Variations in health care (by place, race, class, gender)
Idea: Inequalities in people's health and how they are treated are associated with place, race, class, gender, even after conditioning on other relevant variables.
Guidelines for annotations
Notes and annotations from 2007 course
Initial notes from PT
Annotations on common readings
Alter, et al (2007)
The article was about a research study to assess whether the Canadian Universal Health Care System provides equal access to healthcare based upon a citizen’s needs rather than income.
In turn, Alter and his colleagues measured the effects of socioeconomic status to invasive cardiac procedures and mortality after a heart attack.
They examined the performance in Ontario by measuring income and access to invasive cardiac procedures and mortality one year after a heart attack.
The sample size consisted of 51,591 patients and was from patients from April 1994 to March 1997.
Alter and his colleagues controlled for variables such as age, sex, co-morbidities, etc. The data was acquired from the Ontario Myocardial Infarction Database, Hospital discharge abstracts, insurance claims, the Ontario Registered Persons Database, and the 1996 official Canadian census data. The hospitals were characterized into four categories; no onsite facilities, onsite facility, for angiography only and onsite facility for angiography and revascularization. The neighborhoods were divided into five categories according to median personal income acquired from the 1996 official Canadian census. The study concluded that despite Canada’s Universal Health Care System a individual’s socioeconomic status affected access to cardiac services and increased the prevalence of mortality. (KP, 10/20/09)
Krieger, et al. (2005)
The article was about obtaining a “true” picture of United States Socioeconomic status and racial/ethnic health inequalities. To examine this relationship the Public Health Disparities Geocoding Project was started. Many times socioeconomic data is lacking in US public health surveillance systems 70% of the 467 US public health objectives for the 2010 lack socioeconomic targets. The Public Health Geocoding Project categorizes individuals in relation to the socioeconomic characteristics of the area in which they reside. Krieger and her team used Massachusetts and Rhode Island resident from the 1990 census. In addition they obtained data from the Massachusetts Department of Public Health and Rhode Island Department of Health. A total of 760,000 records were included in the data sets, of which 98% were geocoded to the census tract. They analyzed the data in 5 ways; determined number of cases and population in each census tract, calculated relevant age annual incidence rate or proportion, compared the rates with Healthy People 2010, quantified and graphed outcomes of socioeconomic gradient, and explored the causation of socioeconomic inequalities in health due to racial/ethnic health disparities. As a result, they demonstrated the importance of the routinely monitoring US socioeconomic inequalities in health by using geocoding. In addition, they concluded that for women and men in all racial/ethnic groups, socioeconomic deprivation contributes to racial/ethnic health disparities in more than half of the cases studied (KP, 10/20/09)
Annotated additions by students
The article by Davey-Smith was helpful in recognizing challenges in working with socioeconomic variables, as he cautions researchers against ignoring difference and confounders. Specifically, he advises against using ethnicity as a proxy for socioeconomic position, and advocates for incorporating both in quantitative models. He also advocates in favor of incorporating more than just one socioeconomic variable, as well as looking at socioeconomic position over the life-course. In some ways this article seemed an extension of the discussion in class last week – can one really argue for car ownership, or housing tenure/quality as an indicator of income? Of SES? Of causality re: health outcomes? He concludes that the complexity of macro-effects re: social determinants requires both a broader view and inclusion of variables pertaining to individual outcomes, so as to avoid the mistake of focusing solely on individual socioeconomic position – as it limits the scope of research and findings.
Marmot and Wilkinson argue that researchers should look beyond material privation to examine psychosocial effects on variation in health outcomes, particularly relative deprivation concerning individual agency and control. They make the case, quite persuasively, that material goods by way of consumption and income serve as a proxy for reduced subordination and increased security in the workplace, and thus greater self-esteem and satisfaction. Marmot and Wilkinson once again assert their hypothesis that it is relative inequality of income, rather than absolute inequality of income, that leads to stress and disparate health outcomes. They respond to Lynch’s attack on their theory by citing the structural foundations for psychosocial effects and advocate in favor of structural change, as compared to placing the onus on individuals to change their circumstances.
The article by Wright et al on asthma incidence and severity in the context of neighborhood characteristics continues the theme of variation in health outcomes attributable to SES context. In this study of asthma among children in low-income urban settings, the researchers found a correlation between asthma, stress, and exposure to violence that suggests the need for addressing these intervening variables. One of the most interesting findings in this study is that smoking was not found to be associated with asthma attack incidence, which the authors attributed to smoking serving as a stress/tension release. This flies in the face of current public health calls for ‘smokefree’ housing policies in an attempt to reduce asthma incidence and severity among low-income children in public housing. (posted by AH 10-19-09)
The Cost Conundrum (Gawande, The New Yorker, 2009)
The past year has brought more talk about healthcare than we have seen in years, since the President Clinton’s attempt to pass universal healthcare failed a decade and a half ago. Those who are working to get the healthcare reform criticize the healthcare system in the United States for being too expensive, limited in coverage, and mainly in service of the wealthy – private insurance companies and doctors who rather provide care for their own profit than for the health benefit of their patients.
Gawande in his article describes, compares, and contrast those who are making “big bucks” in medical “industry” (McAllen, Texas) and those who are making strides to provide more efficient care while having the patient (not the profit) come first (The Mayo Clinic). In the McAllen model, the cost is high due to the overuse of medicine from over-treating patients and over-prescribing tests and procedures. In the Mayo model, high-quality-low-cost-let’s think and work together is the main principle. In the Mayo model a doctor is a health care provider; in the McAllen model a doctor is more a businessman than a provider. What do these models mean for patients? The care patients receive is according to their means, not according to their (health) needs. It is according to what their insurance covers, not according to what their health condition requires. What do these models mean for doctors? Other than their conscience and the desire to treat patients, the incentives to take the “patient-comes-first” route are scarce. And more they see doctors who make “big bucks” more they will feel challenged to follow the Hippocrates Oath. In this challenge Gawande sees the main problem healthcare in the U.S. is facing, not in the choice between public or private insurance (DBJ, 2009).
The theme of this week seems to be less of variations or disparities, and more of how to measure, track, and talk about those variations. Alter’s (1999) article shows that socioeconomic (SES) status affects quality of care in Canada, where health care is universally accessible, pointing to other, more structural or sociological sources of inequity in care. Krieger et al (2005) propose a method of tracking SES so that health outcomes and objectives can take this critically important factor into account. Krieger et al point out the need for baseline data, a common, usable metric, and data collection and provide a means for acquiring those things.
Dunn’s (2007) article is more of an introduction to other studies and conversations about these issues. Dunn and the other authors he writes about recognize the need to contextualize health, most notably in neighborhoods (which is similar to SES). This point cannot be overstated. Historically, epidemiology has been blind to the structural issues of SES and place, which are very much intertwined with race and ethnicity. These articles point to a new path of epidemiology to not only take these factors into account, but to provide an empirical basis for research that includes discussion on SES and place.
Dunn’s article made me wonder about the difference between neighborhood and community and whether those two concepts can be measured and contrasted. For example, he mentions Stockdale et al (2007)’s article about whether neighborhood factors influence alcohol and drug use, among other things. I am interested in this article’s findings, but also whether this study or others look into the influence of community (a larger, less distrinct, more feeling-dependent identification) and whether community effects those factors as well. (MC '09)
