Clinical epidemiology
Abstract
Global society has reached a level of interdependence wherein there is a need to share healthcare knowledge and deploy resources in the best interests of people everywhere. Clinical and public health professionals can be united in this effort through their common reliance on epidemiology. Clinical epidemiology and its derivative—the evidence-based medicine movement—have many parallels with public health. Indeed, many clinicians with clinical epidemiology training develop research projects and subsequently research programmes that move beyond clinical decision-making to include a population focus.
In response to this global need, the International Clinical Epidemiology Network (INCLEN) programme has trained over 700 physicians and other health specialists at a Master’s degree level in clinical epidemiology, social sciences, biostatistics, or clinical economics. INCLEN has established a global resource network to support fundamental changes in the way physicians, medical educators, and policy makers think about health and disease. INCLEN now has semi-autonomous regional networks in Africa, India, China, Southeast Asia, Latin America, Europe–Mediterranean, and Canada–United States.
A methods framework, the ‘equity–effectiveness iterative loop’, is used to demonstrate the interface between clinical epidemiology and public health, with special attention to ensuring that the disadvantaged are explicitly considered. The focus is on evidence-based, action-oriented epidemiology based upon the health needs of the relevant community. Various examples are used, such as circumcision to prevent male-acquired HIV infection.
The history and evolution of clinical epidemiology as related to public health
There is a growing concern that the costs of medical care may be exceeding the benefits, or at the very least that the same level of healthcare benefits could be achieved at lower costs (Cutler 2004). More importantly, there is clear consensus that inequitable distribution of these benefits abounds. It often seems that the majority of healthcare researchers and providers in the high-income countries have accorded public health a low priority (Tugwell et al. 2006a). According to the Institute of Medicine, the mission of public health is defined as ‘fulfilling society’s interest in assuring conditions in which people can be healthy’ (Institute of Medicine 1988); public health emphasizes a collective action for the ‘public good’ over a ‘private good’. Some of this disconnect can be explained by the separation of clinical medicine and public health, that began almost a century ago.
With the advent and remarkable successes of the germ theory in the nineteenth century, much of clinical medicine became focused on laboratory research—describing the pathophysiology of individual microorganisms—and little research was conducted on the host or environment. In 1916, the Rockefeller Foundation, endowed by John D. Rockefeller and chartered in 1913 for the well-being of people throughout the world, concluded that insufficient attention was being paid to environmental and social factors in disease, and that public health personnel were widely needed.
The Foundation’s solution was to establish schools of public health apart from schools of medicine. The role of medicine was thus to provide care to individuals and investigate disease processes, and the role of public health was to research and impact the determinants of health and disease in populations. Epidemiology, the social sciences, and qualitative methods became the domain of public health, and medicine assumed dominance in clinical treatment advances.
In his 1938 president’s address to the American Society for Clinical Investigation, John Paul introduced the term ‘clinical epidemiology’, which he defined as the application of epidemiological and related methods from a clinical perspective. The shift in the focus of clinical epidemiology from community ecology to individual patients and groups of patients took place in the 1960s. In 1964, Alvan Feinstein, who has been referred to as the father of clinical epidemiology (Fletcher 2001), published an influential series of papers on scientific methodology for clinical medicine in the Annals of Internal Medicine. When asked by an epidemiologist to provide a lecture on his work in epidemiology, Feinstein replied:
… I am not an epidemiologist. Epidemiologists are people who go around collecting useless statistics about the incidence and prevalence of syphilis in Tasmania. I don’t do that.’ A critic stated, ‘Yes, you are; you study groups of people and apply statistics to them: That is what epidemiologists do.’ And I said, ‘Well, if I am, I am a clinical epidemiologist. (Daly 2005)
Kerr White also felt the need to incorporate greater scientific rigour into medical practice. White’s training included medicine and public health, and he felt that epidemiology—the basis of public health—had much to offer clinical medicine. In 1964, White was the founding chair of what became the Department of Health Policy and Management in the School of Public Health at John Hopkins University, where he attempted to bridge the gap between medicine and public health. However, many of his colleagues were threatened by what they perceived as efforts to usurp clinical authority. As noted by White: ‘When proposing a population-based study of hospitals, for example, I was called before the local medical society and accused of trying to mount a communist plot!’ (Daly 2005).
In 1966, David Sackett established the first Clinical Epidemiology Research Unit in the Department of Medicine at the State University of New York at Buffalo, United States. By 1967, he had moved to Canada and established the influential Department of Clinical Epidemiology and Biostatistics at the McMaster Medical School in Hamilton, Canada.
In 1972, Archibald (Archie) Cochrane compiled a series of lectures commissioned and published by the Nuffield Trust—Effectiveness and Efficiency: Random Reflections on Health—that compellingly articulated the need for rigorous scientific evaluation of diagnosis, treatment, and preventive medicine (Cochrane 1972).
The distinct field of public health was beginning to influence and challenge clinical medicine. The creation of these two parallel institutions with separate mandates resulted in what has been termed the schism between medicine and public health. This divide was not to the benefit of either group and in 1974 the Macy Conference reported:
… [T]he health professionals shut themselves up in their schools of public health, and the physicians stayed within the walls of the medical schools and hospitals. The latter felt that public health specialists ‘were no longer doctors’, while the health people believed themselves to be crusaders in a cause they had to win, imposing it if necessary on the community as well as on other physicians who did not understand them. (Bowers 1974)
In 1977, John Knowles, then President of the Rockefeller Foundation, edited a collection of articles on American healthcare, and wrote:
Public health interests have been, and continue to be, isolated from American medical education and practice. Issues that influence health, such as nutrition, family size, population density, environmental mobility, poverty, racism, sexual practices, unemployment, housing, transportation, and the like, are rarely taken into account in any overall calculation of the health needs of the nation. (Knowles 1977)
Seminal work by, in addition to the above individuals, Gene Glass, Richard Peto, Iain Chalmers, Gordon Guyatt, Brian Haynes, and Andrew Oxman developed methods by which to identify, appraise, select, and summarize healthcare literature to inform clinical questions (a systematic review) and, when appropriate, mathematically combine data to provide a more precise estimate of treatment effect (a meta-analysis). However, such efforts were initially not well-accepted within medicine.
In 1979, commenting on organized medicine’s limited use of clinical epidemiology to inform best practices, Archie Cochrane wrote: ‘It is surely a great criticism of our profession that we have not organized a critical summary, by specialty or subspecialty, adapted periodically, of all relevant randomized controlled trials’ (Cochrane 1979).
Formal international dissemination of the principles of clinical epidemiology began in 1980 when Kerr White, who was Deputy Director for Health Sciences of the Rockefeller Foundation, initiated through the foundation, The International Clinical Epidemiology Network (INCLEN). This programme provided young clinicians from low- and middle-income countries with funding to pursue training in clinical epidemiology at Chapel Hill in the United States, McMaster University in Canada, University of Newcastle in Australia, or the University of Pennsylvania in the United States. As of January 2008, INCLEN included 81 clinical epidemiology units/centres in 33 countries (see Box 6.11.1).
In 1982, 44 years after the term was first coined, the first modern textbook in clinical epidemiology was published by Robert Fletcher, Suzanne Fletcher, and Edward Wagner at the University of North Carolina, United States (now in its 4th edition) (Fletcher & Fletcher 2005). Soon after, clinical epidemiology textbooks were published from McMaster University, Canada (now in its 3rd edition) (Haynes 2006) and Yale University, United States (Feinstein 1985) in 1985, from the University of Washington, United States in 1986 (now in its 3rd edition) Weiss 1996), and from McGill University, Canada in 1988 (Kramer 1988).
The United States government implemented a number of initiatives to improve collaboration between the US Public Health Service and the medical establishment, but in 1988 the Institute of Medicine drew attention to the ongoing ‘poor relationships [of public health workers] with the medical profession’ (Institute of Medicine 1988). Similar concerns were raised in Britain through The Report on the Committee of Inquiry into the Future Development of the Public Health Function (Acheson 1988).
In 1991, White advanced two main issues that he felt had limited the development of public health. First, the inability to attract and retain accomplished medical physicians. As an example he noted that the number of physicians employed in US state health departments had decreased by more than 34 per cent from 1979 to 1986. The second barrier he saw was the failure by most physicians to understand the public health perspective. White proposed that four root causes were to blame for this situation (White 1991):
1. The failed attempt to establish public health as a separate profession, apart from medicine
2. Failure to establish epidemiology as a fundamental science for medicine and public health
3. Failure to provide public health and epidemiological training to medical students
4. Failure of medicine and public health to cooperate and coordinate their efforts
Others have argued that the primary barrier to the development of public health was ideological; the belief of many physicians that their mandate was the application of basic science to improve health one individual and one treatment at a time. This ideology translated into financial benefit and increased authority for the clinical enterprise, resulting in the endorsement of what has now become known as ‘eminence-based medicine’ (Fox 1986).
In 1990, Gordon Guyatt approached the Department of Medicine at McMaster University (Canada) with the idea of incorporating a novel approach to educating physicians (Guyatt & Rennie 2002). In an effort to denote the importance that use of current scientific literature would have in this new curriculum, he proposed the term ‘scientific medicine’. Some department members were incensed by the implication that what they were currently teaching was unscientific. This led to a modification of the term to ‘evidence-based medicine’, that relied heavily upon clinical epidemiology. In order to best define evidence-based medicine, it is helpful to quote the conclusions of Sackett et al. (1996), in their well-known article, ‘Evidence-based medicine: What it is and what it isn’t’:
Good doctors use both individual clinical expertise and the best available external evidence, and neither alone is enough. Without clinical expertise, practice risks becoming tyrannized by evidence, for even excellent external evidence may be inapplicable to or inappropriate for an individual patient. Without current best evidence, practice risks becoming rapidly out of date, to the detriment of patients.
Prominent among the efforts to facilitate the implementation of evidence-based medicine in clinical practice is the Cochrane Collaboration—named in honour of Archie Cochrane—and formed under the leadership of Iain Chalmers in 1993. This collaboration is comprised of an international representation of patients (or professional consumers who claim to speak for patients), clinicians, and methodologists that aims to prepare systematic reviews of the effects of healthcare interventions. These reviews are targeted for updating every 2 years (although this is variable), and their results are available through paid subscription via the Internet (http://www.cochrane.org). Evidence-based medicine is not without its detractors (Straus et al. 2007; Feinstein 1997), but in 2006, the British Medical Journal designated evidence-based medicine one of the 15 greatest medical breakthroughs since 1840 (Ferriman 2007).
The separation of public health and medicine was the result of a deliberate attempt to create two professions. This divide was maintained, in part, due to social and political factors. It now seems evident that close collaboration between medicine and public health will serve society far better. The shift inherent in evidence-based medicine, from the authoritative approach of healthcare to one that attempts to use clinical epidemiology to incorporate the best evidence currently available, represents a challenge for all healthcare providers; however, the global society has reached a level of interdependence wherein there is a need to share healthcare knowledge and deploy resources in the best interests of people everywhere. Clinical and public health professionals and public health can be united in this effort through their common reliance on epidemiology (Institute of Medicine 2007).
The equity–effectiveness loop: A framework for the interface between clinical epidemiology and public health
Clinical epidemiology and its derivative—the evidence-based medicine movement—have many parallels with public health. Indeed many clinicians with clinical epidemiology training develop research projects and subsequently research programmes that move beyond clinical decision-making to include a population focus. This stimulated one of us (PT) and his colleagues to address this process systematically through an iterative measurement loop framework (Tugwell et al. 1985). The focus was on evidence-based, action-oriented epidemiology based upon the health needs of the relevant individuals and their community.
This has recently been updated with an ‘equity lens’ to ensure that the disadvantaged are explicitly considered (Tugwell et al. 2006a) (see Fig. 6.11.1). What was the stimulus to incorporate explicit attention by clinical epidemiologists to the disadvantaged? This was due to the realization that average improvement can hide important inequitably worse health effects amongst the disadvantaged. For example, impressive gains in health during the twentieth century, showing dramatic increases in average life expectancy in rich and poor countries (World Health Organization 1999), would meet the criteria we initially recommended in the 1985 paper, but there is a critical component missing. These averages obscure the fact that health in both high- and low- income settings is unevenly distributed according to socioeconomic position; health and longevity are highest for the richest, and decrease steadily with decreasing socioeconomic status (Wilkins & Adams 1983; Wilkinson 1996). Many of these inequalities are avoidable, and hence unfair.
These social gradients in health, or socioeconomic inequalities in health, are pervasive in all countries of the world (Diderichsen et al. 2001) and hold true for most diseases, injuries, and health behaviours. Modern health policy must increasingly be oriented not only to the production of health, but also the distribution of health (Gwatkin 2003). For example, the Millennium Development Goals state the need to include poor people in the benefits of development (United Nations 2001).
Therefore, the ‘equity–effectiveness iterative loop’, as its name implies, has been expanded to provide a logical way to apply an ‘equity lens’ as one moves from assessing needs through to assessing effectiveness, and cost-effectiveness, of interventions, leading to the development and evaluation of evidence-based health policy. This framework integrates the concepts of individual risk and socioeconomic status with intervention effectiveness from a population health perspective.
Step 1: Burden of illness
This step measures the burden of illness and its gradient by socioeconomic status. This includes downstream (individual), and upstream (societal) determinants of health (biological, cultural, political, psychosocial, and environmental).
Step 2: Differential equity–effectiveness
Controlled studies provide estimates of efficacy and effectiveness; efficacy measures how well an intervention can work in ideal circumstances (Sackett et al. 1985). Effectiveness measures how well an intervention works in real settings and systems at the community level. Community effectiveness is often substantially lower than efficacy because of a staircase effect with four ‘steps’: (1) the result of lower awareness, access, or coverage; (2) screening, diagnosis, or targeting; (3) compliance of providers; and (4) adherence of consumers. Poor people may have circumstances that reduce efficacy at all four steps and therefore a greater staircase effect may be observed compared to the least poor people. There is a need to assess equity issues across each step to identify barriers to implementation related to gradients in wealth.
Step 3: Economic evaluation
This step assesses the efficiency (health benefits such as number of disability-adjusted life years avoided for a specific cost that includes direct, indirect, and where possible intangible costs) of the intervention. Assessing the efficiency requires adequate evidence of efficacy and valid estimates of cost. Assessing the equity issues related to cost-effectiveness implies a trade-off between cost efficiency and population health equity. Priority funding of interventions with the best cost-effectiveness ratios might increase differences between the richest (or least poor) and poorest because the cost of reaching poor people may be higher and health benefits may be lower. Four approaches to this have been proposed (Drummond 2006). One promising method to assess equity issues related to cost-effectiveness is the development of an equity and quality-adjusted life year (EQ-QALY), as a complement to established measures of the difference between rich and poor, such as the concentration index (Wagstaff 2002).
Step 4: Knowledge translation and implementation
Translation of knowledge is defined as the process that transfers research results from producers of knowledge to its users, for the benefit of the population. Moving beyond the traditional domain of academic publication, it comprises three interlinked components of uptake and translation: Exchange, synthesis, and ethically sound application of knowledge (Birdsell et al. 2002). This step entails uptake and translation of knowledge into action (Birdsell 2002; Davis et al. 2003; Grimshaw et al. 2001).
There is therefore a need to develop new, effective means of packaging and communicating evidence on effectiveness across wealth gradients to the different policy, community, and practitioner groups or individuals responsible for each of the components of community effectiveness—access, diagnostic accuracy, compliance of providers, and adherence of consumers (Giuffrida 2000; Briggs et al. 2001; Mowatt et al. 2001; Zwarenstein & Bryant 2000). Evidence that interventions using knowledge translation are efficacious is currently lacking in most sectors. One exception is the work of INCLEN (see Box 6.11.1), which is developing methods explicitly to consider equity issues in developing and applying clinical guidelines (INCLEN 2004). By targeting the wealth gradient in knowledge translation strategies, we support the operational research agenda for optimizing the benefits to the poor of key interventions.
Steps 5 and 6: Monitoring and reassessment
Monitoring identifies the importance of process assessments and intermediate outcomes to assess success in affecting mortality and morbidity by socioeconomic group and deciding whether further remediable need exists; if so, an additional iteration of the equity–effectiveness loop is needed. The Whitehall cohort study, for example, showed that, even with equitable access to cardiac care, the social deprivation gradient still produces disparities in outcomes (Britton et al. 2004), indicating a need to tackle other causes, or ‘steps’, of disparities.
Step-by-step through the equity–effectiveness loop
We elaborate on the application of the equity–effectiveness loop using examples of interventions such as male circumcision to prevent transmission of HIV and other health challenges, particularly in low- and middle-income countries.
Male circumcision, the removal of the foreskin and thus removal of many of the Langerhans cells in the male genital area, has been examined as a prevention strategy since the early 1990s (Wawer et al. 2005). Given the rate that the HIV/AIDS pandemic is increasing, it is clear that we need to implement preventive strategies based on the highest quality evidence. Current strategies promoted for the prevention of HIV infection include education, condoms, sterile injection equipment, and abstinence. Other interventions, such as vaccines, pre-exposure prophylaxis, and microbicides are under study and their effectiveness is still unproven. Most recently, male circumcision has been promoted as a strategy to reduce infection (Wawer et al. 2005; Mills & Siegfried 2006). We will examine this issue in detail using the equity–effectiveness loop.
Step 1: Burden of illness
The HIV/AIDS pandemic now affects around 33 million people infected worldwide, with reason to believe that this will continue to increase unless effective interventions can be developed to slow the rate of transmission. Important gradients are present in HIV-affected populations across different social groups such as differing religions. Religion, even after adjusting for lifestyle risk factors across education and income, appears to predict HIV prevalence (Drain 2005). Higher income and higher education appear to present a paradoxical increase in HIV incidence in most African countries (Malawi National AIDS Commission, http://www.aidsmalawi.org.mw). The effect of HIV on the community impacts on the economy due to orphanhood, grandmothers caring for villages, a lack of health worker supply, and inadequate supplies of effective drugs.
Step 2: Efficacy, community effectiveness, and differential equity–effectiveness
In initial demographic analyses, investigators observed that populations in sub-Saharan Africa with large populations of the males circumcised had comparatively lower HIV/AIDS indices than populations that did not traditionally circumcise males (Drain et al. 2006).
Previous observational studies of males in HIV prevalent areas gave a relative risk of infection among circumcised men that ranged from 0.22 (95 per cent CI, 0.10–0.46) based on cohort studies to 0.41 (95 per cent CI, 0.35–0.49) based on cross-sectional studies. The biological rationale for circumcision as a preventative measure comes from the knowledge that the foreskin is highly susceptible to skin ruptures and possible permeability and has many Langerhans cells. These cells are believed to capture HIV-1 virions through site receptors that bind to antibody-coated virus, thus acting as reservoirs for the virus and serving as a site of replication.
Even with the high relative odds and a plausible biological explanation for a protective effect, observational studies sometimes provide unreliable estimates of the effectiveness of an intervention due to known and unknown confounding (Guyatt & Rennie 2002). In the HIV and circumcision example, an obvious confounding variable is that circumcision was predominantly practised in countries with large Islamic populations, and evidence from countries with lower Islamic populations demonstrated that Islamic populations tend to have lower prevalence of HIV/AIDS than their corresponding non-Islamic populations (Drain et al. 2006). With this conundrum, public health officials were, rightly, hesitant to recommend male circumcision on a population-wide level, given that the studies were observational and not randomized trials, (Siegfried et al. 2005). Figure 6.11.2 displays the widely accepted classification of evidence for implementing interventions in public health decision-making (Guyatt et al. 2000).
For a public health decision to be evidence-based, policy-makers should follow clear steps in: (1) formulating a question; (2) seeking the best available evidence; (3) determining the quality of that evidence; and (4) determining the applicability of the evidence to their settings.
Formulating a question: Any well-designed question should be composed of the following five criteria–PICOT:
1) Population
2) Intervention
3) Control
4) Outcome
5) Time-duration of the trial
To frame the scenario of male circumcision for HIV/AIDS prevention, we propose the following structured question:
In sexually active males (population) that have been circumcised (intervention), are rates of HIV infection (outcome) over 2 years (time–duration of the trial) importantly different from sexually active males that have not been circumcised (control)?
This structured question addresses all of the important components of an answerable question. Note that we state ‘importantly different’ rather than the more conventional ‘statistically’ different as there are examples where statistically significant results do not result in clinical or meaningful differences (Redelmeier et al. 1996; Wells et al. 2001).
As displayed in Fig. 6.11.2, the first choice for assessing a quantitative benefit of an intervention are systematic reviews or meta-analyses of randomized clinical trials (RCTs). Next are single well-conducted RCTs. Next are single observational studies, and so on. We are aware that the demographic and observational studies on male circumcision have existed since the 1990s and provide compelling rationale for considering the introduction of population-wide male circumcision, but we are also cautious that these observational studies may be vulnerable to confounding variables, such as religious behaviours (Drain et al. 2006). As a result, we would ideally like to find a meta-analysis of randomized trials or several well-conducted RCTs.
Seeking the best available evidence: For the purpose of this chapter, we searched the following electronic databases (from inception to 8 October 2007): MedLine (via PubMed); EMBASE; CINAHL; AMED; Cochrane CENTRAL; Cochrane Database of Systematic Reviews; and Relief Web (a database of freely accessible non- governmental organization reports and news reports).
Using the search terms ‘HIV AND circumcision AND random*’, we identified three RCTs published since 2005 in the biomedical literature (Auvert et al. 2005; Bailey et al. 2007; Gray et al. 2007). Next, we appraised the applicability and internal and external validity of each of these RCTs.
Critical appraisal: Three of the trials described RCTs were published in sub-Saharan Africa. All were conducted among sexually active males. All of the trials had a non-circumcised control group and all of the trials assessed HIV infection as their primary outcome. At this stage, all of the trials appear to provide us with information that addresses our initial question. Next, we determined if the trial quality was sufficient to provide strong inferences about the effectiveness of male circumcision for reducing HIV/AIDS.
In order to determine trial quality, we suggest using the users’ guide, ‘How to appraise an article about therapy’, published by the McMaster EBM Working Group in 1994 (Guyatt et al. 1994). This tool has been modified for a variety of fields, but the utility of it remains (see Box 6.11.2).
Using the questions from the Users’ Guide, we determined the quality and utility of each RCT (see Table 6.11.1).
Table 6.11.1 Study characteristics and results
Author, year | Design | Setting | Population | n | Outcomes Relative risk (95% confidence intervals) | ||
|---|---|---|---|---|---|---|---|
Intervention | Control | ||||||
Auvert et al. (2006) | Randomized controlled trial | Orange Farm, South Africa | Males between 18 and 24 years | 3274 | 20/1546 | 49/1582 | 0.42 (0.25–0.70) |
Bailey et al. (2007) | Randomized controlled trial | Kisumu, Kenya | Males between 18 and 24 years | 2784 | 22/1391 | 47/1393 | 0.47 (0.28–0.77) |
Gray et al. (2007) | Randomized controlled trial | Rural Rakai district, Uganda | Males between 15 and 49 years | 4996 | 22/2387 | 45/2430 | 0.50 (0.30–0.83) |
Trial 1 (Auvert et al. 2005)
Primary guides on methods
In 2005, the first RCT of male circumcision was published to great media attention (Table 6.11.1) (Auvert et al. 2005). This study enrolled sexually active males between the ages of 18 and 24 years. The study randomly assigned the participants to receive surgical circumcision or remain uncircumcised. The study used urn randomization and allocation was concealed using opaque envelopes. Participants in the active arm were circumcised using the forceps guided method within a week of randomization. All data was analysed using an intent-to-treat approach, in addition to examining only those patients that underwent treatment according to the study arm they were randomized to (a per-protocol approach). The trial was stopped early due to benefit after the Data and Safety Monitoring Board’s (DSMB) first interim analysis. The average duration of participants in the trial was 18.1 months (range 13–21).
Secondary guides on methods
Participants and health workers were not blinded to group allocation, although efforts were made to ensure confidentiality of the participant’s circumcision status. Groups were largely similar at the start of the trial and almost all (∼90 per cent) were sexually active at the start of the trial.
What were the results?
The proportion of participants lost to follow-up was 8 per cent (251/3128), with 3.2 per cent (100) in the active group and 4.8 per cent (151) in the control groups (P = 0.0016). Among those lost to follow-up, none were HIV-positive at their last completed follow-up visit. By the end of the study, 92 active group participants were not circumcised and 114 control group participants were circumcised. During the study, 20 active group participants and 49 control group participants acquired HIV infection, corresponding to incidence rates of 0.85 per 100 person-years (95 per cent CI, 0.55–1.32) and 2.1 per 100 person years (95 per cent CI, 1.6–2.8) in the intervention and control groups, respectively. The Relative Risk (RR) of HIV infection corresponds to 0.42 (95 per cent CI, 0.25–0.70) in favour of the intervention. Adverse events among the circumcised group were minor and included pain at the site (12/1495), excessive bleeding (9/1495), and haematoma (9/1495).
Will the results help in caring for a target population?
The participants enrolled in the trial likely represent males through sub-Saharan Africa, and there was a suitably broad set of inclusion criteria to represent males of different ethnic groups and religions. The trial reported adequately all important outcomes. We should be wary of the large effect size given that trials stopped early may inflate effect sizes (Montori et al. 2005). The question of whether the treatment benefits outweigh the harms and costs must be interpreted in a culturally specific manner. Will the broad population accept this intervention? Will perceived protection affect other established preventative measures such as condom use?
Trial 2 (Bailey et al. 2007)
Primary guides on methods
The next available trial (Bailey et al. 2007) was published simultaneously with the third RCT (Table 6.11.1) (Gray et al. 2007) in February 2007, although results had been available in the media since early January 2007. A total of 2784 men aged 18–24 years in Kisumu, Kenya, were randomly assigned to an intervention group (circumcision, n = 1391) or a control group (delayed circumcision, n = 1393). Most men identified themselves as unskilled workers, farm labourers, or fishermen (n = 1653, 59 per cent); 632 (23 per cent) were students. Randomization employed stratification for established risk factors. Allocation concealment was ensured through the use of opaque envelopes. Circumcision generally occurred within 24 h of randomization, through the forceps guided method. Participants were counselled to refrain from sexual activity for at least 30 days after the procedure. The trial used an intent-to-treat analysis as well as per-protocol analyses. As with the first trial, this trial was stopped early by the DSMB due to benefit at the third interim analysis. The average duration of participants in the trial was 24 months (range: 18–24).
Secondary guides
Participants and medical staff were not blinded to group allocation. However, data analysts and HIV testers were blinded. Groups were similar at the beginning of the trial and were treated equally throughout, with the exception of the intervention.
What were the results?
Overall, follow-up for HIV status was incomplete for 240 (8.6 per cent) participants: 126 (4.5 per cent) in the circumcision group and 114 (4.1 per cent) in the control group. Circumcision provided a RR of 0.41 (95 per cent CI, 0.24–0.70) protective effect against HIV infection compared with the control group, and a Relative Risk Reduction (RRR) of 0.40 (95 per cent CI, 0.23–0.68) protective effect, after adjustments for non-adherence and for those individuals who were found to be HIV-positive at baseline. By the end of the trial, 18 participants were not circumcised in the active group and 12 participants from the control group were circumcised. Further, 4 participants were found to have been HIV-positive at baseline and so were excluded from the analysis. Adverse events among the circumcised group were minor and included bleeding (5/1391), infection (4/1391), and delayed healing (3/1391).
Will the results help in caring for the target population?
The participants in this trial are representative of the general Kenyan male population and were of mixed professional status and sexual history. It is likely that the results from this trial have applicability to populations throughout East Africa. Although stopped early, the trial largely completed its intended duration
Trial 3 (Gray et al. 2007)
Primary guides on methods
The final study (Gray et al. 2007) that met our eligibility criteria was published along with Bailey et al. (2007) in February 2007. The trial randomized 4996 uncircumcised, HIV-negative men aged 15–49 years in rural Rakai district, Uganda. Men were randomly assigned to receive immediate circumcision (n = 2474) or delayed circumcision for 24 months (n = 2522). Randomization occurred using urn randomization, with opaque envelopes to conceal allocation. Participants in the active group were circumcised within an average of 2 days post-randomization. Circumcision was done with the sleeve procedure, and patients were counselled to abstain from sexual intercourse until complete wound healing had occurred. The primary trial outcome, HIV infection, was assessed using intention-to-treat analysis for participants that provided a PCR-negative HIV test at study initiation. The trial was again terminated by the DSMB due to benefit at the second interim analysis, at which time 44 per cent of all participants had completed the entire trial.
Secondary guides
Participants and healthcare workers could not be blinded. Groups were similar at the beginning of the trial and were treated equally throughout, with the exception of the intervention.
What were the results?
A total of 114 (2.3 per cent) active group participants and 115 (2.3 per cent) control group participants were lost to follow-up over the trial duration. Further, 146 participants in the active group did not receive circumcision and 33 men in the control group were circumcised outside of the trial parameters. There were 22 HIV infections in the active group and 45 infections in the control group. This corresponds to an HIV incidence over 24 months of 0.66 cases per 100 person-years in the intervention group and 1.33 cases per 100 person-years in the control group. The rate of all adverse events related to surgery in the intervention group was 7.6 per cent (178 events in 2328 surgeries) and most were mild. The severe adverse events included one wound infection, two haematomas that required re-exploration and ligation of active bleeding vessels, one wound disruption due to external cause, and one case of severe postoperative herpetic ulceration not involving the surgical wound requiring hospitalization.
Will the results help me in caring for my population?
The population enrolled represents a relatively rural area of Uganda with broad inclusion criteria representing many different employment and sexual histories. It is likely that the results from this trial have applicability to other populations in East Africa and possibly throughout sub-Saharan Africa. The trial was stopped prior to completion and one should be cautious that the effects may be inflated (Montori et al. 2005).
Making sense of the study findings
The studies we included all reported that circumcision promoted important effects in preventing HIV infection. In general the studies were well reported and apparently well conducted. All three trials were stopped early due to benefit, prior to their intended completion. There is reason to believe that studies that stop early are at risk for yielding inflated effect sizes. In an analysis of 143 RCTs that stopped early, Montori et al. (2005) identified that studies that stopped prior to intended completion, and had small numbers of events, yielded much higher effect sizes. The fewer events that had accrued at the time investigators terminated their trial, the higher the risk of overestimation. In the study by Montori et al., when comparing the trials with events fewer than the median number (66) to those above the median, they found the odds ratio for a magnitude of effect greater than the median (RRR of 47 per cent) was 28 (95 per cent CI, 11–73 per cent). Given the small number of events within each of our specific trials, we felt it was best to pool the studies to create a meta-analysis, thereby increasing the number of events available for interpretation.
Meta-analysis of included studies
We included data from all three RCTs in a meta-analysis to pool the effect size estimates in order to determine the likely effect size across all three trials.
In order to pool across studies, we need to be aware of the total sample size in each group, from each trial, that was exposed. We also need to determine the number of patients in each group that acquired the outcome (HIV infection). Table 6.11.1 displays these findings from our three included studies.
Meta-analysis is a statistical tool that can be relatively easy, as in this case where there are only three trials with clear outcomes, or more difficult (and may require the assistance of a statistician) when there is a more substantive number of trials and when differences across trials are examined using advanced methods such as sensitivity analyses, meta-regression, or sub-group analyses.
We pooled our data with Stats Direct, (StatsDirect Ltd. Manchester, UK, version 2.1), using a random effects model, namely the DerSimonian–Laird random effects method, which recognizes and anchors studies as a sample of all potential studies, and incorporates an additional between-study component to the estimate of variability (DerSimonian & Laird 1986). This method provides more conservative estimates than the fixed-effects model. We calculated the I2 statistic and associated 95 per cent confidence interval for the pooled effect size, the percentage of between-study variability that is due to true differences between studies (heterogeneity) rather than sampling error (chance) (Higgins & Thompson 2002). We considered an I2 value greater than 50 per cent to reflect substantial heterogeneity. The meta-analysis is plotted on a forest plot, with RR less than 1 indicating findings in favour of male circumcision (Fig. 6.11.3).
By combining data from the three RCTs, we found a pooled RR of 0.44 per cent (95 per cent CI, 0.33–0.60, P = <0.0001, I2 = 0 per cent, 95 per cent CI 0–35 per cent), corresponding to a RRR of 56 per cent (95 per cent CI, 40–67 per cent). There was no observable heterogeneity across the trials. Another way of interpreting results is the number needed to treat (Laupacis 1988). The number needed to treat to prevent one HIV infection was 58 (95 per cent CI: 48–81).
Applying the findings to the target populations
Our meta-analysis provides us with a strong inference that male circumcision may reduce HIV infection amongst men. We need to now determine whether it is practical and feasible to pursue male circumcision across the African population. When we consider that sexually active males in sub-Saharan Africa represent the group most responsible for infections in women and children, we recognize that reducing male likelihood of infection may protect partners and other groups exposed to sexual activity with the males. For that reason, the intervention is appealing.
We recognize, however, that implementing circumcision across a broad population is fraught with difficulties related to spiritual, religious, and other value-laden decisions (Ngalande et al. 2006). Even in North America, where circumcision was routine until the 1990s, there is now growing opposition to circumcision, and such arguments need to be considered in a compassionate and public health context (Auvert et al. 2006).
We also recognize that there are harms related to circumcision. Many males in sub-Saharan Africa are circumcised in traditional ceremonies and there is concern about a lack of sterility of the tools used and the potential for serious adverse events and potentially even HIV infection through exposure to unsterile equipment (Mills et al. 2006a). Further, in our analysis above, we found that minimizing adverse events in a sterile setting require considerable medical experience, which may not exist in settings with few healthcare workers, to ensure that the risk of infection and serious bleeding from the wound are minimized.
Differential effectiveness. Many controlled studies provide estimates of efficacy i.e. how well an intervention can work in ideal circumstances. The equity–effectiveness loop takes efficacy as the anchor point representing the maximum benefit that can be achieved. Community effectiveness is the measure of how well an intervention does work when delivered in real-life settings and systems at the community level. This ‘community effectiveness’ is often substantially lower than the expected efficacy because of four systemic factors: Awareness/access/coverage, screening/diagnosis/targeting, provider compliance, and consumer adherence.
Circumcision is a very good example of this
Indeed, we are concerned that media and international agencies are sending out the wrong message—that circumcision reduces infection chances by 60 per cent (WHO/UNAIDS 2007). This is based on the efficacy data above. In support of the stability of the estimates is the fact that these efficacy trials were held in comparatively different economic conditions within the threshold of ‘poverty-affected’ populations (i.e. the poor populations of these countries are probably similar, even though GDPs may vary). Examining whether effect sizes within our meta-analysis may differ according to location (see Fig. 6.11.3), we can see that all trials displayed a similar rate of effectiveness. It is also remarkable that our pooled analysis is almost identical to the pooled cross-sectional estimate of 0.41 (95 per cent CI, 0.35–0.49) (Siegfried 2005).
However, in reality, the results should be presented as absolutes—which shows a much less impressive impact, i.e. the absolute risk reduction (ARR) is low, ARR 0.014 (95 per cent CI, 0.07–0.2), corresponding to a Number Needed to Treat of 72 (95 per cent CI, 50–143). This means that, contrary to media reports, 72 circumcisions will need to be conducted in order to prevent one infection over a period of 2 years. When stated in this stark contrast to the media reports, the individual protection is low, but the population effects are high.
The reasons for this dilution of the efficacy results from the trials are several. First, the trials we included in our meta-analysis were all held in settings with high levels of HIV infection, and patients were provided with education about HIV prevention and on reducing their number of sexual partners. Although such intensive education programmes may be included as ministries of health in Africa begin to roll out circumcision as a prevention tool, historically there has been a persistent, yet inadequate, response to prevention strategies in the past. Second, a potential difference between the trial settings and many settings in Africa is the age and specific practice of circumcision. In the included trials, young sexually active males were enrolled. Participants had to be uncircumcised prior to enrolment. In many settings in Africa, even those with high prevalence of HIV, circumcision is commonplace, and traditional healing practices and religion will either initiate circumcision at birth or as a coming of age ceremony. While circumcising at birth is arguably the safest surgical procedure for the patient, and reduces the likelihood of failing the intervention (i.e. having sexual intercourse before the scarring has healed), it would take a generation to determine its population effect. Rwanda, one country in Africa with a comparatively high HIV rate, has chosen to implement the widespread circumcision of adult males in an effort to bring about an immediate public health effect (Anon 2007). Circumcision will be available in ministry clinics free of charge. As a response to unprofessional circumcision strategies, such as traditional healers, many ministries of health have mandated that circumcision aspects of rituals be performed at clinic settings. The success of this mandate will be challenging, but appears to be accepted in some settings (Mills et al. 2006c).
We do not anticipate that consumer adherence will be a major challenge in circumcision given that the intervention requires only a one-time procedure. The immediate adherence challenges related to circumcision will be abstaining from sexual activity during the healing period.
Step 3: Economic evaluation
This step assesses the efficiency (health benefits [number of lives saved, number of quality]/disability-adjusted life years avoided) obtained for a specific cost (direct, indirect, and where possible the intangible costs expressed in monetary units such dollars, euros, or pounds) of the intervention. That is, whether the intervention is being delivered to those who would benefit from it with an optimal use of resources. Assessment of efficiency should not be done in the absence of adequate evidence of efficacy and valid estimates of cost. Application of an equity lens to this step implies a trade-off between cost-efficiency and population health equity. Priority funding of interventions with the best cost-effectiveness ratios might increase rich–poor differences because the cost of reaching the poor may be higher (e.g. distance to care) and health benefits may be lower. The concept of using equity-adjusted traditional utility metrics such as an Equity Adjusted Quality Adjusted Life Year (EQ-QALY) needs to be developed, as described above.
Applying this to the circumcision example, we need to recognize that there are costs related to any intervention and we need to determine if the cost of the intervention will make it inaccessible for some participant groups. In a cost-effectiveness analysis related to the Auvert et al. RCT, assuming full coverage of the male circumcision intervention in a South African province, with a 2005 adult male prevalence of 25.6 per cent, 1000 circumcisions would avert an estimated 308 (80 per cent CI, 189–428) infections over 20 years (Kahn et al. 2006; Williams et al. 2006). The estimated cost was US$181 (80 per cent CI, US$117–US$306) per HIV infection averted, and net savings were US$2.4 million (80 per cent CI, US$1.3 million–US$3.6 million). With a lower HIV prevalence of, say, 8.4 per cent, the estimated cost per HIV infection averted was US$551 (80 per cent CI, US$344–US$1071) and net savings were US$753 000 (80 per cent CI, US$0.3–US$1.2 million). This cost-effectiveness evaluation estimated that each circumcision costs US$47, a price that we would consider to be exceptionally high in sub-Saharan Africa, and that a widespread circumcision roll-out would substantially reduce those costs. The cost of lifetime clinical care in this assessment was US$11 948 with antiretroviral access and US$3793 without. Nonetheless, the cost of HIV/AIDS treatment in sub-Saharan Africa remains outside of the abilities of many nations, indicating that any costs saved through HIV infections averted would be an important saving for national health budgets.
Step 4: Implementation
The circumcision example emphasizes the importance of explicit consideration of the values and cultural acceptability. Individual rights overtook common public health strategies when, in the 1980s context of a lack of treatment, testing for HIV/AIDS was made voluntary, and disclosure of HIV status was viewed as an individual’s right (Bayer & Fairchild 2006). Far be it for us to question the validity of this approach; we simply wish to display here that community values outweighed stricter public health strategies such as widespread testing or quarantine (Fairchild & Bayer 2004). As we see almost 30 years later, a greater emphasis is now being placed on community protection rather than individual rights and the United States as well as the World Health Organization now recommend routine testing of individuals using an opt-out approach (Mills & Chong 2006; Mills et al. 2006b).
The science of implementation overlaps substantially with the recently burgeoning ‘knowledge translation’ initiatives—see the section below.
Steps 5 and 6: Monitoring and reassessment
Monitoring identifies the importance of process assessments and intermediate outcomes (putting the human and physical resources in place, monitoring the identification, and treatment of those at risk/in need). The main purpose of these steps in the iterative loop is to assess success in affecting mortality and morbidity by socioeconomic group and deciding whether further remediable need exists; if so there needs to be an additional iteration of the loop.
Monitoring of any scaled up programme for circumcision to reduce HIV/AIDS is essential; just applying the evidence alone will be insufficient. There is compelling evidence that some strategies in sexual health have previously had negative outcomes when public health decision-makers were too eager to apply new interventions. The use of nonoxynol-9 spermicidal lubricant, for example, was widely marketed to prevent pregnancies during the 1980s and 1990s. It was also believed to potentially reduce HIV infection. However, when multiple trials were conducted, it became apparent that the intervention was harmful and may have contributed to increased infections (Wilkinson 2002). Recent microbicide trials to reduce HIV infection have had similarly disappointing results, despite the tremendous media attention and policy-maker promises of imminent effectiveness (WHO 2007). It is clear that if circumcision is to be implemented widely, there will be a need for intense monitoring and dealing with problems as they arise (Singh & Mills 2005).
An example of an approach to monitor HIV and other health programmes is the Equity Gauge of the Global Equity Gauge Alliance, funded by Rockefeller Foundation. This group has developed ‘equity gauges’ as a means of tracking gaps in health at the national or sub-national levels (McCoy et al. 2003). This approach to equity includes three pillars: (1) measuring key indicators, (2) public participation, and (3) advocacy. Inclusion of all three pillars will ensure that information is acted upon. Equity Gauges have been or are being developed in Bangladesh, Chile, China, Ecuador, Kenya, South Africa, Thailand, Uganda, Zambia, and Zimbabwe (McCoy 2003).
WHO has developed a Health Metrics Network to enable performance-based monitoring of interventions and health systems. The Network will aim to build transparency and accountability, and ensure that policy decisions are based upon evidence. Equity is central to the proposed data indicators (AbouZahr & Boerma 2005).
Our stepwise illustration of the equity–effectiveness loop concludes here, showing the interface between clinical epidemiology and public health.
We will finish the chapter by describing a few other features of this interface.
Knowledge translation and innovation
This is an aspect of clinical epidemiology that has developed fairly recently, stimulated in part by being singled out for funding by agencies such as the Canadian Institutes of Health Research. This is highly relevant to public health and warrants discussion here.
We define knowledge translation as ‘the synthesis, exchange, and application of knowledge by relevant stakeholders to accelerate the benefits of global and local innovation in strengthening health systems and improving people’s health’ (WHO 2006). Essentially, this involves uptake and translation of knowledge into action. To do this, tailored interventions are needed to reach a range of target groups, including researchers, local and national policy makers, professionals, affected communities, industry, media, and the general public.
However, one of the primary challenges facing these initiatives is to determine the most effective strategies to promote the use and application of research. A useful framework to provide direction for determining what strategies can work is the Ottawa Model of Research Use (OMRU), developed by (Graham and Logan 2004).
This framework describes an evidence-based approach to selecting and tailoring strategies to promote the application of research. The framework consists of six key elements that should be assessed, monitored, and evaluated before, during, and after any knowledge translation effort (see Fig. 6.11.4). Three of the key elements relate to barriers and supports: (1) the structural, social, patients, and economic influences within the practice and policy environment; (2) the attitudes, knowledge, motivation, and skills of potential adopters or target audiences; and (3) the perceptions of the research evidence and innovation developed. The other three key elements are: (4) The implementation intervention strategies for diffusing, disseminating, or implementing research findings; (5) the adoption and use of the innovation; and (6) the impact or outcomes of research use.
Below, we illustrate the importance and the evidence available in knowledge translation in low- and middle-income countries, using each of the OMRU components (Santesso & Tugwell 2006). We propose that each step is assessed across the socioeconomic gradient using the equity-oriented knowledge translation cascade (Fig. 6.11.5), thus ensuring that interventions do indeed benefit the disadvantaged (Tugwell et al. 2006b).
1. Perceptions of the research evidence and innovation: Assessing and evaluating the external factors, such as the characteristics of the users and the environment they work in, is relatively straightforward. Assessing the perceptions of the research evidence, however, is less so. These latter characteristics relate primarily to how the evidence was created and the ease of application. For example, Logan and Graham (2004) explain that if the research evidence has been produced in a rigorous or transparent way or by credible developers, it may be more readily applied. But if the evidence is difficult to apply, not compatible with usual practice or not seen as advantageous, it may not be applied. Negative feelings about the evidence may also hinder its application. Haines et al. (2004) relate these concepts to knowledge transfer to the public. They found that the public is wary of evidence that is not congruent with existing cultural values.
This emphasizes the importance of adapting innovations/ interventions to potential users and the setting, and creating an innovation or intervention that is perceived positively. An example of a clinical initiative that gives a priority to local applicability is the INCLEN’s Knowledge ‘Plus’ Program, which is based on the premise that developing, providing access to, and equipping healthcare professionals to use locally appropriate and equitable guidelines enhances knowledge translation in low- and middle-income countries. The development of these guidelines, or Knowledge Plus Packages, is through a transparent and systematic process according to local health priorities and the healthcare environment, and involves healthcare professionals and stakeholders to capture local or ‘tacit knowledge’. Knowledge Plus Packages for TB, acute respiratory infections, and hyperlipidaemia have been developed and are being applied in Colombia, India, and the Philippines. Tools to locally adapt the guidelines have also been developed for other settings (INCLEN 2004).
2. Potential adopters: Attitudes, knowledge, motivation, and skills. While many of the characteristics and barriers to research use in potential adopters may be similar between high- and low-income countries, some characteristics present unique challenges in low- and middle-income countries. ‘Evidence-based health policy’ and ‘evidence-based medicine’, although now widely included in curricula and continuing professional development in industrialized countries, are only just beginning to be included in the training of health policymakers and professionals in low- and middle-income countries, with a few exceptions such as in the medical schools with INCLEN clinical epidemiology units or centres (see Box 6.11.1). Incentives to change behaviour may be quite different between high- and low-income countries: Peer recognition in a small community with few peers may not be a strong motivator for change, and monetary incentives may carry more weight (Santesso & Tugwell 2003).
3. Policy and practice environment: Knowledge translation cannot take place in a vacuum. It has been recognized that there needs to be minimal human resources, financing, drugs, and supply systems before effective interventions can be delivered. Once these systems are available, then a second order of barriers need assessing including the overall policy environment, political instability, and the quality of governance (Travis et al. 2004).
At this second level, a systematic review of decision making in healthcare management and policy making from Lavis et al. (2004) identifies that conflicts and rivalries between elected officials and civil servants may decrease the application of research evidence. Similarly, interviews with key informants in knowledge translation in low- and middle-income countries emphasize that the application of research may be hindered when the political environment is corrupt and unstable and when there is a lack of financial resources (Santesso & Tugwell 2003). At a clinical level, research evidence may be competing with institutionalized habits, superstitions, traditions, and cultures (Cutler 2004).
4. Determining the implementation intervention strategies: According to OMRU, once researchers have assessed the salient barriers and supports for knowledge translation, they can determine the best strategy to ensure the application of knowledge in their potential users. Research evidence and experiences evaluating knowledge translation strategies for potential user groups also informs decisions about the best strategies to use. A widely used evidentiary resource for knowledge translation evidence is available from The Cochrane Collaboration Review Group on Effective Practice and Organization of Care. This group has developed specific methods for finding, analysing, and synthesizing the evidence about what knowledge translation strategies work to change or improve the behaviour of healthcare professionals to ensure the application of knowledge.
5. The adoption and use of the innovation: Evidence from many reviews of the implementation of research shows that the majority of knowledge translation strategies to improve care that are targeted to healthcare professionals and managers are moderately successful. Unfortunately, it is difficult to point to any one type of intervention that works all of the time since the research shows that there is considerable variation in success within and across interventions (Eccles et al. 2005).
This variation may be especially relevant when trying to determine if this research into knowledge transfer strategies is applicable to low- and middle-income countries. Siddiqi and colleagues (2005) reviewed the literature in low- and middle-income countries; they suggest that the success of these interventions is highly dependent on local factors.
Overall, they found few studies evaluating interventions to influence health professionals’ practice to improve healthcare conducted in low- and middle-income countries. Using the limited evidence from the few studies that were conducted in low- and middle-income countries, the authors report that audit and feedback were effective at improving professional practice but the improvement was short-term and the studies poorly designed; education meetings and educational outreach were effective, particularly when the local needs and barriers to change were addressed. Interventions to involve consumers in public participation have been studied but the success of these interventions to improve the use of research is not clear (Haines et al. 2004).
6. Measuring and monitoring outcomes of knowledge translation strategies: This addresses the same issue as in Steps 5 and 6 of the equity–effectiveness loop in the previous sections. OMRU emphasizes the importance of measuring and monitoring the use of the research/innovation and the resulting health outcomes of knowledge translation strategies as distinct from each other but lie on a continuum. First, it is important to determine whether the research is used or consulted by the target audience or is in place and reaching the target audience—outcomes to measure its use are necessary. If the research is used then the next question to ask is: Does it result in improved health outcomes—the primary objective of any knowledge translation activity?
The distinction between the two is illustrated in the use of bed nets in Tanzania to prevent malaria. The Ifakara Health Research and Development Centre in Tanzania has developed social marketing strategies, which include intense promotion of the use of insecticide-treated mosquito nets in local communities, for example, through play skits performed in a community. The consistent use of, care of, and reapplication of insecticide to the bed nets are obviously key in whether the bed nets decrease malaria and mortality in those communities. Therefore, the researchers assessed the numbers of people who had bed nets, the number of nets in bad condition (e.g. more than seven large holes), how often people treated the nets with the insecticide, and how nets were washed (Nathan 2004). This information provided feedback on whether the innovation was being used as intended and ultimately on whether the innovation would need to be modified and how. To measure the health outcomes, another study was conducted using demographic surveillance techniques. This time the prevalence of malaria and anaemia were measured to indicate the success of the social marketing as a knowledge translation strategy to communities (Killeen 2007).
Other interfaces between clinical epidemiology and public health
Other functions of clinical epidemiology that contribute or interface with public health include screening programmes, health technology assessment, and standards setting (through practice guidelines and quality improvement).
Screening is an important part of primary care, typically done by family physicians and nurse practitioners. Good examples of an evidence-based approach to this include the Canadian Task Force on the Periodic Health Exam (Canadian Task Force on Preventive Healthcare (http://www.ctfphc.org), the US Preventive Services Task Force (http://www.ahrq.gov/clinic/uspstfix.htm), and guidelines for disease promotion in immigrants and refugees in Canada (Pottie et al. 2008). Many of these screening clinical actions need to be integrated with public health services.
Health Technology Assessment, defined by Battista as the bridge between science and policy, also bridges clinical epidemiology and public health. Indeed one of us (PT) has used the same stepwise iterative loop approach to systematizing the methods (Tugwell et al. 1995).
Standards setting through practice guidelines and quality assurance in clinical epidemiology also has many parallels with public health, especially public health units. Clinical guidelines are only as good as the evidence and judgements they are based on. For example, the international GRADE initiative has been developed for users of clinical practice guidelines and other recommendations to provide them with the information needed to know how much confidence they can place in the recommendations. Systematic and explicit methods of making judgements can reduce errors and improve communication. The GRADE system grades the quality of evidence and the strength of recommendations that can be applied across a wide range of interventions and contexts. Judgements about the strength of a recommendation require consideration of the balance between benefits and harms, the quality of the evidence, translation of the evidence into specific circumstances, and the certainty of the baseline risk. It is also important to consider costs (resource utilization) before making a recommendation. Inconsistencies among systems for grading the quality of evidence and the strength of recommendations reduce their potential to facilitate critical appraisal and improve communication of these judgements. This system for guiding these complex judgements balances the need for simplicity with the need for full and transparent consideration of all important issues. These are being used not only for clinical guidelines but also at the policy level, such as for the World Health Organization to assess the evidence for anti-virals for avian influenza (Schunemann et al. 2007).
Conclusion
In summary:
♦ The separation of clinical medicine and public health in the early twentieth century led to a schism between clinical care and the collective action for ‘public good’ over ‘private good’. Epidemiology, the social sciences, and qualitative methods became the domain of public health, and medicine assumed dominance in clinical treatment advances.
♦ Clinical epidemiology has adopted many public health methods; conversely, it has contributed to methods for tackling public health problems.
♦ Clinical epidemiology training has led to many clinicians and other healthcare providers developing research programmes addressing public health problems.
♦ Equity, with a special focus on equity in effectiveness, has recently become a focus of clinical epidemiology, and complements the longstanding concern for this in public health.
♦ The equity–effectiveness framework is just one approach for systematically organizing the clinical epidemiological approach to public health problems.
♦ Clinical epidemiology has a contribution to make to public health.
References
AbouZahr, C. and Boerma, T. (2005). Health information systems: The foundations of public health. Bulletin of World Health Organization, 83, 578–83.
Find This Resource
Acheson, D. (1988). Public Health in England: The report of the Committee of Inquiry into the future development of the public health function. HMSO, London.
Find This Resource
Auvert, B., Sobngwi-Tambekou, J., Taljaard, D. et al. (2006). Corresponding Authors’ reply to: The Potential Impact of Male Circumcision on HIV in Sub-Saharan Africa, PLoS Medicine, 3, e67.
Find This Resource
Auvert, B., Taljaard, D., Lagarde, E. et al. (2005). Randomized, controlled intervention trial of male circumcision for reduction of HIV infection risk: The ANRS 1265 Trial. PLoS Medicine, 2, e298.
Find This Resource
Awasthi, S. and INCLEN Childnet Zinc effectiveness for diarrhea group (2006a). Zinc supplementation in acute diarrhea is acceptable, does not interfere with oral rehydration, and reduces the use of other medications: a randomized trial in five countries. Journal of Pediatric Gastroenterology and Nutrition, 42, 300–5.
Find This Resource
Awasthi, S., Verma, T., and Agarwal, M. (2006b). Danger signs of neonatal illnesses. Bulletin of World Health Organization, 84, 819–26.
Find This Resource
Bailey, R.C., Moses, S., Parker, C.B. et al. (2007). Male circumcision for HIV prevention in young men in Kisumu, Kenya: A randomised controlled trial. Lancet, 369, 643–56.
Find This Resource
Bayer, R. and Fairchild, A.L. (2006). Changing the paradigm for HIV testing the end of exceptionalism. The New England Journal of Medicine, 355, 647–9.
Find This Resource
Birdsell, J.M., Atkinson-Grosjean, J., and Landry, R. (2002). Knowledge translation in two new programs: Achieving “The Pasteur effect”. Canadian Institutes of Health Research, Ottawa, Canada.
Find This Resource
Briggs, C.J., Capdegelle P., and Garner, P. (2001). Strategies for integrating primary health services in middle- and low-income countries: effects on performance, costs and patient outcomes. Cochrane Database Systematic Reviews, 4, CD003318.
Find This Resource
Britton, A., Shipley, M., Marmot, M. et al. (2004). Does access to cardiac investigation and treatment contribute to social and ethnic differences in coronary heart disease? Whitehall II prospective cohort study. British Medical Journal, 329, 318–20.
Find This Resource
Bowers, J. (1974). Schools of Public Health in Latin America. Report of a Macy Conference, Medellin, Colombia, November 17–19, 1974. Josiah H. Macy, Jr. Foundation, New York.
Find This Resource
Cochrane, A.L. (1972). Effectiveness and efficiency: Random reflections on Health Services. Nuffield Provincial Hospitals Trust, London.
Find This Resource
Cochrane, A.L. (1979). 1931–1971: A critical review, with particular reference to the medical profession. In Medicines for the year 2000, pp. 1–11. Office of Health Economics, London.
Find This Resource
Cutler, D.M. (2004). Your money or your life: Strong medicine for America’s health care system. Oxford University Press, New York.
Find This Resource
Daly, J. (2005). Evidence-based medicine and the Search for a Science of Clinical Care. University of California Press, London.
Find This Resource
Davis, D., Evans, M., Jadad, A. et al.(2003). The case for knowledge translation: shortening the journey from evidence to effect. British Medical Journal, 327, 33–5.
Find This Resource
DerSimonian, R. and Laird, N. (1986). Meta-analysis in clinical trials. Controlled Clinical Trials, 7, 177–88.
Find This Resource
Diderichsen, F., Evans, T., and Whitehead, M. (2001). The social basis of disparities in health. In Challenging Inequities in Health. From Ethics to Action (eds. T. Evans et al.), pp. 13–23. Oxford University Press, New York.
Find This Resource
Drain, P.K., Halperin, D.T., Hughes, J.P. et al. (2006). Male circumcision, religion, and infectious diseases: An ecologic analysis of 118 developing countries. BMC Infectious Diseases, 6, 172.
Find This Resource
Eccles, M., Grimshaw, J., Walker, A. et al. (2005). Changing the behaviour of healthcare professionals: The use of theory in promoting the uptake of research findings. Journal of Clinical Epidemiology, 58, 107–12.
Find This Resource
Egger, M., Smith, G.D., and Altman, D.G. (eds.) (2001). Systematic reviews in health care: Meta-analysis in context. BMJ Publishing Group. London, (ISBN:0-7279-1488-X).
Find This Resource
Fairchild, A.L. and Bayer, R. (2004). Public health. Ethics and the conduct of public health surveillance. Science, 303, 631–2.
Find This Resource
Feinstein, A.R. and Horwitz, R.I. (1997). Problems in the “evidence” of “evidence-based medicine”. The American Journal of Medicine, 103, 529–35.
Find This Resource
Feinstein, A.R. (1985). Clinical epidemiology; the architecture of clinical research. WB Saunders, Philadelphia.
Find This Resource
Ferriman, A. (2007). BMJ readers choose the “sanitary revolution” as greatest medical advance since 1840. BMJ, 334, 111.
Find This Resource
Fletcher, R. (2001). Alvan Feinstein, the father of clinical epidemiology, 1925–2001. Journal of Clinical Epidemiology, 54, 1188–90.
Find This Resource
Fletcher, R.W. and Fletcher, S.W. (2005). Clinical epidemiology: The essentials (4th edn.). Lippincott Williams & Wilkins, Philadelphia.
Find This Resource
Fox, D.M. (1986). Health policies, health politics: The British and American Experience, 1911–1965, Princeton University Press, Princeton.
Find This Resource
Giuffrida, A., Gosden, T., Forland, F. et al. (2000). Target payments in primary care: Effects on professional practice and health care outcomes. Cochrane Database of Systematic Reviews, 3, CD000531.
Find This Resource
Graham, I. and Logan, J. (2004). Translating research: Innovations in knowledge transfer and continuity of care. The Canadian Journal of Nursing Research, 36 (2), 89–103.
Find This Resource
Gray, R.H., Kigozi, G., Serwadda, D. et al. (2007). Male circumcision for HIV prevention in men in Rakai, Uganda: a randomised trial. Lancet, 369, 657–66.
Find This Resource
Grimshaw, J.M., Shirran, L., Thomas, R. et al. (2001). Changing provider behavior: An overview of systematic reviews of interventions. Medical Care, 39(Suppl 2), I12–I45.
Find This Resource
Guyatt, G.H. and Rennie, D. (2002). User guides to the medical literature. A manual for evidence-based clinical practice. AMA Press, Chicago.
Find This Resource
Gwatkin, D. (2003). How well do health programmes reach the poor? Lancet, 361, 540–1.
Find This Resource
Haines, A., Kuruvilla, S., and Borchert, M. (2004). Bridging the implementation gap between knowledge and action for health. Bulletin of the World Health Organization, 82, 724–31.
Find This Resource
Haynes, R.B., Sacket, D., Guyatt, G. et al. (2006). Clinical epidemiology: How to do clinical practice research. Lippincott, Williams, Wilkins, Philadelphia.
Find This Resource
Higgins, J.P. and Thompson, S.G. (2002). Quantifying heterogeneity in a meta-analysis, Statistics in Medicine, 21, 1539–58.
Find This Resource
Institute of Medicine (1988). The future of public health. National Academy of Health, Washington D.C.
Find This Resource
Killeen, G.F., Tami, A., Kihonda, J. et al. (2007). Cost-sharing strategies combining targeted public subsidies with private-sector delivery achieve high bednet coverage and reduced malaria transmission in Kilombero Valley, southern Tanzania. BMC Infectious Disease. 7, 121.
Find This Resource
Knowles, J.H. (1977). Doing better but feeling worse: Health in the United States. Norton, New York.
Find This Resource
Kramer, M.S. (1988). Clinical epidemiology and biostatistics: A primer for clinical investigators and decision-makers. Springer-Verlag, Berlin.
Find This Resource
Laupacis, A., Sackett, D.L., and Roberts, R.S. (1988). An assessment of clinically useful measures of the consequences of treatment. The New England Journal of Medicine, 318, 1728–33.
Find This Resource
Lavis, J.N., Posada, F.B., Haines, A. et al. (2004). Use of research to inform public policymaking. Lancet, 364, 1615–21.
Find This Resource
Marmot, M. (2001). Inequalities in health. The New England Journal of Medicine, 345, 134–6.
Find This Resource
McCoy, D., Bambas, L., Acurio, D. et al. (2003). Global equity gauge alliance: Reflections on early experiences. Journal of Health and Population Nutrition, 21, 273–87.
Find This Resource
Mills, E.J. and Chong, S. (2006). Lesotho embarks on universal HIV testing. HIV/AIDS policy & law review / Canadian HIV/AIDS Legal Network, 11, 27–8.
Find This Resource
Mills, E.J., Nachega, J.B., Bangsberg, D.R. et al. (2006a). Adherence to HAART: a systematic review of developed and developing nation patient-reported barriers and facilitators. PLoS Medicine, 3, e438.
Find This Resource
Mills, E.J., Nachega, J.B., Buchan, I. et al. (2006b). Adherence to antiretroviral therapy in sub-Saharan Africa and North America: A meta-analysis. JAMA, 296, 679–90.
Find This Resource
Mills, E. and Siegfried, N. (2006). Cautious optimism for new HIV/AIDS prevention strategies. Lancet, 368, 1236.
Find This Resource
Mills, E., Singh, S., Wilson, K. et al. (2006c). The challenges of involving traditional healers in HIV/AIDS care. International journal of STD & AIDS, 17, 360–3.
Find This Resource
Montori, V.M., Devereaux, P.J., Adhikari, N.K. et al. (2005). Randomised trials stopped early for benefit: A systematic review. JAMA, 294, 2203–9.
Find This Resource
Mowatt, G., Grimshaw, J.M., Davis, D.A. et al. (2001). Getting evidence into practice: The work of the Cochrane effective practice and organization of care group (EPOC). The Journal of continuing education in the health professions, 21, 55–60.
Find This Resource
Nathan, R., Masanja, H., Mshinda, H. et al. (2004). Mosquito nets and the poor: Can social marketing redress inequities in access? Tropical medicine & international health: TM & IH, 10, 1121–6.
Find This Resource
Ngalande, R.C., Levy, J., Kapondo, C.P. et al. (2006). Acceptability of male circumcision for prevention of HIV infection in Malawi. AIDS and Behaviour, 10, 377–85.
Find This Resource
Peedicayil, A., Sadowski, L.S., Jeyaseelan, L. et al.: IndiaSAFE Group. Spousal physical violence against women during pregnancy. BJOG: An International Journal of Obstetrics and Gynaecology, 111, 682–7.
Find This Resource
Pottie, K., Robinson, V., Feightner, J. et al. (2008). Evaluation of evidence based literature and formulation of recommendations for clinical preventative guidelines for immigrant and refugees in Canada. CMAJ In press 2008.
Find This Resource
Redelmeier, D.A., Guyatt, G.H., and Foldstein, R.S. (1996). On the debate over methods for estimating the clinically important difference. Journal of Clinical Epidemiology, 49, 1223–4.
Find This Resource
Sackett, D.L., Rosenberg, W.M., Gray, J.A. et al. (1996). Evidence based medicine: What it is and what it isn’t. BMJ, 312, 71–2.
Find This Resource
Santesso, N. and Tugwell, P. (2006). Knowledge translation in developing countries. Journal of Continuing Education in the Health Professions, 26, 87–96.
Find This Resource
Schunemann, H.J., Hill, S.R., Kakad, M. et al. for the WHO Rapid Advice Guideline Panel on Avian Influenza (2007). Rapid Advice Guidelines for pharmacological management of sporadic human infection with avian influenza A (H5N1) virus. The Lancet Infectious Diseases, 7, 21–31.
Find This Resource
Siddiqi, K., Newell, J., and Robinson, M. (2005). Getting evidence into practice: What works in developing countries? International journal for quality in health care: journal of the International Society for Quality in Health Care / ISQua, 17, 447–54.
Find This Resource
Siegfried, N., Muller, M., Deeks, J. et al. (2005). HIV and male circumcision—a systematic review with assessment of the quality of studies. The Lancet Infectious Diseases, 5, 165–73.
Find This Resource
Singh, J.A. and Mills, E.J.(2005). The abandoned trials of pre-exposure prophylaxis for HIV: What went wrong? PLoS Medicine, 2, e234.
Find This Resource
Straus, S., Haynes, R.B., Glasziou, P. et al. (2007). Misunderstandings, misperceptions, and mistakes. Evidence Based Medicine, 12, 2–3.
Find This Resource
Travis, P., Bennett, S., Haines, A. et al. (2004). Overcoming health-systems constraints to achieve the Millennium Development Goals. Lancet, 364, 900–6.
Find This Resource
Tugwell, P., Bennett, K., Sackett, D.L. et al. (1985). The measurement iterative loop: A framework for the critical appraisal of need, benefits and costs of health interventions. Journal of Chronic Diseases, 38, 339–51.
Find This Resource
Tugwell, P., de Savigny, D., Hawker, G. et al. (2006a). Applying clinical epidemiological methods to health equity: The equity effectiveness loop. BMJ, 332, 358–61.
Find This Resource
Tugwell, P., Robinson, V., Grimshaw, J. et al. (2006b). Systematic reviews and knowledge translation. Bulletin of the World Health Organization, 84, 643–51.
Find This Resource
Tugwell, P., Sitthi-Amorn, C., O’Connor, A. et al. (1995). Technology assessment. Old, new, and needs-based. International Journal of Technology Assessment in Health Care, 11, 650–62.
Find This Resource
Wagstaff, A. (2002). Inequality aversion, health inequalities and health achievement. Journal of Health Economics, 21, 627–41.
Find This Resource
Wawer, M.J., Reynolds, S.J., Serwadda, D. et al. (2005). Might male circumcision be more protective against HIV in the highly exposed? An immunological hypothesis. Aids, 19, 2181–2.
Find This Resource
Weiss, N.S. (1996). Clinical epidemiology: The study of the outcome of illness. Oxford University Press, Oxford.
Find This Resource
Wells, G., Beaton, D., Shea, B. et al. (2001). Minimal clinically important differences: review of methods. The Journal of Rheumatology, 28, 406–12.
Find This Resource
White, K.L. (1991). Healing the Schism: Epidemiology, medicine, and the public’s health, pp. 12–13. Springer-Verlag, New York.
Find This Resource
Wilkins, R. and Adams, O. (1983). Healthfulness of life. Institute for Research on Public Policy, Montreal.
Find This Resource
Wilkinson, D., Tholandi, M., Ramjee, G. et al. (2002). Nonoxynol-9 spermicide for prevention of vaginally acquired HIV and other sexually transmitted infections: Systematic review and meta-anlysis of randomised controlled trials including more than 5000 women. The Lancet Infectious Diseases, 2, 613–17.
Find This Resource
Wilkinson, R.G. (1996). Unhealthy societies. The afflictions of inequality. Routledge, London and New York.
Find This Resource
World Health Organization (1999). The World Health Report 1999: Making a difference. Geneva, ISBN 92 4 156194 7.
Find This Resource
World Health Organization (WHO) Bridging the ‘Know–Do’ Gap, Meeting on Knowledge Transition in global Health. Geneva.
Find This Resource
Zwarenstein, M. and Bryant, W. (2000). Interventions to promote collaboration between nurses and doctors. Cochrane Database of Systematic Reviews, 2, CD000072.
Find This Resource