Introduction

Besides differences in access to health services and coverage of associated costs varying between countries because of legislation, organisation, tradition, and practices [1], differences in the offered healthcare exist. Healthcare is provided in two more-or-less separate domains: conventional medicine and complementary and alternative medicine (CAM), with varying legal and organisational specifics. Conventional medicine is delivered within an organisational and legal framework of publicly operated, financed, or regulated facilities by licensed healthcare professionals or private businesses run by licensed healthcare providers with particular responsibilities. CAM is provided within a more liberal marketplace with less public regulation and financial coverage, and the services defined as CAM vary between contexts and reflect national medical traditions, cultures, and healthcare policies [2]. According to WHO, “complementary or alternative medicine refers to a broad set of health care practices that are not part of that country’s own tradition or conventional medicine and are not fully integrated into the dominant health-care system” [3]. It follows that CAM is rooted outside modern science, but may serve as a complement to the conventional medicine and thus in some countries included in medical curricula [3]. As such, the definition proves ambiguous and inconsistent, and a healthcare practice may be defined as CAM in one country, while in other countries considered conventional and provided by licensed professionals. Also, a therapy may be conventional or CAM according to the provider’s legal status, irrespective of the nature of the therapy itself, or it may, due to research or health-policy decisions, change status from alternative to conventional medicine.

It follows that the use of CAM varies between nations and cultures. Within developed countries, approximately one-third of the general population use CAM within 12 months [2, 4, 5], with a two-fourfold higher proportion of CAM use among those with diagnosed health issues compared to healthy individuals [2]. Both the nature and severity of the health issues affect the CAM use, and in particular the presence of musculoskeletal disorders leads to more CAM use compared to most other diagnostic groups [2, 6, 7], whereas the importance of severity relates to the context [8]. Patients suffering from back pain (BP) frequently report intolerable pain intensity or chronic pain, which are reasons similarly known for using CAM [9, 10], and studies investigating the importance of sociodemographic and -economic factors report that women, middle-aged individuals, and those with higher socioeconomic status tend to use CAM more frequently [2, 9, 11].

The use of CAM is rising worldwide and the economic impact is increasing [3], and a timely question must be whether certain treatment needs are overlooked in the conventional healthcare relative to specific disorders.

The impact of diverse spinal disorders in terms of pain and physical impairments, which may constitute different reasons for using CAM besides the degree of CAM, is only sparsely investigated. Among patients with low back pain (LBP), Rondoni et al. [12] and Murthy et al. [9] found that a longer duration of pain was associated with a stronger tendency towards the use of CAM, while Sibbritt et al. [8] found that severe pain was associated with a lower tendency to use CAM in favor of the general practitioner. To the best of our knowledge, this is the first study to investigate the association between specific spinal diagnoses and the use of CAM.

The aims of this study were (1) to investigate CAM use among patients with BP referred to secondary sector care in Denmark, (2) to identify CAM-associated demographics, clinical findings, and ICD-10 diagnoses of spinal disorders, and (3) to identify the rationales for the use of CAM under the assumption that these may vary according to ICD-10 diagnosis.

Materials and methods

Study design

This exploratory single-centre cross-sectional study was based on data from the clinical registry MySpineData (MiRD) (described below), self-reported data on the use of CAM, and clinical data from the Electronic Medical Records (EMR). The reporting adheres to the Strengthening the Reporting of Observational Studies in Epidemiology guidelines [13].

Setting and participants

The study population comprised a consecutive series of patients with BP referred to a public, free-access outpatient medical spine department in Southern Denmark. All patients were referred by general medical practitioners, chiropractors, or other hospital departments and had persistent spinal pain or spine-related functional disability. Patients were 18 years or older. The period of inclusion was from June 2021 to September 2021.

Patients were considered ineligible for participation if they didn’t complete the entrance registry MiRD, didn’t speak/understand Danish, were unable to receive secure digital mail, or were diagnosed outside of the following diagnostic groups: degenerative lumbar/cervical stenosis, lumbar/cervical disc herniation, non-specific lumbar/cervical pain, or spinal pain following trauma/chronic, complex pain (Appendix A in ESM 1: ICD-10 codes contained within the four diagnostic groups).

Data sources and variables

When entering for clinical assessment, patients were invited to fill out the internet-based MiRD registry that collects data on symptoms, overall health, and daily functioning [14]. The following variables were retrieved from the MiRD registry to describe the study population: age, gender, intensity of the pain (11-point Numeric Rating Scale (NRS) [15], functional disability (Oswestry disability index score (ODI) [16] or neck disability index score (NDI-10) [17] for lumbar and cervical spine, respectively), Fear-Avoidance Beliefs Questionnaire (FABQ) item 4 [18], and health-related quality of life (Euro-QoL (EQ VAS) [19]. Furthermore, self-reported data on educational level, other pain-causing conditions, and use of pain medication were retrieved from MiRD. Marital status and ICD-10 diagnostic codes were retrieved from the EMR. Please see Appendix B in ESM 2 for a description of the initial data handling.

This study defined CAM as a treatment given outside the conventional Danish healthcare system by non-licensed healthcare providers. In Denmark, conventional providers relevant to BP are general practitioners, physiotherapists, and chiropractors, and a targeted questionnaire excluding these practitioners and focusing on the current BP was administered to gather information on the use of CAM within 12 months before the clinical assessment (Appendix C in ESM 3). The items and the response options were based on a national background population survey [20], and the type of CAM included addresses four categories of CAM: traditional Asian medical system, alternative medical systems, manual body-based therapies and mind–body therapies [2]. The access to conventional medicine and CAM is equal and without significant capacity challenges.

Considerations of sample size and missing data

The sample size was based on previously reported use of CAM in adult, Danish populations [20, 21], with an expectation of ≥ 30% of secondary sector patients with LBP having used CAM. With a desired precision of 0.05 and correction for anticipated non-response or missing data, the calculated sample size was n = 378.

Data management and statistical analysis

Statistical analysis

Presenting the study population, normally distributed variables were expressed as means and standard deviations, non-parametric data as medians and interquartile ranges, and categorical data as frequencies and proportions. The types, extent, and reasons for CAM use were expressed as frequencies and proportions. Univariate logistic regression analysis estimated the associations between the outcome event (CAM use) and the independent variables. All associations were expressed as Odds Ratios (ORs) with a 95% confidence interval (CI). Using stepwise backward elimination based on a Likelihood ratio test, we tested all variables in a multivariable regression analysis. Independent variables were tested for intercorrelation, and variables with moderate or strong correlation (R2 > 0.5) were excluded prior to the multivariable analysis. Frequencies and proportions of pre-printed response options for reasons for seeking CAM were reportedThe associations between reasons for seeking CAM and spinal diagnostic ICD-10 codes were examined with Chi-squared test.

Analyses were performed using Stata 16.0 (StataCorp, College Station, TX) with an alpha level of 0.05.

Ethics and data control

The Region of Southern Denmark legal department approved the study and use of data from MiRD and EMRs. Danish law exempts questionnaire studies from applications for ethical approval [22]. Upon completing the study-specific questionnaire attached to the MiRD questionnaire, the patients gave written consent to use their data. All data were merged and stored in accordance with the Danish Open Administration Act, the Danish Act on Processing of Personal Data, and the Health Act.

Results

Participants

Complete data were collected on 432 patients, corresponding to 76% of eligible patients within the inclusion period (n = 570) (Fig. 1).

Fig. 1
figure 1

Flowchart describing the selection of the study population

Full size image

Characteristics stratified by spinal region are presented in Table 1. In total, 65% were diagnosed with non-specific lumbar or cervical pain, 18% with degenerative stenosis, 14% with disc herniation, and 3% with BP following trauma or chronic, complex pain.

Table 1 Characteristics of the study population stratified by region (low back and neck) and by the use of complementary and alternative medicine/healthcare (CAM)
Full size table

Use of CAM and self-reported reasons for seeking CAM

In total, 23.8% (n = 103) of the study population reported use of CAM for their BP within 12 months prior to the clinical assessment. Patients with upper spine symptoms (± extremities) used CAM insignificantly more frequently than those with LBP (29% vs. 23%) (Table 1). The types, extent, and reasons for CAM use are presented in Table 2. Twenty-two participants (24%) reported using therapies not predefined in the questionnaire, however the majority of the participants used more types of CAM, and only 6 participants exclusively used one or more non-predefined CAM types. The majority reported two (25%) or three (22%) reasons for seeking CAM and we concluded that the prelisted response options were exhaustive, as only six patients stated other reasons for seeking CAM, which were thematically covered by the listed options (data not shown).

Table 2 The types, extent, and reasons for CAM use
Full size table

Patients with nonspecific ICD-10 diagnoses as per the a priori definition (Appendix A in ESM 1) were the most frequent users of CAM with 74%; 12.5% of CAM users were diagnosed with degenerative stenosis; 12.5% diagnosed with disc herniation, and 1% with spinal pain following trauma/chronic, complex pain. The Chi-squared test indicated no statistically significant association between the reasons for seeking CAM and the diagnoses grouped as described.

The association between clinical characteristics and seeking CAM

The proportion of missing data varied between variables (see Table 3), with a maximum of 20% for “Educational level”. Missing values were excluded from analyses on a case-by-case basis.

Table 3 Associations between patient characteristics and the probability of seeking CAM
Full size table

In univariate analyses, female patients were 1.64 times more likely to seek CAM (OR 1.64 (1.03–2.61)), and each additional year of age was associated with a 3% decrease probability for CAM use (OR 0.97 (0.96–0.98)) (Table 3). With multivariable analysis, gender and age remained significantly associated with the outcome with slightly varying ORs (Table 2). The ODI and NDI scores were not included in the multivariable analysis due to a strong intercorrelation with other variables EQ-VAS, use of pain killers, and pain intensity, and during the stepwise analysis the remaining variables were excluded.

Discussion

The percentage of patients with BP using CAM in this hospital setting (23.8%) is similar to that in the background population in the Region of Southern Denmark (23.5%) [21] and European background populations (25.9%) [2]. While earlier studies associate BP to a higher use of CAM [6], this is not the case in our study. Chiropractic and other manual therapies are often categorized as CAM and rank among the most frequently used treatments for BP [2, 9, 23], whereas in Denmark, provided by licensed health professionals, they are conventional practices covered by national healthcare legislation and with some public reimbursement, which most likely explain the results found in this study. A secondary effect might be that the coverage of such therapies or the free access to general practitioners and hospitals reduces the reduces the incentive to seek CAM not covered by public reimbursement [24]. Thirdly, the global prevalence of CAM used for BP ranges from 6 to 76% (mean 34%) [9], where regions with strong cultural influences and subsequently more varied healthcare offerings are characterized by high CAM prevalences [23]. Due to cultural homogeneity in Denmark, the importance of traditional and cultural influence was not investigated.

This study did not prove statistically significant associations between the diagnostics, the reasons for seeking CAM, and the use of CAM. However, in the univariate analysis, CAM use was significantly associated with non-specific lumbar or cervical pain. For these conditions, specific treatment is unavailable [24], leaving ample room for patients to explore CAM as part of an individual strategy to manage an often chronic and fluctuating condition.

The reported reasons for using CAM align with previous findings [2] and stress a perceived need for managing symptoms more than gaining basic healthcare. This is underlined by 67% of the CAM users partly using CAM to increase general well-being or as a preventative measure.

A surprising 20–30% of the patients used CAM due to disbelief in conventional healthcare or inaccessibility. We speculate if this is associated with the high prevalence of non-specific diagnoses and, as described above, a perceived lack of symptom-relieving treatment. However, no significant association was established. The only variables significantly associated with CAM use in the multivariable analysis were gender and age, which aligns with previous findings [2, 9, 23]. That is, the use of CAM seems more related to individual characteristics than to the underlying health problem.

Strengths and limitations

A strength of our study was the secondary sector study population, allowing us to investigate whether clinically important symptomatology impacted the use of CAM. Also, the specialised hospital setting should ensure a higher degree of uniformity in the use of diagnostics besides the collection of and accessibility to clinical data compared to settings such as private practice.

Methodologically, being a single-centre study, the validity and generalizability of the findings should not be overstated, and multi-centre studies are warranted. However, as touched upon in the Introduction, any scientific investigation of CAM will be particular to the context, affecting the generalizabiliy. Specifically, the Danish classification of chiropractic as conventional healthcare and the culturaFl and legal differences it reflects, limits the comparability of our results to those in other contexts. Future studies with similar research questions could benefit from examining associations with specific therapies regardless of their belonging to conventional medicine or CAM and address the dominant reason for CAM use to optimize the examination of associations. The CAM therapies presented in the questionnaire (Appendix C in ESM 3) were not exhaustive but were based on the most frequently used therapies in similar studies. Based on clinical experience, we would risk a mix-up between professions during answering by including more response options.

Finally, the participants reported CAM use within 12 months. This introduces the risk of recall bias regarding the type of CAM practitioners visited and the reason for consulting.

Conclusions

In this group of secondary sector patients with spinal pain, we found neither association between the use of CAM or reasons for CAM use and the ICD-10 diagnoses of spinal disorders, nor associations between the investigated clinical findings and CAM use. Independent variables significantly associated with more CAM use were the female gender and younger age. Results indicate that having non-specific BP, in particular, but not significantly, led to CAM use, possibly being individual management of often chronic and fluctuating symptoms. Clinicians should recognise this personal choice in a timely manner, but there is no evidence to support active recommendations for CAM.