Article Text
PDF
Abstract
Background Errors associated with failures in filing, actioning and communicating blood test results can lead to delayed and missed diagnoses and patient harm. This study aimed to audit how blood tests in primary care are filed, actioned and communicated in primary care, to identify areas for patient safety improvements.
Methods UK primary care clinicians were recruited through the Primary Care Academic CollaboraTive (PACT). PACT members audited 50 recent sets of blood tests from their practice and retrospectively extracted data on blood test result coding, actioning and communication. PACT members received a practice report, showing their own results, benchmarked against other participating practices.
Results PACT members from 57 general practices across all four UK nations collected data on 2572 patients who had blood tests in April 2021. In 89.9% (n=2311) they agreed with the initial clinician’s actioning of blood tests; 10.1% disagreed, either partially (7.1%) or fully (3.0%).
In 44% of patients (n=1132) an action (eg, ‘make an appointment’) was specified by the filing clinician. This action was carried out in 89.7% (n=1015/1132) of cases; in 6.8% (n=77) the action was not carried out, in 3.5% (n=40) it was unclear. In the 117 cases where the test result had not been actioned 38% (n=45) were felt to be at low risk of harm, 1.7% (n=2) were at high risk of harm, 0.85% (n=1) came to harm.
Overall, in 47% (n=1210) of patients there was no evidence in the electronic health records that results had been communicated. Out of 1176 patients with one or more abnormal results there was no evidence of test communication in 30.6% (n=360). There were large variations between practices in rates of actioning and communicating tests.
Conclusion This research demonstrates variation in the way blood test results are actioned and communicated, with important patient safety implications.
- Primary care
- Diagnosis
- Patient-centred care
- Patient safety
Data availability statement
No data are available.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
Over 100 million blood tests are ordered by general practitioners (GPs) in England every year.
Patients who have a blood test need to know what the results show and what to do next; if this is not communicated effectively then patients could come to harm due to delays in follow-up testing, diagnosis or treatment.
WHAT THIS STUDY ADDS
When primary care clinicians retrospectively reviewed the electronic health records of 2572 patients who had recent blood tests, around 10% disagreed with the initial clinician’s actioning of test results.
Out of the 1132 patients where an action (such as ‘book an appointment’) was specified, there was evidence in the electronic health records that this did occur in 89.7% (varying between 45.2% and 100% in participating practices).
In 47% of patients (n=1210) there was no evidence in the electronic health records that results had been communicated to the patient.
Around 50% of participating practices who completed a follow-up questionnaire had used their benchmarked results to stimulate quality improvement (QI) activities, practice learning or educational activities.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
This research demonstrates variation in the way blood test results are actioned and communicated to patients, with important patient safety implications.
We have shown that using a collaborative model of research in primary care can help stimulate QI and could help widen participation in research beyond traditionally ‘research active’ general practices.
Background
Errors associated with failures in filing, communicating and actioning of abnormal blood test results can lead to delayed and missed diagnoses, which are a common reason for litigation in primary care.1 2 The WHO has identified that rates of test result follow-up are suboptimal, leading to serious lapses in care.3 Methods for communicating test results have changed significantly over recent years, with rapid increases in the use of text messages and online systems of communicating since the COVID-19 pandemic.4
Reviews of studies quantifying failures in test result follow-up, found between 6.8% and 61.9% of laboratory tests are reportedly not followed up in US settings, with no relevant UK studies identified.5 Surveys and qualitative research have demonstrated that most UK general practices rely on patients contacting the practice for their test results, with a lack of fail-safe mechanisms.6–8 Tools have been developed for individual practices to audit and improve blood test handling,9 yet dissemination of this shared learning between general practices is challenging.
Test handling systems are complex, involving multiple systems (hospital IT systems and primary care electronic health records) and multiple people (the patient, clinicians and administrative staff). In the UK, test results are usually received electronically into the primary care medical record, where they are viewed by the requesting clinician in their ‘pathology inbox’. The clinician can then add comments and recommended actions, such as instructions to book a follow-up appointment. The patient can receive their result online, by text, by telephone or face to face, and results can be communicated by reception staff or clinicians. Communication of test results can be instigated by the general practice, or by the patient; sometimes this will vary depending on the result. There is significant variation between general practices in how these systems of test communication are managed in practice.7
The Primary Care Academic CollaboraTive (PACT) is a new UK-wide network of primary care health professionals from England, Wales, Scotland and Northern Ireland, who collectively take part in primary care research and quality improvement (QI) projects that seek to improve patient care. There are over 600 current members of PACT, who have been recruited via social media, electronic newsletters and promotion at primary care conferences and meetings. PACT membership is free and voluntary; members receive a monthly newsletter, with no expectation to participate in PACT activities. Current membership comprises 21% general practitioners (GPs), 55% GP trainees and 7% other primary care clinicians; the remainder are students and non-clinical researchers. PACT members collect data for projects in their individual practices and the data is combined to increase the power and generalisability of the results. Practice-level data, benchmarked against other practices taking part, can be used by PACT members to identify areas for QI.
The aim of this study was to use PACT to audit how blood tests in primary care are filed, actioned and communicated in primary care, to identify areas for improvements in patient safety.
Methods
Recruitment
Full details of the methods have been published previously.10 Primary care clinicians including GPs, GP trainees, nurses and allied health professionals were eligible to participate; medical students and non-clinicians were not eligible. Information about the study was disseminated via the PACT newsletter, social media, National Institute for Health and Care Research Clinical Research Networks, GP teaching newsletters and the ‘Next Generation GP’ newsletter. It was not necessary to be a PACT member to express an interest in the study, but if a practice chose to take part, the person delivering the project was asked to join PACT (hereafter referred to as ‘PACT members’). PACT members completed an online expression of interest and consent form, a GP partner or practice manager was then required to complete a practice agreement form.
Purposive sampling was used to recruit the first seven pilot GP practices to include a range of PACT team members and a range of electronic health records systems (EMIS, SystmOne and Vision). All PACT members who expressed an interest were then invited to take part, with the aim of recruiting at least 50 practices.
Training
PACT members were required to watch two short training videos and code three fictitious clinical cases using REDCap (version 3) prior to commencing data collection (see online supplemental box 1 for details). A pass mark of >70% for each of the three test cases was set; PACT members scoring ≤70% were required to repeat the training until a pass mark was achieved.
Supplemental material
Patient sampling
Each PACT member was provided with an automated search of the electronic health records system which was used to randomly select 50 eligible patients. Eligible patients were those aged ≥18 years old, who had a blood test in primary care during April 2021. This period was chosen pragmatically to capture usual practice following the early waves of the COVID-19 pandemic, and to allow sufficient time for follow-up prior to the blood bottle shortages in August 2021.11 Pregnant women were excluded manually by PACT members due to biochemical differences in reference ranges for routine blood tests.
Data collection
The PACT member reviewed the electronic health records of each patient and manually extracted anonymised data into a REDCap (version 3) database. Data collected include patient demographics (age and gender), information on reasons for testing and type of clinician who requested blood tests (published separately12). PACT members were asked to categorise results into normal (‘all tests within the laboratory specified reference ranges’), borderline (‘one or more tests slightly outside the laboratory specified reference range’) or abnormal (‘one or more tests definitely outside of the normal range’). Borderline tests were considered ‘normal’ for the purpose of this analysis. PACT members reviewed the electronic health records to extract data on how the blood test results were coded by the filing GP: what actions, if any, were specified; whether there was evidence that steps were taken to ensure any intended action was carried out; whether the intended action was actually carried out; and whether in the clinical opinion of the PACT member, the blood test results were actioned appropriately. In cases where test results were not filed or actioned appropriately, PACT members were asked whether, in their clinical opinion, the patient was at risk of harm. PACT members were also asked if there was evidence in the electronic health records that the blood test results were communicated to the patient (see online supplemental figure 1 for details of data collection). No patient identifiable information was extracted from the patient records, PACT members reviewing the patient notes were all clinical members of staff working within the participating practices.
Supplemental material
Practice report
All participating practices received a practice report, with their own data, benchmarked against other participating practices. PACT members completed an evaluation form at the end of data collection, and a second follow-up feedback form around 2 months later, to determine whether participation in the study had led to practice learning or QI activities.
Analysis
Results were analysed using simple descriptive statistics, stratified according to whether tests were normal or abnormal. All free text comments were read and re-read, those which were not relevant to patient safety, test actioning or test communication were excluded. The remainder were classified into broad categories and summarised using tables to illustrate the number of comments within each category and illustrative quotes.
Patient and public involvement
A patient and public involvement and engagement panel comprising five people with lived experience of blood testing in primary care were involved in planning the study. The original objectives focused on using PACT to explore the reasons for blood testing (published separately12), however, patients were particularly interested in exploring how test results were communicated and shared personal experiences of difficulties in accessing test results and knowing what to do with these results. This led to changes in the research objectives and data collection tools to capture test result actioning and communication.
Results
Eligibility and consent forms were received from 149 GP practices, from which a total of 57 were recruited. 92 GP practices which expressed an initial interest did not complete the relevant study documentation required for participation or withdrew from the study (figure 1). Recruited practices came from England (n=46), Scotland (n=4), Wales (n=5) and Northern Ireland (n=2); demographics of participating practices are shown in table 1. Practice level index of multiple deprivation showed that a higher number of practices were recruited in more deprived than less deprived areas, including 19.3% in the most deprived areas. Participating PACT members included GP trainees (n=32), first five GPs (within 5 years of qualifying as a GP) (n=5), more experienced GPs (n=15), a physician associate (n=1), nurse practitioners (n=2) and practice pharmacists (n=2). Participating PACT members were required to complete three training cases; all participants exceeded the predefined pass mark of 70% on the first attempt for each case, with mean scores of 95.4% for case 1 (range 73%–100%), 93.7% for case 2 (range 80%–100%) and 94.2% for case 3 (range 73%–100%). After exclusions, data on a total of 2572 patients were included in the analyses (figure 1). Age and gender of included patients are shown in table 1, 58% of patients were female.
Participant flowchart. *Total figure slightly higher than expected because: 2 practices had problems with the automated searches identifying a large number of ineligible patients and therefore ‘topped up’ their dataset; 5 practices completed 51 and one practice completed 53 data collection forms without explanation; for 3 small practices the search returned less than 50 eligible patients. GP=general practice.
Demographics of participating practices (n=57) and patients (n=2572)
Filing of blood test results
Table 2 shows the action specified by the GP filing the blood test results, stratified according to whether all blood test results were normal or if one or more tests in the panel were coded as ‘abnormal’ by the PACT member. PACT members were asked ‘in your opinion, was the actioning made by the filing GP appropriate?’. In 89.9% (n=2311) the PACT member, with the benefit of hindsight, agreed with the actioning of the test result; 10.1% disagreed, either partially (n=183) or fully (n=78). In the 261 cases in which PACT members felt blood tests had been inappropriately actioned; 1.9% (n=5) were thought to be at high risk of harm, 41.8% (n=109) at low risk of harm, 54.8% (n=143) at no risk of harm and 1.5% (n=4) were unsure. Where PACT members indicated risk of patient harm, they were given the option of providing further information via a free text box; overall 211 free text comments were provided which are categorised and summarised in online supplemental table 1. Participants provided optional details of any inappropriate actions carried out by the filing GP. Three broad themes of inappropriate action were identified (listed in order of frequency): documentation errors, clinical errors and communication errors. Within the theme of documentation errors, the most frequently occurring issue was that medical plans were not clearly documented in the electronic health records. The most frequent issue occurring within the theme of inadequate medical care was that GPs failed to conduct further investigations following receipt of abnormal test results. The most frequent issue regarding poor communication with patients was that test results were not communicated to patients.
Supplemental material
Action coded by filing GP for normal and abnormal blood test results
Actioning of blood test results
In the 1132 cases where an action was specified by the filing GP, PACT members were asked ‘is there evidence that steps were taken to ensure the intended action was carried out?’. In 89.5% (1,013) of cases, steps were taken to action results, however, in 10.5% (119) the PACT member could not find any evidence in the electronic health records that steps were taken to ensure that the intended action was carried out. PACT members were then asked to confirm whether the action did indeed take place; overall in 89.7% (1015) of cases the action specified by the filing GP did occur, in 6.8% (77) of cases the action specified by the filing GP did not occur, in 3.5% (40) of patients the PACT member was not able to tell if the action had taken place or not. Rates of actioning of blood test results varied between 45.2% and 100% in participating GP practices (figure 2A).
(A) Proportion of tests actioned as intended, stratified by GP practice (n=57); (B) proportion of test results communicated to patients*, stratified by participating GP practices *where test result communication was evidenced in the GP electronic health records.
In the 117 cases where blood tests were not actioned as intended by the filing GP, PACT members were asked whether the patient was felt to be at risk of harm because of the unactioned test result. Overall, in 56.4% (n=66) of these unactioned patients were felt to be at no risk of harm, 38.5% (45) were felt to be at low risk of harm, 1.7% (2) were felt to be at high risk of harm, 0.85% (1) patient came to harm and in 2.6% (3) patients the PACT member was unsure about the risk of harm. Optional free text boxes were available for PACT members to provide anonymised information about risks of harm. A total of 24 free text responses were received, which were categorised into three broad themes (listed in order of frequency): inappropriate actioning of test results, poor communication of test results to patients and patients did not respond to communication. These themes, subcategories and examples from the participant data are summarised in online supplemental table 2.
Supplemental material
Communication of blood test results
Table 3 shows the methods of blood test communication, stratified according to whether test results were normal or abnormal. Text messages were used to communicate test results in 8.3% (n=213) of patients; this varied between 0% and 64% in participating practices (see online supplemental figure 2). Overall, in 47% (n=1210) of cases there was no evidence in the medical records that test results had been communicated to patients. In patients with one or more abnormal test results (n=1176), there was no evidence of test communication in 30.6% (n=360). The overall proportion of blood test results which were communicated to patients varied between 12.2% and 80.5% across participating GP practices (figure 2B).
Supplemental material
Method of blood test communication for normal and abnormal blood test results
QI activities triggered by data collection activities
54 out of 57 participating practices completed the initial evaluation form, and 34 out of 57 completed the follow-up feedback form. Overall, 19 out of 54 practices (35.2%) who completed the initial evaluation report had identified learning event(s) as a result of participating in the study. A further 16 out of 54 (29.6%) had not yet identified any learning events but were planning to do so. At the time of follow-up 18 out of 36 (50%) had used the study for QI activities, practice learning or educational activities. Participants gave details of any learning points that arose from taking part in this study, which are summarised in online supplemental table 3. They also gave details of any planned or completed QI projects that were undertaken in response to the findings of this study (see online supplemental table 4). Three broad categories of learning points were identified: administrative errors, inappropriate medical care and high prevalence of unnecessary tests. Examples of QI projects undertaken to tackle these issues include discussing the results of this study with the wider practice team, making amendments to the administrative processes associated with ordering blood tests, improving communication of test results to patients and amending specific medical protocols.
Supplemental material
Supplemental material
Discussion
This research demonstrates variation in clinical practice and risks of error and patient harm when filing, actioning and communicating blood test results in primary care. When PACT members reviewed the medical records (with the benefit of hindsight) 10.1% disagreed with how the results were actioned, either partially (n=183) or fully (n=78). For patients where an action was specified by the coding GP (n=1132) this action was not carried out in 6.8% of cases. In 3.5% it was not possible to tell from the medical records if the action had been carried out or not. There was no evidence of test result communication in 30.6% of patients with one or more abnormal blood test result results.
Although only a minority of patients were felt to be at high risk of harm or came to harm, our finding that around 10% of test results may not be actioned fully appropriately suggests that low risk errors in filing, actioning and communicating test results are common. With over 100 million blood tests ordered by GPs in England every year,13 this has important implications for patients and clinicians.
Our findings are in keeping with previous UK medical protection organisation’s database analysis which demonstrated system hazards in management of laboratory tests in 83% of 647 GP practices.14 Content analysis of 50 UK clinical negligence claims involving test result management systems in general practice found that just under half of cases involved a failure to notify patients of an abnormal test result, while 36% involved a test result not being actioned by a doctor.15
Strengths and limitations
Previous research has used anonymous event reporting by clinicians,16 and analysis of medical indemnity claims14 to categorise types of error associated with primary care blood testing. Both methods are subject to reporting bias; only incidents which reach a reporting threshold will be recognised, and near-misses are likely to be under-reported. A strength of this study was the ability to identify potential errors and near-misses in filing and actioning of test results in a randomly selected sample of primary care blood tests. Using the PACT network, we were able to recruit a broad range of GP practices in terms of deprivation, geography and ethnic diversity across the four UK nations. Using clinicians to extract data from the medical records allowed us to obtain clinical information which could not otherwise be obtained from routinely collected data. Test results which were deemed by PACT members to be ‘borderline’ (‘slightly outside the laboratory specified reference range’) were grouped together with normal results; as borderline results are likely to be common incidental findings, which may not require actioning. However, it is important to recognise that these could also reflect early manifestations of underlying disease; for example, borderline thrombocytosis is associated with an increased risk of cancer,17 so even normal or peri-normal results may have important implications which could require communication and action.
The main limitation of this study is the reliance on clinical judgement to determine whether test results were actioned ‘appropriately’, due to a lack of evidence based guidelines for actioning test results. Although good practice statements for results handling exist,18 they are not suitable for the purpose of retrospective audit. As a result, some of variability between GP practices identified may reflect differences in coding and clinical judgement between PACT members. The data shows how the clinical opinions of PACT members (who had the benefit of hindsight) varies compared with the clinician who originally filed and actioned the test results. We did not aim to provide expert independent opinion on how tests should be filed and actioned; our findings should be interpreted as demonstrating the variation in clinical practice in how tests are coded and actioned, and the potential risks of this variability. All PACT members were clinicians working in the practice, and over half were GP trainees, who might have found it more difficult to question clinical decisions made by senior colleagues/experienced GPs, potentially leading to under-estimates of rates of harmful events.
The data collection is also reliant on the quality of documentation in the electronic health records. It is possible that non-clinical staff members may not document in the electronic health records when they inform patients of their test results. Equally, clinicians dealing with multiple problems at follow-up appointments, may not document that blood test results were discussed. This may lead to underestimates of the rates of blood test communication. However, our finding that there is no documentation of test result communication in around 30% of patients with an abnormal test result has important medico-legal implications; in cases of complaints or litigation clinicians would be reliant on documentation to provide evidence that they had safely communicated an abnormal result.
We aimed to make the study as inclusive as possible; inviting all practices which expressed an interest to participate. The fact that practices self-selected to participate could potentially introduce bias, for example, practices who thought they managed blood tests well might preferentially sign up, meaning our results might underestimate the magnitude of the problem with test actioning and communication.
We achieved a broad range of participating practices in terms of geography and demographics, including a slight over-representation of the most deprived practices who are often less well represented in research; however, the findings are unlikely to be generalisable to different countries or different healthcare systems.
Conclusion
This research demonstrates variation in the way blood test results are actioned and communicated to patients, with important implications for patients and clinicians.
This study also has important implications for researchers, demonstrating the success of the PACT model, which opens up opportunities for future collaborative PACT research projects in primary care. We have shown that PACT members can collect data which requires clinical interpretation of the GP electronic health records. We have also shown that sharing benchmarked results with participating practices can help stimulate QI and could help widen participation in research beyond traditionally ‘research active’ practices.
These findings suggest that further research and QI is needed to ensure patient safety in primary care systems for actioning and communicating test results. This could include local or regional QI activities to ensure safe systems for following up abnormal blood tests appropriately, and safe systems for communicating results to patients. This is particularly important in the current context of a move towards a digital first approach to pathology communication, with the NHS in England recently rolling out online access to test results via the NHS App.19 Patient online access to test results could help improve transparency and test result communication, but also risks further shifting responsibility onto patients, and could make it more difficult for clinicians to know whether patients have reviewed their results or not.
The variation in clinical practice demonstrated in this study suggests that this is not a uniform problem, indicating that improvements should be feasible in those practices with lower rates of actioning and communicating test results. Work has already been done to develop good practice statements18 and care bundles for test result handling9; these could be used by GP practices and primary care commissioning bodies as tools to leverage change.
Data availability statement
No data are available.
Ethics statements
Patient consent for publication
Ethics approval
Ethical approval was obtained from the University of Bristol Faculty of Health Sciences Research Ethics Committee (FREC) reference 116947 and HRA approval was obtained via the Integrated Research Application System (reference 291283).
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
X @drjessicawatson, @salmanwaqar
Collaborators Primary care Academic CollaboraTive (PACT): Ola Abdellatif, Anthony Adaeze, Josephine Adebayo-Begun Fosu, Ifeoluwa Adebayo-Begun, Edemanwan Etuk, Tega Ayara, Emily Brown, Lucy Bradley, Laura Black, Georgina Butler, Mohammed Bux, Maeve Caviston, Alexandra Caulfield, Lakshmi Chandrasekaran, Anna Ciolek, Iain Cromarty, Priya Deol, Efioanwan Damisa, Clare Etherington, Chloë Evans, Rik Fijten, Ben Goude, Sarita Gorolay, Katy Horder, Alice Howe, Nicole Hurlin-Jones, Scott Jamieson, Ryan Jennison, Neeta Khare, Elizabeth Lovegrove, Peter Lillie, Jonathan Lucas, Imogen Llewellyn, Shane McAnearney, Andrzej Murawski, Alison MacTavish, Nkeiruka Sylvia Nwafor, Akinwale Olakunde, Emmanuel Okenyi, Ngozi Okeke, Tanuka Palit, Win Paing, Lucie Parry, Sarah Pocknell, Alice Rhodes, Edward Shattock, Mutiat Salawu, Hammad Shaikh, Mohammed Subhi, Jonathan Stewart, Liesl Smith, Gayle Thorpe, Paul van den Bosch, Catherine Vassallo, Jane Wilcock, Kimberly Thomas, Emma Wyatt-Haines, Oluwatoyin Yusuff.
Contributors JW came up with the original study idea. JW, AB, PD, IB-B, SH, SWDM, SW and PFW all contributed significantly to the design of the work and development of data collection tools. SH lead the patient and public involvement. JW analysed the quantitative data, AR-H analysed the qualitative free text data. JW wrote the first draft of the paper. JW, AB, PD, IB-B, SH, SWDM, SW, AR-H and PFW all revised the draft paper for important intellectual content and approved the final version for publication. JW is the guarantor. All PACT collaborative authors were involved in data collection and had the opportunity to comment on draft versions of the paper and approve the final version for publication.
Funding This study was funded by NIHR Research Capability Funding (RCF) from Bristol, North Somerset and South Gloucestershire CCG (RCF21/22-1JW) and was also supported by the NIHR Applied Research Collaborative West (project number P517).
Disclaimer The views expressed are those of the authors and not necessarily those of the NIHR or the Department of Health and Social Care.
Competing interests PD, IBB, SH, SWDM, and SW were members of the PACT committee during the conduct of this study, as well as being co-investigators. JW is current Chair of the PACT committee.
Patient and public involvement Patients and/or the public were involved in the design, or conduct, or reporting, or dissemination plans of this research. Refer to the Methods section for further details.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.