Results Handling in Primary Care

Results Handling in Primary Care

1 Summary

Ordering tests is a normal part of clinical life. Every day,X rays and investigations are requested, andbodily fluids and biopsy specimens are sampled and transported from healthcare sites to laboratories. These events are commonplace in healthcare.

However, there is a fundamental question in relation to test ordering and results handling that all general practices need to consider:

“do you know whether the results of tests are returned and acted on in a timely manner?”

There is evidence to suggest that in the UK, practices do not always have adequate systems to track requests for tests, to record the results of tests, or to confirm whether follow-up action has taken place.Some practices also lack protocols on how to inform patients of results.

Poor test follow-up, missed results and delay are likely to have a greater impact on patients whose test results point to the need for timely treatment. However, the evidence suggests that the majority of general practices do not have effective processes in place to notify patients of normal results either.

In this context, harm can be understood in terms of failure to diagnose, misdiagnosis, delayed diagnosis and failure tooptimally manage patients. In general terms the impact of each of these is likely to effect the morbidity of patients. However, it is difficult to get a rigorous measure of the harm caused. Nevertheless it is easy to identify the type of harm that may arise at various stages in the results handling cycle.

For the development of improved results handling processes in primary care, the research highlights a number of key issues. While electronic systems provide the potential to save time, they must be integrated with existing systems so that clinicians can access test results as part of the patient’s medical record. As a further time saving test results need to be flagged according to urgency and in a manner that makes it apparent that action is required. There is evidence to suggest electronic and manual systems in practices do not always work effectively.

Ideally there should be clearly defined and understood system establishing who is primarily responsible for reviewing all results and bringing these results to the attention of the relevant clinician and similarly . The results handling process within each practice needs to clearly identify who is responsible to receiving and actioning results.

This report provides some guidance for the development of results handling processes within general practice. Where appropriate, UK Standards and Guidelines are hyperlinked in this report. There is an extract of a conceptual model of the results handling process developed by the Medical and Dental Defence Union of Scotland. Screen shots of commercially available results handling systems are also provided to demonstrate the key features within these.

2 What is the scale of the problem?

While laboratory processes are tightly regulated the pre- and post-analytical process are not controlled in the same way. These processes are less prescriptive and are more heavily influenced by human factors. In the absence of an applicable standard for handling results these processes are more prone to fail.

A study looking at 400 Clinical Risk Self Assessments undertaken by practices in the UK between 2004 and 2006 found that 84% of practices had risksassociated with test results. The most common issues identified were:

No system to record returned results.
Test results are not recorded onto the computer.
No system to ensure follow up with patients.
Non-clinical staff are allowed to inform patients of their results (Wilson, 2007, pp6-7).

Failure rates

Poon et al (2004) from a survey of General Practices in the USfound that only 52% of practices keep a record of tests ordered and only 32% have a system to identify is a patient has missed a test. Only 41% of practices reported feeling even somewhat satisfied with their results management system (p2225). Looking at the prevalence of processes to manage results and low levels of satisfaction among clinicians with existing systems, there appears to be plenty of scope for improvement.

A range of studies looking at hospital inpatients demonstrate that even in a controlled healthcare environment there is a failure to follow-up results in 1% – 22% of patients (Callen, 2011, p195).

Looking at the same issue from a different perspective the rate of failure to follow-up per test requested, identified an error rate ranging from 20% to 62%. It should be noted, however, that some of these figures may be over-calculations as the most prominent method for determining these fail rates are case note reviews, where it is possible that follow-up occurred but was not documented (Callen, 2011, p195).

Poor Test Follow-up

Poor test follow-up has been identified as one of the major processes contributing to unsafe patient care. Missed or delayed diagnosis is the principle risk of poor results handling. Suboptimal clinical outcomes may result and create potential legal issues for clinicians (Callen, 2011, p194).

An analysis by the Medical Protection Society (MPS)reports that the single most common reason for legal action against GPs is delay or failure to diagnose. Of 1000 consecutive claims for medical negligence against GP 539 claims were delay, failure or wrong diagnosis (Silk, no date, p4). Considering the important role of results handling in diagnosis, this analysis points to an area that needs more attention. This finding is mirrored in analysis of US malpractice cases which show that about 25 per cent of diagnosis-related malpractice cases are due to failures in systems to follow-up results (Poon et al 2004 p 2223).

Wahls et al (2007) report that ‘missed’ results leading to clinically important treatment delays arean important and likely underappreciated source of diagnostic error.The most commonly missed results included:

imaging studies (29 percent),
clinical laboratory (22 percent),
anatomic pathology (9 percent),
other (40 percent).

The most common diagnostic delays were for cancer (34 percent), endocrine problems (26 percent), cardiac problems (16 percent), and others (24 percent). While these results are based on an American study they may highlight areas for greater attention in the UK.

Moreover, a study conducted in the US found that 75 per cent of GPs did not routinely notify patients of normal test results. Fewer than 25 per cent had a reliable system in place to identify when a patient was overdue for follow-up of abnormal results (Poon et al 2004 p 2223).

Delay

Research literature from developed countries suggests that:

in an inpatient setting over 25% of treatments were delayed by twice the clinically recommended time because of poor test-results follow-up (Kuperman GJ et al. 1999 cited in WHO 2008 p.77).
only51% of laboratory results indicating potentially life threatening conditions were followed-up by appropriate treatment (Tate KE, 1990 cited in WHO 2008 p.77).
One survey of primary care providers whose patients were discharged with actionable test-results showed that only 61% of results were communicated to primary care (Roy CL et al. 2005 cited in WHO 2008 p.77).

Delays in receiving laboratory test results and dissatisfaction with the management of laboratory test result information are commonly reported. However, the literature above suggests that the processes within each healthcare provider’s control have a greater impact on the timeliness of clinical follow-up.

Record keeping

Cullen (2011) acknowledges that recordkeeping is a problem with results handling. It is noted that the failure rates listed above may be over-estimations, but that the poor record keeping on test results and follow-up action may make the problem seem greater.

Poon et all (2004) compared systems for handling results between general practices in the US and report that patient records documented patient notification of results in between 58% and 85% of patients. Patient records contained the clinician response to the result between 47% and 84% of the time. Follow-up for abnormal results was recorded for between28% and 55% of patients (p 2223).

3 Where is harm occurring?

The number of research studies looking at harm stemming from failures with results handling is limited. In a review of studies it was found that failure to follow-up gave rise to:missed diagnoses ofmalignancy, hypothyroidism,hyperthyroidism, osteoporosis;inappropriate or unnecessary antibiotics prescribed; microbiological results which necessitated thestarting or changing of antibiotic therapy; and positiveserological test results for Helicobacter pylori(Callen, 2011, p195).

The results handling process in NHS Borders has been mapped at a high level and is outlined in figure 1 below. The section in the greyed box describes the laboratory process. This process is outside the focus this paper. The pre and post analytical processes are the focus of this paper.

Most laboratories in the developed world are part of some kind of quality assurance and accreditation scheme. These are most commonly voluntary scheme, some are mandatory. In the UK, the National External Quality Assessment Service (UKNEQAS) and Clinical Pathology Accreditation (CPA) (UK) Ltd fulfil these roles. Membership of these schemes is voluntary.

These assurance and accreditation schemes look at individual test results and oversee the process of delivering these results by measuring performance against agreed standards.

Evidence suggests that regulatory controls are highly effective in ensuring that procedures within the laboratory analytical process are accurate and produce the right results. Laboratory performance is inspected and audited against established standards every two years. As a consequent errors occurring within the laboratory process are very rare and occur in only 0.03% of all test results(Valenstein PN et al, 2004,pp 979–996).

In the absence of robust literature identifying harm, a description of possible errors that could lead to harm has been developed to illustrate the risks associated with results handling.

Figure 1: The Results Handling Cycle

Pre-analytical stages

Requestor decision to do test

There has been little research of any kind on requestor decisions on which test to do. The decision will be based on the diagnostic training the requestor has had, the existence of local, regional or national guidelines (including QOF) and to a certain extent what is actually available locally. Ease of requesting will also tend to enter into the equation. The existence of pre-set profiles of laboratory tests (eg U&Es) means that the requestor may get tests where there is no real indication or need in most cases (eg bicarbonate in primary care). Although not a direct effect on patient care, this would tend to suggest that some laboratory testing is a ‘reflex action’ which does not actually reflect what tests the patient actually needs.

Request ordering

This can be done by filling in a request form or by electronically test ordering. The traditional hand-written request form has often a list of pre-set profiles (for ease of requesting and avoidance of legibility problems). The down side of pre-set profiles is the possibility of over-requesting or receiving tests the patient does not need. Electronic test ordering can also have pre-set profiles (with the same issues as above), but there is the potential in some ordering systems to have intelligent requesting guidelines to assist the requestor in test selection (Westbrooke et al, 2006,p533).

In some instances, it is also apparent that the requestor is not actually the named requestor on the form. It could be a practice nurse following a GP lead protocol and using the GP’s name as the requestor. Additionally it could be other healthcare workers who request a test for the care of the patient with the GP identity. Further confusion is also possible if the requestor asks for a copy of the results to go to the GP (who might not be aware of any healthcare issue). This can lead to issues of responsibility of actioning abnormal test results. The General Medical Council’s (GMC) Good Medical Practice Guidelines on sharing information with colleaguesattempt to shed light on the responsibility of the requestor in this situation.

Phlebotomy

This can be done by the requestor themselves, but is commonly done by a third party phlebotomist. Issues here are the time delay between requests being made and the blood being withdrawn and the presence or absence of tracking between requestor and phlebotomist. We are unaware of any studies that investigate sample tracking at this stage.

Record Test

Record that the test has been ordered (ie clinician decides tests should be done). Also record when the relevant specimen is then taken.

Sample storage

Most laboratories will give local guidelines on how to store samples. Very little information is available on any audits that have been done on this.

Transport of sample to lab

This is a key and critical part of the process. Many of the tests offered by the laboratory are time critical for analysis. Most routine tests should have samples reaching the laboratory at least within six to eight hours. In some specific cases where tests are particularly sensitive patients are invited to hospital for phlebotomy to ensure rapid transport of sample to lab. Generalisable literature on this is sparse although local audit will be commonplace (including author’s laboratory).

InternalHospital transport

This will vary from lab to lab and will be entirely determined by local infrastructure. However, it should be take account in any overall systems approach particularly when looking at turn around times of test requests to reports. Again there is no generalisable evidence of this.

Laboratory Stages

This is very much part of the Clinical Pathology Accreditation regulatory system. Most laboratories in the UK will follow the CPA Standards. This involves detailed documentation of adoption of standards which are reviewed annually and inspected on site every two-years. In many ways this gives us a model of standardisation that we may wish to bring forward when looking at pre- and post-analytical laboratory practices.

The activities within the analytical process include:

Laboratory sample processing and documentation of sample
Analytical validation
Clinical validation
Report production
Transmission/transport of report

For logistical reasons, usually related to a turnaround time, some GP practices have taken on in-house laboratory testing. This is commonly known as ‘point of care testing’ (POCT). This has significant quality issues. Most POCT instruments in primary care do not follow CPA Guidelines. Therefore the performance and quality of the results produced are unlikely to be of the same standards. This does not mean they will not be adequate, but the regulatory controls of process will not be as defined, or inspected. This leaves a potential non-audited aspect of testing within the local health community which may have a deleterious effect on patient safety. There are national guidelines from the Medicines and Healthcare productsRegulatory Authority (MHRA) on the recommended use of POCT.

Post-analytical stages

Receipt of report

There has been some research done on turnaround time of results to the requestor. These have been mainly questionnaire based looking at requestor and/or patient satisfaction. There was no reference to any expected standards, these results were all subjective judgments.

There was some evidence in literature from the US of dissatisfaction with the inability to track when the results are likely to be available.

There was also a significant issue with the communication of abnormal results in some of the studies.

In the UK the Royal College of Pathologists with the RCGP have published guidelines on telephoning results on commonly requested tests. This is broken down to telephoning results in normal working hours and out of hours. This is a common standard used in laboratories throughout the UK.

Requestor responsibility

Where the requestor is not the patient’s GP, GMC Guidelines suggest that:

“52 If you provide treatment or advice for a patient, but are not the patient's general practitioner, you should tell the general practitioner the results of the investigations, the treatment provided and any other information necessary for the continuing care of the patient, unless the patient objects.”

Confusion can arise where the requestor is not available, where copy results are sent to a third party or indeed if the requestor is not aware that they are the named requestor.

There is no generalisable evidence of how extensive these problems are, but the Guidelines can be used as standards within the local health community.

Recognition and appropriate action

Clearly the recognition of the results and the appropriate action required will depend on the training and experience of the requestor. In this sense the ‘appropriateness’ of requestors may be challenged. Most laboratories would expect requestors to be medical practitioners or specifically trained healthcare professionals who require test results for their practice.

It is also a reality that some requesting is done on behalf of the traditionally recognised requestors with their knowledge. However, inappropriate requesting by non-trained professionals is another area that may warrant attention. The problems of resource management, responsibility of results, subsequent clinical management of patients and finally patient safety are not recognised in the literature.

Patient care

At the end of the laboratory cycle the test results should lead to appropriate clinical care.

Conceptual model for results handling

The Medical and Dental Defence Union of Scotland (MDDUS) has developed a model for Results Handling. With a high-level description of the process, the model poses questions relevant to each stage, that challenge practices to consider the robustness of their processes.

4 Successful interventions for managing results in practice