Merits and caveats of propensity scores to adjust for confounding

2018 ◽  
Vol 34 (10) ◽  
pp. 1629-1635 ◽  
Author(s):  
Edouard L Fu ◽  
Rolf H H Groenwold ◽  
Carmine Zoccali ◽  
Kitty J Jager ◽  
Merel van Diepen ◽  
...  
Keyword(s):  
Risk Factors ◽  
The Netherlands ◽  
Health Outcomes ◽  
Observational Data ◽  
Propensity Scores ◽  
Cooperative Study ◽  
Dialysis Patients ◽  
Proper Understanding ◽  
The Past ◽  
Using Data

Abstract Proper adjustment for confounding is essential when estimating the effects of treatments or risk factors on health outcomes in observational data. To this end, various statistical methods have been developed. In the past couple of years, the use of propensity scores (PSs) to control for confounding has increased. Proper understanding of this method is necessary to critically appraise research in which it is applied. In this article, we provide an overview of PS methods, explaining their concept, advantages and possible disadvantages. Furthermore, the use of PS matching, PS adjustment and PS weighting is illustrated using data from the Netherlands Cooperative Study on the Adequacy of Dialysis (NECOSAD) cohort of dialysis patients.

Survey Research in Disaster Public Health

2000 ◽  
Vol 15 (1) ◽  
pp. 65-71 ◽  
Author(s):  
Kimberley I. Shoaf ◽  
Corrinne Peek-Asa
Keyword(s):  
Public Health ◽  
Risk Factors ◽  
Health Outcomes ◽  
Data Sources ◽  
Health Impacts ◽  
Northridge Earthquake ◽  
Injury Rates ◽  
Telephone Interviews ◽  
The Past ◽  
Human Losses

AbstractIntroduction:While much has been learned during the past three decades of research in the disaster field, there still are some major gaps in knowledge. The need for more and better research on the health aspects of disasters is especially noted. Often, research into the health aspects has been anecdotal in nature and suffers from poor documentation of human losses. However, there are valid research methodologies that can be adapted to better document losses, evaluate interventions, and set priorities for investments to reduce the burden on the health of the population caused by disasters.Methods:A number of data sources are used to demonstrate the potential uses of surveys in disaster health. The majority of the examples reflect data collected by telephone interviews following earthquakes in California.Results:By using comparable instruments, it is possible to track the changes in preparedness levels across time. Similarly, it is possible to compare injury rates or other health impacts across time, place, and disaster type. In addition, risk factors can be identified for health outcomes. For example, in the Northridge earthquake, those over age 60 years were three times more likely to be hospitalized or die as a result of injuries than were those aged 20–59 years. Interventions can be evaluated. Slightlyless than half of the respondents of the El Niño study had heard messages about preparing for the on-coming weather and their preparedness levels were not significantly different from those who had not heard about preparing for the weather.Conclusion:Surveys are useful tools for identifying and evaluating the health impacts of disasters.

Risk factors for sporadic Shiga toxin-producingEscherichia coliO157 and non-O157 illness in The Netherlands, 2008–2012, using periodically surveyed controls

2014 ◽  
Vol 143 (7) ◽  
pp. 1360-1367 ◽  
Author(s):  
I. H. M. FRIESEMA ◽  
M. SCHOTSBORG ◽  
M. E. O. C. HECK ◽  
W. VAN PELT
Keyword(s):  
Risk Factors ◽  
Risk Factor ◽  
The Netherlands ◽  
Shiga Toxin ◽  
Age Groups ◽  
Meat Products ◽  
Farm Animals ◽  
Severe Illness ◽  
Main Risk Factor ◽  
Using Data

SUMMARYShiga toxin-producingEscherichia coli(STEC) infections have been associated with severe illness. Ruminants are seen as the main reservoir and the major transmission route is considered to be foodborne. In The Netherlands, a case-control study was conducted, using data collected during 2008–2012. Patients were interviewed and controls completed a self-administered questionnaire. Patients travelling abroad were excluded from the analyses. STEC O157 and non-O157 were examined separately and differentiated into two age groups (<10 years, ⩾10 years). We included 130 O157 cases, 78 non-O157 cases and 1563 controls. In both age groups of O157 patients, raw spreadable sausage was the main risk factor for infection. For STEC non-O157 cases aged <10 years, contact with farm animals was the main risk factor and in non-O157 cases aged ⩾10 years, consumption of beef was the main risk factor. During 2008–2012, risk factors for STEC infections in the Dutch population differed between age groups and serogroup categories, and were related to eating meat and contact with farm animals. Advising the public about the risks of consuming raw or undercooked meat (products) and hygiene habits in case of contact with farm animals, could help in the prevention of STEC infections.

Predictors of poor outcome in chronic dialysis patients: The Netherlands cooperative study on the adequacy of dialysis

2000 ◽  
Vol 35 (1) ◽  
pp. 69-79 ◽  
Author(s):  
Maruschka P. Merkus ◽  
Kitty J. Jager ◽  
Friedo W. Dekker ◽  
Rob J. de Haan ◽  
Els W. Boeschoten ◽  
...  
Keyword(s):  
The Netherlands ◽  
Chronic Dialysis ◽  
Cooperative Study ◽  
Dialysis Patients ◽  
Poor Outcome

scholarly journals Core human values and their interactions with pro-Tobacco factors on cigarette smoking

2008 ◽  
Vol 6 (1) ◽  
pp. 23-39 ◽  
Author(s):  
Xinguang Chen ◽  
Xiaolan Tang ◽  
Xiaoming Li ◽  
Bonita Stanton ◽  
Hanwu Li
Keyword(s):  
Risk Factors ◽  
Multivariate Analysis ◽  
Cigarette Smoking ◽  
Sensation Seeking ◽  
Human Values ◽  
Smoking Rate ◽  
The Past ◽  
Terminal Values ◽  
Instrumental Values ◽  
Using Data

More effective tobacco control requires new data on factors that are not explicitly related to smoking but are influential, such as “Terminal Values” regarding desirable end-states of existence and “Instrumental Values” regarding desirable modes of conduct. Association analysis was conducted among 36 Core Values (18 Terminal and 18 Instrumental) derived from Rokeach’s Value Survey, three risk factors (protobacco media, smoking peers and sensation-seeking), and cigarette smoking using data collected from a sample of 334 medical students in China. The participants were 18 to 24 years old (47% female) and 18.4% of them smoked in the past 30 days. Multivariate analysis indicated that cigarette smoking was negatively associated with nine Terminal Values (e.g., a Sense of Accomplishment and Self-Respect) and ten Instrumental Values (e.g., Clean and Self-Controlled). As expected, when the endorsed number of values/total value scores increased from low to high, the 30-day smoking rate declined from 32.6% - 75.0% to 13.5% - 15.9% (p < .01). The odds ratios (OR) for the endorsed Terminal Values and the total value scores were 0.50 (p < .01) and 0.64 (p < .01) respectively, and the ORs for the endorsed Instrumental Values and the total value scores were 0.42 (p < .01) and 0.44 (p

scholarly journals Health Recommendations and Selection in Health Behaviors

2020 ◽  
Vol 2 (2) ◽  
pp. 143-160 ◽  
Author(s):  
Emily Oster
Keyword(s):  
Health Behavior ◽  
Health Behaviors ◽  
Health Outcomes ◽  
Observational Data ◽  
Research Finding ◽  
Good Health ◽  
Vitamin Supplementation ◽  
Positive Health ◽  
Using Data ◽  
New Research

Consider a case in which a new research finding links a health behavior with good health outcomes. A possible consequence is take-up of this behavior among individuals who engage in other positive health behaviors. If this occurs, later analyses of observational data may be biased by the change in selection. This paper evaluates these dynamic biases in empirical settings. Using data from vitamin supplementation and diet, I show that selection responds endogenously to health recommendations. These results highlight how spurious findings on health behaviors can be self-reinforcing. (JEL I12)

Time Trends in Social Class Mortality Differentials in the Netherlands, 1820–1920

2009 ◽  
Vol 33 (2) ◽  
pp. 119-153
Author(s):  
Frans van Poppel ◽  
Roel Jennissen ◽  
Kees Mandemakers
Keyword(s):  
Social Class ◽  
The Netherlands ◽  
Adult Mortality ◽  
Living Standards ◽  
Theoretical Issue ◽  
Historical Period ◽  
The Past ◽  
Long Term Trends ◽  
Using Data

The question whether socioeconomic status gradients in adult mortality have changed over a broad historical period has become an important political and theoretical issue but is hard to test. In this article we study long-term trends in social inequality in adult mortality by using data for 2 (of the 11) provinces of the Netherlands for the period 1812–1922. We apply indirect estimation techniques, which have been developed for the analysis of mortality patterns in countries with deficient data. Our article shows that indeed there was a clear social class gradient in mortality, with the elite having higher survival chances between ages 35 and 55 than the middle class and farmers. Differences were even more apparent in comparison with workers. Over time there was a strong convergence among social classes in mortality levels. The implications of our results for the dominant views on the change in living standards in the past are discussed.

A Historical Overview of Health Disparities and the Potential of eHealth Solutions

2008 ◽  
pp. 678-685
Author(s):  
Michael Christopher Gibbons
Keyword(s):  
Breast Cancer ◽  
Risk Factors ◽  
Health Disparities ◽  
Health Status ◽  
Health Outcomes ◽  
Ethnic Groups ◽  
High Frequency ◽  
Scientific Evidence ◽  
Historical Overview ◽  
The Past

Over the past decade, a rapidly expanding body of scientific evidence has been put forth documenting differences in health status among U.S. racial and ethnic groups. Evidence has also mounted suggesting that these differences may be related to both medical and nonmedical determinants. Internationally, however, neither the evidence nor the realization of a link between nonmedical sociobehavioral factors and health outcomes is new. The earliest reported observation of a hypothesized association between socioenvironmental risk factors and health outcomes occurred in Italy over three centuries ago when Bernardino Ramazzini detailed an unusually high frequency of breast cancer in Catholic nuns (Wilson, Jones, Coussens, & Hanna, 2002). Not long thereafter, in 1775, British surgeon Sir Percival Pott reported a cluster of scrotal cancer cases among British chimney sweeps (Wilson et al., 2002).

Sign in / Sign up

Export Citation Format

Share Document