Article Text
Abstract
Background Patients with cancer are at risk of injury during treatment. Some injuries are preventable, but prevention requires knowledge about the hazards.
Aims To identify hazards and injuries relating to Danish patients with cancer (types and severity) and to test three different methods of identifying cancer-specific hazards.
Methods Adverse events in cancer care were identified through reports from healthcare staff to the Danish Patient Safety Database, a retrospective chart review using the Global Trigger Tool, and reports to the Danish Cancer Society from patients with cancer and their relatives. Events were typed using the draft International Classification for Patient Safety. Severity was assessed by Safety Assessment Coding, the National Coordinating Council for Medication Error Reporting and Prevention (NCC MERP) and qualitative assessment.
Results 2429 cancer-related adverse event reports were found on the Danish Patient Safety Database, 260 adverse events were identified using the Global Trigger Tool, and 151 safety events were reported by patients and their relatives. Cancer-specific adverse events and general safety problems were identified. In most cases injury to patients was temporary but severe and permanent injury occurred with a relatively high frequency.
Conclusion Patients with cancer are at risk of injury from cancer treatment procedures and as a consequence of problems related to administrative processes and communication. Types of identified events varied according to the methods used, and each method added new information. Further research on patient safety in cancer care and safety-enhancing activities is needed.
- Safety
- medical errors
- methods
- cancer
- patient participation
- incident reporting
- medical error
- medication safety
- risk management
Statistics from Altmetric.com
- Safety
- medical errors
- methods
- cancer
- patient participation
- incident reporting
- medical error
- medication safety
- risk management
Introduction
Patient safety is about the active management of risk. To improve safety in cancer care, risks need to be known and actions prioritised based on safety data, but there are methodological challenges associated with generating safety data.
Patients with cancer generally receive potent treatments in regimens involving various healthcare settings and several healthcare professionals. This complex and specialised treatment means that patients are at relatively high risk of iatrogenic harm.1
A disease-specific approach to patient safety is not widespread and knowledge of safety issues in cancer care is sparse. On this basis, and recognising the methodological challenges involved in monitoring patient safety, the Danish Cancer Society conducted three studies. The aim of the studies was to characterise the types of adverse events identified via different data sources in order to provide a knowledge base for safety improvement in Danish cancer care.
Methods
Three studies were conducted:
Analysis of reports from healthcare professionals on cancer-specific adverse events identified on the national Danish Patient Safety Database (DPSD).2
Analysis of adverse events identified in a retrospective chart review using the Global Trigger Tool (GTT).3
Analysis of safety events reported to the Danish Cancer Society by patients with cancer and their relatives.4
Danish Patient Safety Database
The DPSD was introduced in 2004 to support the aim of the Danish Act on Patient Safety: learning from adverse events.5 The Act required licensed healthcare professionals working in Danish hospitals to report adverse events to the DPSD. Use of the DPSD is mandatory, confidential and non-punitive. Reports are filed electronically by frontline personnel using a combination of categorical data and narrative to gather information on the person reporting the event, the context, event description, suspected causes and proposed preventive actions. The hospital risk manager edits and supplements this information with a Safety Assessment Code (SAC) and event type classification before forwarding it to the National Board of Health. Cancer-specific event reports for this study were retrieved in text searches using 39 search terms in all reports from 2004 to mid 2008.2
Global Trigger Tool
A chart review using the GTT was done in cooperation with four hospital wards performing different types of cancer surgery (gynaecological, breast, gastrointestinal and pulmonary) and one general oncology ward. The wards were located at five hospitals in different parts of Denmark. The review was carried out according to the Institute for Healthcare Improvement protocol.6 The only change was that cancer diagnosis was added to the GTT inclusion criteria. Inpatient charts were randomly sampled during a 6-month period in 2008 in each ward.
Patient reporting
The Danish Cancer Society implemented a reporting system for patients with cancer and their relatives on its website. The report form was almost identical to the DPSD report form but also gathered information on recognition of the event and subsequent follow-up. Reports filed during the period December 2008 to October 2009 were collected and analysed.
All events were categorised according to type using the draft International Classification for Patient Safety (ICPS).7 The classification was modified primarily to allow for coding of events related to ‘patient–provider communication’, but also to clarify that ‘documentation’ includes electronic documentation, that is the electronic medication record, and to separately categorise problems with medications being prescribed or administered at the wrong time. Inter-rater reliability of classification of events by type was tested on a subset of 100 reports from the DPSD. The κ coefficient was used as the statistical measure of inter-rater agreement among three reviewers.
Severity was assessed using the existing SAC in the DPSD study. The National Coordinating Council for Medication Error Reporting and Prevention (NCC MERP) was used in the GTT study according to the GTT protocol.8 9 As a result of the information gathered from patient reporting, a qualitative assessment of events was made categorising them as physical, psychological and/or social. Categorisation was done by a doctor, a pharmacist or a registered nurse.
Results
DPSD study
A total of 2429 cancer-specific event reports, corresponding to 5% of all events recorded on the database, were identified. The events were primarily related to medication, clinical processes, documentation and clinical administration. Major event types and subgroups are shown in table 1.
The agreement between the three reviewers two by two had κ values of 0.69, 0.69 and 0.68 respectively. The κ value expressing the agreement between all three of the reviewers was 0.68.
In relation to severity, 1468 events (60%) were assigned a SAC 1, 579 events (24%) a SAC 2 and 91 events (4%) a SAC 3. For 291 events (12%) SACs were missing.
GTT study
Injury was identified 260 times during the review of 572 charts. Most cases were related to clinical processes, healthcare-associated infections and medication (table 2).
Regarding severity, 185 events (71%) were categorised in NCC MERP category E, 72 (28%) in category F and 3 (1%) in category G. No events were assigned to category H or I.
Patient-reporting study
A total of 102 reports by patients or their relatives describing 151 events were analysed. Events were related to clinical processes, medication, communication, clinical administration or infections (table 3). Of the 151 events, 135 took place at hospitals and 16 in primary care; 15 of the latter concerned delay in diagnosis.
Events experienced by patients led to physical, psychological and social consequences. For 83 events, staff allegedly knew they had occurred and for 35 events (42%) patients felt they had the opportunity to discuss the event with staff.
Discussion
The studies do not produce overall incidence rates, but as shown in figure 1, findings across the three studies indicate that major hazards in cancer care are related to
medication (chemotherapy, pain treatment)
clinical processes (surgical complications, radiotherapy)
healthcare-associated infections (airway, urinary tract, wound)
documentation and clinical administration (delayed assessment of images, handling of charts, referrals, booking)
communication and information (incomplete, wrong or differing).
During the past 10 years three national programmes have been developed in order to improve Danish cancer care. None of these include patient safety. This could be because of the lack of an overview of safety hazards in cancer care. A systematic gathering of safety information could be a first step to allowing patient safety to be included in the planning of cancer care in Denmark.
A disease-specific approach to patient safety is not widespread. Most of the scientific literature about safety in cancer care concerns specific treatments and the focus has mainly been on hazards relating to medication. Our results confirm the findings of other studies identifying specific risks relating to chemotherapy.10–13 We have not been able to find much scientific literature on the other identified safety hazards relating to cancer care.
Our studies support the literature and show that the events identified are closely related to the method used and reflect the fact that information from charts, staff event reports and patient event reports is not identical but complementary. Differences in safety focus between data sources are also shown.14 15 In addition, various data sources contain information from different parts of the cancer care pathway. In our studies safety problems related to hospital treatment were the most frequent, whereas safety data on primary care and transitions were clearly under-represented. Interpretation of the results of the mixed methods approach should take into account the pros and cons associated with the various methods, but can improve our overall understanding of safety challenges across the continuum of care.
Severity of events reported across the studies shows that most injuries appear to be temporary, although there are indications that serious and permanent injury is more frequent than in the general population.16
As in all reporting systems, there appears to be under-reporting of information on the DPSD. A challenge of using data from the DPSD was that patients with cancer were not easily identified because of a lack of disease registration. According to Danish legislation, the reporting of adverse events on the DPSD in the study period was limited to the hospital sector. From October 2010 the national adverse event reporting system was expanded to the primary sector and pre-hospital care. National reporting by patients and relatives is expected by 2011. Both initiatives support the need for data across the continuum of care, but present some new challenges in terms of analysis and action.
The use of the GTT to monitor patient safety in cancer care is limited by the fact that it is a generic tool specifically focused on hospital care. According to the methodology, only events related to in-hospital treatment and information that is documented on patient charts is included. Hence the use of the GTT to achieve an overall understanding of hazards in cancer care is suboptimal, because some of the treatment is delivered in an outpatient setting, including chemotherapy and radiation therapy.
Patient and relative reporting depends on awareness of the reporting system and needs continuous promotion. Also, safety terminology is a challenge. Words that are used in the healthcare system do not necessarily make sense to patients. We invited patients to report ‘adverse events’, explained as ‘unfortunate things that were not predicted or errors that were about to happen’ and/or ‘incidents that should not occur in relation to treatment’. A subset of reports was shown to a hospital risk manager and a senior hospital physician who stated that all reports are relevant in relation to safety and/or quality improvement. The reports complement knowledge gained from inside the healthcare system, that is, about communication, information and suboptimal coordination. They stress the importance of patient follow-up after serious adverse events and describe both psychological and social consequences.
Categorising events by type is a first step in understanding disease-specific safety hazards. Various taxonomies exist but the ICPS was chosen partly to allow for exchange of knowledge internationally, and partly because a modification of the ICPS is used in the revised version of the DPSD. The experience of using the ICPS in these projects has been communicated to the National Board of Health in order to use the lessons learned when modifying the taxonomy.
Conclusion
Our studies identified safety problems specific to treating patients with cancer (ie, risks related to chemotherapy and radiotherapy) and hazards similar to those seen in the general patient population (ie, risk of nosocomial infections, surgical complications, errors in documentation and clinical administration, delayed diagnosis). The post-event dialogue between patients and healthcare professionals is also an area that needs improvement. Our studies confirm that combining multiple methods increases our understanding of hazards and support the use of this approach in prioritising preventive action. The disease-specific approach creates a platform for change and allows for dialogue between healthcare professionals involved in cancer care. This approach also offers the opportunity to make safety in cancer care a national issue and not only a matter for local improvement.
Acknowledgments
The authors thank the Danish Board of Health for cooperation on the analysis of the Danish Safety Database. They also thank the five wards that worked with them on the chart review. Last but not least they thank patients and relatives who shared their safety experiences with the Danish Cancer Society.
Footnotes
Funding Two of the three studies have been carried out under a grant from IMK Almene Fond.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.