McMillan AM, Landorf KB, Barrett JT, Menz HB, Bird AR. Diagnostic imaging for chronic plantar heel pain: a systematic review and meta-analysis.  J Foot Ankle Res. 2009 Nov 13; 2:32. (PubMed ID:19912628)

WHY did the authors undertake this study?
Chronic plantar heel pain (CPHP) is a prevalent presenting podiatric complaint, with plantar fasciitis often implicated among the range of differential diagnoses.  Although imaging analyses often form part of the diagnostic plan, there are relatively few published critical analyses of these different options.  The objective of this article was to review the diagnostic imaging options for CPHP, and to evaluate these study findings with a meta-analysis design when appropriate.

HOW did they attempt to answer this question?
A systematic electronic medical research review was undertaken to identify all original, peer-reviewed articles reporting diagnostic imaging findings in patients with CPHP compared to independent control groups.

In order to determine if meta-analysis evaluation was appropriate, the methodological quality and diversity of the included studies were assessed by two blinded authors with a modified version of the Quality Index, as well as I2 and Chi2 statistics. 

WHAT were the specific results?
A total of 764 studies were initially identified, with 23 of these meeting inclusion criteria for review.  Three variables were determined to be appropriate for meta-analysis:  thickness of the proximal plantar fascia, ultrasound echogenicity of the proximal plantar fascia, and evidence of plantar calcaneal spur.

The thickness of the proximal plantar fascia was most commonly measured with ultrasonography, and meta-analysis of the pooled data demonstrated that patients with CPHP had two statistically significant findings.  First, the plantar fascia of CPHP patients was 2.16 mm thicker compared to controls, and second, patients with CPHP were more likely to have thickness values greater than 4.0 mm.

A statistically significant difference was also found with respect to the ultrasound echogenicity of the proximal plantar fascia.  Patients with CPHP were 200 times as likely to demonstrate hypoechogenicity compared to control patients.

And finally, the presence of plantar calcaneal spur on plain film x-ray was found to be statistically significant.  Patients with CPHP were 8 times as likely to show evidence of subcalcaneal spur when compared to control patients.

HOW did the authors interpret these results?
From these results, the authors concluded  that imaging analysis does play a role in the diagnosis of CPHP.  Specifically, the thickness and ultrasound hypoechogenicity of the proximal plantar fascia, as well as the presence of plantar calcaneal spur on plain film radiograph are strongly associated with plantar heel pain syndrome.

Lipsky BA, Polis AB, Lantz KC, Norquist JM, Abramson MA.  The value of a wound score for diabetic foot infections in predicting treatment outcome: a prospective analysis from the SIDESTEP trial. Wound Rep Regen. 2009 Sep/Oct; 17(5): 671-77.

WHY did the authors undertake this study?
Among the goals of any classification system is to be both descriptive and predictive.  The authors of this study hypothesized that a newly developed diabetic foot infection score could, in addition to being descriptive, determine the type and urgency of required antibiotic and surgical intervention, as well as predict clinical outcomes.

HOW did they attempt to answer this question?
The authors analyzed prospectively collected data from the SIDESTEP trial, which enrolled patients with moderate to severe diabetic foot infections.  A diabetic foot infection (DFI) wound score was developed by the primary author consisting of two components:  wound volume (area, depth, and undermining) and wound parameters (local signs/symptoms of infection).  

Multiple statistical analyses, including internal measures of the DFI score variation, correlation and validity, were performed to objectively assess this new classification.

WHAT were the specific results?
The new DFI wound score had a good correlation with clinical outcome as the rate of favorable clinical response progressively decreased with higher baseline wound scores.  For example, 94.9% of patients with a baseline score <12 had a favorable outcome, whereas only 77.0% of patients with a baseline score of >19 had a favorable outcome. 

In addition, all of the wound volume and wound parameter components demonstrated consistent correlation numbers with the exception of purulent and non-purulent drainage across the study.

HOW did the authors interpret these results
From these results the authors reached two important conclusions.  The first was that this new wound score appears to be a valid instrument, significantly correlated with the already validated University of Texas Wound Classification.  The second was that the presence or absence of purulent drainage or nonpurulent drainage did not appear to be a helpful factor in describing the wound.  Their overall conclusion was that this new score may be a useful measure of diabetic foot infections used for clinical practice and in research.

Let’s take a closer look at the topic of the meta-analysis study design, like that which was utilized by the McMillan study.  A meta-analysis is a specific type of systematic review that uses original statistical techniques to summarize results.  Very often in the lower extremity orthopedic literature we simply don’t have large clinical trials.  There are several reasons for this, but the important concept is that the sample sizes of our study populations are very low compared to say, a more medically-based study that examines the effect of a lipid-lowering drug on thousands of patients.  Personally, I can only think of a couple podiatric studies off the top of my head that reach over a hundred patients in the sample population, let alone thousands.

The problem with this is that is can be difficult to reach definitive conclusions from multiple, smaller studies.  There is a lot more information to work through, and you almost have to try and piece the data together.  It would be easier to draw definitive conclusions from fewer, larger studies.  In addition to having less material to work through, the statistical power of the studies is greater because of the larger sample sizes.  This is the idea behind most meta-analyses.  They attempt to combine the data from multiple, smaller studies into one large study with a statistically powerful sample size.  As the statistical power of a study increases, you can be more confident in the results.

However, you obviously just can’t accept a meta-analysis at face value, but must examine it from a critical analysis standpoint.  We’ll just hit the basics here, but one of the most important things is to ensure that the authors have performed a completely exhaustive literature review.  This is critical because the literature review is essentially the “data collection” portion of a meta-analysis.  This is where the raw numbers for the original meta-analysis statistical techniques will come from.  The authors should verify that they only include studies that are of high quality and similar enough to each other that direct comparison is possible.

Let’s evaluate how McMillan et al accomplished this.  Did they perform an exhaustive literature review?  Yes, five electronic bibliographic databases were searched with a total of 764 citations identified for possible inclusion.  These were narrowed down to a total of 23 that were included in the study.  Did they narrow it down to those of high quality?  Yes, they performed a Quality Index score on all articles to ensure high methodological quality to those that were included.  Their mean score of 55% indicated a moderate quality.  Did they narrow it down to those similar enough to each other to allow for direct comparison?  Yes, they performed two tests of heterogeneity (I² and Chi²) prior to inclusion for the meta-analysis.  In other meta-analyses, you may see the κ (kappa) statistic and intraclass correlations used for these comparisons.

For those interested in learning more about this, I recommend an article published by Mohit Bhandari entitled Users’ guide to the surgical literature: how to use a systematic literature review and meta-analysis Dr. Bhandari has published a series of articles on the topic of specific critical analysis of the orthopedic literature that are well worth reading.