Textbook of Psychiatry/Diagnosis & Classification
This chapter explains what is meant by a psychiatric diagnosis, methods for making diagnoses, and aspects of diagnostic reliability, validity, and utility. Psychiatric and somatic comorbidities are elucidated. It includes a section on the influence of traditional medicine for most of the world’s population. It provides an overview of diagnostic interviews and screening questionnaires.
Historical development of psychiatric diagnoses[edit| edit source]
What is a diagnosis? The word stems from dia (Greek) meaning through and gnosis (Greek) meaning knowledge, or the establishing of the nature of a disease. Making diagnoses is as old as medical history.
Diagnoses described in ancient times still hold, for example clinical depression was described by Aretaeus (81-138), who practiced medicine in Rome and Alexandria. The physician Ibn Zohr-Avenzoar (1092-1162) in Morocco described in his clinical treatment guideline acute delirium, melancholia and dementia among other psychiatric disorders, and also reported the first known account of suicide in melancholics. In 1286, Le Maristane (hospital) Sidi Frej was built in Fes, Morocco, for psychiatric patients, and was a model for the first mental asylum in the western world, in Valencia, Spain, in 1410.
The term neurosis was created by the Scottish neurologist William Cullen in 1769 to label patients with nervous symptoms that occurred without an obvious organic cause. Chronic alcoholism was described by Magnus Huss in Stockholm in 1849. The German psychiatrist and neuropathologist Wilhelm Griesinger (1817-1868) laid the modern foundation of psychiatric classification in 1845, publishing a monograph on diseases of the brain. He proposed a unitary concept of psychosis. Subsequently Emil Kraepelin in Munich (1856-1926), the forefather of contemporary scientific psychiatry, split this unitary concept of psychosis into two distinct forms, based on symptom patterns, that he called manic depression and dementia praecox. The Swiss psychiatrist Eugen Bleuler (1857-1939) renamed the latter schizophrenia, having determined that this disorder did not necessarily progress to dementia.
French psychiatrists made important early contributions to psychiatric diagnoses. Tourette’s syndrome was first described in 1885 by the neurologist Giles de la Tourette (1857-1904). He also described anorexia nervosa in 1890. Paul Hartenberg (1871-1949) eloquently described social anxiety disorder in his monograph Les Timides et la timidité in 1901.
After the second world war, the validity of psychiatric diagnoses were questioned by the United States military, since many recruits had been considered unfit for soldier duty by psychiatrists. Many combat soldiers were discharged on psychiatric grounds. There was no consensus on how to make psychiatric diagnoses. In the absence of an agreed classification, epidemiological research was not possible.
There were many thought leaders on the merits of making diagnoses. Sigmund Freud (1856-1939) postulated unconscious conflicts as the source of mental ill health, while the Swiss-born psychiatrist Adolf Meyer (1866-1950), influential in the United States, advocated that such ill health was a personality reaction to psychological, social, and biological factors. In Scotland, Ronald Laing (1927-1989) launched the "antipsychiatric" idea in 1955 that psychosis was a reaction to a cold family environment that produced a false "id," and he popularized an example case history which described a "schizophrenogenic mother." He argued that psychiatric diagnoses rested on false grounds in that they were was solely based on the patient’s conduct, without external validators. The Hungarian-born American psychoanalyst Thomas Szasz (1920-) advanced the idea that psychiatric disorders are a myth, or social branding. He was embraced by the scientology movement in 1968, whose originator L. Ron Hubbard (1911-1986) created in 1950 the business of dianetics, the doctrine of the Church of Scientology, as an alternative to psychoanalysis.
The 1950s and 1960s brought widespread criticism of psychiatric diagnoses, a movement that coincided with the civil rights movement of the 1960s, and that particularly targeted the grounds for involuntary commitment to psychiatric care by means of diagnoses. When, in an experiment, several psychiatrists were asked to diagnose the same patient, it was obvious that they represented different schools of thought that did not share a common set of definitions. This challenging of the intellectual ground of psychiatry had profound effects on the allocation of resources, shifting from institutionalization to outpatient voluntary care in the United States and in Europe. In Italy all involuntary care was declared unlawful in 1978.
Two psychiatrists at Washington University in St. Louis then decided to bring sense into psychiatric diagnoses: Samuel Guze (1923-2000) and Eli Robins (1921-1995). In 1970 they published a paper on a criteria-based diagnosis of schizophrenia. This seminal paper became the intellectual basis for the 3rd version of the Diagnostic and Statistical Manual of Mental Disorders (DSM-III) that was published in 1980 by the American Psychiatric Association. This fundamentally new classification was based on a consensus of clinical criteria. Also, the DSM-III did not assume etiological factors; it was based on a consensus among academic psychiatrists about the typical symptoms of a disorder and its prognosis. In 1987 and in 1994 this classification was revised, based on 150 literature surveys, and 12 field studies with more than 6000 diagnostic interviews. Work on its 5th version is ongoing, and it is to be published in 2012 (http://www.psych.org/MainMenu/Research/DSMIV/DSMV.aspx).
The DSM classification applies 5 perspectives on a patient: Axis 1 disorders (for example major depressive episode, anorexia nervosa), Axis 2 personality disorders, and neurodevelopmental disorders), Axis 3 somatic disorders (for example diabetes mellitus, traumatic brain injury), Axis 4 current stressors (for example having been raped, bereavement), and Axis 5 global assessment of function.
One current ambition in revising the DSM classification is to pay more attention to ethnicity in understanding how symptoms may present. Gender differences will also be elucidated. The most important change in DSM-V will be the inclusion of dimensions in diagnoses; for example, how severely ill a patient is with schizophrenia or depression.
By international convention most countries use the International Classification of Diseases (ICD) in making all diagnoses (somatic and psychiatric) in routine health care. This classification is produced by the World Health Organization. The current ICD-10 classification is quite similar to that of DSM-IV. The WHO is currently working its 11th revision of the ICD. With regard to the psychiatric diagnoses there is a joint effort with the DSM-V developers to use similar principles and standards. The revision process was formulated in 2007 and the draft version will be tested in field trials (http://www.who.int/classifications/icd/ICDRevision/en/index.html).
These efforts have advanced the reliability of psychiatric diagnoses to standards similar to those of other disciplines. Methods for external validation have emerged in recent years. For example, functional magnetic resonance imaging (fMRI), and other in vivo imaging techniques, allow the study of how the amygdala reacts to an anxiety provocation in a subject with an anxiety disorder. Imaging techniques reveal disturbed CNS networks in subjects with schizophrenia, and pronounced structural aberrations in the lateral and medial parts of the temporal and frontal lobe. Untreated depression has been shown to cause cerebral shrinking. The efficacy of serotonergic medications depends on neurogenesis. Latency to rapid eye movement sleep is correlated with clinical symptoms of depression. Amyloid, a protein in the plaques in Alzheimers disease, has been detected in vivo in patients in a PET study. The effects of antipsychotic and antidepressant drug treatments can be correlated with symptom reduction, cerebral blood flow, and brain metabolite ratios.
Psychiatric comorbidity[edit| edit source]
The criteria in the DSM-IV classification are not always specific for the disorder. Therefore, epidemiological studies produce high rates of comorbid psychiatric conditions, especially if subjects are monitored longitudinally rather than cross-sectionally (lifetime or 12-month prevalence vs. point prevalence). These are consequences of criteria-based classification that need to be accounted for in selecting subjects for research and treatment.
Subjects with a primary anxiety disorder may develop a secondary depression, causing them to seek treatment. Treating the depression uncovers the underlying primary disorder. Anxiety subjects may also self-medicate with alcohol and other substances that are anxiolytic and be diagnosed with a substance use disorder. A patient with schizophrenia may develop a depression, and unless that is properly diagnosed the antipsychotic medication may be unnecessarily increased. A patient with recurring depressive episodes may eventually develop a manic episode, thus altering the diagnosis from unipolar depression to bipolar disorder. Subjects with substance use disorders may develop psychotic reactions to e.g., cannabis or amphetamine that may mimic schizophrenia. Since subjects with schizophrenia tend to seek various drug effects, the effects of cannabis or alcohol may cause psychiatric symptoms per se. There are many more instances of comorbidity that need to be understood.
Personality disorders[edit| edit source]
An issue with the DSM-IV classification is the distinction between axis I disorders and axis II personality disorders. Personality, cognitive style, and social attitudes are moderately or highly heritable according to adoption and twin studies. There is even a genetic contribution to being religious or antisocial, and to the amount of time spent watching television! Personality traits are stable and genetically determined throughout life, and are modifiable only by serious effects such as a neurodegenerative disease, severe substance use, a traumatic brain injury, a brain tumor, or a severe generalized medical condition. One such famous case is Phineas Gage, a railroad worker who survived an iron rod that passed through his frontal lobes in 1848 and caused a pronounced personality change.
There have been many theories since Hippocrates to explain how personalities are shaped. The current explanatory model is the five-factor model that describes a person along five different dimensions, e.g., being curious or rigid, dependable or careless, as well as degrees of self-confidence, stubbornness, shyness, and extroversion. In the DSM-IV classification, personality disorders are assessed categorically, based on clinical assessments of cognition, affectivity, interpersonal functioning, and impulse control. If a person exhibits stable traits that deviate from the norms of the subject’s ethnic group, they may be deemed a personality disorder. There are 11 DSM-IV personality disorders divided into three clusters. Personality disorders occur in about 10 per cent of population samples, and in about a third of clinical samples.
The distinction between axis 1 and axis 2 disorders is sometimes unclear. A patient with a serious axis I disorder may qualify for a personality disorder diagnosis, e.g., long-standing social anxiety disorder may be regarded as a phobic personality disorder if sufficiently impaired. Yet, such a patient may respond well to treatment. A subject with high-functioning autism or Asperger syndrome may be regarded as having a schizotypal personality disorder. Attention Deficit Hyperactivity Disorder (ADHD) may be confused with antisocial personality disorder. In the work groups for the DSM-V, these issues may cause a fundamental change in dealing with axis II personality disorders.
Somatic comorbidity[edit| edit source]
Somatic disease may cause or aggravate psychiatric disorders. For example, a patient with diabetes mellitus who has taken too much insulin may present confused or agitated in the emergency room because his blood sugar is too low. A patient with hypothyroidism or hyperthyroidism or hyperparathyroidism usually has anxious or depressive symptoms. Patients with acute intermittent porphyria may become psychotic, and are always anxious. Depression is known as a risk factor for acute myocardial infarction, and can add to the risk of cardiovascular complications. Patients with stroke often develop anxiety and depression. Such manifestations of somatic disease are important to recognize, and they are diagnosed on axis III in the DSM-IV.
Premenstrual dysphoria is an intermittent cluster of symptoms among which irritability and dysphoria are the most disturbing. It develops following ovulation and reaches a peak until menstruation occurs, obviously governed by hormonal variations across the menstrual cycle.
Multiple sclerosis can present with psychotic symptoms and mood elevations including euphoria. Wilson’s disease is a disorder of copper metabolism that can cause rapid mood swings and cognitive dysfunction. Systemic lupus (SLE) can present with confusion and psychotic symptoms. Pernicious anemia (deficiency of vitamin B12) may present with depressive symptoms, memory deficits and sometimes confusion.
The medical model – is it useful?[edit| edit source]
The scientific community assumes that there is a molecular basis for psychopathology, and that symptoms are produced that can be elicited, quantified and classified by interviewing and observing a subject. This medical model was critiqued in the 1950s and 1960s, causing thought leaders to argue for external causation rather than disorders of the brain. Psychiatry was also abused for political purposes. Sane political dissidents in the Sovjet Union were sentenced by courts to be diagnosed and incarcerated in mental asylums and given tranquilizers (for some this may have been a better alternative than imprisonment).
Early support for the medical model came from twin studies that showed a strong genetic contribution to schizophrenia and bipolar disorder. Neurosyphilis, first defined in 1672, and thought to be an immoral disease, was determined to be an infectious disease in 1913. The Austrian psychiatrist Julius Wagner-Jauregg was awarded the Nobel Prize in 1927 for having shown that neurosyphilis could be treated by infecting the patient with malaria, and in 1943 patients began treatment with penicillin. The dramatic effects of lithium on mania were elucidated in the 1950s. The equally dramatic effect of chlorpromazine on delusions and hallucinations in schizophrenia was also discovered in the 1950s. With regard to anxiety, a break-through in 1964 was the finding by Donald F. Klein that imipramine could extinguish panic attacks, previously believed to stem from unconscious parental conflicts.
In recent years, the medical model has gained support from neuroimaging studies. The model has proven useful in that the benefits and hazards of psychotherapies and psychotropic medications have been shown in randomized controlled trials for which subjects with these diagnoses have been selected. The regulatory bodies of the world base their research protocols and marketing approvals on the ICD-10 and DSM-IV nosologies. Good Clinical Practice, the code by which treatment studies are undertaken, assumes that subjects are selected based on structured diagnostic interviews and that validated measures of changes in symptoms and functioning are applied (see below). The courts pronounce verdicts on forensic psychiatric assessments that are based on the medical model. The medical model is the basis for clinical research into the genetics, etiology, pathogenesis, epidemiology, treatments, and outcomes of psychiatric disorders.
The medical model is often poorly understood by lay persons in politics, administration and the media. It is attacked by the scientology movement and other antipsychiatric movements that refuse to acknowledge the scientific basis for psychiatric disorders. No wonder that the public is so confused, and that stigma against psychiatric disorders is so prevalent in many societies. Traditional medicine This paragraph is a brief excursion into the domain of traditional medicine and how it relates to psychiatric diagnoses. Examples of this interface are given.
The overwhelming majority of the world population will primarily be diagnosed and treated in traditional medicine that was developed locally by indigenous peoples:
Traditional medicine is the sum total of the knowledge, skills and practices based on the theories, beliefs and experiences indigenous to different cultures, whether explicable or not, used in the maintenance of health as well as in the prevention, diagnosis, improvement or treatment of physical and mental illness. (World Health Organization, 2000).
Complementary and Alternative Medicine (CAM) are recently developed therapies, often in opposition to evidence-based medicine:
… a broad domain of healing resources that encompasses all health systems, modalities, and practices and their accompanying theories and beliefs, other than those intrinsic to the politically dominant health systems of a particular society or culture in a given historical period (The Cochrane Collaboration, 2000).
There are approximately 500 000 certified practitioners of Traditional Chinese Medicine (TCM) in China, and additionally folk herbalists and "magic witch doctors," serving 56 ethnic nationalities with widely differing beliefs about illness causation, and with much stigma toward psychiatric disorders. While patients with obviously disorganized behavior will be admitted to psychiatric hospitals, those with lesser morbidities are primarily dealt with in TCM. Diagnoses are flexible from one day to another, and based on listening, observing, questioning, and pulse-taking. Religious healers, although forbidden, may apply fortune-telling, handwriting analysis, and palm-reading. They try to counteract evil spirits and repair relationships with ancestors. There is a Chinese Classification of Mental Disorders (CCMD-3) written in Chinese and in English in 2001, that includes about 40 ethnic diagnoses. One is shenjing shuairuo which emphasizes somatic complaints and fatigue, as in the ICD diagnosis neuroasthenia. Another is koro (an excessive fear of genitals and breasts shrinking back into the body).
At healing shrines in India, e.g., at the temples of Balaji in Rajasthan, most subjects have a diagnosable psychiatric illness including psychosis and severe depressive episodes. Healers name it spirit illness, and prescribe offerings and rituals at the temple and at home.
In Japan, Morita therapy was devised by a psychiatrist, and draws from Zen Buddhism, aiming to make people accept their destiny, and live with the symptoms that are similar to social anxiety disorder in the DSM-IV. There is a period of absolute rest, then a period of light work, followed by a period of normal work. In studies more than one half of all patients, including those with schizophrenia, had seen a traditional healer or shaman (yuta) before seeking psychiatric treatment. Taijin Kyofosho (anthropophobia, phobia of eye contact) is a culture-bound syndrome, rooted in consideration for others, loyalty to the group, protecting a vertical society, mutual dependence, a sense of obligation, and empathy.
In the Xhosa tribe in South Africa, amafufuynana and ukuthwasa are culture-bound syndromes that overlap with the DSM-IV criteria for schizophrenia. Both include delusions, hallucinations, and bizarre behaviour. A young person with ukuthwasa is a candidate to become a traditional healer, as he/she can communicate with ancestors, and resisting such a calling may cause illness. There is often a family history of schizophrenia and other psychiatric disorders among those with ukuthwasa. Amafufuynana is believed to be caused by sorcery.
In Quichuas, an Amerindian nation in South America, someone who suffers from anxiety or depression according to the DSM classification is diagnosed as the victim of sorcery or bad spirits.
In the United Kingdom, South Asian patients, including Muslims from Pakistan, frequently seek traditional healers (hakims), practicing Unani Tibbia that stems from Jundishapur south of Teheran. Psychiatric disorders are treated with herbs, diets, and amulets with holy words from the Koran, or the patient is referred to a mullah. Such treatments are often conducted in tandem with biomedical treatments. African-Caribbean patients employ counter-measures including religious rituals and magic (Obeah - witchcraft), having consulted divine healers from the Pentecostal or other churches.
In Italy, the Catholic Charismatic Renewal, sanctioned by the Pope, stems from the Pentecostal cult and includes 300 000 believers. Illness, according to the Catholic doctrine of 2000, is closely related to Evil; it can be God’s punishment for sins, and healing by God can be obtained by collective prayer that produces exalting salvation and jubilation.
These are some brief examples of the multitude of traditional explanations and treatments that are used for the large majority of the world’s population. Traditional healers are a major force in global mental health, as about 40 per cent of their clients suffer from mental illnesses. A psychiatrist trained in evidence-based medicine thus needs to develop an understanding of the large influence of such faiths on patients with psychiatric disorders, even in technologically advanced societies, and need to adjust for it to establish a therapeutic alliance and improve the chances of a favorable outcome.
Structured diagnostic interviews and screening questionnaires[edit| edit source]
Many structured diagnostic interviews have been tested over the years. The first was the Present State Examination (PSE) in Great Britain in the 1950s that was integrated into the Schedules for Clinical Assessment in Neuropsychiatry (SCAN, see below). The Mental Status Examination was developed in the United States in the 1960s.
Diagnostic interviews differ in scope and the qualifications of the interviewer, and in being based on ICD or DSM classification. Some are comprehensive and designed to find all psychiatric morbidity in general population samples, in primary care, or in tertiary care. Others deal primarily with e.g., affective disorders, substance use disorders, or personality disorders. Web-based case finding questionnaires are being developed to encourage people to seek treatment, as most individuals with conditions (such as substance use disorders, anxiety disorders, depression) amenable to treatment are not receiving any kind of treatment. Self-rating symptom scales are available for case-finding in e.g., the reception area of an outpatient unit, or to assess symptom changes in treatment studies.
Below are short descriptions of some currently used instruments.
The MINI Neuropsychiatric Interview was developed by David Sheehan and Yves Lecrubier as an efficient tool for the experienced mental health worker to look for 15 psychiatric diagnoses in an interview that takes about 30 minutes: Affective, anxiety, psychotic, substance use, eating, and antisocial personality disorders as well as current suicidality. The subject is instructed to simply answer yes or no to each question. Each section has one or a few lead-in questions, and in-depth questions in case there is a positive response. It is essential that the subject understands the questions, so the interviewer may have to repeat them or explain them. The questions are purposely overinclusive (false positives) so that cases will not be missed. It is critical that the interviewer has clinical judgment to assess the value of the subject’s responses. Since somatic diseases may have caused the symptoms (such as a brain tumor, thyroid disease, or adverse effects of medications and substances), a physician must validate the interview results. An experienced nurse or psychologist or mental health worker may do the actual interview. The MINI is the most common interview in drug treatment studies, and is available in over 40 languages. The English MINI version 6.0 was updated in 2009. It can be down-loaded without charge from www.medical-outcomes.com.
The Composite International Diagnostic Interview 3.0 (CIDI) is a fully structured non-clinical interview intended for use in general population surveys .
The CIDI-SAM (SAM is for Substance Abuse Module) is a structured interview that ascertains DSM-III, DSM-III-R, Feighner, RDC and ICD-10 diagnoses for alcohol, tobacco and nine classes of psychoactive drugs. It was designed at the request of the World Health Organization to expand the substance abuse sections of the CIDI. The SAM module takes an average of 45 minutes to complete. 
The Schedules for Clinical Assessments in Neuropsychiatry (SCAN) is a semi-structured clinical interview to assess major mental disorders  in clinical settings.
Schedules for Affective Disorders and Schizophrenia (SADS) has been produced in several versions since 1975, and can take up to 3 hours to complete by a trained clinician. It is the basis for the Structured Clinical Interview for Diagnosis (SCID I and SCID II) that is also an expert instrument.
The Personality Diagnostic Questionnaire (PDQ-4) holds 99 true/false items to screen for 11 DSM-IV personality disorders .
The General Health Questionnaire (GHQ-12) was developed in the 1970s for self-screening in primary care, public health surveys, and other settings with lower degrees of psycho-pathology. GHQ-12 asks if 12 symptoms have been present in recent weeks much more than usual, rather more than usual, no more than usual or not at all. Total scores derived using the Likert method (3-2-1-0) range from 36 to zero with higher scores denoting greater morbidity. It has proved reliable, stable and valid when tested in numerous primary care and hospital settings with a sensitivity and specificity versus CIDI of 79% and 77% respectively at cutpoint 11/12.
Another self-screen questionnaire is the Hospital Anxiety Depression Scale (HADS), developed in the UK to find cases with symptoms of anxiety and depression. It consists of 14 items that a subject can respond to within a few minutes, for example prior to a physician visit.
The Clinical Interview Schedule (CIS) was developed to assess anxiety, depression and somatization. The revised version (CIS-R) has been used in population surveys by lay interviewers.
The Kessler Psychological Distress Scale (K-10) checks if 10 mental symptoms have been present in the last 4 weeks for all, most, some, a little or none of the time. It was designed for use in general health surveys and has proved reliable and valid in surveys in the United States and in Australia .
Legal issues and psychiatric diagnoses[edit| edit source]
The courts in most societies take a diagnosis of a psychiatric disorder into account before passing sentence. Usually the court will order that a subject undergoes a forensic psychiatric examination to determine whether there is a severe psychiatric disorder, and whether the subject can be held accountable for his actions. Does a subject with schizophrenia or antisocial personality disorder understand the consequences of his actions for other people and for society? Did the mother kill her child because of a depression, or because she was under the influence of auditory hallucinations? If there is an indisputable organic brain disease is the subject to be held accountable for a crime? These are evaluations that require an experienced, professional, thorough and highly regulated psychiatric assessment. The law varies between nations, and the court may order commitment to psychiatric care, or a prison term or both.
In many societies doctors are responsible by law to report if a patient is deemed unfit to possess a fire arm, or unfit to have a driver’s license, or to have custody of a child. Such reports require a careful psychiatric diagnosis.
Medical records[edit| edit source]
In most countries, the history and mental health status examination should result in a clinical evaluation of the patient and at least one psychiatric diagnosis, all of which make up the core of the patient’s medical record (chart). This may be a preliminary or definite diagnosis. For example, a patient presenting with typical symptoms of schizophrenia can be given a preliminary diagnosis that is confirmed after 6 months, because of the duration criterion in DSM-IV.
The physician can be held accountable to a disciplinary board if the diagnostic procedure is not properly recorded. The diagnosis is the basis for justifying treatments and perhaps involuntary care.
Records are still written by hand or typed in many countries. Increasingly in Europe and in the United States there is a move to electronic medical records. This is in the interest of administrators and regulators to hold physicians accountable and to increase patient safety. Insurers have a stake in psychiatric diagnoses to assess the risk of a potential subject for a health insurance or retirement plan. If records contain valid and reliable information about the patient’s diagnosis, treatments, suicidal risk, and risk for aggression it will increase the quality of care. If all of the patient’s health care contacts (the emergency room, primary care unit, psychiatric clinic) are eligible to read the patient’s record it will increase patient safety, and reduce unnecessary investigatory procedures. There are opportunities for longitudinal case studies, research, and allocation of health care resources.
The potential drawback with a unifying electronic medical record is that it will be at the expense of person integrity and privacy. Particularly, a psychiatric record will contain highly sensitive information that should not be accessible to insurers and employers. Patients should have the option to decline such a unifying medical record that can otherwise be read by all eligible users of a computerized record system.
Suggested reading[edit| edit source]
Anne Farmer, Peter McGuffin, Julie Williams. Measuring Psychopathology. Oxford University Press 2002.
Samuel B. Guze. Why Psychiatry is a Branch of Medicine. Oxford University Press, 1992.
Mario Incayawar, Ronald Wintrob, Lise Bouchard, Goffredo Bartocci (eds.). Psychiatrists and Traditional Healers. Unwitting Partners in Global Mental Health. Wiley-Blackwell, 2009.
Donald W. Goodwin, Samuel B. Guze. Psychiatric Diagnosis - 4th Edition. Oxford University Press, 1989.