Ronald S. Hosek, PhD
Edward F. Owens, MS
Lawrence V. Steinle, DC

Life Chiropractic College, Marietta, Georgia

An abiding issue in chiropractic has been how to adequately conduct blinded studies of the effects of specific adjustments versus placebos. Such studies are thought to be the only adequate way to judge the psychogenic component of patient response to chiropractic. This issue really consists of two separate problems: finding a true placebo and obtaining blinding. Most DC's will argue that any form of interaction involving contact may cause a change, implying that there can be no real placebo. Blinding can exist at three levels: patient, doctor and assessor/analyzer. Blinding of the latter is not thought to be a problem. Blinding of the patient can be achieved in the chiropractic setting, since it is possible to find naive patients who can be convinced that any form of physical interaction with the DC is an adjustment. It is virtually impossible to blind the doctor who will be contacting the patient. It is useful to draw the distinction here between placebo adjustments, which are inactive, and "sham" adjustments which are active but non-specific. In practice, designs using shams are more feasible, since patient blinding can be achieved while the effects of specific versus non-specific adjustments can be compared. Unfortunately, the contributions of psychogenic effects to outcome can only be inferred.

One of the ways that has been suggested for getting around these problems is to use an electrically-activated adjusting instrument which has a stylus whose excursion can be controlled. This idea presupposes that the depth (or force) of the thrust is an important biomechanical element in an adjustment; More specifically, it assumes that light or non-contact adjustments should have no effect or much less effect on typical clinical variables than should moderate to heavy thrusts, and hence should serve as acceptable placebos, or at worst, shams. Because the thrust is fast on such instruments, most patients can not discern the difference between deep and shallow excursions and are more apt to acknowledge the motor noise as the cue for the adjustment. Further, if the excursion could be preset or remotely controlled without the knowledge of the DC, blinding of the doctor could be achieved as well. Recently, with the development of a new device, The Sid Williams Research Instrument (SWRI), testing of this idea has become possible.

This instrument is a prototype which delivers an electronically controlled thrust and which senses the positional, force and velocity parameters of the adjusting stylus and makes them available to a computer for monitoring, storage, analysis and display. With this device, it is possible to study a wide variety of questions about precision adjusting and material properties.

The goal of the present study was to determine whether no-excursion, no-contact adjustments can in fact be used as placebos. We therefore chose to study whether changes in the adjustive thrust are important to outcome. Thrust was assumed to consist of two parts: preload and excursion. Preload is defined as the static force between the stylus and the tissue of the neck prior to delivery of the thrust. Excursion is defined as the distance the stylus moves during the thrust. Generally, the greater the preload and excursion, the more force encountered by the stylus. This work hypothesizes that the greater the preload and excursion, the more effective would be the adjustment.


Subjects for this experiment were volunteers obtained from the patient population of the Life College Clinic. Participation was contingent upon passage of the clinic physical and signing an informed consent. The experiment itself basically involves assessing each subject by xray, leg check and palpation, adjusting and then reassessing (fig. 1).

Upon acceptance, each subject was assigned randomly to one of four possible adjustment groups or a no-adjustment control group (fig. 2). The latter group received all assessments but did not undergo set-up or thrust; instead the subject was asked to walk up and down a hall for two minutes. Except for controls, subjects were blinded to their group assignments but were aware that group differences existed. The groups reflect the combinations of excursion and preload used: Group 0 has Zero preload and Zero excursion; Group 1 has 0 preload and 1/8 inch excursion; Group 2 has 1/4 lb preload and 0 excursion; Group 3 has 1/4 lb preload and 1/8 inch excursion. Note that in the Zero preload groups, the stylus is moved back far enough so that it never contacts the skin. The preload-excursion combination was set on the instrument at this time. The control group, not shown, got no adjustment.

After assignment, subjects were given a standard set of three upper cervical xrays: the lateral, nasium and vertex. In taking these films, placement was considered critical and was carefully monitored by two doctors. Following development, the xrays were analyzed using the Life Cervical method, resulting in a side of laterality, line of drive, rotational couple and formula for placing a stress bar to prestress the cervical spine. The analysis also yields variables which describe the atlas position and the cervical curve. These measures were used in combination to form an index of structural change in this study, in a way that will be described in detail later.

A pre-adjustment supine leg check was performed according to a typical upper cervical protocol. This involves having the subject start from a standing position, push up onto the table and recline as straight as possible. The assessor provides a small amount of headward pressure, gently removes inversion or eversion and then reads the length difference at the heel-last line. The leg check was always monitored and checked by a second assessor. Leg check results were scored as integers representing the number of 16ths of an inch short; a left leg short 3/16 would be scored as a minus three (-3), while a short right leg would yield a positive score; even legs would be scored as 0.

A subjective assessment was obtained by palpating bilaterally at the levels of the C1 and C2 vertebrae for pain, tenderness or spasm. Earlier work has demonstrated the reliability of this measure. Palpation is scored as a number which consists of the sum of qualities at each site. For example, a subject reporting pain and tenderness at C1 left and spasm at C2 right would be scored as a 3. The presence of no pain, tenderness or spasm would be scored as a zero; having all three qualities at all four sites would be scored as a 12.

The results of the palpation, leg check and x-ray analysis were recorded on individual data collection sheets coded for each subject for group membership.

Next, the adjustment was given. The subject was placed with the side of atlas laterality up and the opposite mastoid or temple supported on the headpiece according to the side of atlas listing. Headpiece, shoulderpiece and stressbar heights were set according to the results of the analysis of the lower cervical curve. The adjusting head was moved into place and the appropriate preload, either 0 or 1/4 lb, is applied. Delivery of the thrust was accomplished by activating a foot switch. Following the delivery of one thrust, the computer was programmed to plot the force and excursion data sampled during the actual thrust. At this time, the subject was given a post-adjustment leg check and palpation assessment, followed by a set of post-adjustment xrays.


In this study, we observed leg length difference, palpation and cervical xrays before and after a single adjustment of one of four types or no adjustment. For each variable, the major point of interest was how it has changed with the adjustment. To display such changes we performed the following transformation on all pre-post data:

PVC = (PRE - POST) / PRE X 100

The numbers generated by this transformation we call PERCENTAGE CHANGE VARIABLES, or simply PCV's. They let us see in a percentage fashion how close we have come to correcting or reducing a variable to its normal state. For instance, if the pre palpation score was 6 and the post palpation score was 0, this transformation yields a result of 100. This value means the adjusment has changed the variable to its normal or optimally correct value. Numbers less than 100 but greater than 0 mean a less-than-perfect correction. Greater than 100 means overcorrection. Numbers less than 0 mean changes in the wrong direction or a worsening.


Looking across a population of 50 subjects representing 10 individuals in each group (including the control) we see that there is no pattern evident in the distribution of the PCV's for leg length except for the last 10 which represent the controls (fig. 3a). Recall that negative values of the PCV mean things have gotten worse, i.e., the short leg got even shorter. Values of 100 represent even legs on the post check. Values greater than 100 mean that a short leg has gone long. We see that both over-corrections as well as worsenings occured. For the most part the controls showed no change. Examining group averages reveals that group 1 (0 preload, 1/8" excursion) showed the most change, although this is not significantly different that the other adjustment groups (fig 3b). The same conclusions can be drawn looking at the results of palpation viewed by group (fig. 4a). Again, only the control group appears to be markedly different, in that almost no pre-post change is seen. Viewing the averages shows that group differences are not significant (fig. 4b).

The most objective of the assessments used in this study was the xray. To obtain a measure of pre to post structural change, an index was used which was evolved from angles obtained during the analysis. This index consists of the average of the following four PCV's: (1) ATOCC, (2) ATAX, (3) CCL and (4) SSL. These variables represent changes in (1) atlas rotational position relative to occiput as seen on vertex, (2) atlas rotational position relative to axis as seen on vertex, (3) the angle between the atlas plane line and the skull center line as seen on nasium, and (4) the angle between the atlas plane line and the cervical spine as seen on nasium, respectively.

Examining this index by group, it appears as though there is an uneven distribution with no pattern (fig 5a). In this case, the control group is not seen to be markedly different from the other groups. Group averages exhibit no significant differences (fig. 5b). The control value is slightly less than the other group values but not by a significant amount.


These results are preliminary but they cast great doubt on the validity of using non-contact machine adjusting as a placebo or sham in blinded studies. The results instead suggest two interesting possibilities: either that low force, even off-the-skin adjusting is effective, or that the results may be attributable to the unreliability of the measures used.

The observation that single-thrust low force/no force adjusting is effective is not new; in fact, some adherents of at least one upper cervical technique group (Atlas Orthogonalists) routinely adjust with an instrument off the skin with excellent results as judged by criteria comparable to those used in this study. Their results are futher bolstered by the fact that their patients achieve clinical improvement over time; they are not limited to a one-time assessment such as was done here.

It is appropriate to question the degree to which these results are skewed by xray placement and analysis errors, especially given the control group outcome. Some of these errors may be considered as relatively constant, such as those due to xray misalignment, placement biases and methods of point choice. We have endeavored to eliminate these by maintaining aligned xrays and using more than one individual for placement and analysis decisions. Errors of these types would tend to show up on xrays as laterality biases or persistent distortions. We have not seen this. Another type of error, that due to random processes, should average itself out across a group of films. While we have attempted to reduce xray artifacts to a minimum, we acknowledge that no system is perfect. The fact that real changes were seen in the no-adjustment control group of magnitudes comparable to the adjusment groups suggests that the x-ray findings are questionable. That the changes seen in the control group were, on the average, lower suggests that the adjustments may have had some effect; exactly how much can not be assessed without accounting for the variance introduced by the x-ray uncertainty. The possibility of real effects can not be totally discounted, however, because of the differences seen in leg check and palpation between the adjustment and control groups.

The present results, while interesting, only serve to provide us with a number of new questions to answer. If the stylus does not need to touch the skin, how do adjustive forces reach the cervical complex to cause structural changes which reduce derangements? To answer this question, a number of issues have to be addressed. The first issue that needs to be resolved is the degree to which subject placement on the adjusting table affects the outcome. This can be studied by repeating the present work, up to and including the set-up, but performing no adjustment. The position of the subject can be documented by using the stylus position sensors to monitor landmarks on the head and body and then recording the measured headpiece and stressbar forces. In this fashion, the effects of different setup positions as determined by different headpiece, shoulderpiece and stressbar combinations can be studied.

Beyond this, it remains still to determine how much force might have been transmitted to the cervical complex through other means. One way this might have happened is by vibration set up by the adjusting cam which could have been transmitted through the headpiece to the cervical complex. Another possibility would be that of forces generated by neck muscles reflexively responding to vibration or to the sound of the cam and motor. To evaluate these possibilities, it will be necessary to analyze the force curves generated by the headpiece and stressbar transducers. We have these curves from the present study and are currently performing such an analysis.


The present work demonstrates that an adjusting instrument may be used as part of experimental designs where blinding of the doctor, patient and data analyzer is important. The results, however, do not support the hypothesis that low force/no force instrument adjusting will serve as an adequate placebo. Further work utilizing more reliable measures and a longer time frame is seen as necessary to validate these results.