Journal of the National Cancer Institute Advance Access originally published online on March 25, 2008
JNCI Journal of the National Cancer Institute 2008 100(7):483-491; doi:10.1093/jnci/djn066
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
© The Author 2008. Published by Oxford University Press.
ARTICLES |
Design and Analysis of Group-Randomized Trials in Cancer: A Review of Current Practices
Affiliations of authors: Divisions of Epidemiology (DMM) and Health Behavior and Health Promotion (CMA), College of Public Health, Department of Family Medicine, College of Medicine (JL), and Comprehensive Cancer Center (CMA), The Ohio State University, Columbus, OH; Quantitative Sciences and Data Management Branch, Division of HIV/AIDS Prevention, National Center for HIV, Viral Hepatitis, STD, and TB Prevention, US Centers for Disease Control and Prevention, Atlanta, GA (SLP); Division of Public Health and Environment, RTI International, Research Triangle Park, NC (JLB)
Correspondence to: David M. Murray, PhD, Division of Epidemiology, College of Public Health, The Ohio State University, Columbus, OH (e-mail: dmurray{at}cph.osu.edu).
 |
ABSTRACT |
|---|
 Top Abstract Context and CaveatsMethodsResultsDiscussionFundingReferencesNotes |
|---|
Background: Previous reviews have identified problems in the design and analysis of group-randomized trials in a number of areas. Similar problems may exist in cancer research, but there have been no comprehensive reviews.
Methods: We searched Medline and PubMed for group-randomized trials focused on cancer prevention and control that were published between 2002 and 2006. We located and reviewed 75 articles to determine whether articles included evidence of taking group randomization into account in establishing the size of the trial, such as reporting the expected intraclass correlation, the group component of variance, or the variance inflation factor. We also examined the analytical approaches to determine their appropriateness.
Results: Only 18 (24%) of the 75 articles documented appropriate methods for sample size calculations. Only 34 (45%) limited their reports to analyses judged to be appropriate. Fully 26 (34%) failed to report any analyses that were judged to be appropriate. The most commonly used inappropriate analysis was an analysis at the individual level that ignored the groups altogether. Nine articles (12%) did not provide sufficient information.
Conclusions: Many investigators who use group-randomized trials in cancer research do not adequately attend to the special design and analytic challenges associated with these trials. Failure to do so can lead to reporting type I errors as real effects, mislead investigators and policy-makers, and slow progress toward control and prevention of cancer. A collaborative effort by investigators, statisticians, and others will be required to ensure that group-randomized trials are planned and analyzed using appropriate methods so that the scientific community can have confidence in the published results.
Prior knowledge Group-randomized trials pose distinct analytic challenges, and the validity of the statistical methods used by investigators in the field of cancer research to analyze these trials was unknown. Study design Group-randomized trials on cancer prevention and control were identified by searching the peer-reviewed literature with sets of key words. Each article was coded according to a set of criteria pertaining to the analytic procedures used and their reporting. Contribution Most reports of group-randomized trials in cancer research analyzed in this study used analytical methods that were inappropriate to the assignment by group or failed to provide sufficient information for their assessment. Most articles did not document appropriate methods for sample size calculations, and many ignored group randomization altogether. Implications Failure to attend to the design and analytic challenges posed by interventions that operate at a group level is likely to cause type I errors to be identified as real effects and mislead investigators. A collaborative effort by investigators, statisticians, and others will be needed to ensure that group-randomized trials are planned and analyzed appropriately. Limitations The study did not critique the trials in terms of more general design and analytic issues that were not specific to group-randomized trials.
|
Group-randomized trials, sometimes called cluster-randomized trials, are comparative studies in which investigators randomly assign identifiable groups to conditions and observe individual members of those groups to assess the effects of an intervention (1,2). In this context, an identifiable group refers to any group that is not constituted at random so that there is some physical, social, or other connection among its members. As an example, consider a trial in which worksites are randomly assigned to an intervention to encourage workers to quit smoking and increase their physical activity or to control conditions and workers within those worksites are observed over time to assess the effects of the intervention. The worksites are the groups randomized to the study conditions, and the workers are the members observed to assess the effects of the intervention.
Just as the randomized clinical trial is the gold standard in public health and medicine when randomization of individuals to study conditions is possible, the group-randomized trial is the gold standard in public health and medicine when randomization of individuals is not possible. Group randomization will be required whenever investigators seek to evaluate interventions that operate at a group level, manipulate the social or physical environment, or cannot be delivered to individuals. In this context, it is not surprising that group-randomized trials have become increasingly common in cancer research because so many trials in cancer involve interventions that meet at least one of these conditions.
Group-randomized trials pose design and analytic challenges distinct from those posed by randomized clinical trials. The connections among the group members create the likelihood of positive correlation among observations taken from members of the same group (3). Such intraclass correlation invalidates the usual independence assumption and threatens the validity of the analytic methods typically used in randomized clinical trials; application of those methods in a group-randomized trial will yield a type I error rate that is inflated, often badly (4–7). Furthermore, when the number of groups in each condition is limited, df and power available for a valid test of the intervention will be limited (1,2,4,8–12). In addition to these analytic challenges, simple random assignment of a limited number of groups to each condition may not evenly distribute all potential confounders, thereby jeopardizing the internal validity of the trial (1,2).
Consideration must be given to these challenges as the group-randomized trial is planned and analyzed to support valid inference. Clear reporting is also important, with attention to the extension of the CONSORT statement for group-randomized trials (13). Previous reviews have identified problems in group-randomized trials in areas such as prevention of adolescent drug use and cardiovascular risk reduction (14–16), and similar problems may exist in cancer research (17). Here we provide what is to our knowledge the first comprehensive review of the design and analytic methods used in group-randomized trials in cancer research to assess the state of the practice in this area. Our goals are to summarize current practices with regard to design and analytic methods used in these trials, to identify the strengths and weaknesses in those methods, and to offer recommendations that will improve the design and analysis of future group-randomized trials in cancer research.
|
Methods |
|---|
|
TopAbstractContext and CaveatsMethodsResultsDiscussionFundingReferencesNotes |
|---|
We conducted a review of group-randomized trials that were focused on cancer prevention and control, including studies having as the primary outcome cancer risk factors, cancer morbidity, or cancer mortality, that were published in the peer-reviewed literature during the period 2002–2006, inclusive. The methods used in this review were based on methods used in a recent review of group-randomized trials in the general public health literature (14). In brief, we searched for articles in Medline and PubMed using six sets of key words: "cancer and worksite," "cancer and workplace," "cancer and school," "cancer and community," "cancer and church," and "cancer and practice." An additional term "randomized" was included in the Medline search using the parameters "cancer and practice" due to the large number of citations initially identified. The search yielded 92 possible group-randomized trials in cancer research from 45 journals.
Group-randomized trials were defined as studies that randomly assigned identifiable groups to conditions and obtained observations from individual members of those groups. Articles reporting the results of studies in which groups were not randomly assigned to study conditions were excluded, as were studies involving observations at only the group level, rather than the individual level. Articles were also excluded if they lacked a clear statement that all groups were randomly assigned to conditions.
After these exclusions, we reviewed 75 primary articles (18–92) in 41 journals. We reviewed 20 additional articles (93–112) cited as background articles in one of the primary articles because the design of a group-randomized trial is often described in such an article.
Each article was reviewed by three authors (DM, SP, and CA; or DM, JB, and JL) to characterize the design and analytic features of the study. Assigned authors coded each article independently using a process adapted from Varnell et al. (14). In particular, we sought to determine whether the article included sample size calculations and analyses that reflected the data structure of the trial. We reviewed articles to determine whether articles included evidence of taking group randomization into account a priori in establishing the size of the trial, such as the expected intraclass correlation, group component of variance, or variance inflation factor (8). If no such evidence was found, articles were reviewed to determine whether articles claimed that variance was inflated to account for the expected intraclass correlation.
Table 1 [adapted from Varnell et al. (14)] presents the criteria used to judge whether the different analytic approaches were appropriate. Methods that were considered to be appropriate included but were not limited to mixed-model regression approaches, including analysis of variance (ANOVA) or analysis of covariance (ANCOVA) (1,113,114) and random coefficient models (1,114,115); two-stage analyses (analysis on a summary statistic computed at the level of the group including randomization-based tests) (116,117); and generalized estimating equations (118,119). Because each of these methods can be applied incorrectly, we established additional criteria for rating analyses as appropriately applied; these depend on the design of the study, the assumptions underlying the analytic method, and the robustness of the method to violations of these assumptions (Table 1). Mixed-model ANOVA or ANCOVA was considered to be appropriate if only one or two time points were included in the analysis, if variation at the condition level was assessed against variation at the group level, and if df were based on the number of groups. If more than two time points are included in the analysis of a group-randomized trial, a random coefficient analysis preserves the nominal type I error rate, whereas mixed-model ANOVA or ANCOVA does not (115); thus, random coefficient analyses were considered to be appropriate if more than two time points were included in the analysis but mixed-model ANOVA or ANCOVA was not. Two-stage approaches were considered to be appropriate if the second stage was conducted at the group level and df were based on the number of groups. Analysis based on generalized estimating equations was considered to be appropriate if there were enough groups to provide 40 or more df for the test of the intervention effect or if special steps were taken to correct the downward bias in the empirical sandwich estimator when there were fewer than 40 df (120). Several articles reporting less common methods referenced articles outlining these methods; we reviewed those articles for evidence that the analytic method described was suitable.
|
Because many articles presented more than one analytic strategy to evaluate intervention effects, we reviewed each article to determine whether all of the analytic approaches used to evaluate intervention effects were appropriate, whether some were appropriate, or whether none were appropriate. Also, each of the analytic approaches was recorded, along with any justifications offered in articles that reported inappropriate analytic strategies. Disagreements for any of the coding decisions were resolved through discussion among the authors assigned to review that article.
|
Results |
|---|
|
TopAbstractContext and CaveatsMethodsResultsDiscussionFundingReferencesNotes |
|---|
The Studies
Table 2 lists the peer-reviewed journals in which the reports of group-randomized trial in cancer research that we located were published during the period 2002–2006, inclusive. The focus of the journals ranged widely across medicine, public health, addiction research, behavioral medicine, and health promotion. Fifteen (20%) of the 75 articles were published in Preventive Medicine, five (6.7%) were published in the American Journal of Public Health, and the remaining 55 articles were distributed among 39 other journals, with no more than three (4.0%) in any single journal. Thirty-five (47%) articles were based on trials supported by the National Cancer Institute (NCI); six others (8.0%) were based on trials supported by other institutes within the National Institutes of Health (NIH).
|
Design Characteristics
Table 3 presents the design characteristics from the 75 articles. Most (88%) employed a design with just two conditions (eg, intervention vs control). A tendency to assign groups to just two conditions has been observed in other content areas (14) and reflects the complexity and cost often involved even in simple group-randomized trials. Most (60%) used a priori matching and/or stratification in their design, consistent with current recommendations (1,2,120). Most (85%) were conducted in medical clinics, schools or colleges, worksites, or communities or neighborhoods.
|
Half of the trials (51%) included nine or more groups per condition, which is often the minimum number needed to have 80% power given modest intervention effects (eg, 0.25 SD units) and the magnitude of intraclass correlation typically observed in public health and medicine. Four trials (5.3%) included only a single group per condition. Such a design does not permit a valid analysis without strong and untestable assumptions and is not recommended where the goal is to make inferences regarding intervention effects (121).
The number of members per group ranged from an average of just less than two to nearly six thousand. A modest majority (55%) used a design with just one time point, and 84% used a design with one or two time points. Most (84%) focused on primary or secondary prevention. Of the former, 91% targeted individuals with an unknown or mixed personal cancer history, and of the latter, 45% targeted individuals with no personal history of the target cancer. The range of primary outcome variables was broad, although the most common were screening, tobacco use, dietary variables, and knowledge or attitudes.
Sample Size Methods
In 35 (47%) of the articles, there was no mention of sample size calculations. Nine (12%) articles reported a power analysis without providing details as to the methods used. Six (8.0%) based their sample size on individuals, ignoring groups altogether. Six (8.0%) reported that variance had been inflated to account for the expected intraclass correlation but provided no further detail. Only 18 (24%) reported appropriate methods for sample size calculations and included an intraclass correlation, group component of variance, or variance inflation factor.
Analytic Methods
Among the 75 articles, 34 (45%) reported only analyses that were judged to be appropriate given the design of the study (Table 4); 15 of these were supported by grants from the NCI and six by grants from other institutes within the NIH. The analytic methods used represented all of the methods described in Table 1 as appropriate under the proper conditions. Mixed model regression methods were used most often, though analyses based on generalized estimating equations or two-stage methods were used in 12 studies.
|
Six (8.0%) articles reported some analyses that were judged to be appropriate given the design of the study and some that were not; half were supported by grants from the NCI. In this group, the most commonly used inappropriate analysis was an analysis at an individual level that ignored the groups altogether.
Twenty-six (35%) articles reported only analyses that were judged to be inappropriate given the design of the study; half of these were funded by grants from the NCI. In this group as well, the most commonly used inappropriate analysis was an analysis at the individual level that ignored the nested groups altogether. There were also a number of studies that conducted analyses based on generalized estimating equations with fewer than 40 df when the beneficial asymptotic properties were unlikely to hold.
Nine articles (12%) did not provide sufficient information to judge whether their analytic methods were appropriate; three of these were supported by grants from the NCI. Often these articles made reference to an appropriate method but did not provide enough detail to determine whether the method had been used appropriately. We classified these articles as indeterminate.
Other Reporting Issues
Only six of the 75 studies (8.0%) reported an intraclass correlation coefficient for their primary outcome variable, despite repeated calls for routine publication of this information (eg, 1, 2, 120, 122). Two (2.7%) failed to report the number of groups per condition or the number of members per group. Failure to report intraclass correlations makes it more difficult to plan future studies, and failure to provide all the details of the design makes it difficult to fully evaluate the current study. The recent CONSORT statement (13) may help in this regard, although our review included articles published up until the end of 2006 and thus 2 publication years after that statement appeared.
|
Discussion |
|---|
|
TopAbstractContext and CaveatsMethodsResultsDiscussionFundingReferencesNotes |
|---|
Our review identified 75 articles published in 41 journals that reported intervention results based on a group-randomized trial related to cancer or cancer risk factors between 2002 and 2006. The number of articles and their distribution across so many journals underscore how widespread the use of interventions requiring group-randomized trials for their evaluation has become in cancer research.
Unfortunately, our findings indicate that many investigators conducting these trials do not adequately attend to the special design and analytic challenges they pose. For example, only 24% of the articles reported enough information on sample size calculation to assure the reader that the calculations had been done properly. That does not mean that the remaining studies were planned incorrectly; the investigators might have used appropriate methods but failed to report them in their articles. Even so, one-third of the studies included five or fewer groups per condition, suggesting that appropriate sample size methods were not used in a substantial proportion of the studies included in this review.
Further evidence of the lack of attention to the requirements in the design and analysis of these trials is that 26 (34%) of the 75 articles reported no analytic methods judged appropriate to assess intervention effects and 21 (81%) of these reported methods that either ignored the group completely or performed subgroup analyses, methods that have been widely discredited for more than a decade (eg, 7). Given that 88% of these articles also reported statistically significant intervention effects, there is reason to be concerned that many of these reported effects may be type I errors. Continued reporting of spurious effects misleads investigators and policymakers and slows progress toward control and prevention of cancer.
In the field of psychotherapy research, Baldwin et al. (123) examined 33 articles that provided the primary evidence that had placed an intervention on the American Psychological Association’s list of Empirically Supported Treatments and found that none had properly accounted for the group-level intraclass correlation or limited df inherent in the design used in each of the 33 articles. Using published intraclass correlation estimates from similar studies, they made post hoc corrections to estimate the effect size and statistical significance that might have been obtained using an appropriate analysis. Depending on the intraclass correlation estimate used, 18%–60% of the studies no longer had any statistically significant intervention effects after the correction. If a similar pattern held for group-randomized trials in cancer research, the ramifications would be substantial. Making matters worse, Baldwin et al. (123) dealt with a study design [individuals randomized to receive treatment in groups design, cf. (124)] in which the adverse impact of the intraclass correlation is less than it would be in a group-randomized trial so that the situation could be more severe in cancer research.
Further evidence of the poor attention to the design and analytic issues is found in the explanations offered by investigators who used inappropriate methods. Many ignored the group altogether and argued that their study was not large enough to support an analysis at the group level or that their intervention was designed to change outcomes in individuals. These arguments ignore the design of the study and leave the investigator at great risk of reporting type I errors as intervention effects. Some examined the intraclass correlations in their data and judged them small enough to ignore. The problem with this approach is that any positive intraclass correlation will inflate the type I error rate, with the level of inflation related both to the magnitude of the intraclass correlation and the average number of members per group included in the analysis. The prudent course is to retain all random effects associated with the design and sampling plan (1,2); after all, if the intraclass correlation really is close to zero, the extra variation will be limited, and if it is not, the test will be adjusted appropriately. Others have recognized that the number of groups was limited and analyzed their data at the subgroup level instead, presumably to increase the df for the test of the intervention effect. An example would be a trial in which only a few schools were randomly assigned to each condition and the analysis was performed at the level of the classroom. The problem with this approach is that it rests on the strong assumption that the intraclass correlation at the subgroup level captures all of the intraclass correlation at the group level; unfortunately, this assumption is usually not testable, and simulation studies have shown that if it is violated, there is a risk of an inflated type I error rate (121). Others used methods known to be valid asymptotically, for example, methods based on generalized estimating equations, but did so when the df for the test of the intervention effect were quite limited so that the beneficial asymptotic properties were unlikely to apply. A number of studies have shown that the type I error rate is inflated under these conditions (115,125–127).
Given the extensive literature on the analysis of group-randomized trials that has developed over the past 30 years, and particularly over the past 10 years, the low usage of appropriate statistical methods and the high usage of discredited ones may seem surprising. On the other hand, earlier reviews have reported similar results in the general public health literature (14–16), so the situation may be no worse in cancer research than elsewhere. Even so, additional education is needed for investigators, grant reviewers, and journal reviewers and editors to take full advantage of the information that is readily available regarding the proper design and analysis of these trials.
Investigators need not become methodologists or statisticians to improve the methods used in their studies. They can benefit substantially by collaborating with methodologists or statisticians who know these issues well, in the same way they benefit by collaborating with interventionists who understand their part of the health promotion and disease prevention research process. Ten years ago, it was difficult to find methodologists or statisticians who were familiar with the design and analytic issues involved in group-randomized trials, but that is no longer the case. Instead, expertise in this area is available at many of the major research universities and institutes. It is increasingly common, for example, to find statisticians trained in mixed-model regression methods and in methods based on generalized estimating equations; such statisticians may not have worked on group-randomized trials before, but they can easily adapt their training. Increasingly, statisticians and methodologists are training in the methods for group-randomized trials directly, as reflected by the growing literature in this area. There has been a large increase in the interest in these trials in educational research in the past few years, promoted in part by the new Institute of Education Sciences of the US Department of Education; as a result, an increasing number of statisticians and methodologists working in educational research are becoming familiar with their design and analysis (eg, 128). At this point, investigators looking for experts to collaborate on their projects should be able to do so, but they need to make the effort.
Reviewers for agencies funding group-randomized trials and for journals publishing these studies should ensure that studies are properly planned and that their reports provide evidence of that planning, both in terms of sample size estimation and in terms of data analysis. Editors and scientific review administrators also need to set an appropriate standard, and they can take an important step by assigning articles or grant applications for these trials to a statistician or other methodologist familiar with the special design and analytic issues facing these studies.
Although we tried to identify all group-randomized trials in cancer research published between 2002 and 2006, we may have missed some articles; as such, our review may be incomplete. In addition, we limited our review to the design and analytic features that were specific to these trials; thus, we did not attempt to critique the articles for more general design and analytic issues, and our review may be incomplete in that regard as well. Finally, we did not critique the studies in terms of their intervention programs, and the results of any trial will depend on the quality of the intervention as well as on the quality of the design and analysis.
Group-randomized trials remain the gold standard for studies designed to evaluate an intervention that operates at a group level, manipulates the social or physical environment, or cannot be delivered to individuals. The issue is not whether to use these trials, or even how to use them, but rather to ensure that they are planned and analyzed using appropriate methods so that we can have confidence in their published results. The results of this review indicate that we need to do a better job in the area of cancer research.
|
Funding |
|---|
|
TopAbstractContext and CaveatsMethodsResultsDiscussionFundingReferencesNotes |
|---|
National Cancer Institute (R01CA116487 to D.M.M.); American Cancer Society (SIRSSG05253 to D.M.M.).
|
NOTES |
|---|
|
TopAbstractContext and CaveatsMethodsResultsDiscussionFundingReferencesNotes |
|---|
The findings and conclusions in this paper are those of the authors and do not necessarily represent the views of the Centers for Disease Control and Prevention.
The funding agencies played no role in the design of the study; the collection, analysis, or interpretation of the data; the decision to submit the manuscript for publication; or the writing of the manuscript.
|
REFERENCES |
|---|
|
TopAbstractContext and CaveatsMethodsResultsDiscussionFundingReferencesNotes |
|---|
1. Murray DM. Design and Analysis of Group-Randomized Trials (1998) New York: Oxford University Press.
2. Donner A, Klar N. Design and Analysis of Cluster Randomization Trials in Health Research (2000) London, UK: Arnold.
3. Kish L. Survey Sampling (1965) New York: John Wiley & Sons.
4. Cornfield J. Randomization by group: a formal analysis. Am J Epidemiol (1978) 108(2):100–102.
5. Zucker DM. An analysis of variance pitfall: the fixed effects analysis in a nested design. Educ Psychol Meas (1990) 50(4):731–738.[Abstract]
6. Murray DM, Wolfinger RD. Analysis issues in the evaluation of community trials: progress toward solutions in SAS/STAT MIXED. J Community Psychol (1994) 140–154. (CSAP special issue).
7. Murray DM, Hannan PJ, Baker WL. A Monte Carlo study of alternative responses to intraclass correlation in community trials: is it ever possible to avoid Cornfield's penalties? Eval Rev. (1996) 20(3):313–337.
8. Donner A, Birkett N, Buck C. Randomization by cluster: sample size requirements and analysis. Am J Epidemiol (1981) 114(6):906–914.
9. Feldman HA, McKinlay SM. Cohort versus cross-sectional design in large field trials: precision, sample size, and a unifying model. Stat Med (1994) 13(1):61–78.[Web of Science][Medline]
10. Hsieh FY. Sample size formulae for intervention studies with the cluster as unit of randomization. Stat Med (1988) 6(11):1195–1201.
11. Koepsell TD, Martin DC, Diehr PH, et al. Data analysis and sample size issues in evaluations of community-based health promotion and disease prevention programs: a mixed-model analysis of variance approach. J Clin Epidemiol (1991) 44(7):701–713.[CrossRef][Web of Science][Medline]
12. Murray DM, Hannan PJ. Planning for the appropriate analysis in school-based drug-use prevention studies. J Consult Clin Psychol (1990) 58(4):458–468.[CrossRef][Web of Science][Medline]
13. Campbell MK, Elbourne DR, Altman DG. CONSORT statement: extension to cluster randomised trials. Br Med J (2004) 328(7441):702–708.
14. Varnell SP, Murray DM, Janega JB, Blitstein JL. Design and analysis of group-randomized trials: a review of recent practices. Am J Public Health (2004) 94(3):393–399.
15. Donner A, Brown KS, Brasher P. A methodologic review of non-therapeutic intervention trials employing cluster randomization, 1979–1989. Int J Epidemiol (1990) 19(4):795–800.
16. Simpson JM, Klar N, Donner A. Accounting for cluster randomization: a review of primary prevention trials, 1990 through 1993. Am J Public Health (1995) 85(10):1378–1383.
17. Murray DM, Varnell SP, Blitstein JL. Design and analysis of group-randomized trials in cancer prevention and control. In: Handbook of Behavioral Science and Cancer—Miller S, Bowen D, Croyle R, Rowland J, eds. Washington, DC: Behavioral Science and Cancer. In press.
18. Aitken JF, Youl PH, Janda M, Lowe JB, Ring IT, Elwood M. Increase in skin cancer screening during a community-based randomized intervention trial. Int J Cancer (2006) 118(4):1010–1016.[CrossRef][Web of Science][Medline]
19. Amaro S, Viggiano A, Di Costanzo A, et al. Kaledo, a new educational board-game, gives nutritional rudiments and encourages healthy eating in children: a pilot cluster randomized trial. Eur J Pediatr (2006) 165(9):630–635.[CrossRef][Web of Science][Medline]
20. Ausems M, Mesters I, van Bruekelen G, De Vries H. Short-term effects of a randomized computer-based out-of-school smoking prevention trial aimed at elementary schoolchildren. Prev Med (2002) 34(6):581–589.[CrossRef][Web of Science][Medline]
21. Baranowski T, Baranowski J, Cullen KW, et al. 5 a day achievement badge for African-American boy scouts: pilot outcome results. Prev Med (2002) 34(3):353–363.[CrossRef][Web of Science][Medline]
22. Bonhauser M, Fernandez G, Puschel K, et al. Improving physical fitness and emotional well-being in adolescents of low socioeconomic status in Chile: results of a school-based controlled trial. Health Promot Int (2005) 20(2):113–122.
23. Boyer CB, Shafer M-A, Shaffer RA, et al. Evaluation of a cognitive-behavioral, group, randomized controlled intervention trial to prevent sexually transmitted infections and unintended pregnancies in young women. Prev Med (2005) 40(4):420–431.[CrossRef][Web of Science][Medline]
24. Braun KL, Fong M, Kaanoi ME, Kamaka ML, Gotay CC. Testing a culturally appropriate, theory-based intervention to improve colorectal cancer screening among Native Hawaiians. Prev Med (2005) 40(6):619–627.[CrossRef][Web of Science][Medline]
25. Buller DB, Andersen PA, Walkosz BJ, et al. Randomized trial testing a worksite sun protection program in an outdoor recreation industry. Health Educ Behav (2005) 32(4):514–535.
26. Buller DB, Reynolds KD, Yaroch A, et al. Effects of the Sunny Days, Health Ways curriculum on students in grades 6 to 8. Am J Prev Med (2006) 30(1):13–22.[CrossRef][Web of Science][Medline]
27. Campbell MK, James A, Hudson MA, et al. Improving multiple behaviors for colorectal cancer prevention among African American church members. Health Psychol (2004) 23(5):492–502.[CrossRef][Web of Science][Medline]
28. Campbell MK, Tessaro I, DeVellis B, Benedict S, Kelsey K, Belton L. Effects of a tailored health promotion program for female blue-collar workers: health works for women. Prev Med (2002) 34(3):313–323.[CrossRef][Web of Science][Medline]
29. Coleman EA, Lord J, Heard J, et al. The Delta project: increasing breast cancer screening among rural minority and older women by targeting rural healthcare providers. Oncol Nurs Forum (2003) 30(4):669–677.[Medline]
30. Corbie-Smith G, Ammerman AS, Katz ML, et al. Trust, benefit, satisfaction, burden: a randomized controlled trial to reduce cancer risk through African-American churches. J Gen Intern Med (2003) 18(7):531–541.[CrossRef][Web of Science][Medline]
31. Coyne K, Xu R, Raich P, et al. Randomized, controlled trial of an easy-to-read informed consent statement for clinical trial participation: a study of the Eastern Cooperative Oncology Group. J Clin Oncol (2003) 21(5):836–842.
32. Dubey V, Mathew R, Iglar K, Moineddin R, Glazier R. Improving preventive service delivery at adult complete health check-ups: the Preventive health Evidence-based Recommendation Form (PERFORM) cluster randomized controlled trial. BMC Fam Pract (2006) 7:44.[CrossRef][Medline]
33. Elliott TE, Elliott BA, Regal RR, et al. Lake Superior Rural Cancer Care Project, Part III: provider practice. Cancer Pract (2002) 10(2):75–84.[CrossRef][Web of Science][Medline]
34. Emmons KM, Stoddard AM, Fletcher R, et al. Cancer prevention among working class, multiethnic adults: results of the healthy directions-health centers study. Am J Public Health (2005) 95(7):1200–1205.
35. English DR, Burton RC, del Mar CB, Donovan RJ, Ireland PD, Emery G. Evaluation of aid to diagnosis of pigmented skin lesions in general practice: controlled trial randomised by practice. Br Med J (2003) 327(7411):375–380.
36. Ferreira MR, Dolan NC, Fitzgibbon ML, et al. Health care provider-directed intervention to increase colorectal cancer screening among veterans: results of a randomized controlled trial. J Clin Oncol (2005) 23(7):1548–1554.
37. Fitzgibbon ML, Stolley MR, Schiffer L, Van Horn L, KauferChristoffel K, Dyer A. Hip-Hop to Health Jr. for Latino Preschool Children. Obesity (2006) 14(9):1616–1625.[CrossRef][Web of Science][Medline]
38. French SA, Story M, Fulkerson JA, et al. Increasing weight-bearing physical activity and calcium-rich foods to promote bone mass gains among 9–11 year old girls: outcomes of the Cal-Girls study. Int J Behav Nutr Phys Act (2005) 2(1):8.[CrossRef][Medline]
39. Fry A, Cull A, Rush R, et al. A randomised controlled trial of breast cancer genetics services in South East Scotland: psychological impact. Br J Cancer (2003) 89(4):653–659.[CrossRef][Web of Science][Medline]
40. Ganksy SA, Ellison JA, Rudy D, et al. Cluster-randomized controlled trial of an athletic trainer-directed spit (smokeless) tobacco intervention for collegiate baseball athletes: results after 1 year. J Athl Train (2005) 40(2):76–87.[Medline]
41. Ganz PA, Farmer MM, Belman MJ, et al. Results of a randomized controlled trial to increase colorectal cancer screening in a managed care health plan. Cancer (2005) 104(10):2072–2083.[CrossRef][Web of Science][Medline]
42. Glanz K, Elliott T, Steffen A, O'Riordan D. Diffusion of an effective skin cancer prevention program: design, theoretical foundations, and first-year implementation. Health Psychol (2005) 24(5):477–487.[CrossRef][Web of Science][Medline]
43. Glanz K, Geller AC, Shigaki D, Maddock J, Isnec MR. A randomized trial of skin cancer prevention in aquatics settings: the Pool Cool program. Health Psychol (2002) 21(6):579–587.[CrossRef][Web of Science][Medline]
44. Glazebrook C, Garrud P, Avery A, Coupland C, Williams H. Impact of a multimedia intervention "Skinsafe" on patients’ knowledge and protective behaviors. Prev Med (2006) 42(6):449–454.[CrossRef][Web of Science][Medline]
45. Goodwin JS, Satish S, Anderson ET, Nattinger AB, Freeman JL. Effect of nurse case management on the treatment of older women with breast cancer. J Am Geriatr Soc. (2003) 51(9):1252–1259.[CrossRef][Web of Science][Medline]
46. Gorin SS, Ashford AR, Lantigua R, et al. Effectiveness of academic detailing on breast cancer screening among primary care physicians in an underserved community. J Am Board Fam Med (2006) 19(2):110–121.
47. Gritz ER, Tripp MK, James AS, et al. An intervention for parents to promote preschool children's sun protection: effects of Sun Protection is Fun! Prev Med (2005) 41(2):357–366.[CrossRef][Web of Science][Medline]
48. Hamilton G, Cross D, Resnicow K, Hall M. A school-based harm minimization smoking intervention trial: outcome results. Addiction (2005) 100(5):689–700.[CrossRef][Web of Science][Medline]
49. Hennrikus DJ, Jeffery RW, Lando HA, et al. The SUCCESS project: the effect of program format and incentives on participation and cessation in worksite smoking cessation programs. Am J Public Health (2002) 92(2):274–279.
50. Holloway RM, Wilkinson C, Peter TJ, et al. Cluster-randomised trial of risk communication to enhance informed uptake of cervical screening. Br J Gen Pract (2003) 53(493):620–625.[Web of Science][Medline]
51. Holt CL, Klem PR. As you go, spread the word: spiritually based breast cancer education for African American women. Gynecol Oncol (2005) 99(3 suppl 1):S141–S142.[CrossRef][Web of Science][Medline]
52. Katz DA, Muehlenbruch DR, Brown RL, Fiore MC, Baker TB. Effectiveness of implementing the agency for healthcare research and quality smoking cessation clinical practice guideline: a randomized, controlled trial. J Natl Cancer Inst (2004) 96(8):594–603.
53. Kim YH, Sarna L. An intervention to increase mammography use by Korean American women. Oncol Nurs Forum (2004) 31(1):105–110.[CrossRef][Web of Science][Medline]
54. Kralj B, Iverson D, Hotz K, Ashbury FD. The impact of computerized clinical reminders on physician prescribing behavior: evidence from community oncology practice. Am J Med Qual (2003) 18(5):197–203.
55. Lazovich D, Parker DL, Brosseau LM, et al. Effectiveness of a worksite intervention to reduce an occupational exposure: the Minnesota Wood Dust Study. Am J Public Health (2002) 92(9):1498–1505.
56. Luckmann R, Savageau JA, Clemow L, Stoddard AM, Costanza ME. A randomized trial of telephone counseling to promote screening mammography in two HMOs. Cancer Detect Prev (2003) 27(6):442–450.[CrossRef][Web of Science][Medline]
57. Lytle LA, Kubik MY, Perry C, Story M, Birnbaum AS, Murray DM. Influencing healthful food choices in school and home environments: results from the TEENS study. Prev Med (2006) 43(1):8–13.[CrossRef][Web of Science][Medline]
58. McCambridge J, Strang J. The efficacy of single-session motivational interviewing in reducing drug consumption and perceptions of drug-related risk and harm among young people: results from a multi-site cluster randomized trial. Addiction (2004) 99(1):39–52.[CrossRef][Web of Science][Medline]
59. Meredith LS, Yano EM, Hickey SC, Sherman SE. Primary care provider attitudes are associated with smoking cessation counseling and referral. Med Care (2005) 43(9):929–934.[CrossRef][Web of Science][Medline]
60. Michielutte R, Sharp PC, Foley KL, et al. Intervention to increase screening mammography among women 65 and older. Health Educ Res. (2005) 20(2):149–162.
61. Munro JF, Nicholl JP, Brazier JE, Davey R, Cochrane T. Cost effectiveness of a community based exercise programme in over 65 year olds: cluster randomized trial. J Epidemiol Community Health (2004) 58(12):1004–1010.
62. Myers RE, Turner B, Weinberg D, et al. Impact of a physician-oriented intervention on follow-up in colorectal cancer screening. Prev Med (2004) 38(4):375–381.[CrossRef][Web of Science][Medline]
63. Pate RR, Ward DS, Saunders RP, Felton G, Dishman RK, Dowda M. Promotion of physical activity among high-school girls: a randomized controlled trial. Am J Public Health (2005) 95(9):1582–1587.
64. Powe BD, Ntekop E, Barron M. An intervention study to increase colorectal cancer knowledge and screening among community elders. Public Health Nurs (2004) 21(5):435–443.[CrossRef][Web of Science][Medline]
65. Rapp K, Buchele G, Jahnke AG, Weiland SK. A cluster-randomized trial on smoking cessation in German student nurses. Prev Med (2006) 42(6):443–448.[CrossRef][Web of Science][Medline]
66. Resnicow K, Kramish Campbell M, Carr C, et al. Body and soul: a dietary intervention conducted through African-American churches. Am J Prev Med (2004) 27(2):97–105.[CrossRef][Web of Science][Medline]
67. Ringdal GI, Jordhoy MS, Kaasa S. Family satisfaction with end-of-life care for cancer patients in a cluster randomized trial. Int J Cancer (2002) 24(1):53–63.
68. Roetzheim RG, Christman LK, Jacobsen PB, et al. A randomized controlled trial to increase cancer screening among attendees of community health centers. Ann Fam Med (2004) 2(4):294–300.
69. Roetzheim RG, Christman LK, Jacobsen PB, Schroeder J, Abdulla R, Hunter S. Long-term results from a randomized controlled trial to increase cancer screening among attendees of community health centers. Ann Fam Med (2005) 3(2):109–114.
70. Ruffin MT, Gorenflo DW. Interventions fail to increase cancer screening rates in community-based primary care practices. Prev Med (2004) 39(3):435–440.[CrossRef][Web of Science][Medline]
71. Saitz R, Horton NJ, Sullivan LM, Moskowitz MA, Samet JH. Addressing alcohol problems in primary care: a cluster randomized, controlled trial of a systems intervention. The screening and intervention in primary care (SIP) study. Ann Intern Med (2003) 138(5):372–382.
72. Sankaranarayanan R, Nene BM, Dinshaw KA, et al. A cluster randomized trial of visual, cytology and human papillomavirus screening for cancer of the cervix in rural India. Int J Cancer (2005) 116(4):617–623.[CrossRef][Web of Science][Medline]
73. Sankaranarayanan R, Rajkumar R, Theresa R, et al. Initial results from a randomized trial of cervical visual screening in rural south India. Int J Cancer (2004) 109(3):461–467.[CrossRef][Web of Science][Medline]
74. Schroy PC, Glick JT, Geller AC, Jackson A, Heeren T, Prout M. A novel educational strategy to enhance internal medicine residents' familial colorectal cancer knowledge and risk assessment skills. Am J Gastroenterol (2005) 100(3):677–684.[CrossRef][Web of Science][Medline]
75. Schulze A, Mons U, Edler L, Potschke-Langer M. Lack of sustainable prevention effect of the "Smoke-Free Class Competition" on German pupils. Prev Med (2006) 42(1):33–39.[CrossRef][Web of Science][Medline]
76. Sejr HS, Osler M. Do smoking and health education influence student nurses' knowledge, attitudes, and professional behavior? Prev Med (2002) 34(2):260–265.[CrossRef][Web of Science][Medline]
77. Siminoff LA, Gordon NH, Silverman P, Budd T, Ravdin PM. A decision aid to assist in adjuvant therapy choices for breast cancer. Psychooncology (2006) 15(11):1001–1013.[CrossRef][Medline]
78. Sorensen G, Barbeau E, Stoddard AM, Hunt MK, Kaphingst K, Wallace L. Promoting behavior change among working-class, multiethnic workers: results of the healthy directions—small business study. Am J Public Health (2005) 95(8):1389–1395.
79. Sorensen G, Stoddard AM, LaMontagne AD, et al. A comprehensive worksite cancer prevention intervention: behavior change results from a randomized controlled trial (United States). Cancer Causes Control (2002) 13(6):493–502.[CrossRef][Web of Science][Medline]
80. Stoddard AM, Fagan P, Sorensen G, Hunt MK, Frazier L, Girod K. Reducing cigarette smoking among working adolescents: results from the SMART study. Cancer Causes Control (2005) 16(10):1159–1164.[CrossRef][Web of Science][Medline]
81. Taylor VM, Jackson JC, Yasui Y, et al. Evaluation of an outreach intervention to promote cervical cancer screening among Cambodian American women. Cancer Detect Prev (2002) 26(4):320–327.[CrossRef][Web of Science][Medline]
82. Thomas DB, Gao DL, Ray RM, et al. Randomized trial of breast self-examination in Shanghai: final results. J Natl Cancer Inst (2002) 94(19):1445–1457.
83. Thompson B, Coronado G, Chen L. Celebremos La Salud! A community randomized trial of cancer prevention (United States). Cancer Causes Control (2006) 17(5):733–746.[CrossRef][Web of Science][Medline]
84. Vallerand AH, Riley-Doucet C, Hasenou S, Templin T. Improving cancer pain management by homecare nurses. Oncol Nurs Forum (2004) 31(4):809–816.[CrossRef][Web of Science][Medline]
85. Vinker S, Nakar S, Rosenberg E, Kitai E. The role of family physicians in increasing annual fecal occult blood test screening coverage: a prospective intervention study. Isr Med Assoc J (2002) 4(6):424–425.[Web of Science][Medline]
86. Walsh JME, Salazar R, Terdiman JP, Gildengorin G, Perez-Stable EJ. Promoting use of colorectal cancer screening tests. Can we change physician behavior? J Gen Intern Med (2005) 20(12):1097–1101.[CrossRef][Web of Science][Medline]
87. Walsh MM, Hilton JF, Ellison JA, et al. Spit (smokeless) tobacco intervention for high school athletes: results after 1 year. Addict Behav (2003) 28(6):1095–1113.[CrossRef][Web of Science][Medline]
88. Weller D, May F, Rowett D, et al. Promoting better use of the PSA test in general practice: randomized controlled trial of education strategies based on outreach visits and mailout. Fam Pract (2003) 20(6):655–661.
89. Wilson BJ, Torrance N, Mollison J, et al. Cluster randomized trial of a multifaceted primary care decision-support intervention for inherited breast cancer risk. Fam Pract (2006) 23(5):537–544.
90. Yi JK, Luong KN-T. Apartment-based breast cancer education program for low income Vietnamese American women. J Community Health (2005) 30(5):345–353.[CrossRef][Web of Science][Medline]
91. Young JM, D'Este C, Ward JE. Improving family physicians' use of evidence-based smoking cessation strategies: a cluster randomization trial. Prev Med (2002) 35(6):572–583.[CrossRef][Web of Science][Medline]
92. Zhu K, Hunter S, Bernard LJ, et al. An intervention study on screening for breast cancer among single African-American women aged 65 and older. Prev Med (2002) 34(5):536–545.[CrossRef][Web of Science][Medline]
93. Aitken JF, Elwood JM, Lowe JB, Firman DW, Balanda KP, Ring IT. A randomised trial of population screening for melanoma. J Med Screen (2002) 9(1):33–37.
94. Fitzgibbon ML, Stolley MR, Dyer AR, VanHorn L, KauferChristoffel K. A community-based obesity prevention program for minority children: rationale and study design for Hip-Hop to Health Jr. Prev Med (2002) 34(2):289–297.[CrossRef][Web of Science][Medline]
95. Ganz PA, Farmer MA, Belman M, et al. Improving colorectal cancer screening rates in a managed care health plan: recruitment of provider organizations for a randomized effectiveness trial. Cancer Epidemiol Biomarkers Prev (2003) 12(9):824–829.
96. Gortmaker SL, Peterson K, Wiecha J, et al. Reducing obesity via a school-based interdisciplinary intervention among youth. Arch Pediatr Adolesc Med (1999) 153(4):409–418.
97. Haire-Joshu D, Brownson RC, Schechtman K, Nanney MS, Houston C, Auslander W. A community research partnership to improve the diet of African Americans. Am J Health Behav (2001) 25(2):140–146.[Web of Science][Medline]
98. Holt CL, Kyles A, Wiehagen T, Casey C. Development of a spiritually based breast cancer educational booklet for African American women. Cancer Control (2003) 10(5 suppl):37–44.[Medline]
99. Hunt MK, Fagan P, Lederman R, et al. Feasibility of implementing intervention methods in an adolescent worksite tobacco control study. Tob Control (2003) 12(suppl 4):iv40–iv45.
100. Jordhoy MS, Fayers P, Loge JH, Ahlner-Elmqvist M, Kaasa S. Quality of life in palliative cancer care: results from a cluster randomized trial. J Clin Oncol (2001) 19(18):3884–3894.
101. Jordhoy MS, Fayers P, Saltnes T, Ahlner-Elmqvist M, Jannert M, Kaasa S. A palliative-care intervention and death at home: a cluster randomised trial. Lancet (2000) 356(9233):888–893.[CrossRef][Web of Science][Medline]
102. Lytle L, Murray DM, Perry C, et al. School-based approaches to affecting adolescents’ diets: results from the TEENS study. Health Educ Behav (2004) 31(2):270–287.
103. Lytle LA, Perry CL. Applying research and theory in program planning: an example from a nutrition education intervention. Health Promot Pract (2001) 2(1):68–80.[Abstract]
104. Myers RE, Turner B, Weinberg D, et al. Complete diagnostic evaluation in colorectal cancer screening: research design and baseline findings. Prev Med (2001) 33(4):249–260.[CrossRef][Web of Science][Medline]
105. Okamura T, Tanaka T, Babazono A, et al. The high-risk and population strategy for occupational health promotion (HIPOP-OHP) study: study design and cardiovascular risk factors at the baseline survey. J Hum Hypertens (2004) 18(7):475–485.[CrossRef][Web of Science][Medline]
106. Powe BD. Promoting fecal occult blood testing in rural African American women. Cancer Pract (2002) 10(3):139–146.[CrossRef][Web of Science][Medline]
107. Reuben DB, Bassett LW, Hirsch SH, Jackson CA, Bastani R. A randomized clinical trial to assess the benefit of offering on-site mobile mammography in addition to health education for older women. AJR Am J Roentgenol (2002) 179(6):1509–1514.
108. Taylor VM, Schwartz SM, Jackson JC, et al. Cervical cancer screening among Cambodian-American women. Cancer Epidemiol Biomarkers Prev (1999) 8(6):541–546.
109. Thomas DB, Gao DL, Self SG, et al. Randomized trial of breast self-examination in Shanghai: methodology and preliminary results. J Natl Cancer Inst (1997) 89(5):355–365.
110. Thompson B, Coronado GD, Solomon CC, McClerran DF, Neuhouser ML, Feng Z. Cancer prevention behaviors and socioeconomic status among Hispanics and non-Hispanic whites in a rural population in the United States. Cancer Causes Control (2002) 13(8):719–728.[CrossRef][Web of Science][Medline]
111. Tilley BC, Glanz K, Kristal AR, et al. Nutrition intervention for high-risk auto workers: results of the Next Step Trial. Prev Med (1999) 28(3):284–292.[CrossRef][Web of Science][Medline]
112. Tilley BC, Vernon SW, Myers RE, et al. The Next Step Trial: impact of a worksite colorectal cancer screening promotion program. Prev Med (1999) 28(3):276–283.[CrossRef][Web of Science][Medline]
113. Brown H, Prescott R. Applied Mixed Models in Medicine (2006) 2nd ed. Chichester, UK: John Wiley & Sons, Inc.
114. Raudenbush SW, Bryk AS. Hierarchical Linear Models (2002) 2nd ed. Thousand Oaks, CA: Sage Publications, Inc.
115. Murray DM, Hannan PJ, Wolfinger RD, Baker WL, Dwyer JH. Analysis of data from group-randomized trials with repeat observations on the same groups. Stat Med (1998) 17(14):1581–1600.[CrossRef][Web of Science][Medline]
116. Gail MH, Mark SD, Carroll RJ, Green SB, Pee D. On design considerations and randomization-based inference for community intervention trials. Stat Med (1996) 15(11):1069–1092.[CrossRef][Web of Science][Medline]
117. Murray DM, Hannan PJ, Varnell SP, McCowen RG, Baker WL, Blitstein JL. A comparison of permutation and mixed-model regression methods for the analysis of simulated data in the context of a group-randomized trial. Stat Med (2006) 25(3):378–388.
118. Liang KY, Zeger SL. Longitudinal data analysis using generalized linear models. Biometrika (1986) 73(1):13–22.
119. Zeger SL, Liang K-Y. Longitudinal data analysis for discrete and continuous outcomes. Biometrics (1986) 42(1):121–130.[CrossRef][Web of Science][Medline]
120. Murray DM, Varnell SP, Blitstein JL. Design and analysis of group-randomized trials: a review of recent methodological developments. Am J Public Health (2004) 94(3):423–432.
121. Varnell SP, Murray DM, Baker WL. An evaluation of analysis options for the one group per condition design: can any of the alternatives overcome the problems inherent in this design? Eval Rev. (2001) 25(4):440–453.
122. Murray DM, Blitstein JL. Methods to reduce the impact of intraclass correlation in group-randomized trials. Eval Rev. (2003) 27(1):79–103.
123. Baldwin SA, Murray DM, Shadish WR. Empirically supported treatments or type-1 errors?: a revaluation of group-administered treatments on the empirically supported treatments list. J Consult Clin Psychol (2005) 73(5):924–935.[CrossRef][Web of Science][Medline]
124. Pals SP, Murray DM, Alfano CM, Shadish WR, Hannan PJ, Baker WL. Individually randomized group treatment studies: are the most frequently used analytic models misleading? Am J Public Health. In press.
125. Bellamy SL, Gibberd R, Hancock L, et al. Analysis of dichotomous outcome data for community intervention studies. Stat Methods Med Res. (2000) 9(2):135–159.
126. Feng Z, McLerran D, Grizzle J. A comparison of statistical methods for clustered data analysis with Gaussian error. Stat Med (1996) 15(16):1793–1806.[CrossRef][Web of Science][Medline]
127. Mancl LA, DeRouen TA. A covariance estimator for GEE with improved small-sample properties. Biometrics (2001) 57(1):126–134.[CrossRef][Web of Science][Medline]
128. Raudenbush SW, Martinez A, Spybrook J. Strategies for improving precision in group-randomized experiments. Educ Eval Policy Anal (2007) 29(1):5–29.[CrossRef]
Manuscript received October 16, 2007; revised January 21, 2008; accepted February 20, 2008.
Editorial about this Article
Related Article in JNCI
CiteULike 
Connotea 
Del.icio.us What's this?
J Natl Cancer Inst 2008 100: 452-453.
J Natl Cancer Inst 2008 100: 447.
This article has been cited by other articles:
|
|
 |
|
  R. Passalacqua, C. Caminiti, F. Campione, F. Diodati, R. Todeschini, G. Bisagni, R. Labianca, M. Dalla Chiesa, R. Bracci, M. Aragona, et al. Prospective, Multicenter, Randomized Trial of a New Organizational Modality for Providing Information and Support to Cancer Patients J. Clin. Oncol., April 10, 2009; 27(11): 1794 - 1799. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Schrag Evaluating the Impact of Organizational Changes in Health Care Delivery: Challenges in Study Design J. Clin. Oncol., April 10, 2009; 27(11): 1744 - 1745. [Full Text] [PDF]
|
|
|
|
 |
|
 V. M. Taylor, T. T. Nguyen, J. C. Jackson, and S. J. McPhee Cervical Cancer Control Research in Vietnamese American Communities Cancer Epidemiol. Biomarkers Prev., November 1, 2008; 17(11): 2924 - 2930. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. R. Church Separation of Health and Statistics J Natl Cancer Inst, April 2, 2008; 100(7): 452 - 453. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||