Associations, Predictions, Causes, and Interventions


Idea: Relationships among associations, predictions, causes, and interventions run through all the cases and controversies in this course. The idea introduced in this session is that epidemiology has two faces: One from which the thinking about associations, predictions, causes, and interventions are allowed to cross-fertilize, and the other from which the distinctions among them are vigorously maintained, as in "Correlation is not causation!" The second face views Randomized Control Trial (RCTs) as the "gold-standard" for testing treatments in medicine. The first face recognizes that many hypotheses about treatment and other interventions emerge from observational studies and often such studies provide the only data we have to work with. What are the shortcomings of observational studies we need to pay attention to (e.g., systematic sampling errors leading to unmeasured confounders)?

Initial notes on the Cases

From PT: Ridker et al. show that the conventional risk factors for heart disease in women (as combined in the Framingham score) identify many women as of intermediate risk who are higher or lower risk. The new Reynolds Risk Score does a much better job, primarily it seems by including the risk marker cReactive Protein. Both scores are based on observations not randomized trials.
Jick presents evidence that statin treatment was associated with lowered risk of dementia but the Alzheimer Research Forum presents the more recent assessment (using RCTs) that statins are not protective against dementia. The discrepancy seems to be undetected bias in which patients get prescribed statins.
The case of hormone replacement therapy as a protection against heart disease (Stampfer 1990) is another, more significant instance of mismatch of observational results and RCTs -- see Stampfer 2004 & Pettiti for analyses of the discrepancy.
Lawlor et al take on the case of vitamin C and Davey-Smith & Ebrahim provide a quick review of a number of cases.
Try to get a handle on the different kinds of explanation for the discrepancies, including physician bias in who gets prescribed a treatment, residual confounders, and reverse causation. Gordis may help here.



Substantive statement

This week’s readings present us with several challenges that speak to the fundamental nature of epidemiological inquiry and even the research endeavor in general. How do we design studies that allow us to answer discrete questions about particular health outcomes, processes, or behaviors while taking into account the myriad factors that may affect our object? How do we account for all possible variables while also constructing a theoretically tight and sound model with enough explanatory power to yield a useful result? How do we adapt when more and contrary research becomes available? How do we construct truly transdisciplinary approaches that allow us to view varying methods as complementary rather than in competition, e.g. qualitative and quantitative; population health and clinical health; randomized controlled trials and observational study, etc.

These overarching questions and many others arise when reviewing the readings for this week’s class and considering the larger implications for the field of epidemiology. From these broad questions, the readings begin to help us delve deeper into some of the more specific considerations we must take into account to address their implications.

Much of the debate that we can elicit from the readings pertains to differential findings about the effects of certain treatment mechanisms for cardiovascular disease and dementia, primarily the tendency of results gleaned from one method of research, such as randomized controlled trials, to contradict the results of previously conducted studies. For example, the article from the Alzheimer Research Forum details a speech given by the representative of a health care system indicating that several prospective (observational) studies conducted to examine the efficacy of statin prescription on preventing and/or treating dementia were issuing findings contradicting the results of earlier case-control studies, such as that described by Jick et al., that exhibited a negative correlation between prescribing statins and the risk of developing dementia. Following the article are several comments by other researchers and interested parties trying to tease out the reasons for the conflicting findings. The Stampfer and Colditz articles review numerous studies examining the effects of estrogen-replacement therapy, all of which reach different conclusions – some only slightly different, others demonstrating opposite results. Ridker et al. determine that conventional means of predicting cardiovascular risk are not sound enough to restrain a proliferation of false positives; thus they develop a new set of algorithms designed to improve the prediction, with the addition of a new biomarker for evaluation. Finally, Davey Smith and Ebrahim give us an overview of studies exemplifying some of the pitfalls necessary to account for when tackling epidemiologic research before outlining their own alternative approach to common methodologies.

These debates and questions do much to serve the field of research in pushing researchers to regularly re-evaluate their assumptions and discover innovative means of dissecting disease and its correlates. Such debate may also gain the attention of policy makers and others in a position to advance (or detract from) the progress of health science and promotion. However this prompts me to make two important points, which taken together pose something of a conundrum to the public health field: 1) epidemiological research and its surrounding dialectic do not exist in a vacuum; policies at all levels are enacted using the information generated therein, and 2) because epidemiology is concerned with real threats to the health of populations, it is not sufficient to wait around for the perfect study to be conducted before taking action to prevent and/or treat disease occurrence.

As Rose argues: “Certainty is not a prerequisite for action. A sick patient can expect from the physician only a reasonable confidence that the diagnosis is right and that the treatment is likely to do more good than harm. Prevention should be judged in the same way, so that action may then proceed alongside continuing research and evaluation, recognizing that new evidence may mean a change of policy.” (1)

To explore the two caveats above a bit further we need only refer again to the readings for this week. The example of prescribing statins for staving off dementia is salient. Findings indicating that this method would indeed be preventive encouraged physicians to begin prescribing statins, people concerned for their health to begin taking them of their own accord, and communities in general to gain hope for a way to manage dementia. Though perhaps not indicative of formal policy change, the findings of the initial research in this area have certainly been employed in a widespread manner, such that when contrary findings began to be leaked, an uproar ensued in the research and community circles most concerned with dementia – a phenomenon we can see in the number of responses to the Alzheimer Research Forum article. However it would be difficult to fault those physicians, patients, advocates and researchers who placed emphasis on the benefits of statins; the initial research was seemingly sound, both theoretically and methodologically. The researchers had no mal-intent; instead they had a limited body of reference literature with which to work and an innovative idea to explore. Without such work, future researchers would not have had a model to disprove. Thus we can refer again to Rose’s point about certainty.

On the other hand, this is not to argue that researchers do not have a real responsibility in conducting the soundest research possible and anticipating possible reactions to their findings once released. For example, Judy’s annotated reference for this week provides an excellent example of an area in which research is fraught with the usual measurement issues as well as profound emotional, political and ethical considerations, an area which also illustrates the conundrum described above. As Judy points out in her annotation, suicide has such a low base rate that it is very difficult to measure in the first place. Add to that questions of suicidality (i.e. suicidal thoughts and behavior) which has no clear definition in the field and already researchers have a major challenge on their hands. Beyond these concerns are issues of data availability and integrity, the ethical issues Judy mentions with regard to conducting RCTs, the emotional component to studying suicide, and many other challenges.(2, 3) Fortunately these obstacles have only delayed research in the area of suicide prevention rather than stopping it entirely.

One avenue that has seen much research activity and public debate in recent years is the relationship between selective serotonin reuptake inhibitors (SSRIs) and suicidality. The question is whether SSRIs are protective against suicide or if they actually trigger suicidal thoughts. The reference population for this question has generally been children and adolescents, and the results of selected RCTs (certain results demonstrated a very slight up tick in probability of attempting suicide in experimental groups of adolescents) were alarming enough to prompt the Food and Drug Administration (FDA) to issue a black box warning on SSRI prescriptions – one of the most stringent actions the FDA can take with regard to prescription medications. (4-7)

However, if we dig deeper we find that almost all research that has previously been conducted on the influence of SSRIs on depression and suicidality has been conducted with adults. There is little understanding of how SSRIs interact in adolescent brains and even less understanding of the interaction in child brains. (8, 9) In fact there is little agreement on whether certain mental illnesses that may be treated with SSRIs can even be diagnosed before a certain threshold age; for example it is only very recently that some mental health professionals have acknowledged the possibility of diagnosing bipolar disorder in children and adolescents. (10) Furthermore, subsequent researchers have argued that the reporting of results for some of the implicated studies includes outliers in the overall findings such that the suicide risk has been inflated for political reasons. (6, 7) The FDA being a political body may also have responded to outside pressure to issue the black box warning given the high level of emotion associated with suicide deaths. Additionally, the media cacophony surrounding the issue has drowned out the voices of prominent mental health professionals who point out that any medication prescribed for mental illness must be taken with great precaution and that the issue may well be the level of monitoring between doctor and patient rather than any threat intrinsic to this class of medications.

These two examples offer some orientation to the practical implications involved when developing theories, designing studies and interventions, and offering predictions or arguments related to disease/disorder etiology, incidence and prevalence. Given this more conceptual look at these various considerations, I hope that we can all come to class with an eye to understanding some of the more technical aspects of epidemiological inquiry beginning with some of the points raised in the articles for this week. In other words, what are some of the details that contribute to solid research design and what are some of the traps to avoid on the way?

References

  1. Rose G. The Strategy of Preventive Medicine. Oxford: Oxford University Press, 1992.
  2. Maris RW, Canetto SS, McIntosh JL, et al. eds. Review of Suicidology. New York: The Guilford Press, 2000.
  3. De Leo D, Evans R. International Suicide Rates And Prevention Strategies. Cambridge: Hogrefe & Huber Publishing, 2004.
  4. Hall W, Lucke J. How have the selective serotonin reuptake inhibitor antidepressants affected suicide mortality? The Australian and New Zealand Journal of Psychiatry 2006;40:941-50.
  5. Hamrin V, Scahill L. Selective serotonin reuptake inhibitors for children and adolescents with major depression: current controversies and recommendations. Issues in Mental Health Nursing 2005;26:433-50.
  6. Healy D. Did regulators fail over selective serotonin reuptake inhibitors? British Medical Journal 2006;333:92-5.
  7. Rihmer Z, Akiskal HS. Do antidepressants t(h)reat(en) depressives? Toward a clinically judicious formulation of the antidepressant-suicidality FDA advisory in light of declining national suicide statistics from many countries. Journal of Affective Disorders 2006;94:3-13.
  8. Cheung A, Emslie G, Mayes T. Review of the efficacy and safety of antidepressants in youth depression. Journal of Child Psychology and Psychiatry 2005;46:735-54.
  9. Gibbons R, Brown C, Hur, K, , Marcus S, et al. Early evidence on the effects of regulators' suicidality warnings on SSRI prescriptions and suicide in children and adolescents. American Journal of Psychiatry 2007;164:1304-6.
  10. Wozniak J. Recognizing and managing bipolar disorder in children. Journal of Clinical Psychiatry 2005;66:18-23.

Response to statement

Lousia does a wonderful job in not only summarizing the readings but uncovering the true complexities that accompany research. The increasing desires for complex designs have caused researchers to overlook simple factors that could bias their results. The stress placed on the reliability and validity of research has threatened the meaning of the findings and the importance of interventions. The readings for this week also bring in many aspects of last weeks discussion and reading that question the use and development of concrete definitions for factors such as SES or race that could cause bias.

Lawlor and Ridker do shed some light on the debate of “nature” versus “nurture” as they call for the addition of parental factors and childhood SES into the design that could account for the effect of life long events on health outcomes. These readings really seem to justify the need for environment and genetic contribution to differences by accounting for both.

Overall, the readings lead to questions surrounding the ability to control the research design. Some of the control issues seem trivial but are repeated over and over. One example would be the dementia patients that are included in the study as “newly diagnosed” but could have mild to severe dementia. This problem is also seen in women who seek estrogen therapy as they might go to the doctor more than someone that is not seeking estrogen therapy after menopause. These two trivial design problems can account for a large part of the error and a need for a sturdy sample collection. These problems also underline the readings in that researchers really need to take a step back and look at their findings and understand it from a lay person’s point of view. They also need to back off the need for a “significant”, “reliable”, or “valid” finding to look at the characteristics of the sample and the need for intervention (JN).



Class Participation Aides


Examples of Studies Using Randomized Controlled Trials

Woodall WG, Delaney HD, Kunitz SJ, Westerberg VS, Zhao H. Randomized Trial of a DWI Intervention Program for First Offenders: Intervention Outcomes and Interactions With Antisocial Personality Disorder Among a Primarily American-Indian Sample. Alcoholism: Clinical & Experimental Research 2007; 31(6):974-987.

Background: Randomized trial evidence on the effectiveness of incarceration and treatment of first-time driving while intoxicated (DWI) offenders who are primarily American Indian has yet to be reported in the literature on DWI prevention. Further, research has confirmed the association of antisocial personality disorder (ASPD) with problems with alcohol including DWI. Methods: A randomized clinical trial was conducted, in conjunction with 28 days of incarceration, of a treatment program incorporating motivational interviewing principles for first-time DWI offenders. The sample of 305 offenders including 52 diagnosed as ASPD by the Diagnostic Interview Schedule were assessed before assignment to conditions and at 6, 12, and 24 months after discharge. Self-reported frequency of drinking and driving as well as various measures of drinking over the preceding 90 days were available at all assessments for 244 participants. Further, DWI rearrest data for 274 participants were available for analysis. Results: Participants randomized to receive the first offender incarceration and treatment program reported greater reductions in alcohol consumption from baseline levels when compared with participants who were only incarcerated. Antisocial personality disorder participants reported heavier and more frequent drinking but showed significantly greater declines in drinking from intake to post-treatment assessments. Further, the treatment resulted in larger effects relative to the control on ASPD than non-ASPD participants. Conclusions: Non-confrontational treatment may significantly enhance outcomes for DWI offenders with ASPD when delivered in an incarcerated setting, and in the present study, such effects were found in a primarily American-Indian sample.

Kronborg H, Væth M, Olsen J, Iversen L, Harder I. Effect of early postnatal breastfeeding support: a cluster-randomized community based trial. Acta Pædiatrica 2007; 96(7):1064-1070.

Aim: To assess the impact of a supportive intervention on the duration of breastfeeding. Design and setting: A community based cluster-randomized trial in Western Denmark. Subjects: Fifty-two health visitors and 781 mothers in the intervention group, and 57 health visitors and 816 mothers in the comparison group. Intervention: Health visitors in the intervention group received an 18-h course. The intervention addressed maternal psychosocial factors and consisted of 1–3 home visits during the first 5 weeks post-partum. Health visitors in the comparison group offered their usual practice. Main outcome measure: Duration of exclusive breastfeeding during 6 months of follow-up. Results: Mothers in the intervention group had a 14% lower cessation rate (HR = 0.86 CI: 0.75–0.99). Similar results were seen for primipara, and multipara with previously short breastfeeding experience. Mothers in the intervention group received their first home visit earlier, had more visits and practical breastfeeding training within the first 5 weeks. Babies in the intervention group were breastfed more frequently, fewer used pacifiers, and their mothers reported more confidence in not knowing the exact amount of milk their babies had received when being breastfed. Conclusion: Home visits in the first 5 weeks following birth may prolong the duration of exclusive breastfeeding. Postnatal support should focus on both psychosocial and practical aspects of breastfeeding. Mothers with no or little previous breastfeeding experience require special attention.

Examples of Prospective Cohort Studies

Laaksonen E et al. Socioeconomic circumstances and common mental disorders among Finnish and British public sector employees: evidence from the Helsinki Health Study and the Whitehall II Study. International Journal of Epidemiology 2007; 36(4): 776.

Background: Common mental disorders do not always show as consistent socioeconomic gradients as severe mental disorders and physical health. This inconsistency may be due to the multitude of socioeconomic measures used and the populations and national contexts studied. We examine the associations between various socioeconomic circumstances and common mental disorders among middle-aged Finnish and British public sector employees. Methods: We used survey data from the Finnish Helsinki Health Study (n=6028) and the British Whitehall II Study (n=3116). Common mental disorders were measured by GHQ-12. The socioeconomic indicators were parental education, childhood economic difficulties, own education, occupational class, household income, housing tenure and current economic difficulties. Logistic regression analysis was the main statistical method used. Results: Childhood and current economic difficulties were strongly associated with common mental disorders among men and women in both the Helsinki and the London cohort. The more conventional indicators of socioeconomic circumstances showed weak or inconsistent associations. Differences between the two cohorts and two genders were small. Conclusions: Our findings emphasize the importance of past and present economic circumstances to common mental disorders across different countries and genders. Overall, our results suggest that among employee populations, the socioeconomic patterning of common mental disorders may differ from that of other domains of health.


Examples of Cross-sectional and Case Control Studies

Freedman VA, Schoeni RF, Martin LG, Cornman JC. Chronic Conditions and the Decline in Late Life Disability. Demography 2007;44(3): 459-477.

Using data from the 1997–2004 National Health Interview Survey (NHIS), we examine the role of chronic conditions in recent declines in late-life disability prevalence. Building upon prior studies, we decompose disability declines into changes in the prevalence of chronic conditions and in the risk of disability given a condition. In doing so, we extend Kitigawa's (1955) classical decomposition technique to take advantage of the annual data points in the NHIS. Then we use respondents' reports of conditions causing their disability to repartition these traditional decomposition components. We find a general pattern of increasing prevalence of chronic conditions accompanied by declines in the percentage reporting disability among those with a given condition. We also find declines in heart and circulatory conditions, vision impairments, and possibly arthritis and increases in obesity as reported causes of disability. Based on decomposition analyses, we conclude that heart and circulatory conditions as well as vision limitations played a major role in recent declines in late-life disability prevalence and that arthritis may also be a contributing factor. We discuss these findings in light of improvements in treatments and changes in the environments of older adults.

Pasic J, Russo JE, Ries RK, Roy-Byrne PP. Methamphetamine Users in the Psychiatric Emergency Services: A Case-Control Study. American Journal of Drug & Alcohol Abuse 2007;33(5): 677-686.

The purpose of this study is to examine the sociodemographic, clinical, and service use characteristics of patients with positive methamphetamine (MA) urine toxicology and compare with non-MA users seen in an urban Psychiatric Emergency Services (PES). One hundred twenty patient charts were extracted for demographics, mode of arrival, clinical information, medication treatment of MA-intoxication, and disposition. Compared with non-MA patients, MA patients were significantly younger, male, referred by police, with cardiac symptoms, psychosis, dysphoria, past substance use, and were less likely to have a diagnosis of Schizophrenia, a past psychiatric history/hospitalization, and a history of suicide attempts. Subsequent hospitalization rates did not differ. MA patients treated with medications more readily accepted the referral to chemical dependency treatment. This study shows that hypertension and tachycardia upon arrival to the PES, symptoms of dysphoria and psychosis, past substance use and not having the diagnosis of Schizophrenia are all related to methamphetamine use.


Annotated additions by students


Moller, Hans-Jurgen (2006) Evidence for beneficial effects of antidepressants on suicidality in depressive patients: A systematic review, European Archives Psychiatry Clinical Neuroscience, 256:329-343

The role of antidepressants in suicide prevention is a major public health question, given the high prevalence of both depression and depression related suicidality. Randomized control group studies seem to be the best basis for statements about the suicide risk of certain anti-depressants. However, there are many methodological pitfalls inherent in the design of these studies. The low basal prevalence of suicidal behavior has the consequence of it being almost impossible to perform a control group study with an adequate statistical power to differentiate between the outcome results of two treatment groups as the suicide numbers are small, even in a large control group study. Most studies for ethical reasons exclude patients with serious suicide thoughts. The principal ethical aim of all studies is to avoid harm to patients and this conflicts with the scientific objectivity of these studies. This eliminates a lot of potential suicides from the studies and almost all exclusion criteria includes co-morbidity with other personality disorders that may aggravate suicide risks. Because of these limitations other epidemiological analyses such as naturalistic follow up studies have been used with impressive results.

Along with other countries, Sweden, United States, Hungary and Denmark have conducted large population suicide studies. Controlling for unemployment and alcohol abuse, there appeared to be a decrease in suicide rates in the population that followed the widespread use of antidepressants. A large Swiss cohort (406) of uni-polar and bi-polar depressed hospital patients (admitted in a two year time frame) was followed over a 34 year period by Angst et al., and the results were published in 2002.. The mortality rate was 76%, which is considerably higher than the standardized mortality rate for the general population and rates for suicide, and circulatory disorders was twofold higher than in the general population. The data is suggestive of a positive drug effect and certainly it raises the hypothesis of a positive effect of a long term medication with antidepressants or a combination of antidepressants with neuroleptics or lithium. Angst cautions that this follow up study has to be interrupted with the greatest caution. However he maintains that the presented results have a certain power of persuasion with respect to the suicide-prophylactic effect of antidepressants in the long-term treatment of patients with mood disorders. These studies present a good argument in favor of an epidemiological naturalistic approach and cohort studies to ascertain if antidepressants curtail the suicide rate. (JG)

Samet, J.M. & Munoz, A. (1998). Evolution of the cohort study. Epidemiologic Reviews, 20 (1): 1-14.
This review article provides an historical introduction to the cohort study design. Beginning with an interesting tidbit about the origin of the word "cohort" (from the Latin cohors, referring to warriors and the notion of a group of people proceeding in time) the terms and definitions that have been used to describe this type of study design are reviewed. The evolution of the cohort design dates back to the life tables developed by Graunt and Halley, but the real impetus for collecting data on mortality and disease came from the emergence of the modern insurance industry. As a result of his work on tuberculosis, Wade Hampton Frost was credited with developing methodological advances relating to the analysis of vital statistics data as well as the use of the retrospective design.

The authors date the late 1940s and 1950s as the beginning of the contemporary era of cohort studies. The landmark Framingham study, the survivors of the atomic bombs in Japan, and the Colorado Plateau uranium miners study all began around this time. By the 1970s, large multisite studies such as the Multicenter AIDs cohort study aimed to be representative of the national population. The article details the development of multivariate methods of analysis and multiple regression techniques before concluding with an extended example of the application of these methods to the Multicenter AIDs Study. (JC)

Elliott, J. & Shepherd, P. (2006). Cohort profile: 1970 British Births Cohort (BCS70). International Journal of Epidemiology, 35: 836-843.
This article describes the aims and development of one of the key multidisciplinary studies that is ongoing in the UK at this time. The BCS70 aims to track all the babies born in England, Scotland, and Wales in a single week in 1970, some 17,500 subjects at its start. Due to the very long study period involved, it is instructive to learn how the funding and responsibility for the study has changed hands over the course of the 30+ years.

The BCS70 was designed as a multipurpose study that – in addition to health – has collected information on the subjects’ family backgrounds, education, and on social development. The authors discuss difficulties inherent in following up participants, noting that when they left school other methods were needed to maintain contact, one of which is an annual birthday card mailed to all the subjects. The article briefly discusses what has been measured in the study; key findings that have been published (300 publications to date); prenatal and perinatal antecedents of health problems; social circumstances and health outcomes; and predictors of adult health status – all referenced in the bibliography.

Of particular interest is the fact that as this is the 3rd in a series of four British Birth Cohort Studies, cross-cohort comparisons have been made. Such studies have examined health outcomes such as hay fever, eczema, asthma, psychological disorders, and health problems and illness in adulthood. The main weakness identified by the authors of this cohort study is that it does not have the ethnic diversity that is found in today’s population. In addition, spotty funding in the 1980s and 1990s had an effect on the timing and some of the content of the follow ups. Strengths include the large study sample and standardization of objective measures, tests and scales. Data from the BCS70 is freely available to non-commercial users. The next follow up is planned for 2008 when the cohort members turn 38 years old. (JC)

Lawlor, D. et al (2005). Childhood socioeconomic position, educational attainment, and adult cardiovascular risk factors: the Aberdeen Children of the 1950s cohort study. American Journal of Public Health, 95 (7): 1245-1251. This study sought to assess the associations of childhood socioeconomic position (SEP) with cardiovascular risk factors (smoking, binge alcohol drinking, and being overweight) and to examine the role of educational attainment and cognitive function in these associations in a cohort born in Aberdeen, Scotland, between 1950 and 1956. The original cohort participants numbered 12,150 children. Information about their physical characteristics and the course of their mothers’ pregnancies were obtained from the Aberdeen Maternity & Neonatal Databank. The cohort was reinitiated in 1999 and the data for the present study was collected from a health questionnaire mailed to the surviving 7184 cohort members between 2000 and 2002.

The researchers found that strong graded associations existed between social class at birth (SEP defined by father’s occupation) and smoking, binge drinking and being overweight. Adjustment for educational attainment completely attenuated these associations leading the researchers to conclude that educational attainment is an important mediating factor in the relations between socioeconomic adversity in childhood and smoking, binge drinking and being overweight in adulthood. In the discussion section of the article, the astonishing claim is made that: “Simply on the basis of knowledge of the occupation of a participant’s father at the time of the participant’s birth, one could have predicted whether this individual would be likely to smoke, engage in binge drinking, and be overweight 40 years after he or she took part in the initial survey in 1962” (p. 1249).

I would like to hear from others who read this article whether such a claim is warranted, notwithstanding the fact that other articles cited by the author contain similar findings. (JC)



Fonda, S. J., Bertrand, R., O’Donnell, A., Longcope, C., & McKinlay, J. B. (2005). Age, hormones, and cognitive functioning among middle-aged and elderly men: Cross-sectional evidence from the Massachusetts male aging study. The Journal of Gerontology, 60A(3), 385-390.

It is an example of a cross-sectional study on cognition. Many types of hormones were used in the study as independent variables to look at the effect on cognition. There were some strong relationships among some hormones and cognition (results found on page 388 followed by questions).
Age was strongly associated with cognitive decline (result found on page 387). The study done at one point of time compared cognition and age differences, and it showed that older participants showed lower cognitive functioning than younger participants on the test. In their study, they concluded that “hormones do not mediate the age-cognition relationship.” (km)

Col NF, Pauker SG. The discrepancy between observational studies and randomize trials of menopausal hormone therapy: did expectations shape experience? Ann Intern Med 2003;139:923-929

The effect of hormone therapy on CHD has been shown through observations and RCT to show that women that choose hormone therapy were healthier than women that did not choose hormone replacement. This data comes from the Nurse Health study (NHS). The differences in the findings can be accounted for beliefs held by the patients, clinicians, and investigators. Future studies should look the effect of HT on CHD of users and non users to understand the real impact of HT on CHD (JN).

DCRI: Research Activities: Landmark Projects www.dcri.duke.edu/research/landmrkproj

This web site lists several of the RCT trials I was referring to during class last week. It is designed and monitored by Duke Clinical Research Institute (DCRI). This institute focuses primarily on the following topics

1. Acute Coronary Syndromes

2. Cardiovascular Cath Lab

3. Endocrinology

4. Gastroenterology

5. Heart Failure

6. Nephrology

What I really like about this website is the ongoing account of follow-up research for this disease process. Each step focused on a particular aspect of either anticoagulation or thrombolytic therapy; utilization of these drugs in relationship to coronary artery disease and then the comparison with different coronary procedures utilized in the cath lab. Perhaps I really like it because a). It’s my specialty and/or b). It’s something concrete (measurable) that I can grasp. If you would like to read more of each of these RCT’s, the Institute has listed the journal articles that have documented each of these trials. (SA)