Christensen JC, Jennings MM.,  Normal and abnormal function of the first ray.   Clinics in Podiatric Medicine and Surgery. 2009 Jul; 26(3).

WHY did the authors undertake this study?
One of the challenges of foot and ankle surgery is matching the correct procedure to the underlying pathology. Since Morton’s “Metatarsus ativicus” article in 1927, our profession has spent decades working to objectify normal and abnormal function of the first ray and medial column.  The authors of this review attempt to summarize this work and provide clinically relevant diagnostic techniques.

HOW did they attempt to answer this question?
The authors provide a concise and well-organized literature review on a number of clinically-relevant topics including static/dynamic arch stabilization, the windlass mechanism, first ray hypermobility, metatarsal length patterns, hallux valgus pathomechanics, and hallux limitus pathomechanics.

WHAT were the specific results?
Although we typically consider hallux abductovalgus a primarily transverse plane deformity, this article demonstrates how the sagittal and frontal plane mechanics of the entire medial column contribute to 1st metatarsal-phalangeal joint sequela.  A complete understanding of the function of the first ray is required in order to appreciate the mechanics of the remainder of the forefoot and rearfoot.

Scherer SS, Pietramaggiori G, Mathews JC, Orgill DP.  Short periodic applications of the vacuum-assisted closure device cause an extended tissue response in the diabetic mouse model. Plastic and Reconstructive Surgery.  2009 Nov; 109(3): 943-50

WHY did the authors undertake this study?
The application and popularity of negative pressure wound therapy  (NPWT) in the treatment of wounds has outpaced our ability to understand its exact mechanisms of action.  One of its variables is the use of timed-intervals, as opposed to continuous application.  The objective of this study is to determine the effects of reduced time applications of negative pressure wound therapy. 

HOW did they attempt to answer this question?
The primary outcome measures of this study were a variety of measurements of wound healing:  total wound area, granulation tissue area, angiogenesis (as measured with platelet endothelial cell adhesion molecule 1), and cell proliferation (as measured by pan-endothelial cell marker Ki-67). 

Wounds were created on diabetic mice and divided into five treatment groups:  (1) occlusive dressing for 7 days, (2) continuous NPWT for 7 days, (3) a single 6-hour application of NPWT immediately after wound creation followed by occlusive dressing for 7 days, (4) a single 12-hour application of NPWT after wound creation followed by occlusive dressing for 7 days, and (5) three 4-hour periodic applications of NPWT immediately after wound creation followed by days 2 and 4. 

WHAT were the specific results?

With respect to total wound area, the 6-hour and 12-hour treatment groups demonstrated statistically significant reductions in wound area when compared to the continuous NPWT application group.  With respect to the formation of granulation tissue, the 12-hour and 4-hour periodic application treatment groups demonstrated statistically significant increases in granulation tissue when compared to the occlusive dressing group.  With respect to angiogenesis, the continuous NPWT and 4-hour periodic application treatment groups demonstrated statistically significant vascular effects compared to the occlusive dressing group.  And with respect to cell proliferation, all NPWT groups demonstrated statistically significant increases in proliferation rates compared to the occlusive dressing group. 

HOW did the authors interpret these results
From the results of this study, the authors concluded that any intervention of NPWT has some positive effects on wound healing, even with relatively short initial application.  They note that this information could lead to changes in the recommended clinical protocols with the use of this device, and possibly improvements in patient compliance. 

Let’s take a brief and introductory look at the topic of variance analysis as it relates to the Scherer et al article.  Their study design incorporated one control group (continuous occlusive dressing) and a total of 4 experimental groups (with varying time applications of NPWT).  Things can get somewhat more complicated from a statistical standpoint, and from a critical analysis standpoint, when there are more than 2 groups. 

When there are only a total of two groups in a study, things are fairly straight forward because there are only a total of two things being compared. A common example of this is pre- and post-operative radiographic measurements.  Is the 1st intermetatarsal angle the same or different after a HAV procedure?  We have two measurements (pre- and post-operative values) of one variable (1st intermetatarsal angle).  In this example there is not much confusion about what we are comparing!

However, when there are more than two groups, it can be difficult to keep track of what comparisons are happening.  The Scherer et al article had 5 different groups with the measurement of 4 different variables.  For each of the variables, are we comparing:  whether the groups are at all different from each other?;  if the individual experimental groups are different from the control group?;  or if the individual experimental groups are different from the other experimental groups?

Let’s use the concrete example of granulation tissue formation from this article.  The authors appropriately chose to use a kind of analysis of variance measurement to determine if each of the experimental groups was different than the control group.   They found two statistically significantly different differences:

• The 12-hour NPWT group produced a statistically significant greater volume of granulation tissue when compared to the control occlusive dressing group.
  
  
• The 3 x 4-hour NPWT group produced a statistically significant greater volume of granulation tissue when compared to the control occlusive dressing group. 

It is important to consider both what this information is telling us, and what this information is NOT telling us.  It is telling us that there are individual statistically significant differences between the 12-hour and 3 x 4-hour NPWT groups when compared to the control occlusive dressing group.  It is also telling us that there are not statistically significant differences between the continuous NPWT and 6-hour NPWT groups when compared to the control occlusive dressing.  However, it is NOT telling us that the 12-hour and 3 x 4-hour NPWT groups produced statistically significant greater volume of granulation when compared to the continuous and 6-hour NPWT groups. 

In the previous example, and for 3 out of the 4 variables of the Scherer study, each experimental group was compared to the control group, and not to each other.  However, the authors switched it up on us though for the 4th variable, wound area reduction.  Here they found statistically significant differences between the 6-hour and 12-hour NPWT groups when compared to the continuous NPWT group.  So the experimental groups were compared to the other experimental groups, and not to the control group.  In fact, there were no statistically significant differences between the 6-hour and 12-hour NPWT groups and the control group.
It is not inappropriate to switch comparisons in a study like this, but it does require focus on the part of the reader to appreciate what exactly is being compared. One thing that I do when there are multiple groups with multiple variables in a study, is always ask myself the question “Compared to what?”.  Whenever a statistically significant difference is reported, I ask myself and write down in the margins what exactly it is significantly different from.  This seems like a simple piece of advice, but I think you’ll find yourself surprised how difficult it is to answer this questions sometimes.