Design

A comprehensive search of peer-reviewed articles published in English between 2000 and 2015 reporting new empirical research findings of key aspects of MT use among patients with migraine and non-migraine headache disorders was undertaken. Databases searched were MEDLINE, AMED, CINAHL, EMBASE and EBSCO. The key words and phrases used were: ‘headache’, ‘migraine’, ‘primary headache’, ‘cephalgia’, ‘chronic headache’ AND ‘manual therapy’, ‘spinal manipulation’, ‘manipulative therapy’, ‘spinal mobilization’, ‘chiropractic’, ‘osteopathy’, ‘massage’, ‘physical therapy’ or ‘physiotherapy’ AND then ‘prevalence’, ‘utilization’ or ‘profile’ was used for additional searches against the previous terms. The database search was accompanied by a hand search of prominent peer-reviewed journals. All authors accessed the reviewed literature (data) and provided input to analysis.

Due to the focus of the review, literature reporting randomized control trials and similar clinical research designs were excluded as were articles identified as letters, correspondence, editorials, case reports and commentaries. Further searches were undertaken of the bibliographies in the identified publications. All identified articles were screened and only those reporting new empirical findings on MT use for headache in adults were included in the review. Articles identified and selected for the review were research manuscripts mostly within epidemiological and health economics studies. The review includes papers reporting MT use pooled with the use of other therapies, but only where MT patients comprised a large proportion (as stated) of the included study population. Results were imported into Endnote X7 and duplicates removed.

Search outcomes, analyses and quality appraisal

Figure 1 outlines the literature search process. The initial search identified 3286 articles, 35 of which met the inclusion criteria. Information from each article was organized into a review table (Table 1) to summarise the findings of the included papers. Information is reported under two selected headache groups and within each individual MT profession - chiropractic, physiotherapy, osteopathy and massage therapy – where sufficient detail was available.

Authors Year	Country/ Region	Population/ Profession	Study Method	Sample size	Themes †	Prevalence use based on Headache Groupings
Ailliet et al 2010 [65]	Europe (Belgium)	Manual Therapy population/Chiropractic	Postal questionnaire by chiropractors	517 patients	1	Headache: Chiropractic 1.9%
Bethell et al 2013 [76]	North America	General population	Secondary analysis of national survey	2411	2
Bigal et al 2008 [19]	North America	General population	Longitudinal study following a cohort of headache sufferers	Chronic migraine (520), Episodic migraine (9424)	1	Chronic migraine: Chiropractic 36.2% Physiotherapist 13.3% Episodic migraine: Chiropractic 25.7%, Physiotherapy 4.2%
Brown et al 2014 [61]	Australia	Manual Therapy population/Chiropractic	Cross-sectional survey completed by patients	486	1	Headaches: Chiropractic 5.5%
Cherkin et al 2002 [62]	North America	Manual Therapy population/Chiropractic	Practitioner completed questionnaire	2550	1	Headaches: Chiropractic Massachusetts 4.6%, Arizona 6.4%
Cooke et al 2010 [49]	North America	General population	Telephone survey to public	1210	1	Migraine: Chiropractic 6%, Massage 2%, Osteopathy 1%
Coulter et al 2002 [66]	North America	Manual Therapy population/Chiropractic	Patient questionnaires	1275	1	Headaches: Chiropractic 4.0%
Brown et al, 2013 [69]	Australia	Manual Therapy population/Chiropractic	Cross-sectional general population survey questionnaire	757	1	Headache: Chiropractic 45.5%
French et al 2013 [64]	Australia	Manual Therapy population/Chiropractic	Cross-sectional observational practitioner survey	4464	1	Headaches: Chiropractic 4%
Gaul et al 2009 [70]	Europe (Germany/Austria)	Headache clinic population	Questionnaire based patient survey	432	1,2,3,4	Mixed primary headaches: Massage 46.1%, Physiotherapy 27.8%
Gaul et al 2011 [72]	Europe (Germany/Austria)	Headache clinic population	Questionnaire-based survey	448	1,2	Migraine (78.5%): Physiotherapy 18.7%, Massage 56.4%
Gaumer G 2006 [56]	North America	General population	Random telephone survey	800	1	Headaches: Chiropractic 5.3%
Goksel et al 2014 [73]	Europe (Turkey)	Headache clinic population	Patient questionnaire through interview	110	1,2,4	Migraine (64.6%): Massage 51%
Hartvigsen et al 2003 [68]	Europe (Denmark)	Manual Therapy/Chiropractic	Questionnaire data collected by practitioners	1897 patients	1	Headache: Chiropractic 4%
Jackson P 2001 [63]	North America	Manual Therapy population/Chiropractic	Postal questionnaire to chiropractors	1500	1	Headaches: Chiropractic 15.4%
Kristoffersen et al 2012 [20]	Europe (Norway)	General population	Cross-sectional epidemiological survey	405	1,2	All Primary Headaches: Chiropractic 28% Physiotherapy 52%
Kristoffersen et al 2013 [79]	Europe (Norway)	General population	Cross-sectional epidemiological postal survey and clinical interview	253 primary 82 secondary	4
Lambert et al 2010 [77]	Europe (UK)	Headache clinic population	Self-administered questionnaire	92	2,3
Lyngberg et al 2005 [1, 52]	Europe (Denmark)	General population	Medical doctor interviews	740	1	Mostly migraine: Chiropractic 9% Physiotherapy 5%
Malone et al 2012 [71]	North America	General population	On-line survey via migraine website	2735	1	Migraine: Massage 29.7%
Minen et al 2014	North America	Headache clinic population	Secondary analysis of baseline questionnaire data	225	1,2	Migraine with/without aura: Chiropractic 27.1%, Massage 18.2%, Physiotherapy 4.9%
Morin et al 2014 [54]	North America (Quebec)	Manual Therapy population	Prospective survey	1402	1	Migraine: Osteopathy 1.7% Headaches: Osteopathy 2.7%
Ndetan et al 2009 [57]	North America	General population	Secondary Survey analysis	31248	1	Headache: Chiropractic 15.1%
Orrock P 2009 [75]	Australia	Manual Therapy population/Osteopathy	Mailed practitioner questionnaire	2238 patient records	1	Headache: Osteopathy 10%
Ossendorf et al 2009 [60]	Europe (Germany)	Pain clinic population	Physician-administered structured interview and questionnaires	288 (136 with Headache)	1,4	Headache: Chiropractic 22%, Physiotherapy 35%, Osteopathy 9%, Massage 54%
Rossi et al 2005 [53]	Europe (Italy)	Headache clinic population	Physician-administered structured interview	481	1,2,3,4	Migraine: Massage 10.1%, Chiropractic 8.9%, Osteopathy 2.7%
Rossi et al 2006 [58]	Italy	Headache clinic population	Physician-administered structured interview	110	1,2,3,4	Headache (CTTH): Chiropractic 21.9%, Massage 17.8%
Rossi et al 2008 [59]	Europe (Italy)	Headache clinic population	Physician administered structured interview	100	1,2,4	Headache (cluster): Chiropractic 12%, Acupressure 12%
Rubinstein et al 2000 [67]	Europe (Netherland)	Manual Therapy population/Chiropractic	Retrospective patient questionnaires	833	1	Headache: Chiropractic 7%
Sanderson et al 2013 [21]	USA, Canada, UK, Germany, France and Australia	General population	Web-based screening questionnaire	16663	1,3	Chronic migraine: 10% USA Canada 10%, France/UK 0%, Germany 1%, Australia 14% Episodic Migraine: USA 7% Canada 4%, France/UK 1%, Germany 6%, Australia 14%
S von Peter et al 2002 [78]	North America	Headache clinic population	Patient interview using a standardized questionnaire	73	2,3,4	Tension, Migraine (27%) and other headaches: Chiropractic 15.1%, Massage 42.5%
Vukovic et al 2010 [48]	Europe (Croatia)	General population	Random cross-sectional survey questionnaire	616	1	Migraine: Chiropractic 9.5%, Physiotherapy 19.4% Tension headache: Chiropractic 4.0%, Physiotherapy 12.2%
Wells et al 2010 [51]	North America	General population	National cross-sectional survey sample	23,393	1	Migraine 18.5% and Headaches 15.7%: Chiropractic/massage pooled
Wells et al 2011 [50]	North America	General population	National cross-sectional survey sample	23,393	1,3	Migraine: Chiropractic 15.4%, Massage 15.1%
Xue et al 2008 [55]	Australia	General population	Cross-sectional telephone survey	1067	1	Headaches: Chiropractic 9.3%

Themes †: 1 = MT prevalence use, 2 = Profile and motivations, 3 = Concurrent use, 4 = Self-reported effectiveness

Prevalence of MT use

Thirty-one of the reviewed articles with a minimum sample size (>100) reported findings regarding prevalence of MT use. The prevalence of chiropractic use for those with migraine ranged from 1.0 to 36.2% (mean: 14.4%) within the general population [19–21, 48–52] and from 8.9 to 27.1% (mean: 18.0%) within headache-clinic patient populations [53, 54]. The prevalence of chiropractic use for those reported as headache ranged from 4 to 28.0% (mean: 12.9%) within the general population [20, 48, 51, 55–57]; ranged from 12.0 to 22.0% (mean: 18.6%) within headache/pain clinic patient populations [58–60] and from 1.9 to 45.5% (mean: 9.8%) within chiropractic patient populations [61–69].

The prevalence use of physiotherapy for those with migraine ranged from 9.0 to 57.0% (mean: 24.7%) within the general population [19, 20, 48, 52] and from 4.9 to 18.7% (mean: 11.8%) within headache-clinic patient populations [54, 70]. The prevalence use of physiotherapy for those reported as headache ranged from 12.2 to 52.0% (mean: 32.1%) within the general population [20, 48] and from 27.8 to 35.0%% (mean: 31.4%) within headache/pain clinic populations [60, 70].

Massage therapy use for those with migraine ranged from 2.0 to 29.7% (mean: 15.6%) within the general population [49, 50, 71] and from 10.1 to 56.4% (mean: 33.9%) within headache-clinic populations [53, 54, 72, 73]. Massage/acupressure use for those reported as headache within headache/pain clinic patient populations ranged from 12.0 to 54.0% (mean: 32.5%) [58–60, 70].

Osteopathy use for those with migraine was reported as 1% within the general population [49]; as 2.7% within a headache-clinic patient population [53] and as 1.7% within an osteopathy patient population [74]. For headache the prevalence was 9% within a headache/pain clinic population [60] and ranged from 2.7 to 10.0% (mean: 6.4%) within osteopathy patient populations [74, 75].

The combined prevalence rate of MT use across all MT professions for those with migraine ranged from 1.0 to 57.0% (mean: 15.9%) within the general population; ranged from 2.7 to 56.4% (mean: 18.4%) within headache-clinic patient populations and was reported as 1.7% in one MT patient population. The combined prevalence rate of MT use across all MT professions for those reported as headache ranged from 4.0 to 52.0% (mean: 17.7%) within the general population; ranged from 9.0 to 54.0% (mean: 32.3%) within headache-clinic patient populations and from 1.9 to 45.5% (mean: 9.25%) within MT patient populations.

Profile and motivations for MT use

While patient socio-demographic profiles were not reported within headache populations that were exclusively using MT, several studies report these findings where MT users made up a significant percentage of the non-medical headache treatments utilized by the study population (range 40% – 86%: mean 63%). While findings varied for level of income [58, 70] and level of education, [70, 72, 73] this patient group were more likely to be older [70, 72], female [20], have a higher rate of comorbid conditions [58, 70, 76] and a higher rate of previous medical visits [20, 58, 70] when compared to the non-user group. Overall, this group were reported to have a higher level of headache chronicity or headache disability than non-users [20, 54, 58, 70, 72, 77].

Several studies within headache-clinic populations report patient motivations for the use of complementary and alternative headache treatments where MT users made up a significant proportion of the study population (range 40% – 86%: mean 63%) [58, 70, 72, 78]. From these studies the most common motivation reported by study patients was ‘seeking pain relief’ for headache which accounted for 45.4% – 84.0% (mean: 60.5%) of responses. The second most common motivation was patient concerns regarding the ‘safety or side effects’ of medical headache treatment, accounting for 27.2% – 53.0% (mean: 43.8%) of responses [58, 70, 72]. ‘Dissatisfaction with medical care’ accounted for 9.2% – 35.0% (mean: 26.1%) of responses [58, 70, 72].

A limited number of reviewed papers (all from Italy) report on the source of either the referral or recommendation to MT for headache treatment [53, 58, 59]. From these studies, referral from a GP to a chiropractor ranged from 50.0 to 60.8% (mean: 55.7%), while referral from friends/relatives ranged from 33.0 to 43.8% (mean: 38.7%) and self-recommendation ranged from 0 to 16.7% (mean: 5.6%). For massage therapy, referral from a GP ranged from 23.2 to 50.0% (mean: 36.6%), while referral from friends/relatives ranged from 38.4 to 42.3% (mean: 40.4%) and self-recommendation ranged from 7.7 to 38.4% (mean: 23.1%). For acupressure, referral from a GP ranged from 33.0 to 50.0% (mean: 41.5%), while referral from friends/relatives was reported as 50% and self-recommendation ranged from 0 to 16.6% (mean: 8.3%). One study reported findings for osteopathy where referral from both GP’s and friends/relatives was reported as 42.8% and self-recommendation was reported as 14.4%. Overall, the highest proportion of referrals within these studies was from GPs to chiropractors for chronic tension-type headache (56.2%), cluster headache (50%) and migraine (60.8%).

Concurrent use and order of use of headache providers and related communication of MT users

Several studies report on the concurrent use of medical headache management with complementary and alternative therapies. In those studies where the largest percentage of the patient population were users of MT’s (range 57.0% – 86.4%: mean 62.8%), [58, 70, 78] concurrent use of medical care ranged between 29.5% and 79.0% (mean: 60.0%) of the headache patient population.

These studies further report on the level of patient non-disclosure to medical providers regarding the use of MT for headache. Non-disclosure ranged between 25.5 and 72.0% (mean: 52.6%) of the patient population, with the most common reason for non-disclosure reported as the doctor ‘never asking’, ranging from 37.0 to 80.0% (mean: 58.5%). This was followed by a patient belief that ‘it was not important for the doctor to know’ or ‘none of the doctor’s business’, ranging from 10.0 to 49.8% (mean: 30.0%). This was followed by a belief that either ‘the doctor would not understand’ or ‘would discourage’ these treatments, ranging from 10.0 to 13.0% (mean: 11.5%) [53, 77].

One large international study reported the ordering of the typical provider of headache care by comparing findings between several countries for migraine patients [21]. Primary care providers followed by neurologists were reported as the first and second providers for migraine treatment for nearly all countries examined. The only exception was Australia, where those with chronic migraine selected chiropractors as typical providers at equal frequency to neurologists (14% for both) while those with episodic migraine selected chiropractors at a greater frequency to neurologists (13% versus 5%). Comparatively, chiropractors were selected as the typical provider for those with chronic migraine by 10% in USA and Canada, 1% in Germany and 0% for UK and France. Chiropractors were selected as the typical provider for those with episodic migraine by 7% in USA, 6% in Germany, 4% in Canada and by 1% in both the UK and France.

Future research

Despite the strong need for more high quality research to assess the efficacy of MT as a treatment for headache, the substantial use of MT brings attention to the need for more public health and health services research within this area of headache management. The need for this type of research was identified in a recent global report on the use of headache-related healthcare resources [18]. Furthering this information can lead to improvements in healthcare policy and the delivery of healthcare services.

The substantial use of physical therapies such as MT has been under-reported within many of the national surveys reporting headache-related healthcare utilization [3, 5, 83–85]. Regardless, the role of physical therapies in headache management continues to be assessed, often within mainstream and integrated headache management settings [86–89]. Continuing this research may further our understanding of the efficacy and outcomes associated with a more multidisciplinary approach to headache management.

Further to this is the need for more research to understand the healthcare utilization pathways associated with those patients who use MT in their headache management. Little is known about the sociodemographic background, types of headaches, level of headache disability and comorbidities more common to this patient population. In turn, such information can provide insights that may be valuable to provider clinical decision-making and provider education.

Limitations

The design and findings of our review has a number of limitations. The design of the review was limited by a search within English language journals only. As a result, some research on this topic may have been missed. While the quality scoring system adopted for this review requires further validation, the data we collected was limited by the low to moderate quality of available papers which averaged 6.4 out of 10 points (Table 3). The low scoring was largely due to significant methodological issues and the small sample size associated with much of the collected papers. Much of the data on this topic was heterogeneous in nature (telephone, postal surveys and face-to-face interviews). There was a lack of validated practitioner and patient questionnaires to report findings, such as for questions on prevalence, where the time frames utilized varied between ‘currently’, ‘last 12 months’ and ‘ever’.

Data on the prevalence of MT use for headache was limited particularly within individual MT provider populations when compared to data found within the general population and headache-clinic populations. Many studies assessed the use of MT for headache without identifying headache types. Only one study inside an MT population had reported the percentage of patients attending for reasons of migraine alone (osteopathy). The prevalence of MT use for headache was reported most within chiropractic patient population studies, however information was limited on the types of headache. We found no studies reporting the prevalence of headache patients within physiotherapy or massage therapy patient populations using our search terms.

A lack of data for some themes necessitated providing findings pooled with users of other non-medical headache providers. Data within many geographical regions was very limited with the most limited data was on the source of referral to MT headache providers (three papers from Italy only). These limitations support the call for more research to be focused exclusively within MT populations and different regional areas before stronger conclusions can be drawn.