Dataset Information

Choosing a design to fit the situation: how to improve specificity and positive predictive values using Bayesian lot quality assurance sampling.

ABSTRACT:

Background

Lot Quality Assurance Sampling (LQAS) is a provably useful tool for monitoring health programmes. Although LQAS ensures acceptable Producer and Consumer risks, the literature alleges that the method suffers from poor specificity and positive predictive values (PPVs). We suggest that poor LQAS performance is due, in part, to variation in the true underlying distribution. However, until now the role of the underlying distribution in expected performance has not been adequately examined.

Methods

We present Bayesian-LQAS (B-LQAS), an approach to incorporating prior information into the choice of the LQAS sample size and decision rule, and explore its properties through a numerical study. Additionally, we analyse vaccination coverage data from UNICEF's State of the World's Children in 1968-1989 and 2008 to exemplify the performance of LQAS and B-LQAS.

Results

Results of our numerical study show that the choice of LQAS sample size and decision rule is sensitive to the distribution of prior information, as well as to individual beliefs about the importance of correct classification. Application of the B-LQAS approach to the UNICEF data improves specificity and PPV in both time periods (1968-1989 and 2008) with minimal reductions in sensitivity and negative predictive value.

Conclusions

LQAS is shown to be a robust tool that is not necessarily prone to poor specificity and PPV as previously alleged. In situations where prior or historical data are available, B-LQAS can lead to improvements in expected performance.

SUBMITTER: Olives C

PROVIDER: S-EPMC3600627 | biostudies-literature | 2013 Feb

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Choosing a design to fit the situation: how to improve specificity and positive predictive values using Bayesian lot quality assurance sampling.

Olives Casey C Pagano Marcello M

International journal of epidemiology 20130201 1

<h4>Background</h4>Lot Quality Assurance Sampling (LQAS) is a provably useful tool for monitoring health programmes. Although LQAS ensures acceptable Producer and Consumer risks, the literature alleges that the method suffers from poor specificity and positive predictive values (PPVs). We suggest that poor LQAS performance is due, in part, to variation in the true underlying distribution. However, until now the role of the underlying distribution in expected performance has not been adequately e ...[more]

PMID: 23378151

Similar Datasets

Project description:BackgroundCardiovascular diseases (CVDs) are the leading cause of mortality in India. India has rolled out Comprehensive Primary Health Care (CPHC) reforms including population based screening for hypertension and diabetes, facilitated by frontline health workers. Our study assessed blood pressure and blood sugar coverage achieved by frontline workers using Lot Quality Assurance Sampling (LQAS).MethodsLQAS Supervision Areas were defined as catchments covered by frontline workers in primary health centres in two districts each of Uttar Pradesh and Delhi. In each Area, 19 households for each of four sampling universes (males, females, Above Poverty Line (APL) and Below Poverty Line (BPL)) were visited using probability proportional to size sampling. Following written informed consent procedures, a short questionnaire was administered to individuals aged 30 or older using tablets related to screening for diabetes and hypertension. Using the LQAS hand tally method, coverage across Supervision Areas was determined.ResultsA sample of 2052 individuals was surveyed, median ages ranging from 42 to 45 years. Caste affiliation, education levels, and occupation varied by location; the sample was largely married and Hindu. Awareness of and interaction with frontline health workers was reported in Uttar Pradesh and mixed in Delhi. Greater coverage of CVD risk factor screening (especially blood pressure) was seen among females, as compared to males. No clear pattern of inequality was seen by poverty status; some SAs did not have adequate BPL samples. Overall, blood pressure and blood sugar screening coverage by frontline health workers fell short of targeted coverage levels at the aggregate level, but in all sites, at least one area was crossing this threshold level.ConclusionCVD screening coverage levels at this early stage are low. More emphasis may be needed on reaching males. Sex and poverty related inequalities must be addressed by more closely studying the local context and models of service delivery where the threshold of screening is being met. LQAS is a pragmatic method for measuring program inequalities, in resource-constrained settings, although possibly not for spatially segregated population sub-groups.

Project description:India has experienced a significant increase in facility-based delivery (FBD) coverage and reduction in maternal mortality. Nevertheless, India continues to have high levels of maternal health inequity. Improving equity requires data collection methods that can produce a better contextual understanding of how vulnerable populations access and interact with the health care system at a local level. While large population-level surveys are valuable, they are resource intensive and often lack the contextual specificity and timeliness to be useful for local health programming. Qualitative methods can be resource intensive and may lack generalizability. We describe an innovative mixed-methods application of Large Country-Lot Quality Assurance Sampling (LC-LQAS) that provides local coverage data and qualitative insights for both FBD and antenatal care (ANC) in a low-cost and timely manner that is useful for health care providers working in specific contexts. LC-LQAS is a version of LQAS that combines LQAS for local level classification with multistage cluster sampling to obtain precise regional or national coverage estimates. We integrated qualitative questions to uncover mothers' experiences accessing maternal health care in the rural district of Sri Ganganagar, Rajasthan, India. We interviewed 313 recently delivered, low-income women in 18 subdistricts. All respondents participated in both qualitative and quantitative components. All subdistricts were classified as having high FBD coverage with the upper threshold set at 85%, suggesting that improved coverage has extended to vulnerable women. However, only two subdistricts were classified as high ANC coverage with the upper threshold set at 40%. Qualitative data revealed a severe lack of agency among respondents and that household norms of care seeking influenced uptake of ANC and FBD. We additionally report on implementation outcomes (acceptability, feasibility, appropriateness, effectiveness, fidelity, and cost) and how study results informed the programs of a local health non-profit.

Dataset Information

Choosing a design to fit the situation: how to improve specificity and positive predictive values using Bayesian lot quality assurance sampling.

Background

Methods

Results

Conclusions

Publications

Choosing a design to fit the situation: how to improve specificity and positive predictive values using Bayesian lot quality assurance sampling.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets