Chronic Low Back Pain (CLBP) is commonly defined as ‘persistent’ or ‘sustained’ low back pain lasting more than three months or twelve weeks [1-6]. The symptoms are diverse and include: impaired physical function such as limited range of motion: memory impairment: irritability: cognitive dysfunction (such as poor concentration); and psychological symptoms (such as anxiety and depression) [7].

There are several types of CLBP treatments, including: medication, physiotherapy, chiropractics and osteopathy, back surgery, complementary and alternative medicine, and multidisciplinary therapy [5,7]. These treatments provide varying levels of pain relief.

Non-adherence to treatments in CLBP contributes to increased healthcare expenditure and social costs. Increasingly, evidence suggests that improved patient satisfaction could increase adherence to advised treatment plans, potentially reducing such costs to society [5,8,9].

Thus, treatment satisfaction is an important factor to evaluate in routine clinical practice to help improve and maintain patient wellbeing. An evaluation of treatment satisfaction is likely to be of benefit in any CLBP survey or clinical trial.

There are many definitions of patient satisfaction. A widely accepted definition is that satisfaction or dissatisfaction is an attitude or opinion expressed by patients concerning their clinical experiences [5,10]. There is some discussion, however, over whether it is a unitary or multidimensional concept [10-12].

This review aims to:

i. Examine treatment satisfaction and dissatisfaction in patients with CLBP.

ii. Establish what definitions exist in the literature on treatment satisfaction and dissatisfaction.

iii. Identify factors associated with treatment satisfaction in patients with CLBP.

iv. Assess the adequacy of questionnaires that measure treatment satisfaction and dissatisfaction.

v. Evaluate the quality of treatment satisfaction studies, and ultimately to establish the level of treatment satisfaction in patients with CLBP.

Disclosure

This systematic review was adapted from a PhD conducted by Dr. Rofail at Brunel University.

Literature search

The University of York NHS Centre for Reviews and Dissemination guidelines were followed to conduct the systematic search [13]. Six electronic databases were used: (1) PubMed (index of medical literature), (2) PsychINFO (index of psychological literature), (3) Embase (indexes of journals in biomedical and pharmacological literature), (4) CINAHL (Cumulative Index to Nursing and Allied Health Literature), (5) Web of Knowledge (A key specialist database), and (6) PROQOLID (Patient- Reported Outcomes Quality of Life Instruments Database).

The search was implemented between January 1990 and September 2014 using the following keywords and commands: “chronic low back pain OR CLBP” AND “treatment” OR “therapy” OR “medication” OR “analgesic” OR “opioid” OR “physiotherapy” OR “physical therapy” AND “satisfaction” OR “dissatisfaction”.

The journals of Pain, Physiotherapy, and Spine were hand searched separately in 2014 to identify any publications that may not yet have been entered into databases.

Eligibility criteria

The inclusion criteria were developed to identify articles that measured treatment satisfaction or dissatisfaction in CLBP. To meet these criteria, articles needed to specify in the title or abstract that their studies included patients with CLBP, and present original data of patient satisfaction and / or dissatisfaction with treatments.

Studies were excluded if CLBP was not the primary focus; and where treatments were invasive methods such as surgery, injections, implantable drug delivery systems, or acupuncture. Other excluded studies were those that did not document treatment satisfaction scores, or focused on other types of satisfaction e.g. life satisfaction. Qualitative studies, reviews, posters, letters or commentaries, and non- English papers were also excluded. Only studies which met all of the pre-defined inclusion and none of the exclusion criteria were further explored.

Description of studies

Information was extracted about the aim, design and method, treatment or therapy, sample size, outcome measures, statistical analysis, and main findings including statistical significance.

Data quality and analysis

A qualitative review of studies was conducted with a focus on strengths and weaknesses of the methods used. The lack of data and high variation between studies meant that a meta-analysis of the data was not appropriate.

The quality of studies was initially assessed through a ‘hierarchy of evidence’ which classified studies based on the properties of their design. It is a grading of bias that increases gradually downwards. The studies were then assessed using a 10-item checklist modified for this review based on a systematic review of patient satisfaction with antipsychotic medications [5,14]. The ten items on the checklist were (1) whether or not studies had explicit a priori aims, (2) definitions of the size of the population under investigation, (3) sample size calculation, (4) justification that sample was representative of population, (5) specification of inclusion/exclusion criteria, (6) demographic details, (7) justification of reliable and valid satisfaction instruments, (8) an original questionnaire, (9) response/drop-out rates, and (10) discussion about the generalizability of results [5].

Selection of studies

A total of 789 articles emerged from the initial search. Of these, 705 were excluded e.g. search terms not in title or abstract (n=394), non- English abstracts (n=13), and duplicates from use of various electronic databases (n=298). Eighty-four abstracts met the inclusion criteria and were given a more detailed evaluation which then excluded a further 57 studies after evaluation of full text (e.g. search terms not in results), leaving 27 studies included in the review (Figure 1).

Figure 1: Flow diagram of study selection process.

Description of included articles

Table 1 presents the final list of reviewed articles, and summarises the design, participants, treatments/interventions and satisfaction measures used in the included studies.

Table 1: Characteristics of included CLBP studies.

The definition of CLBP was quite diverse in terms of duration of chronicity and included pain lasting ≥ three months or ≥ three months constant/intermittent LBP [15-17], and in some instances ≥ three months continual/recurrent LBP [18]. One study defined CLBP as having had pain for ≥ 6 weeks [18,19].

Impact on activities of daily living or functioning was a main criterion of inclusion for three studies [20-22]. In one study, it was specified that patients must be experiencing disabling symptoms [21], another stated that patients had to be on sick leave for six weeks or more [20], and another stipulated that patients have experienced CLBP and activity limitations every day for the last three months [22].

There were two studies that included other groups of patients such as neck pain [16], or acute pain [23]. These studies were included because the results of each group were separate from the CLBP group.

The treatments and interventions studied (Table 1) included medical/GP care, chiropractic care, osteopathy and physiotherapy and drug therapy.

Quality of studies

The quality of studies was varied. The checklist used indicated a range of 5 to 7 out of 10, with a mean number of points lost of 4.1 (Table 2), reflecting an overall low to fair quality of studies. The studies performed best for ‘explicit a priori aims’ (included in all studies), followed by ‘inclusion/exclusion of studies’ (included in 96% of studies). However, 22% of studies provided a sample size calculation, and only 11% of studies discussed the generalizability of results [5].

Table 2: Quality of included CLBP studies

Design

Longitudinal cohort designs (48%, n=13) comprised the majority of study designs [5,19,21,23-30]. Others were randomised controlled trials (33%, n=9) [15,17,18,31-36], before and after studies (11%, n=3) [37], and cross-sectional surveys (7%, n=2) [5,38].

Samples

All studies, with the exception of one, specified the definition or diagnostic criteria used for CLBP [28], but the criteria used to select CLBP populations varied.

Statistical analyses

Descriptive and inferential statistics were used in all studies. Descriptive statistics used across studies included measures of central tendency such as mean, median, mode as well as measures of dispersion such as range, minimum, maximum and standard deviation. Inferential statistics frequently reported were analysis of variance, chi squared tests [17], and regression analyses [5,18,19].

Definitions of patient satisfaction with treatments in CLBP

None of the 27 studies included in the review operationally defined patient satisfaction or dissatisfaction with treatments in CLBP. Frequently, no distinction was made between concepts of satisfaction and dissatisfaction, methods used to measure these concepts tended to imply equality between concepts, and there were no discussions around whether or not the concepts were considered to be one-dimensional or multidimensional.

Questionnaires used to measure patient satisfaction with treatment in CLBP

The majority of studies reported satisfaction using a self-reported measure. However, a telephone interview was used in three of the studies. Results indicated that there is no agreed standard for measuring treatment satisfaction or dissatisfaction in CLBP.

Seven percent of studies (n=2) used the Cherkin and McCornack Satisfaction Questionnaire [19,28]. The contents of this instrument relates to patients’ satisfaction with care. The Cherkin and McCornack questionnaire comprises of 11 items, scored on a 5 point Likert-Type scale from ‘strongly agree’ to ‘strongly disagree’. The items were developed based on clinical expert opinion rather than direct patient input, and the questionnaire is administered by telephone whereby the interviewer reads out loud 17 statements to patients. It is unknown if cognitive debriefing was performed to ensure that patients understand the questions. Further, whilst factor analysis revealed three main factors of the Cherkin and McCornack questionnaire (namely Information, Caring, and Effectiveness), the percentage of variance attributable to each factor was not reported. The measure has good internal consistency but other psychometric properties were not documented, such as test-retest reliability [5].

In another study, a modified version of a 27-item osteopathic and GP management satisfaction questionnaire was used [39,40]. The questionnaire was originally developed in patients with ‘chronic illness’. Further details regarding the types of chronic illness, the number of patients involved in the development of the instrument, and whether or not individuals with CLBP were included in the instrument’s development were not available.

Five studies used questionnaires that measured satisfaction with different features of treatment, such as aspects of the therapeutic relationship or a specific rehabilitation programme [5,18,21,27,29,41]. These questionnaires tended to be one-dimensional and appeared to lack specificity in relation to important aspects of treatment satisfaction (e.g. involvement in treatment decisions).

Satisfaction instruments not only varied with regards to number of items included but also with regard to the response options used. Notable examples included questionnaires based on Cherkin & McCornack’s satisfaction questionnaire using a 5-point Likert-type scale [19,28], as well as 15 studies that used different types of a single global satisfaction item [15-17,20,24-26,30,31,34-37,42,43]. Even global satisfaction items were very diverse as they often focused on different aspects of CLBP and its treatment. For example, one such rating focused on satisfaction related to pain relief [31], whilst another measured back satisfaction [5,43].

Further, the response scales also varied for global satisfaction assessments from 5-point Likert-type scales [16,20,24,42], to scales that ranged from 0-10 or 0-100 [23,25], where higher scores indicated a higher level of satisfaction. Some global satisfaction items were rated according to 4 categories: ‘very satisfied’, ‘satisfied’, ‘not satisfied and not disappointed’, and ‘unsatisfied’ [5,30].

Results of studies measuring treatment satisfaction and dissatisfaction in CLBP

It is difficult to make comparisons between the satisfaction results of the 27 studies because for many of the studies satisfaction was not their focus, and because study design and measurement of satisfaction varied so greatly. In spite of these difficulties, a number of key points emerged when satisfaction results were compared (Table 3). Firstly, a majority of studies reported positive satisfaction with treatment data. A second finding was that chiropractic care appeared to be favoured by patients over medical care [19,28].

Table 3: Patient satisfaction data in included CLBP studies.

Factors associated with patient satisfaction

Four studies explored satisfaction and its association to other concepts [5,22,26,28,40]. Factors related to satisfaction included: pain [26], disability [40], age, employment status, narcotic use [22], and doctor type. However, the results of these studies were mixed. For example, whilst Hazard et al. [26] found an association between satisfaction and pain/disability, Pincus et al. [40] found no association between satisfaction and symptoms including duration, pain intensity and disability.

One study also reported factors that did not seem to be associated with satisfaction [40]. These were: osteopath’s competence, quality of care, efficacy, and number of appointments, demographic characteristics (sex, work status, and ethnicity), and psychological factors (depression, anxiety and coping style).

This review aimed to: i) explore patient satisfaction and dissatisfaction with treatments for CLBP; ii) establish definitions of treatment satisfaction and dissatisfaction in CLBP; iii) provide an overview of questionnaires used to measure the concepts; iv) establish the quality of studies reviewed; v) determine the level of patient satisfaction or dissatisfaction with treatments in CLBP; and vi) identify factors associated with treatment satisfaction or dissatisfaction in CLBP.

The results revealed a handful of studies concerned with patient satisfaction with treatments for CLBP. None operationally defined the concept or tried to verify if the concept is one-dimensional or multidimensional. Therefore further research is warranted to establish what the concept of patient satisfaction or dissatisfaction with treatment is in CLBP and how it relates to other concepts such as adherence to treatment.

The majority of studies reported satisfaction using a self-reported measure. There were three studies that used telephone interviews performed by a physician. This may have resulted in potential bias incurred by the way that statements are read to patients, or a social desirability effect as a result of the patient’s desire to please the physician.

The Cherkin and McCornack questionnaire was used in 2 out of 27 studies (7%) [19,28]. Closer observation indicated that it was not developed to measure treatment satisfaction in CLBP. The development of the questionnaire was informed by clinical expert opinion and there was little evidence of patient input. Physicians and chiropractors are involved in daily management of patients with CLBP and are likely, as a result, to have a good appreciation of the difficulties faced by patients. However, developing a questionnaire without patient input may have led to missing key concepts/questions important to patients. Further, although a factor analysis had been performed on the instrument, there were limited details to determine the appropriateness of analyses and decisions made. With regards to psychometrics, the instrument demonstrated good internal consistency and concurrent validity. However, there was no reference made to discriminant validity, testretest reliability, or sensitivity to change over time. Given that a comprehensive and well developed instrument to measure treatment satisfaction in patients with CLBP could not be found in the literature, there is a need to develop and validate such an instrument.

Several studies in the review used global ratings of treatment satisfaction. In general, these require patients to average their assessment of for example, ‘satisfaction with therapy’ into one single rating at each time point. Given that various factors may be related to and influence patient satisfaction with treatment, the use of a global rating scale may be inappropriate. Further, it is questionable as to whether patients can reliably average all of the factors related to treatment satisfaction into a single item. Whilst global ratings are brief and easy to administer, the lack of consistency introduced when patients make their evaluations based potentially on different criteria may in the end affect reproducibility of scores. In addition, global ratings of satisfaction are subject to ceiling effects and can disguise or hide aspects of dissatisfaction [5,44,45]. Furthermore, global ratings are mainly thought to be less informative than disease-specific or treatment-specific questionnaires, which are multidimensional [45,46]. As a result, global rating scales are at times included in studies to aid interpretation.

This review included several types of studies, ranging from cohort studies to double blind prospective randomised controlled trials (RCTs). It is notable that none of the studies included in the review provided a rationale for the utility of the questionnaires used. The domain coverage and content of questionnaires were often briefly described, though most studies did not report the psychometric properties of the instruments. Consequently, it is questionable whether the studies were truly assessing treatment satisfaction in CLBP or a different concept. Furthermore, the mixed quality of studies results in difficulties making comparisons and generalisations. Whilst few studies explored patient satisfaction with treatments in CLBP, there were mainly positive attitudes towards treatment. However some caution is necessary, as their patients may have been susceptible to the Hawthorne effect, since data was usually collected under clinical conditions. To minimise such bias, future studies could be performed in a non-clinical setting, and patients interviewed by an independent researcher who is unaware of the study objectives.

Whilst some factors associated with treatment satisfaction and dissatisfaction were identified, there was some disagreement between findings in studies [5,26,40]. Consequently, additional research is needed to obtain greater certainty regarding factors associated with treatment satisfaction or dissatisfaction.

The results of this review indicate that, in general, studies do not distinguish between satisfaction and dissatisfaction. Interviews with patients with CLBP and input from clinicians and researchers is likely to provide greater understanding of these concepts. There is also a requirement to develop a valid and reliable measure of satisfaction with treatments in patients with CLBP. A process of patient interviews to elicit concepts important to individuals with CLBP should be carried out in the development of any future measure of patient satisfaction and dissatisfaction in CLBP. Through the use of this new measure, further research could be carried out to test hypotheses regarding factors associated with patient satisfaction and dissatisfaction.

Patient satisfaction and dissatisfaction with treatment in CLBP are important concepts that should be assessed in routine clinical practice, clinical trials and surveys. Current instruments are not based on an operational definition of what comprises patient satisfaction with treatment [5]. Although the Cherkin and McCornack Satisfaction Questionnaire has been used in two studies, these do not provide a rationale for why it was used, nor do they define patient satisfaction. Where satisfaction measures have been used in CLBP, psychometric properties have rarely been reported.

This review underlines a requirement to develop an instrument specific to patient satisfaction with treatments in CLBP that is based on a standard operational definition and a conceptual framework with good psychometric properties. Measuring patient satisfaction enables health professionals to target aspects of treatment that cause patients distress (such as the experience of side effects), and may also aid to improve and maintain the patient’s health. Moreover, evaluating patient satisfaction with medication could indicate if best practice is being met and provide a point of reference for clinical practice and future research. Finally, the ability to validly and reliably measure patient satisfaction and dissatisfaction with treatments in CLBP means analyses can be performed to explore how these concepts relate to concepts like healthrelated quality of life and adherence to treatment regimens.

The authors declare no conflict of interest. This systematic review was adapted from a PhD conducted by Dr. Rofail at Brunel University and partly funded by Mapi Values. Roche Products LTD did not provide any financial support.