Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE)

1. Abstract

The Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE) is an internationally recognized, collateral-source psychometric screening instrument developed by Anthony F. Jorm in 1989. Engineered to circumvent the substantial socio-demographic biases—most notably education, pre-morbid intellectual quotient (pre-morbid IQ), language barriers, and socioeconomic status—that compromise traditional direct cognitive performance tests such as the Mini-Mental State Examination (MMSE), the IQCODE assesses intra-individual cognitive change over a standardized 10-year retrospective interval. The instrument gathers behavioral observations from a knowledgeable collateral informant, such as a spouse, adult child, sibling, or dedicated caregiver who has sustained consistent personal contact with the older individual.

The original instrument comprises 26 items (with a widely validated 16-item short form developed in 1994), each evaluating everyday manifestations of memory, executive functioning, mental speed, language comprehension, and practical problem-solving. Each item is rated on an anchored 5-point Likert scale ranging from 1 (Much improved) to 5 (Much worse), with a rating of 3 denoting Not much change. Psychometric investigations conducted across clinical memory clinics, acute geriatric wards, epidemiologic community surveys, and cross-cultural cohorts consistently document extraordinary internal consistency (Cronbach’s alpha typically ranging between .93 and .97 for both the 26-item and 16-item versions) and high test-retest reliability ($r > .75$ to $.96$ across diverse test-retest windows). Exploratory and confirmatory factor analyses demonstrate that the tool is essentially unidimensional, dominated by a pervasive general cognitive decline factor. ROC curve analyses indicate robust diagnostic accuracy for detecting dementia, major neurocognitive disorder, and secondary neurodegenerative processes, routinely demonstrating sensitivity and specificity values exceeding 80% to 88% at established cut-off thresholds (typically 3.30 to 3.60 depending on the clinical context).

2. Keywords

IQCODE, Informant Questionnaire on Cognitive Decline in the Elderly, dementia screening, collateral report, cognitive decline, Alzheimer's disease, Anthony F. Jorm, psychometrics, neurocognitive assessment, geriatric psychiatry

3. Authors

The primary creator and principal developer of the Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE) is:

• Anthony Francis Jorm, PhD, DSc, FASSA — Emeritus Professor, Melbourne School of Population and Global Health, University of Melbourne, Victoria, Australia; formerly of the NHMRC Social Psychiatry Research Unit, The Australian National University, Canberra, ACT, Australia. Contact: [email protected].

Key initial co-investigators, methodological contributors, and statistical collaborators who participated in early psychometric validation studies include:

• Patricia A. Jacomb, BA (Hons) — NHMRC Social Psychiatry Research Unit, The Australian National University, Canberra, Australia.
• Ailsa E. Korten, BSc — NHMRC Social Psychiatry Research Unit, The Australian National University, Canberra, Australia.
• Helen Christensen, PhD, FASSA — Scientia Professor of Mental Health, Black Dog Institute, University of New South Wales (UNSW), Sydney, Australia; formerly of The Australian National University.
• Ruth Scott, MB, BS, FRANZCP — Geriatric Psychiatrist, Mental Health Services for the Elderly, Woden Valley Hospital, Canberra, ACT, Australia.

4. Purpose

The fundamental clinical and epidemiological purpose of the IQCODE is to provide a valid, reliable, and bias-free assessment of cognitive change in older adults by capitalizing on observational behavioral reports from a reliable collateral informant. Traditional performance-based cognitive screening instruments—exemplified by the Mini-Mental State Examination (MMSE), the Montreal Cognitive Assessment (MoCA), and the Addenbrooke's Cognitive Examination (ACE)—require direct interactive participation from the patient. Consequently, performance scores are heavily confounded by the patient’s pre-morbid intellectual ability, years of formal education, socioeconomic background, primary language fluency, test anxiety, and sensory deficits such as profound visual or hearing impairments. In populations with minimal or no formal literacy, direct testing can yield alarmingly high false-positive rates for dementia, whereas in highly educated individuals, high cognitive reserve can mask clinically meaningful decline, yielding false negatives.

The IQCODE directly circumvents these structural psychometric confounders by shifting the frame of reference from normative interpersonal comparison (evaluating how an individual performs relative to a population norm) to ipsative longitudinal comparison (evaluating how an individual functions relative to their own baseline performance a decade prior). The questionnaire asks the informant to explicitly compare the individual's current everyday functioning with their performance 10 years earlier across a diverse array of ecologically valid behaviors. These behaviors include remembering recent conversations, finding personal belongings, navigating familiar domestic electronics, managing financial transactions, handling arithmetic problems, and making prudent daily decisions.

Clinical applications of the IQCODE are extensive and diverse:

• Primary Care Triage: General practitioners frequently encounter patients who deflect or mask subtle deficits; administering the IQCODE to an accompanying family member permits rapid detection of objective decline without inducing defensive reactions in the patient.
• Secondary and Tertiary Memory Clinics: The IQCODE forms an indispensable pillar of comprehensive neurocognitive diagnostic protocols, providing an independent source of collateral history that complements direct neurocognitive testing, neuroimaging (e.g., MRI, amyloid-PET), and biological fluid biomarkers.
• Delirium and Acute Hospital Geriatrics: In emergency room and post-operative orthopedic settings, older patients frequently exhibit acute cognitive disturbances (delirium). Direct tests administered in these states capture acute encephalopathy rather than baseline dementia. Because the IQCODE explicitly anchors questions to a 10-year baseline preceding the acute hospitalization, it allows clinicians to differentiate pre-existing chronic neurodegenerative decline from pure acute delirium.
• Cross-Cultural and Low-Literacy Epidemiological Surveys: Because the items index everyday domestic, social, and functional activities rather than abstract paper-and-pencil tasks, the IQCODE can be deployed across non-Western, illiterate, or socioeconomically disadvantaged populations with minimal adaptation.

5. Psychological Construct

The psychological construct measured by the IQCODE is perceived functional-cognitive decline relative to pre-morbid baseline. While formal neuropsychology traditionally conceptualizes human cognition through compartmentalized neurocognitive domains—such as episodic memory, semantic memory, working memory, executive function, visuospatial processing, and instrumental activities of daily living (IADLs)—the IQCODE operationalizes these dimensions as they manifest in observable, ecologically valid everyday life.

Episodic and Short-Term Memory

This facet assesses the encoding, consolidation, and delayed retrieval of temporally indexed biographical events and recent daily occurrences. In early-stage Alzheimer's disease, neurofibrillary pathology predominantly infiltrates the transentorhinal cortex and hippocampus, producing rapid anterograde amnesia. In the IQCODE, this underlying neurobiological breakdown is assessed by items such as:

• Item 4: Remembering things that have happened recently.
• Item 5: Recalling conversations a few days later.
• Item 8: Remembering what day and month it was.
• Item 10: Remembering where to find things which have been put in a different place than usual.

Semantic Memory and Crystallized Knowledge

Unlike episodic memory, crystallized knowledge and semantic memory—stored throughout anterior temporal and association neocortices—tend to remain relatively preserved in normal aging and only deteriorate as neurodegeneration becomes diffuse. The IQCODE includes items sensitive to this domain to distinguish between normal age-associated changes and pervasive cortical pathology:

• Item 1: Recognizing the faces of family and friends (prosopagnosia/severe person-knowledge deficit).
• Item 2: Remembering the names of family and friends.
• Item 15: Remembering things that happened to him/her when he/she was young.
• Item 17: Understanding the meaning of unusual words.

Executive Function, Problem-Solving, and Fluid Intelligence

Executive functioning encompasses supervisory attentional control, working memory, cognitive flexibility, abstract reasoning, and strategic decision-making, mediated by frontostriatal circuits. Everyday functional manifestations of executive decline assessed by the IQCODE include:

• Item 11: Adjusting to any change in his/her day-to-day routine (cognitive flexibility and set-shifting).
• Item 13: Learning to use a new gadget or machine around the house (procedural acquisition, working memory, fluid adaptation).
• Item 22: Making decisions on everyday matters (judgment, ecological reasoning).
• Item 26: Using his/her intelligence to understand what’s going on and to reason things through (global fluid reasoning).

Instrumental and Practical Arithmetic Competence

Higher-order instrumental functioning requires the integration of numerical cognition, sequencing, sustained attention, and prospective memory. Deficits in complex IADLs often serve as the diagnostic demarcation point between Mild Cognitive Impairment (MCI) and manifest dementia:

• Item 23: Handling money for shopping.
• Item 24: Handling financial matters, e.g., the pension, dealing with the bank.
• Item 25: Handling other everyday arithmetic problems, e.g., knowing how much food to buy, knowing how long between visits from family or friends.

6. Theoretical Framework

The IQCODE is grounded in a convergence of cognitive neuropsychology, the psychometrics of latent growth and ipsative measurement, and cognitive reserve theory as articulated by Yaakov Stern and colleagues.

Classical Test Theory and the Confound of Pre-morbid Ability

Under Classical Test Theory (CTT), an observed score on any direct cognitive examination ($X$) reflects the true latent cognitive ability ($T$) plus measurement error ($E$), represented as $X = T + E$. However, in cross-sectional cognitive screening, $T$ represents the individual's current absolute cognitive capacity, which is inherently a composite of two divergent latent factors: (1) pre-morbid intellectual endowment and cognitive reserve accumulated across the lifespan ($T_{\text{pre}}$), and (2) the acquired pathological cognitive decline attributable to neurodegenerative disease ($\Delta T_{\text{path}}$):

$$X = (T_{\text{pre}} – \Delta T_{\text{path}}) + E$$

Because an examiner administering an MMSE or MoCA cross-sectionally observes only $X$, it is statistically impossible to distinguish whether a low score of 22/30 represents a brilliant scholar who has suffered massive catastrophic decline ($\Delta T_{\text{path}} = 10$) or an individual with minimal educational attainment whose lifelong baseline score was 22/30 ($\Delta T_{\text{path}} = 0$). Jorm realized that by altering the measurement paradigm to an explicit collateral retrospective change scale, the instrument isolates the delta term directly:

$$\text{IQCODE Item Score} propto \Delta T_{\text{path}} + E_{\text{informant}}$$

Cognitive Reserve Theory

Cognitive reserve posits that individual differences in brain networks, synaptic density, and educational/occupational complexity provide resilience against brain damage, allowing individuals to sustain neurodegenerative neuropathology without demonstrating overt clinical impairment on standard cross-sectional tests. Consequently, standard psychometric tests often misclassify high-reserve individuals as normal until neuropathology is advanced, while misclassifying low-reserve individuals as impaired. The theoretical architecture of the IQCODE neutralizes cognitive reserve as a confounding variable because an informant evaluates an individual’s performance strictly against the individual’s own historical personal standard. A university professor who can no longer balance their financial accounts (Item 24) or manage a home computer (Item 12) will be identified as having declined, even if their residual verbal knowledge enables them to score 28/30 on an MMSE.

Informant Report as an Ecological Window

Direct cognitive testing represents a brief snapshot (typically 10 to 45 minutes) of an individual’s performance in an artificial, quiet, structured clinical environment under one-on-one supervision. This artificial structure often compensates for executive and attention deficits that manifest prominently in unstructured, noisy real-world environments. Collateral informants observe the individual longitudinally across diverse naturalistic ecological settings—under conditions of fatigue, multitasking, domestic stress, and social complexity. Therefore, the informant serves as an ecological observer whose longitudinal observations aggregate thousands of naturalistic behavioral trials over time, maximizing the ecological validity of the clinical assessment.

7. Validity

The construct, criterion, convergent, and discriminant validity of the IQCODE has been scrutinized across hundreds of international peer-reviewed investigations spanning more than three decades.

Criterion and Diagnostic Validity

Criterion-related validity has been demonstrated by comparing IQCODE scores against gold-standard clinical diagnoses of dementia formulated according to DSM-III-R, DSM-IV, DSM-5, and ICD-10 criteria. In a landmark systematic review and meta-analysis published by Jorm (2004) covering dozens of studies, the area under the receiver operating characteristic curve (AUC-ROC) for the IQCODE across memory clinics, community settings, and general hospital wards routinely fell between .85 and .95.

• Community Cohorts: In general population samples, the IQCODE typically exhibits an AUC of .84 to .90, with optimal cut-off scores ranging between 3.30 and 3.40 yielding sensitivity and specificity profiles of approximately 80% to 85%.
• Clinical and Hospital Populations: In acute medical or geriatric rehabilitation wards where prevalence is higher, cut-off scores between 3.40 and 3.60 yield sensitivity values between 82% and 93% and specificity values between 78% and 88%.
• Predictive Validity: Lewis et al. (1994) demonstrated that in acute general medical inpatients who did not meet formal diagnostic criteria for dementia at baseline, elevated IQCODE scores significantly predicted the subsequent incidence of clinical dementia over longitudinal follow-up. Furthermore, Jorm et al. (2000) validated informant ratings against true objective longitudinal change on neuropsychological test batteries measured over a 7- to 8-year span, proving that collateral reports track verified objective cognitive decline rather than merely informant mood.

Convergent and Concurrent Validity

The IQCODE correlates substantially with established direct cognitive screening batteries and clinical staging instruments:

• Correlations with the Mini-Mental State Examination (MMSE) consistently range from $r = -.50$ to $r = -.78$ in clinical samples (negative due to opposing scoring directions: higher IQCODE = worse decline; lower MMSE = worse impairment).
• Correlations with the Clinical Dementia Rating (CDR) scale and the Global Deterioration Scale (GDS) are exceptionally robust, frequently exceeding $r = .70$ to $.82$.
• Correlations with functional status assessments—specifically Lawton Instrumental Activities of Daily Living (IADL) and the Bristol Activities of Daily Living Scale—range between $r = .65$ and $r = .80$, substantiating that the IQCODE measures functional expressions of cognitive deterioration.

Discriminant Validity and Independence from Confounding Variables

A primary psychometric advantage of the IQCODE over direct cognitive tests is its documented independence from patient education, pre-morbid intellectual quotient (e.g., as estimated by the National Adult Reading Test / NART), and socioeconomic status. Multiple regression analyses by Jorm and Jacomb (1989) and subsequent multinational investigators confirmed that whereas MMSE scores exhibit substantial variance explained by years of education ($R^2$ frequently exceeding .15 to .25 in lower-education populations), IQCODE scores correlate negligibly with formal education ($r$ typically between -.05 and .08, non-significant). However, research highlights that informant characteristics—specifically informant depression, high caregiver burden, or personality traits—can introduce minor variance (accounting for approximately 2% to 5% of score variance), indicating that while patient demographic bias is eradicated, clinician awareness of caregiver distress is warranted.

8. Reliability

The IQCODE exhibits exceptional internal consistency and test-retest reliability across diverse cultural, linguistic, and clinical contexts.

Internal Consistency

Due to the high mutual inter-correlation among items measuring functional-cognitive decline, internal consistency estimates using Cronbach's alpha are uniformly high:

• Full 26-Item Version: Alpha coefficients across community and clinical samples consistently fall between $\alpha = .93$ and $\alpha = .97$ (e.g., Jorm & Jacomb, 1989; Jorm, 2004).
• Short 16-Item Version: In developing the 16-item short form, Jorm (1994) showed that reduction of items did not compromise reliability, with Cronbach’s alpha maintaining levels of $\alpha = .94$ to $.96$.
• Split-Half Reliability: Guttman split-half coefficients and Spearman-Brown reliability estimates routinely exceed $.90$.

Test-Retest and Inter-Rater Reliability

Temporal stability evaluated across intervals where true cognitive deterioration is negligible (e.g., several days to a few weeks) demonstrates high reproducibility:

• Test-Retest Stability: Jorm and Jacomb (1989) documented a test-retest correlation coefficient of $r = .96$ across a 2- to 3-day interval in a hospital sample, and $r = .75$ over a 1-year interval in a community sample of non-demented older adults (where true mild age-related variance naturally enters over 12 months).
• Inter-Rater Reliability: When two separate informants (e.g., a spouse and an adult child) evaluate the same elderly individual independently, inter-rater reliability correlations typically range between $r = .60$ and $r = .80$. Lower inter-rater agreement typically stems from differential exposure to the patient (e.g., co-residing spouse witnessing financial mismanagement versus a non-co-residing adult child visiting intermittently).

9. Factor Analysis

Extensive exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) have evaluated whether the IQCODE captures multidimensional sub-domains (e.g., memory vs. executive function vs. language) or operates as an essentially unidimensional scale.

Exploratory Factor Analyses (EFA)

Initial principal components and principal axis factoring analyses conducted by Jorm and Jacomb (1989) on the 26-item instrument revealed a dominant first unrotated general factor accounting for approximately 50% to 60% of the total variance, with an eigenvalue far exceeding that of any subsequent factor (eigenvalue 1 > 12.0; eigenvalue 2 < 1.8). While forced oblique or varimax rotations can yield secondary factor solutions—often segregating into (1) everyday memory/learning, (2) acquiring new skills/handling gadgets, and (3) handling money and arithmetic—these secondary factors correlate extremely highly with one another ($r > .65$ to $.80$). This led Jorm to conclude that the questionnaire is essentially unidimensional in clinical utility.

Confirmatory Factor Analyses (CFA) and Model Fit

Subsequent psychometric investigations employing structural equation modeling have tested one-factor, multi-factor, and bifactor models:

• Unidimensional Model Fit: A strictly unidimensional model frequently yields adequate to good fit indices when minor residual covariances between similarly phrased items (e.g., Item 23 and Item 24 regarding financial handling) are accounted for: Comparative Fit Index ($ ext{CFI}) > .94$, Tucker-Lewis Index ($ ext{TLI}) > .93$, and Root Mean Square Error of Approximation ($ ext{RMSEA}) < .065$.
• Bifactor Solutions: Bifactor modeling confirms that a general cognitive decline factor ($G$-factor) accounts for over 80% of the common variance ($ECV > .80$), with hierarchical omega ($\omega_h$) exceeding .90. Specific group factors (e.g., memory, verbal, numerical executive) capture negligible unique reliable variance, reaffirming that reporting a single aggregated mean item score is psychometrically justified.
• Factor Loadings: In both EFA and CFA models, standardized factor loadings across the 26 items are uniformly strong, almost all exceeding $lambda = .65$. The items demonstrating the highest factor loadings on general cognitive decline consistently include Item 4 (remembering recent events, $\lambda \approx .86$), Item 5 (recalling conversations, $\lambda \approx .84$), Item 14 (learning new things in general, $\lambda \approx .88$), and Item 26 (using intelligence to reason things through, $\lambda \approx .87$). The lowest, though still satisfactory, loadings are typically observed for Item 15 (remembering childhood events, $\lambda \approx .45-.55$) and Item 17 (understanding unusual words, $\lambda \approx .50-.60$).

10. Instrument / Measurement Tool

• Instrument Name: Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE).
• Primary Developer: Anthony F. Jorm (1989, 1994).
• Instrument Type: Collateral-report / Informant-rated psychometric rating scale.
• Format / Administration Route: Paper-and-pencil self-administered questionnaire, postal survey, in-person clinical interview, or standardized telephone administration. Can be administered verbally by a trained clinician or completed independently by the informant.
• Item Count:
  • Full Version: 26 items.
  • Short Form: 16 items (indicated with an asterisk in the scale items).
• Target Population: Older adults (typically aged 65 years and older) suspected of experiencing cognitive decline, mild cognitive impairment, or dementia, evaluated via an informant who has known them well for at least 10 years.
• Retrospective Time Frame: Explicitly evaluates the subject's current performance compared with 10 years ago.
• Response Scale (Anchors): 5-point Likert scale:
  • 1 = Much improved
  • 2 = A bit improved
  • 3 = Not much change
  • 4 = A bit worse
  • 5 = Much worse
• Scoring Rules:
  • Each item is scored from 1 to 5 based on the response.
  • Total Score Calculation: Sum the item scores and divide by the total number of items answered (producing a mean item score ranging from 1.00 to 5.00).
  • $$\text{Mean Score} = \frac{\sum \text{Item Scores}}{\text{Number of Answered Items}}$$
  • A mean score of exactly 3.00 indicates no perceived change overall compared to 10 years ago.
  • Mean scores greater than 3.00 indicate general cognitive decline.
  • Mean scores less than 3.00 indicate general cognitive improvement.
  • Handling Missing Data: If items are omitted (e.g., if an item is not applicable because the individual never handled financial matters or used household gadgets), the sum is divided by the number of valid items completed. If more than 3 to 5 items are missing on the 26-item version (or more than 2 items on the 16-item short form), the assessment should be interpreted with extreme caution or discarded.
• Clinical Cut-Off Thresholds:
  • General Community Screening: Scores between 3.30 and 3.40 balance sensitivity (~85%) and specificity (~80%) for detecting clinical dementia.
  • Geriatric / Hospital Inpatient Units: Scores between 3.40 and 3.60 optimize diagnostic accuracy, minimizing false positives in populations with high medical comorbidity.
  • Scores of 3.00 to 3.29 generally represent normal age-associated changes or absent functional decline.

11. Permissions & Fee and Test Year

The Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE) was initially published in 1989 by Anthony F. Jorm and Patricia A. Jacomb. In 1994, Professor Jorm published the validated 16-item short form. To facilitate global clinical assessment, geriatric research, and cross-cultural dementia screening, the author placed the IQCODE in the public domain for non-commercial clinical and academic research purposes.

Clinicians, researchers, and health organizations are permitted to reproduce, translate, and utilize both the 26-item full version and the 16-item short version without payment of royalty fees or formal copyright licensing charges, provided appropriate citation and academic attribution are maintained. The official versions and authorized translations are archived through academic repositories, health technology assessments, and clinical guides (e.g., Ian McDowell's Measuring Health: A Guide to Rating Scales and Questionnaires, Oxford University Press). Commercial organizations intending to bundle the IQCODE into proprietary commercial software suites or commercial clinical trial platforms should contact the author or the relevant institutional technology transfer office to confirm licensing terms.

12. References

Jorm, A. F. (1994). A short form of the Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE): Development and cross-validation. Psychological Medicine, 24(1), 145–153. https://doi.org/10.1017/s003329170002691x

Jorm, A. F. (2004). The Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE): A review. International Psychogeriatrics, 16(3), 275–293. https://doi.org/10.1017/s1041610204000390

Jorm, A. F., Christensen, H., Korten, A. E., Jacomb, P. A., & Henderson, A. S. (2000). Informant ratings of cognitive decline in old age: Validation against change on cognitive tests over 7 to 8 years. Psychological Medicine, 30(4), 981–985. https://doi.org/10.1017/s0033291799002275

Jorm, A. F., & Jacomb, P. A. (1989). The Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE): Socio-demographic correlates, reliability, validity and some norms. Psychological Medicine, 19(4), 1015–1022. https://doi.org/10.1017/s0033291700005742

Jorm, A. F., Scott, R., Cullen, J. S., & MacKinnon, A. J. (1991). Performance of the Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE) as a screening test for dementia. Psychological Medicine, 21(3), 785–790. https://doi.org/10.1017/s0033291700022418

Jorm, A. F., Scott, R., & Jacomb, P. A. (1989). Assessment of cognitive decline in dementia by informant questionnaire. International Journal of Geriatric Psychiatry, 4(1), 35–39. https://doi.org/10.1002/gps.930040107

Lewis, B., Harwood, D., Hope, T., & Jorm, A. F. (1994). Can an informant questionnaire be used to predict the development of dementia in medical patients? International Journal of Geriatric Psychiatry, 9(12), 941–945. https://doi.org/10.1002/gps.930091203

McDowell, I. (2006). Measuring Health: A Guide to Rating Scales and Questionnaires (3rd ed.). Oxford University Press. https://doi.org/10.1093/acprof:oso/9780195165678.001.0001

13. Items of the Scale (Questionnaire)