An expert consensus on a precise definition of agitation in psychiatric illness is currently lacking. However, there are fairly consistent definitions in the medical literature. These include “exceeding restlessness associated with mental distress” and “excessive motor activity associated with a feeling of inner tension.”¹⁰ The word “agitation” is used not only for nonaggressive behavior such as repetitive questioning and pacing, but also for physically and verbally aggressive outbursts. It is likely that in the future it will be possible to define the parameters of psychomotor agitation and related symptoms more specifically.¹¹ This would be useful because agitation is a common warning signal that often precedes violence.¹²

Official incident reports from psychiatric hospitals tend to underestimate inpatient violent behavior, and even more so self-harm and property damage.¹³ The inpatient aggression prevalence rates differ substantially across studies with gross international and interhospital variation.¹⁴ It has been estimated that 3% to 25% of psychiatric inpatients exhibit violent behavior while hospitalized.¹⁵, ¹⁶, ¹⁷ A meta-analysis of all studies using the same method of measuring aggression found the number of aggressive incidents
to vary considerably between acute admission wards, ranging from 0.4 to 33.2 incidents (average 9.3) per patient year.¹⁸ A UK study recruiting patients with first-episode psychosis found that almost 40% were aggressive at first contact with services and more than half of these were physically violent.¹⁹ One study estimated that in a 12-month period at an acute psychiatric hospital ward a nurse would have a one in ten chance of receiving any kind of injury as a result of patient’s physical aggression.²⁰

There are several explanations for the reported variation in prevalence of violent acts performed by psychiatric inpatients. The most obvious explanation is the lack of common understanding and standardization not only across different studies but also among staff as to what constitutes an act of aggression or violence.²¹ Agitation is even harder to define, yet combined with the threat of violence it is the most common reason for coercive measures such as seclusion or restraint, as well as for hospitalization itself.²² Obviously, prevalence is also dependent on the tasks and functions of the ward in question. Many of the studies measuring inpatient aggression have been carried out on high-risk wards thereby yielding skewed results. There are also major differences in local and national policies and cultures that contribute to patient and staff behavior on hospital wards. Studies of disruptive patient behavior reveal that conflict behavior is not only ubiquitous but also heterogeneous. One typology created seven categories: the angry absconder, the angry refuser, the absconding misuser, the protestor, the self-harmer, the abstainer, and the medication ambivalent. Only the first two were associated with aggressive behavior.²³

Only a few epidemiologic studies have been done to assess the prevalence of agitation.¹¹ Agitation is common among emergency patients with psychoses, and it has been estimated that 21% of psychiatric emergency visits may involve agitated patients with schizophrenia.²⁴

Most studies of inpatient aggression have been carried out on adult wards and have not emphasized older people in particular. The overall impression is of a high incidence of usually low-impact aggression among people with dementia.²⁵ There are even fewer studies about violence on child and adolescent psychiatric units. In spite of the lack of epidemiologic data, the existence of the problem is proved by the published practice guidelines on management of acute aggressive behavior in young people in inpatient treatment facilities.²⁶

Observer-rated scales have been developed for the quantitative measurement of violence during inpatient treatment. The Staff Observation Aggression Scale (SOAS) consists of five columns, each pertaining to specific aspects of aggressive behavior: the provocation, means used by the patient during the aggression, target of the aggression, consequences, and measures taken to stop aggressive behavior.²⁷ SOAS has been revised (SOAS-R) in order to develop a finer-grained severity scoring system.²⁸ The Modified Overt Aggression Scale (MOAS) collects the most serious incidents during the past week. It includes four dimensions: verbal and physical aggression, property damage, and self-inflicted harm.²⁹ The Social Dysfunction and Aggression Scale (SDAS) consists of nine items (SDAS-9) covering outward aggression and two items (SDAS-2) covering inward aggression.³⁰ All these scales intercorrelate highly.³¹

Some instruments have been developed to measure agitation. Positive and Negative Syndrome Scale-Excited Component (PANSS-EC) is a five-item subscale of the frequently used PANSS.³² The Behavioral Activity Rating Scale (BARS) for acutely agitated psychotic patients scores behavioral activity from one (difficult or unable to arouse) to seven (violent).³³

Agitation is a common symptom of schizophrenia, bipolar disorder (manic and mixed episodes), and dementia. In particular, almost half of the patients with Alzheimer disease have agitation,³⁴ and behavioral disturbances are very common.³⁵ Agitation may also occur in catatonic states due to various causes and as an adverse effect of psychiatric medication.³⁶ The pathophysiology of agitation in these different states is not well understood.¹¹ In patients with psychosis, proposed mechanisms for agitation have included hyperdopaminergia in the basal ganglia, increased norepinephrine tone, and a reduction of inhibitory γ-aminobutyric acid influences.³⁷ Among patients with dementia, agitation is believed to be of multifactorial etiology.³⁸

A number of theories have been developed to explain the reasons for inpatient aggression. These have been narrowed down to three models: the internal, the external, and the interactional models. The internal model sees mental illness as the cause of aggressive behavior. The external model is based on the assumption that environmental factors, like the type of regime on the ward or provision of privacy, contribute to the occurrence of aggressive incidents. The interactional model emphasizes the relationship between staff and patient. Patients perceive environmental conditions and poor communication to be a significant precursor of aggressive behavior. Nurses, in comparison, see the patients’ mental illness to be the main reason for aggression. Although both groups are unsatisfied with the strategies used to control disruptive behavior, the suggestions for actions differ according to the assumed underlying cause. The use of medication and seclusion is supported by staff, but much less so by the patients, who prefer negotiation and de-escalation.³⁹

Several studies have reported that the two most common types of patient assailants are the older male with a diagnosis of schizophrenia, past histories of violence toward others and substance use disorder, and the younger patient with personality disorder and past histories of violence toward others, personal victimization, and substance use disorder.⁴⁰,⁴¹ Other patient characteristics have also been associated with the threat of inpatient violence.¹³,⁴², ⁴³, ⁴⁴ Little is known about the risk of aggression among those with affective psychoses. A study of patients with first-episode psychosis found those with a diagnosis of mania to be almost three times more likely to be aggressive at first contact than patients with schizophrenia.¹⁹ From a clinical viewpoint it is useful to know that there are groups of repetitively violent patients who are responsible for a disproportionately large number of aggressive incidents.⁴⁵

An interesting recent finding associated lifetime exposure to a life-threatening traumatic event with the use of seclusion or restraint during the hospital stay. The authors concluded that previous exposure to traumatic events enhances the risk of revictimization and retraumatization during inpatient treatment.⁴⁶ This view is often supported by the psychiatry survivor movement, that is, by patients who have experienced the existing mental health laws and practices as oppressive.

There is evidence that adverse incidents are more likely during and after weeks of high numbers of male admissions, during weeks when other incidents also occur, and during weeks of high regular staff absence through leave and vacancy.⁴⁷ Overcrowding⁴⁸ and boredom on wards⁴⁹ are associated with increased amount of violent incidents. Patients associate aggression on ward to negative and controlling staff attitudes.⁵⁰ There is some evidence that staff with poor clinical and interpersonal skills are at increased risk of violence.⁵¹

Many theories of inpatient violence see aggressive incidents as a result of complicated interactions between patient, staff, and environmental stressors. For instance, repeated inpatient aggression could be the result of a vicious circle: the patient’s violent behavior increases the environmental and communicational stress on the patient and contributes to the risk of another violent outburst.⁵²

There are three types of risk factors: empiric, theoretic, and clinical. Predicting inpatient violence differs from predicting aggressive behavior in the community.⁵³ Purely theoretic predictors with high correlation with violence like poverty and segregation are useless from the clinical standpoint. Clinical approaches to violence prediction have been considered only slightly better than by chance,⁵⁴ but there are also reports of moderate accuracy.⁵⁵ Research in clinical violence prediction is difficult for several reasons. Assessment of the violence risk is judged to be one of the key competencies by many psychiatrists and is based on intuition and clinical experience, both hard to measure.⁵³ However, from the clinical viewpoint the true task of the clinician is to prevent violence, not to predict it. Predicted aggression leads to preventive actions, and therefore the prediction appears incorrect. Typical difficulties in clinical assessment of the risk of aggressive acts include vagueness of what is being assessed, relying on illusory correlations, failure to incorporate situational information, and ignoring statistical base-rate information.⁵⁶

There has been some controversy over the usefulness of clinical versus actuarial prediction of aggression. In clinical judgment, the information about the probability of violence is processed inside the head of the decision maker, whereas in actuarial methods conclusions are drawn solely on the basis of established empirical knowledge.⁵⁵ On the basis of empirical findings, by far the best predictor of future violence is previous violent behavior.⁵³ The clinical approach involves expert clinical
judgments of several factors in relation to the individual and the situation. Unaided clinical prediction of violent recidivism after hospital discharge does not function well, but clinical decision making can be accurate in estimating short-term aggression during psychiatric hospital care.⁵⁵ When making these clinical judgments, nurses rely on their personal knowledge of the patient, but also “tune in” on potentially violent situations as a whole and try to search causes for the violent behavior.⁴ There is no difference in the accuracy of violence prediction between psychiatric nurses and psychiatrists.⁵⁷ Recently, a combination of clinical and actuarial approaches has been recommended. The emphasis is on developing evidence-based guidelines that promote systemization and consistency, but that are flexible enough to take account of specific cases and contexts.⁵⁶

Some violence-related rating instruments to predict inpatient violence have proved to be useful, for example, the Bröset Violence Checklist (BVC).⁴⁵ The BVC assesses the presence of six observable patient behaviors, namely whether the patient is confused, irritable, boisterous, verbally threatening, and attacking objects. The sensitivity and specificity results indicate that BVC discriminates between the violent and nonviolent patients over the next 24-hour prediction period. It has shown to be 63% accurate in predicting that violence would occur within the next 24 hours and 92% accurate in predicting that violence would not occur.⁵⁸ BVC has been combined with the Visual Analog Scale (VAS) with good results.⁵⁹

The Short-Term Assessment of Risk and Treatability (START)⁶⁰ is a new structured professional assessment aimed to guide assessment and management of diverse population of mentally disordered patients and to act as a clinical indicator of treatment progress.⁶¹ START is specifically intended to be completed and used by an interdisciplinary team. It includes 20 dynamic risks and strength-related factors that give information concerning multiple risk domains relevant to everyday psychiatric clinical practice (e.g., risk to others, suicide, self-harm, self-neglect, substance abuse, unauthorized leave, and victimization). Preliminary findings have been promising.

The evidence regarding interventions to treat violent or agitated behavior is limited.⁶² It is also difficult to decide which interventions to use for which patients. Both agitation and aggression should be treated according to the underlying cause of the behavior when possible. For the present, treatment is usually identical for both agitation and aggression. Studies examining factors motivating inpatient aggression found that violent episodes can be categorized into three classes with different underlying causes.⁶³,⁶⁴ The most common reason underlying an assault was disordered impulse control, often in situations that involved directing a patient to change his unwanted behavior and refusal of a patient request. Aggressive incidents due to organized assaults came second, and assaults motivated by psychosis were the least common.⁶⁴ Psychotically motivated assaults were due to positive psychotic symptoms, psychotic confusion, and disorganization.⁶³ All these should be treated differently and according to the cause. Psychosocial and pharmacological interventions for these states are discussed in the subsequent text and in Chapters 2 and 3, respectively.