No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Alternative Strategies for Selecting Population Controls Comparison of Random Digit Dialing and Targeted Telephone Calls
Abstract
This project was carried out to identify a valid framework for selecting controls to be used in a population-based case-control study of breast cancer, and to compare participation rates and characteristics between women contacted using a standard random digit dialing (RDD) strategy and those who were sent a letter of presentation prior to telephone contact (targeted telephone calls, TTC).
Twelve hundred women, ages 20-74, were sampled from the Department of Motor Vehicles (DMV) and Health Care Financing Administration (HCFA) records. Women for whom telephone numbers were obtained (N = 771) were randomly assigned to RDD or TTC. The respondents participated in a brief telephone interview. Odd ratios (OR) and their 95% confidence intervals (CI) were used to estimate differences in characteristics of the respondents between the two contact strategies.
Telephone numbers were obtained for 79% of women aged > or = 55 years and for only 38% of women aged < 55 years. Interviews were obtained for 48% of women for whom we obtained telephone numbers, and for 77% of women for whom eligibility was confirmed via telephone contact. Participation of target women appeared to be higher for the TTC than the RDD group (42% vs. 35%, p = 0.054). Among respondents who were > or = 55 years old, those in the TTC group were 80% more likely (OR = 1.8, 95% CI: 0.9-3.4) to report a serious medical condition than women in the RDD group, 60% less likely (OR = 0.4, 95% CI: 0.2-1.0) to report having used oral contraceptives, and 80% less likely (OR = 0.2, 95% CI: 0.1-0.5) to report having had breast surgery.
Characteristics of respondents differed according to method of contact. These differences, along with the sampling frame used, should be considered when interpreting findings of case-control studies.
... An unexpected, but notable finding came from the researcher experience which suggested that email contact prior to the recruitment phone call, regardless of group allocation, helped promote participants' engagement with the study. This finding is contrary to the Cochrane Review by Mapstone et al. [30] who reported that prewarning participants was not beneficial to enrolment, examining methods including sending a letter a week before [31], a postcard prior to sending a questionnaire [32] and leaving a voicemail [33]. However, none of these studies investigated the use of email to pre-warn participants, and our finding may be attributed to a shift in type of contact preference towards email. ...
Background
Reaching recruitment targets in randomised controlled trials is a challenge. Media tools are increasingly used to engage participants, yet there is a paucity of research into the use of video to optimise recruitment. We therefore tested whether adding a participant information video clip to a standard participant information sheet improved recruitment into a parenting trial.
Methods
One hundred seven participants were randomised to receive either a participant information sheet ( n = 51) or an informational video clip ( n = 56) as part of an email contact following a screening phase. All participants went on to receive the information sheet as part of the existing consent procedure.
Results
The video condition did not increase the odds of recruitment into the trial, such that those in the video condition were significantly less likely to participate in the main trial (OR = 0.253, CI = 0.104–0.618, p = 0.003).
Conclusion
The introduction of a video clip into the recruitment stages of a parenting trial did not lead to an improvement in recruitment; however, the small sample size precludes definitive inferences. We offer reflections on challenges encountered in implementing the SWAT and suggestions for other researchers seeking to embed recruitment SWATs into similar trials.
Trial registration
Current controlled trials ISRCTN 58327365 . Registered on 19 March 2015.
SWAT registration
SWAT 106; Effects of a video clip on recruitment into a randomised trial. Registered on 20 December 2016.
... These include-but are not limited to-snowball sampling, in which participants are asked to name other potential participants [3,4]; and respondent-driven sampling, which combines snowball sampling with statistical weighting to account for nonrandom sampling [5][6][7]. Other studies have used advertisements in the form of postings, online ads, radio, or TV; or direct outreach, for example "cold-calling" potential participants [8]. Both opt-in and opt-out methods have also been used. ...
Background
Participant recruitment is an ongoing challenge in health research. Recruitment may be especially difficult for studies of access to health care because, even among those who are in care, people using services least often also may be hardest to contact and recruit. Opt-out recruitment methods (in which potential participants are given the opportunity to decline further contact about the study (opt out) following an initial mailing, and are then contacted directly if they have not opted out within a specified period) can be used for such studies. However, there is a dearth of literature on the effort needed for effective opt-out recruitment. Methods
In this paper we describe opt-out recruitment procedures for two studies on access to health care within the U.S. Department of Veterans Affairs. We report resource requirements for recruitment efforts (number of opt-out packets mailed and number of phone calls made). We also compare the characteristics of study participants to potential participants via t-tests, Fisher’s exact tests, and chi-squared tests. ResultsRecruitment rates for our two studies were 12 and 21%, respectively. Across multiple study sites, we had to send between 4.3 and 9.2 opt-out packets to recruit one participant. The number of phone calls required to arrive at a final status for each potentially eligible Veteran (i.e. study participation or the termination of recruitment efforts) were 2.9 and 6.1 in the two studies, respectively. Study participants differed as expected from the population of potentially eligible Veterans based on planned oversampling of certain subpopulations. The final samples of participants did not differ statistically from those who were mailed opt-out packets, with one exception: in one of our two studies, participants had higher rates of mental health service use in the past year than did those mailed opt-out packets (64 vs. 47%). Conclusions
Our results emphasize the practicality of using opt-out methods for studies of access to health care. Despite the benefits of these methods, opt-out alone may be insufficient to eliminate non-response bias on key variables. Researchers will need to balance considerations of sample representativeness and feasibility when designing studies investigating access to care.
... Via this approach, we achieved an overall survey completion rate of 67%, with only 3% declining participation, and required a median (IQR) of 3 (2, 5) contact attempts among respondents. This combined 67% survey completion rate, with individual response rates of 49% in the mail trial and 51% in the telephone trial, compares favorably with other studies [10,22,45]. Comparison of response rates between the mail versus telephone trials within our study was not conducted since participants in telephone trial represented nonresponders from mail trial or participants without a valid mailing address, and thus a different population from the mailing trial. ...
Effective strategies for contacting and recruiting study participants are critical in conducting clinical research. In this study, we conducted two sequential randomized controlled trials of mail- and telephone-based strategies for contacting and recruiting participants, and evaluated participant-related variables' association with time to survey completion and survey completion rates. Subjects eligible for this study were survivors of acute lung injury who had been previously enrolled in a 12-month observational follow-up study evaluating their physical, cognitive and mental health outcomes, with their last study visit completed at a median of 34 months previously.
Eligible subjects were contacted to complete a new research survey as part of two randomized trials, initially using a randomized mail-based contact strategy, followed by a randomized telephone-based contact strategy for non-responders to the mail strategy. Both strategies focused on using either a personalized versus a generic approach. In addition, 18 potentially relevant subject-related variables (e.g., demographics, last known physical and mental health status) were evaluated for association with time to survey completion.
Of 308 eligible subjects, 67% completed the survey with a median (IQR) of 3 (2, 5) contact attempts required. There was no significant difference in the time to survey completion for either randomized trial of mail- or phone-based contact strategy. Among all subject-related variables, age <=40 years and minority race were independently associated with a longer time to survey completion.
We found that age <=40 years and minority race were associated with a longer time to survey completion, but personalized versus generic approaches to mail- and telephone-based contact strategies had no significant effect. Repeating both mail and telephone contact attempts was important for increasing survey completion rate.Trial registration: NCT00719446.
... telephone numbers we called were nonworking. This rate is much lower than the RDD rates reported in other studies (11,17,24,25), which range from 9.3% to 29.6%. In addition, nonresidential numbers accounted for only 2% of the numbers called from the InfoUSA.com ...
The selection of controls is an important methodological consideration for case-control studies. Neighborhood-matched control selection is particularly crucial for studies of vector-borne disease, such as Lyme disease, for which risk is intrinsically linked to geographical location. The matching of case-control pairs on neighborhood can help control for variation in ecological risk factors that are tied to geographical location, like vector and host habitat in the peridomestic environment. Random-digit dialing has been used to find neighborhood controls by using the area code and exchange of the case to generate lists of potential control households. An alternative to random-digit dialing is the purchase of residential telephone numbers from a commercial marketing database. This report describes the utility of the InfoUSA.com (InfoGroup, Papillion, Nebraska) commercial marketing database for neighborhood control recruitment in a Lyme disease case-control study in Connecticut during 2005-2007.
... Außerdem unterscheiden sich Personen, die telefonische Anfragen nur ungern beantworten , von solchen, die bereitwillig an Umfragen teilnehmen [18]. So werden etwa junge Frauen seltener durch Zufallsziffernanwahl ermittelt als andere [19]. Aber auch wenn diese Eigenarten der telefonischen Erfassung zu Bias führen können [15], sind die auf diese Weise selektierten Kontrollgruppen für die Referenzpopulation doch weitgehend repräsentativ [17]. ...
In einem alten Film wurde der Komiker Groucho Marx gefragt: "Groucho, how is your wife?" Groucho witzelte: "Com-pared to what?" 1 Auch wenn seine Ant-wort ("Verglichen womit?") nach heuti-gem Ermessen sexistisch anmuten mag, so passt sie doch genau zu der Frage, die sich bei den Ergebnissen von Fall-Kon-troll-Studien stellt, nämlich im Vergleich wozu? Denn die Validität von Schluss-folgerungen hängt davon ab, ob sich ei-ne geeignete Vergleichsgruppe finden lässt. Mit anderen Worten: Die Verwen-dung nicht so geeigneter Kontrollgrup-pen hat schon vielen wissenschaftlichen Arbeiten die Basis entzogen. Kontrollgruppen sind ein aussagekräfti-ges und seit langem bewährtes Instru-ment der Forschung. Die allererste Do-kumentation einer Vergleichsgruppe findet sich schon in der Bibel im Buch Daniel [1]. Daniel und seine drei Freun-de, die vom babylonischen König Nebu-kadnezar gefangen gehalten wurden, nahmen am Hof des Königs an einer zehntägigen "Studie" teil, in der ge-sunde Ernährung mit dem königlichen Speiseplan verglichen wurde. Am Ende schienen Daniel und seine Freunde ge-sünder zu sein als die babylonischen Ju-gendlichen, die das übliche Essen zu sich nahmen. Im Laufe der Jahre ist die-se "Studie" nicht nur wegen ihrer un-zureichenden Expositionsdauer, son-dern auch wegen der wahrscheinlich göttlichen Begleitintervention in die Kritik geraten. Jedenfalls sind Kontroll-gruppen erst einmal für mehrere Jahr-tausende aus der Literatur verschwun-den! An James Linds Untersuchung über die Behandlung von Skorbut aus dem Jahre 1747 hat sich das Interesse an gleich-zeitig beobachteten Kontrollgruppen erneut entzündet [2]. Trotz ihrer gerin-gen Größe (sechs Behandlungsgruppen mit je zwei Seeleuten) zeigte die Studie, wie nützlich die Nahrungsergänzung mit Zitrusfrüchten ist. In Studien ohne Randomisierung kann das Zusammen-stellen einer angemessenen Kontroll-gruppe manchmal eine wahre Heraus-forderung darstellen. Wir werden hier die Rolle von Kontrollgruppen in Fall-Kontroll-Studien erläutern, die beson-deren, mit ihrer Auswahl verbundenen Probleme beschreiben und verschiede-ne Implikationen erörtern.
... In our study, the overall rate of contact with participants by the end of the 12-week standardized recruitment protocol was relatively high at 72% (of whom 95% provided informed consent), compared to the 6 -60% rates experienced in other trials of recruitment strategies (8,23). We attribute this relatively high rate of contact to our repeated contact attempts and the use of different methods to contact patients. ...
To study the effect of different mail- and phone-based strategies, along with patient- and research-related factors, on the time to contact with research participants.
A prospective evaluation of a 12-week standardized protocol (embedded with two randomized trials of mail- and phone-based strategies) for contacting existing research participants for recruitment into a related study.
Of 146 participants, 87 were eligible for contact via the standardized protocol, and 63 (72%) of these were successfully contacted within 12 weeks after multiple mail- and phone-based efforts. Using Cox proportional hazards regression analysis, the different mail and phone strategies showed no significant difference in the time to contact with participants. Of 34 patient- and research-related factors evaluated, only two were independently associated with time to contact among all 146 participants: (1) participants having their last visit conducted outside of the research clinic because of patient illness/condition had a longer time to contact and (2) those with a self-reported chronic fatigue history had a shorter time to contact.
Few patient characteristics and research-related factors accurately predict time to contact. Repeated attempts using different strategies are important for successful and timely contact with study participants.
... This problem also is similar to telephone surveys using RDD 210 . In recent studies using RDD, estimated response rates from 28% to 35% have been reported 209,211 . In an e-mail study of a defined group of employees in Alaska 91% of questionnaires were returned 59 . ...
... The actual process of selecting a representative sample is an ongoing challenge in both casecontrol and cohort studies. Technology, privacy laws, and regulations are severely limiting the use of established methods of selection, such as random digit dialing (RDD) (Brogan et al., 2001) and Department of Motor Vehicles (DMV) records (Funkhouser et al., 2000 ). Hospitalbased and neighborhood recruitment strategies can be time-consuming, and are more likely to produce samples that are not representative of the general population (Wacholder et al., 1992). ...
Neurodevelopmental disabilities affect 3–8% of the 4 million babies born each year in the U.S. alone, with known etiology for less than 25% of those disabilities. Numerous investigations have sought to determine the role of environmental exposures in the etiology of a variety of human neurodevelopmental disorders (e.g., learning disabilities, attention deficit-hyperactivity disorder, intellectual disabilities) that are manifested in childhood, adolescence, and young adulthood. A comprehensive critical examination and discussion of the various methodologies commonly used in investigations is needed. The Hershey Medical Center Technical Workshop: Optimizing the design and interpretation of epidemiologic studies for assessing neurodevelopmental effects from in utero chemical exposure provided such a forum for examining these methodologies. The objective of the Workshop was to develop scientific consensus on the key principles and considerations for optimizing the design and interpretation of epidemiologic studies of in utero exposure to environmental chemicals and subsequent neurodevelopmental effects. (The Panel recognized that the nervous system develops post-natally and that critical periods of exposure can span several developmental life stages.) Discussions from the Workshop Panel generated 17 summary points representing key tenets of work in this field. These points stressed the importance of:
Organizations and individuals usually have well defined objectives at any given point of time and invest their resources systematically to fullfill the objectives by setting their goal and formulating and following a strategy to chase their goal. Various strategies are used to chase the goal in both organizational and individual cases include competitive strategies or red ocean strategies, monopoly or blue ocean strategies, sustainability or green ocean strategies, survival or black ocean strategies, and mixed or white ocean strategies. The complexities of business decisions after globalization of business and technification of business processes, winning or sustaining or even surviving in current business is considered as an intensive challenge for organizations and also to individuals. In such environment called the turbulent business or social environment for organizations and individuals respectively, a new strategy for survival called “Alternative Strategy” is proposed in this paper and the concept of the strategy, its importance in the current business environment for organizational decisions are discussed. The paper also looks into the understanding of the Generic strategies, their applicability and constraints while identifying the importance of alternative strategy. Eight postulates are developed to support the concept of alternative strategy, and an alternative strategy model is suggested using lateral thinking techniques. The model consists of evaluating and comparing the alternative strategy with primal strategy using ABCD analysis framework. Such model of alternative strategy can be used at operational level, tactical level, and strategic level of any organization to realize its objectives.
Various strategies are used to chase the goal in both organizational and individual cases include competitive strategies or red ocean strategies, monopoly or blue ocean strategies, sustainability or green ocean strategies, survival or black ocean strategies, and mixed or white ocean strategies. The complexities of business decisions after globalization of business and technification of business processes, winning or sustaining or even surviving in current business is considered as an intensive challenge for organizations and also to individuals. In such environment called the turbulent business or social environment for organizations and individuals respectively, a new strategy for survival called "Alternative Strategy" is proposed in this paper and the concept of the strategy, its importance in the current business environment for organizational decisions are discussed. The paper also looks into the understanding of the Generic strategies, their applicability and constraints while identifying the importance of alternative strategy. Eight postulates are developed to support the concept of alternative strategy, and an alternative strategy model is suggested using lateral thinking techniques. The model consists of evaluating and comparing the alternative strategy with primal strategy using ABCD analysis framework. Such model of alternative strategy can be used at operational level, tactical level, and strategic level of any organization to realize its objectives.
Researchers do many untested things to try to increase recruitment into studies. Many studies fail to recruit their planned number of participants. Using effective ways to improve recruitment would be of benefit to the research community and society. The reviewers did not find any trials of methods aimed at researcher collaborators or ethics committees. Fifteen trials of methods aimed at participants were identified, but the results are not easily generalizable.
The increasing popularity of the telephone interview as a research method may be a reflection of broader social change and technological advances, with increased use and acceptability of telecommunications to support healthcare and service industries in general. Despite its widespread use there are few definitions of the term. Studies which directly compare telephone and face-to-face interviewing tend to conclude that telephone interviewing produces data which are at least comparable in quality to those attained by the face-to-face method. While it has been used for large survey studies, in nursing research the telephone interview is used predominantly in smaller-scale qualitative studies, where contact has already been made with the participants. The telephone interview was used in a study by one of the authors (EC) which explored the experience of postoperative pain. Issues relating to ethical considerations, reliability, validity, limitations and analysis are explored. The use of the approach for pain research is reflected upon before considering the wider applications that are available for this method of data collection in healthcare practice.
IntroductionData Sources for Analyses of Secular TrendsData Sources for Case–Control StudiesUsing Randomized Clinical Trials as Postmarketing Surveillance StudiesAdditional Databases Useful for Pharmacoepidemiologic ResearchConclusions
References
Evaluation will continue to play an important role for health care managers in all coun- tries, particularly in light of population-based
planning, for several reasons. First, the organiza- tion of formal and informal alliances in the health care industry and
the competition among them will prompt managers to continuously evaluate morbidity, mortality, and disability trends associated
with service delivery and treatment of health problems in order to modify or discard services and treatments that are not
efficient, effective, and of high quality. Second, the number and size of social programs have grown over time with the broadening
scope of government involvement in health and social service issues. The application of epidemiological methods is useful
for monitoring the health of the population and identifying changes over time. Epidemiology is based on the premise that health
problems that can be identified often can be corrected. Last, programs addressing the needs of aging populations will proliferate
and necessitate evaluation to assess changing needs, assure that expectations are met, and promote achievement of the desired
health outcomes.Epidemiological methods can be used effectively to target identified needs and to assess the implications
of programs or interventions for health care policy.
Recruiting participants to trials can be extremely difficult. Identifying strategies that improve trial recruitment would benefit both trialists and health research.
To quantify the effects of strategies to improve recruitment of participants to randomised controlled trials.
We searched the Cochrane Methodology Review Group Specialised Register - CMR (The Cochrane Library (online) Issue 1 2008) (searched 20 February 2008); MEDLINE, Ovid (1950 to date of search) (searched 06 May 2008); EMBASE, Ovid (1980 to date of search) (searched 16 May 2008); ERIC, CSA (1966 to date of search) (searched 19 March 2008); Science Citation Index Expanded, ISI Web of Science (1975 to date of search) (searched 19 March 2008); Social Sciences Citation Index, ISI Web of Science (1975 to date of search) (searched 19 March 2008); and National Research Register (online) (Issue 3 2007) (searched 03 September 2007); C2-SPECTR (searched 09 April 2008). We also searched PubMed (25 March 2008) to retrieve "related articles" for 15 studies included in a previous version of this review.
Randomised and quasi-randomised controlled trials of methods to increase recruitment to randomised controlled trials. This includes non-healthcare studies and studies recruiting to hypothetical trials. Studies aiming to increase response rates to questionnaires or trial retention, or which evaluated incentives and disincentives for clinicians to recruit patients were excluded.
Data were extracted on the method evaluated; country in which the study was carried out; nature of the population; nature of the study setting; nature of the study to be recruited into; randomisation or quasi-randomisation method; and numbers and proportions in each intervention group. We used risk ratios and their 95% confidence intervals to describe the effects in individual trials, and assessed heterogeneity of these ratios between trials.
We identified 27 eligible trials with more than 26,604 participants. There were 24 studies involving interventions aimed directly at trial participants, while three evaluated interventions aimed at people recruiting participants. All studies were in health care. Some interventions were effective in increasing recruitment: telephone reminders to non-respondents (RR 2.66, 95% CI 1.37 to 5.18), use of opt-out, rather than opt-in, procedures for contacting potential trial participants (RR 1.39, 95% CI 1.06 to 1.84) and open designs where participants know which treatment they are receiving in the trial (RR 1.25, 95% CI 1.18 to 1.34). However, some of these strategies have disadvantages, which may limit their widespread use. For example, opt-out procedures are controversial and open designs are by definition unblinded. The effects of many other recruitment strategies are unclear; examples include the use of video to provide trial information to potential participants and modifying the training of recruiters. Many studies looked at recruitment to hypothetical trials and it is unclear how applicable these results are to real trials.
Trialists can increase recruitment to their trials by using the strategies shown to be effective in this review: telephone reminders; use of opt-out, rather than opt-in; procedures for contacting potential trial participants and open designs. Some strategies (e.g. open trial designs) need to be considered carefully before use because they also have disadvantages. For example, opt-out procedures are controversial and open designs are by definition unblinded.
In the Multiple Environmental and Genetic Assessment of risk factors for venous thrombosis (MEGA study), a large population-based case-control study, we investigated lifestyle factors as risk factors for venous thrombosis. Overweight, smoking and alcohol consumption were addressed and pregnancy and the postpartum period were evaluated in women. Due to the large sample size of the study it was possible to investigate the joint effect of these risk factors with important genetic risk factors for venous thrombosis such as the factor V Leiden and the prothrombin 20210A mutation. In addition to these lifestyle related risk factors, two polymorphisms within the promoter region of the protein C gene were studied as risk factors for venous thrombosis and the influence of genotypic variation on plasma protein C levels was assessed. Finally, we described our experience with the inclusion of two different control groups in the MEGA study.
Protein C is an important inhibitor of blood coagulation. We investigated the effect of two polymorphisms within the promoter region of the protein C gene (C/T at position 2405 and A/G at position 2418) on risk of venous thrombosis and on plasma protein C levels. In addition the combined effect of the two polymorphisms with factor V Leiden and oral contraceptive use was investigated. Previous studies on these polymorphisms were small and were not able to investigate synergistic effects. In the Multiple Environmental and Genetic Assessment of risk factors for venous thrombosis (MEGA study), protein C levels were determined in 2,043 patients with venous thrombosis and 2,857 control subjects, and the two polymorphisms in 4,285 patients and 4,863 control subjects. The CC/GG genotype was associated with the lowest protein C levels. Compared to carriers of the TT/AA genotype - a genotype associated with higher protein C levels - the risk of venous thrombosis in CC/GG carriers was 1.3-fold increased (95% confidence interval 1.09-1.48). The combination of factor V Leiden with the CC/GG genotype led to a 4.7-fold increased risk, compared to non-carriers with the TT/AA genotype. Oral contraceptive use together with the CC/GG genotype resulted in a 5.2-fold increased risk. In conclusion, the CC/GG genotype is associated with lower levels of protein C and an elevated risk of venous thrombosis compared to the TT/AA genotype. There is no clear synergistic effect of the CC/GG genotype with factor V Leiden or oral contraceptive use on thrombotic risk.
This paper presents methods and operational results of a population-based case-control study examining the effects of oral contraceptive use on breast cancer risk among white and black women aged 35-64 years in five U.S. locations.
Cases were women newly diagnosed with breast cancer during July 1994 through April 1998. Controls were identified through random digit dialing (RDD) using unclustered sampling with automated elimination of nonworking numbers. Sampling was density-based, with oversampling of black women. In-person interviews were conducted from August 1994 through December 1998. Blood samples were obtained from subsets of cases and controls, and tissue samples were obtained from subsets of cases. A computerized system tracked subjects through study activities. Special attention was devoted to minimizing exposure misclassification, because any exposure-disease associations were expected to be small.
An estimated 82% of households were screened successfully through RDD. Interviews were completed for 4575 cases (2953 whites; 1622 blacks) and 4682 controls (3021 whites; 1661 blacks). Interview response rates for cases and controls were 76.5% and 78.6%, respectively, with lower rates for black women and older women.
The methodologic details of this large collaboration may assist researchers conducting similar investigations.
Challenges to random digit dialing have been documented, but the reliability of random digit dialing outcomes from telephone number calling, household identification, and enumeration has never been addressed, despite its potential to bias population representativeness by affecting completeness of coverage. The authors explored interobserver reliability of calls to numbers generated by random digit dialing for a 1990-1996 population-based case-control study in San Francisco, California, area women, using data from a quality control effort in which 122 of 4,890 random digit dialing numbers were assigned to a second interviewer for recontacting within 4 months. The 34 numbers discrepant between the first and second calls did not differ from the 88 unchanged outcomes, and reliability was good (kappa = 0.65, 95% confidence interval: 0.55, 0.75). Eligibility (an adult woman in the household) was confirmed for nine of 11 eligible households. However, six of 29 households originally ineligible because of gender were eligible on recontact, and eligible residences rose from 24% to 39% between the two calls, although the two groups of eligible women did not differ in age or race. This underenumeration of women by random digit dialing confirms prior observations, although interviewer differences or changes in respondents or household composition between the first and second calls may have contributed. Recontact of gender-ineligible households may improve completeness of random digit dialing coverage for female populations.
Since 1998, serum folate levels have increased nationally after mandatory fortification of cereal grain products with folic acid. Whether serum folate levels have increased among all women has not been well-studied. Identifying characteristics of women with lower serum folate levels would also be helpful in designing educational campaigns. Data for this report were collected during January 2000-January 2001. During 2000, blood samples were collected from 1,059 women aged 18-45 years who attended six family planning clinics in Georgia and analyzed for serum folate levels. This sample included women aged 18-25 years (60%), black women (41%), and women who had a high school education or less (49%). The median serum folate level (8.9 ng/mL) among this population was lower than the median of women of childbearing age (13.0 ng/mL) who participated in the 1999-2000 National Health and Nutrition Examination Survey (NHANES). In logistic regression analysis, women who were black (odds ratio [OR] = 2.4; 95% confidence interval [CI] = 1.48, 3.96), who smoked (OR = 2.1; 95% CI = 1.26, 3.43), or who used Depo-Provera contraceptive injection (manufactured by Pharmacia Corporation, Peapack, New Jersey) (OR = 2.3; 95% CI = 1.15, 4.62) were more likely to be ranked in the lowest quartile (< or = 62 ng/mL) of serum folate concentrations when compared with the highest quartile (> 12.4 ng/mL). Women who consumed cereal regularly (OR = 0.4; 95% CI = 0.26, 0.62) or folic acid supplements (OR = 0.2; 95% CI = 0.09, 0.30) were the least likely to be in the lowest serum folate quartile. This study indicates that certain women are at greater risk for having lower serum folate levels, including women who are black, smokers, Depo-Provera users, and those less likely to eat cereal regularly or to take folic acid supplements. In Georgia, these data are useful in defining target populations (e.g., black women and smokers) for folic acid education campaigns because public health officials can develop contextually appropriate messages and outreach approaches for targeting women for folic acid interventions. Ongoing surveillance of serum folate status among women can guide future intervention efforts.
The use of folic acid is a critical component in preventing birth defects. Health-care providers should take advantage of all health-care visits to counsel not only women at high risk (i.e., those with a history of having an infant with a neural tube defect [NTD]) but all women regarding the importance of folic acid use. A study conducted in Texas confirmed that white and Hispanic mothers were equally likely to recall receiving postpartum advice to use folic acid; however, Hispanic women were much less likely to use folic acid, compared with white women. This report covers data from May 2000 through November 2001. A study was conducted in Texas to determine whether women at high risk recall and follow recommendations to use folic acid. The study included 195 women at high risk and 223 control mothers who gave birth to infants without birth defects. These women participated in a telephone interview for a population-based case-control study of NTDs. Approximately 56.4% (110 of 195) of mothers who had infants affected by an NTD recalled receiving postpartum advice to use folic acid, compared with 25.6% (57 of 223) of control mothers (p < 0.01). Among nonpregnant case mothers, 54 (32.7%) of 165 reported regular use of supplements containing folic acid, and 53 (25.2%) of 210 nonpregnant control mothers reported this behavior (p = 0.11). Among case mothers, use of folic acid was significantly higher for whites (64.7%) versus Hispanics (16.5%) (p < 0.001); for women with some college education (57.1%) versus no college education (20.2%; p < 0.001); for women who were trying to get pregnant (66.7%) versus those using birth control (38.3%) or reporting using no contraceptive method (18.8%) (p = 0.001); and for women who reported receiving advice to use folic acid (40.9%) versus those who did not (22.2%; p = 0.01). Findings from this study support the need to implement NTD recurrence prevention activities in Texas. Data also identify a need for educational strategies in Texas that target Hispanic women at high risk, especially those who primarily speak Spanish. Further efforts should be made to determine why Hispanic women have low rates of folic acid use (e.g., the cost of vitamins and language and cultural barriers). On the basis of a review of research and current practice, recommendations developed by the Public Health Service include 1) women at risk for a recurrent NTD-affected pregnancy should take 0.4 mg of folic acidper day; and 2) if a woman at high risk is planning a pregnancy, she should consult her physician regarding taking the higher dose of 4.0 mg per day.
In the United States, sampling women of reproductive age from the general population for research purposes is a challenge. Even more difficult is conducting a population-based study of couples attempting pregnancy to assess fecundity and fertility or related impairments. To address the problem of obtaining representative samples from the population in order to study such health-related issues, a commercially and readily available CD-ROM telephone directory was used and tested as a sampling framework for studies aimed at enrolling gravid women aged 18–44 years. A self-administered questionnaire (SAQ) was mailed to a stratified random sample of 10 005 (3%) households in Erie County, NY, USA.
Overall, 17% of the questionnaires were undeliverable despite updating all addresses with residential software before mailing. Thirteen per cent (n = 1089) of the households returned completed questionnaires, of which 35% (n = 377) were completed by women aged 18–44 years. Using 1990 census information for zip code, respondents were more likely to be white and to have higher median household incomes than non-respondents. Of the 377 women who completed the questionnaire, 79% had been pregnant at least once, 5% reported being unable to become pregnant, and 16% reporting never trying to become pregnant. Despite the overall low response to the SAQ, the sampling framework captured a diverse group of women of reproductive age who reported various fecundity and fertility outcomes. The use of low-cost commercially available software linked to census data for selecting samples of women or couples for reproductive and perinatal research may be possible; however, oversampling of households, use of incentives and follow-up of non-respondents is needed to ensure adequate sample sizes.
Cysticercosis is an increasingly important disease in the United States, but information on the occurrence of related deaths is limited. We examined data from California death certificates for the 12-year period 1989-2000. A total of 124 cysticercosis deaths were identified, representing a crude 12-year death rate of 3.9 per million population (95% confidence interval [CI] 3.2 to 4.6). Eighty-two (66%) of the case-patients were male; 42 (34%) were female. The median age at death was 34.5 years (range 7-81 years). Most patients (107, 86.3%) were foreign-born, and 90 (72.6%) had emigrated from Mexico. Seventeen (13.7%) deaths occurred in U.S.-born residents. Cysticercosis death rates were higher in Latino residents of California (13.0/106) than in other racial/ethnic groups (0.4/106), in males (5.2/106) than in females (2.7/106), and in persons >14 years of age (5.0/106). Cysticercosis is a preventable cause of premature death, particularly among young Latino persons in California and may be a more common cause of death in the United States than previously recognized.
In March 2000, a large outbreak of gastroenteritis occurred in a community where a regional computer network provides free Internet access for 42% of the households. We conducted an epidemiologic investigation using the Internet for data collection. Norovirus was identified in stool samples of nine patients but not in the municipal water supply. Of households with access to the network, 19% participated in the survey. The overall attack rate by household was 63%. Drinking water from the nonchlorinated community water system was associated with illness (relative risk [RR] 1.6; 95% confidence interval [CI] 1.1 to 2.2); drinking water only from a private well was associated with decreased likelihood of illness (RR 0.3; 95% CI 0.1 to 0.8). Data collection through the Internet was efficient. Internet surveys may become more common in epidemiologic investigations and have the potential to provide data rapidly, enabling appropriate public health action. However, methods should be developed to increase response rates and minimize bias.
To address a major concern in pharmacoepidemiology studies related to whether the characteristics of responders are different from those who refuse to participate.
We compared utilization of postmenopausal hormone replacement therapy (HRT) in women who agreed to participate in a telephone interview on HRT utilization and in women who refused to participate in the telephone interview. Information on HRT utilization among responders and refusers was independently available to us from a claims database (the Healthcare Management Alternatives, HMA, in Philadelphia), showing drugs dispensed to these patients.
Out of a random sample of 213 women selected from the claims database whom we contacted, 154 (72.3%) women agreed to participate and 59 (32.7%) women refused. Among the 154 women who agreed to participate, 79 (51.3%, 95% CI: 43.1-59.4%) were shown by the database to have been dispensed an HRT during the 15-month period January 1995 through March 1996. Among the 59 women who refused to participate, 30 (50.8%, 95% CI: 37.5-64.1%) were shown by the database to have been dispensed an HRT during the same period.
Thus, we have evidence that use of HRT was almost identical in responders and nonresponders.
Hodgkin lymphoma (HL) risk has been linked with higher social class and lower parity, but our prior population-based case-control study in adult women had unexpected null findings for these variables. Because subject participation was 87% for cases but 65% for random digit-dialing (RDD) controls, we examined representativeness of our controls and the impact of detected bias on prior results.
Using data from RDD enumeration, abbreviated interviews with nonparticipating controls, and the US census, we compared participating and nonparticipating RDD controls across several age groups and then recomputed odds ratios for risk factor associations adjusted for bias.
The 325 RDD control participants were younger, more likely to be white, better educated, and of lower birth order and lower parity than the nonparticipants. Adjustment of odds ratios for bias strengthened previously null findings for education and for parity, breast-feeding and miscarriages in young adult women; these latter changes eliminated previously apparent age modification of risks.
Selection bias in female RDD controls resulted from differential participation by socioeconomic factors, varied with age, and produced underestimations of several associations in young women, including reproductive factors. Thus, our prior conclusions of etiologic irrelevance for some study variables may have been inaccurate.
Use of control (comparison) groups is a powerful research tool. In case-control studies, controls estimate the frequency of an exposure in the population under study. Controls can be taken from known or unknown study populations. A known group consists of a defined population observed over a period, such as passengers on a cruise ship. When the study group is known, a sample of the population can be used as controls. If no population roster exists, then techniques such as random-digit dialling can be used. Sometimes, however, the study group is unknown, for example, motor-vehicle crash victims brought to an emergency department, who may come from far away. In this situation, hospital controls, neighbourhood controls, and friend, associate, or relative controls can be used. In general, one well-selected control group is better than two or more. When the number of cases is small, the ratio of controls to cases can be raised to improve the ability to find important differences. Although no ideal control group exists, readers need to think carefully about how representative the controls are. Poor choice of controls can lead to both wrong results and possible medical harm.
This article compares the recruitment costs and participant characteristics associated with the use of probability and nonprobability sampling strategies in a longitudinal study of older hemodialysis patients and their spouses. Contrasts were made of people who accrued to the study based on probability and nonprobability sampling strategies. Probability-based sampling was more time-efficient and cost-effective than nonprobability sampling. There were no significant differences between the respondents identified through probability and nonprobability sampling on age, gender, years married, education, work status, and professional job status. Respondents from the probability sample were more likely to be Protestant and less likely to be Catholic than those from the nonprobability sample. Respondents from the probability sample were more likely to be Black, whereas those from the nonprobability sample were more likely to be White. There are strengths and shortcomings associated with both nonprobability and probability sampling. Researchers need to consider representativeness and external validity issues when designing sampling and related recruitment plans for health-related research.
Alzheimer disease research has focused on detecting the earliest signs of cognitive decline and efforts are ongoing to develop biomarkers and cognitive measures that reliably distinguish between nondemented and demented individuals. However, little is known about factors that may directly or indirectly influence screening behavior of older community-dwelling adults. We describe an iterative process for the development and formative evaluation of a questionnaire about dementia knowledge and screening behaviors in older adults to understand the psychosocial factors underlying intention to obtain dementia screening to profile individuals manifesting intention to undergo dementia screening compared to those who will not. The Behavioral Model of Health Services Use was used as a conceptual framework for a questionnaire with constructs from the Health Belief Model, Theory of Reasoned Action and Self-Efficacy. After pretesting, we used a random dialing strategy to test our questionnaire on a final sample of 1024 older Missourians. Internal consistency and construct validity were examined. Pretesting identified several potential problems that were improved with rewording. Cronbach alpha was greater than 0.6 (range 0.62 to 0.92) in all but one construct testing dementia knowledge, suggesting good to excellent internal consistency. Convergent (construct) validity was assessed using confirmatory factor analysis. All constructs but 3 demonstrated good validity. Addressing these issues will allow researchers to identify unique characteristics based on age, race, sex, socioeconomic differences and geographic location, and characterize barriers to screening programs to more effectively develop targeted community-based interventions.
Research studies are essential to improving healthcare. However, many fail to recruit their planned number of participants. There are many interventions that researchers try to improve recruitment. Finding which ones are effective would be of benefit to the research community and society.
To quantify the effects of strategies to improve participation in research studies.
We aimed to find all randomized and quasi-randomized controlled trials of strategies to improve recruitment to research studies. We searched nine electronic databases and manually searched the reference lists of relevant trials. We assessed the eligibility of each trial using pre-defined criteria.
Randomized and quasi-randomized controlled trials of methods to increase recruitment in research studies. This includes non-healthcare studies. Studies that required only questionnaire completion were excluded.
We extracted data on the method evaluated, nature of the population, nature of the study to be recruited into, randomisation or quasi-randomisation method, allocation concealment, numbers and proportions in each arm. We used risk ratios and their 95% confidence intervals to describe the effects in individual trials, and assessed heterogeneity of these ratios between trials.
We identified 15 eligible trials, including a total of 33,719 participants. All strategies were aimed at participants for healthcare studies. No strategies were identified at the level of researcher collaborators or ethics committees. Because of heterogeneity between trials and within strategies, the results were not synthesised. Trials of monetary incentives, an additional questionnaire at invitation and treatment information on the consent form demonstrated benefit; these specific interventions from individual trials are not easily generalizable.
On this evidence, it is not possible to predict the effect most interventions will have on recruitment. Funders and researchers should ensure that the evaluation of recruitment strategies are incorporated into research studies.
The results of studies investigating the relationship of smoking with venous thrombosis are inconsistent. Therefore, in the MEGA study, a large population-based case-control study, we evaluated smoking as a risk factor for venous thrombosis and the joint effect with oral contraceptive use and the factor V Leiden mutation. Consecutive patients with a first venous thrombosis were included from six anticoagulation clinics. Partners of patients were asked to participate and additional controls were recruited using a random digit dialing method. Participants completed a standardized questionnaire. Individuals with known malignancies were excluded from the analyses, leaving a total of 3,989 patients and 4,900 controls. Current and former smoking resulted in a moderately increased risk of venous thrombosis (odds ratio (OR)(current) 1.43, 95% confidence interval (CI95) 1.28-1.60, OR(former) 1.23, CI95 1.09-1.38) compared with nonsmoking. Adjustment for fibrinogen levels did not substantially change these risk estimates. A high number of pack-years resulted in the highest risk among young current smokers (OR(>or=20 pack-years) 4.30, CI95 2.59-7.14) compared with young nonsmokers. Women who were current smokers and used oral contraceptives had an 8.8-fold higher risk (OR 8.79, CI95 5.73-13.49) than nonsmoking women who did not use oral contraceptives. Relative to nonsmoking noncarriers, the joint effect of factor V Leiden and current smoking led to a 5.0-fold increased risk; for the prothrombin 20210A mutation this was a 6.0-fold increased risk. In conclusion, smoking appears to be a risk factor for venous thrombosis with the greatest relative effect among young women using oral contraceptives.
The causes of lung cancer among nonsmokers are not clearly understood. To further evaluate the relation between passive smoke exposure and lung cancer in nonsmoking women, we conducted a population-based, case-control study.
Case patients (n = 618), identified through the Missouri Cancer Registry for the period 1986 through 1991, included 432 lifetime nonsmokers and 186 ex-smokers who had stopped at least 15 years before diagnosis or who had smoked for less than 1 pack-year. Control subjects (n = 1402) were selected from driver's license and Medicare files.
No increased risk of lung cancer was associated with childhood passive smoke exposure. Adulthood analyses showed an increased lung cancer risk for lifetime nonsmokers with exposure of more than 40 pack-years from all household members (odds ratio [OR] = 1.3; 95% confidence interval [CI] = 1.0, 1.8) or from spouses only (OR = 1.3; 95% CI = 1.0, 1.7). When the time-weighted product of pack-years and average hours exposed per day was considered, a 30% excess risk was shown at the highest quartile of exposure among lifetime nonsmokers.
Ours and other recent studies suggest a small but consistent increased risk of lung cancer from passive smoking. Comprehensive actions to limit smoking in public places and worksites are well-advised.
Driver's license lists are infrequently used for population-based sampling, presumably because of suspicions of poor population coverage. The 1990 Iowa driver's license list was compared with the 1990 census to evaluate coverage by 5-year age group, sex, resident county, and urbanicity. Coverage exceeded 90% among 15- to 74-year-old men and 15- to 64-year-old women, with uniform coverage by county and county urbanicity group in these age ranges. In Iowa, these lists are convenient and cost-effective and appear to be representative for 25- to 64-year-olds. The representativeness of driver lists in regard to other factors and in other geographic regions deserves further evaluation.
Although random digit dialing (RDD) is an accepted and commonly used method of sampling populations for controls in case-control studies, there have been surprisingly few attempts to compare the accuracy of RDD with that of the best traditional alternative, i.e., household area surveys. The aim of the present study was to compare a variety of characteristics of control subjects selected by ROD and area survey methods. All data were gathered through in-person interviews of both types of control subjects. The area survey identified a population-based sample of 20- to 79-year-old residents of two Washington State counties in 1978 and 1979. Control groups for three case-control studies of bladder cancer, gynecologic cancers, and multiple myeloma were drawn from this area sample, for a total of 240 control subjects. Controls aged 21-64 years from the same counties were identified for the National Bladder Cancer Study using ROD telephone sampling during the same time period. There were 134 respondents in the ADD control group. Overall, the two control groups selected by these two different methods yielded similar estimated frequencies of various population characteristics. The small differences observed in some age/sex subgroups and the statistical significance of the overall measure of association for occupational exposure to organic substances may be attributed to multiple compar isons.
The association between oral contraceptive use and breast cancer risk was examined using data from a case-control study of breast cancer in Long Island, New York. Cases were defined as female residents of Nassau and Suffolk Counties between the ages of 20 and 79, diagnosed with breast cancer between January 1, 1984 and December 31, 1986. Age- and county-matched controls were selected from driver's license files. Among all women under age 70 at diagnosis, there was no association between oral contraceptive use and breast cancer; there was, however, a positive association in the subgroup ages 20-49 (adjusted odds ratio = 1.68, 95% CI: 1.16-2.42). Risk increased with increasing duration of use, but did not differ between women who first used oral contraceptives before the first pregnancy and those who first used them later, or between women who first used oral contraceptives before age 25 and those who first used them at a later age. Risk also appeared to increase with number of years of use before the first pregnancy or before age 25, although numbers were small. History of benign breast disease did not influence risk. The association of breast cancer risk with oral contraceptive use appeared stronger in women from Suffolk County than Nassau County
Background
Sampling frames and mode of contact and administration of questionnaires are important factors contributing to response rates and selection bias in population-based research. The purpose of this paper is to evaluate whether contact by mail before contact by telephone increases response rate, and to assess the concurrent validity of telephone surveys for collecting health research and service data.
Methods
Two thousand households were randomly selected from electronic white pages. Half were randomly allocated to receive or not to receive an explanatory letter before telephone contact. Interviewers were blinded to whether a household received a letter. Respondents aged 18 years or over were randomly selected from within each household using a Kish grid and interviewed by telephone.
Results
The overall response rate was 68 per cent [confidence interval (CI) 66–70]. The response rate of those who received the letter was 76 per cent (CI 73–79), and of those who did not receive the letter was 60 per cent (CI 56–63). Use of the Kish grid to select randomly a respondent decreased the response rate by less than 10 per cent. The internal validity of the data was as follows: in a 10 per cent sub-sample, the Kish grid had been correctly applied in 93 per cent of households, and in 99 per cent of households the exclusion criteria had been correctly adhered to. The external validity was as follows, comparisons with data obtained from the same reference population using similar instruments administered face-to-face revealed no meaningful or significant differences in population estimates.
Conclusions
Mail-out before telephone contact greatly increases response rates at low cost. Telephone surveys can yield valid, useful data for health research and service evaluation.
Young adult blacks and whites aged 18–30 years of both low and high educational levels were recruited through random digit dialing to participate in the Chicago, Illinois, portion of a longitudinal study, Coronary Artery Risk Development in Young Adults (CARDIA). Overall, 31 % of randomly selected persons eligible to participate had unlisted telephone numbers—about 50% of black men and women and 11% and 17% of white men and women, respectively. There was no difference in proportions of numbers unlisted by educational level, except for white men, who were more likely to have unlisted numbers at a low educational level than at a high educational level. There was no consistent pattern of differences in rates of participation across race, sex, or education subgroups for unlisted and listed numbers, and there were no significant differences for selected health measures, except smoking. The findings suggest that in Chicago, there is a potential bias in estimates of sociodemographic characteristics from the exclusion of unlisted numbers, but it is likely to be insignificant if recruitment is stratified according to race, sex, and education. Within strata, there was little bias with respect to the attributes measured. Ideally, to guard against possible bias, random digit dialing is recommended as the preferred way to select a representative population-based sample. Am J Epidemiol 1992;135:697–709.
The purpose of this case-control study was to evaluate potential risk factors for prostatic hypertrophy. The cases were 910 residents of Rhode Island who had a partial or total prostatectomy that was not related to cancer in the years 1985-1987. The controls were 2,003 members of the source population who were selected from a list of holders of Rhode Island driver's licenses or a roster of older Americans compiled by the Health Care Financing Administration. Cases and controls were interviewed by telephone. The risk of prostatic hypertrophy was elevated in Jewish men compared with Protestants and Catholics and in blacks compared with whites. Risk was reduced in ever-married compared with never-married men, in men who had left school at age 16 years or more compared with those who had left earlier, and in relatively tall or relatively heavy men. Coffee drinking and cigarette smoking were inversely but only weakly related to prostatic hypertrophy. There was a relatively strong, although irregular, inverse relation of beer drinking to prostatic hypertrophy. The associations of spirits and wine consumption with prostatic hypertrophy were weak.
Control groups selected by random digit dialing are frequently used in case-control studies. Concern about the potential for bias in these control groups has been expressed, primarily because of low response rates. This study compares the characteristics of a hypothetical control group consisting of 341 men and women aged 40-74 years, selected by random digit dialing and participating in an interview in 1990, with the characteristics of 15,563 men and women aged 40-74 years who participated in a privately conducted census in the same upstate New York county in 1989. For most measures, no differences were seen between the random digit dialing sample and the census population. However, the hypothetical control group was more likely to have had their cholesterol checked in the past 2 years and was somewhat more likely to have had other screening tests as well. In addition, the hypothetical control group was somewhat better educated. The results suggest that, at least in this setting, control groups selected by random digit dialing are representative of the general population in most respects; however, caution should be used when studying the relation between screening tests and disease occurrence by means of case-control studies using controls selected by random digit dialing.
The association between oral contraceptive use and breast cancer risk was examined using data from a case-control study of breast cancer in Long Island, New York. Cases were defined as female residents of Nassau and Suffolk Counties between the ages of 20 and 79, diagnosed with breast cancer between January 1, 1984 and December 31, 1986. Age- and county-matched controls were selected from driver's license files. Among all women under age 70 at diagnosis, there was no association between oral contraceptive use and breast cancer; there was, however, a positive association in the subgroup ages 20-49 (adjusted odds ratio = 1.68, 95% CI: 1.16-2.42). Risk increased with increasing duration of use, but did not differ between women who first used oral contraceptives before the first pregnancy and those who first used them later, or between women who first used oral contraceptives before age 25 and those who first used them at a later age. Risk also appeared to increase with number of years of use before the first pregnancy or before age 25, although numbers were small. History of benign breast disease did not influence risk. The association of breast cancer risk with oral contraceptive use appeared stronger in women from Suffolk County than Nassau County.
A study of health behaviors in four communities in the western United States in 1988 provided the opportunity to compare two methods of sampling elderly respondents for a telephone interview. The Polk telemarketing lists were used to identify 1,407 respondents aged 65 years and older in four communities, where 253 respondents in the same age group were also identified by the method of random digit dialing. Individuals identified from the Polk lists received a letter prior to the initial telephone contact. The overall response rate was 49.3% for random digit dialing and 57.3% for the Polk lists. On the average, the identification of one elderly respondent using the Polk lists required about 20-25% as much interviewer time per subject identified as was required by the random digit dialing method. The elderly identified by the Polk lists were significantly older than those identified by random digit dialing, and the proportions of the Polk sample who were married, white, or had an income of greater than $10,000 were slightly higher than those of the random digit dialing sample. Among 40 variables measuring various health behaviors, indicators of health status, and participation in health-related programs and classes, only three differed significantly between the two samples. The authors conclude that sampling from commercial telemarketing lists was an efficient method of identifying elderly respondents and that in these four communities, the estimates of health behaviors and health status were comparable with those obtained by random digit dialing techniques.
Nasca PC (New York State Department of Health, Bureau of Cancer Epidemiology, Albany, New York 12237, USA), Bap-tiste MS, Field NA, Metzger BB, Black M, Kwon CS and Jacobson H. An epidemiological case-control study of breast cancer and alcohol consumption. International Journal of Epidemiology, 1990; 19: 532–538.
A case-control study of breast cancer and alcohol consumption was conducted with 1617 patients diagnosed with a primary cancer of the breast between 1982 and 1984 in 18 New York State counties. For each case, one control, matched for year of birth and county of residence, was selected from the driver's license files of the New York State Department of Motor Vehicles. Breast cancer risk was shown to increase as daily consumption of alcohol increased, with a risk of 1.37 (95% CI = 1.07, 1.75) observed among women who consumed 15 or more grams of alcohol per day. Breast cancer risk did not appear to be related to the total number of years a woman drank or to be restricted to specific types of alcoholic beverages. The data suggest that this may be higher in women who began drinking at a later age. The increased risk associated with alcohol consumption, observed in the current study, persisted within strata of various breast cancer risk factors.
Our analysis of response rates for community controls approached in a retrospective study of melanoma indicates that there was an 11 per cent improvement in response rate when the contact letter was changed from inviting participation in a "healthy study" with no mention of cancer to a format which explicitly explained that the study concerned cancer. This improvement was observed for all age-sex groups except younger males. The result suggest that higher response rates may be achieved in studies of cancer if fuller disclosure of the study purposes is made.
Response rates are an important component of epidemiologic research. The purposes of this study are (a) to evaluate how response rates are defined and calculated for control subjects in epidemiologic case-control studies, and (b) to explore factors that may impact response in epidemiologic studies. Our results show that the method of control subject selection has an impact on study response. Gender of respondent does not appear to impact response rates. However, response rates are generally worse for individuals less than 45 years old. Methods used to calculate response have a great impact on "response rate"; therefore, it is important for researchers to define exactly what the reported response rates represent and how they are derived so that data can be interpreted appropriately.
The epidemiology of large-bowel cancer suggests a role for endocrine factors in its development. Although analytic studies have not consistently provided evidence for an association between reproductive history and large-bowel cancer, some relatively small studies have observed a reduced risk among women using postmenopausal hormone replacement therapy (HRT).
This study was planned to evaluate more precisely the relationship between HRT and the risk of colon and rectal cancers.
Female residents of Wisconsin aged 30-74 years with a diagnosis of colon or rectal cancer within 2 years were identified through a statewide tumor registry. Control subjects were randomly selected from lists of licensed drivers if the case subjects were less than 65 years old and from lists of Medicare beneficiaries if they were 65-74 years old. Information on post-menopausal hormone replacement use, medical history, and family history was obtained in telephone interviews. After premenopausal women were excluded, 694 case subjects and 1622 control subjects remained for analysis. The odds ratios and 95% confidence intervals (CIs) obtained from conditional logistic regression models were used to estimate relative risks (RRs). All RRs were adjusted for age, family history of large-bowel cancer, use of screening sigmoidoscopy, and recent alcohol consumption.
Compared with postmenopausal women who never used HRT, recent users had an RR of 0.54 (95% CI = 0.36-0.81) for colon cancer and an RR of 0.91 (95% CI = 0.54-1.55) for rectal cancer. This inverse association was observed among users of both estrogen only and combined estrogen and progestin preparations. Decreasing time since last use was inversely associated with colon cancer risk (P for trend < .001). The effect of HRT appeared to be stronger among women at lower absolute risk of colon cancer, particularly among women with lean body mass.
Use of HRT was associated with a statistically significant reduced risk of colon cancer. In contrast, no statistically significant relationship was observed for rectal cancer. Given the widespread use of postmenopausal hormones and the morbidity and mortality from adenocarcinoma of the bowel in women, these findings may have potentially important public health implications.
We compared telephone and household surveys as methods for obtaining information about breast cancer screening. The study population was comprised of women age 40 years or older who were permanent residents of two large areas of Florida. We contacted women using random digit dialing for the telephone survey. We used a stratified multistage design for the household survey. Response rates were 49% (telephone) and 77% (household). Distributions of most screening (mammography, clinical examination, and breast self-examination) and demographic variables were comparable for the two surveys; income was higher in the telephone survey. Both surveys underestimated the proportion of older women, based on 1990 Census data.
Insufficient response rates are a frequent problem in mailed epidemiologic health surveys. As part of a health survey of 1,235 young adults conducted from November 1992 through January 1993 in Geneva, Switzerland, a randomized factorial trial was carried out to assess the contributions of two tactics to increase response rates. The first incentive was the promise of sending 10 Swiss francs (7 US dollars) to respondents; the second was a red postcard, mailed 2 days after the questionnaire, which reminded potential participants to complete the questionnaire. The most effective strategy according to interim analysis was to be selected for follow-up mailings. Two weeks after the first mailing, response rates were 65% for those who had received both incentives, 57% for those who had been offered the money reward only, 54% for recipients of the reminder card, and 48% for those who had received neither incentive (p based on chi 2 (3 df) test < 0.001). The relative hazards of responding were 1.34 (95% confidence interval 1.15-1.55) for the money reward and 1.22 (95% confidence interval 1.05-1.41) for the reminder card. Follow-up mailings to all nonrespondents included both incentives. At the end of data collection, 1,007 persons (82%) had returned the questionnaire. The final response rates were 83, 84, 82, and 78% in the four groups, respectively (p = 0.29). Both the promise of a monetary reward and a reminder postcard considerably improved early response rates. The use of both incentives in follow-up mailings brought final response rates above 80%, except for those who had received no initial incentive. Testing various combinations of incentives in the early stages of a mailed health survey may be a generally useful strategy.
The purpose of this study was to investigate how response rates to a postal questionnaire are affected by title and length of the survey instrument. Five questionnaires, which differed according to title and length, were designed. Each questionnaire was mailed to a random sample of one thousand Norwegian women aged, 35-49 years. A total of 3,106 questionnaires were returned (62.1%). The highest response rate (70.2%) was achieved by a two-page questionnaire entitled "Women and Cancer". An otherwise identical questionnaire entitled "Oral Contraceptives and Cancer" had a response rate of 60.7%. Questionnaires entitled "Women and Cancer" with a length of four and six pages had a response rate of 62.8% and 63.3%, respectively. The four page questionnaire entitled "Women, Lifestyle and Health" had the lowest response rate of 57.1%. This study shows that in a general population of Norwegian women the title of a postal questionnaire influences the response rate. The results indicate that although the shortest questionnaire had the highest response rate, the most extensive survey instrument did not have the lowest response rate. The distribution of risk factors for breast cancer did not vary according to response rate or design of questionnaire. The overall findings of this study suggest that the benefits from the increased information obtained from extensive postal questionnaires out-weighs a potential non-response bias due to a somewhat lower response rate.
Social marketers use the tools of selling to promote good nutrition, regular checkups, and other positive behavior. Instead of preaching, they try to understand and change the complex motivations that lie behind risky activities like smoking and unsafe sex. AIDS prevention is the vanguard of social marketing in the U.S., but the practice is likely to spread for a simple reason: it works.
This report describes Medicare eligibility requirements; the processes to establish Medicare entitlement; types of coverage; the composition of the enrolled population; and outlines some differences in measurement techniques used in a decennial census in contrast to Medicare enrollment.
Current Medicare enrollment figures do not represent a complete count of any segment of the total United States population. Some persons age 65 and over are not eligible for Medicare; others are eligible but not entitled. However, the Medicare enrollment figures are frequently used as surrogate counts of the aged population because they provide excellent sources of detailed demographic and geographic information for a large proportion of those age 65 and over. The data are produced semi-annually, based on daily updates.
The aim of this study was to estimate the association between diabetes and its complications and at-fault automobile crashes among older drivers.
This was a population-based case-control study. Case subjects were drivers aged > or = 65 years who had been involved in a crash during 1996 in which they were at fault. Two control groups were selected: 1) crash-involved not-at-fault subjects and 2) non-crash-involved subjects. Telephone interviewers collected information on demographic characteristics, driving habits, diabetes sequelae and treatment, other chronic medical conditions, and visual function.
Overall, there was no association between diabetes and at-fault crash involvement. The adjusted odds ratio (OR) for diabetes was 1.1 (CI 0.7-1.9) when case subjects were compared with either control group. However, the adjusted OR for diabetes was 2.5 (0.9-7.2) among subjects who had been involved in a crash in the 4 years preceding 1996, while it was only 0.9 (0.5-1.7) among those who had not. There was no evidence of an association between treatment modalities and at-fault crash involvement. Case subjects were, although not significantly (P = 0.25), more likely (OR 2.4) to report neuropathy compared with both control groups, and retinopathy was not associated with increased crash risk.
This study provides no evidence that older drivers with diabetes are at increased risk for automobile crashes. There remains the possibility that those with diabetes who have more severe disease or have had multiple crashes are at increased risk.
If you want to know what factors influence response rate, why not ask?
- Myers Be Leitner Dw
- K Sardeson
- Keller
- Pfg
- Chang
- Tj
Leitner DW, Myers BE, Sardeson K, Keller PFG, Chang TJ. If you want to know what factors influence response rate, why not ask?
Health diaries: mun Health
- Norman Gr Mcfarlane Ah
- Streiner
- Dl
- Neale
Norman GR, McFarlane AH, Streiner DL, Neale K. Health diaries: mun Health. 1990;44:316–320.
Effect of contact letter on control CA Recruiting survey respondents to mailed surveys: Controlled trials response rates in cancer studies
- Walter
- Sd
- Marrett Ld
Walter SD, Marrett LD, Mishkel N. Effect of contact letter on control CA. Recruiting survey respondents to mailed surveys: Controlled trials response rates in cancer studies. Am J Epidemiol. 1988;127:691–694.
Chance of free dinner increases Proceedings of the Survey Research Methods Section Assoc; response to mail questionnaire
- Douglas A B Woodward
- Miles
Woodward A, Douglas B, Miles H. Chance of free dinner increases Proceedings of the Survey Research Methods Section. Am Stat Assoc; response to mail questionnaire. Intl J Epidemiol. 1985;14:641–642. 1979:359–362.
Timing payments to subjects of mail surveys: tudes of participants in a case-control study. J Epidemiol Commun Cost-effectiveness and bias
- M Schweitzer
- Asch
- Da
Schweitzer M, Asch DA. Timing payments to subjects of mail surveys: tudes of participants in a case-control study. J Epidemiol Commun Cost-effectiveness and bias. J Clin Epidemiol. 1995;48:1325–1329.
Random digit dialing in Chicago CARDIA
- Orden
Sampling elderly in the community
- Psaty
Diabetes and automobile crashes in the elderly
- McGwin