This news release issued by the American Osteopathic Association describes a relatively rigorous randomized controlled trial that compared the impact of osteopathic manipulation, light touch massage and standard antibiotic and mobilization treatment on pneumonia outcomes in some older, hospitalized patients. The outcomes measured — length of stay and overall deaths — showed statistically significant lower rates in both measures for subsets of 375 patients who participated in the study, and the study protocols incorporated two separate, accepted and complementary means of sorting out such factors as severity of illness and treatment compliance that make a reasonable case for comparable study groups.
That said, the release did not do justice to the study author’s own acknowledgments of their research’s limitations or weaknesses. Nor did it reflect the investigators’ reasoned call for further research to better pin down demonstrated benefits from osteopathic manipulation. In this way, the release and the AOA missed a very good opportunity to address a) the widespread skepticism that greets much osteopathic medicine and research; and b) questions among the public and professionals about how and why adjunctive medical practices work, and under what conditions they should be used to treat patients. Instead, the release treats the potential benefits as a proven done deal. Those who take the time (and have the capacity) to examine the complicated graphic chart that accompanies the release might get more out of the release, but it’s unlikely that many of the intended audiences for the release will plow through it.
The new report is essentially several subgroup analyses but the release doesn’t mention the existence of the earlier study that reported the primary outcomes. This is a theme we’ve seen in other news releases such as here and here.
Pneumonia, whether acquired in the hospital or elsewhere, is an often lethal and widespread disease in elderly hospitalized patients, as the research paper and release itself makes clear. Antibiotic use, and evidence-based treatment protocols that work on getting most patients with pneumonia out of bed for frequent if limited periods of time have already greatly reduced lengths of stay and in-hospital deaths from bacterial forms of the disease. But because antibiotic overuse is increasingly risky, and because pneumonia accounts for repeated hospitalizations in some elderly and immunocompromised patients, those at risk for pneumonia and their caregivers are rightly interested in any non-pharmaceutical and non-invasive means to further improve outcomes.
The release acknowledges that osteopathic manipulative treatment (OMT) was designed in the pre-antibiotic era and it isn’t clear how OMT provides benefit. The current release focuses on new analyses of a prior publication from 2010. The original study reported, “Intention-to-treat (ITT) analysis (n = 387) found no significant differences between groups. Per-protocol (PP) analysis (n = 318) found a significant difference between groups (P = 0.01) in LOS [length of stay].” The release could have been clearer about what’s new in this report released 6 years later. Indeed, the ITT analyses also showed no effect in any subgroup in the new study. More discussion also could have been focused on differences between OMT and the light-touch group in addition to the conventional care group. The light-touch may reflect a “sham” version of OMT and how results differ between the light touch and conventional care groups may differentiate between the extra time that may have been spent with the patient and any specific effect of OMT itself.
The release does not address the costs of osteopathic adjunct therapy even in general terms. Readers would have no idea, for example, if medical insurers might cover the cost for in-hospital or post-discharge osteopathic care. And it might have been highly useful to include easy-to-get information about the average cost of an extra day or two in the hospital for pneumonia treatment. Given that payment for patients hospitalized with pneumonia is typically based upon a bundled fee, the costs of providing OMT may be offset if length of stay is reduced.
The release briefly but adequately covers the major quantitative findings. For example, the 1.1 reduction in hospital stay mentioned in the release is an absolute reduction. The release would have been greatly strengthened by including more of the results in the narrative that are confined to the graphic. Describing the actual length of stay in each group, which is provided in the graphic, as opposed to just the amount of the reduction would provide some context for how meaningful the benefit it is. (i.e. a 1-day reduction on a 30-day stay being less impressive than a 1-day reduction on a 4-day stay). It also could have given more attention to a comparison of the full manipulative therapy and “light touch” therapy groups. For the mortality data, it wasn’t clear from the text if they are referring to an absolute or relative risk reduction.
Although it’s commonly accepted that osteopathic manipulation is “harmless,” the release should have addressed any potential complications, discomforts, cost issues or other side effects.
Here is the category with perhaps the biggest deficiencies. The release is based on a study of subgroup analyses from another already published study and that really needed to have been highlighted in the release. The primary outcomes study (mentioned in the introduction of the current paper) found no difference in intention-to-treat analyses (ITT) but some differences in per-protocol ones. (ITT analysis means all patients who were enrolled and randomly allocated to treatment are included in the analysis. Per-protocol analysis is a comparison of treatment groups that includes only those patients who completed the treatment originally allocated.)
The authors of the study go to some lengths to discuss the limitations and weaknesses of their findings, explicitly noting that “possible explanations” for their positive results could include a “physiologic response to attention and touch,” but also “early mobilization” created by the need to position the patients for manipulation and also “anticipatory belief in the treatment” (a kind of placebo effect). The release also does not offer any explanation as to what mechanisms (physiologic, psychological, etc.) might be responsible for the reported benefits. Nor does it give any description of what the manipulation therapy is composed of. The researchers clearly call for additional research to sort out the “niche” applications of adjunctive osteopathic manipulation, and which patients at highest risk would most benefit. The release, however, quotes the study author as saying only that “this study should encourage physicians to use their osteopathic techniques when treating older patients with pneumonia,” a blanket assertion not supported by the data.
The table highlights that no statistically significant differences were found in any intention-to-treat analyses — the most rigorous ones in a randomized trial. Only in per-protocol analyses do some statistically significant differences emerge. Even there the results don’t fully make sense. Length of stay is lower with OMT in younger patients than those with more severe disease (PSA class IV). This would in some ways seem contradictory.
No mongering. Pneumonia in the hospitalized elderly is a serious issue. However, this study also included younger patients and those with less severe pneumonia sick enough to be hospitalized.
We gave this a pass for barely meeting the criteria. The release notes that the trial was funded by a consortium of foundations including the Foundation for Osteopathic Health Services. It doesn’t specifically disclose potential conflicts of interest but directs readers to the study for “full details.”
When we looked at the disclosures, it was clear that the researchers were aligned with or officially affiliated with osteopathic institutions and foundations.
The release explains the comparison groups in the study and offers the graphic chart. This category would have been improved with more detail, particularly what exactly is involved the treatment arms (OMT and light touch).
The release notes that osteopathic manipulative therapy is established practice for osteopaths and presumably could be available in hospitals more broadly.
But we thought the availability of the procedure could have been made much more clear. How would a patient know if their provider was an osteopath? What would it take to implement this more broadly? Could non-osteopaths be taught to do this? Could non-physicians? It’s implied that osteopathic physicians who take care of patients hospitalized with pneumonia could do this. But do patients really know whether their physician is an osteopath?
The release doesn’t state that this study was a follow-up to a previously published report. It does say that the current study focuses on subgroup analyses and that may imply that the main analyses were previously released. But it’s doubtful a lay person or even a professional would know this.
The release gets to the “news” and describes the outcomes, but the release would have benefited from additional context and caveats as noted above.
The release really needed a quote or other statements qualifying the call for essentially immediate and wide use of manipulation in elderly patients with pneumonia. The statement related to the benefits needed some moderation. In a randomized controlled trial, only intention-to-treat analyses of the primary outcomes really count. It doesn’t appear that this study found any significant analyses in primary, ITT analyses. As such, all these subgroup analyses are hypothesis generating and would require confirmation in a new trial focused on the subgroups demonstrating OMT benefit.