Inclusion and exclusion criteria

The abstracts of the studies found were reviewed and included if they reported original quantitative data describing patient or mental health staff attitudes to LAIs. As it was expected that most studies would be at the lower end of the hierarchy of study design, no quality threshold for inclusion of studies was set.

Assessment of studies

All included studies were assessed according to a hierarchy of evidence. ⁷ The studies were also assessed according to a 13-item checklist constructed by Walburn et al, ^{Reference Walburn, Gray, Gournay, Quraishi and David1} and subjected to the same basic statistical analysis to identify the strengths and weaknesses of each study and to allow comparison with the quality of studies found in the previous review.

Quality of studies

All the studies were in the form of a cross-sectional survey, which is level IV evidence in the hierarchy of study design. ⁷ The assessment of the studies against the checklist designed by Walburn et al (Table 2) revealed a varied performance, with a range of scores from 4 to 11 (out of a maximum of 13). ^{Reference Walburn, Gray, Gournay, Quraishi and David1} Eleven of the 12 studies had explicit aims and details of inclusion and exclusion criteria, but none carried out a sample size calculation. Other areas of poor performance were justification of response/drop-out rate and justification that the sample was representative of the population. Important demographic details were included in all studies – for example, years working in psychiatry and number of patients in their care currently receiving LAIs (in staff studies), and (in patient studies) diagnosis, age, gender, and current and past medication formulations.

Table 2 Quality analysis of all included studies

Study	Explicit a priori aims	Definition/size of population under investigation	Sample size calculation	Justification that sample is representative of population	Inclusion/exclusion criteria stated	Demographic details	Research independent of routine care/practice	Justification of validity/reliability of measures	Original questionnaire available	Response/drop-out rate specified	Justification of response/drop-out rate	Discussion of generalisability	Statement of source of funding	Marks lost	Percentage of maximum quality score %
Bradstreet (2004) Reference Bradstreet and Norris10	–	–	–	–	–	+	+	–	+	N/A	N/A	–	+	7	31
Castle (2002) Reference Castle, Morgan and Jablensky11	+	+	–	–	+	+	+	–	–	–	–	–	–	8	39
Harris (2007) Reference Harris, Lovell and Day19	+	–	–	–	+	+	+	+	+	–	–	–	–	7	46
Heres (2006) Reference Heres, Hamann, Kissling and Leucht18	+	–	–	–	+	+	+	–	+	+	–	+	+	5	62
Heres (2007) Reference Heres, Schmitz, Leucht and Pajonk12	+	–	–	+	+	+	–	–	+	+	–	+	+	5	62
Lambert (2003) Reference Lambert, Brennan, Castle, Kelly and Conley17	+	+	–	+	+	+	+	–	+	+	–	+	–	4	69
Patel (2003) Reference Patel, Nikolaou and David13	+	+	–	–	+	+	+	+	+	+	–	+	+	3	77
Patel (2005) Reference Patel, De Zoysa, Baker and David14	+	–	–	–	+	+	+	+	+	+	–	+	–	5	62
Patel (2008) Reference Patel, De Zoysa, Bernadt and David9	+	–	–	+	+	+	+	+	+	+	–	–	+	4	69
Patel (2009) Reference Patel, De Zoysa, Bernadt and David8	+	–	–	–	+	+	+	–	+	–	–	–	+	7	46
Patel (2008) Reference Patel, Yeung, Haddad and David15	+	–	–	–	+	+	+	+	+	+	–	+	–	5	62
Patel (2009) Reference Patel, Haddad, Chaudhry, McLoughlin, Husain and David16	+	+	–	+	+	+	+	+	+	+	+	+	–	2	85
Totals	11/12	4/12	0/12	4/12	11/12	12/12	11/12	6/12	11/12	8/11	1/11	7/12	6/12	Mean 5.2	Mean 59.2

+, present; –, absent; NA, not applicable

The mean percentage of maximum-quality scores is higher for studies published since mid-1999 compared with the studies reviewed by Walburn et al (59% v. 44%). ^{Reference Walburn, Gray, Gournay, Quraishi and David1}

Patient attitudes

Of the five studies that contained quantitative data addressing patient attitudes to LAIs, one elicited a predominantly positive attitude, two were negative and two were neutral (Table 3). The two neutral studies revealed no difference in total scores on the 10-item Drug Attitude Inventory (DAI–10) and Beliefs about Medication Questionnaire (BMQ) between individuals on oral and LAI maintenance antipsychotics, leading the authors to suggest that the attitudes held by a group of voluntary out-patients regarding treatment were similar for those given LAIs and those taking tablets. ^{Reference Patel, De Zoysa, Bernadt and David8,Reference Patel, De Zoysa, Bernadt and David9} In contrast, when asked detailed questions about their attitudes regarding medication adherence (using the Rating of Medication Influences questionnaire), those in the LAI group (v. oral) had higher non-adherence scores and more self-stated reasons for non-adherence. ^{Reference Patel, De Zoysa, Bernadt and David9} A subsample analysis of DAI–10 scores by Patel et al found that those who had previously switched from LAIs to oral medication retained negative views regarding LAIs. ^{Reference Patel, De Zoysa, Bernadt and David8} Embarrassment or shame associated with receiving an LAI concerned a significant proportion of those questioned, and just under half of participants who had current or previous experience of LAIs felt they were forced to start receiving LAIs. ^{Reference Patel, De Zoysa, Bernadt and David8}

Table 3 Patient preferences and attitudes to long-acting injections: data extraction

Study	Positive	Neutral	Negative
Bradstreet (2004) Reference Bradstreet and Norris10			Overall weighing up the pros and cons of antipsychotic LAIs, 43% rated them as unhelpful compared with 38% who rated them as helpful
Castle (2002) Reference Castle, Morgan and Jablensky11			The depot antipsychotic group had the highest proportion of individuals rating their medication as unhelpful, with 7% of those on depot with insight and 23% of those on LAI without insight rating their medication as unhelpful
Heres (2007) Reference Heres, Schmitz, Leucht and Pajonk12	Of those currently on LAI 73% agreed with this treatment modality; 54% of the total group agreed that it is more convenient to receive LAI every 2–4 weeks than to take tablets daily
Patel (2008) Reference Patel, De Zoysa, Bernadt and David9		Those on LAI scored significantly higher on the ROMI non-adherence factors (LAI 15.65, oral 14.37) but there was no difference for ROMI adherence factors or BMQ
Patel (2009) Reference Patel, De Zoysa, Bernadt and David8		Attitudes regarding current medication tested using DAI–10 altered for formulation-specific questions were not different between oral and LAI groups; 43% of those on LAI preferred it to tablets and 26% had no preference

BMQ, Beliefs about Medicine Questionnaire; DAI–10, 10-item Drug Attitude Inventory; LAI, long-acting injection; ROMI, Rating of Medication Influences

Two studies reported a negative view of LAI medication. In both studies LAIs were unfavourably compared with other treatments. Bradstreet & Norris surveyed the opinions of individuals who had changed psychiatric medication within the preceding 3 years, and found that 43% of respondents who had used LAIs rated them as unhelpful compared with 38% who rated them as helpful. ^{Reference Bradstreet and Norris10} When asked specifically about symptom relief, only just over half of respondents rated LAIs as helpful. All antipsychotic LAIs (first- and second-generation) were grouped together in these results and performed poorly in comparison with other grouped medications such as all first-generation antipsychotics and all second-generation antipsychotics.

Castle et al similarly found that the LAIs group (only conventional or first-generation depots at that time) had the highest proportion of individuals rating their medication as unhelpful. ^{Reference Castle, Morgan and Jablensky11} Patients without insight were also noted to rate the usefulness of their medication lower than those with insight across all agents.

Patient preference

Data directly related to patient preference for mode of antipsychotic administration were found in two studies. ^{Reference Patel, De Zoysa, Bernadt and David8,Reference Heres, Schmitz, Leucht and Pajonk12} Heres et al questioned 300 in-patients just before discharge about their preferences in the mode of administration of antipsychotic treatment, taking earlier LAI experience into account. ^{Reference Heres, Schmitz, Leucht and Pajonk12} A predominantly positive attitude towards LAIs was conveyed with the main result found that the acceptance of LAI treatment increases with experience. ^{Reference Heres, Schmitz, Leucht and Pajonk12} Only 23% of patients naive to LAIs considered them acceptable, compared with 45% of those previously given LAIs and 73% of those currently receiving LAIs. Looking at the entire group, 40% of patients fully or possibly agreed with LAI treatment. Oral maintenance treatment was still felt to be much more acceptable, with 88% fully or possibly agreeing with the use of tablets.

Patel et al found that patients tended to prefer the current formulation of their antipsychotic: 43% of those currently on LAIs preferred LAI medication to oral, whereas 26% had no preference and 30% preferred an oral formulation. ^{Reference Patel, De Zoysa, Bernadt and David8} In contrast, 92% of those prescribed oral medication preferred tablets and only 6% preferred LAIs. Looking at the group as a whole, 18% preferred LAIs, 71% tablets and 11% had no preference.

Staff attitudes

Four of the seven studies addressing staff opinions found predominantly positive attitudes, with one study being neutral and two negative (Table 4). Using a questionnaire specifically designed for the study, Patel et al found that a large proportion of psychiatrists considered LAIs to have positive aspects, for example 71% of psychiatrists surveyed believed that LAIs are part of a patient-centred approach and 69% believed that the good aspects of LAI medication outweigh the bad. ^{Reference Patel, Nikolaou and David13} Nevertheless, a substantial minority of psychiatrists surveyed felt that LAIs were old-fashioned and stigmatising. Patel et al noted that overall the route of administration did not appear to be the major determinant of psychiatrists' attitudes, as the majority agreed that they would be persuaded to prescribe LAIs or depots if they were associated with fewer side-effects, in patients for whom adherence was an issue.

Table 4 Staff attitudes to long-acting injections: data extraction

	Attitude
Study	Positive	Neutral	Negative
Harris (2007) Reference Harris, Lovell and Day19			Of the mental health staff surveyed, 58% felt that oral medication is the option of choice if the service user agrees
Heres (2006) Reference Heres, Hamann, Kissling and Leucht18			Of the psychiatrists surveyed, 65% and 71% considered SGA–LAI and FGA–LAI respectively are inappropriate treatment for first-episode psychosis; 68% felt that neither FGA–LAIs nor SGA–LAIs are an appropriate option after relapse; 65% of respondents with schizophrenia or schizoaffective disorder had never been offered an LAI
Lambert (2003) Reference Lambert, Brennan, Castle, Kelly and Conley17		Fifty-nine per cent of all staff felt LAIs were used at the correct rate. Significant variation in attitudes to LAIs among different health professionals was evident
Patel (2003) Reference Patel, Nikolaou and David13	Of the psychiatrists surveyed, 71% believed LAIs are part of a patient-centred approach and 69% believed the good aspects of LAI medication outweighed the bad
Patel (2005) Reference Patel, De Zoysa, Baker and David14	Of CPNs surveyed, 79% believed LAIs are part of a patient-centred approach and 84% believed the good aspects of LAI medication outweighed the bad
Patel (2008) Reference Patel, Yeung, Haddad and David15	Of those surveyed, 72% felt LAIs are a patient-centred approach to treatment; 83% felt that the good aspects of LAI outweighed the bad
Patel (2009) Reference Patel, Haddad, Chaudhry, McLoughlin, Husain and David16	Of those surveyed, 89% and 75% regarded LAIs as associated with better adherence and lower relapse rates respectively; 72% felt that the good aspects of LAI outweighed the bad

CPN, community psychiatric nurse; FGA, first-generation antipsychotic; LAI, long-acting injection; SGA, second-generation antipsychotic

Similar results were found by Patel et al when the same questionnaire was applied to a group of community psychiatric nurses (CPNs) in the UK. ^{Reference Patel, De Zoysa, Baker and David14} The majority of the CPNs surveyed reported favourably on carrying out the LAI administration role and had favourable attitudes towards antipsychotic LAIs. Similarly, a significant minority held negative attitudes, such as believing that prescribing LAIs is coercive, stigmatising and old-fashioned. The authors noted that this study revealed interprofessional differences, with CPNs having less favourable attitudes to LAIs regarding patient autonomy and coercion compared with psychiatrists.

The same questionnaire from the above two studies by Patel et al was used in two further comparative studies. In the first, ^{Reference Patel, Yeung, Haddad and David15} a group of CPNs from Hong Kong were compared with the UK CPNs from the earlier study. Patel et al illustrated that although the Hong Kong CPNs were mostly positive about LAIs, overall they were much less positive than their UK CPN counterparts. The majority felt that most patients always prefer oral medication and 40% felt that force is sometimes required during LAI administration. A quarter of respondents admitted that they would feel ambivalent, negative or concerned if a patient requested an LAI.

In the second study a group of consultant psychiatrists' attitudes were assessed and compared with the previous psychiatrists' results 5 years earlier. ^{Reference Patel, Haddad, Chaudhry, McLoughlin, Husain and David16} Consultants in north-west England had a range of views regarding LAIs – only 4% of psychiatrists rated LAIs as first choice for long-term maintenance treatment in schizophrenia, and most respondents reported that their use of LAIs had decreased or not changed over the past 5 years. ^{Reference Patel, Haddad, Chaudhry, McLoughlin, Husain and David16} The majority of those surveyed regarded LAIs as associated with better adherence and lower relapse rates. The most commonly stated LAI prescribed over the past year was risperidone, and having more second-generation antipsychotics available in long-acting form was the most commonly cited factor that would persuade these psychiatrists to use more LAIs. The authors noted that concerns about stigma and autonomy were less than in the study of senior psychiatrists in the south London area conducted 5 years earlier, ^{Reference Patel, Nikolaou and David13} but worries about patient acceptance persisted.

Lambert et al questioned simultaneously both psychiatrists and allied health professionals regarding their attitudes to LAIs. ^{Reference Lambert, Brennan, Castle, Kelly and Conley17} Attitudes to the current rate of LAI use varied between different mental health professionals, with 53% of medical staff, 18% of nurses and 24% of allied health staff feeling that LAIs were being used too much; 59% of the total group felt they were being used at about the right rate. Medical staff tended not to consider patient preference as an indication for LAIs and gave lower ratings to limitation of patients' rights as a problem compared with nursing staff. Ninety-four per cent of medical staff and 81% of nursing staff reported feeling confident using oral antipsychotics, compared with 65% and 78% of medical and nursing staff respectively who reported feeling confident using LAIs.

Heres et al adopted a different approach by questioning psychiatrists as to their reasons for not prescribing a first- or second-generation antipsychotic LAI (FGA–LAI or SGA–LAI) to their patients diagnosed with schizophrenia or schizoaffective disorder. ^{Reference Heres, Hamann, Kissling and Leucht18} The majority of psychiatrists surveyed considered LAIs to be inappropriate treatment for first-episode psychosis. The high cost of the one SGA available in LAI form was important in their decision not to prescribe it in 71% of respondents and 68% felt that both FGA–LAIs and SGA–LAIs were not an appropriate option after relapse. Presumed adequate adherence with oral medication and the perceived poorer control of antipsychotic effect with LAIs compared with the oral administration of the identical drug were commonly cited as reasons for not prescribing LAIs. Participants reported prescribing LAIs to only 20% of their patients diagnosed with schizophrenia or schizoaffective disorder but also admitted that 65% of their patients with schizophrenia and schizoaffective disorder had never been offered LAI treatment.

The aim of the study by Harris et al was to develop an inventory assessing the attitudes and beliefs of mental health practitioners towards maintenance neuroleptic treatment. ^{Reference Harris, Lovell and Day19} Only one question pertained specifically to LAIs: 58% of the mental health staff surveyed felt oral medication was the option of choice if the service user agrees.

Correlation of attitude and knowledge

Heres et al found that psychiatrists aged 50 years and older offered LAI treatment to significantly more of their patients than did younger colleagues. ^{Reference Heres, Hamann, Kissling and Leucht18} The 2003 study by Patel et al revealed the positive correlation between the attitudes of psychiatrists towards LAIs and the extent of their knowledge regarding LAIs. ^{Reference Patel, Nikolaou and David13} A similar study in 2008 showed again that psychiatrists' attitudes were positively correlated with knowledge. ^{Reference Patel, Haddad, Chaudhry, McLoughlin, Husain and David16} Psychiatrists who had decreased their overall use of LAIs over the preceding 5 years had significantly lower scores on the side-effects knowledge subscale than those whose use of LAIs had remained unchanged or increased.

Patel et al illustrated that CPNs who had more than 10 patients receiving LAIs had more favourable attitudes than those who had fewer than 10 patients on LAIs, and that the degree of knowledge concerning the side-effects of these injections was again positively correlated with current attitudes to LAIs. ^{Reference Patel, De Zoysa, Baker and David14} Interestingly, a similar study of CPNs in Hong Kong showed that nurses who currently administered LAIs were more likely to have less favourable LAI-specific attitudes than those who did not, but their degree of knowledge concerning side-effects of LAIs still positively correlated with favourable attitudes to LAIs. ^{Reference Patel, Yeung, Haddad and David15}

What have we learned since Walburn's study?

Some of the criticisms put forward by Walburn et al in 2001 have been addressed. ^{Reference Walburn, Gray, Gournay, Quraishi and David1} Attempts have been made to recruit a more representative sample of people taking antipsychotic medication, with attempts at the recruitment of patients attending general practice surgeries, those staying at shelters and hostels, ^{Reference Castle, Morgan and Jablensky11} through acute in-patient admission, ^{Reference Heres, Schmitz, Leucht and Pajonk12} and through wide distribution of a questionnaire through voluntary organisations and online. ^{Reference Bradstreet and Norris10} Although this still does not address the almost impossible task of gaining the opinion of those who refuse consent to take part, it does allow for a broader view than that of only those who regularly attend out-patient clinics. Also, individuals prescribed different treatment formulations were asked their opinion about formulation preference, ^{Reference Patel, De Zoysa, Bernadt and David8,Reference Heres, Schmitz, Leucht and Pajonk12} including those previously given LAIs and now taking oral medication.

Heres et al found that patient acceptance of LAIs increased with duration of treatment. ^{Reference Heres, Schmitz, Leucht and Pajonk12} Patel et al reported that patients generally preferred their current formulation, ^{Reference Patel, De Zoysa, Bernadt and David8} but in both studies across all groups oral medication was still much more acceptable to patients, and those who had previous experience of LAIs but were currently taking an oral antipsychotic tended to find LAIs less appealing and retained negative views about them. ^{Reference Patel, De Zoysa, Bernadt and David8,Reference Heres, Schmitz, Leucht and Pajonk12} The reason for this as well as the reasons for switching treatment are unclear and an important area of further investigation.

The attitudes of different disciplines within the mental health team have been assessed during more recent years, and attitudes to LAIs have been correlated with frequency of use and knowledge of LAIs. Studies carried out by Patel et al illustrate that those with more service users on LAIs, both nurses and psychiatrists, have more positive attitudes to LAIs, with the one exception being CPNs working in Hong Kong. ^{Reference Patel, Nikolaou and David13–Reference Patel, Haddad, Chaudhry, McLoughlin, Husain and David16} However, all staff surveys by Patel et al revealed that those with greater knowledge regarding LAIs had more positive attitudes. This is perhaps hardly surprising. Heres et al also showed that psychiatrists over 50 years old offered LAIs to more patients than did younger psychiatrists, ^{Reference Heres, Hamann, Kissling and Leucht18} which could be explained by the extra experience older psychiatrists may have with FGA–LAIs balancing the marketing of SGA-orals in the past decade. Heres et al also identified inconsistencies, with the majority of respondents stating that LAIs were frequently rejected by patients when offered, but then also admitting that 65% of their patients with a diagnosis of schizophrenia and schizoaffective disorder had never been offered LAIs. ^{Reference Heres, Hamann, Kissling and Leucht18} This raises questions regarding medical paternalism and the extent of patient choice – more education and shared decision-making was advocated by Heres et al. ^{Reference Heres, Hamann, Kissling and Leucht18}

Lambert et al found varied attitudes between different members of the mental health team, with all team members being more confident using oral atypicals than LAIs for chronic schizophrenia – these authors also concluded that more education was required. ^{Reference Lambert, Brennan, Castle, Kelly and Conley17}

No data regarding the opinion and attitudes of next of kin, general practitioners or those working in voluntary organisations toward LAIs have been identified in this review. It has been shown that relatives are well aware of the risks of discontinuing clozapine, ^{Reference Angermeyer, Boitz, Loffler, Muller and Priebe20} and are aware of the positive and negative effects of antipsychotics taken by their family members, ^{Reference Angermeyer and Matschinger21} which may influence treatment decisions. Further research into carers' opinions regarding mode of treatment of antipsychotic medication would be helpful.

Impact of SGA–LAIs

Patel et al noted a reported drop in the prescribing rate of LAIs over the past 5 years during the period risperidone LAI has been available. ^{Reference Patel, Haddad, Chaudhry, McLoughlin, Husain and David16} Heres et al reported that the cost involved in the use of risperidone LAI and the fact that it was the only SGA available in LAI form had a role in the decision not to prescribe, ^{Reference Heres, Hamann, Kissling and Leucht18} and Patel et al speculated that more psychiatrists would prefer to use SGA–LAIs currently but felt limited by local restrictions. ^{Reference Patel, Haddad, Chaudhry, McLoughlin, Husain and David16} A study in Japan where only FGA–LAIs are available questioned why psychiatrists did not prescribe LAIs. ^{Reference Kanazawa, Tsutsumi, Kawashige, Nishimoto and Yoneda22} Although the lack of availability of SGA–LAIs was cited by only 13.5% of respondents, over 80% wished to use SGA–LAIs in the future. The worldwide impact of the earlier introduction of SGA–orals should not be underestimated, as these have been heavily marketed, arguably to the detriment of the older antipsychotics, including FGA–LAIs, although this has not been systematically reviewed. Second-generation antipsychotic oral medication marketing campaigns initially focused on the lower incidence of extrapyramidal side-effects with these drugs, implying greater adherence and hence a reduced need for FGA–LAIs. Nevertheless, up to 30% of individuals on maintenance antipsychotic medication receive an LAI formulation, and despite evidence of a reduced relapse rate on LAIs as opposed to equivalent oral medications, ^{Reference Haddad, Taylor and Niaz23} there does not appear to be a consensus that LAIs as a group are underutilised.

There is little evidence to support the idea that risperidone LAI has greatly changed patient or staff attitudes to all LAIs, although Heres et al found that reasons not to prescribe an FGA–LAI or depot were more common than reasons not to prescribe risperidone LAI. ^{Reference Heres, Hamann, Kissling and Leucht18} Mental health staff seemed to hold similar views regarding coercion and stigmatism before and after the introduction of risperidone LAI, ^{Reference Patel, Haddad, Chaudhry, McLoughlin, Husain and David16} and the main anxiety of staff is the presumption of patient non-acceptance. ^{Reference Patel, Haddad, Chaudhry, McLoughlin, Husain and David16}

Has the quality of the studies changed?

Following the earlier review by Walburn et al, ^{Reference Walburn, Gray, Gournay, Quraishi and David1} a number of studies have investigated patient and staff attitudes to LAIs. Using Walburn's checklist, there appears to have been some improvement in the quality of studies since mid-1999, and in contrast with previous findings there appears to be a trend towards lower quality being associated with more negative results.

Patient choice

In the UK the ‘patient choice’ agenda and Creating a Patient-Led NHS encourage patient autonomy and involvement in treatment and care decisions, ^{Reference Samele, Lawton-Smith, Warner and Mariathasan24,25} which is compatible with a recovery model for mental health, promoting a move away from medical paternalism and towards individuals making decisions about what services best meet their personal needs. ^{Reference Samele, Lawton-Smith, Warner and Mariathasan24} However, it is clear from the new research on staff attitudes that mental health staff often simply assume that patients will not want an LAI and so a large proportion do not even consider discussing this treatment formulation with them.

Ethical concerns regarding coercion associated with using LAIs persist among psychiatrists and nurses alike. ^{Reference Patel, Nikolaou and David13–Reference Patel, Haddad, Chaudhry, McLoughlin, Husain and David16} It has been suggested that possible underutilisation of LAIs in the USA (as suggested by lower LAI prescribing rates than in many other countries) may reflect concerns over ‘choice’, ²⁵ medicolegal anxieties of the prescriber, or even a lower level of community-based services for individuals with schizophrenia compared with elsewhere. However, others view LAIs as promoting voluntarism or choice in individuals who prefer LAIs owing to their awareness of reduced relapse rates and non-adherence, ^{Reference Pereira and Pinto27} and LAIs can be viewed as an ethical option by allowing the patient's mental illness to be adequately treated and therefore promoting chances of full recovery. ^{Reference Pereira and Pinto27}

Challenges associated with LAI prescribing

Summary

Since the review by Walburn et al, ^{Reference Walburn, Gray, Gournay, Quraishi and David1} new research has illustrated the attitudes and preferences of patients regarding different antipsychotic formulations, as well as of those previously given LAIs who have switched to tablets. For patients, the overall preference for LAIs ranges from 18% to 40%. ^{Reference Patel, De Zoysa, Bernadt and David8,Reference Heres, Schmitz, Leucht and Pajonk12} Preference for LAIs increased with familiarity or duration of exposure, ^{Reference Heres, Schmitz, Leucht and Pajonk12} and some but not all patients were aware of the potential benefits of LAIs. ^{Reference Patel, De Zoysa, Bernadt and David9} Voluntary patients on maintenance antipsychotics respond similarly when questioned about their attitudes to antipsychotic medication whether they are on LAI/depot or oral formulations, suggesting no intrinsic problem with LAI. ^{Reference Patel, De Zoysa, Bernadt and David8,Reference Patel, De Zoysa, Bernadt and David9} However those who have switched from LAIs to oral medications retain negative views about LAIs. ^{Reference Patel, De Zoysa, Bernadt and David8} Also, when all different types of psychiatric medication are compared, patients suggest LAIs are least preferred. ^{Reference Bradstreet and Norris10,Reference Castle, Morgan and Jablensky11}

Results from recent staff surveys in the UK revealed generally positive views about LAIs, with staff being aware of the positive aspects associated with their use, ^{Reference Patel, Nikolaou and David13–Reference Patel, Haddad, Chaudhry, McLoughlin, Husain and David16} although there is evidence that staff often assume that patients will not accept LAIs and large proportions of patients are never offered this treatment. ^{Reference Heres, Hamann, Kissling and Leucht18} There appears to be a close correlation between attitudes and knowledge regarding LAIs in both community nurses and psychiatrists, highlighting a need for continuing education and updating of skills. Finally, there are not enough data since the introduction of risperidone LAI to determine whether this first SGA–LAI has had an impact on attitudes generally.

The articles reviewed add further weight to previous studies but still are at a low level of evidence in the hierarchy of study design. Future work could focus on the attitudes of family members towards LAIs v. orals; whether offering a full, informed choice improves either adherence or outcomes; and whether negative views regarding LAIs correlate with being detained or restrained at the time of injection.