Title. Making patients safer: nurses’ responses to patient safety alerts.
Aim. This paper is a report of a study to determine whether action required by patient safety alerts was effectively taken.
Background. Over the last 10 years, there has been a growing awareness of the number of patients unintentionally harmed in the course of their treatment. Safety alerts are designed to reduce the incidence of adverse events by removing these predisposing factors.
Method. A multi-method study was carried out in 20 acute, two mental health, four ambulance and 15 primary care provider organizations in the United Kingdom in 2006–2007 using surveys, interviews with senior managers and front-line staff, collection of documentary evidence and equipment audit. The implementation of three safety alerts for nursing action is reported.
Findings. Most staff were aware of the dangers posed by gloves to staff with latex allergy, but only 20% were aware of the types of common equipment that posed a danger to sensitive patients. Almost 40% of nurses were unable to give a correct acidity value to allow nasogastric feeding to commence. One alert, on needle-free intravascular connectors, was distributed in only a few organizations as the term used was unfamiliar at all levels of the organization.
Conclusion. Healthcare providers have succeeded in setting up successful systems to disseminate alerts to middle management level, but there is evidence that implementation of recommendations by nurses is sub-optimal.
• In the last 10 years, international concern with patient safety has developed and safety alerts feature in many healthcare systems.
• Adverse events occur when latent and active errors align in a given situation, as is the case when the wrong patient receives a drug (active error) on an understaffed ward (latent error).
• Concern has been expressed that some healthcare provider organizations in the United Kingdom (UK) have been slow to act on some alerts issued through the Safety Alert Broadcast System.
What this paper adds
• National Health Service organizations in UK have set up robust systems to disseminate safety alerts, but there is little evidence of attempts made to audit implementation.
• Many ward managers are not consistently aware of the subject matter of alerts and consequently have failed to implement safer practice recommendations.
• Problems in the drafting of alerts include repetition of previously addressed subject matter and the use of terminology unfamiliar to recipients.
Over the last 10 years, there has been a growing awareness of the number of patients unintentionally harmed in the course of their treatment. Following the publication of a seminal report in the United States of America (USA) (Institute of Medicine 1999), which estimated the extent of mortality and morbidity caused by preventable adverse events, the global healthcare community began to examine strategies successfully used in the airline and other high risk industries to reduce the incidence of fatal events. A key feature of the airline industry’s safety strategy is the identification and eradication of latent errors by the establishment of an anonymous reporting system run by a central body charged with developing solutions and disseminating them across the industry (Reason 1997).
Over the last 10 years, in his work on human error and its prevention, Reason (1990, 1997) has consistently argued that errors are the result of the alignment of latent and active failures. Latent failures, which are frequently embedded in managerial systems, do not have an immediate obvious adverse effect but may contribute to one many years later. Active failures, on the other hand, have immediate observable consequences. Adverse events occur when latent and active errors align in a given situation, as is the case where the wrong patient receives a drug (active error) on an understaffed ward (latent error). Safety alerts, such as those issued by the Joint Commission for the Accreditation of Healthcare Organizations (JCAHO) in USA and Australia Council for Safety and Quality in Health Care are designed to reduce the incidence of adverse events by removing these predisposing factors.
Following the publication of key policy documents by the UK Department of Health (DH) (Department of Health 2000, 2001), a system for recording adverse events was set up by the newly formed National Patient Safety Agency (NPSA). It was intended that this agency, along with the Medicines and Healthcare Products Regulatory Agency (MHRA) and DH Estates and Facilities, would issue safety alerts to the National Health Service (NHS), based on reports of actual and potential (near miss) events both from the NHS and from international sources. These alerts are designed to reduce the likelihood of the recurrence of the error.
It was quickly realized that if patient safety was to be optimized it was not sufficient simply to disseminate alerts. An electronic system was devised to allow tracking of the uptake of the advice and subsequent action taken by the receiving organizations. The Safety Alert Broadcast System (SABS) was developed in 2004 as a channel for the dissemination of alerts prepared by the issuing agencies. In each healthcare provider organization, a named person (the SABS liaison officer) is responsible for receiving and circulating alerts and confirming that the necessary action has been taken.
From the setting up of SABS in 2004 until 31 December 2007, a total of 365 alerts were issued: 286 by the MHRA, 31 by the NPSA, 45 by DH Estates and Facilities (DHEF) and three by the Department of Health. Normally, an alert is graded with regard to urgency and includes some background information and a clear statement of actions required. Each also carries a list of people to whom the alert should be distributed and gives a date by which action must be started and a second by which action should be completed.
Evidence of action in response to safety alerts is limited and the findings mixed. Although most of the formal evidence comes from the UK, conclusions can be drawn from other data. As an example, JCAHO’s (1998) sentinel event alert on wrong site surgery was followed by the publication of a universal protocol 5 years later because of continuing reports of surgery involving the wrong patient, procedure or site. Rhodes et al. (2006), studying the NPSA alert on the same subject in the UK, found varied awareness and ‘little evidence of formal or standard policies’ of the need to mark the site of proposed surgical intervention at a time when the patient was awake and aware (Rhodes et al. 2006, p. 2). The 2004 annual report by the UK Chief Medical Officer (CMO 2004) revealed a sluggish response by healthcare provider organizations to a variety of alerts, and an influential report (National Audit Office 2005) indicated that these organizations had been slow to take advantage of the learning opportunities offered by the systems which had been established over recent years. On the other hand, Lankshear et al. (2005) found that pharmacists had taken fast and effective action in response to one of the first alerts issued by the NPSA, removing supplies of concentrated potassium chloride from all but high dependency clinical areas.
The aim of the study was to determine whether action required by patient safety alerts was effectively taken by National Health Service Trusts in England.
The study design was informed by our two previous national studies (Sheldon et al. 2004, Lankshear et al. 2005). The first phase of the multi-method study was a survey of all SABS Liaison Officers (SLOs) in healthcare provider organizations and telephone interviews with Strategic Health Authorities (SHAs) staff with responsibility for monitoring performance in this area and is reported elsewhere (Lankshear et al. 2008). The second phase involved visits, undertaken between June 2006 and 2007, to 41 NHS organizations to explore the management of the alerts and trace the action taken on a number of ‘tracker’ alerts. In this phase, we undertook semi-structured interviews with managers involved in implementation and the frontline staff who would be expected to have taken action. Finally, we collected evidence of compliance with the ‘tracker’ alerts by auditing the availability of policy documents and equipment.
Protecting people with allergy associated with latex
National Patient Safety Agency (NPSA)
26 May 2005
Reduction of harm caused by misplaced nasogastric feeding tubes
21 February 2005
Patient Safety Alert 05
Needle-free intravascular connectors
Medicines and Healthcare Regulatory Agency (MHRA)
17 May 2005
We approached a total of 80 organizations randomly selected from a list stratified by size and geographical location, successfully recruiting 20 acute provider organization and 15 primary care, four ambulance and two mental health provider organizations (n = 41).
In all organizations, we interviewed senior managers (n = 193), such as Medical Directors, Directors of Nursing, Clinical Governance Leads, Chief Pharmacists, Risk Managers, SABS Liaison Officers (n = 39) and others as indicated about the workings of the SABS system and the process of adoption and dissemination of the various categories of alert. We also undertook 181 structured interviews with ward managers, 63 with district nurses and 22 with paramedics.
Managers’ semi-structured interviews were recorded on a digital voice recorder and transcribed. Key responses were also collected on recording sheets to allow for quantitative analysis, but respondents were then encouraged to give more detail and describe specific local problems and solutions. Interviews with ward-based staff were brief, structured and recorded on recording sheets only, although these did have space to record comments and responses to open questions. We also collected evidence of the implementation of alerts by collecting policies which were later analysed and by recording observations of equipment and patient records on data capture sheets.
The proposed study was submitted to a multi-site NHS research ethics committee whose decision was that the work constituted health service evaluation and therefore did not require ethics committee approval. Research governance approval was gained from participating organizations. The prior consent of managers was obtained, but the convenience sample of ward managers was approached by researchers on the day of␣the visit only, although in most organizations the purpose of our visit had been made clear to staff in advance, along with the expectation of managers that staff would participate.
The early senior manager interviews were recorded, transcribed and analysed using thematic analysis techniques (Boyatzis 1998) until saturation was reached (34 interviews). Interview transcripts were coded independently by two experienced qualitative researchers (AJL and JH) and themes were identified. Discrepancies were resolved by discussion. The resultant coding frame was used in the analysis of all subsequent qualitative data collected using the question schedules. Quantitative analysis was undertaken using descriptive statistics and the accuracy of these data was subject to rigorous double-entry checking methods.
Credibility was addressed by the use of methodological triangulation which, within a constructivist paradigm, has been likened to laying one net upon another so that the holes in one net (limitations of the method) are partially obscured by the twine (methodological strengths) of another (Lincoln & Guba 1985). Transferability was promoted by the random selection of trusts and the relatively large number of respondents in comparison with other evaluative work of this type. Question schedules were piloted in one acute health care, one primary care and one ambulance provider organization to ensure usability, and two training days were held with data collectors to maximize reliability. Dependability was addressed by the transparency of the process. Both full report and all technical appendices are available on the Patient Safety Research Portfolio site at the University of Birmingham http://www.pcpoh.bham.ac.uk/publichealth/psrp/.
Dissemination of the alerts
Sixty-six per cent of ward managers reported receiving alerts by email, usually ‘within hours or days’ of their arrival in the organization. Likewise, most ward managers (68%) and community nurse managers (66%) received their safety alerts in electronic form, but a majority (53%) made alerts available to other staff in hard copy, frequently by placing a hard copy in a folder which others were expected to access.
Implementation of the alerts
The Patient Safety Information issued by the NPSA in May 2005 required NHS organizations to replace latex-containing equipment with safer alternatives ‘wherever possible’ and asked them to identify and take steps to protect patients with an allergy, pointing out that the lives of allergic patients could be put at risk by exposure to latex. The alert did not list equipment containing latex, although this is, in fact, extensive and includes such everyday items as mattress covers, sphygmomanometers, oxygen masks, urinary catheters and resuscitation equipment (ambu-bags).
Senior managers in all organizations in the study reported that they had begun to tackle the issue of latex allergy some time before the NPSA alert was published. We found that, in some organizations, earlier advice focusing on the risk to staff from latex gloves had given rise to confusion at both senior management and ward level. In one acute care and four primary care organizations, despite senior manager assurances that alert requirements had been incorporated into policy, the audit showed that the policy dealt only with the risk to staff posed by gloves. Despite a completion date of 31 January 2006, revised policies in four organizations visited in the year to June 2007 were still in draft form, having been neither ratified nor distributed. In some organizations, the alert itself had not been circulated to staff, as policy ratification was seen as a necessary prerequisite to the notification of staff.
A number of organizations claimed to have made boxes of latex-free equipment available, but few staff were aware of these and in some cases the boxes had not been acquired at the time of the visit. No primary care organization had made latex-free boxes available to community staff, although it should be noted that we did not investigate community hospitals.
All nurse respondents without exception were aware that the alert or local policy addressed the problems of staff in relation to latex-containing gloves, but very few (19%) knew of the possibility of patient reactions or whether they had immediate access to latex-free versions of specific equipment (oxygen masks, syringes, nasogastric tubes). Groups of respondents who had a higher-than-average awareness of latex allergy and its management worked in theatres, accident and emergency departments, intensive care units and children’s wards, where ‘champions’ had frequently carried out major work to identify latex-containing equipment and to find alternatives. In addition to asking staff about the availability of latex-free alternatives, we noted the actual availability of equipment. Whereas only 45% of respondents thought that they had access to latex-free oxygen masks and IV lines, the reality was that over 70% of wards actually had access to latex-free versions of this equipment. Also, a high proportion of syringes (87%) and a majority of blood pressure cuffs (60%) were found to be free from latex. On the other hand, the audit demonstrated that only 48% of ward areas had access to an ambu-bag that was marked latex-free.
We found respondents who either believed that an organization ‘moving towards latex-free status’ was entirely latex-free (4) or others (3) who evinced blind faith in the organization to provide equipment that would not present a threat:
You assume equipment is safe for use. (Community nurse)
They wouldn’t provide this stuff if it wasn’t safe. (Ward manager)
We found latex-free markings on equipment to be an unsatisfactory source of information for practitioners, because in some cases such markings did not exist (on mattresses) or because the information was given in small print amongst many lines of other information, or was too small to read (see Figure 1). Therefore, even if warned that a patient was highly allergic to latex, staff might have no access to information to determine whether a specific item was safe for use. At present, there is no universal symbol indicating whether an item of equipment does or does not contain latex, although international negotiations are ongoing (NPSA, personal communication).
Both mental health organizations were found to have a glove policy which dealt only with the care of affected members of staff. It was, however, pointed out that gloves are used more frequently in this setting for non-clinical procedures such as searching patients. Two of the four ambulance organizations had taken the decision to purchase entirely latex-free equipment and, having taken this step, had not circulated the alert to staff nor warned them about the issue.
The naso-gastric tube alert, issued by the NPSA in February 2005, was explored in acute and primary care organizations only. The alert required managers ‘to provide staff, carers and patients in the community, with information on correct and incorrect testing methods’. It specifically banned the use of the ‘whoosh test’ (auscultation following the injection of air through the tube), and reinforced earlier messages to the NHS that the use of litmus paper was unreliable and should be replaced by pH paper. X-ray was to be used where results were equivocal.
Five of the 15 primary care trusts did not circulate the alert, deeming it to be irrelevant. The research visits revealed that 15 acute and five primary care organizations claimed to have written or revised a policy to reflect the NPSA guidance. Whilst the majority accurately reflected the NPSA guidance, one acute trust’s policy, distributed to all wards, specifically advocated the use of the ‘whoosh test’ in a phrase that omitted one key negative word:
The whoosh test should be used in isolation.
In this organization, all senior staff were confident that the issue had been addressed effectively because their nutrition specialist nurse was at the forefront of practice. Three other acute organizations were found to be dependent on The Royal Marsden Hospital Manual of Clinical Nursing Procedures (Dougherty & Lister 2004), whose 6th edition predated the NPSA safety alert.
It was noted that pH papers were only available on 36% of wards, while litmus papers were seen on 9% (16) of wards visited. A total of 114 (67%) acute organization nurses (AONs) and 20 (31·7%) district nurses (DNs) stated that they would check that a nasogastric tube was in the stomach rather than in the lungs at the time of insertion by testing gastric aspirate using pH indicator paper, with 13 (8%) AONs and 10 (4·8%) DN interviewees using the auscultation method (whoosh test). Fifteen (9%) AONs and 21 (33·3%) DNs stated that they used litmus paper and 69 (41%) AONs and 18 (28·6%) DNs used X-ray. Several respondents qualified their responses by stating that they might use X-ray if no aspirate was obtainable, if there was any doubt, or for certain patient groups. Interviewees were also asked to state the accepted range for pH paper to allow a feed to be given. Of the 130 respondents who replied to this question, 16% (21) said they had never used it, 22 (17%) AONs and nine (14·3%) DNs said the range was between 1 and 3, 42 (32%) AONs and 12 (19·0%) DNs gave the range as between 4 and 6 and five (4%) AONs and three (5%) DNs that it was from 7 to 9. Forty (31%) AONs and 22 (34·9%) DNs said that they did not know. Several nurses named the required colour of the reagent strip but could not state the actual pH value.
Needle-free intravascular connectors
The needle-free intravascular connectors (NFIVCs) alert, issued by the MHRA in 2005, warned staff that manufacturers had recently changed their instructions in respect of these small devices, which are inserted into IV cannulae to allow the administration of medication. The instructions drew attention to the frequency with which these devices should be changed and the correct processes for disinfection before use. The alert, however, did not make reference to the word ‘infection’ for legal reasons.
Only three (6%) senior managers in acute organizations were able to state how this alert had been handled within organizations. In most, the issue had seemed to pass unrecognized as a device that was used in almost every ward. When asked whether community nurses gave intravenous drugs, 13 (44·8%) community nurse managers claimed that they did, whilst 12 (41·4%) said they did not and one did not know. Respondents said that community nurses were flushing lines frequently. Six respondents (20·7%) from this group stated that they distributed the alert, three (10·4) did not and eight (34%) did not know.
Sixty-three per cent of ward managers interviewed used NFIVCs, although at ward level, again, few respondents recognized the term used in the alert, referring to the devices as ‘bungs’, or by their trade name. A series of questions were asked about the use of NFIVCs. In response to the question ‘Is there a prescribed time period after which you have to change NFIVCs?’, 19% believed this was 24 hours or less, 46% between 24 and 72 hours, and 18% between 3 and 7 days. A further 12% did not know or did not believe that there was time period.
In response to the question ‘How do you keep track of how often a connector has been used or how long it has been in use?’, 33% marked this on the patient’s care plan, 15% on the chart at the foot of the bed, 7% on the medicine chart, and 21% said it was not an issue as they used a connector only for a short time. A further 17% gave other explanations, such as marking the date and time on dressings or fluid charts or on the connectors, or having a fixed regime for changing them.
In response to the question ‘Which disinfectant do you use before attaching a syringe or giving set to a valve connector’, 94% said they used an alcohol wipe. In response to the question ‘Are you aware of any particular instructions regarding contact time with disinfectant’, 9% said not <2 seconds, 15% between 3 and 5 seconds, 12% between 6 and 10 seconds and 58% said they just wiped the connector. Seven per cent said that they did not use anything. While nine commented that they did not know, or were not aware of any guidance.
In the ward audit, we attempted and failed in all but one case to locate instructions for use, as suggested by the alert. This failure is attributable to the fact that these small devices are delivered to wards by means of a supply top-up system, which delivers the required number into plastic containers. The instructions are normally discarded with the empty box at the end of the supply round.
Over the last 9 years, the systematic publication of safety alerts has become a feature of many healthcare systems across the world, yet the literature gives little evidence of the uptake of their requirements in practice, or of their impact on the number of adverse events reported. This study is therefore an important addition to the literature.
Despite undertaking a robust study we recognize that, by the nature of the methodology employed, the study has a number of limitations. First, all the organizations agreed to participate and it would be reasonable to assume that there may have been some degree of self-selection involved; the volunteers might have had greater confidence in their systems than those who declined to join the study. However, the evidence that we collected indicated considerable variation in both quality of process and outcomes. Second, all the organizations knew that we were visiting and on which alerts we were focusing and could possibly have taken action in advance of the visit. We believe, however, that our audit trail of policies and records and observation of equipment lent weight to our findings. Finally, due to the inevitable time lag between receiving funding, starting the study and undertaking site visits, the alerts aged and memories became weaker. In the last months of the study, we were interviewing staff about alerts issued two years earlier.
Whilst the incidence of latex allergy in the population is unknown, the evidence is that it is the growing (Meric et al. 1998, Bowyer 1999, Lieberman 2002) NHS organizations, castigated before the study for not having done enough to alert staff (NPSA 2004), appeared to have taken slow and relatively ineffective action to address this shortcoming. The timescale required for policy development was unacceptably long, especially in cases where the original alert had not been distributed pending managerial action. Ward managers, highly aware of the risk to colleagues with a latex sensitivity, appeared, with the exception of those working in operating theatres, paediatrics and other high dependency areas, to be relatively unaware of the risk to patients. It appeared that in some organizations, earlier advice from the Medical Devices Agency (MDA 1996), Department of Health (HSC 1999/999) and Health and Safety Executive (HSE 2002) had caused senior managers and others to assume that action had already been taken on the issues dealt within the NPSA alert.
Professional awareness of the requirements of the nasogastric tube alert is also disappointing, especially in the light of the data released by the NPSA in January 2007 concerning seven deaths and four near misses caused by misplaced nasogastric tubes in children and adults:
In two cases the whoosh test was used; one case where it was the sole test and one where it was also used and a subsequent x-ray for deterioration was misinterpreted. In one case the gastric aspirate was possibly contaminated through squirting the aspirate onto the paper wrapping of the syringe and then running the reagent strip through the aspirate. All other cases including the four near misses involved misinterpretation of X-rays. (NPSA 2007)
Poor uptake of the alert on NFIVCs appears to stem from the alert writers’ failure to ensure that the term used would be understood by recipient organizations, and from equipment distribution systems at ward and clinic level that make instructions unavailable to staff. In addition, the alert writers were prevented from using the word infection in the alert (MHRA, personal communication) after a challenge by the manufacturers of the devices. Two years later, the Scottish safety action notice (NHS Scotland 2007) was able to refer to the specific danger, citing the epic2 guidelines on hospital infection (Pratta et al. 2007).
It appears that, for a variety of reasons, many nurses are missing or being denied the opportunity to learn from adverse events, including deaths that have occurred in other parts of the NHS. As a result, they are failing to protect patients and are guilty of falling into a trap described by Cohen:
Unfortunately there are too many in health care who feel that if it hasn’t happened to them the adverse events of others do not apply. (JCAHO 2001)
Sadly, the profession is not alone in failing to grasp the opportunity to make patients safer. A report in the New England Journal of Medicine stated that neither physicians nor healthcare users understood the scale of the problem of adverse incidents (Blendon et al. 2002). The UK National Audit Office report (National Audit Office 2005) commented on the results of an interactive campaign conducted on behalf of the NPSA through the doctors.net website, which concluded that 60% of junior doctors appeared to have no awareness of the work of that agency. In addition, the work by Lankshear et al. (2005b) indicated that junior doctors were far less likely to be aware of the NPSA’s alert on the dangers of concentrated potassium chloride solution than nurses, and the findings with regard to correct identification of surgery site on both sides of the Atlantic indicate that even senior doctors cannot be relied upon to accept and implement safety recommendations, despite the fact that to do so would render them less likely to litigation (JCAHO 2003, CMO 2004, Rhodes et al. 2006).
More than half a million doctors and nurses are currently employed in the UK NHS. Ensuring that all of them receive notification of alerts issued by the statutory agencies is a major undertaking. Whilst the adoption of an electronic dissemination system within the NHS has ensured better dissemination to middle managers (those whose work revolves around information technology systems), it has had less impact on frontline staff, whose access to computers may be less certain. Once the dissemination problem is resolved, ensuring that these same alerts are read and then put into practice is infinitely more difficult. The literature on the problems of introducing practice based on safety, like research evidence, tells us that staff have to be persuaded to ‘buy into’ new practices (Leape 2005). Research into the management of hazards in a variety of industries indicates that safety precautions are more likely to be implemented if they are convenient to apply, non-compliance is socially unacceptable, the consequences are perceived as important and the chance of detection is high (Williams 1996). What the airline industry has achieved, but the British NHS has thus far failed to do, is ensure that all of these optimizing conditions are met.
It is consistently argued in the safety literature that people will, for a variety of reasons, make errors and managers are exhorted to put sufficient safeguards in place to ensure that harm does not result from these. It is clear from this study that the very process of managing and implementing these alerts is not error-free, and that in many organizations, these have been unsuccessful in consistently changing behaviour and improving patient safety. In USA, compliance with Sentinel event alerts is a component in acquisition of JCAHO accreditation. This year, in UK, the Healthcare Commission will require evidence of compliance with safety alerts as part of its NHS monitoring role, a move likely to raise the profile of alert implementation and as a result increase patient safety.
AL and KL were responsible for the study conception and design. AL, JH and PL performed the data collection. AL, KL, JH and RS performed the data analysis. AL and KL were responsible for the drafting of the manuscript. AL made critical revisions to the paper for important intellectual content. KL provided statistical expertise. AL and KL obtained funding. RS provided administrative, technical or material support. AL and KL supervised the study.