RAND Medical Outcomes Study – 20 Items

Abstract

The Medical Outcomes Study 20-Item Short-Form Health Survey (MOS SF-20) is a seminal, patient-reported outcome measure developed by the RAND Corporation and the Medical Outcomes Study research group under the scientific leadership of Anita L. Stewart, Ron D. Hays, and John E. Ware, Jr. Designed to evaluate generic health-related quality of life (HRQoL) with minimal respondent burden, the SF-20 operationalizes core dimensions of functional status and well-being derived from the World Health Organization’s multidimensional definition of health. The instrument consists of 20 items distributed across six core health concepts: Physical Functioning (6 items), Role Functioning (2 items), Social Functioning (1 item), Mental Health (5 items), Current Health Perceptions (5 items, including an initial general health rating), and Bodily Pain (1 item). Items utilize varying discrete response scales measuring limitation duration, frequency, severity, and categorical agreement. Psychometric evaluation across diverse clinical and general populations demonstrates high internal consistency (Cronbach’s alpha ranging from 0.76 to 0.88 for multi-item subscales) and strong construct, convergent, and discriminant validity. Scores are transformed to a standardized 0–100 scale, where higher values designate superior health status. As the direct empirical antecedent to the widely implemented SF-36 and SF-12 instruments, the MOS SF-20 remains a foundational benchmark in health econometrics, epidemiological surveys, chronic disease monitoring, and clinical trials where brief yet psychometrically rigorous generic health assessment is required.

Keywords

Medical Outcomes Study, MOS SF-20, Health-Related Quality of Life, RAND Corporation, Physical Functioning, Mental Health Inventory, Patient-Reported Outcome Measures, Psychometrics, Health Status Measurement, Functional Assessment

Authors

The MOS SF-20 was developed through the collaborative efforts of principal investigators in health services research and psychometrics within the RAND Corporation and the New England Medical Center:

• Anita L. Stewart, Ph.D. — Professor Emeritus at the Institute for Health & Aging, School of Nursing, University of California, San Francisco (UCSF); lead investigator on health status assessment methodologies in the Medical Outcomes Study.
• Ron D. Hays, Ph.D. — Distinguished Professor of Medicine at the David Geffen School of Medicine at UCLA, Affiliated Adjunct Behavioral Scientist at RAND; primary architect of scoring algorithms, reliability metrics, and construct validation protocols for RAND health surveys.
• John E. Ware, Jr., Ph.D. — Professor and Director of Outcomes Studies at the Department of Quantitative Health Sciences, University of Massachusetts Medical School; former Senior Scientist at RAND; pioneer of the Health Insurance Experiment (HIE) and developer of the Short-Form health profiling system.

Institutional affiliations historically responsible for the genesis of the tool include the RAND Corporation (Santa Monica, CA, USA) and the Health Institute at the New England Medical Center (Boston, MA, USA).

Purpose

The primary purpose of the MOS SF-20 is to provide a brief, psychometrically sound, self-administered questionnaire capable of quantifying multidimensional generic health status across both general populations and patients with varied chronic medical and psychiatric conditions. Originating during the large-scale Medical Outcomes Study in the mid-1980s, the instrument was engineered to address a critical methodological dilemma in clinical epidemiology: comprehensive health surveys (such as the original 108-item MOS functioning and well-being battery or the Sickness Impact Profile) imposed prohibitive respondent burden, resulting in missing data and survey fatigue, while ultra-brief single-item global indicators lacked the sensitivity, reliability, and multidimensionality necessary to detect subtle clinical differences or shifts over time.

The instrument fulfills several key objectives across healthcare research and clinical delivery:

• Population Health Monitoring: Providing normative epidemiological benchmarks to monitor health inequalities, chronic disease prevalence, and population-level functional health burdens.
• Comparative Effectiveness and Outcomes Research: Facilitating cross-disease comparisons of treatment effectiveness, therapeutic interventions, and pharmaceutical outcomes on a standardized metric, independent of specific pathology.
• Clinical Practice and Screening: Enabling rapid point-of-care assessment to detect unrecognized functional impairment, pain severity, or psychological distress in primary care and outpatient specialty settings.
• Health Economics and Policy Modeling: Supplying generic functional status inputs to assess disease burden, evaluate healthcare system redesigns, and inform health resource allocations.

Importantly, the SF-20 was explicitly designed to balance brevity (requiring an average completion time of only 3 to 5 minutes) with rigorous psychometric properties, ensuring that physical, emotional, social, and pain dimensions could be captured without sacrificing construct fidelity.

Psychological Construct

The MOS SF-20 operationalizes Health-Related Quality of Life (HRQoL) as a multidimensional construct encompassing objective functional performance, subjective self-evaluations, and affective well-being. Rather than conceptualizing health merely as the absence of disease, the construct mirrors the biopsychosocial framework by delineating six discrete, correlated dimensions:

1. Physical Functioning (6 Items)

Physical functioning measures the extent to which an individual’s health limits performance of basic, intermediate, and vigorous physical activities of daily living. Items assess severe functional limitations (eating, dressing, bathing, toileting), mobility limitations (walking one block, walking uphill, climbing stairs), intermediate physical tasks (bending, lifting, stooping, moderate tasks like moving a table or bowling), and strenuous activities (running, heavy lifting, engaging in strenuous sports). The construct distinguishes acute short-term deficits from chronic limitations by utilizing a temporal duration response metric (limited for more than 3 months, limited for 3 months or less, not limited at all).

2. Role Functioning (2 Items)

Role functioning quantifies the degree to which physical and emotional health status impedes an individual’s capacity to fulfill primary social and economic roles, including employment, housekeeping, or academic schooling. The items assess both complete cessation of role activities (keeping the individual from working or attending school) and qualitative restrictions (inability to do certain kinds or amounts of work or chores) over acute versus chronic time horizons.

3. Social Functioning (1 Item)

Social functioning captures the frequency and magnitude with which physical or mental health problems interfere with normal social interactions, interpersonal relationships, and community engagements (e.g., visiting friends or close relatives). Evaluated over a 4-week recall window on a 6-point continuum from “All of the time” to “None of the time,” this dimension gauges the social isolation and interpersonal disruption induced by ill health.

4. Mental Health (5 Items)

Derived from the validated Mental Health Inventory (MHI-5), this subscale evaluates psychological well-being and affective distress during the preceding month. The construct incorporates two opposing affective poles: psychological distress (manifested through anxiety [“very nervous person”] and depressive symptomatology [“downhearted and blue,” “felt so down in the dumps that nothing could cheer you up”]) and positive psychological well-being (manifested through tranquility [“calm and peaceful”] and general happiness [“happy person”]).

5. Current Health Perceptions (5 Items)

Current health perceptions reflect the respondent’s subjective, cognitive evaluation of their overall health status. This incorporates a general rating item (evaluating health from “Excellent” to “Poor”) and four Likert-type statements examining perceived vulnerability, health stability, and subjective illness (“I am somewhat ill,” “I am as healthy as anybody I know,” “My health is excellent,” “I have been feeling bad lately”). This dimension captures prognostic self-perceptions that often predict future healthcare utilization and mortality beyond objective clinical biomarkers.

6. Bodily Pain (1 Item)

Bodily pain measures the perceived intensity and severity of physical discomfort experienced over the past 4 weeks, graded along a 6-point continuum from “None” to “Very Severe.” It reflects somatic symptom burden and interference directly attributable to physical pain.

Theoretical Framework

The conceptual architecture of the MOS SF-20 is rooted in the World Health Organization (WHO) 1948 constitution, which defined health as “a state of complete physical, mental, and social well-being and not merely the absence of disease or infirmity.” The operationalization of this paradigm within the RAND Health Insurance Experiment (HIE) and subsequently the Medical Outcomes Study transformed this philosophical definition into standardized, psychometrically verifiable constructs.

The theoretical framework rests upon several foundational pillars:

• The Generic Health Profiling Model: Stewart, Ware, and Hays postulated that health outcomes must be evaluated via generic profiles capable of functioning across all clinical conditions, demographic strata, and treatment modalities. Generic profiling ensures comparability across diverse diagnostic groups (such as hypertension, diabetes, coronary heart disease, and major depressive disorder), permitting researchers to quantify the relative functional impact of dissimilar pathologies.
• The Functioning and Well-Being Paradigm: The framework explicitly bifurcates health outcomes into observable behavioral functioning (what an individual can physically and socially do) and subjective well-being (how an individual internally feels). Physical and role functioning reflect behavioral capacity, whereas mental health, bodily pain, and health perceptions reflect subjective internal states.
• Classical Test Theory (CTT) and Scaling Assumptions: The instrument assumes that latent health attributes can be captured using multi-item summated rating scales. Scale construction adheres to the criterion of summated ratings (Likert method), where items within a postulated dimension demonstrate roughly equal variances, linear associations with the underlying latent trait, and substantial item-convergent validity.

Validity

The psychometric validity of the MOS SF-20 has been extensively evaluated across thousands of medical outpatients, psychiatric cohorts, and general population surveys.

Construct and Convergent Validity

Construct validity was established through multitrait-multimethod correlation matrices during the Medical Outcomes Study validation trials (Stewart et al., 1988). Scale items correlate substantially higher with their hypothesized dimensions than with competing dimensions, confirming item-convergent validity (correlations generally exceeding 0.50 after correcting for item-scale overlap). Convergent validity is evidenced by strong correlations between SF-20 Physical Functioning and external functional measures, such as the Karnofsky Performance Scale ($r = 0.65$ to $0.75$) and physician ratings of functional capacity ($r > 0.55$). Similarly, the SF-20 Mental Health subscale correlates strongly ($r > 0.70$) with established psychiatric instruments, including the Beck Depression Inventory (BDI) and the Center for Epidemiologic Studies Depression Scale (CES-D).

Discriminant and Known-Groups Validity

The SF-20 reliably differentiates among patient populations with varying disease burdens. In seminal validation trials (Stewart et al., 1989; Wells et al., 1989):

• Patients with serious chronic medical conditions (e.g., congestive heart failure, chronic obstructive pulmonary disease) scored significantly lower on Physical Functioning ($p < 0.001$) compared to healthy controls or patients with mild conditions (e.g., uncomplicated hypertension).
• Patients diagnosed with clinical depression showed marked, selective decrements in the Mental Health subscale (mean scores below 50), while physical functioning remained relatively preserved unless complicated by medical comorbidity.
• The instrument effectively discriminated levels of disease severity: patients classified as New York Heart Association (NYHA) functional class III/IV had Physical Functioning scores averaging 30 to 40 points lower than NYHA class I patients.

Predictive and Criterion Validity

Longitudinal investigations confirm that lower baseline scores on the SF-20 Current Health Perceptions and Physical Functioning subscales independently predict one-year and five-year mortality, hospitalization rates, and primary care utilization, controlling for age, biological markers, and physician-assessed comorbidity indices.

Reliability

The reliability of the MOS SF-20 has been demonstrated using both internal consistency and test-retest methodologies across diverse clinical and community samples.

Internal Consistency

Cronbach’s alpha ($lpha$) coefficients published in the definitive MOS psychometric studies (Stewart et al., 1988; McHorney et al., 1992) consistently surpass the accepted threshold of 0.70 for group comparisons:

• Physical Functioning (6 items): $lpha = 0.86$ to $0.88$
• Role Functioning (2 items): $lpha = 0.81$ to $0.87$
• Mental Health (5 items): $lpha = 0.84$ to $0.88$
• Current Health Perceptions (5 items): $lpha = 0.76$ to $0.86$

Single-item scales (Social Functioning and Bodily Pain) cannot be evaluated via internal consistency metrics; however, their reliability was confirmed through alternate-form correlations with expanded parent batteries from the MOS, demonstrating single-item reliability coefficients between 0.65 and 0.76.

Test-Retest Stability

Short-term test-retest reliability across 2-week to 4-week intervals among clinically stable chronic outpatients yields intraclass correlation coefficients (ICCs) ranging from 0.73 (Current Health Perceptions) to 0.85 (Physical Functioning). Floor and ceiling effects are well-documented: while the Physical Functioning subscale exhibits minor ceiling effects in robust young adult samples, it provides fine-grained discriminative power among geriatric and chronically ill clinical populations.

Factor Analysis

Both exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) corroborate the theoretical architecture of the SF-20, revealing a robust multidimensional structure organized around two overarching higher-order domains: Physical Health and Mental Health.

Exploratory Factor Structure

Principal components analysis with orthogonal (Varimax) and oblique (Promax) rotations consistently extracts distinct factors corresponding to the multi-item subscales. When a two-factor solution is forced:

• Factor 1 (Physical Health Component): Highly loaded by Physical Functioning ($0.75 – 0.85$), Role Functioning ($0.65 – 0.78$), and Bodily Pain ($0.60 – 0.72$).
• Factor 2 (Mental Health Component): Highly loaded by Mental Health items ($0.78 – 0.88$) and Social Functioning ($0.58 – 0.70$).
• Cross-Loading: Current Health Perceptions loads moderately across both factors ($0.45 – 0.55$ on Physical, $0.40 – 0.50$ on Mental), empirically confirming its role as an integrative, holistic health indicator.

Confirmatory Factor Analysis and Model Fit

Modern structural equation modeling (SEM) confirms the adequacy of the six-factor first-order model and the hierarchical two-factor higher-order model across national cohorts. Typical goodness-of-fit indices reported in literature include:

• Comparative Fit Index (CFI): $0.93 – 0.96$
• Tucker-Lewis Index (TLI): $0.92 – 0.95$
• Root Mean Square Error of Approximation (RMSEA): $0.048 – 0.062$ ($90%\text{ CI } [0.043, 0.067]$)
• Standardized Root Mean Square Residual (SRMR): $0.041 – 0.054$

These findings substantiate that the 20 items capture discrete yet interdependent facets of functional health without unacceptable structural redundancy.

Instrument / Measurement Tool

• Test Type: Generic Health-Related Quality of Life (HRQoL) Profile / Patient-Reported Outcome Measure (PROM).
• Format: Self-administered paper-and-pencil or digital questionnaire; also adaptable for trained interviewer-assisted administration.
• Number of Items: 20 items.
• Subscale Breakdown:
  • Physical Functioning: 6 items (Items 2a, 2b, 2c, 2d, 2e, 2f)
  • Role Functioning: 2 items (Items 4, 5)
  • Social Functioning: 1 item (Item 6)
  • Mental Health: 5 items (Items 7, 8, 9, 10, 11)
  • Current Health Perceptions: 5 items (Items 1, 12a, 12b, 12c, 12d)
  • Bodily Pain: 1 item (Item 3)
• Response Formats: Categorical options tailored by construct, including 3-point duration limitations, 6-point pain severity, 6-point frequency ratings, and 5-point Likert true/false scales.
• Scoring and Transformation Rules:
  • Step 1 (Recoding / Reverse Scoring): Items phrased in opposite directions must be recoded prior to aggregation so that higher numerical values systematically represent better health. For instance, in Item 1 (General Health), “Excellent” is scored 5 and “Poor” is scored 1. In negative mental health items (Items 7, 9, 11), “None of the time” receives 6 points, whereas “All of the time” receives 1 point. In positive items (Items 8, 10), “All of the time” receives 6 points.
  • Step 2 (Raw Scale Calculation): For each multi-item scale, the raw score is calculated as the algebraic mean of the completed items within that subscale.
  • Step 3 (Linear Transformation to 0–100 Scale): Raw subscale scores are transformed linearly to a standard range from 0 (poorest possible health / maximum limitation) to 100 (optimal health / no limitation) using the standard RAND transformation formula:
    $$\text{Transformed Score} = \frac{\text{Actual Raw Score} – \text{Lowest Possible Raw Score}}{\text{Possible Raw Score Range}} \times 100$$
• Administration Time: Approximately 3 to 5 minutes.

Permissions & Fee and Test Year

• Publication Year: Originally published in 1988 by Anita L. Stewart, Ron D. Hays, and John E. Ware, Jr.
• Copyright and Licensing: The MOS SF-20 was developed as part of the public domain research initiative within the RAND Health Program. Unlike proprietary commercial revisions (such as SF-36v2 or SF-12v2 licensed by QualityMetric), the core RAND MOS instruments are available in the public domain for research, academic, and clinical purposes without royalty fees or licensing costs.
• Usage Conditions: Researchers are free to utilize, administer, and reproduce the MOS SF-20 provided proper academic attribution is credited to the RAND Corporation and original Medical Outcomes Study authors. Alteration of wording, item order, or response categories without explicit documentation is discouraged to preserve psychometric comparability against historical normative datasets.

References

• Bennett, S. J., Oldridge, N. B., Eckert, G. J., Embree, J. L., Browning, S., Hou, N., & Deer, M. (2002). Discriminant properties of commonly used quality of life measures in heart failure. Quality of Life Research, 11(4), 349–359. https://doi.org/10.1023/A:1015509923830
• 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
• McHorney, C. A., Ware, J. E., Jr., Rogers, W., Raczek, A. E., & Lu, J. F. (1992). The validity and relative precision of MOS short- and long-form health status scales and Dartmouth COOP charts to discriminate among levels of medical and psychiatric sickness. Medical Care, 30(5), MS253–MS265. https://doi.org/10.1097/00005650-199205001-00025
• Stewart, A. L., Greenfield, S., Hays, R. D., Wells, K., Rogers, W. H., Berry, S. D., McGlynn, E. A., & Ware, J. E., Jr. (1989). Functional status and well-being of patients with chronic conditions: Results from the Medical Outcomes Study. JAMA, 262(7), 907–913. https://doi.org/10.1001/jama.1989.03430070055030
• Stewart, A. L., Hays, R. D., & Ware, J. E., Jr. (1988). The MOS Short-Form General Health Survey: Reliability and validity in a patient population. Medical Care, 26(7), 724–735. https://doi.org/10.1097/00005650-198807000-00007
• Stewart, A. L., Hays, R. D., & Ware, J. E., Jr. (1992). Health perceptions, energy/fatigue, and health distress measures. In A. L. Stewart & J. E. Ware, Jr. (Eds.), Measuring functioning and well-being: The Medical Outcomes Study approach (pp. 143–172). Duke University Press.
• Ware, J. E., Jr., Kosinski, M., & Keller, S. D. (1996). A 12-Item Short-Form Health Survey: Construction of scales and preliminary tests of reliability and validity. Medical Care, 34(3), 220–233. https://doi.org/10.1097/00005650-199603000-00003
• Wells, K. B., Stewart, A., Hays, R. D., Burnam, M. A., Rogers, W., Daniels, M., Berry, S., Greenfield, S., & Ware, J. (1989). The functioning and well-being of depressed patients: Results from the Medical Outcomes Study. JAMA, 262(7), 914–919. https://doi.org/10.1001/jama.1989.03430070062031

13. Items of the Scale (Questionnaire)