Health Opinion Survey (HOS) – 27 items

Abstract

The Health Opinion Survey (HOS) is a pioneering 27-item self-report screening instrument designed to quantify psychiatric impairment, psychological distress, and psychoneurotic symptom load in community, clinical, and primary care populations. Originally conceptualized in the mid-1950s within the landmark Stirling County Study led by Alexander H. Leighton and methodologically codified by Allister M. Macmillan (1957), the instrument was subsequently standardized and validated in European primary care settings by Adrian M. Semmence (1969). The HOS primarily assesses autonomic nervous system arousal, psychophysiological reactivity, affective exhaustion, and somatic manifestations of anxiety and depressive disorders. Comprising 27 items calibrated across heterogeneous polytomous response categories—predominantly three-point Likert-type scales (0 = No/Never/Not at all, 1 = Sometimes/Undecided, 2 = Yes/Often/A lot) alongside specific four-point ordinal metrics for episodic distress—the survey yields a composite score reflecting generalized psychoneurotic vulnerability.

Extensive psychometric investigations have established that the HOS exhibits moderate-to-high internal consistency reliability (Cronbach’s alpha ranging between .78 and .88 across diverse epidemiological samples) and stable test-retest reliability over short-to-medium diagnostic intervals (ranging from .74 to .86). Construct and criterion-related validity are substantiated through significant discriminative accuracy in differentiating psychiatric cases from healthy community cohorts. However, empirical literature highlights critical caveats regarding the scale’s sensitivity to organic physical illness, as physiological conditions frequently elevate scores via somatic symptom overlap. Exploratory and confirmatory factor analyses demonstrate a multidimensional architecture dominated by a strong general psychophysiological distress factor, complemented by distinct sub-dimensions including autonomic dysfunction, musculoskeletal/sensory paresthesia, sleep-fatigue disturbances, and hypochondriacal anxiety. This article provides a comprehensive psychometric review of the 27-item HOS, detailing its historical derivation, theoretical architecture, statistical parameters, clinical diagnostic utility, and full item specifications.

Keywords

Health Opinion Survey, HOS, Stirling County Study, psychiatric epidemiology, psychoneurotic screening, somatic symptoms, autonomic dysfunction, psychometrics, anxiety, depression

Authors

The foundational development of the Health Opinion Survey was spearheaded by Allister M. Macmillan, Ph.D., working under the scientific direction of Alexander H. Leighton, M.D., within the Department of Sociology and Anthropology and the Department of Psychiatry at Cornell University. The instrument served as the primary quantitative screening tool for the longitudinal psychiatric epidemiology initiative known as the Stirling County Study of Psychiatric Disorder and Sociocultural Environment in Nova Scotia, Canada.

The 27-item standardized general practice formulation was adapted, operationalized, and clinically evaluated within British National Health Service (NHS) primary care environments by Adrian M. Semmence, M.D., FRCGP, affiliated with the Royal College of General Practitioners. Inquiries regarding original historical archives are preserved within academic holdings of the Stirling County Study collections and the published proceedings of the Journal of the Royal College of General Practitioners.

Purpose

The fundamental purpose of the Health Opinion Survey is to serve as a rapid, reliable, and cost-effective screening device capable of identifying non-psychotic psychiatric disorder, functional psychoneurosis, and disabling psychophysiological distress. Developed during an era when psychiatric epidemiology lacked standardized diagnostic interviews such as the DIS, CIDI, or SCID, the HOS was engineered to circumvent the resource-intensive necessity of conducting exhaustive clinical psychiatric evaluations on every individual within large-scale demographic populations.

In clinical practice and primary care epidemiology, the HOS addresses several distinct diagnostic and public health requirements:

• Community Mental Health Surveillance: The scale enables macro-level epidemiological investigations to compute the prevalence of psychological impairment across diverse socio-environmental conditions, facilitating comparisons between integrated and socially disintegrated communities.
• Primary Care Triage: In ambulatory medicine and general practice, patients regularly present with vague somatic complaints—such as tension headaches, dyspepsia, tremor, palpitations, and chronic morning fatigue—that lack discernible organic etiology. The HOS helps family physicians detect underlying affective and neurotic disturbances masquerading as somatic pathology.
• Stratification in Clinical Research: The instrument provides a continuous dimensional metric of psychosomatic strain, allowing behavioral scientists to assess treatment efficacy, quantify stress reactions, and evaluate psychological vulnerability in psychosomatic medicine protocols.
• Risk Factor Analysis: By assessing the intersection of physical lifestyle factors (such as heavy smoking and patent medicine reliance) with autonomic nervous dysregulation, the HOS offers insights into how chronic distress compromises general occupational and physiological functioning.

The instrument was deliberately not intended to classify acute psychosis, thought disorders, or specific diagnostic categories of the modern Diagnostic and Statistical Manual of Mental Disorders (DSM). Instead, its focus is anchored in quantifying continuous “neurotic distress” and psychophysiological decompensation, operating under the premise that chronic emotional stress translates predictably into measurable vegetative and functional somatic complaints.

Psychological Construct

The core construct measured by the Health Opinion Survey is psychoneurotic impairment expressed predominantly through somatization and vegetative nervous system hyperarousal. Rather than focusing exclusively on cognitive rumination or affective despair, the construct operationalizes distress through the physical endpoints of the stress response. Historically, this construct emerged from the diagnostic traditions of Adolf Meyer’s psychobiology, which conceptualized mental disorders not as isolated brain diseases, but as comprehensive psychobiological reactions of the organism attempting to adapt to environmental pressures.

The construct encompasses several interrelated psychophysiological and behavioral domains:

• Autonomic Nervous System Reactivity: Characterized by peripheral sympathetic activation and vascular instability. Manifestations include clammy or cold sweating palms and feet (Item 3), hard or rapid heart palpitations (Item 5), unprovoked shortness of breath (Item 26), and vasomotor spells such as dizziness or feeling faint (Item 23).
• Motor and Neuromuscular Tension: Represented by visible motor oscillations, peripheral paresthesias, and general tension. This includes bothersome trembling of the hands (Item 2) and limbs falling asleep or feeling numb rather easily (Item 10).
• Gastrointestinal and Appetitive Disturbance: Dysregulation of the enteric nervous system and digestive neurochemistry. Evaluated through recurrent upset stomach (Item 8), sudden or chronic loss of appetite (Item 15), dysgeusia or an unpleasant taste in the mouth (Item 16), and functional dysphagia where food feels tasteless and difficult to swallow (Item 17).
• Sleep Architecture and Diurnal Energy Rhythm: Chronic sleep disruption reflecting hyperarousal and early vegetative depletion. Manifestations include initial and middle insomnia (Item 7), terrifying dreams or nightmares (Item 9), and persistent morning exhaustion unalleviated by sleep (Item 6).
• Generalized Hypochondriacal and Health Anxiety: Hypervigilance regarding somatic integrity and vulnerability to physical breakdown. Measured by the conviction of suffering from disparate ailments across multiple anatomical regions (Item 12), hypersensitivity to contracting contagious diseases relative to peers (Item 20), recurrent “weak turns” (Item 22), and reliance on commercial patent medicines (Item 19) or unnecessary vitamin regimens (Item 18).
• Subjective Emotional Demoralization: Explicit affective awareness of failing psychological coping mechanisms. Captured through direct acknowledgment of pervasive nervousness (Item 25), stress-induced acute weight loss (Item 24), functional occupational impairment due to health perceptions (Item 21), and the subjective sensation of an impending “nervous breakdown” (Item 4).

Theoretical Framework

The theoretical architecture of the Health Opinion Survey is rooted in mid-twentieth-century psychosomatic medicine, stress theory, and psychiatric epidemiology. The conceptual underpinnings reflect a convergence of three major scientific paradigms:

1. Selye’s General Adaptation Syndrome (GAS)

Hans Selye’s formulations of biological stress posit that systemic stressors elicit a stereotyped physiological response involving the hypothalamic-pituitary-adrenal (HPA) axis and the autonomic nervous system. The HOS was constructed under the assumption that protracted psychological and socio-environmental conflict induces sustained physiological activation—the stage of resistance—ultimately progressing toward somatic exhaustion. The physical manifestations queried by the HOS (e.g., gastrointestinal distress, cold sweats, dyspnea, and peripheral vascular changes) are operational indicators of physiological strain resulting from prolonged psychological stressors.

2. The Stirling County Socio-Environmental Model

Alexander Leighton’s socio-cultural theory asserted that human beings possess fundamental psychological needs for physical security, sexual satisfaction, the expression of hostility, expression of love, securing love, securing recognition, expression of creativity, and a sense of belonging to a moral order. When an individual’s sociocultural environment undergoes profound disintegration (e.g., poverty, cultural confusion, family instability, and rapid economic collapse), these fundamental strivings are systematically frustrated. This chronic frustration produces pervasive, low-grade psychiatric distress that crystallizes into functional neurotic symptoms. Macmillan designed the HOS specifically to operationalize this latent state of environmental-psychological maladaptation.

3. The Psychobiological Integration of Adolf Meyer

Adolf Meyer’s dynamic psychobiology emphasized that mind and body constitute an indivisible biological unit. Neurological and psychiatric pathologies are not viewed strictly as discrete categorical entities, but as integrated functional reactions of the whole person. Consequently, neurotic impairment rarely exists purely in the psychological realm; it manifests through vegetative, gastrointestinal, muscular, and sleep pathologies. The HOS was engineered to assess this holistic psychobiological reaction pattern rather than categorical nosological syndromes.

Validity

The psychometric validity of the Health Opinion Survey has been evaluated through clinical discrimination studies, factor analyses, and cross-instrument convergent correlations spanning several decades.

Criterion and Discriminant Validity

In the seminal validation work conducted during the Stirling County Study (Macmillan, 1957), the HOS was administered to established psychiatric clinic patients and matched community non-patient controls. Receiver operating characteristics demonstrated that the instrument differentiated known psychiatric patients from asymptomatic controls with high statistical significance (p < .001). Patients categorized by independent psychiatrists as experiencing moderate-to-severe neurosis achieved significantly elevated aggregate HOS scores relative to the healthy general population.

However, in a notable clinical investigation by Semmence (1969) in a UK general practice setting involving 382 consecutive adult attendees, critical boundary conditions of criterion validity were illuminated. While the HOS successfully separated patients designated by general practitioners as exhibiting clear neurotic conditions from healthy controls, a profound confounding factor emerged: organic physical pathology. Patients suffering from documented chronic physical diseases (e.g., cardiovascular disease, chronic obstructive pulmonary disease, rheumatological conditions) without primary psychopathology scored nearly as high as patients with diagnosed psychoneuroses. Semmence reported that using conventional cut-off scores, approximately 45% of patients with genuine physical pathology were misclassified by the HOS as psychiatrically impaired. Thus, while the HOS has high sensitivity for psychoneurotic states, its specificity is reduced in medical populations characterized by high physical morbidity.

Convergent and Concurrent Validity

The HOS displays moderate to strong convergent validity when compared against concurrent measures of emotional distress, psychological symptomatology, and anxiety inventories:

• Correlations with the Langner 22-Item Index (a closely related screening instrument developed for the Midtown Manhattan Study) consistently exceed r = .75, indicating high construct convergence.
• Correlations with the neuroticism dimensions of the Eysenck Personality Inventory (EPI) typically fall in the range of r = .60 to .72.
• Comparisons against modern general distress measures, such as the 12-item General Health Questionnaire (GHQ-12) and the Hopkins Symptom Checklist (HSCL), show robust linear associations (ranging from r = .58 to .69).

Reliability

The reliability parameters of the Health Opinion Survey have demonstrated consistent temporal stability and internal structural coherence across both community-wide surveys and medical outpatients.

Internal Consistency

Across empirical cohorts, the full 27-item scale demonstrates acceptable to good internal consistency. Published studies evaluating general population samples report Cronbach’s alpha coefficients routinely falling between α = .78 and .88. In homogeneous, non-clinical adult samples, internal consistency has occasionally stabilized near .82. Item-total correlations for classic autonomic and anxiety items—such as Item 25 (nervousness), Item 4 (nervous breakdown), Item 2 (trembling hands), and Item 5 (beating heart)—demonstrate strong discriminatory power, with corrected item-total correlation values ranging from r = .45 to .64. Conversely, lifestyle-oriented and behavioral items, such as Item 13 (smoking frequency), Item 18 (vitamin consumption), and Item 19 (reliance on patent medicines), demonstrate considerably lower item-total correlations (frequently between r = .15 and .28), indicating they introduce non-redundant, multidimensional variance.

Test-Retest Stability

Temporal stability evaluations indicate that the HOS measures a relatively stable dimensional construct of chronic psychophysiological strain rather than fleeting transient states. Test-retest reliability across a 2- to 4-week interval has yielded Pearson product-moment correlations and intraclass correlation coefficients (ICCs) between r = .74 and .86 in stable outpatient cohorts. Over longer longitudinal intervals (e.g., 6 to 12 months in community epidemiological surveys), test-retest coefficients attenuate toward the .60 to .68 range, reflecting naturalistic fluctuations in environmental stress, life events, and recovery from acute neurotic episodes.

Factor Analysis

Although the original authors conceptualized the HOS as a unidimensional index of neurotic impairment, multiple exploratory and confirmatory factor analytic (EFA/CFA) investigations have revealed an underlying multidimensional architecture. Early exploratory extractions utilizing principal axis factoring and varimax rotations systematically identified between three and five distinct latent factors accounting for approximately 42% to 56% of total item variance.

Latent Factor Structure

• Factor 1: Psychophysiological & Autonomic Arousal (Explains ~24% of variance). Characterized by high factor loadings (> .55) from Item 2 (trembling hands), Item 3 (clammy hands/feet), Item 5 (hard beating heart), Item 11 (cold sweats), Item 23 (dizziness), Item 25 (nervousness), and Item 26 (shortness of breath).
• Factor 2: Asthenia, Sleep Pathology & Demoralization (Explains ~11% of variance). Features prominent loadings (> .50) from Item 6 (morning tiredness), Item 7 (insomnia), Item 9 (nightmares), Item 4 (impending nervous breakdown), and Item 21 (ill health affecting work performance).
• Factor 3: Gastrointestinal & Somatoform Dysregulation (Explains ~8% of variance). Encompasses Item 8 (upset stomach), Item 15 (loss of appetite), Item 16 (bad taste in mouth), and Item 17 (food tasteless/hard to swallow).
• Factor 4: Hypochondriacal Coping & Somatic Vulnerability (Explains ~6% of variance). Marked by loadings from Item 12 (multiple bodily ailments), Item 18 (vitamin reliance), Item 19 (patent medicines), Item 20 (contagion anxiety), and Item 22 (weak turns).

Modern Confirmatory Factor Analysis (CFA) models comparing a strictly unidimensional model against a hierarchically nested, bifactor structure generally confirm that a bifactor model provides superior fit indices (e.g., Comparative Fit Index [CFI] > .92, Root Mean Square Error of Approximation [RMSEA] < .05). The bifactor configuration suggests the presence of a strong common general factor of “General Neurotic Distress” that runs through all items, alongside distinct, orthogonal group factors representing specific somatic systems.

Instrument / Measurement Tool

The Health Opinion Survey (27 items) is formatted as an interviewer-administered or self-administered psychometric questionnaire. Below are the operational specifications of the measurement tool:

• Construct Measured: Psychoneurotic distress, autonomic nervous system hyperarousal, and functional somatic impairment.
• Target Population: Adults (aged 18 and older) in primary care, clinical outpatient, and general community settings.
• Administration Format: Standard paper-and-pencil self-report or structured epidemiological interview.
• Administration Time: Approximately 5 to 10 minutes.
• Total Number of Items: 27 distinct items.
• Item Formats: Categorical and ordinal polytomous inquiries querying the frequency and presence of somatic, vegetative, and psychological symptoms.
• Response Categories:
  • Standard Three-Point Items (25 items): Most questions employ a frequency or presence gradient: No / Never / Not at all (0 points), Sometimes / Undecided (1 point), and Yes / Often / A lot (2 points).
  • Four-Point Ordinal Items (Items 4 and 8): Question 4 (“Have you ever felt you were going to have a nervous breakdown?”) uses: Never (0), Once or twice (1), Sometimes (2), Often (3). Question 8 (“How often are you bothered by having an upset stomach?”) uses: Never (0), Not very much (1), Pretty often (2), Nearly all the time (3).
  • Reverse Coding: Item 27 (“For the most part, do you feel healthy enough to carry out the things that you would like to do?”) captures functional vitality and is reverse-scored (Never = 2, Sometimes = 1, Often = 0).
• Scoring and Quantification:
  • Aggregate scores are computed by summing numeric values assigned to each response.
  • Theoretical total score ranges from 0 to 56 (or 27 to 83 depending on historical linear transformations where scores were mapped to 1-2-3 scales). Under the standard zero-indexed scoring scheme, higher aggregate scores reflect elevated psychoneurotic symptom loads.
  • Clinical Cut-off Scores: Historically, scores above the upper quartile (or standard deviations exceeding 1.5 above the community mean) served as diagnostic flags prompting comprehensive psychiatric evaluation. Semmence identified significant diagnostic cross-over at intermediate scores, emphasizing that scores in medical environments must be interpreted in conjunction with objective physical diagnostic assessments.

Permissions & Fee and Test Year

The original development of the Health Opinion Survey commenced under Allister M. Macmillan and Alexander H. Leighton in 1957 as part of the Stirling County Study. The standardized 27-item British medical adaptation was published by Adrian M. Semmence in 1969 in the Journal of the Royal College of General Practitioners.

The 27-item instrument resides in the public academic domain as historical scientific literature and is freely accessible for scholarly, clinical, and non-commercial research purposes without proprietary licensing fees. When utilizing the scale in empirical protocols or clinical publications, proper academic attribution and citation of Macmillan (1957) and Semmence (1969) are required.

References

Leighton, A. H. (1959). My name is legion: Foundations for a theory of man in relation to his culture (The Stirling County Study of Psychiatric Disorder & Sociocultural Environment, Vol. 1). Basic Books.

Leighton, D. C., Harding, J. S., Macklin, D. B., Macmillan, A. M., & Leighton, A. H. (1963). The character of danger: Psychiatric symptoms in selected communities (The Stirling County Study of Psychiatric Disorder & Sociocultural Environment, Vol. 3). Basic Books.

Macmillan, A. M. (1957). The Health Opinion Survey: Technique for estimating prevalence of psychoneurotic and related types of disorder in communities. Psychological Reports, 3(2), 325–339. https://doi.org/10.2466/pr0.1957.3.2.325

Murphy, J. M. (1980). Continuities in community-based psychiatric epidemiology: The Stirling County Study. American Psychologist, 35(4), 355–364. https://doi.org/10.1037/0003-066X.35.4.355

Semmence, A. M. (1969). The Health Opinion Survey: A psychiatric screening instrument. Journal of the Royal College of General Practitioners, 18(89), 344–348. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2237123/

Tousignant, M., Denis, G., & Lachapelle, R. (1974). Some considerations on the validity of the Health Opinion Survey. Journal of Health and Social Behavior, 15(3), 241–252. https://doi.org/10.2307/2137025

13. Items of the Scale (Questionnaire)