The Lawton and Brody a Half Century Later: A Case for Contemporary Revision

Structured Abstract

Objective:

To re-evaluate the psychometric properties and contemporary relevance of the 1969 Lawton and Brody Instrumental Activities of Daily Living (IADL) Scale in older adults.

Methods:

Data from 3,552 participants and co-participants in the Texas Alzheimer’s Research and Care Consortium (TARCC) cohort were analyzed. Lawton and Brody iADL Scale items were analyzed in a sample with no cognitive disorder, mild cognitive impairment (MCI), and dementia. Graded response models assessed item discrimination and difficulty, with differential item functioning (DIF) analyses examining demographic effects. Expert reviewers evaluated item content for cultural and technological relevance.

Results:

Items assessing shopping, food preparation, and transportation were most sensitive to mild functional impairment, while the total scale had limited detection of subtle functional decline. Statistical DIF emerged on 75% of items, whereas clinically meaningful DIF for several demographic groupings emerged on the transportation item. Expert reviewers identified outdated content, particularly for telephone and financial items, reflecting technological shifts. Broader changes in normative behaviors and sample demographics may further challenge scale applicability.

Conclusions:

The Lawton and Brody IADL Scale remains foundational but shows limited sensitivity for early functional decline. Outdated items and evolving socioeconomic and technological contexts underscore the need for refinement to enhance unbiased early detection and maintain longitudinal comparability. Updating the instrument to reflect modern demands could strengthen functional assessment in ADRD research and clinical trials.

Summary

The daily life of an older adult has fundamentally changed in the half century since Lawton and Brody published their scale. Cultural norms around daily activities have shifted, digital technologies have transformed how people shop, manage finances, communicate, and navigate their environments, and our clinical understanding of cognitive decline has expanded to encompass stages of impairment that were not recognized when the instrument was created. What has not changed is the central goal the Lawton and Brody iADL scale pursued: to understand, and ultimately preserve daily functioning and independence, the thing that matters most to older adults and their families. Our findings highlight both the limitations of legacy measures in capturing these changes in modern contexts, but also highlights the opportunity to leverage advances in psychometric methods to build instruments that evolve alongside the people and contexts they are designed to serve.

Background

Measurement of instrumental activities of daily living (iADLs) remains a cornerstone of clinical care and research in Alzheimer’s disease and related disorders (ADRD). Functional status, as indexed by iADLs, is incorporated in current ADRD diagnostic and staging frameworks (Albert et al., 2011; Jack Jr. et al., 2024; McKhann et al., 2011), reflects real-world outcomes that matter to patients and caregivers (Reilly & Harding, 2025; Tochel et al., 2019), and informs regulatory decisions regarding the approval of new treatments (Edgar et al., 2019).

Among many approaches developed to assess iADLs, caregiver questionnaires such as the Lawton and Brody scale (Lawton & Brody, 1969) are widely used. The scale alone has over 23,000 citations at the time of this writing in 2026. However, more than five decades after its original publication, there are three broad reasons to re-evaluate the Lawton and Brody iADL scale in light of significant shifts in the ADRD landscape.

First, with the advent of disease modifying treatments, the field has increasingly focused clinical and research attention on earlier symptomatic stages of cognitive decline. For example, in recent clinical staging criteria (Jack et al, 2024), transitional cognitive decline and mild cognitive impairment (MCI) are explicitly described as conditions with minimal and early functional impact, respectively, despite relative preservation of independence. To date, few psychometric studies have examined how well the Lawton and Brody iADL scale captures such subtle functional changes before frank dependence in iADLS occurs. Further, it is important to note that these diagnostic entities were not recognized concepts at the time of the scale’s creation.

Existing research suggests such limitations of the scale in earlier disease stages. For example, the scale as a whole neither differentiated among MCI subtypes (diagnosed on the basis of clinical judgment and cognitive screening), nor distinguished those with MCI from those who were cognitively unimpaired (Burton et al., 2009). At the item level, in a study of 202 individuals with dementia or a related condition diagnosed by clinical examination, McGrory and colleagues (2014) found that the shopping item was best at identifying subtler functional changes, with other items being endorsed only in the presence of more significant functional impairment. However, relatively little is known at the scale or item level about how the Lawton and Brody iADL scale performs in the mildest stages of ADRD functional impact.

Second, broad changes in demographics as well as cultural and societal norms have occurred in the United States (US) in the past 50+ years. Along with these shifts have come reconsideration of optimum approaches to neuropsychological assessment in heterogenous populations (Fujii, 2023). To highlight one example of how shifts in societal norms impacts assessment, consider Lawton and Brody’s approach to evaluating the impact of gender on iADL performance: “For women, the maintenance of earlier life levels of adequacy in such tasks as shopping, cooking, and manner of doing laundry may be the best means of assessing general competence. While the list of such representative activities is smaller for men, one can still differentiate their performance of tasks such as the use of transportation, or handling money, as the basis of measuring competence” (Lawton & Brody, 1969, p.180). Indeed, their item level data from the development sample supported these gender-linked differences in item performance at the time of the scale’s creation.

However, as explored more fully in an earlier publication (González & Benge, 2024), assessment of iADLs is likely representative of a complex interaction between a person’s actual ability, the need/expectation to perform a given task, and experience with a given activity. These latter factors are in turn influenced by changing societal, cultural, and economic forces. For example, compared to 1969, there has been a marked increase in workforce participation among women (Albanesi, 2023), which necessitates reconsideration of household activities as the primary markers of functional ability for women. In addition, the frequency with which men cook has been on the rise over the same time period (Taillie, 2018), suggesting that these items might have become useful indicators of functional ability for some men. These differences in expected daily activities amongst men and women may also differ by cultural context (Brini et al., 2022).

From a psychometric standpoint, differences in how items perform in subgroups of the population are typically explored through differential item functioning (DIF) analyses (Penfield & Camilli, 2006). Two studies have evaluated DIF by gender and age groups in the US on measures similar to the Lawton and Brody iADL scale (Fleishman et al., 2002; LaPlante, 2010). These studies suggested that while DIF exists for comparable iADL items, it was of minimal practical impact. Two additional studies investigated DIF on iADL measures in other cultures. In contrast to findings in the US, DIF was more notable in a Korean sample when evaluated by gender, age, and education (Jang & Kawachi, 2019). Further, a Norwegian sample employing similar items revealed noteworthy DIF for living situation (Lutomski et al., 2016). Thus, DIF analyses on iADL measures exploring educational, language, and ethnoracial groups appears to be lacking.

A third major change since the scale’s initial development has been rapid technological advancements that transform how iADLs are performed. Technologies such as the internet and smartphones now broadly influence daily activities such as shopping, managing finances, and communicating with others. While a digital divide for older adults exists, it is closing (Charness & Boot, 2022; Huxhold et al., 2020). Indeed, older adults now frequently report using digital approaches for a broad range of daily tasks, most notably for shopping and financial management (Benge et al., 2024; Mis et al., 2025). Though using technologies may be harder for those impacted by ADRD, digital daily activities remain applicable to over 80% of those presenting for ADRD evaluations in a memory disorders clinic (Benge, Ali, et al., 2024; Benge et al., 2020).

This technological shift could impact iADL measures such as the Lawton and Brody in numerous ways. Item content may be difficult for patients or clinicians to interpret in today’s context. For example, the legacy scale refers to looking up and dialing phone numbers, an activity that likely has different connotations and even cognitive underpinnings in an age of paper phone books versus online searches. Does difficulty with this task reflect the same thing in 2026 as it did in 1966? Relatedly, balancing checkbooks or writing checks may have reduced in importance to financial management in an age of online accounts (Greene et al., 2020). Does “dependence” for check writing carry the same meaning for functional independence as someone who relies on automatic bill pay to manage finances? Further, new digital activities have also come online in recent years. For example, searching the internet for information, texting/emailing someone, and accessing health records through portals were not considerations for daily tasks in the 1960s, but are now commonplace and increasingly critical tasks in a post-pandemic age (Benge et al, 2024).

In light of shifts in the ADRD clinical enterprise, changing socioeconomic impacts on daily activities, and technological approaches to daily tasks, re-evaluating the Lawton and Brody iADL scale is warranted. Re-evaluating and, based on that information, refining, instruments is a compelling alternative over using one of the many psychometrically strong new iADL measures for three broad reasons (Schalet et al., 2021). First, the original scale and its variants continue to be widely used in research and clinical practice (Graf, 2008). In the setting of increasing efforts to leverage and harmonize large existing datasets to increase data discovery in ADRD and neuropsychology more broadly (Tate et al., 2023), developing tools that build upon and can draw psychometrically comparable inferences across cohorts has a distinct advantage over scale replacement. Such approaches are already widely used to harmonize cognitive scales which evolve and change over time and across study sites (Monsell et al., 2016), but are less commonly done with iADL measures. Advanced techniques such as linking and equating item pools may allow iADL tools to add and augment domains covered (González et al., 2025) and pursue such an approach. Second, there is growing recognition that shifts in society, environment, and diagnostic needs will continue to occur over time, further highlighting the value of evolving measures over time without losing existing data. Understanding, refining, and harmonizing existing measures is in line with modern measurement psychometric approaches to patient reported outcomes (see Schalet et al., 2021 as well as the broader PROSetta Stone project www.prosettastone.org). Third, there are pragmatic considerations for building and revising existing measures, such as regulatory familiarity, the embedded nature of instrumentation within health record workflows, and cross-national comparisons with legacy measures.

In light of these shifts in the ADRD landscape and the potentially utility for revising legacy measures, in this manuscript we seek to address the following aims:

1. What items from the Lawton and Brody iADL scale are most informative of minimal/early functional changes in ADRD? To this end, we use an item response theory (IRT) approach, seeking to evaluate response patterns and item information at earlier functional changes in the disease, with consideration for learning how and what daily activities may inform the next generation of even milder/presymptomatic presentations of the disease.

2. How does the instrument perform across heterogenous demographically-defined groups? We utilize DIF to determine if the items differentially measure functional ability according to the following dichotomized groupings: language (English vs. non-English and predominantly Spanish), ethnoracial group (non-Hispanic White vs. other ethnoracial identification, including Hispanic identifying individuals of any race), biological sex, and educational background (≤12 years vs.>12 years).

3. Critically review item content using an expert panel to identify items that may be missing, evaluate item clarity and relevance, and suggest potential modifications.

Our goals are to identify current strengths and weaknesses of this seminal instrument and to identify directions for its evolution for the 21^st century and beyond.

Methods

Participants

Data for this study was derived from the Texas Alzheimer’s Research and Care Consortium (TARCC) database. This multi-site, collaborative research project involved recruiting older adults with or without known neurodegenerative disease for annual comprehensive assessments at seven Texas-based medical schools from 2007–2018. Additional information about the broader TARCC project and data request information can be found at https://www.txalzresearch.org/. To maximize the spectrum of functional abilities captured for the current analysis, we selected data from the most recent study visits for all participants where a Lawton and Brody iADL measure was completed by a co-participant (i.e., collateral). Instruments were administered by trained clinical or technical staff at each site. All participants gave written informed consent for participation in the study at their host institution.

Instruments

Lawton and Brody iADL scale modifications:

Co-participants completed the Lawton and Brody iADL measure (Lawton & Brody, 1969; Vergara et al., 2012) with slight modifications made by the original TARCC study team (not the authors of this paper) based upon their experience and translations of the instrument agreed upon across sites (see final instrument utilized at https://www.txalzresearch.org/wp-content/uploads/2019/02/tarcc-6.0-forms-annotated.pdf). These modifications included the co-participant being given the option of endorsing that the item was not applicable or had never been done by the participant for items concerning food preparation, housekeeping, laundry, finances, and medications. Second, wording changes in the response options relative to the published 1969 version included:

Ability to Use Telephone:

The 1969 version’s most impaired item was “Does not use telephone at all.” The TARCC version expanded this item to five response options which included “Does not use telephone at all under own initiative, but may talk if put on line,” and a new, more severe option was added: “Incapable of using telephone.”

Shopping:

The second response option was changed from “Shops independently for small purchases” to “Shops independently for a limited number of purchases (3 or less).”

Food Preparation:

The second response option added to “if supplied with ingredients” the phrase ”or given supervision or reminding.”

Housekeeping:

In the first response option, the parenthetical example “(e.g., ‘heavy work-domestic help’)” was dropped from the original. The most severe option was changed from “Does not participate in any housekeeping tasks” to “Is unable to participate in any housekeeping tasks”.

Laundry:

Minor wording changes were made to the second response option (e.g., “stockings” was replaced with “socks or some items”).

Mode of Transportation:

The 1969 version contained five response options including “Does not travel at all” as the most severe anchor. The TARCC version reduced this to four options by removing “Does not travel at all.” In addition, the second response option was revised from “Arranges own travel via taxi, but does not otherwise use public transportation” to “Arranges own travel via taxi/bus, but does not drive own car”.

Ability to Handle Finances:

The phrase “balances checkbook” was added to the parenthetical list in the first (independent) response option.

Responsibility for Own Medications:

The second response option was expanded from “Takes responsibility if medication is prepared in advance in separate dosages” to include “or if reminded.”

Cognitive Staging:

While a variety of neuropsychological and clinical evaluations were included in the TARCC database, we utilized Clinical Dementia Rating (Morris, n.d.) global score to categorize participants into no cognitive disorder (CDR= 0), mild cognitive impairment (CDR =.5), and dementia (CDR >=1) stages. It is possible that the same personnel who administered the CDR may have administered the iADL measure and the frequency with which this occurred was not tracked in this data set; however, the CDR was scored independently and is used for staging purposes.

Statistical Analysis and Reporting

Item level distributions were summarized overall and by CDR group; stage effects were evaluated with one-way MANOVA. Group contrasts were made to focus on the initial shift between normal cognition and MCI.

Unidimensionality was tested with the lavaan package (version 0.6–21) using diagonally weighted least squares estimation for ordinal indicators. The total instrument was scored by using summation of the ordinal scores, retaining not applicable scores only for descriptive purposes. A Graded Response Model (GRM) was estimated using the mirt package (version 1.45.1) in R.

Next, DIF analyses were performed. A logistic ordinal regression evaluation of multigroup variation in GRM parameters was used, as recommended and implemented in NIH’s Patient-Reported Outcomes Measurement Information System (PROMIS) and related lordif package (version 0.4.2). Omnibus DIF evaluation also included separate evaluation of uniform and nonuniform DIF. Chi-square values beyond a .01 alpha level were used to screen for statistical differences in item performance among groups. However, since chi-square is sensitive to sample size (Menard, 2002), we additionally implemented evaluation of change in McFadden’s pseudo-R² as a measure of effect size and practical significance of DIF. ΔR² values ≥ .02 were considered meaningful but small, and ΔR²≥ .07 was considered a large DIF effect (Jodoin & Gierl, 2001). Analyses were conducted across four demographic contrasts: sex (male vs. female), primary language (English vs. non-English and predominantly Spanish speaking), education level (≤12 years vs.>12 years), and a dichotomized race/ethnicity (non-Hispanic White vs. other), as these represented the largest subgroups of our sample. Of note, some individuals will be represented in more than one of these subgroups.

Microsoft Co-Pilot (based in part on Chat GPT model 4.0 and 5.0) and Anthropic (model 4.6) was utilized for coding assistance, language clarification, and formatting purposes during the analysis and manuscript reporting process. The use of these tools did not replace core activities by the researcher or author responsibilities, and was conducted with human oversight.

Content Review

Prior to analyzing the quantitative data, the authors (5 neuropsychologists, 1 social worker, and 1 psychometrist, all with a minimum of 5 years’ experience working with older adults with neurodegenerative diseases) served as the expert review panel and independently reviewed the items of the measure. Via an online survey form with structured questions and free text response fields, the panel gave feedback on item clarity/understandability, potential modifications needed to apply to a heterogenous patient population, the relevance of the item content, impact of technologies on this item in modern contexts, and any further recommendations for revising this item. In addition, the panel asked to identify domains the current scale did not cover, and additional reflections on technological difficulties in daily tasks described by their patients/collaterals. The study team reviewed responses for the purposes of this manuscript, but formal thematic analysis was not conducted on this data.

Data and Code Availability.

Data can be requested with appropriate approvals from the Texas Alzheimer’s Research and Clinical Care Consortium Website https://www.txalzresearch.org/. SPSS and R code used in the analyses are available by request to the corresponding author.

Results

Sample and Descriptive Information

Table 1 summarizes demographic descriptive data for participants and collaterals/co-participants. Table 2 displays item level responses from the Lawton and Brody iADL scale for the sample as a whole, and supplemental tables (S1-S3) describe iADL item level response descriptives and frequencies in the no cognitive disorder, MCI, and dementia groups.

Table 1:

Descriptive Data for the Sample (N=3552)

Participant Characteristics		%	Mean	SD
Age			73.5	9.7
Years of Education			13.3	4.3
Ethnicity	Non-Hispanic	62.9
	Hispanic	37.1
Race	White	93.0
	Black or African American	4.1
	American Indian or Alaskan Native	0.2
	Native Hawaiian or Other Pacific Islander	0.1
	Asian	2.6
Language	English	81.8
	Spanish	17.5
	Other or Unknown	0.7
Clinical Dementia Rating	No Cognitive Impairment	42.0
	Mild Cognitive Impairment	26.5
	Dementia	31.4
Co Participant Characteristics		%	Mean	SD
Age			63.5	14.5
Years of Education			11.8	5.9
Ethnicity	Non-Hispanic	64.3
	Hispanic	35.7
Race	White	78.5
	Black or African American	3.6
	American Indian or Alaskan Native	0.2
	Asian	0.5
	Missing, Other, or Unknown	17.1
Relationship to Participant	Spouse/Partner	46.9
	Child	22.9
	Sibling	3.9
	Other Relative,	3.0
	Friend or Neighbor,	8.7
	Paid Caregiver/Provider	0.2
	Other or Unknown	14.3
Lives with Subject	No	45.3
	Yes	54.7

Table 2:

Item level Descriptive Data for the Lawton and Brody (N=3552)

	%	Mean	SD
Telephone		1.65	1.20
Operates telephone on own	72.2
Dials a few numbers	8.5
Answers but does not dial	6.1
Talks on phone if given	8.4
Incapable of using	4.9
Shopping		1.81	1.14
Independent	63.0
Limited purchases	7.8
Needs to be accompanied	14.8
Completely unable	14.4
Food Preparation		1.84	1.24
Not applicable, never did	5.1
Plans, prepares, and serves	62.2
Prepares if provided ingredients	5.3
Heats and serves prepared meals	8.0
Needs to have meals prepared	19.4
Housekeeping		1.87	1.38
Not applicable, never did	4.8
Maintains house alone	60.7
Performs light tasks	12.8
Performs light tasks, unclean	5.5
Needs help with all	5.7
Unable to participate	10.4
Laundry		1.48	0.80
Not applicable, never did	7.2
Does personal laundry completely	66.7
Launders small items	8.0
All laundry done by others	18.0
Transportation		1.91	1.32
Travels independently/drives	65.1
Arranges via taxi or bus	5.1
Travels on transportation when accompanied	3.2
Only travels with others	26.6
Finances		1.63	0.84
Not applicable, never did	2.6
Manages independently	59.1
Manages day to day purchases but needs help with major	15.3
Incapable of handling money	23.0
Medications		1.58	0.84
Not applicable	1.3
Is responsible for taking	64.1
Takes if prepared/reminded	11.5
Incapable of managing	23.1

Summarizing across this information, a one-way MANOVA revealed a statistically significant effect of clinical stage on each of the IADL items (Pillai’s Trace = 0.851, F(16, 6214) = 287.37, p < .001) with the expected pattern of increasing severity of cognitive impairment stage associated with worse iADL performance across items (all p values < .001). Group differences between normal cognition and MCI on total/mean scores were modest, which is consistent with the instrument’s original design to index dependence rather than these more subtle changes. Stage stratified item distributions are provided in supplemental tables S1-S3.

When considering the earliest clinical transition (from normal to MCI), we specifically evaluated which items showed the strongest effect size. The largest differences were observed for items assessing shopping (no cognitive disorder: M=1.04, SD=0.27; MCI: M=1.32, SD=0.69; Δ=0.28); transportation (no cognitive disorder: M=1.06, SD=0.33; MCI: M=1.34, SD=0.82; Δ=0.28); and finances (no cognitive disorder: M=1.03, SD=0.19; MCI: M=1.31, SD=0.54; Δ=0.28).

Item Response Theory Analyses

Establishing Unidimensionality.

A confirmatory factor analysis (CFA) was conducted to evaluate the unidimensionality of the iADL scale using a diagonally weighted least squares (DWLS) estimator appropriate for ordinal data. The one-factor model demonstrated excellent fit to the data, χ²(20) = 134.30. Incremental fit indices indicated near-perfect model fit (CFI = 1.00; TLI = 1.00). Absolute fit indices further supported this conclusion, with a root mean square error of approximation (RMSEA) of .043, 90% CI [.036, .050], and a standardized root mean square residual (SRMR) of .009. The test of close fit for RMSEA ≤ .05 was supported (p = .954), whereas the test for poor fit (RMSEA ≥ .08) was rejected (p < .001). Collectively, these results provide strong evidence that the iADL items reflect a single underlying latent construct, supporting the unidimensional structure of the scale.

Item Parameters:

IRT parameters (Table 3) suggest that all the Lawton and Brody iADL items demonstrated exceptionally high discrimination (range = 5.58–8.44). When evaluating the threshold parameters for individual responses for each item, these clustered around 0.3 and 1.7 on the latent trait scale (θ). Pragmatically, 0 for this latent variables represents average functioning in the sample and higher values reflect greater functional impairment.

Table 3:

Item Parameters

Item	Discrimination Parameter	Threshold Parameters (Difficulty)
		1	2	3	4
Telephone	5.999	0.637	0.938	1.182	1.694
Shopping	8.437	0.329	0.589	1.114	—
Food Preparation	8.413	0.416	0.605	0.894	—
Housekeeping	5.875	0.366	0.823	1.039	1.296
Laundry	5.706	0.621	0.942	—	—
Mode of Transportation	5.578	0.390	0.56	0.683	—
Finances	7.282	0.269	0.789	—	—
Medications	6.805	0.401	0.796	—	—

Note: N=3552

Test information curves (Figure 1A) peaked approximately .3 to 1.5 standard deviations above the mean, where higher values indicate worse functioning. This suggests that the scale’s greatest precision in the impaired range of functioning and comparatively less information at the normal/MCI boundary. The initial endorsement for difficulties with these items (moving from no difficulty to the first difficulty rating) provides information in a relatively narrow window of underlying functional ability (figure 1B). Similar to the mean level analysis, difficulties with finances and shopping are again revealed to be the most sensitive items to early functional impairment.

Figure 1.

1A represents item information curves for all Lawton and Brody IADL scale items; 1B represents the category curves for each item’s first response option. In both graphs, theta is represented such that higher values indicate more functional impairment. Sample size is N=3552.

Figure 2 illustrates the distribution of estimated functional ability (θ) as measured by the Lawton & Brody iADL scale in individuals with no cognitive diagnosis, mild cognitive impairment, and dementia. Individuals with normal cognition (blue) cluster at the far left of the continuum, indicating higher functional ability, while those with mild cognitive impairment (yellow) also tend to cluster at values near zero. In contrast, individuals with dementia (green) are concentrated toward the right, reflecting greater functional difficulty. This graph illustrates the limitations to the scale at capturing subtle functional difficulties in those with no cognitive disorder or mild cognitive impairment. The scale primarily differentiates individuals with more substantial functional impairment and may provide limited precision for detecting subtle functional changes among those with normal cognition or MCI (or transitional states between them).

Figure 2.

Person item map showing functional ability (X axis) where greater impairment is represented by higher values. Frequency of underlying score is represented by diagnostic group on the Y axis. Sample size is N=3552.

Differential Item Functioning

Starting with gender, four items were flagged as having both uniform and nonuniform DIF (see Table 4); however, only items addressing laundry and transportation demonstrated a clinically meaningful degree of DIF, with ΔR² of .02 and .03 respectively, primarily driven by a uniform DIF finding. It is important to note that purifying for DIF in total theta estimates did not result in meaningful differences in estimates (theta change was not ≥ 0.1). However, there was a trend for the driving item in women to have less information at the lower end of ability spectrum, and higher ratings of more impairment, before purification.

Table 4:

Differential Item Functioning by Dichotomized Demographic Variables

		Uniform DIF		Any DIF		Non-Uniform DIF
Dichotomized Grouping Variable	IADL Item	χ2 p	ΔR2	χ2 p	ΔR2	χ2 p	ΔR2
Sex	Ability to Use Telephone	0.04	0.00	0.12	0.00	0.63	0.00
	Shopping	0.91	0.00	0.87	0.00	0.60	0.00
	Food Preparation	0.03	0.00	0.00	0.00	0.00	0.00
	Housekeeping	0.00	0.00	0.00	0.00	0.00	0.00
	Laundry	0.00	0.01	0.00	0.02	0.00	0.00
	Mode of Transportation	0.00	0.02	0.00	0.03	0.00	0.01
	Finances	0.03	0.00	0.10	0.00	0.87	0.00
	Medications	0.97	0.00	0.58	0.00	0.30	0.00
Language	Ability to Use Telephone	0.00	0.00	0.00	0.00	0.18	0.00
	Shopping	0.15	0.00	0.35	0.00	0.86	0.00
	Food Preparation	0.00	0.00	0.00	0.00	0.69	0.00
	Housekeeping	0.14	0.00	0.27	0.00	0.50	0.00
	Laundry	0.40	0.00	0.65	0.00	0.68	0.00
	Mode of Transportation	0.00	0.01	0.00	0.02	0.00	0.01
	Finances	0.97	0.00	0.98	0.00	0.86	0.00
	Medications	0.06	0.00	0.15	0.00	0.56	0.00
Education	Ability to Use Telephone	0.00	0.00	0.00	0.00	0.54	0.00
	Shopping	0.06	0.00	0.17	0.00	0.99	0.00
	Food Preparation	0.00	0.00	0.00	0.00	0.95	0.00
	Housekeeping	0.56	0.00	0.00	0.00	0.00	0.00
	Laundry	0.39	0.00	0.50	0.00	0.42	0.00
	Mode of Transportation	0.00	0.01	0.00	0.02	0.00	0.01
	Finances	0.28	0.00	0.52	0.00	0.71	0.00
	Medications	0.04	0.00	0.12	0.00	0.97	0.00
Race/Ethnicity	Ability to Use Telephone	0.03	0.00	0.06	0.00	0.42	0.00
	Shopping	0.00	0.00	0.00	0.00	0.01	0.00
	Food Preparation	0.06	0.00	0.17	0.00	0.97	0.00
	Housekeeping	0.61	0.00	0.32	0.00	0.15	0.00
	Laundry	0.15	0.00	0.26	0.00	0.43	0.00
	Mode of Transportation	0.00	0.01	0.00	0.02	0.00	0.02
	Finances	0.03	0.00	0.06	0.00	0.27	0.00
	Medications	0.17	0.00	0.18	0.00	0.22	0.00

Note: Values for Chi-Square are bolded if p<.01; McFadden’s Pseudo R2 is bolded if value greater than or equal to .02; item name per contrast is bolded if both those criteria are met for one form of DIF

In regards to the race/ethnicity dichotomization, while all items were flagged as statistically significant, again only the driving item approached meaningful significance (ΔR²=.02), primarily powered by nonuniform DIF. Purification did not result in meaningful differences in theta estimates, but there was a trend for minoritized individuals to be rated as having less impairment and to have the DIF be most prominent at the lowest end of ability spectrum

A similar pattern emerged for language, where three items were flagged as DIF, but only the transportation item was flagged for both uniform and non-uniform DIF to a borderline meaningful way (ΔR²=.02). Purification did result in meaningful differences in theta ratings for many cases, with a tendency for Spanish-speaking individuals to have been rated as having less functional impairment before correcting for DIF (DIF most prominent at lower end of theta spectrum).

As it relates to education, only the transportation item bordered on having a meaningful DIF (ΔR²=.02). Purification did not result in meaningful differences in theta estimates, but there was a similar trend of the largest differences being at the lower end of the theta spectrum.

Content Evaluation

Overall, clinicians viewed the item content from the original Lawton and Brody iADL scale as still broadly relevant with the exceptions of the current telephone use (83.3%) and laundry (66.7%) items. That being said, broad themes for item refinement were suggested, which are summarized in Table 5.

Table 5:

Summary of Expert Feedback on Lawton and Brody iADL scale Items

Domain	Key modernization changes	Applicability to diverse contexts	Technology-specific considerations	Refinements to response options
Ability to Use Telephone	Use “phone/smartphone” instead of “telephone”; explicitly include texting, contact list use, answering/swiping, and scam handling.	Differentiate preference vs. ability (does not use vs. cannot use); account for longstanding family roles.	Unknown numbers/scams; finding the right app; silencing/notifications; charging; photos.	Add options for “does not use” vs “cannot use”; grade initiation vs. response; include “with reminders.”
Shopping	Either separate or explicitly include online shopping; clarify accompanied vs. independent and shopping vs. payment steps.	Allow “historically never shopped”; consider cultural/family role norms and access to accounts/transport.	Online purchasing pitfalls (duplicates/impulsivity); passwords; payment methods (cash/card/device).	Clarify “small purchases”; add options for online vs. in-person; note reasons (cognitive vs. physical).
Food preparation	Clarify “adequate meals” vs. ability to prepare; include microwaves/heat-and-eat; recipe-following and delivered meals.	Account for culture, lifelong roles, and living settings (e.g., facilities providing meals); include “never cooked.”	Microwave and appliance learning; recipe apps; safety (stove off).	Anchor on preparation skills (not nutrition); add safety cues; include NA (“never cooked”).
Housekeeping	Rename from “housekeeping” to inclusive chores/maintenance; include yardwork/dishwasher; separate physical vs. cognitive limits.	Role/gender norms; facility vs. patient tasks; note aberrant changes (e.g., hoarding).	Learning modern appliances; growth of automation (e.g., robot vacuums).	Avoid dichotomies that conflate task complexity; clarify what “maintenance” entails; add NA (“never did”).
Laundry	Define steps (sort, wash, transfer, dry, fold, put away); remove outdated garment distinctions.	Family role division and living situation; allow NA (“never did this”).	Learning new/smart washer/dryer setups.	Score partial step completion; modernize wording (remove “stockings”).
Transportation	Include rideshare/public transit app use; grade GPS reliance; separate driving safety vs. accompaniment/homebound.	Urban/rural contexts; voluntary driving cessation; sensory/motor limitations; historic non-drivers.	App navigation (Uber/Lyft); car tech; early decline markers (getting lost).	Provide parallel anchors for driving and for public transit/rideshare; add airport/air travel navigation as optional subitem.
Medications	Distinguish organizing/refilling from taking; incorporate reminders/alarms, pill packs/dispensers; add supplements.	Facility-managed meds; family habits; evaluate capability vs. habit.	Smartphone/watch alarms; electronic dispensers; prepackaged systems (PillPak).	Add level for “takes correctly when prepared”; capture error frequency; include reminders.
Finances	Explicitly include online banking/autopay/passwords/scams; separate day-to-day purchases, major decisions, and taxes.	Marital/family role norms; trust in banks; autopay prevalence.	Password resets; phishing/scams; website/app changes; payment apps.	Replace generic “incapable of handling money” with specific capability anchors; include autopay setup competency.

Revisions were suggested for activities of daily living that have been impacted by technology. For example, the existing telephone item was recommended to be modernized both in the terms used and the specific activity that is queried. Examples felt to be clinically relevant included explicit inclusion of texting and not just voice calls, using contact lists, swiping/answering challenges, and dealing with scam/unknown callers. In a similar vein, reviewers recommended separately querying about online shopping, an activity viewed by clinicians as potentially having distinct risks (buying duplicates or impulsively buying items) and different cognitive demands (remembering passwords, navigating websites, novel payment methods) not imagined by the original authors of the measure. In terms of transportation, querying how using GPS, navigating ridesharing apps, or situations where the participant is driving despite being unsafe were all mentioned as refinements that may be of value. Finally, adjusting the financial measure to encapsulate autopay, online banking, and scams was felt to be relevant as well.

Sociocultural changes in certain iADLs and physical contributions to iADL task performance were also highlighted by raters. For example, the existing housekeeping item may need to be broadened to specify what aspect of household maintenance is being queried, and the recognition that physical/sensory issues might contribute to task completion or perceived difficulty. Societal changes in eating/meal behaviors were noted as relevant as well, with raters suggesting the food preparation item could be clarified to include the advent of microwavable meals and greater specificity on what was meant by “adequate”.

In terms of querying about additional domains not assessed in the current Lawton and Brody iADL scale, recommendations included adding items focused on managing calendars and appointments, general computer and technology use, and querying social engagement and hobbies. Specific technological problems not captured by the current measure included difficulties with recalling passwords, smartphone user interfaces (specifically silencing/swiping), problems with smart TV’s and the TV remote, hearing aids, and adapting to app/website changes.

Discussion

The Lawton and Brody iADL Scale remains a widely used and historically significant tool for assessing functional independence in older adults. Revisiting this measure over a half century from its initial publication, our findings suggest a) item performance at earlier stages of disease presentation may be sub-optimal for detecting subtler functional impairment b) the item assessing transportation may not perform in the same manner across various subgroups, and c) changes to the technological and sociocultural context over the past decades may necessitate the need for updating the items. We consider each of these points in turn.

How does the Lawton and Brody perform in earlier stages of impairment?

Consistent with the instrument’s original purpose (namely describing independence in daily task performance), the instrument demonstrated strong unidimensionality and item discrimination separating dementia from earlier stages of decline. The existing scale demonstrates high information (a ≈ 5.6–8.4) within a narrow band of functional impairment (information peaking roughly θ≈0.3–1.5, where higher θ reflects worse function). In practical terms, the scale sharply distinguishes moderate impairment but is comparatively insensitive to subtle functional changes at the normal/MCI boundary (and presumably any intervening stage like transitional cognitive decline), where scores cluster at floor. This is not surprising, as the measure was designed to capture actual functional dependence rather than subtler difficulties or changes, as concepts such as MCI and transitional cognitive decline were not clinical constructs at the time of the scale’s development.

The items assessing transportation, shopping, and financial management in particular were most sensitive to earlier stages of functional decline, similar to what was reported by McGrory and colleagues (2014). However, the relatively narrow range of the underlying dimension of functional impairment captured by these items suggests insensitivity to earlier stages of functional decline. For example, the vast majority of both normal and MCI participants scored a zero on the scale. On the one hand, this is good inasmuch as individuals with MCI should be functionally independent day to day. However, the ability to capture concepts such as independent but more difficult or independent but making errors is lacking. These more fine grained distinctions are critical for characterizing concepts such as “transitional cognitive decline” (Jack Jr. et al., 2024), the subtle changes that precede MCI. As ADRD research moves increasingly to prevention and asymptomatic/early stage disease interventions, quantification of subtle functional decline will be critical.

With that in mind, it is telling that items assessing transportation use, shopping and financial management are also among the items clinicians identified as most impacted by technological innovations. It is possible that using passive activity monitoring of these domains could yield more sensitive indicators of earlier functional decline. Such work has begun to be demonstrated with functional measures of online shopping (Woods et al., 2017), online financial management (Wild et al., 2022) and passive monitoring of driving behaviors (Chen et al., 2025). We believe informant reports are likely to remain scalable, readily used in clinic and research settings, and familiar to clinicians, scientists, and regulatory groups alike. Hybrid approaches that capture, collate, and interpret these many data sources to arrive at more precise indicators of functional status are needed as our field evolves.

We believe that if attempts are made to refine the Lawton and Brody for continued use in modern context, care should be taken to prioritize the development of items and response options that could be more sensitive to the earliest transitions in functional capacity, as well as work with caregivers and informants to find language that helps them convey these subtle functional impacts that precede true dependence in a given life domain.

Does the Lawton and Brody perform the same across demographic subgroups?

Our evaluation included subsamples defined by sex, primary language, education, and a dichotomized race/ethnicity variable. The items performed comparably in our predominantly Spanish speaking subsample, adding to our understanding of the psychometric properties of this version of the instrument (Vergara et al., 2012). Within these groups, DIF was observed for six items, but the magnitude was minimal for most items

It is notable that the transportation item showed a small degree of clinically significant DIF across subgroups evaluated. This is consistent with prior literature that showed DIF on the transportation item (Jang & Kawachi, 2019; LaPlante, 2010). Pragmatically, this finding implies that the item may function differently in these different groups, or at least clinicians may be perceiving the behavior differently in these subgroups. It is possible that refining items to capture specific aspects of transportation (most likely driving for residents of Texas) may be necessary to address this phenomena. Alternatively, it may be that contextual factors explain this difference (Bruderer-Hofstetter et al., 2020). For example, not being able to take the written test for a driver’s license due to language or educational factors could make items perform in an aberrant way psychometrically. Additionally, regional differences in transit options, particular transportation uses cases, sprawl, and population density that covary with demographics may also be a contributing contextual factor. While the DIF was relatively small in magnitude, it was one of the three items that was most sensitive to earlier stages of the disease, so this is a domain worth refining further if the goal is to capture milder functional impacts in an accurate and generalizable fashion.

To that end, DIF would be problematic if it contributed to diagnostic misclassification or inequitable care planning, which were not evaluated in this study. Nonetheless, it does seem possible to develop instruments that are less prone to DIF across groups. For example, the Amsterdam iADL measure was developed with careful consideration of sociocultural and technological factors, and has shown limited DIF across gender, culture, age and educational groups in 8 different countries (Dubbelman et al., 2020).

How relevant are Lawton and Brody items to current practice and research?

While expert reviewers rated item domains from the Lawton and Brody as broadly still applicable, it was clear that several items are increasingly mediated by digital technologies. As noted previously (Benge et al., 2024), the impact of the internet and computer revolutions have radically impacted nearly every day to day task. This raises the question about how our definition of functional impairment needs to shift as well. At what point does technological skill and literacy become a requirement for performing instrumental tasks? Does a technologically enriched environment make daily tasks harder due to increased complexity, or easier due to automation of tasks? How do bidirectional relationships between technological engagement and cognition impact both domains (Benge & Scullin, 2025)? These are empirical and theoretical questions for future studies, but they have practical significance. If a goal of a neuropsychological evaluation is to understand not just cognitive performance, but how that performance impacts real world functioning, one would argue it is a critical research agenda for our field.

Similarly, shifts in social roles and expectations have, and will continue to, alter the relevance and interpretation of iADL items. The current sample primarily includes individuals born in the post- World War II generation, with relatively distinctive sociocultural factors. Subsequent generations witnessed a number of marked changes, including multiple waves of immigration to the US, increasing participation of women in the workforce, and changes in wealth distribution. How these factors will influence both between and within generational iADL performances remains to be seen.

Implications for refining measures of instrumental activities

As we discuss in the introduction, the continued refinement and update to legacy instruments is critical, even in the presence of new instruments with sound psychometric properties are available We believe that the ability to leverage data across clinical and research databases, developing a measurement science that expects continuing adaptation and improvements in instruments as opposed to developing static instruments, and helping legacy clinical and regulatory guidelines adapt to changing daily functions and demands is critical to making assessments relevant in the future. To this end, our results suggest that the Lawton and Brody IADL Instrument, as it was developed half a century ago, likely has significant limitations. It does not capture many now-relevant aspects of daily living, and what it does capture may be insensitive to the subtle changes most important to the field currently and in the future. Thus, we suggest a two-pronged approach to refining functional assessment tools.

First, the broader context of iADL performance should be considered. This includes considering development samples that capture a broad and representative sample and employing psychometric techniques that account for potential subgroup differences. Further, environmental factors such as degree of technological requirements/utilization in the environment, the role of non-cognitive factors (i.e. sensory and motor status), and social networks should be considered when evaluating functional status (Bruderer-Hofstetter et al., 2020).

Second, instrument development for activities of daily living should be a dynamic and not static process. As shown by González and colleagues (González et al., 2025), it is possible to use linking and equating techniques to add and refine instruments to legacy measures. This helps the field to both leverage existing data resources and cohorts while also advancing our knowledge of how real brains are functioning in the real world. We are therefore pursuing measurement modernization approaches that clarify wording on the basis of feedback from caregivers, links to legacy items, and developing complementary questions that can supplement domains not covered by the original instrument.

While innovating though, it is important that revisions should preserve the original purpose of the measure: the capacity to identify and score major areas of functional dependence. While extending sensitivity to earlier functional transitions is a key goal, the ability to capture frank dependence in core daily living domains must be retained, as the utility of this scale has indeed led the measure to be a cornerstone of clinical practice and ADRD research for over five decades.

Limitations

This study is limited by the geographic and demographic scope of the sample. Data were drawn from a Texas-based consortium of academic medical centers, and the most represented subgroups include non-Hispanic White and Hispanic individuals and English/Spanish language speakers. Notably, Black or African American individuals comprised 4.1% of our sample, which is below their representation in the Texas population (approximately 13%), limiting the generalizability of DIF findings for this group. Future work would benefit from continued exploration of the measure in additional representative samples, particularly multi-group DIF as our sample was restricted to dichotomous DIF analyses given the primary subgroups available in our sample. We further note the lack of cross-validation with other functional measures. There are alternative ways of scoring the Lawton and Brody iADL measure (Lawton et al., 1982), and the fact that the current version used was slightly different than the original version may limit the generalizability of these results to different instantiations of the measure. We note that data collection on this project ended pre-pandemic era, and tech usage in older adults that could impact iADL performance might have increased since the time the data was collected further impacting interpretation of these items by collateral reporters. The qualitative data was gathered informally from the investigator team, which also did not include collateral reporters for their input. Our future work includes formal thematic analyses with collateral reporters in an attempt to provide more user focused and real world revisions to wording and themes.