|The effectiveness of interventions to promote mammography among women with historically lower rates of screening
|Legler J, Meissner HI, Coyne C, Breen N, Chollette V, Rimer B K
To describe mammography intervention research efforts in diverse populations; to determine which intervention characteristics explain variation in intervention effectiveness; and to assess the effectiveness of particular types of interventions and this benefit for specific populations of women with historically lower use of mammography (older women, women with low incomes and non-Caucasian women).
A 'parent' database of articles that described screening mammography interventions (see Other Publications of Related Interest no.2) was established from initial searches of MEDLINE, the Science Citation Index and PsycINFO between 1984 and 1997. These were updated by further searches of MEDLINE, the Social Sciences Citation Index and EMBASE for studies published in the English language between 1984 and August 2000. The reference lists in reviews and other published articles were also checked. National Cancer Institute-funded investigators with breast- screening intervention grants were contacted for pre-publication manuscripts or published articles that had not been entered onto journal databases.
Study designs of evaluations included in the review
Studies with an experimental or quasi-experimental design, which included a control group, were eligible for inclusion in the review.
Specific interventions included in the review
Interventions consisting of strategies designed to increase the use of mammography in asymptomatic women in the target population, such as tailored and untailored reminder letters, telephone calls and counselling, were eligible for inclusion in the review. The interventions were categorised by Rimer's intervention typology (see Other Publications of Related Interest no.1) into:
individual directed, e.g. one-to-one counselling, tailored and untailored letters and reminders, and telephone counselling, which was associated with either a health care or community setting;
system directed, e.g. provider prompts;
access enhancing, e.g. transportation to appointments, facilitated scheduling, mobile vans, vouchers and reduced cost mammograms; social network, e.g. peer leaders and lay health advisors; community education;
mass media campaigns; and
multi-strategy or social networking, i.e. combinations of the interventions listed above. Interventions that exclusively focused on physicians or office systems as a means to increase mammography use were excluded. The control groups were characterised as those whose members received either no intervention, usual care (i.e. the standard care provided to clinic-based populations, although in some cases this may have been quite comprehensive), minimal intervention (e.g. mailed informational pamphlets), or some other nonbreast cancer intervention identified by the authors.
Participants included in the review
Women were eligible for inclusion in the review if they were older (defined as at least 60 years of age), asymptomatic of breast disease, had a high school education or less, were in receipt of a low income (as defined by the study authors), were non-Caucasian (i.e. were members of an ethnic or racial group) and/or lived in a rural or an inner city area.
Outcomes assessed in the review
The outcomes were study-specific adherence rates, as defined by the study author(s). These were generally defined as the receipt of a mammogram, either by self-report or a verified report in a clinical database (e.g. chart-reviews, medical record) and/or medical claims, within a specified number of months, which varied from study to study. When the outcomes were assessed as several points in time, the first outcome was designated as the primary outcome unless the study authors identified some other outcome as primary. Only outcomes involving diverse populations were included among the comparisons.
How were decisions on the relevance of primary studies made?
The authors do not state how the papers were selected for the review, or how many of the reviewers performed the selection.
Assessment of study quality
The authors do not state that they assessed validity.
For consistency, one co-author entered the data for every article; the data were double-entered by another co-author to ensure reliability. Any discrepancies were identified and reconciled by two of the co-authors, and then subjected to approval by all of the team co-authors. Data were extracted on: the study authors; the characteristics of the nature of the intervention according to category (Rimer's intervention typology); dates of the period of the intervention; year in which the data were collected and the year of publication; characteristics of the intended recipients of the intervention (e.g. individual or groups of women); sub-populations of interest (e.g. older, low income, ethnic minority, and urban or rural women); delivery methods (e.g. in-person, mail, telephone, TV/radio/print); and settings (community, clinic, worksite, church, and housing unit). Data were also extracted on the number of different strategies reported to ascertain the variety and, to a lesser extent, the intensity of the intervention approaches. Estimates of the intervention effect for each study comparison, along with 95% confidence intervals (CIs), were computed using the difference in the post-intervention adherence rates between the intervention and control groups. A parallel analysis was carried out using log odds ratios (ORs). When a pre-test post-test design was used with baseline mammography rates reported, the intervention effect was calculated as the difference in the changes from pre- to post-intervention for the intervention and control groups. When a pre-intervention survey was conducted to ascertain study eligibility, only the post-intervention rates were used to calculate the intervention effects. When only the post-intervention screening rates were used, the SEs for the estimated intervention effects were approximated using square roots. When the difference in the pre-intervention screening rates was available, the approximate variance was calculated.
Methods of synthesis
How were the studies combined?
The data were combined both narratively and quantitatively. Intervention descriptors were summarised across all strategies used to reflect the salient features of an intervention group's experience. Key study level characteristics were aggregated and summarised to present a picture of the literature, and the data were pooled in a statistical meta-analysis.
The combined ORs and 95% CIs were estimated, based upon a random- effects model for each intervention type. The combined intervention effect estimates for each group of studies were adjusted for the number of months used for the outcome, by including the number of months centred at 12 months in the model. The combined effects are, therefore, estimates of the excess percentage of screening in the intervention group for mammograms obtained in the past 12 months.
How were differences between studies investigated?
Heterogeneity was investigated using study-level characteristics, with mixed model regressions weighted to accommodate with- and between-study variation. Heterogeneity of the intervention effects was tested using the Q statistic of DerSimonian and Laird (see Other Publications of Related Interest no.3). When the result of the Q test was significant, a random-effects model was used to accommodate the heterogeneity. Potentially influential studies and outliers were removed from the analyses to gauge their impact on the study results. Analyses were also performed with and without the international studies. The combined ORs were estimated for the following subgroups: older women; women with low income; and non-Caucasian women.
Results of the review
Thirty-eight studies (at least 26,532 women) were included in the review.
Significant heterogeneity was found among intervention effects associated with comparisons based upon the single most intense intervention from each entire study (Q=218, p<0.001, compared with a chi-squared with 34 degrees of freedom).
For access-enhancing intervention strategies, pooled data from 14 studies produced a statistically-significant increase in mammography use of 18.9% (95% CI: 10.4, 27.4; OR 2.3, 95% CI: 1.7, 3.1).
For individual-directed strategies associated with a health care setting, data from 15 studies produced a statistically-significant increase in mammography use of 17.6% (95% CI: 11.6, 24.0; OR 2.5, 95% CI: 1.9, 3.4).
For individual-directed strategies associated with a community setting, 13 studies produced an increase in mammography use of 6.8% (95% CI: 1.8, 11.8; OR 1.3, 95% CI: 1.0, 1.6), which was of marginal statistical significance.
For community education, 14 studies produced a statistically- significant increase in mammography use of 9.7% (95% CI: 3.9, 15.6; OR 1.5, 95% CI: 1.2, 1.9).
For media campaigns, 6 studies produced an increase in mammography use of 5.9% (95% CI: 0.3, 11.5; OR 1.3, 95% CI: 1.0, 1.8), which was of marginal statistical significance.
For social network, 7 studies produced an increase in mammography use of 5.8% (95% CI: -0.2, 11.9; OR 1.4, 95% CI: 1.0, 2.0), which was of marginal statistical significance.
The use of multiple intervention types was also found to be effective, with pooled effects from 26 studies averaging 13.3% overall (95% CI: 8.6, 18.0). With the exception of social network interventions, the estimated intervention effects were significantly greater than zero for all of the groupings, although each grouping exhibited significant heterogeneity.
For combinations of interventions, the pooled data from 9 studies indicated that the strongest combination of approaches used access-enhancing and individual-directed strategies, which resulted in a 26.9% increase in mammography use (95% CI: 9.9, 43.9). The pooled data from 5 studies showed that a combination of access-enhancing and system-directed interventions resulted in a 19.4% increase (95% CI: 8.2, 30.6).
As a reflection of secular trends in mammography screening, the control screening rates increased significantly over time (p=0.02); this was parallelled with a concomitant decline in intervention effects (p=0.04).
In 11 studies with larger proportions of older women, the estimated intervention effects tended to be greater (7.9%, 95% CI: 10.5, 25.4, p=0.01) and the control group rates lower (p=0.02). The pattern was similar in 26 studies where the comparisons consisted of more than 40% low income women (12.7%, 95% CI: 7.3, 18.1), although these were not statistically significant. For 24 studies of comparisons where non-Caucasian women were more than 40%, the estimated effects were 12% (95% CI: 6.7, 17.4). If the comparisons were limited to at least 40% African-American women, the estimated effects (16 studies) were 11.6% (95% CI: 6.4, 16.7). Significant heterogeneity was found for each of these groupings. Regression analyses explained the variation between studies by indicators of the use of access-enhancing approaches.
Further results from the different statistical analyses were provided in the review.
There was evidence of substantial effects for categories of interventions directed at subsets of populations that have been historically underserved. Whereas racial or ethnic differences in mammography use are no longer evident, significant differentials in terms of education, income, health insurance coverage and having a usual source of health care, still persist. Therefore, access-enhancing strategies are an important compliment to individual- and system-directed interventions for women with historically lower rates of mammography screening, who may lack the resources to readily learn about or obtain these services.
The review question and the study selection criteria were stated clearly. The literature search seemed reasonably comprehensive with efforts made to find additional material. However, limiting the search to papers published in English may mean that some studies were missed. The authors provided relatively little information on the methodology of the literature selection process, and it is unclear whether the primary studies were assessed for validity. However, there was ample information on the data extraction process and analyses. An appropriate range of statistical tests seems to have been undertaken, although it was not reported whether publication bias was tested for. There was ample presentation of the results, both in the text and graphically, and a good discussion of the findings. While no formal conclusions were offered, these are readily interpreted from the 'Discussion' section, and seem valid in the light of the data the authors present.
Implications of the review for practice and research
Practice: The authors state that access-enhancing strategies are an important complement to individual- and system-directed interventions for women with historically lower rates of mammography screening, who may lack the resources to readily learn about or obtain these services. The authors further state that practitioners and policy makers should be encouraged to select and promote efficacious interventions.
Research: The authors state that future research should measure the effects of the different strategies used, and their intensity, to evaluate the effectiveness of each element that comprises an intervention package. The authors further state that researchers should build on the cumulative knowledge base to design the next generation of behavioural interventions.
Legler J, Meissner HI, Coyne C, Breen N, Chollette V, Rimer B K. The effectiveness of interventions to promote mammography among women with historically lower rates of screening. Cancer Epidemiology, Biomarkers and Prevention 2002; 11(1): 59-71
Other publications of related interest
1. Rimer BK. Mammography use in the US: trends and the impact of interventions. Ann Behav Med 1994;16:317-26. 2. Meissner HI, Breen N, Coyne C, Legler JM, Green DT, Edwards BK. Breast and cervical cancer screening interventions: an assessment of the literature. Cancer Epidemiol Biomark Prev 1998;7:951-61. 3. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986;7:177-88.
Subject indexing assigned by NLM
Aged; Breast Neoplasms /ethnology /prevention & Case-Control Studies; Clinical Trial; Confidence Intervals; Educational Status; Female; Health Promotion; Incidence; Mammography /statistics & Mass Screening /statistics & Middle Aged; Minority Groups; Poverty; Regression Analysis; Rural Population; Socioeconomic Factors; United States /epidemiology; World Health; control; numerical data; numerical data
Date bibliographic record published
Date abstract record published
This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.