Prevention of sickness absence through early identification and rehabilitation of at-risk patients with musculoskeletal disorders (PREVSAM): short term effects of a randomised controlled trial in primary care

Abstract

Purpose

To evaluate short-term effects of the PREVention of Sickness Absence for Musculoskeletal disorders (PREVSAM) model on sickness absence and patient-reported health outcomes.

Methods

Patients with musculoskeletal disorders were randomised to rehabilitation according to PREVSAM or treatment as usual (TAU) in primary care. Sickness absence and patient-reported health outcomes were evaluated after three months in 254 participants.

Results

The proportion of participants remaining in full- or part-time work were 86% in PREVSAM vs 78% in TAU (p = 0.097). The PREVSAM group had approximately four fewer sickness benefit days during three months from baseline (p range 0.078–0.126). No statistically significant difference was found in self-reported sickness absence days (PREVSAM 12.4 vs TAU 14.5; p = 0.634), nor were statistically significant differences between groups found in patient-reported health outcomes. Both groups showed significant improvements from baseline to three months, except for self-efficacy, and only the PREVSAM group showed significantly reduced depression symptoms.

Conclusions

The findings suggest that for sickness absence, the PREVSAM model may have an advantage over TAU, although the difference did not reach statistical significance at the p < 0.05 level, and similar positive effects on patient-reported health outcomes were found in both groups. Long-term effects must be evaluated before firm conclusions can be drawn.

IMPLICATIONS FOR REHABILITATION

• Early identification of at-risk patients and team-based rehabilitation within primary care to prevent sickness absence and long-term problems due to acute/subacute musculoskeletal disorders has been scarcely studied.

• The PREVSAM model provides a framework for team-based interventions in primary care rehabilitation.

• The PREVSAM model may be used in the management of acute/subacute musculoskeletal disorders in the prevention of sickness absence.

Keywords:

• Musculoskeletal pain
• physiotherapy
• occupational therapy
• interdisciplinary teamwork
• primary care
• secondary prevention
• sickness absence

Introduction

Musculoskeletal disorder (MSD) is a term frequently used to describe a broad range of conditions with musculoskeletal pain and decreased physical function that are very common in the general population. It was estimated in 2019 that over one billion adults globally have an MSD that would benefit from rehabilitation [1]. Even though acute episodes of MSD usually improve rapidly, recurrences are common and approximately 25% of adult patients’ primary care consultations are due to MSDs of any type [2].

First line treatment for MSD is suggested to be self-management advice, exercise therapy, and psychosocial interventions and MSDs are to a large extent managed in primary care [3,4]. Physiotherapy is suggested as first line treatment for MSD when self-management is not enough [5]. The aim of physiotherapy is: to promote health; prevent illness and injuries; and maintain, regain and maximise optimal movement potential, health-related quality of life, and participation in social life [6]. Physiotherapists as primary assessors for MSDs in primary care has at least as positive health effects as general practitioners [7]. Moreover, patients with MSD have reported high satisfaction with physiotherapy as a first line treatment [8].

Several consistent recommendations have been identified that could be used as guidelines in MSD care, for example a person-centred approach, a clinical examination, assessing psychosocial factors and work-related issues [9]. A biopsychosocial approach is recommended when explaining development of chronic pain [10] and work-related causes for MSD [11]. Work activities and work conditions can contribute both to the onset and progression of pain problems. One of the most common causes for work disability and sickness absence is MSDs, in Sweden representing about 17% of all sick leave in January 2023 [12]. The initial intensity of pain itself did not seem to be related to the likelihood for extended pain duration or for sickness absence in a group of patients with neck and low back pain [13]. Nor did pain itself correlate with work engagement, i.e., experienced work ability and occupational satisfaction, in women with MSD, while psychosocial and lifestyle factors significantly correlated with work engagement [14]. Comorbidity with anxiety, depression, and stress-related conditions is common when having chronic pain [15]. Therefore, the need for screening for risk factors and prevention of long-term disability is essential to reduce individual suffering and sickness absence. Early interventions may also reduce societal costs for medical care, reduced work ability and sickness benefits.

Obstacles towards why physiotherapists within primary care may not always discuss work-related factors with the patient have been identified. It may be due to the belief that assessing work-related issues is not the physiotherapist’s responsibility, perceived lack of competence, or organisational factors [16]. A lack of clarity about occupational health and a wish for guidelines for work-related issues have been reported among physiotherapists [17]. Moreover, even though patients find it important to address work issues with their physiotherapists, they may not raise the issue themselves [18]. Therefore, it is important that health professionals prioritise discussing this topic with patients. Clinical guidelines for optimising work participation suggest that patient management involve: a strong therapeutic alliance with therapist-patient agreement on goals and interventions; a physical performance examination; an individual treatment plan; screening of psychosocial risk factors and work demand; and, if needed, communication with the workplace [19]. Dropkin et al. recommend prompt treatment of work-related MSD using an extended team within primary care in cooperation with the patient’s employer [20]. Involving work-related factors is important as, in Sweden, employers have responsibility for the work environment regulated in law [21]. An early workplace dialogue, in addition to structural physiotherapy for MSD patients in primary care, has been shown to improve work ability statistically significantly compared to a control group [22]. Moreover, workplace interventions, service coordination to ensure communication among different stakeholders and interventions to improve the patients’ health conditions are recommended [23].

To facilitate identifying those at risk for long-term pain problems, screening tools have been developed to identify psychological factors which may increase the risk for sickness absence in those with MSD [24]. The prognostic screening tool Örebro Musculoskeletal Pain Screening Questionnaire in an original version (ÖMPSQ) [25] and a short form, ÖMPSQ-SF [26], are predominant in the literature. The instrument has been shown to be ‘excellent’ to identify risk for prolonged absenteeism due to low back pain [24]. Fear of pain or reinjury, beliefs in severity of health conditions, catastrophising, depression, poor problem solving, low return-to-work expectancies and lack of confidence in performing work-related activities are risk factors for long-term work disability. However, it is important to bear in mind that psychological risk factors covariate with other risk factors such as older age, physically demanding work, societal factors such as social isolation, and reimbursement systems [27].

Interdisciplinary treatment is recommended for rehabilitation of chronic pain, with teams consisting of at least three different health professions with a biopsychosocial understanding of the patient’s pain problem [28]. Physiotherapy and occupational therapy are common within rehabilitation for chronic pain in primary care together with psychological treatment [29]. The disciplines have a holistic view on the patient and contribute to the rehabilitation of MSD with different perspectives [6,30]. In occupational therapy, activities and occupations are core components with the aims to clarify the physical, emotional, cognitive and social demands of activities and occupations [30].

In summary, evidence suggests identifying vulnerable subgroups to initiate secondary prevention of pain with a biopsychosocial and person-centred approach before a transition towards persistent pain and a disability state may occur [9,27,31]. To meet these requirements, and hereby aiming to reduce individual suffering and societal costs, the rehabilitation model “Prevention of sickness absence through early identification and rehabilitation of at-risk patients with MSD” (PREVSAM) was designed [32]. The PREVSAM model’s purpose is to prevent sickness absence and development of persistent pain in a group of patients with MSD who have been identified to be at risk for sickness absence. Levels of psychosocial problems and the way they influence personal pain conditions need to be identified and assessed to provide targeted interventions to patients with MSD in primary care rehabilitation [33]. Furthermore, it must be clear when the patient can effectively be treated within primary care rehabilitation, and when more specialised psychological support is needed [34]. For those at risk, the PREVSAM model’s interdisciplinary teamwork intends to provide targeted interventions for the MSD and for psychosocial factors, including work-related factors, and hereby prevent development of long-term problems and sickness absence.

The study hypothesis was that, in a short-term perspective (three months), early identification of risk factors and rehabilitation according to the PREVSAM model would be more effective than treatment as usual (TAU) in prevention of sickness absence and in improving health-related outcomes in at-risk patients with MSDs.

The overall aim of this study was to evaluate the trial’s short-term effects (three months) of the PREVSAM model on sickness absence and patient-reported health outcomes. Specific aims were: (1) to evaluate the proportion of patients remaining at work, number of sickness benefit days, and self-reported sickness absence and (2) to evaluate patient-reported outcomes: psychological risk factors associated with sickness absence, perceived work ability, physical disability, pain status, symptoms of anxiety and depression, general and pain self-efficacy, and health-related quality of life.

Materials and methods

Study design

A two-armed randomised controlled trial evaluating the effects of the PREVSAM model compared with TAU. The trial was registered with ClinicalTrials.gov, Protocol ID: NCT03913325. This article presents three months’ follow-up results of sickness absence and patient-reported work ability and health outcomes. Findings are reported according to the CONSORT Checklist.

Setting

The trial was mainly conducted at public and private primary care rehabilitation clinics from May 2019 to June 2023 in Sweden’s second largest regional council Region Västra Götaland, which provides healthcare services to 1.7 million inhabitants. In the region, all existing public rehabilitation clinics and most private rehabilitation clinics within the reimbursement system “Care choice rehabilitation” were informed about the trial through managements team contacts and were invited to participate. The “Care choice” means that patients choose preferred rehabilitation clinic, and no referral is needed. The smaller Region Värmland was also invited which, like the Region Västra Götaland, has open access with equal conditions for public and private rehabilitation clinics.

Participants

Eligible participants were adults of working age, (aged ≥18 years, not fully retired) with MSD lasting less than three months. They were invited to participate when seeking care from a primary rehabilitation clinician for their MSD, if they scored 40 points or more on the ÖMPSQ-SF which indicate increased risk for sickness absence. Having a paid job or receiving parental benefits, unemployment benefits, or student aid were mandatory.

Exclusion criteria were “red flags” indicating risk of serious pathology, pain not primarily generated from the musculoskeletal system, severe mental illness, ongoing substance abuse disorders, pregnancy, being on sick leave due to the MSD for more than a month during the last year, fully retired or full disability pension, or inadequate Swedish literacy for the trial’s questionnaires.

Participants were told that the trial compared TAU with a more extensive team-based rehabilitation and were aware of their group allocation.

Procedure of randomisation

Participants were recruited and randomised at the rehabilitation clinic either at their first visit or by telephone a few days later by the responsible physiotherapist or occupational therapist. The participants were allocated by sealed opaque envelopes to either PREVSAM or TAU. Block-randomisation made by a computer programme in blocks of six (three of each) was used.

Treatment provided

Differences and similarities regarding treatment provided are presented in Table 1 for the PREVSAM model and TAU.

Table 1. The PREVSAM model’s components and differences towards treatment as usual within primary care rehabilitation.

PREVSAM model	Treatment as usual
Early identification of psychological risk factors
A person-centred approach with a biopsychosocial understanding including a medical history with a broad mapping of the patient’s situation	May include a person-centred approach. Medical history focusing on the MSD
Physiotherapy assessment	Physiotherapy assessment
Occupational therapy assessment	May occur in small numbers
A team meeting to establish a joint health plan with SMART goals	May include an individualised rehabilitation plan
Identification and clarification of the patient’s attitudes towards responsibility for the management of musculoskeletal pain and a clear division of the team members’ responsibilities
Possibility of early access to psychological treatment (optional)	May have psychological treatment elsewhere, however no formal collaboration with the rehabilitation clinic
Possibility to address work-related factors and contact with the workplace (optional)

The PREVSAM model

Both pragmatic and theoretical considerations guided the design of the PREVSAM model, for the primary care setting [32]. The PREVSAM model requires early identification of psychological risk factors and interdisciplinary teamwork with a person-centred approach (Table 1). The physiotherapist and the occupational therapist make their individual assessments and, based on these assessments and a biopsychosocial understanding of the patient’s situation, the health care professionals and the patient establish a joint individual health plan. Listening to the patients’ narrative, identifying important elements and their priorities, and discussing healthy lifestyle behaviours are emphasised. The health plan should include the patient’s goals and be inspired by the “SMART” approach, i.e., be specific, measurable/meaningful, achievable/activity-based, realistic/relevant and timed [35]. Moreover, clarifying of responsibilities and whether other contacts need to be made, such as with a psychotherapist or the workplace, are documented in the health plan. If needed, the patient’s general practitioner is contacted for medical assessment. The PREVSAM model provides a framework for team-based rehabilitation. Therefore, treatments to be provided are not specified. The rehabilitation should comprise evidence-based treatments commonly used in primary care rehabilitation.

Treatment as usual

In the control group, TAU, mainly physiotherapy treatment is provided within primary care, but occupational therapy may also be provided. The physiotherapist and/or occupational therapist responsible for the patient’s care use evidence-based interventions and provide sessions in line with their clinical judgement and available resources.

Primary outcome

The primary outcome was sickness absence, measured by collecting registry data on sickness benefits days from the Swedish Social Insurance Agency (SSIA) and by patient-reported data.

Sickness absence is defined as “absence from work that is attributed to sickness by the employee and accepted as such by the employer” [36]. In Sweden, the employee or the employer are liable for the costs of short-time sickness absence, the employee for one qualification day and the employer for the remaining days of the first two weeks. Periods longer than two weeks are compensated by the tax-funded social security system through sickness benefits from SSIA.

However, in Sweden, all residents with an income are insured by the Swedish sickness insurance system [37]. The income may come from salary, parental benefit, student aid, or unemployment benefit. This means that it is not mandatory to be employed to be eligible for sickness benefits. When sick and not being able to take care of children, study or actively seek a new job it is thus possible to apply for sickness benefits. A doctor’s certificate is mandatory to apply for sickness benefits.

Sickness benefit days data on the individual level (including diagnoses) were provided by the SSIA for sick-leave spells longer than 14 days. Data for number of gross and net sickness benefit days during the three months’ follow-up period were retrieved from the MicroData for Analysis of Social insurance (MiDAS) database. According to the funder’s request, we followed the outcome measures in the Social Insurance Report 2016:9 [38].

Responses to weekly sent text messages via mobile phone (SMS-Track ApS, Denmark, www.sms-track.com) were used to measure self-reported sickness absence. Validity and reliability of text messaging for research data collection have been established [39]. The participants were asked to answer the question “Have you been absent from work during the past week due to your musculoskeletal pain, and if so for how many days?” and to enter a number from 0 to 7. They were informed that the number of days was regardless of whether they were absent the whole day or part of the day.

Self-reported sickness absence via SMS-track and sickness benefit days were operationalised as number and proportion of individuals with and/or without sickness absence during the follow-up period.

Secondary outcomes

Secondary outcomes were patient-reported work ability and the health outcomes listed below. To collect the patient-reported outcome measures (PROMs), an e-mail with a unique link to a web-based survey (esMaker.net, ©Entergate, Halmstad, Sweden) was sent to the participant at inclusion (baseline), and 1, 3, 6 and 12 months after inclusion. The data collected at inclusion (baseline), and at three months after inclusion were used in this short-term follow-up.

Patient-reported risk associated with sickness absence

The instrument ÖMPSQ-SF was used primarily as a screening instrument. It was filled out at the first visit to either a physiotherapist or an occupational therapist, at the rehabilitation centre, for screening and this measure was used as baseline for those accepting to participate in the trial. It was included in the web-based survey’s follow-ups. ÖMPSQ-SF consists of ten items. The first item refers to duration of pain. The following 9 items each contain a statement or an assertion to be rated on an 11-point Likert scale from 0 = absence of impairment, to 10 = severe impairment, however three items are reversed [26]. The ÖMPSQ-SF was also used as a secondary outcome measure to evaluate changed risk associated with sickness absence, including proportion lowering their risk by scoring <40p at follow-up. Validity and reliability of the instrument have been established [25].

Patient-reported work ability

Work Ability Score (WAS) is a single-item question of the Work Ability Index (WAI), “current work ability compared with the lifetime best.” The participant is asked to rate his/her current work ability compared to his/her lifetime best on a 11-point Likert scale anchored by 0 = “Cannot work at all” and 10 = “My ability to work is at its best.” This question has been found to correspond well to WAI categories of the valid and reliable full instrument [40].

Patient-reported physical disability

The Disability Rating Index (DRI), is a self-assessment measure of experienced physical disability in patients with MSDs. It includes twelve everyday activities that most people perform or can imagine. The questions are arranged in order of increasing physical demands, where the participants mark on a 100-mm visual analogue scale (VAS) in accordance with her/his presumed ability to perform the daily physical activities in question. The anchor points are 0 = “without difficulty” and 100 = “not at all.” The total score (index) is the mean of all 12 items. The instrument has shown good reliability, validity and responsiveness [41].

Patient-reported pain status

Pain intensity during the previous seven days was measured using an 11-point numeric rating scale (PNRS) anchored by 0 = “no pain at all” and 10 = “worst imaginable pain.” To measure self-reported pain, the PNRS has been shown to have better responsiveness, ease of use, compliance rate and applicability compared to a visual analogue scale and a verbal rating scale [42]. Even though PNRS is recommended and widely used to measure pain intensity, inconsistent results for measurement properties are reported [43]. Pain frequency during the previous week was measured using a four-point Likert scale from 1 = “always/nearly always” to 4 = “rarely.”

Patient-reported pain self-efficacy

PSEQ-2 is a two-item short form of the Pain Self-Efficacy Scale (PSEQ), which is a valid, reliable and well-established measure of changes of beliefs held regarding the possibility to carry out certain activities despite pain [44]. From 0 = “not at all confident” to 6 = “completely confident” is rated on a seven-point Likert scale. PSEQ-2 is translated into Swedish, PSEQ-2SV and culturally adapted [45].

Patient-reported general self-efficacy

The General Self-Efficacy scale (GSE) assess the strength of an individual’s belief in his/her own ability to respond to difficult situations and to execute a course of action to a desired outcome [46]. It consists of 10 items rated on a four-point Likert scale from 1 = “not at all true” to 4 = “exactly true.” The Swedish translation of the GSE has proven to be a valid and reliable indicator of perceived general self-efficacy [47].

Patient-reported anxiety and depression symptoms

The HADS is a validated 14-item measure of self-reported symptoms of anxiety (7 items) and depression (7 items) [48,49]. Each item is scored from 0 to 3, where 0 = “indicating absence of symptoms or presence of positive features” and 3 = “maximal presentation of symptoms or the absence of positive features.”

Patient-reported health-related quality of life

The European Quality of Life, EQ VAS, is a validated measure to assess health-related quality of life [50,51]. EQ VAS is a vertical scale anchored between 0 = the “worst imaginable health” and 100 = the “best imaginable health.”

Baseline measurements

Data regarding age, sex, marital status, income and place of birth were retrieved from the web-based survey, while diagnoses were collected from the interdisciplinary team at each rehabilitation clinic.

Data analysis and statistical methods

Analysis was by intention to treat. Descriptive data are presented as mean (standard deviation), median (interquartile range) for continuous variables and number and percentage for categorical variables. Between-group comparisons were analysed using the tests Mann-Whitney U test, independent t-test and chi-square test, depending on data level. Within-group comparisons were calculated with the non-parametric Wilcoxon signed rank test. For all comparisons, differences where p < 0.05 were considered statistically significant. Data were analysed using IBM SPSS Statistics, version 28.0.1.1.

Results

Eight rehabilitation clinics participated in the trial. Four clinics were located in a larger city, one in a small town, and the remaining three in rural municipalities within commuting distance (about 30–60 km) from the big city. Half of the clinics were smaller with about 10 employees, while the others were bigger with about 20 employees or more. A total of 261 patients were recruited and randomised to either the PREVSAM intervention or TAU. The external dropouts consisted of 7 patients (Figure 1). From the remaining 254 participants, 129 in the PREVSAM group and 125 in TAU, data on background characteristics and treatment provided were collected, and short-term data from baseline to three months from SMS-track and PROMs were analysed. Four participants did not permit data to be retrieved from the MiDAS database. Results from the remaining 250 participants, 128 in the PREVSAM group and 122 in TAU, were analysed regarding sickness absence data from SSIA from baseline to three months. The response rate for the web-based survey was 98,5% at baseline and 82,9% at the three months follow-up. However, all PROMs were not always answered leading to additional missing data.

Figure 1. Flow chart of participants’ enrolment and allocation.

A flow chart showing the process from enrolment to final cases in the analyses.

Description of participants

There were no statistically significant differences between the PREVSAM group and TAU at baseline regarding characteristics (Tables 2 and 3) except for occupations. The classification of occupation was made according to the Swedish Standard Classification of Occupations 2012 (SSYK 2012) based on the participants’ description of the work they do [52]. More participants in the PREVSAM group described work classified as managers or requiring advanced level of higher education. On the other hand, in the TAU group, more participants described having potentially more physically demanding work.

Table 2. Baseline demographics regarding sex, place of birth, marital status/living situation, and age.

Characteristics baseline	PREVSAM group n = 129		TAU n = 125		Between groups analysis
					p Value (χ2-test)
Sex	n	%	n	%	0.924
Women	87	67.4	85	68
Men	42	32.6	40	32
Place of birth	Missing = 2		Missing = 5
Within or outside Sweden					0.582
Born in Sweden	112	88.2	103	85.8
Born outside Sweden	15	11.8	17	14.2
Living situation	Missing = 2		Missing = 4
Living alone or with others					0.625
Single	24	18.9	20	16.5
Cohabitating (with partner, children, friend and/or parents)	103	81.1	101	83.5
Age at start					(Independent t-test)
Mean, median (min–max)	47.8, 49 (20–67)		47.0, 49 (21–66)		0.669

Note: Variable numbers of participants are owing to internal missing data.

Table 3. Baseline demographics regarding education level, occupation, and income.

Characteristics baseline	PREVSAM group n =129 n    %		TAU n = 125 n    %		Between groups analysis
					p Value (χ2-test)
Highest completed education	Missing = 2		Missing = 5		0.791
Primary school or equivalent	11	8.7	7	5.8
Secondary school	57	44.9	56	46.7
College, university including higher academic degree	59	46.5	57	47.5
Occupation according to The Swedish Standard Classification of Occupations 2012 (SSYK 2012)	Missing = 2		Missing = 5		0.008
Managers	11	8.5	3	2.4
Occupations requiring advanced level of higher education	35	27.1	19	15.2
Occupations requiring higher education qualifications or equivalent	7	5.4	21	16.8
Administration and customer service clerks	15	11.6	11	8.8
Service, care, and shop sales workers	30	23.3	36	28.8
Agricultural, horticultural, forestry and fishery workers	1	0.8	0	0.0
Building and manufacturing workers	8	6.2	13	10.4
Mechanical manufacturing and transport workers, etc.	6	4.7	8	6.4
Elementary occupations	11	8.5	9	7.2
Students	3	2.3	0	0.0
Approximate household income before tax	Missing = 2		Missing = 6		0.125
0–14 999 Swedish crowns	6	4.7	1	0.8
15 000–29 999 Swedish crowns	14	11.0	22	18.5
30 000–44 999 Swedish crowns	33	26.0	31	26.1
≥45 000 Swedish crowns	74	58.3	65	54.6

Note: Variable numbers of participants are owing to internal missing data. Boldface indicates statistical significance at p < 0.05.

The team members reported the diagnoses according to ICD-10 they used in their medical records, and each patient had from one up to five diagnoses (Table 4). To have an MSD was mandatory and subsequently the most common was having an M diagnosis, i.e., “Diseases of the musculoskeletal system and connective tissue.” MSD can also be caused by injuries and classified with an S diagnosis; “Injuries, poisonings, and certain other consequences of external causes.” When pain is unspecified, the R diagnoses “Symptoms, signs of disease and abnormal clinical and laboratory findings not elsewhere classified” were used. B and G diagnoses, such as late effects of polio and myotonic diseases, were rare forms of MSD. Even though the inclusion criteria were to have psychological risk factors according to the ÖMPSQ-SF, the use of F diagnoses “Mental and behavioural disorders” was relatively low in both groups while the use of Z diagnoses “Factors of importance for the state of health and for contacts with health care” was more common in the intervention group compared to the control group. A few participants had H diagnoses as well because of vertigo and G diagnosis because of tension type headache.

Table 4. ICD-10 diagnoses reported by the team members in the PREVSAM group and in TAU.

ICD-10	PREVSAM group n = 129%	TAU n = 125%
A00-B99 Certain infectious diseases and parasitic diseases	0.0	0.8
F00-F99 Mental and behavioural disorders	3.1	3.2
G00-G99 Diseases of the nervous system	0.8	1.6
H60-H95 Diseases of the ear and the mastoid process	0.0	1.6
M00-M99 Diseases of the musculoskeletal system and connective tissue	87.6	84.8
R00-R99 Symptoms, signs of disease and abnormal clinical and laboratory findings not elsewhere classified	11.6	16.0
S00-T98 Injuries, poisonings, and certain other consequences of external causes	12.4	8.0
Z00-Z99 Factors of importance for the state of health and for contacts with health care	10.9	3.2

The diagnoses reported by the team members were not always equal to the diagnoses the physicians had used in the sickness certificate found in the MiDAS database. The differences may be due to another reason when seeking care or a change in condition as the diagnoses were set at different times and/or professional differences in which diagnoses are used. Sickness absence from the database were analysed in three rounds; first only sickness absence because of diagnoses related to the musculoskeletal system, i.e., M and S diagnoses was analysed. In the next round, F and R diagnoses were added. In the final step, all sickness absence was analysed regardless of the reason.

Primary outcome

Data on sickness benefit days from the SSIA were clearly skewed as most of the participants did not have any registered sick leave days for the three months following baseline. In the PREVSAM group, 86%–79% remained in full- or part-time work, compared with 78%–70% in TAU (at most 27 cases with sickness benefit days compared to 36 cases, respectively), not reaching statistically significant differences between the groups at p < 0.05. See Table 5 for proportions and confidence intervals. There were approximately four more gross and net sickness benefits days in TAU, but the difference between groups was not statistically significant (Table 6).

Table 5. Comparison of proportions of patients remaining in full- and part time work (no registered sickness benefit days) between the PREVSAM group and TAU.

	PREVSAM group, n = 128		TAU, n = 122
Time period	Proportion  95% CI		Proportion  95% CI		Proportion difference	p Value
Three months following baseline, M and S-diagnoses	0.859	0.788; 0.909	0.779	0.697; 0.843	0.080	0.097
Three months following baseline, M, S, F and R-diagnoses	0.828	0.753; 0.884	0.738	0.653; 0.808	0.090	0.082
Three months following baseline, all diagnoses	0.789	0.711; 0.851	0.705	0.619; 0.779	0.084	0.126

Notes: M “Diseases of the musculoskeletal system and connective tissue.” S “Injuries, poisonings, and certain other consequences of external causes.” F “Mental and behavioural disorders.” R “Symptoms, signs of disease and abnormal clinical and laboratory findings not elsewhere classified.”

Table 6. Comparisons of registered sickness benefit days between the PREVSAM group and TAU.

Time period	PREVSAM group, n = 128				TAU, n = 122
	Mean      SD     Median   Q1;Q3				Mean      SD     Median   Q1;Q3				p Value
Three months following baseline, M and S-diagnoses gross days	7.13	20.9	0	0;0	10.81	25.3	0	0;0	0.106
Three months following baseline, M and S-diagnoses net days	6.15	19.2	0	0;0	10.18	24.4	0	0;0	0.097
Three months following baseline, M, S, F and R-diagnoses gross days	7.69	21.2	0	0;0	11.79	25.6	0	0;4	0.087
Three months following baseline, M, S, F and R diagnoses net days	6.66	19.4	0	0;0	10.93	24.5	0	0;4	0.078
Three months following baseline, all diagnoses gross days	8.63	22.4	0	0;0	12.40	25.7	0	0;9.25	0.126
Three months following baseline, all diagnoses net days	7.60	20.9	0	0;0	11.55	24.7	0	0;9	0.112

Notes: Mann–Whitney U Test, SD: standard deviation, Q1;Q3: first and third quartile. M “Diseases of the musculoskeletal system and connective tissue.” S “Injuries, poisonings, and certain other consequences of external causes.” F “Mental and behavioural disorders.” R “Symptoms, signs of disease and abnormal clinical and laboratory findings not elsewhere classified.”

Self-reported sickness absence registered weekly by SMS-track was clearly skewed with approximately half the participants reporting no absence (Table 7). Slightly more patients in the PREVSAM group than in TAU reported sickness absence, but the mean number of self-reported sickness days was higher in the TAU group (Tables 7 and 8). None of the differences were statistically significant.

Table 7. Comparison of proportion of patients with self-reported sickness absence between the PREVSAM group and TAU.

	PREVSAM group, n = 128		TAU, n = 122
Time period	Proportion   95% CI		Proportion   95% CI		Proportion difference	p Value
Three months following baseline, self-reported sickness absence for inclusion condition	0.539	−0.060;0.185	0.475	−0.060;0.184	0.064	0.314

Table 8. Comparison of self-reported sickness absence days between the PREVSAM group and TAU.

Time period	PREVSAM group, n = 128				TAU, n = 122
	Mean ­    SD     Median   Q1;Q3				Mean ­    SD     Median   Q1;Q3				p Value
Three months following baseline, gross self-reported sickness absence days for inclusion condition	12.43	23.18	2.00	0.00;11.75	14.49	27.04	0.00	0.00;15.00	0.634

Note: Mann–Whitney U Test, SD: standard deviation; Q1;Q3: first and third quartile.

Weekly self-reported sickness absence is shown in Figure 2 for both groups. Most sickness absence was reported at the beginning of the study period. Missing values were common and slightly more frequent in TAU than in PREVSAM, with 51% and 48%, respectively, having one or more missing weekly reports (p = 0.201). Data from the Social Insurance Agency was imputed for missing values, when possible, for sickness absence over 14 days. Many of the missing values are likely zero based on surrounding values but this could not be confirmed from other sources.

Figure 2. Mean self-reported sick days per week for the PREVSAM and TAU groups.

Shows two lines representing PREVSAM respectively TAU regarding weekly mean of self-reported sick days per week.

Secondary outcomes

There were no statistically significant differences between groups in patient-reported work ability and health outcomes. Both groups showed statistically significant and clinically meaningful improvements in the within-group comparison for change over time from baseline to the three-month follow-up, except for general self-efficacy and depression symptoms (Table 9). Only the PREVSAM group showed significantly reduced depression symptoms. Both groups reported satisfactory levels of general self-efficacy already at baseline which remained over time.

Table 9. Differences in patient-reported health outcomes within and between groups.

	PREVSAM group		Within-group analysis	TAU		Within-group analysis	Between-groups analysis
	n = 129		Wilcoxon signed rank test	n = 125		Wilcoxon signed rank test	Mann–Whitney U test
Instrument	Baseline Median (Q1;Q3)	Three months follow-up Median (Q1;Q3)	p Value	Baseline Median (Q1;Q3)	Three months follow-up Median (Q1;Q3)	p Value	p Value
ÖMPSQ-SF	51 (45;58.5) n = 129	39 (29;49) n = 103	<0.001	52 (45;59) n = 121	42 (30;52) n = 86	<0.001	0.34
WAS	6 (4;8) n = 126	8 (6;9) n = 107	<0.001	6.5 (3;8) n = 120	7.5 (5;9) n = 86	<0.001	0.48
DRI	38.17 (20.75;53.75) n = 103	21.71 (4.92;40.5) n = 92	<0.001	40.17 (18.25;59.25) n = 98	25.38 (8.25;44.88) n = 78	<0.001	0.78
PNRS	5 (3;6) n = 127	3 (1;5) n = 107	<0.001	5 (3;7) n = 120	3 (1;4) n = 88	<0.001	0.33
Pain frequency	2 (1;3) n = 127	3 (2;4) n = 107	<0.001	2 (1;3) n = 120	3 (2;4) n = 88	<0.001	0.23
PSEQ-2SV	8 (6;10) n = 67	10 (8;12) n = 70	0.003	8 (6;10) n = 66	10 (8;12) n = 55	<0.001	0.34
GSE	31 (27;36) n = 67	31 (29;36) n = 68	0.24	31 (28;34) n = 66	32 (30;34) n = 55	0.20	0.94
HADS-A	7 (4;12) n = 123	6 (3;10) n = 104	<0.001	6 (4;10) n = 119	5 (3;9) n = 84	0.047	0.45
HADS-D	4 (2;8) n = 125	2.5 (1;7) n = 106	<0.001	3 (1;6) n = 120	2.5 (1;6) n = 86	0.36	0.76
EQ VAS	51.5 (40;70) n = 124	69 (51;80) n = 107	<0.001	56 (40;71) n = 120	70 (48.25;80) n = 84	0.006	0.69

Note: ÖMPSQ-SF: Örebro Musculoskeletal Pain Screening Questionnaire Short Form; WAS: Work Ability Score; DRI: Disability Rating Index; PNRS: Pain Numeric Rating Scale; Pain frequency: four-point Likert scale from 1 = “always/nearly always” to 4 = “rarely”; PSEQ-2: Pain Self-Efficacy Questionnaire two-items; GSE: General Self-Efficacy Scale; HADS: Hospital Anxiety Depression Scale; EQ VAS: The European Quality of Life Visual Analogue Scale. Boldface indicates statistical significance at p < 0.05.

Discussion

Most patients had no sickness absence during this short-term follow-up and significant improvements in patient-reported work ability and health outcomes were found in both groups. The results suggest that both team-based rehabilitation and TAU can be used for patients with acute/subacute MSDs who exhibit psychological risk factors.

According to this trial’s short-term results, a higher proportion in the PREVSAM group remained in full- or part-time work and had fewer registered sickness benefit days compared with TAU. The opposite was reported in a previous RCT, in which early multidisciplinary assessment was associated with increased sickness absence during the first three months [53]. However, that intervention did not include any treatment and was not described as person-centred teamwork. After assessment, those who were judged to be able to benefit from treatment were referred by the GP to standard healthcare resources. Contrary to the PREVSAM model’s person-centred approach with goal-targeting rehabilitation, the extensive assessments provided may, as discussed by the authors, have focused more on the patients’ symptoms and problems.

Both groups reduced their self-assessed risk associated with sickness absence. In the PREVSAM group the risk decreased by 24% and in the control group by 19%. In the PREVSAM group, the median decreased to below the trial’s cut-off value of 40p on ÖMPSQ-SF at the three-month follow-up. Screening as routine for MSDs may have disadvantages, such as being burdensome or leading to false results [34]. A score of 48 points on ÖMPSQ-SF is suggested as the optimal cut-off value to best balance between false positives (misclassifying as high risk) and false negatives (misclassifying as low risk) [54]. This trial’s cut-off value of 40 points may therefore have been too low. However, median score was 51 in the PREVSAM group and 52 in TAU. Screening for risk factors is suggested to reduce long-term disability in medium- and high-risk patients, and prevent over-treatment of low-risk patients but requires that the therapists have biopsychosocial knowledge and competency to handle risk factors [33]. Using a person-centred approach is suggested to be a cornerstone for all MSD management [31]. If the healthcare professional only checks if the total score is above cut-off for increased risk, important information may be overlooked. In rehabilitation according to the PREVSAM model, both what the patient reports works well and issues from the screening results should be considered, using a person-centred approach. This information is important when establishing a health plan. Moreover, this person-centred approach should ensure that the health plan is neither too demanding or comprehensive, nor insufficient.

Regarding perceived work ability, the median was increased by two points to 8 in the PREVSAM group and by one point to 7.5 in the TAU group at the three-month follow-up, indicating that both groups reached a moderate to good work ability according to accepted classifications of the instrument [40]. In prior research WAS as work ability measure has been suggested to measure other aspects of work ability than sickness absence, as patients may be working despite experiencing decreased work ability [55].

On the score of DRI, both groups reported improvement, the intervention group by 43.1% and the control group by 36.8%. Changes greater than 10% on DRI have been suggested as a minimal important difference at a group level [56]. Both groups reported a decrease by two points on the 11-point Likert scale for current pain, which is considered an important clinical improvement [57]. Similar improvement in PNRS after 13 weeks in the intervention group receiving cognitive functional therapy, was seen in Kent et al.’s study [58]. Contrary to our results, their control group receiving TAU did not report significant improvements. This may be due to differences in TAU and that their participants had chronic low back pain. In this group treatment mainly focuses on increasing function rather than decreasing pain. However, it is important to bear in mind that the pain experience may be influenced by other things than pain itself [43].

A desirable level of pain self-efficacy was reported in both groups already at baseline, according to the interpretation guide from Nicholas et al. [44], and was increased at follow-up. The reported level of general self-efficacy was high from baseline to follow-up in both groups. The self-reported high levels of self-efficacy already at baseline is a notable finding as prior research has suggested self-efficacy to be relevant for clinical outcomes in patients with MSDs in a primary care setting [59,60]. For patients with MSDs, having low pain levels and low pain self-efficacy at baseline, is associated with similar or worse outcome than having high baseline pain and high pain self-efficacy [60]. As increased psychological risk factors were an inclusion criterion, it is interesting that the participants’ self-reported anxiety and depression symptoms were not troublesome on a group level, however present in some individuals. The bidirectional relationships between symptoms of depression, anxiety and pain are highlighted in prior research. The same applies to treatment, and consequently adequate treatments of pain as well as depression and anxiety are needed [61]. Both groups reported rather low health-related quality of life on EQ VAS at baseline compared to a general Swedish population reporting 88.8 points [62]. Both groups reported a higher health-related quality of life at the three-month follow-up compared to baseline, an increase of 34% from baseline in the PREVSAM group and 25% in TAU at the three months’ follow-up; however, the median was still below the general population. This finding may reflect the complexity of our participants’ health problems as EQ VAS is related to health-related behaviours, diseases, and self-reported conditions [63].

Methodological considerations

Conducting research on complex interventions for participants with complex health issues within primary care is a major challenge. To our knowledge, few RCTs conducted in primary care have addressed sickness absence due to MSD and the interventions have seldom yielded significantly reduced days of sickness absence [53,64–66], with one exception [22].

The randomisation was successful except for demographic differences in occupation. Slightly more patients described having a physically demanding work in the TAU group. Obstacles to work when having MSDs may be due to demanding work tasks and to psychosocial factors [67]. On the other hand, those describing a more demanding intellectual job or personnel responsibilities in the PREVSAM group also have work-related risk factors for MSDs and psychosocial factors. Prior research reports that 20% to 60% of office workers suffer from MSDs related to upper limb or neck [68]. Moreover, being a manager may also include physically demanding work, for example when working in construction, warehouse, or restaurant industries. Furthermore, they may have had secondary pain problems, such as sleep difficulties. During the COVID-19 pandemic, quality of sleep has been reported sub-optimal in office workers [69]. Some reported working additional hours to compensate for the self-perceived submaximal work performance. However, stress is associated with MSD symptoms in both occupational groups [70].

Evaluating prevention of sickness absence involves a statistical challenge, i.e., how to measure something that does not occur. In these short-term analyses, we analysed proportions who remained in full- or part time work, and gross and net days for those who were absent from work. As the first two weeks rely on patient-reported data, the missing data complicated the analysis as we cannot know whether participants were working or not, if not reporting sickness absence. A related limitation was that the patient-reported sickness absence data only provided information on absence, not whether it was a whole day or part of the day. We analysed the absence, regardless of full or part time, as reduced work ability leading to sickness absence, in line with prior research [22]. Marmot et al. [71] concluded that spells of sickness absence longer than seven days, are more likely to be health-related than shorter spells. This indicates that analysing registered sickness benefits days is a more reliable measure; however, as the aim was to prevent all sickness absence, not only sickness benefit days, we decided to include self-reported sickness absence. The non-parametric rank sum test was used due to data skewness as relatively few per group had sickness absence, which in turn means that each individual’s value significantly affects the results.

The minimal important difference is suggested by Jaeschke et al. [72,p.408] to be “the smallest difference in score in the domain of interest which patients perceive as beneficial and which would mandate, in the absence of troublesome side effects and excessive cost, a change in the patient’s management.” Which treatment effects that patients with MSDs consider large enough to make it worthwhile to participate in treatments are discussed by Christiansen et al. [73] and suggested to be at least 20% of additional improvement on pain and disability compared with natural recovery. As both groups received treatments in our trial, it is not clear to what degree these improvements are due to treatments provided or are a result of natural recovery. We used non-parametric tests in the analysis of differences in the PROMs between baseline and follow-up assessments. Both groups reported statistically significant improvements from baseline to the three-month follow-up that we consider clinically meaningful improvements in most of the PROMs, discussed above.

Strengths of the current study are the randomised controlled design, the objective measurement of sickness benefits days, the use of established PROMs, and that it was conducted within clinical settings in primary care and thereby reflects clinical reality. However, weaknesses may exist because different circumstances in this real-life setting may have affected the treatment fidelity. The PREVSAM intervention required active involvement from both the professionals and the patients. To recognise each patient’s unique situation, mapping his/her resources and personal goals is suggested to empower patients to cope with MSDs [74]. Organisational factors, such as a high workload with time pressure and perceived lack of support from colleagues and managers, have been reported as barriers to teamwork [75,76] and may have affected the conditions at some clinics for a team-based, person-centred approach in the present trial.

Some other factors may have influenced the results of this trial. Firstly, it was mostly conducted during the COVID-19 pandemic and from 11 March 2020, until April 2022, the qualification day for sickness absence which was not normally reimbursed was temporarily reimbursed by the Swedish Social Insurance Agency. Moreover, the temporary regulations meant that a doctor’s certificate was not required during the first 14 sick days to receive sick pay, in contrast to the usual requirement of providing a doctor’s certificate from the eighth sick day [77]. Additionally, during six months in 2020 and nine months in 2021, a doctor’s certificate was not required until the 21st day of a sickness spell. While the temporary regulations for sick leave during the COVID-19 pandemic complicated the analysis, they should have affected both groups to the same extent and, therefore, should not have influenced the results. Moreover, the possibility to work from home may have reduced sickness absence, not only for infection symptoms but also for MSDs and minor mental illness. Due to the pandemic, there was a high level of absenteeism among both patients and staff. Scheduled visits had to be cancelled at short notice and could not always be postponed to a later date. Deviations from the PREVSAM model may have led to the care provided being too similar in the intervention and control groups to show significant differences between the groups.

Secondly, the treatment periods were longer in PREVSAM compared with TAU, which means that approximately 77% of the patients in PREVSAM had not yet completed their course of treatment within this short-term follow-up compared to 60% in TAU. As the PREVSAM intervention was team-based and thus more comprehensive compared with TAU, the longer treatment periods are not surprising but may have affected the short-term results, most likely to the disadvantage of the PREVSAM model.

Thirdly, patients agreeing to participate in a study may differ from the average patient in clinical reality. The target group of patients of working age with MSD, screening 40 points or above on the ÖMPSQ-SF is, according to our study’s result, a heterogenous group. The study participants differ to some degree from Swedish society in general. In Sweden, about 15% of the population were born outside the European Union [78], whereas in this sample only about 5% were born outside Europe. The participants’ education level and income level are higher than the average in Sweden and the participants are married or co-habiting to a greater extent [78,79]. The study participants in both groups being more resourceful than the average MSD patient within primary care rehabilitation may limit the transferability of the study’s result to other clinical contexts.

Fourthly, although no significant differences in sickness absence or patient-reported outcomes compared to TAU were found, some benefits may be present. In qualitative studies, the PREVSAM model was well received by both healthcare professionals and patients, appreciating teamwork [Ekhammar accepted, Ekhammar, in manuscript]. However, concerns were also raised. The screening was considered important to capture risk factors, but a person-centred medical history was considered necessary to identify the need for team-based rehabilitation in those at risk.

Finally, the response rate decreased over time, both regarding patient-reported sickness absence and patient-reported health outcomes. Internal dropout is a major concern in research. Although we believe our study’s response rates for the patient-reported outcomes stand up well in a research context [80,81], they may still have affected the results. Prior research suggests reasons for not answering text messages may be being less motivated when having a poorer prognosis or low interest in the research regardless of health status [75]. Therefore, it is unknown whether the dropouts have affected the results in favour of either of the compared groups.

Conclusions

Even though identified as being at increased risk for sickness absence, the majority of participants, in both groups, remained in full- or part-time work during and after their rehabilitation. For the primary outcome, sickness absence, the PREVSAM model may have an advantage over TAU, but the difference did not reach statistical significance at p < 0.05 level. The findings suggest similar, clinically meaningful improvements in patient-reported health outcomes with both rehabilitation according to the PREVSAM model and with TAU. Long-term effects need to be evaluated before firm conclusions can be drawn.

Future research

The addition of team-based interventions to TAU will be further explored in long-term follow-ups (six months and one year) as well as in qualitative studies of the healthcare professionals’ and the patients’ experiences. A process evaluation study will further explore whether the results may be due to the intervention itself or the way it has been implemented. A cost-effectiveness study will evaluate whether the PREVSAM model is cost-effective in relation to TAU, for example, whether the trend towards fewer sickness benefit days may compensate for the costs of increased number of visits.

Ethics approval

This study was approved by the Swedish Ethical Review Authority [Registration] No 2019-00832/1253-18, amendments 2020-00417, 2021-04275, 2022-06191-02, 2024-02233-02.

Consent form

The principles of the Helsinki declaration were followed, and written informed consent was obtained from the participants.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data are available on reasonable request.

Additional information

Funding

The study was funded by FORTE (grant no 2018-01250 and 2019-01732).