We’ve done several columns on test results slipping through the cracks to result in delayed diagnosis of cancer (see links at the end of this article). But until now, we knew of no comprehensive review of events across the entire continuum of care that lead to delayed diagnosis of cancer. Now the NPSA (UK) has issued a report “Delayed Diagnosis of Cancer: Thematic Review” that does just that. And while there are differences in the health care delivery systems between the US and the UK, most of the lessons in that report are applicable to our health care system.
The NPSA report was put together based upon a literature review, analysis of reports in their incident reporting system(s), and focus groups of stakeholders. They basically break down the delays into (1) patient delays (2) doctor delays (3) system delays.
Under patient delays the overwhelming influence leading to delays in diagnosis is symptom recognition and interpretation. Particularly when initial symptoms are vague patients are likely to dismiss them or attribute them to everyday explanations. For instance, symptoms of oral cancer are often attributed to dental problems or infection or symptoms of pharyngeal cancer are often attributed to the common cold or infection. Other contributing patient-related factors are fear and anxiety. Fear of cancer and fear of embarrassment are often noted in the literature. Even guilt may cause delays (eg. a patient feels guilty that one of their behaviors may have predisposed them to cancer). Lastly, socio-demographic factors and ethnicity may play roles. Older age has been identified as a risk factor and that may be due, in part, to the fact that older individuals have been less likely to speak up and challenge their physicians. Some cultures also do not allow male physicians to fully examine female patients, another factor that may delay diagnosis.
On the physician side, misattribution of symptoms is a common cause for delayed diagnosis. Many of the early symptoms of cancer may be quite vague and patients may have several comorbidities that could explain such symptoms. However, the study identified a change in the pattern of visits as a common presentation of cancer. This means that a patient who usually is seen infrequently now has an increase in visits for the same symptoms. Communication problems (between doctor and patient) are also frequently noted in cases of delayed diagnosis. Particularly with pediatric cases, physicians often minimized or discounted concerns that the parents had about their child. And lack of good systems for tracking tests and referrals remain a significant problem (which we’ll discuss under system issues).
The report notes that relatively little is actually known about delays occurring on the primary care side because there is very limited reporting from primary care into their incident reporting systems. That, in fact, was recognized as one of their lessons learned.
System issues were readily uncovered in their review of incident reports. For pathology, they noted pre-lab problems and in-lab problems each accounting for about 45% of cases, with less than 10% of incidents being related to post-lab problems. Pre-lab problems included mistakes in labeling of specimens, getting specimens to the lab, and getting them to the lab in a timely fashion while they were still able to be properly evaluated. In-lab problems included delays in getting reports out or reports where malignancies were erroneously reported as benign. Examples included cases where a pathologist confused a specimen with that of a different patient. Post-lab issues were primarily those related to the reports being lost, misfiled, or otherwise never acted upon.
For radiology, where risk of delay was high, failure to communicate abnormal test results was the primary concern. These are the same issues we have discussed in detail in our prior columns (see links at the end of this article). But they did have a couple new ones that are really unintended consequences of technology. Both pertained to PACS (the picture archives and communications systems). A real value of PACS is the ability to easily call up old images for comparison. However, in one example provided, the radiologist reported an earlier examination rather than the current one. In another, a patient came for a CT-guided biopsy and the relevant images had not been placed on PACS. We have also mentioned previously that when you are looking at anything on the computer screen, PACS included, it is very easy to think you are looking at one patient when, in fact, information about a different patient is being displayed. That is why you always want the name of the patient and other identifiers displayed at the top of every screen. Actually, a radiologist should do a “mini-timeout” before every report (asking himself questions like “Is this the correct patient?”, “Is this the correct study?”, Is this the correct date?” etc.).
But about a fifth of their incidents were related to failures of communication. This could be failure of the pathologist or radiologist to be sure the ordering physician actually received the report of the abnormal finding, or miscommunications when several physicians within a practice were seeing the patient on different visits, or miscommunications between primary care physicians and specialists, or cancellations of visits (whether patient-related or provider-related).
From their focus groups and workshops, which included patients as well as providers and other stakeholders, communication arose as the overriding theme. Patients especially felt that their primary care providers often failed to listen to them or failed to spend sufficient time with them. They often felt that the physician (or nurse practitioner) was too quick to dismiss their symptoms or label the patient as “menopausal” or “neurotic”. Such patient labeling is an issue we will probably discuss in an upcoming Tip of the Week since we have seen numerous serious incidents in which labels applied to patients likely biased their care and led multiple healthcare workers to overlook important aspects of care.
The focus groups also discussed in detail the “passive patient” issue. That is the patient who returns over and over with symptoms that have continued yet fails to complain or request a second opinion. They pointed out that the physician often reinforces the passive patient role rather than empowering the patient to speak up.
The report stresses non-verbal aspects of communication. We already mentioned the critical change in pattern of visits. That is when a patient suddenly starts making more visits than he/she usually makes and should be a sign that certain symptoms are not being adequately addressed. This could especially be a problem in those settings where multiple caregivers provide care for the patient (though sometimes the new provider was a “set of fresh eyes” that saw a diagnosis previously overlooked). Maybe there are ways we can “flag” such changes in visit patterns electronically to alert the provider in some fashion that this might be a higher risk scenario.
But communication across transitions of care may be even more of an issue. We’ve noted that the emergency department is a particularly vulnerable area because both the patient and the physician may be long gone by the time the results come back. That problem is accentuated when you cannot track the patient down because the phone number given was incorrect. A recent MSNBC story on inaccurate patient phone numbers for patients seen in ER’s noted “a combination of delayed test results and bad contact information — sometimes from fearful patients who deliberately give fake names and numbers — is forcing some emergency room officials to resort to people-finders, registered letters and law enforcement visits to deliver their diagnoses.”
It may be just as difficult for hospitalists, who may never see a patient again after discharge. You have heard us harp on the need to include a section in every discharge summary for “test results pending”. We still think that should be included in every case. However, identifying the “pending test results” is not an easy task. We tried it one day. We pulled charts of several patients whom a hospitalist was discharging and tried to identify what tests had been done and match them with final reports. It took forever! So how is a busy hospitalist (or housestaff member) who is taking care of 10-15 inpatients going to complete this task? The only way is likely to be writing a computer script that does the search for you. But even that may prove difficult because it is very likely to show lots of erroneous instances, in which case alert fatigue will likely cause the hospitalist to ignore all instances.
Note also that your good hospitalist programs have the hospitalist call the patient by phone 2-3 days after discharge to see how they are doing and whether they have any problems or questions. That may also be a good time to followup on pending test results.
And remember that Singh et al found that the well-intentioned process of sending test results to both the ordering physician and another physician (such a the primary care physician) actually increased the likelihood that no one would follow up on the abnormal results (see our Patient Safety Tip of the Week October 13, 2009 “Slipping Through the Cracks”).
The NPSA review makes the following five recommendations (our paraphrasing):
If you are like us, you looked at those recommendations and said “Where’s the beef?!” Too much fluff and no good action items. So we’ll have to look elsewhere. We shouldn’t say it had no action items. They did note a few. One is to give the patient an information card when tests are done. It should outline what tests have been performed, when to expect the results will be available, who to contact if a result has not been conveyed to them within a specified period of time, and what to do if the test result is negative but the symptoms persist. We like that! We have always recommended that the physician verbally tell the patient those things. But it actually should be very easy to deploy word processing templates that would fill in the dates based on the date the test is done and allow you to give the patient a printout for each test ordered.
But for some more specific action items, the Singh group comes to the rescue again! They have “Eight Recommendations for Policies for Communicating Abnormal Test Results” in an article in this month’s Joint Commission Journal on Quality and Patient Safety (Singh and Vij 2010). While their article talks about compliance with the Joint Commission’s national patient safety goal NPSG.02.03.01 (reporting critical results of tests and diagnostic procedures on a timely basis), it includes not only “panic values” but also abnormal but not immediately life-threatening results such as the suspected tumor on chest radiographs. They discuss the key elements put together over several years in their VA system for communicating abnormal results
The first recommendation is to clearly define terms in your policy such as what are the specific test abnormalities that require either immediate notification or urgent notification (they provide good examples in the article). Secondly, the policy needs to clearly spell out the responsibilities of each provider (remember their key lesson learned previously that ambiguous responsibilities, as seen when dual notification takes place, actually increases the likelihood that no action will be taken).
Thirdly, they call for fail-safe methods for communication of abnormal results. While that could be a verbal communication, they also allow for a computer notification that includes a mandatory response (i.e. that the recipient cannot dismiss that alert without acknowledging its receipt and contents). They do have contingencies for alerting other responsible parties (eg. the responsible attending physician when a resident ordered the test). Particularly good are their recommendations about documentation of the communication that takes place, including verbal read-back, and an algorithm “with escalation to supervisory level” for after-hours situations.
Fourthly, policies should define reporting procedures for both critical and significantly abnormal test values. For any potentially life-threatening abnormal result, a verbal communication should take place. For the significantly abnormal (but not immediately critical) results they do allow a variety of electronic communications. Note that they also discuss that such notifications may not be always necessary on subsequent tests once the initial abnormality has been communicated. For imaging studies, they have specific codes for various findings and provide a table that outlines what verbal and/or electronic notifications must take place for each finding.
The fifth recommendation is specifying the timelines for notification of the results. The sixth deals with timelines for patient notification and discusses the various means of patient communication that should be specified in your policy. The seventh recommendation is that your policy should incorporate feedback from key stakeholders in your organization. For instance, they found that reporting abnormal (but not critical) results to the on-call resident after hours may be counterproductive. And, lastly, the policy should include responsibilities for monitoring the compliance with and effectiveness of the communications.
The recommendations are excellent and you can learn a lot from this article. They should work especially well in an integrated system. However, some barriers still exist. We still have to deal with fragmentation of care and poor communication between disparate providers. In real life, it is still common for patients to seek their cancer care at tertiary cancer centers. So if you are the radiologist at the small community hospital who noted the suspicious lesion on chest films, you may be unable to confirm that anyone actually took action because you (and, all too often, the primary care physician) may lack access to those records from the tertiary care cancer center. In fact, HIPAA might even preclude you from seeking that followup once your relationship with that patient has ended. So you may have to be satisfied with verbally notifying the ordering physician of your findings and hoping that the patient ultimately gets the appropriate management.
See also our other columns on communicating significant results:
National Patient Safety Agency (UK). Delayed Diagnosis of Cancer: Thematic Review. March 2010
Aleccia J. Wrong patient phone numbers trip up ER docs. Fake or inaccurate contacts raise risk when sick patients can't be found. msnbc.com April 26, 2010
Singh H, Thomas EJ, Mani S, et al. Timely Follow-up of Abnormal Diagnostic Imaging Test Results in an Outpatient Setting. Arch Intern Med. 2009; 169(17): 1578-1586.
Singh H; Vij MS. Eight recommendations for policies for communicating abnormal test results. Jt Comm J Qual Patient Saf. 2010;36:226-232