Reducing attrition in panel studies in developing countries.

Zelee Hill

doi:10.1093/ije/dyh060

Home/Archive/PubMed

PubMed · 15131086

Reducing attrition in panel studies in developing countries.

Zelee Hill

View original source ↗Export metadata ↓

Abstract

BACKGROUND: Panel studies offer repeated observations of individuals over time, but the mobility of populations in the developing world often causes attrition in panel studies. Such attrition can cause bias if it is selective but can be reduced by tracking respondents. Tracking in developing countries can be costly and difficult as populations are often highly mobile, infrastructure is poor, structures frequently change, and formal address systems or population records rarely exist. Method In this paper, the attrition and tracking experiences of panel studies in developing countries are reviewed and recommendations made for ensuring effective tracking. Comments Tracking can reduce attrition by up to 45% and is feasible if procedures are locally appropriate, well planned, involve the community, collect as much locating data as possible, and have explicit criteria, and if tracking is done at regular intervals, and interviewers are well trained, supervised, and motivated. CONCLUSION: Attrition is an important issue in panel studies, whilst tracking can be costly it can reduce attrition if effective procedures are used.

Explore related subjects

Bias Data Collection Developing Countries Epidemiologic Research Design Humans Interviews as Topic Longitudinal Studies Patient Dropouts Population Dynamics Population Surveillance Journal Article Research Support, Non-U.S. Gov't Review

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Zelee Hill. 2004-05-06. Reducing attrition in panel studies in developing countries.. https://doi.org/10.1093/ije%2Fdyh060

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related citations

PubMed2026

Assessment of blinding in pharmacotherapy and noninvasive neuromodulation randomized controlled trials for neuropathic pain in adults.

In randomized controlled trials (RCTs), study participants and research personnel are often blinded to minimize biases related to knowing treatment allocation. To determine if blinding was effective, participants may be asked which treatment they believe they received ("treatment guess"). This descriptive review characterized blinding assessment (BA) reporting in pharmacotherapy and neuromodulation neuropathic pain RCTs. Of 288 papers, 36 (12.5%) reported a BA. One paper reported the results of 2 studies, so in total 37 studies with a BA were assessed. Of these, 19 were crossover, 17 parallel, and 1 partial crossover in design. All 37 studies assessed participant blinding, and 10 also assessed investigator blinding. Approximately 27% included an "unsure" answer option for treatment guess, and 38% asked the reason for the guess. There were no clear patterns in BA reporting across time nor based on treatment type. Seventeen trials provided sufficient data to calculate Bang Blinding Index (BI) to determine blinding success. Participants remained blinded (BI = 0 ± 0.2) in 10/17 placebo and 10/17 treatment arms, 6 placebo and 5 treatment arms had a BI > 0.2 suggesting possible unblinding, whereas 1 placebo and 2 treatment arms had a BI < -0.2 suggesting misinformed guessing. Overall, we found that BAs are done in a minority of published neuropathic pain trials and with variable methodology. Given the importance of minimizing risk of bias because of treatment unblinding, future studies should consider including BAs, and further consensus building is necessary to determine if and how BAs should be conducted and interpreted in analgesic clinical trials.

Bias↗

PubMed2006

Multiple imputation for correcting verification bias.

In the case in which all subjects are screened using a common test and only a subset of these subjects are tested using a golden standard test, it is well documented that there is a risk for bias, called verification bias. When the test has only two levels (e.g. positive and negative) and we are trying to estimate the sensitivity and specificity of the test, we are actually constructing a confidence interval for a binomial proportion. Since it is well documented that this estimation is not trivial even with complete data, we adopt multiple imputation framework for verification bias problem. We propose several imputation procedures for this problem and compare different methods of estimation. We show that our imputation methods are better than the existing methods with regard to nominal coverage and confidence interval length.

Bias↗

PubMed2006

Rater bias in a blinded randomized placebo-controlled psychiatry trial.

Rater bias occurs when rater knowledge of treatment assignment modifies the outcome assessment. Raters may be unconsciously influenced by inclinations for or against a particular treatment and consequently may give a more or less generous assessment depending upon these biases. Blinding of raters by keeping raters unaware of treatment assignment is one way to limit bias influencing assessment due to knowledge of treatment assignment. Unblinding may be particularly problematic in efficacy studies comparing placebo to drugs and/or non-drug psychotherapy treatments where subjects may reveal drug side-effects or mention their therapist by name, thus unblinding their treatment assignment. We present a new instrumental variable statistical approach for assessing the association between success in blinding and its impact on efficacy estimates of active drug and/or cognitive behavioural psychotherapy versus placebo in the multicentre comparative treatment study of panic disorder. Despite the uncertainty involved in assessing bias that may be unobserved and unconscious, we will show how to derive a bound for the impact of rater bias.

Bias↗