate measure for diabetes. Annex 4 therefore contains risk prediction charts that do not use cholesterol, but only age, sex, smoking, systolic blood pressure, and presence or absence of diabetes to predict cardiovascular risk. Obesity, abdominal obesity (high waist–hip ratio), physical inactivity, low socioeconomic position, and a family history of premature cardiovascular disease (cardiovascular disease in a ﬁrst-degree relative before the age of 55 years for men and 65 years for women) can all modify cardiovascular risk. These risk factors are not included in the charts, which may therefore underestimate actual risk in people with these characteristics. While including these risk factors in risk stratiﬁcation would improve risk prediction in most populations, the increased gain would not usually be large, and does not warrant waiting to develop and validate further risk stratiﬁca- tion tools. Nevertheless, these (and other) risk factors may be important for risk prediction, and some of them may be causal factors that should be managed. Clinicians should, as in any situa- tion, use their clinical acumen to examine the individual’s lifestyle, preferences and expectations, and use this information to tailor a management programme. The risk prediction charts and the accompanying recommendations can be used by health care professionals to match the intensity of risk factor management with the likelihood of cardio- vascular disease events. The charts can also be used to explain to patients the likely impact of interventions on their individual risk of developing cardiovascular disease. The use of charts will help health care professionals to focus their limited time on those who stand to beneﬁt the most. It should be noted that the risk predictions are based on epidemiological data from groups of people, rather than on clinical practice. However, these objections do not detract from their potential to bring much-needed coher- ence to the clinical dilemmas of how to apply evidence from randomized trials in clinical practice, and of who to treat with a growing range of highly effective but costly interventions. Clinical assessment of cardiovascular risk Clinical assessment should be conducted with four aims: ● to search for all cardiovascular risk factors and clinical conditions that may inﬂuence prognosis and treatment; ● to determine the presence of target organ damage (heart, kidneys and retina); ● to identify those at high risk and in need of urgent intervention; ● to identify those who need special investigations or referral (e. Table 4 Causes, clinical features and laboratory tests for diagnosis of secondary hypertension Causes Clinical features and Investigations Renal parenchymal ◆ family history of renal disease (polycystic kidney), hypertension ◆ past history of renal disease, urinary tract infection, haematuria, analgesic abuse ◆ enlarged kidneys on physical examination ◆ abnormalities in urine analysis – protein, erythrocytes, leucocytes and casts ◆ raised serum creatinine Renovascular ◆ abdominal bruit hypertension ◆ abnormal renal function tests ◆ narrowing of renal arteries in renal arteriography Phaeochromocytoma ◆ episodic headache, sweating, anxiety, palpitations ◆ neuroﬁbromatosis ◆ raised catecholamines, metanephrines in 24-hour urine samples Primary aldosteronism ◆ muscle weakness and tetany ◆ hypokalaemia ◆ decreased plasma renin activity and/or elevated plasma aldosterone level Cushing syndrome ◆ truncal obesity, rounded face, buffalo hump, thin skin, abdominal striae, etc. Physical examination A full physical examination is essential, and should include careful measurement of blood pres- sure, as described below.
Interventions should be evidence-based cheap avalide 162.5 mg free shipping, and they should also consider local needs and resource constraints discount avalide 162.5 mg without prescription. Sufﬁcient resources must be available to provide the intervention to all those identiﬁed as in need discount avalide 162.5 mg mastercard. The major difference is that the likelihood of future clinical events is much greater once disease is established. When the systolic and diastolic values fall in different risk levels, the higher category applies. People who fall exactly on a threshold between cells are placed in the cell indicating higher risk. When the systolic and diastolic values fall in different risk levels, the higher category applies. People who fall exactly on a threshold between cells are placed in the cell indicating higher risk. They include the following: » Behavioural interventions: including those for tobacco cessation, increased physical activity and dietary change, with the promotion of weight loss if appropriate. Together, these may achieve a risk reduction of over 60% in people with established heart disease, and are also a key part of achieving good blood glucose control in people with diabetes (31). A combination of all four of these is expected to reduce the risk of recurrent myocardial inf- arction by 75%. Following successful implemen- diovascular death and account for half tation in these areas, the services were made available across of all cardiovascular deaths. Smokers set a date with the help of their people, international guidelines recom- adviser, and are then supported through the ﬁrst stages of their mend long-term antiplatelet, blood pres- attempt to stop smoking and followed-up after four weeks. A sure lowering and cholesterol lowering large increase in funding was made available and a demanding therapies. However, treatment gaps national target was set: 800 000 smokers to have stopped at the are substantial in all countries, in part four-week follow-up stage by March 2006. It is planned that an because of the cost and complexity of electronic appointments system will be available to smokers to multiple drug use. One strategy that has been proposed Results for the period April 2004–March 2005 show that around to reduce these barriers is a ﬁxed dose 300 000 smokers had successfully stopped at the four-week fol- combination pill (now commonly known low-up stage compared with about 205 000 the year before (an as a polypill). Initial ﬁndings also show that equity of access apparently works in addition to the oth- to treatment is good, although success rates are lower among ers, net beneﬁts are anticipated to be disadvantaged groups. As well as improving clinical outcomes, they simplify distribution of multiple medications, which can be an important advantage in a resource-limited health-care setting. The major challenge remains one of implementation – new strategies are required for the many millions of under-treated individuals with established cardiovascular disease in low and middle income countries. For people with cardiovascular disease in low and middle income countries, access to preventive care is usually dependent upon their ability to pay, and hence it is this large, underserved group that stands to gain most from a polypill (32, 33). Yet in many places, effec- tive interventions for chronic diseases are poorly delivered or are not available at all. In some settings, lack of human, physical and ﬁnancial resources are the major constraining factors. In other settings, resources are available but are used in a fragmented In a rural South African setting, a nurse-led chronic and inefﬁcient manner. Factors to take into account disease management programme for high blood include the following: pressure, diabetes, asthma, and epilepsy was » evidence-based decision support tools can improve established as part of primary health care for a the delivery of effective care for chronic diseases; population of around 200 000 people. The pro- » effective clinical information systems, including gramme included the introduction of: clinic-held patient registries, are an essential tool for provi- treatment cards and registries; diagnostic and ing the continuity of care necessary for chronic management protocols; self-management sup- diseases; port services; and regular, planned follow-up with a clinic nurse. Nurses were able to improve disease control among most of the patients: 68% of patients with high blood pressure, 82% of those with diabetes,109 and 84% of those with asthma (34). Five greater efﬁciency from their health systems health-care facilities, each with a multidisciplinary team of by combining disease management for all staff, were involved in the decision-making and planning of chronic conditions. They enable the » reallocation of ﬁnancial and human resources to facilitate organization of patient information, tracking implementation of these services. Multidisciplinary health-care teams, centred on primary The Secretariat of Health of Mexico health care, are an effective means in all settings of achieving this has launched a “crusade for the goal and of improving health-care outcomes (37 ). It is possible, however, to provide some the implementation of a structured of the core skills from these disciplines in other ways (by training diabetes education programme. It may be possible to provide core trained to adopt a quality improve- aspects of effective health care that in more resourced settings ment methodology. Among the inno- would be provided by health professionals from several different vations in primary health centres disciplines. The while among those receiving usual production of an evidence-based guideline is a resource-intensive care the proportion only increased and time consuming process. Documented foot lines are available for many chronic diseases (see, for example, care education increased to 76% of http://www. For example, simply providing information about the guideline is likely to have little impact, but linking the guideline to workshops or outreach training sessions and providing prompts within medical records are much more likely to change practice (41).