Objective To determine the required components for developing the reporting components of a safety learning system (SLS) for community-based family practice.
Methods Multiple databases were searched for all languages for all types of papers related to medical safety in community practice: Books@Ovid, BIOSIS Previews, CDSR, ACP Journal Club, DARE, CCTR, Ageline, AMED, CINAHL, EMBASE, HealthSTAR, Ovid MEDLINE In-Process, Other Non-Indexed Citations, Ovid MEDLINE, PsycINFO, HAPI and PsycBOOKS. A grey literature search was done in Google.
Results The online search identified 190 papers. English abstracts were read and the full papers (or chapters) were retrieved for 90, of which 18 were deemed appropriate. The grey literature search revealed 18 additional papers, and an additional 12 papers were identified from bibliographies of included papers. The common themes identified from the articles became the main consideration for developing an SLS for family practice and include current and past initiatives, system design, incident reporting form and classification system.
Conclusion There is a small but growing body of literature concerning the requirements for developing the reporting component of an SLS for family practice. For the reporting component of an SLS to be successful, there needs to be strong leadership, voluntary reporting, legal protection and feedback to reporters.
- family practice
- primary healthcare
- safety learning system
Statistics from Altmetric.com
The Health and Safety Executive of the UK outlined five steps to success for managing health and safety: set policy, organise staff, plan and set standards, measure performance and learn from experience (audit and review).1 A safety learning system (SLS) fits well within the last three steps of this framework for managing health and safety.
An SLS is a system that monitors patient safety incident information with analysis to develop and implement improvement strategies to increase patient safety. It comprises two components: a reporting system and continuous learning. A patient safety incident is an event or circumstance that could have resulted, or did result, in unnecessary harm to a patient.2 Both components of an SLS have been implemented in the acute care sector in many countries.3 Very little work on this has been attempted in community-based family practice where most of the patient–physician interactions occur.
Family practice goes by many different terms, including family practice, general practice, family medicine or community practice. Compared with acute care, family practice typically sees patients of lower acuity over extended time frames. Family physicians and their staff develop long-term relationships with patients. These factors influence the type and frequency of hazards and incidents that are reported.4 Hence, it is important that systems be developed specifically for family practice/general practice to capture incidents accurately.
Reviews concerning medical safety in family practice have concentrated on methods for measuring frequency,5 6 types of errors,5–8 classification systems,9a steps to enhance patient safety7 9b and an analysis of interventions.10 None of theses reviews have described the necessary components to develop an effective SLS for family practice. This review describes the reporting component of an SLS (Wolters Kluwer, Alphen aan den Rijn, The Netherlands); a review of the continuous learning component is beyond the scope of this work.
In January 2007, a search of multiple databases was performed. The search used Ovid and included Books@Ovid, BIOSIS Previews, CDSR, ACP Journal Club, DARE, CCTR, Ageline, AMED, CINAHL, EMBASE, HealthSTAR, Ovid MEDLINE In-Process, Other Non-Indexed Citations, Ovid MEDLINE, PsycINFO, HAPI and PsycBOOKS. The search was restricted to the years 1990–2007. Table 1 contains the search strategy and terms used. Titles and abstracts were read to identify papers with focus and information on development of an SLS. Those that focused exclusively on inpatients or on other areas of medical practice were excluded as were those focusing on interventions, level of harm, preventability, incident rate and type of incident. The full text of the resulting English-language papers was retrieved and these, together with the English abstracts of the non-English citations, comprised the initial review set. Additional papers from the grey literature were identified through a Google search (table 1). Finally, all retrieved articles were screened for additional references. Thematic analysis on the resulting set was performed by PS and confirmed by SH for all included papers.
The initial OVID search identified 190 papers. After reviewing the abstracts, the full text of 90 English publications was retrieved. After the full text was reviewed, an additional 74 were excluded. This report concerns the remaining 16 articles, plus two abstracts of non-English papers. The grey literature search revealed 18 additional articles that included reviews, opinion, reports and original research. Finally, an additional 12 papers were identified from bibliographies of included papers.
The common themes identified from the articles became the main consideration for developing the reporting component of an SLS for family practice and included current and past initiatives, system design, incident reporting form and classification system. The relevant findings for each of these themes are discussed below.
Current and past initiatives
Initiatives that have been developed specifically to monitor incidents and or increase patient safety in family practice or acute care systems that have a family practice component have occurred in Australia,11–14 Canada,15 Germany,16 Israel,17 The Netherlands,18 New Zealand,18 Switzerland,19 the UK20 21 and the USA.22–25
Information on system design for a reporting system for family practice was found in six countries: Australia, Denmark, Germany, Israel, the UK and the USA. Design considerations for an effective SLS were encountered at the clinic organisation level of the family practice and at the “reporting” system level.
At the clinic level, Phillips26 discussed changes to increase patient safety and, more recently, Woodward27 modified a model from the National Patient Safety Agency to suit family practice. Seven Steps to Patient Safety for Family Practice (box 1) describes the seven key areas of activity that family practice clinics, teams and staff can work through to safeguard their patients.
Box 1 Seven steps to patient safety for family practice
Step 1. Build a safety culture
Step 2. Lead and support your staff
Step 3. Integrate your risk management activity
Step 4. Promote reporting
Step 5. Involve and communicate with patients and the public
Step 6. Learn and share safety lessons
Step 7. Implement solutions to prevent harm
Adapted from Woodward27
At the reporting level, one theoretical and 10 practical systems were identified. Singh et al28 produced a theoretical framework for engineering a reporting system for family practice that can be integrated with electronic medical records. From the 10 studies for practical systems, approximately 30 design considerations appeared at least once and 17 appeared in more than one study. Those that appeared more than once are listed in descending order of frequency in box 2.
Box 2 Design considerations
Confidential reporting (the identity of the reporter is known, but protected)29–32
There were disagreements among the studies concerning several design elements, including confidential versus anonymous reporting and the need for multiple media for reporting. Six studies used anonymous reporting (the identity of the reporter is unknown) and four used confidential (the reporter is known and this allows for further clarification of the report if necessary); of these studies, one used both.
One of the challenges of a reporting system in community is defining an incident or error.37 Elder et al38 reviewed the literature for definitions of error. They then devised scenarios representative of preventable problems experienced by family physicians and surveyed family physicians as to whether an error or mistake had occurred according to Reason's definition of error.39 They found that not all scenarios were perceived as errors by all physicians. The inconsistency revolved around three questions: “Do I know the outcome and is harm involved?”; “Is the event a common or rare event?” and “Does responsibility lie predominantly with an individual or with a system?” Elder et al present a model portraying the physician decision when assessing whether an event should be classified as an error.
The practical studies cited above have all utilised a separate reporting mechanism from electronic medical records for events. Honigman40 has shown that computerised monitoring can be used to successfully identify adverse events in the outpatient setting when an electronic medical record is used.
Incident reporting form
The basis for all reporting systems is to obtain enough information concerning an incident to decide where to apply improvement efforts. In a high-level overview of the requirements for reporting incidents, Aspden41 developed a framework for the standardised collection and codification of report data. Box 3 illustrates an adapted version of this framework.
Box 3 Domain areas for a common report format41
Who discovered/reported the event
How was the event discovered
The event itself
What happened—type of event
Where in the care process was the event discovered/did the event occur
When did the event occur
Who was involved (roles)
Why (what is the most dominant cause based on preliminary analysis)
Risk assessment, severity, preventability and likelihood of recurrence
Narrative of event—including contributing factors
Product information involved in the event
Patient information—age, gender, diagnosis, procedures and co-morbid conditions.
In the present review, four report forms were identified for the prospective collection of events within family practice on an on-going basis: Advanced Incident Management System,11 42 Applied Strategies for Improving Patient Safety Taxonomy23 43 44 and the National Health Service.45 Each utilised their own forms. The Linnaeus Collaboration/Primary Care International Study of Medical Errors,18 34 46 the American Academy of Family Physicians25 47 and the Threats to Australian Patient Safety12 used very similar forms with only slight variations. Table 2 details the fields occurring in more than one of these forms and the format (free text, fixed choice or both) of the event reports from the above six studies.
Four reporting forms had unique fields: in the National Health Service report,45 time reported, weight, causes, who reported and was it a near miss all appeared; Applied Strategies for Improving Patient Safety Taxonomy anonymous23 43 44 showed patients knowledge of incident and type of practice; Advanced Incident Management System11 42 predicted long-term outcomes (included a supplemental form for describing any further resources utilised); and Threats to Australian Patient Safety12 had a patient who does not speak English. None of the reporting forms gather all the information that is suggested by Aspden (box 3). The National Health Service comes close, but is missing likelihood of reoccurrence and diagnosis.
Multiple classification schemes have been developed to describe safety incidents in family practice (box 4). The Canadian Taxonomy of Error identified six types of errors or adverse events and 10 causal factors.15 The Linneaus/ Primary Care International Study of Medical Errors/American Academy of Family Physicians consisted of four layers: the first comprises two primary categories that distinguish between “Process” errors and “Knowledge and skills” errors, the second layer has 8 categories, the third has 13 categories and the fourth has 21 categories.4 The Joint Commission on Accreditation of Healthcare Organizations Patient Safety Event Taxonomy classified events into five complementary root nodes or primary classifications, which were then divided into 21 sub-classifications, which, in turn, are subdivided into more than 200 coded categories and an indefinite number of uncoded text fields to capture narrative information.48
Box 4 Primary care incident classifications
Applied Strategies for Improving Patient Safety Taxonomy/Victoroff Taxonomy/Dimensions of Medical Outcome43
The Advanced Incident Management System14
Kuzel patient-focused typology50
Kostopoulou multilevel taxonomy of patient safety in family practice51
PRISMA-medical (Prevention and Recovery Information System for Monitoring and Analysis)52
Rubin error classification53
Canadian Taxonomy of Error.15
Because of the lack of standard classification, the WHO was requested to develop global norms, standards and guidelines for quality of care and patient safety, and for the definition, measurement and reporting of adverse events and near misses in healthcare. The WHO responded by developing a report containing information on existing classification schemes, a draft framework for analysis of classification systems and a glossary of patient safety terms.50 The WHO also developed an International Patient Safety Event Classification (ICPS).54 The conceptual framework for the classification system includes the following primary classes: contributing factors/hazards, incident type, incident characteristics, patient characteristics, detection, mitigating factors, patient outcomes, organisational outcomes and ameliorating actions.
An SLS moves beyond an incident reporting system and promotes continuous learning through analysis of incidents and development, implementation and evaluation of improvement strategies. In this review, we have restricted ourselves to only one of the components of an SLS—the reporting component. This review uncovered four themes concerning this component: current and past initiatives, system design, incident reporting form and classification system. Multiple initiatives have been developed around the globe to report incidents and/or increase patient safety for family practice. Australia, the USA and the UK are the most active.
System design considerations were found for the clinic and the reporting system, and both levels should be considered for the successful implementation of a reporting system. At the clinic level, it is vital that leadership is committed to a safety culture.27 Without this commitment, any initiative may have limited success. The crucial considerations for implementing the reporting component of an SLS that were present throughout all the sources we reviewed were: reporting from all clinical and non-clinical personnel, well-defined inclusive incident definition, routine feedback to healthcare community, immunity from legal action and voluntary reporting. The importance of allowing reporting from clinical and non-clinical personnel is reflected by the fact that of the 10 studies that had information on system design, eight studies suggested that reports should be gathered from all personnel. This could be attributable to the differing perspectives of different types of personnel and their ability to pick up varying types of incidents. An incident that is apparent to an office assistant or nurse may not be apparent to a physician. Phillips found that clinicians were more likely to report incidents related to medication, laboratory investigations and diagnostic imaging, and staff were more likely to report communication with patients and appointments.33
A well-defined inclusive incident definition allows for a broader spectrum of reporting11 and more precise profiling of risk factors and high-risk behaviours17 It also allows participants the freedom to use their interpretation of what an incident is in the hopes that near misses (no pt. harm actually occurs) will be recorded.23 Immunity from legal action is rooted in the key principle that incident analysis should serve for learning, not for blaming,17 and it takes away the risk from reporting an incident.35
Voluntary reporting holds the greatest potential for the discovery of system vulnerabilities and improving safety,30 particularly reporting on near misses that will not appear in current mandatory reporting systems.16
At the learning system level, disagreement was apparent over several design considerations, specifically the necessity for anonymity and the reporting method. Information from confidential reports appears to be superior to that from anonymous reports and may be more useful in understanding errors and designing interventions to improve patient safety.17 23 30 32 Phillips stated that a limitation in their study was that they were unable to go back to reporters who introduced their own biases in analysis.33 Because confidential reporting appears to give better results in capturing usable information, confidential reporting (if possible) is recommended. For confidential reporting to be successful, it has been suggested that reporters should have immunity from legal actions and punitive administrative actions. The American Academy of Family Practice plans to use confidential reporting when legal protection is afforded.25 It was not possible to determine the best method for reporting (phone, fax, web) from this body of literature; thus, multiple methods are recommended. It should be noted that the cost of a phone service for smaller-scale operations may be prohibitively expensive. Whatever method is used, it should be readily accessible because systems that rely on the recall of reporters months after the incident are of limited value.13
Other considerations to ponder are a specific system, large-scale implementation, incentives and links to education activities. A specific system for capturing incident reports would be ideal as the only system not specifically designed for this was limited by the number of questions that could be asked about safety issues because the patient safety section was an add-on to their standard monitoring.13 Rather than a small-scale single practice or regional implementation of an SLS, it might be best to opt for a larger-scale (provincial/state/national) implementation. It also may be prudent to include incentives such as additional education credits or reimbursement for reporting as difficulty has been experienced in motivating physicians to report.19 Future directions for the reporting system could include the addition of links to educational activities.12
Multiple systems have been developed for taxonomy and classification. Elder et al24 suggested that clarifying definitions for harm and an improved typology of incidents are needed, and the WHO has taken this activity upon themselves. The ICPS will enable standard reporting of events, and includes the following primary classes: contributing factors/hazards, incident type, incident characteristics, patient characteristics, detection, mitigating factors, patient outcomes, organisational outcomes and ameliorating actions. The ICPS will need to be evaluated in family practice to determine its applicability.
Aspden41 developed a framework for the standardised collection and codification of report data. None of the event reporting forms reviewed gather all the information that is suggested. Two of the elements did not appear on any collection forms: likelihood of reoccurrence and diagnosis. Further research is required to determine if the inclusion of this information is necessary.
There is a small but growing body of literature concerning the requirements for developing the reporting component of an SLS for family practice. For the reporting component to be successful, there needs to be strong leadership, the system needs to be voluntary, legal protection must be afforded, and the users of the system require feedback concerning the incidents and improvement strategies. It is important to reinforce the systems approach of a culture of no blame/shame to improve patient safety. Further investigations are required on the applicability of ICPS to family practice and the required elements for an event reporting form. Also, further review is required on the continuous learning component of an SLS, including the analysis of incidents and the elements necessary for the development, implementation and evaluation of improvement strategies.
The production of this literature review was funded by Canadian Health Services Research Foundation, Canadian Patient Safety Institute and Alberta Heritage Foundation for Medical Research.
Funding Canadian Health Services Research Foundation, 1565 Carling Avenue, Suite 700 Ottawa, Ontario, Canada K1Z 8R1; Alberta Heritage Foundation for Medical Research, Suite 1500, 10104-103 Avenue, Edmonton, Alberta, Canada T5J 4A7; Canadian Patient Safety Institute, Suite 1414, 10235-101 Street, Edmonton, Alberta, Canada T5J 3G1.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.