As in a previous study  aimed at the profession in the US, osteopathic physicians in the present study used a wide range of techniques for treatment of spinal and pelvic somatic dysfunction. Respondents indicated a preference for soft tissue based techniques and – compared to studies from other countries – a greater preference for cranial methods. This utilization of OCF was higher than expected, based on the previous American survey  and studies of osteopaths in the UK and Australia [15, 18]. Survey responses also suggest a gender bias for some of the treatment techniques.
The reported preference of survey respondents for cranial techniques is interesting. Johnson and Kurtz  reported that OCF was ranked last on a list of manual techniques used by osteopathic physicians. In that survey, however, those respondents who listed OMT as a specialty used a broader range of techniques and had a significantly greater preference for OCF than those who did not list OMT as a specialty . Perhaps the similarity of this subgroup with our survey respondents explains the similar preference for cranial approaches. In contrast, osteopaths in the UK and Australia (where OMT is the primary treatment modality) [15, 18] seem less inclined to use OCF. Snapshot studies in those countries report cranial approaches are used on 23% of patients, but those studies also report what techniques were used for patients over a full treatment. Therefore, a treatment specific to spinal or pelvic dysfunction would likely involve even less use of cranial techniques because OCF is emphasized for dysfunction in the "involuntary mechanism" (the mechanism postulated to underlie the cranial rhythmic impulse) and may be used to assess different phenomena than biomechanical function of the spine and pelvis.
The Fryette model of spinal coupled motion was commonly used by respondents. This result is expected since American osteopathic textbooks advocate assessment based on these principles [1, 5, 6, 11]. The use of the Fryette model by osteopaths in other countries has not been examined, but it is expected to be less given that many authors from other countries have not used this model [8–10]. The Fryette model has been criticized for its prescriptive diagnostic labeling and questionable inferences concerning motion restriction from static positional assessment [21, 22]. Recent studies suggest that spinal coupled motions are inconsistent and there is variability between spinal levels and between individuals for motion in the lumbar, thoracic, and, to a lesser extent, cervical spine [22–28]. Lack of consistency in coupled spinal motion should be a concern to those who advocate the Fryette model as a means of predicting triplanar motion restrictions. Given the common use of this model and its endorsement from the Educational Council of Osteopathic Principles (ECOP) , which guides the curriculums of OMM programs, the American profession may wish to re-examine the validity and usefulness of the Fryette model.
Reported preferences for assessment of the pelvis and sacroiliac joints were consistent with the biomechanical and treatment model proposed by Mitchell  and advocated by most American osteopathic texts [1, 5, 6, 11]. The Mitchell model recommends the use of motion tests (typically the flexion tests) to determine the side of the dysfunction and the identification of landmark asymmetry to determine the type of dysfunction. The positive responses for assessment of symmetry of pelvic landmarks in the present study were similar to an Australian study, suggesting that this model is also prevalent outside the US . Motion testing of the sacroiliac joint in the present study was commonly performed, and cranial diagnosis for the pelvis and sacroiliac joint was also popular. The Australian study reported frequent use of motion tests, but a considerably lower use of cranial diagnosis (30% compared to 61% in the present study) .
Sacroiliac joint pain provocation tests, procedures intended to reproduce the patient's familiar pain by stressing the sacroiliac joint and thus implicating it as a pain generator, were used by just over half the respondents. The most popular test, the active straight leg raise, is rarely referred to in the osteopathic literature, and some respondents may have confused it with the straight leg raise for nerve root irritation. ASIS compression (performed in a similar way to the motion test but intended to provoke the complaint) was also reportedly used by nearly half of respondents. These two tests are rarely mentioned in osteopathic texts, but their reliability and validity are supported by the scientific literature [30, 31]. Given these circumstances, it is interesting that they are used by a substantial proportion of respondents. An Australian survey of osteopaths found a comparable use of pain provocation tests , which suggests that both groups use a pragmatic approach to patient care that includes procedures other than those typically recommended by the profession.
Female respondents reported more frequent use of soft tissue technique, muscle energy, and strengthening exercises for treatment of the spine and pelvis, whereas male respondents more frequently used HVLA. Similarly, Johnson et al  found that female respondents were more likely than men to use indirect techniques. These gender preferences may reflect the physical strength required to perform direct techniques, such as HVLA. Alternatively, it is possible that the patient populations seen by male and female practitioners are different, and that the patients seen by female practitioners (who may comprise more women and children) may prefer and request more gentle treatment approaches.
While most respondents in the present study reported that they document physical findings and the type of OMT delivered, a substantial minority do not frequently document their findings or treatment. This represents a potential concern for the profession given the importance of maintaining accurate records of treatment, particularly in the event of adverse reactions or litigation. The common use of the Fryette spinal model explains the documentation of somatic dysfunction using this model's positional notation. Most respondents bill for OMT and use a -25 modifier. By using the -25 modifier, both the patient visit and OMT can be billed at the same visit, providing an economic incentive for OMT use. These findings suggest a healthy amount of business savvy amongst this group and an awareness of efficient billing practices for OMT. Respondents reported a low use of diagnostic imaging prior to OMT, which is consistent with current guidelines that state plain imaging is of little use for non-specific spinal pain .
This study has a number of limitations. The response rate to the survey was relatively low and generalizing these results to the entire AAO membership may be inappropriate. However, according to data provided by the AAO, the demographics of the sampled respondents closely reflect the membership as a whole. Respondents and membership were strikingly similar for gender (69% and 60% male, respectively) as was place of osteopathic training, with KCOM, UNECOM, and NYCOM listed as the most common institutions for both the study respondents and AAO membership (each institution accounting for 11% of respondents compared to 13%, 10%, and 9% of the total membership). Study respondents listed OMM/Neuromusculoskeletal Medicine (60%) and Family Practice/OMT (51%) as their most common specialties, which were also the most common specialties for the AAO membership (36% and 43%, respectively, followed by a list of other specialties which each accounted for less than 5% of members). Although a higher percentage of study respondents than AAO members reported their specialties as OMM/Neuromusculoskeletal Medicine and Family Practice/OMT, this increase may be a result of limiting the survey to only those members who use OMT and of the exclusion of students, interns, and residents. No information was available on years of practice experience for members of the AAO, but the comparisons above support the study sample as being representative of the entire membership.
Members of the AAO were targeted because they have an expressed interest in OMM. Because the use of OMT is diminishing within the osteopathic profession, study respondents do not necessarily represent the broader profession. Additionally, due to the relatively low response rate for the survey, a sample bias may favor more computer savvy respondents. Those practitioners who did not respond to the email invitation or did not have an email address listed with the AAO may potentially have been less proficient in newer methods and, thus, have different treatment preferences. Therefore, caution is necessary when generalizing the results to the membership of the AAO even though similar demographics support this group as representative. Researchers have recommended email and web-based surveys as holding great promise as a fast, inexpensive medium for health research, but comparisons between survey delivery methods have so far demonstrated a greater response rate to postal surveys [33, 34]. A combined postal and email survey would likely have improved the response rate, and this method should be considered for future surveys until the time the growing internet culture favors a better response rate for electronic distribution.
It should be understood that this study surveyed what respondents said they did in practice and was not a snapshot survey or a record of what approaches respondents actually used in practice. The responses are therefore subject to recall bias of the practitioners. Although requiring more resources, a snapshot survey is a worthwhile addition for future studies to provide a more accurate indication of technique preferences in the practice setting.
Although the intention of this study was to determine what treatment approaches osteopathic physicians use, a review of the validity and reliability of these approaches may provide additional context for interpreting study results. For instance, few procedures for segmental dysfunction – with the exception of palpation for tenderness and pain provocation – have acceptable interobserver reliability . For assessment of pelvic and sacroiliac dysfunction, landmark asymmetry and tests that assess sacroiliac motion have been criticized for poor reliability and lack of validity [31, 34–36]. While cranial approaches for the diagnosis and treatment of the spine and pelvis were favored by study respondents, their use is debated within the profession, with criticisms ranging from the biological plausibility of the concept to the lack of examiner reliability and outcome studies [37, 38]. Using the results from this study, which provides information about tests that are in common use, future studies could be designed to improve the reliability of these procedures. Alternatively, the osteopathic profession may want to reconsider the use of these tests.