BMJ Open Study: Self Diagnosis During A Pandemic

 

 

 

Photo Credit CDC Influenza Home Care Guide

 

# 5837

 

We’ve a new study appearing in BMJ Open that looks at just how accurate self diagnosis of influenza during the 2009 pandemic really was.

 

First, a little background.

 

ILIs  . . .  or  Influenza-like Illnesses are among the  most common maladies reported to doctors each year, and while often attributed to `flu’, the causes extend far beyond influenza.

 

The symptoms generally include fever, cough, and body aches  -  but may also commonly include rhinitis, sneezing, headache, fatigue, sore throat, nausea & vomiting, and diarrhea.

 

Influenza A & B, which get most of the headlines, are only responsible for a fraction of the cases of ILI each year. Some estimates put that share as low as 10%.

 

According to the CDC, each year adults (on average) experience 1 to 3 bouts with an ILI, while children may see 3 to 6 flu-like illnesses (cite MMWR)

 

Most of these viral infections are mild, self-limiting, and are almost never identified since testing (beyond, perhaps, a rapid influenza test) is rarely warranted. 

 

Which is why doctors generally refer to ILIs, or Influenza-like Illnesses (or sometimes ARI Acute Respiratory Infection), when making a clinical diagnosis.

 

If an ILI is mild, most people chalk it up to the `common cold’, and if it is more debilitating, figure it is `the flu’.

 

But in reality the spectrum of common respiratory viruses is far more diverse; metapneumovirus, parainfluenzavirus, coronaviruses, respiratory syncytial virus (RSV), enteroviruses, any of the myriad Rhinoviruses  (Common cold), and a number of varieties of adenovirus.

 

 

The prevalence of these non-flu ILI’s complicate the job for doctors and public health officials during a pandemic when limited resources (hospital beds, antivirals, etc) must be prioritized and people must decide whether to go to the ER or self isolate if symptomatic.

 

Which brings us to today’s study, where researchers surveyed more than 1100 people from New Zealand during the 2009 pandemic and compared their perception of whether they caught the flu that year with blood tests showing antibodies to the novel H1N1 virus.

 

 

Self-diagnosis of influenza during a pandemic: a cross-sectional survey

Annemarie Jutel1, Michael G Baker2, James Stanley2, Q Sue Huang3, Don Bandaranayake4

BMJ Open 2011;1:e000234 doi:10.1136/bmjopen-2011-000234

Abstract

Background Self-diagnosis of influenza is an important component of pandemic control and management as it may support self-management practices and reduce visits to healthcare facilities, thus helping contain viral spread. However, little is known about the accuracy of self-diagnosis of influenza, particularly during pandemics.

Methods We used cross-sectional survey data to correlate self-diagnosis of influenza with serological evidence of 2009 pandemic influenza A(H1N1) infection (haemagglutination inhibition titres of ≥1:40) and to determine what symptoms were more likely to be present in accurate self-diagnosis. The sera and risk factor data were collected for the national A(H1N1) seroprevalence survey from November 2009 to March 2010, 3 months after the first pandemic wave in New Zealand (NZ).

Results The samples consisted of 318 children, 413 adults and 423 healthcare workers. The likelihood of being seropositive was no different in those who believed they had influenza from those who believed they did not have influenza in all groups.

Among adults, 23.3% (95% CI 11.9% to 34.7%) of those who reported having had influenza were seropositive for H1N1, but among those reporting no influenza, 21.3% (95% CI 13% to 29.7%) were also seropositive.

Those meeting NZ surveillance or Ministry of Health influenza case definitions were more likely to believe they had the flu (surveillance data adult sample OR 27.1, 95% CI 13.6 to 53.6), but these symptom profiles were not associated with a higher likelihood of H1N1 seropositivity (surveillance data adult sample OR 0.93, 95% CI 0.5 to 1.7).

Conclusions Self-diagnosis does not accurately predict influenza seropositivity. The symptoms promoted by many public health campaigns are linked with self-diagnosis of influenza but not with seropositivity. These findings raise challenges for public health initiatives that depend on accurate self-diagnosis by members of the public and appropriate self-management action.

 

I’ve only printed excerpts from the abstract, but the entire study is open access, and freely available.  The key finding (below) is well illustrated by one of the charts accompanying the study:

 

The likelihood of being seropositive was no different in those who believed they had influenza from those who believed they did not have influenza in all groups.

 

 

In this study, roughly 75% of the people who thought they caught the pandemic flu that fall did not.  While nearly 25% who believed they had not contracted the virus actually had.

 

Even among healthcare workers, the percentage who correctly diagnosed themselves with the pandemic flu was only 33%.

 

And those who reached the conclusion that they had contracted the flu after consultation with a healthcare professional were wrong most of the time as well. 

 

 

Not only can the misidentification of influenza cause serious problems during a pandemic, it also likely skews the public’s perception of the effectiveness of the yearly flu vaccine.

 

Many people complain that in years when they got the shot they still caught the flu. As this study shows, people often mistake non-flu ILIs (which are not covered by the shot) for influenza.

 

The major implication here is that during a pandemic, the majority of people seeking antivirals or medical care due to having flu-like symptoms are likely not to have the pandemic virus.

 

The authors of this study conclude:

 

These findings reinforce public health advice during the pandemic that patients should seek medical care on the basis of disease severity rather than for the purpose of diagnosis.

 

Of course, getting the public to go along with this advice during an influenza pandemic is going to be a lot easier said than done.

»» Read More

A Tale of Two Headlines

 

 

# 4355

 

 

Yesterday, in the wake of the release of a study in the journal Pediatrics on the efficacy of RIDTs (Rapid Influenza Diagnostic Tests), there were a number of news reports that seemed to depict the results in vastly different ways.

 

We’ll look at the headline and lede of two of them.

 

Health Day News saw it this way:

 

Rapid Flu Test Most Accurate for Young Kids

HealthDay News

A widely available rapid influenza diagnostic test is good, but not perfect, in determining whether a child has the flu, a new study shows.

(Continue. . . . )

 

While Medpage Today reported:

 

Rapid Tests Fail at Ruling Out H1N1 in Kids

By Todd Neale, Staff Writer, MedPage Today
Published: February 15, 2010

Rapid influenza tests have poor sensitivity for detecting pandemic H1N1 flu in children, two studies showed.

(Continue . . . )

 

Confused?  

 

One report says the test is `good but not perfect’, while the other states that it has `poor sensitivity’.  

 

In truth, both headlines are correct . . .  despite giving disparate  impressions.     

 

I’ll try to explain why.

 

The two main measures of the accuracy of a diagnostic test are sensitivity and specificity.

• Sensitivity is defined as the ability of a test to correctly identify individuals who have a given disease or condition.

• Specificity is defined as the ability of a test to exclude someone from having a disease or illness.

 

The various RIDTs are designed to show if someone tests positive for the Influenza A or B virus, but not the strain of flu.

 

According to the CDC:

 

The rapid tests vary in terms of sensitivity and specificity when compared with viral culture or RT-PCR. Product insert information and research publications indicate that:

• Sensitivities are approximately 50-70%
• Specificities are approximately 90-95%

 

The study in question today is entitled:

 

Performance of a Rapid Influenza Test in Children During the H1N1 2009 Influenza A Outbreak

Andrea T. Cruz, MD, MPH^a,b, Gail J. Demmler-Harrison, MD^b,c,d, A. Chantal Caviness, MD, MPH, PhD^a, Gregory J. Buffone, PhD^c, Paula A. Revell, PhD^c,

 

 

Here researchers evaluated the performance of RIDTs used at two different hospitals (that were capable of confirmatory rRT-PCR & Viral Culture Testing) over last summer.  They were then able to determine the accuracy of one of 10 FDA approved RIDTs (BinaxNOW) .

 

A couple of excerpts from the abstract.  Then some discussion.

 

Results  . . . With rRT-PCR as the reference, overall test sensitivity was 45% (95% confidence interval [CI]: 43.3%–46.3%) and specificity was 98.6% (95% CI: 98.1%–99%).  . . . .  RIDT sensitivity was significantly higher in young infants and children younger than 2 years than in older children.

Conclusions The RIDT had relatively poor sensitivity but excellent specificity in this consecutive series of respiratory specimens obtained from pediatric patients.

 

 

In other words, if the test said someone had influenza, it was right almost all of the time (Note: This was during a time of high influenza activity). 

 

But the test failed to detect influenza in roughly half the patients that were infected.

 

So the first headline Rapid Flu Test Most Accurate for Young Kids is true, since the report states that `sensitivity was significantly higher in young infants and children younger than 2 years.’

 

As is the second headline - Rapid Tests Fail at Ruling Out H1N1 in Kids – since overall, `The RIDT had relatively poor sensitivity.’

Contrary to popular belief, a 45% sensitivity level doesn’t mean that the test will be right less than half the time.  That depends on the actual incidence of influenza in the community when the test is taken.

 

As we’ve discussed before, influenza makes up but a fraction of all of the ILI’s (Influenza-like-Illnesses) that circulate in a community.   During the summer, that percentage drops into the low single digits, while at the height of flu season, it can approach 50%. 

 

The other ILI culprits include Coronaviruses, metapneumovirus, parainfluenzavirus, respiratory syncytial virus (RSV), any of the myriad Rhinoviruses (Common cold), and some adenoviruses.

 

None of these are influenzas, and so none would be expected to test positive with the RIDT.  

 

The following comes, again, from the CDC.

 

Accuracy Depends Upon Prevalence

The positive and negative predictive values vary considerably depending upon the prevalence of influenza in the community.

• False-positive (and true-negative) influenza test results are more likely to occur when disease prevalence is low, which is generally at the beginning and end of the influenza season.
• False-negative (and true-positive) influenza test results are more likely to occur when disease prevalence is high, which is typically at the height of the influenza season.

Clinical Considerations of Testing When Influenza Prevalence is Low

When disease prevalence is relatively low, the positive predictive value (PPV) is low and false-positive test results are more likely. By contrast, when disease prevalence is low, the negative predictive value (NPV) is high, and negative results are more likely to be true.

 

 

The interpretation of positive results should take into account the clinical characteristics of the case. If an important clinical decision is affected by the test result, the rapid test result should be confirmed by another test, such as viral culture or polymerase chain reaction (PCR).

Clinical Considerations of Testing When Influenza Prevalence Is High

When disease prevalence is relatively high, the NPV is low and false-negative test results are more likely. When disease prevalence is high, the PPV is high and positive results are more likely to be true.

 

 

So, to recap.  

 

If there’s a lot of flu around, and you test positive, there’s a pretty good chance you really have the flu.   

 

If you test negative . . .  well, doctors are urged not to base a diagnosis solely on a negative test.   Somewhere around 30% are probably false negatives.

 

 

 

And If there’s not much flu going around? 

 

Even if you test positive, there’s a pretty good chance you don’t have the flu.  

 

But a negative test (while not exclusionary) is substantially more likely to be true.

 

Obviously, given the ambiguity of all of this, a more accurate RIDT is something that doctors would be eager to get their hands on.  

»» Read More

PLoS Currents: Improving Diagnostic Efficiency Of H1N1

 

 

#3891

 

Now that routine testing for the H1N1 virus has become logistically impossible for most patients, doctors must base their diagnosis largely on the clinical presentation of the patient, the use of RIDTs  (Rapid Influenza Diagnostic Tests), and the current level of flu activity in the community.

 

From PloS Currents, we get a study conducted at the Infectious Diseases and Tropical Medicine department of Marseille university hospital that looked at the RIDT (Rapid Influenza Diagnostic Tests) and clinical signs and symptoms of patients that tested positive via RT-PRC testing for the H1N1 virus.

 

Of the 307 patients evaluated with flu-like symptoms, 31 tested positive for H1N1.  Or an incidence level of about 10%. Further evidence that not all ILI’s over the summer were H1N1 (see ILI’s Aren’t Always The Flu).

 

While many patients without H1N1 presented with a cough, of those that tested positive for the virus, 100% had a cough.  Asthenia (weakness, loss of strength) was present in  97% of cases, and fever in 87%. 

 

Myalgia was significantly higher in non-flu ILI’s (90%) than among the confirmed H1N1 cases (64%)

 

The authors theorize:

. . .  a patient who present with a flu-like illness and cough should systematically be tested by RIDT because the presence of cough is associated with a sensitivity at 100%. This clinical sign alone will not be predictive of the disease as the PPV is low and will stay low even during the epidemic (Figure 3A).

 

However the association of cough, diarrhoea and the lack of myalgia, although not useful in the early phase, may be useful in the acme of the epidemic as the PPV reaches up to 90% when the prevalence exceeds 20%.

 

Although based on a relatively small cohort of positive H1N1 patients (n=31), this study gives us an interesting overview of their clinical signs and symptoms.

 

A few highlights below, but follow the link to read the study in its entirety.

 

 

Improving the diagnostic efficiency of H1N1 2009 pandemic flu: analysis of predictive clinical signs through a prospective cohort.

 

BROUQUI, Philippe; VU HAI, Vinh; NOUGAIREDE, Antoine; LAGIER, Jean-Christophe; BOTELHO, Elisabeth; NINOVE, Laetitia; ZANDOTTI, Christine; CHARREL, Remi N.; DE LAMBALLERIE, Xavier; RAOULT, Didier. Improving the diagnostic efficiency of H1N1 2009 pandemic flu: analysis of predictive clinical signs through a prospective cohort

[Internet]. Version 24. PLoS Currents: Influenza. 2009 Oct 21. Available from: http://knol.google.com/k/philippe-brouqui/improving-the-diagnostic-efficiency-of/qccicvnl16mc/1.

In late June 2009, we set up a dedicated flu-like illness outpatient consultation in the Infectious Diseases and Tropical Medicine department of Marseille university hospital to detect the new A/H1N1 pandemic influenza and to contain efficiently the A/H1N1 infected patients.

 

For 3 months, we compiled data corresponding to a total of 307 patients who presented with a flu-like syndrome. 31 of them were positive for H1N1 pandemic flu through real-time RT-PCR (rRT-PCR); among them, 19 were positive for a rapid influenza detection test (RIDT). We report here the significant clinical characteristics of A/H1N1 pandemic flu patients compared with other flu-like illnesses, which were used to improve the predictive value of the diagnosis in the current epidemiological situation.

 

We found that regardless of the prevalence of A/H1N1 positive cases in the suspected patients, the absence of cough rejects the diagnosis of A/H1N1 infection in 100% of cases. Among patients referred for flu-like illness, those with cough should be tested for A/H1N1 by RIDT.

(Continue . . . )

 

 

 

 

Interestingly, myalgia was significantly associated with “non-Flu” patients (20/46%) with a Relative Risk of 0.42. All other clinical signs including fever, asthenia, headaches, and vomiting were not significantly associated with “Flu patient” (Table 1).

 

 

»» Read More

ILI’s Aren’t Always The Flu

 

 

 

# 3832

 

 

Although nearly all of the influenza circulating in the the US and Canada right now is apparently the novel H1N1 `swine’ flu, it would be incorrect to say that if you have a flu-like illness, you must have the pandemic flu.

Last week’s numbers from the CDC’s FluView, which I’ve plugged into the above spreadsheet so as to generate a graph, show that more than 72% – almost 3/4ths – of all virus samples tested from people who had flu-like symptoms came back negative for influenza.

 

Although testing may miss some cases (samples degrade, viral shedding at the time of sampling may have been low, etc), it is pretty obvious that a lot of flu-like illnesses are caused by something other than influenza.

 

While we tend to talk most about influenza, the truth is, there are a plethora of pulmonary pathogens out there, and telling them apart isn’t easy to do.    The signs and symptoms of all of these viruses can be pretty similar. 

 

In fact, laboratory PCR testing is just about the only reliable way to sort them out.   Rapid tests – commonly found in doctor’s offices - are very unreliable, with sensitivities running only about 50%. 

 

Which is why doctors generally refer to ILIs, or Influenza-like Illnesses (or sometimes ARI Acute Respiratory Infection), when making a clinical diagnosis.  

 

But it is worth noting that most ILI’s are not caused by the influenza virus. 

 

This is something that Tom Nolan mentioned earlier this week in his blog on the BMJ.   He presented a chart, similar to the one I created above, except that it breaks down some of the `other causes’ of ILI’s. 

 

You’ll notice, however, that the bulk remain `UNKNOWN CAUSES’.

 

While it is estimated that, in a normal year, between 7% and 10% of Americans catch the `flu’, far more of us will endure some other kind of respiratory infection. 

If it isn’t flu, you no doubt are wondering what else might it be?

 

Some of the `usual suspects’ would include metapneumovirus, parainfluenzavirus, respiratory syncytial virus (RSV), one of the myriad Rhinoviruses (Common cold), and adenovirus.

 

Among others.

Testing isn’t generally done because of the costs involved, and because knowing the etiology doesn’t really affect treatment.   Bed rest, fever reducers, and plenty of fluids is the usual regimen.

 

A recent study published in the Journal of Medical Microbiology gives us some idea of just how varied these viruses can be:

 

Aetiology of influenza-like illness in adults includes parainfluenzavirus type 4

 

J Med Microbiol 58 (2009), 408-413; DOI: 10.1099/jmm.0.006098-0

Influenza viruses cause significant morbidity and mortality in adults each winter. At the same time, other respiratory viruses circulate and cause respiratory illness with influenza-like symptoms. Human respiratory syncytial virus (HRSV), human parainfluenza viruses (HPIV) and human metapneumovirus have all been associated with morbidity and mortality in adults, including nosocomial infections.

 

This study evaluated 154 respiratory specimens collected from adults with influenza-like/acute respiratory illness (ILI) seen at the Edward Hines Jr VA Hospital, Hines, IL, USA, during two successive winters, 1998–1999 and 1999–2000.

 

The samples were tested for ten viruses in two nested multiplex RT-PCRs. One to three respiratory viruses were detected in 68 % of the samples. As expected, influenza A virus (FLU-A) infections were most common (50 % of the samples), followed by HRSV-A (16 %).

 

Surprisingly, HPIV-4 infections (5.8 %) were the third most prevalent. Mixed infections were also relatively common (11 %). When present, HPIV infections were approximately three times more likely to be included in a mixed infection than FLU-A or HRSV.

 

Mixed infections and HPIV-4 are likely to be missed using rapid diagnostic tests. This study confirms that ILI in adults and the elderly can be caused by HRSV and HPIVs, including HPIV-4, which co-circulate with FLU-A.

 

Interestingly, more than 10% of those tested had two or more concurrent viral infections.

During the height of `flu season’, only between 25%-30% of the samples tested by the CDC usually test positive for influenza.  During the summer, that percentage will drop to the low single digits.

 

The reason for bringing all of this up – besides the fact that it is interesting – is really two-fold.

 

First, since many of these milder ILI’s are commonly perceived by the public as being the `flu’, many people have a false perception of what having real influenza can be like. 

 

And  second, it is entirely possible that some of the people who suspect that they’ve already had the pandemic virus over the summer or last spring – and would therefore be immune – really had one of the other commonly circulating respiratory viruses. 

 

Which means that they may not have acquired the protective antibodies they think they have.

 

Two things to consider, now that the H1N1 vaccine is becoming more widely available.

»» Read More

Lancet: Atypical H1N1 Presentation In Children

 

 

#  3685

 

Viral infections during the winter months are as ubiquitous as they are difficult to identify.  Doctors see them every day, but most of the time, their etiology remains uncertain.  

 

“You’ve picked up a virus.”, may very well be the most common  utterance of a GP during the flu season.

 

Testing can be expensive, inaccurate, and often an exercise in futility.  Most viral infections run their course in 3 to 5 days, and are over and done with before any laboratory tests can come back. 

 

Office tests, such as the RIDT (Rapid Influenza Diagnostic Test) are notorious for their inaccuracies.   The CDC recently released guidance for physicians explaining why a NEGATIVE test should not be used to exclude influenza as a diagnosis.

 

 

Essentially, a positive test will generally be correct.  A negative test may miss 50% of influenza cases, and so doctors are advised to use clinical evaluation to make the diagnosis.

 

The problem, of course, is that the symptoms for influenza are pretty much the same as the symptoms for a great many other viral infections; fever, malaise, body aches, cough.

 

What looks like `flu’ could just as easily be parainfluenza, a rhinovirus, respiratory syncytial virus, metapneumovirus, or adenovirus – to name a few.

 

Complicating matters this year is the novel H1N1 virus, which has swept across the globe over the summer.   While it is a mild-to-moderate illness in the vast majority of patients, in some small percentage of people it produces severe – even life threatening symptoms.

 

With no reliable rapid detection test, physicians (and the public) are called upon to make a diagnosis based on clinical symptoms.  

 

That may prove difficult, as we learn from this Lancet article, since the novel H1N1 virus can produce atypical symptoms in a surprisingly high percentage of patients.

 

A hat tip to Niko on Flutrackers for posting this study.

 

Clinical characteristics of paediatric H1N1 admissions in Birmingham, UK

 

S Hackett a, L Hill a, J Patel a, N Ratnaraja b, A Ifeyinwa b, M Farooqi b, U Nusgen c, P Debenham c, D Gandhi c, N Makwana b, E Smit a d, S Welch a

 

Our experience of the first wave of paediatric H1N1 swine-origin influenza admissions in Birmingham, UK, shows that presentations can be atypical, severity is often associated with underlying disease, and rates of secondary bacterial infection are low.

 

We reviewed the 78 available case notes of 89 children positive for H1N1 influenza by PCR admitted to hospitals across our three Trusts between June 5 and July 4, 2009.

(Continue . . . .)

 

 

A confusing array of symptoms reported in Pediatric Patients

 

In the UK the HPA (Health Protection Agency) produced a case definition of H1N1 for doctors to use to make a clinical diagnosis of the illness.  That called for a temperature ≥38°C or a history of fever and two other symptoms of cough, sore throat, rhinorrhoea, limb or joint pain, headache, vomiting or diarrhoea, or a severe or life-threatening illness

 

According to the researchers, To have followed the HPA algorithm would have meant that 40% of children with H1N1 influenza would not have been diagnosed.

 

It should be noted that this study involved kids sick enough to be admitted to the hospital, and yet 20% of them had no fever, and about 25% no cough.

 

A dilemma, obviously, for physicians. But most are pretty adept at telling when a child is really sick.

 

We don’t have good data on the spectrum of symptoms being experienced by those with milder illness, although it is likely to be similarly nebulous and diffuse.

 

Which may prove a big challenge for individuals and families who must decide, based on symptoms, whether to stay home from work or school to avoid spreading the virus. 

Infected individuals can range from being completely asymptomatic, to mild to moderately symptoms, to severely ill. 

 

And not everyone will present with `classic’ flu symptoms. 

 

Which is the main reason why health officials have accepted that there is no reasonable way to contain, or halt, the spread of influenza in our communities.

 

It is the master of disguise, can operate in `stealth mode’, and is impossible to reliably identify 100% of the time.

»» Read More