How do your patient safety incident reports compare with your peers'?

Transcription

1 How do your patient safety incident reports compare with your peers'? NHS organisation feedback report for Anytown (Mental health) A feedback report for your organisation on patient safety incidents reported to the National Reporting and Learning System within a cluster For any queries, please contact: nrlsxtranet@npsa.nhs.uk

2 Only for new page

3 Contents 1 Introduction... 2 Key Findings How effective are your organisation s links to the NRLS? Your organisation s reporting culture compared with other organisations Further Findings How many patient safety incidents should be expected? How do your organisation s reports of degree of harm compare? How do your organisation s incident types compare? Where did the incidents reported from your organisation occur? How do your organisation s specialties compare? How do your organisation s reporting of medication incidents compare? Reporting information that could identify individuals Can your organisation improve the quality of the data submitted to the NRLS? Contact Us Appendices... Appendix 1: Terms of use Appendix 2: Detailed data Appendix 3: Queries about your data? References

4 List of figures and tables Figure 1: Reports received by month from your organisation Figure 2: Incident rate per thousand bed days Figure 3: Degree of harm to patients Figure 4: Incident type Figure 5: Top level specialties within the cluster Figure 6: Medication incident rate per thousand bed days Figure 7: The stage of the medication process at which the incident occurred Figure 8: Types of medication error Table 1: Incident location Table 2: Reports from your organisation that included person identifiable data Table 3: Summary statistics for the number and rate of incidents reported Table 4: Incident type... Table 5: Incident location Table 6: Degree of harm to patients Table 7a: Top level specialties within the cluster Table 7b: Level one and two specialties split by paediatrics/adult Table 8: The stage of the medication process at which the incident occured Table 9: Types of medication error

5 1 Introduction Purpose of the report This feedback report provides an analysis of patient safety incidents that have been reported from your organisation and collated via the National Reporting and Learning System (NRLS) at the National Patient Safety Agency (NPSA). This feedback report compares your reports of patient safety incidents with incidents reported by other mental health and learning disability organisations. It aims to stimulate your Board, clinical governance or risk team, and clinical staff to have constructive discussions on whether local reporting of patient safety incidents could be further improved, and gives suggestions for local learning to improve patient safety. The information sent by your organisation to the NRLS is vital for enabling the NPSA to identify patterns and themes, in order to provide feedback to the NHS on how patient safety can be improved. Throughout this report we have identified relevant NPSA products for improving safety which have been developed using learning from the NRLS, along with other sources of information. There are many reasons why reports of patient safety incidents differ between individual mental health and learning disability organisations. This feedback report cannot provide you with all the answers - you need to add your knowledge of your organisation s services, patients, and staff - but it can help you to ask the right questions that can lead to ongoing improvements in incident reporting and patient safety. How should your organisation use this report? The NPSA would expect the key findings section within this report to be shared at a Board level meeting, with the full report discussed and any further action planned by the group that leads on patient safety. Your departmental patient safety leads and non-executive director or governor (or in Wales, non-officer) who acts as patient safety champion should also receive copies. The medication section should be shared with your Head of Pharmacy or equivalent role. The NPSA provides a template for local staff newsletters which you can use to share key findings from this report with frontline staff. This can be downloaded from You may also wish to use this report as part of your organisation s self assessment against the Healthcare Standards for England and the Healthcare Standards for Wales. Terms of use Please see the privacy statement, usage rules, and disclaimer in Appendix 1. The National Reporting and Learning System Over 99 per cent of reports to the NRLS are received through electronic links with local risk management systems, which means that, for frontline staff, submitting a report to the NRLS involves no additional time or effort. The NPSA is very grateful to the staff in each organisation who have developed and maintained these electronic links and who regularly upload locally reported incidents. What is included in this report? This report covers all patient safety incidents successfully submitted to the NRLS by 23 November 2007, where the date that the patient safety incident occurred was between 1 April 2007 and 30 September Both reports made via your local risk management system and reports made via the eform are included. Page 1 of 35

6 Which organisations are in my cluster group? Your organisation s cluster group is mental health and learning disability organisations. This group includes organisations providing only mental health services, organisations providing mental health and learning disability services, and a small number of organisations providing only learning disability services. The NPSA uses standard benchmark group used by other parts of the NHS. If you would like to see the names of all organisations within this group, please look at Learning disability organisations need to interpret with care! Ideally, benchmark cluster groups contain broadly similar types of organisations, so that they compare apples with apples. We acknowledge organisations providing learning disability services alone are very different from organisations providing mental health services, but any benchmarking process relies on there being an adequate number of organisations within the cluster group. Therefore where only one or two organisations of a type exist, the only option is to place them within a cluster of the organisations most like them. For learning disability organisations we acknowledge this can mean they are comparing apples with pears. Where is the detail? Key findings are presented in the main report, whilst the detail of numbers, percentages and upper and lower quartiles, etc. can be found in Appendix 2. Do you have a query about your data? If you cannot reconcile this report with the number of incidents you believe your organisation has sent to the NRLS, please see the data query section in Appendix 3. Can we help? We are always willing to help with reporting to the NRLS in any way we can. Our helpdesk number is or guidance and other sources of support can be found at Page 2 of 35

7 2 Key Findings 2.1 How effective are your organisation s links to the NRLS? Successful reporting to the NRLS relies on your organisation regularly uploading patient safety incidents from your local risk management system. We expect uploads to be made at least monthly in all organisations, with fortnightly uploading appropriate in organisations with more than 50 reported patient safety incidents a week, and weekly uploading appropriate in organisations reporting more than 100 patient safety incidents a week. Figure 1 below shows the pattern of uploading from your organisation between 1 April 2007 and 30 September This figure (and this figure only) uses the date you successfully submitted incidents to the NRLS rather than the date the patient safety incident occurred. Figure 1: Reports received by month from your organisation Source: patient safety incident reports successfully submitted to the NRLS during the period 1 April 2007 to 30 September 2007 Page 3 of 35

8 What does Figure 1 tell you? If all six months include broadly similar numbers of incidents this would suggest your organisation has well established systems for regularly reporting to the NRLS, and your local risk management or clinical governance team should be congratulated. If the numbers differ dramatically over the six months, or some months show no reports submitted, it suggests your organisation has not yet established reliable systems for reporting to the NRLS. In our experience, the problems which affect regular uploading of incidents most commonly include: organisations which rely on a single member of staff who understands the local risk management system, and therefore fail to report to the NRLS if the staff member is absent or leaves the organisation; local resource issues, including long term leave or recruitment problems affecting the local entry of data (because if data is not entered for local use it is not available to upload to the NRLS); problems in linking your local risk management system to the NRLS which you have reported to the NPSA, and where we are providing ongoing advice and support to help you resolve them; a pattern of monthly reporting around the first or last day of each calendar month can lead to the appearance of less frequent reporting, for example if reports were submitted on the 1 and 31 October, the 30 November, the 3 and 31 January, etc. The NPSA would expect that if your organisation had not been reporting regularly to the NRLS, this would be known to Board level staff through internal communication before this feedback report was received, and that a plan was in place to establish regular reporting. Irregular reporters interpret with care! If your organisation has not reported regularly to the NRLS, or did not report all incidents that occurred between 1 April 2007 and 30 September 2007 by 23 November 2007, the remainder of this feedback report may be affected. For organisations who caught up with their reporting after 30 September 2007 but before 23 November 2007, these reports will not be shown in Figure 1, but will be used throughout the rest of this feedback report. Page 4 of 35

9 2.2 Your organisation s reporting culture compared with other organisations Directly comparing the number of reports received from your organisation with other organisations can be misleading, as even organisations within the same cluster can vary considerably in size and activity. To make comparison across the cluster more meaningful, we have shown your rate of reported patient safety incidents per thousand occupied bed days in Figure 2 below. The data from your organisation has been highlighted in black. Figure 2: Incident rate per thousand bed days Source: patient safety incident reports successfully submitted to the NRLS where the incident occurred during the period 1 April 2007 to 30 September 2007 Note 1: the data for one organisation in this cluster has been removed, due to an artefact in the bed days data. Note 2: Welsh trusts please note that the above rates are based on bed days for the whole organisation and not just mental health bed days. Page 5 of 35

10 What does Figure 2 tell you? In an organisation with a stronger reporting and learning culture than other organisations in the cluster, we would expect to see a higher rate of reported patient safety incidents. Experience in other industries show that as an organisation s reporting culture matures, staff become more likely to report incidents. 1 Because reports of patient safety incidents are predominantly made by frontline staff, organisations are unlikely to be able to maintain a high rate of reporting unless frontline staff see that reports are used to improve patient safety, and that staff involved in patient safety incidents are treated fairly. However, even in high reporting organisations, research suggests many incidents do not get reported 2, although most organisations connected to the NRLS are showing continued increases in reporting over time. All organisations should be aiming to further increase their reporting rates, and checking that good reporting cultures exist in all departments and all staff groups. If your organisation shows relatively low reporting rates, additional explanations might apply: however hard individual organisations are working on improving reporting rates, half of all organisations will inevitably appear in the lower half of this figure; data may be missing if you have not uploaded patient safety incidents regularly to the NRLS; the population your organisation serves, or the specialties you provide, may mean patient safety incidents are less likely to occur (for example, younger adult patients are less likely to fall than older patients). The NPSA defines a patient safety incident as any unintended or unexpected incident which could have, or did lead to harm for one or more persons receiving NHS funded healthcare. Reporting rates can be affected by local definitions and examples of the kinds of patient safety incidents that should be reported. However, we find most organisations have similarly wide definitions, usually with an additional if in doubt, report proviso. If an organisation has set definitions or examples that result in fewer incidents being reported, this may in itself suggest a difference in reporting culture. If you believe your reporting rates are below average because you have made greater improvements in patient safety than most other organisations in your cluster, you would need to have carried out case note reviews or observational studies to confirm your lower rates were due to improvements in safety rather than to under-reporting. Page 6 of 35

11 3 Further Findings Low reporters interpret with care! If your organisation has reported few patient safety incidents, the rest of this feedback report may be affected small numbers of incidents can make the detailed comparisons which follow misleading, as a single incident can affect proportions (for example, in an organisation that reported only 50 incidents, each incident will represent two per cent of the total, and make a visible difference in figures). 3.1 How many patient safety incidents should be expected? In acute settings, case note reviews and observational studies have helped establish the proportion of admissions experiencing adverse events. In mental health, similarly wide-ranging studies are rare, but there is some information on the frequency of specific incident types. Internationally, the inpatient suicide is estimated at around 14 suicides per 10,000 admissions. 3 Studies based in mental health services serving inner city areas suggest one missing patient episode occurs for every three admissions (although this often consists of multiple missing episodes by a minority of patients, and only a small percentage will not return safely) 4 and one episode of self-harm for every six admissions. 5 Data on medication errors in mental health settings is also limited. Based on reviews of drug charts, around two per cent of individual prescription items were affected by prescribing error. 6 Therefore, even in the better reporting mental health and learning disability organisations, there is still scope to improve reporting and learning. However, it is very unlikely a correct level of reporting of patient safety incidents can ever be established. As more complex treatments are developed for previously untreatable conditions, more opportunities for error can arise, even if safety improvements are happening at the same time. As expectations and standards improve, minor delays in treatment or diagnosis which were previously accepted as normal practice will become perceived as patient safety incidents. Triangulating this report with other information on safety culture Considering this feedback report in combination with other data sources on your organisation s safety culture can be helpful. For English organisations, the Healthcare Commission s staff survey at may be useful. Similar questions are included in the NHS Wales staff survey. A study found that organisations with better scores on items in the staff questionnaire (related to fair treatment of staff reporting errors, encouraging reporting, confidentiality, and preventative action taken) also had higher reporting rates per admission. 7 However, the NPSA would suggest caution in interpreting responses to question 3.4.4(England) or questions 16a-c (Wales) (for example, Have staff observed an error or near miss in the past working month? ) as a higher than average response here could indicate an organisation with staff who are more alert to patient safety incidents, rather than an organisation where more patient safety incidents occur. A study in acute settings comparing local risk management systems with case note review, patient administration data, laboratory results, complaints, claims and inquests found that all were potentially useful sources for identifying some reports of patient safety incidents not recorded by local risk management systems. Most of these data sources will also be useful for mental health settings. More detail can be found under using a broad range of data to monitor patient safety at It is also important to consider how your organisation s reporting changes over time: does your organisation undertake local analysis of year-on-year increases in reporting? Page 7 of 35

12 Can the NPSA help your organisation improve your safety culture? The NPSA provides a variety of tools which can help you improve your local patient safety culture of reporting and learning, including: Seven steps to patient safety advice on building a safety culture MaPSaF a safety culture assessment tool for mental health services The Incident Decision Tree to ensure fair treatment of staff involved in a patient safety incident Engaging clinicians a resource pack to support local initiatives to increase reporting and learning Medical error aimed particularly at encouraging reporting from junior doctors Being Open managing communication with patients, relatives and staff when a patient safety incident has occurred Chief Executives checklist how leadership from the top can influence patient safety Go to for links to all these as well as links to advice on safety culture from other patient safety organisations. Would you like to contact a high reporting organisation? Some high reporting organisations are willing to be contacted by other organisations who are seeking to improve their reporting rates, to discuss any differences in reporting systems, policy and practice. If you would like to contact organisations in the top quarter of Figure 2, please makingcontact@npsa.nhs.uk Page 8 of 35

13 3.2 How do your organisation s reports of degree of harm* compare? Figure 3 below shows the proportions of no harm and more serious incidents reported in your organisation (in black) for comparison with the proportions reported by other organisations in your cluster (in white). This figure uses percentages rather than numbers, so that organisations of different sizes can make comparisons. Some incidents do not have a degree of harm assigned to them, these are patient-group incidents where many patients were involved. In these cases it has not been possible include degree of harm, however only a minority of incidents are affected. Figure 3: Degree of harm to patients Degree of harm Source: patient safety incident reports successfully submitted to the NRLS where the incident occurred during the period 1 April 2007 to 30 September 2007 No harm Low harm Moderate harm Severe harm Death e.g. inpatient went missing, but returned safely - requiring extra observation or minor treatment e.g. fell and grazed arm, dressing applied - causing significant but not permanent harm e.g. inpatient self-harmed, required transfer to A&E for treatment - causing permanent and significant harm e.g. inpatient found unconscious in bath, anoxic brain damage - directly attributable to the patient safety incident e.g. paracetamol levels not checked in overdose patient, fatal liver failure * Figure 3 uses abbreviated definitions; see Seven steps to patient safety (page 100) for full definitions Page 9 of 35

14 What does Figure 3 tell you? If Figure 3 showed your organisation s proportion of incidents of different degrees of harm differed noticeably from other organisations in your cluster, you may wish to consider: are there errors in local coding of degree of harm? (see the quality section below); have local initiatives affected the degree of harm reported? (for example, initiatives to encourage the reporting of near misses may lead to an increased proportion of no harm incidents); do the services provided by your organisation mean you would expect differences in degree of harm? (for example, if your organisation provides specialist services for patients particularly likely to self-harm). Experience in other industries suggests as reporting cultures mature, the proportion of reports of no harm incidents increases, and the number of severe harm incidents decrease. 8 However, organisations should be very cautious in applying the findings of other industries direct to healthcare settings. For example, in the aviation industry, as reporting of near misses increased, the number of fatal incidents reduced. 9 But aviation, unlike healthcare, was starting from a baseline where all fatal aircraft crashes were known, whilst the issues for healthcare are much more complex. For example, a missed diagnosis leading to a potentially preventable death may only be detected if the patient s family consented to a post mortem examination. In other industries, there is usually a pyramid effect on severity, with minor harm incidents greatly outnumbering more serious harm, and serious injury outnumbering fatalities. However, this may not be completely applicable to healthcare; some patient safety incidents are unlikely to result in minor or moderate harm, with the impact most likely to be either no harm or death (for example, if a patient with depression goes missing from inpatient care). An analysis of high reporting organisations found no significant correlation between the number of reports made per admission and the proportions of degrees of harm. 4 Because of this, the NPSA would expect organisations to be seeking to improve local reporting levels of patient safety incidents of all degrees of harm. Are there quality issues affecting your degree of harm coding? Some organisations are making errors in their coding of degree of harm which will affecting their ability to use the data locally, as well as affect this feedback report: Some organisations confuse potential degree of harm with actual degree of harm, leading to near misses where no harm resulted coded as severe. The NPSA requires the degree of harm to reflect the actual, not the potential, degree of harm caused by the patient safety incident. Severe should be used only if the patient is expected to suffer significant permanent harm (e.g. amputation, brain damage). Currently, analysis of free text suggests only around 25 per cent of incidents reported with an outcome of severe harm were correctly coded by the reporting organisation. Page 10 of 35

15 Most mental health and learning disability organisations locally log many types of death for routine investigation, for example all mental health outpatients whose death is reported; a few organisations are in error recording the severity of all such incidents as death. For local analysis of trends, as well as for national learning, it is very important that the recorded degree of harm differentiates between those deaths unrelated to patient safety, and those deaths which are directly attributable to a patient safety incident. Although we appreciate there are grey areas, the following types of incident are unlikely to fit the definition of a patient safety incident and should not usually be reported to the NRLS: natural and expected deaths of inpatients receiving terminal care, for example, the death of a patient with advanced dementia and other illnesses on a care of the dying pathway; deaths of outpatients or former patients from natural causes; actual or apparent suicides of outpatients or former patients EXCEPT in circumstances where a patient safety incident is believed to have contributed to the death (for example, a failure to follow up as planned after discharge from inpatient care); deaths of outpatients or former patients from alcohol abuse or use of street drugs EXCEPT in circumstances where a patient safety incident is believed to have contributed to the death (for example, a delay in access to addiction services); unconfirmed hearsay reports of death, for example client in day services told staff fellow client was reported to have died last Sunday from drug overdose. We recommend you review incidents with an outcome of severe or death in your local risk management system to check whether your organisation is categorising these correctly. Page 11 of 35

16 3.3 How do your organisation s incident types compare? Figure 4 shows incident types grouped into broad categories from your organisation (in black) for comparison with the proportion of each incident type reported by other organisations in your cluster (in white). This figure uses percentages rather than numbers, so that organisations of different sizes can make comparisons. Figure 4: Incident type Incident type (Note: for full data labels please refer to Appendix 2.) Source: patient safety incident reports successfully submitted to the NRLS where the incident occurred during the period 1 April 2007 to 30 September 2007 Page 12 of 35

17 What does Figure 4 tell you? Incident type can provide useful context for the earlier figures on reporting culture. If your organisation was a high (or low) reporter, does that hold true over all incident types, or was your organisation s higher (or lower) reporting rate arising mainly from one or two incident types? Small variations between organisations are expected due to differences in local categories and sub-categories used on their local risk management system. However, if any incident types appear markedly more common or less common in your organisation than in other organisations of your type, it may be useful to ask why: could this be due to differences in inpatient population? (for example, a higher proportion of older patients may be reflected in an increased risk of patient accidents); could your local risk management system be less successful at capturing some types of incident? (for example, in some organisations pharmacists may maintain a separate reporting system that does not feed into their main local risk management system); if your organisation shows a higher proportion of incidents coded as other this may indicate problems with the local incident categories you use, or with the mapping of your local risk management system to the NRLS. The comparison of incident types reported from your organisation with incident types reported from other organisations in your cluster is intended to complement, not replace, the extensive local analysis carried out in most organisations to prioritise and target efforts to improve patient safety, and understand trends. Can the NPSA help with specific incident types? A range of NPSA resources which can help you with specific incident types relevant to mental health and learning disability organisations have been developed, including: Bedrail safety Cleaning Crash calls Patient deterioration Falls prevention Healthcare risk assessment made easy Latex allergy For links to these, please go to The site also links to useful advice from other organisations, including: Transfer of care Missing patients The NPSA is also undertaking a major project on creating a safer environment on mental health wards. For resources and advice on medication safety, please see the medication section later in this report. Page 13 of 35

18 3.4 Where did the incidents reported from your organisation occur? Not all patient safety incidents reported by organisations have occurred in their organisation. The following table compares the location of incidents reported from your organisation with incident location reported from other organisations in your cluster. Table 1: Incident location Incident location across cluster % across cluster from your organisation % from your organisation Ambulance (including call / control centre) 1 <0.1% 0 0.0% Community hospital 27 <0.1% 0 0.0% General / acute hospital 7 <0.1% 0 0.0% Mental health unit / facility 54, % 1, % Not applicable 5 <0.1% 0 0.0% Primary care setting 10 <0.1% 0 0.0% Public place % 0 0.0% Residence / home 2, % 8 0.5% Social care facility 4, % % Other % 0 0.0% Unknown % 0 0.0% Total 63, % 1, % Source: patient safety incident reports successfully submitted to the NRLS where the incident occurred during the period 1 April 2007 to 30 September 2007 Page 14 of 35

19 What does Table 1 tell you? As patients move across healthcare services, patient safety incidents which occurred outside your organisation may be reported within your organisation, because the patient safety incident caused harm that needed acute hospital treatment, or because staff from your organisation were meeting the mental health needs of a patient in acute hospital care, or because your staff pick up possible patients safety incidents made by other community staff working with their patients. For example: attended to assess patient for recurrence of depressive illness based on urgent GP referral almost immediately apparent she was not depressed but physically very ill, hot to the touch, very breathless, emergency ambulance called later contacted GP We recommend using this table as a prompt to review how you manage the investigation and learning from incidents which cut across organisational boundaries, including agreements with other local healthcare organisations on responsibilities for investigating patient safety incidents, and communication and action to reduce the likelihood of recurrence. Cross-reporting will happen in both directions, with primary care organisations or acute organisations potentially reporting patient safety incidents originating in mental health or learning disability organisations. This table can also indicate whether there are differences in the services your organisation provides that could affect comparisons with other organisations in your cluster. For example, mental health organisations will vary in the amount of activity based in inpatient mental health units, and that delivered in patients own homes. Differences in service provision can affect the type and number of patient safety incidents reported. 3.5 How do your organisation s specialties compare? Figure 5 shows the specialties where patient safety incidents occurred from your organisation (in black) for comparison with proportions of each specialty reported by other organisations in your cluster (in white). This figure uses percentages rather than numbers, so that organisations of different sizes can make comparisons. Many different specialties and sub-specialties are proved by mental health and learning disability organisations: this figure shows top level specialties within your cluster. For more detailed sub-group breakdowns please refer to Appendix 2 - table 7. Page 15 of 35

20 Figure 5: Top level specialties within the cluster Specialties (Note: for full data labels please refer to Appendix 2.) Source: patient safety incident reports successfully submitted to the NRLS where the incident occurred during the period 1 April 2007 to 30 September 2007 What does Figure 5 tell you? Variations between your organisation and your cluster in reporting specialties are expected due to differences in local service provision. However, if any other specialties appear to be markedly higher reporters in your organisation than in other organisations of your type, consider: Are there differences in service provision that explain differences in reporting? (for example, not all mental health organisations provide forensic services); Does your organisation have specialties that appear to report less than in other organisations? (if so, do they need support to improve reporting and learning?) The comparison of specialties reporting patterns from your organisation with reporting patterns from specialties in other organisations in your cluster is intended to complement, not replace, the extensive local analysis carried out in most organisations to understand and act on the particular patient safety challenges within different specialties. Page 16 of 35

21 3.6 How do your organisation s medication reports compare? Figure 6 shows patient safety incidents involving medication reported from your organisation (in black) compared to those reported from other organisations in your cluster (in grey). To make comparison across the cluster more meaningful, we have shown your rate of reported medication incidents per thousand bed days. Figure 6: Medication incident rate per thousand bed days Source: patient safety incident reports successfully submitted to the NRLS where the incident occurred during the period 1 April 2007 to 30 September 2007 Note: the data for one organisation in this cluster has been removed, due to an artefact in the bed days data. What does Figure 6 tell you? Even in high reporting organisations, research suggests many medication incidents do not get reported. 2 All organisations should be aiming to further increase their reporting rates, and checking that good reporting cultures exist in all departments and all staff groups involved in medication processes. Page 17 of 35

22 Whatever efforts are being made locally to improve reporting and learning from medication incidents, half of all organisations will inevitably appear in the lower half of this figure. If your organisation is in the lower half of the figure, it may be helpful to consider: Do local reports of medication incidents all get entered on your local risk management system? (for example, in some organisations pharmacists may report to a separate system); Do you enter enough information on medication incidents to successfully upload them to the NRLS? (see the quality section below); Does your medicines management policy include strategies for encouraging reporting and learning from medication incidents? If you believe your reporting rates are below average because you have made greater improvements in medication safety than most other organisations in your cluster, you would need to have carried out prescription card reviews or observational studies to confirm your lower rates were due to improvements in safety rather than to under-reporting. Can the NPSA help your organisation improve medication safety? The NPSA provides a range of resources and advice which can help you improve medication safety. Many additional topics and good practice examples are included in the fourth report of the NPSA s patient safety observatory on medication safety Safety in doses. Go to for links to all of these, and links to advice on medication safety from other organisations. Are there quality issues affecting your medication reports? The stage at which a medication error occurred and the type of medication incident are mandatory fields when reporting medication incidents to the NRLS. If your organisation does not record this information, you will not be able to successfully upload reports of medication incidents to the NRLS. Some descriptions of medication incidents submitted to the NRLS do not contain the name of the medication involved. Ideally organisations should use the specific field to enter the medication name that is provided in most local risk managment systems. If this is not possible, the name of the medication with checked and correct spelling should be included in the free text description of the medication incident. Correctly recording the medication name is vital for national learning, but is also essential for local understanding of what your highest-risk medications are, so that local medication safety improvements can be appropriately targeted. Do you know which five medication types are the most likely to be involved in medication errors in your organisation? Interpret with care! If your organisation has reported few medication incidents to the NRLS in Figure 6, the rest of this feedback report may be affected small numbers of medication incidents can skew the appearance of Figure 7 and Figure 8 as visible differences in proportions can actually be attributed to single medication incident report. Page 18 of 35

23 Figure 7 shows the stage of the medication process at which medication incidents occurred in your organisation (in black) for comparison with the average proportion of each stage in the medication process reported by other organisations in your cluster (in white). This figure uses percentages rather than numbers, so that organisations of different sizes can make comparisons. Figure 7: The stage of the medication process at which the incident occurred (Note: for full data labels please refer to Appendix 2.) Source: patient safety incident reports successfully submitted to the NRLS where the incident occurred during the period 1 April 2007 to 30 September 2007 Stage What does Figure 7 tell you? Small variations between organisations are expected due to differences in their services and in local risk management systems. However, if any stage of the medication process appears markedly more common or less common in your organisation than in other organisations of your type, it may be useful to question why. For example, a higher proportion of errors occurring at the prescribing stage could indicate you have excellent systems for pharmacists to detect, correct, and report prescription errors before they reach the patient, or this could indicate you provide fewer resources to support prescribers. Page 19 of 35

24 Figure 8 shows the type of medication incidents reported by your organisation (in black) for comparison with the average proportion of types of medication incident reported by other organisations in your cluster (in white). This figure uses percentages rather than numbers, so that organisations of different sizes can make comparisons. Figure 8: Types of medication error Types of medication error (Note: for full data labels please refer to Appendix 2.) Source: patient safety incident reports successfully submitted to the NRLS where the incident occurred during the period 1 April 2007 to 30 September 2007 What does Figure 8 tell you? Small variations between organisations are expected due to differences in their services and in local risk management systems. However, if any type of medication incident appears markedly more common or less common in your organisation than in other organisations of your type, it may be useful to question why. For example, a higher proportion of errors where the patient was allergic to treatment might indicate local problems with how allergy status is documented and checked. Page 20 of 35

25 3.7 Reporting information that could identify individuals The NPSA screens free text fields to find information that may potentially identify individuals, so this information can be removed. Table 2 shows the proportion of incidents reported by your organisation that had person identifiable data removed by the NPSA. The table shows the proportions of incidents that contain patient information (address, names, patient numbers or dates of birth), proportions of incidents that contained staff information and the proportion of incidents that contain other identifiable information (third party information or information we are unable to determine the nature of). Please note that no screening process is totally accurate, so the number of incidents below may be an under-estimation of the actual number of incidents that contain person-identifiable information sent by your organisation. Table 2: Reports from your organisation that included person identifiable data Type of person identifiable data % of incidents from your organisation including patient identifiable information <0.1% including staff identifiable information 0.0% including other identifiable information 0.0% Total incidents with identifiable information <0.1% Source: patient safety incident reports successfully submitted to the NRLS during the period 1 April 2007 to 30 September 2007 All organisations, including NHS organisations, are required by common law duty of confidentiality and the Data Protection Act 10 not to send any person-identifiable data to third parties, including the NPSA, without the consent of the patients concerned. The NPSA should only hold person-identifiable information with the person s consent. The reporting mechanism does not include a facility to obtain patient consent. Therefore, the NRLS is designed not to capture person-identifiable data in its reporting dataset. However person-identifiable information is included by some organisations in free text fields. While we make every effort to remove person-identifiable information from incident records, this cannot be guaranteed. Therefore organisations should avoid including person-identifiable data in all fields where free text can be entered, such as the description of the incident, minimising actions and actions taken to prevent reoccurrence. Most organisations successfully avoid this by using descriptors rather than names (e.g. the patient, staff nurse A) at the point of data entry, even if staff used real names in an original handwritten report. Page 21 of 35

26 3.8 Can your organisation improve the quality of the data submitted to the NRLS? We recognise that all organisations have to seek a balance between the amount of information they request when a patient safety incident has occurred, and making reporting as easy and straightforward as possible for frontline staff. Because of this, only a limited number of fields are mandatory when submitting information to the NRLS (failure to complete these fields means the report cannot be successfully uploaded). However, further information adds considerably to the value of reports both for your local learning and for the effectiveness of the NRLS. We would suggest the highest priorities for improvements in data quality are: Including a correctly spelt medication name in reports of medication incidents; Using the degrees of harm of severe or death correctly, including avoiding reporting deaths from natural causes as patient safety incidents and reporting the actual not potential degree of harm; Avoiding the use of staff or patient-identifiers within the free text; Increasing the proportion of reports where the free text section of actions taken to prevent recurrence is completed. Page 22 of 35

27 4 Contact Us We hope this feedback report will stimulate your Board, risk and integrated governance team, and clinical staff to have constructive discussions on whether local reporting of patient safety incidents could be further improved, and that you will be able to use the resources this report signposts to help your organisation improve patient safety. The NPSA and the national network of Patient Safety Managers are willing to help with the interpretation of this report, and with supporting you as much as we can in any local patient safety improvements. For contact details of your local patient safety manager, look under patient safety managers at We will undertake a formal evaluation of this feedback report with a sample of organisations. If you would like to comment on this report, or make suggestions for future reports, please us at nrlsxtranet@npsa.nhs.uk Page 23 of 35

28 Appendices APPENDIX 1: Terms of use The extranet information service website is hosted by the NPSA. It enables you to access aggregated data on patient safety incident reports that have been submitted to the NRLS. Your use of this information is governed by the following privacy statement, usage rules and disclaimer. Privacy statement The NRLS does not seek to collect information on the names of patients or staff involved in an incident, and takes steps to remove person-identifiable data that is inadvertently included in descriptions of the incident as far as possible. The data presented in this and future reports will not contain any information that could directly identify an individual patient or staff member involved in an incident. The NPSA will regularly publish statistics and analyses of patient safety incident data to promote a learning culture and the development of patient safety improvements in the NHS. The NPSA operates under the principle that it will share information with partner organisations where this is beneficial to patient safety, but will inform your Chief Executive first of its intention, and ensure your organisation has first sight of any data. Usage rules Our aim is to provide you with a report focused on your organisation. We will provide notes on interpretation, including where organisations should interpret with caution. Your organisation is responsible for the use of the data and the communication of the information to your staff. The data is made available to you through a secure website with password protection. It is the responsibility of your organisation to ensure that login details and passwords are restricted to authorised users only. You are responsible for authorising the appropriate people within your organisation to access the site. Disclaimer The incidents summarised in this report have been drawn from the NRLS, which supports the goal of the NPSA to make patient care safer. These incidents have been reported to the NRLS by NHS organisations across England and Wales, and are reported through a variety of routes by individual NHS staff through local risk management systems and web based eforms (including an open access eform). The individual reports are not investigated or verified by the NPSA. These incidents are self-reported and so are not necessarily representative of the NHS across England and Wales and therefore should be interpreted with care. Page 24 of 35

29 APPENDIX 2: Detailed data The following tables provide back-up detail for the figures and tables in the main report. All the caveats and cautions which apply to the figures apply to these tables, and they should be read in conjunction with the text supplied within the main report. Table 3 provides summary statistics on the number of safety incident reports. The cluster information relates to incidents that occurred during the period between 1 April 2007 to 30 September Table 3: Summary statistics for the number and rate of incidents reported From your organisation occurring (1Apr Sep2007) 1,614 reported (1Apr Sep2007) 1,269 Bed days (1Oct Mar2006) 74,748 Across your cluster Minimum number of reports 0 Lower quartile 272 Median 750 Upper quartile 1,437 Maximum 4,692 Total number of reports 62,285 Page 25 of 35

30 Table 4: Incident type Incident type across cluster % incidents across cluster from your organisation % incidents from your organisation Access, admission, transfer, discharge (including missing patient) 5, % % Clinical assessment (including diagnosis, scans, tests, assessments) % 0 0.0% Consent, communication, confidentiality % 1 <0.1% Disruptive, aggressive behaviour 12, % % Documentation (including records, identification) % 3 0.2% Implementation of care and ongoing monitoring / review % 6 0.4% Infection control incident % 0 0.0% Infrastructure (including staffing, facilities, environment) % % Medical device / equipment % 0 0.0% Medication 3, % % Patient abuse (by staff / third party) 2, % 0 0.0% Patient accident 21, % % Self-harming behaviour 10, % % Treatment, procedure % 0 0.0% Other 5, % % Total 63, % 1, % Source: patient safety incident reports successfully submitted to the NRLS where the incident occurred during the period 1 April 2007 to 30 September 2007 Page 26 of 35

31 Table 5: Incident location Incident Location Ambulance (including call / control centre) Incident Location across cluster % across cluster from your organisation % from your organisation In vehicle / in transit 1 <0.1% 0 0.0% Community hospital General areas 2 <0.1% 0 0.0% General / acute hospital Mental health unit / facility Inpatient areas 22 <0.1% 0 0.0% Support Services 1 <0.1% 0 0.0% Other 1 <0.1% 0 0.0% Missing 1 <0.1% 0 0.0% Accident & Emergency (A&E) / minor injury unit / medical assessment unit 1 <0.1% 0 0.0% General areas 2 <0.1% 0 0.0% Inpatient areas 2 <0.1% 0 0.0% Support Services 2 <0.1% 0 0.0% Community mental health facility 2, % 3 0.2% Day care services 1, % % General areas 1, % 3 0.2% Inpatient areas 47, % % Outpatient department % 2 0.1% Support Services % 4 0.2% Other % 1 <0.1% Missing 37 <0.1% 0 0.0% Page 27 of 35

32 Table 5: Incident location (Continued) Incident Location Incident Location across cluster % across cluster from your organisation % from your organisation Not applicable 5 <0.1% 0 0.0% Primary care setting Dental surgery 2 <0.1% 0 0.0% GP surgery 7 <0.1% 0 0.0% Health centre / out-of-hours centre 1 <0.1% 0 0.0% Public place % 0 0.0% Residence / home Intermediate care setting 1 <0.1% 0 0.0% Nursing home 1, % 0 0.0% Prison / remand centre 5 <0.1% 0 0.0% Private house / flat etc. 1, % 8 0.5% Other 37 <0.1% 0 0.0% Social care facility Day care services 38 <0.1% 0 0.0% Residential care home 4, % % Other % 0 0.0% Other % 0 0.0% Unknown % 0 0.0% Total 63, % 1, % Source: patient safety incident reports successfully submitted to the NRLS where the incident occurred during the period 1 April 2007 to 30 September 2007 Page 28 of 35