J Manipulative Physiol Ther 2004 (Nov); 27 (9): 547—553 ~ FULL TEXT
Mitchell Haas, DC, Elyse Groupp, PhD, Mikel Aickin, PhD,
Alisa Fairweather, MPH et al.
Center for Outcome Studies,
Western States Chiropractic College,
Objective: To acquire information for designing a large clinical trial and determining its feasibility and to make preliminary estimates of the relationship between headache outcomes and the number of visits to a chiropractor.
Design: Randomized, controlled trial.
Setting: Private practice in a college outpatient clinic and in the community.
Subjects: Twenty-four adults with chronic cervicogenic headache.
Methods: Patients were randomly allocated to 1, 3, or 4 visits per week for 3 weeks. All patients received high-velocity low-amplitude spinal manipulation. Doctor of Chiropractics could apply up to 2 physical modalities at each visit from among heat and soft tissue therapy. They could also recommend modification of daily activities and rehabilitative exercises. Outcomes included 100–point Modified Von Korff pain and disability scales, and headaches in last 4 weeks.
Results: Only 1 participant was insufficiently compliant with treatment (3 of 12 visits), and 1 patient was lost to follow-up. There was substantial benefit in pain relief for 9 and 12 treatments compared with 3 visits. At 4 weeks, the advantage was 13.8 (P = .135) for 3 visits per week and 18.7 (P = .041) for 4 visits per week. At the 12–week follow-up, the advantage was 19.4 (P = .035) for 3 visits per week and 18.1 (P = .048) for 4 visits per week.
Conclusion: A large clinical trial on the relationship between pain relief and the number of chiropractic treatments is feasible. Findings give preliminary support for the benefit of larger doses, 9 to 12 treatments, of chiropractic care for the treatment of cervicogenic headache.
From the FULL TEXT Article:
Headache (HA) is one of the most common ailments. With prevalence in the general population of about 16%,  epidemiologic studies report that 4% of those adults have HAs on a daily basis.  Approximately 7 million people suffer from HAs at least every other day, which translates into millions of lost workdays and severely impacts work efficiency. 
Cervicogenic HA, defined as pain originating in the cervical spine and referred to the head,  became ensconced in the International Headache Society diagnostic classification in 1988.  Although not as prevalent as tension-type headache or migraine, cervicogenic HA nevertheless is diagnosed in 0.4% to 2.5% of the general population. [6, 7] However, in pain clinics, cervicogenic HA occurs in 33.8% of HA patients.  According to a recent survey of complementary and alternative therapy use, chiropractic was frequently selected for the relief of head and neck pain,  accounting for 18 to 38 million manipulations performed annually. [10, 11]
The scientific evidence to support the efficacy/effectiveness of cervical manipulation for the relief of chronic HA has been assessed in systematic reviews of random, controlled trials. [11–14] The reviews conclude that more rigorous studies need to be performed to make definitive recommendations about treatment efficacy. However, the most recent systematic reviews [12, 13] have found higher quality evidence supporting manipulation as an efficacious therapy for cervicogenic HA. Three studies have found manipulation superior to placebo [15, 16] and to a no-treatment control. 
An important question that is not addressed in any of the efficacy studies is the possible effect that treatment dosage may have on HA outcomes. Recommendations for the duration and frequency of chiropractic care vary, [18–20] because there is no evidence-based guideline for optimal treatment. The purpose of this pilot study was to determine the feasibility of a large clinical trial and to gather preliminary data on the dose-response relationship between the number of chiropractic treatments and relief of pain in cervicogenic HA sufferers.
Our pilot study showed the feasibility of a larger randomized trial. Participants were willing to be randomized, follow the treatment protocol, and complete the baseline and mail follow-up questionnaires. Only 1 person dropped out of the study because of disappointment that treatment visits were not an hour long. Noncompliance with the treatment schedule was minimal (1 person). Perhaps the success with these feasibility criteria can be attributed to 3 design strategies. We included 2 baseline visits as a strategy for reducing the dropout rate.  Persons who missed either baseline appointment were excluded from the study under the assumption that they were more likely to be noncompliant. The second strategy was regular telephone calls to patients to maintain rapport and keep them engaged in the study. Participants also received a nominal sum of $5 for each returned questionnaire.
The data suggest preliminary support for a dose-response relationship between cervicogenic HA pain/disability and the number of chiropractic treatments. Larger numbers of visits produced greater benefit. It remains to be determined if the observed benefit is attributable to larger total treatments or greater short-term concentration of care. It also appears that the range of treatments under study fell in the linear range of the dose-response curve with much room for health improvement possible. This implies that more treatments may be required to achieve maximum benefit and saturation of the dose-response curve. However, it must be pointed out that caution is required in interpreting the data. Because of the small sample size, the groups may have been unbalanced in important prognostic variables that could not be controlled in the analysis. Even statistically significant results must therefore be viewed with caution. Confidence intervals were also large, so that effect sizes and the shapes of the dose-response curves are still far from clear. Long-term benefit of care was not evaluated.
It should be noted that the patient population was limited to chronic cervicogenic HA. Persons were excluded for concomitant HAs such as migraine. This can slow recruitment, because many cervicogenic HA patients also suffer migraines.  A trial on mixed HAs may be valuable, because manipulation appears beneficial for both types of HA.  It is interesting to note that medication for migraines has not been successful in treating the cervicogenic component of a cervicogenic/migraine mix. 
Few participants sought additional care outside the study for their HAs. This is encouraging for future studies because there was minimal potential confounding of study results by outside care. Participants were permitted to seek outside care for ethical reasons and because we wanted honest accounting of all treatment. Reporting of outside care could have been compromised by any injunction against it.
Additional lessons were learned about the order of activities at the 2 baseline visits. We started the study with potential participants attending an information session at the first baseline and the physical examination and first treatment at the second visit. For the second half of the study, persons received the screening examination at the first visit. The advantage of the first design is that money is saved on examinations for people who change their minds about joining the study and do not show up for the second visit. One disadvantage is that there may be insufficient time for examination and treatment on the same day. Patients attending satellite clinics run the risk of having to make multiple trips to campus for radiographs. In the second design, individuals are given time to think over participation for several days after they have all clinical information available. We found the second method more logistically viable for a clinical trial.
A large clinical trial on the relationship between cervicogenic HA outcomes and treatments from a chiropractor is feasible. Findings suggest the benefit of 9 to 12 visits over 3 weeks for the treatment of HA/neck pain and disability. A larger number of visits than 12 in 3 weeks may be required for maximum relief and durability of outcomes.