Understanding non-performance reports for instrumental activity of daily living items in population analyses: a cross sectional study

BMC Geriatrics, 2016 · DOI: 10.1186/s12877-016-0235-0 · Published: February 26, 2016

Aging Rehabilitation Research Methodology & Design

Simple Explanation

This study looks at how to deal with survey answers about IADLs when people say they don't do certain activities for reasons other than health. It explores different ways to classify these responses and sees how each method affects estimates of disability prevalence. The researchers tested four different strategies: imputing the missing data, excluding the data, classifying the responses as 'no difficulty,' or classifying them as 'difficulty.' They then compared the results to see if the different methods significantly changed the overall disability estimates. The study found that while there were some small differences, the overall impact on population-level disability estimates was minimal. This suggests that IADL surveys can still be useful for assessing activity limitations, even when many people don't perform certain tasks for non-health reasons.

Study Duration

Not specified

Participants

13,879 non-institutionalized adult Medicare beneficiaries

Evidence Level

Cross-sectional study

Key Findings

1
IADL stage prevalence estimates differed little across the four alternative strategies.
2
Classification as “no difficulty” led to slightly lower, while classification as “difficulty” raised the estimated population prevalence of disability.
3
Results from the high prevalence staging assignment strategy are the most different from the other three methods.

Research Summary

This study evaluated different strategies for classifying “does not do” responses to IADL questions when estimating prevalence of IADL limitations in a national survey. IADL stage prevalence estimates differed little across the four alternative strategies, including multiple imputation, exclusion, classifying as “no difficulty,” or classifying as “difficulty.” The study concludes that IADL items can be used to estimate the prevalence of activity limitation despite high rates of non-performance, with only trivial differences resulting from alternative handling of unrated IADL information.

Practical Implications

Clinical Practice

Clinicians should be aware of potential small differences when using IADL survey data due to different handling of unrated information.

Research

Researchers can be reassured that IADL items are useful for estimating activity limitations despite non-performance, especially when using MCBS data.

Policy

Policy end-users can legitimately use complete case analysis in assigning stages when determining population prevalence estimates of IADL limitations.

Study Limitations

1
Assumptions of Missing at Random on which multiple imputation is based cannot be tested.
2
Lack of a gold standard IADL measure makes it difficult to evaluate which method is further away from the truth.
3
Findings are relevant primarily to the IADL questions as worded in surveys using wording similar to the MCBS.

Your Feedback

Was this summary helpful?

Back to Aging