To evaluate whether an early cognitive–behavioral treatment complementary to a rheumatologic care program, for patients with recent-onset temporary work disability caused by musculoskeletal disorders (MSDs) is effective.
To evaluate whether an early cognitive–behavioral treatment complementary to a rheumatologic care program, for patients with recent-onset temporary work disability caused by musculoskeletal disorders (MSDs) is effective.
Patients with an MSD-related temporary work disability episode from 3–8 weeks' duration who were in a rheumatologic care program were randomized into a control group (rheumatologic care program) or an intervention group (rheumatologic care program plus cognitive–behavioral treatment). Enrollment lasted 24 months and followup lasted 6–24 months. Efficacy variables included duration of temporary work disability episodes, total number of work days saved, relative efficacy, and relative rate to return to work. An economic evaluation was also performed.
One hundred eighty-one patients were included (66 control and 115 intervention patients), generating 222 episodes of MSD-related temporary work disability. Episodes tended to be shorter in the intervention group than in the control group (mean 98 versus 127 days; P = 0.053), with a relative efficacy of 22.9%. There were no differences in duration of the first episode between groups (mean 105 versus 110 days; P = 0.79), but relapse episodes were significantly shorter in the intervention group (mean 63 days versus 197 days; P = 0.0002). Costs were also lower in the intervention group. To save 1 day of temporary work disability, $13.50 had to be invested in the program. Each dollar invested generated a benefit of $4.08. The program had a net benefit of $172,607.
Early cognitive–behavioral treatment complementary to a rheumatologic care program is cost-effective, adds >20% efficacy to the rheumatologic care program, and reduces the duration of relapses.
Musculoskeletal disorders (MSDs) are prevalent, potentially disabling conditions (1) with enormous social costs (2–4). In industrialized societies, they are the main cause of permanent work disability and functional loss in adults (3–8) and the second greatest cause of short-term temporary work disability (9), with productivity losses of up to 1.3% of the US gross national product (10). Costs are divided as direct (health system costs) and indirect (loss of productivity). In Spain, 20% of all sick leave days are related to MSDs, generating a remarkable and increasing burden at the work place (11).
In 1998 we initiated a community-based program for recent-onset temporary work disability due to MSDs based upon specific clinical and therapeutic protocols that were carried out by rheumatologists trained in the basics of sick leave administrative tasks as well as on rehabilitation and occupational aspects of MSDs. Specific protocols were created for different diagnoses, including a 3-level clinical management system for attending to patient evolution. The program achieved a significant increase in the rate of patients returning to work (relative efficacy of 39%), a decrease of nearly 50% in permanent work disability proposals, and was also cost-effective (each dollar invested generated a benefit of $11) and increased patient satisfaction (12). However, the program was unable to achieve the same effectiveness in all patients due to the impact of several factors. We have demonstrated that sociodemographic, clinical, and work factors such as age, female sex, marriage, peripheral osteoarthritis and sciatica, self-employment, unemployment, being a manual worker, and having one's work position covered during sick leave worsen a patient's prognosis and disability, increasing the duration and number of relapse episodes in short-term disability due to MSDs (13, 14).
In terms of the biopsychosocial model, work disability is a complex health problem resulting from a combination of several factors, all of which interact and make the return to an optimal functional status difficult. Though psychological factors were not analyzed in our previous studies, it is known that they are implied in the transition from short- to long-term disability (15–17), and that they can be modified by means of different intervention strategies (18–20).
Several studies of multidisciplinary intervention (including psychological treatment) in active workers with pain and MSD disability have demonstrated the interventions' effectiveness in a patient's status (18–20), but mostly in long-term disability patients, and usually not including outcomes about patients returning to work. Several studies have shown that it is necessary to implement interventions early, before 12 weeks of temporary work disability have occurred (12, 21–23). We observed that what happens during the first 4 weeks of MSD-related temporary work disability can differentiate between patients with good or poor evolution, and that this could be the time period to introduce a complementary tool (12).
The purpose of this study was to evaluate whether an early cognitive–behavioral therapy program, complementary to our routine MSD-related temporary work disability program, had the possibility to modify the transition between acute and chronic disability in daily practice. We measured the duration of all MSD-related temporary work disability episodes, the duration of relapse episodes, and the cost efficacy of the program.
Of the 5.5 million people in Madrid, Spain, 98% receive universal health coverage. Care provision is organized into 11 health districts. The present study was performed in Health District 7, which has a total population of 522,742 inhabitants and an active working population of 179,155 people. It has 17 primary centers, including 2 specialized centers and a tertiary hospital. Our rheumatology service provides care to all inhabitants of Health District 7 at the specialized centers and hospital. All workers who require sick leave are given a temporary work disability initiation form and receive disability compensation payments.
A randomized, controlled, unblinded study was performed. The inclusion and randomization period of the study began in October 2004 and concluded in October 2006. Followups were ≥6 months. Patients who were continuing on sick leave at the end of the study were assigned an end date (April 16, 2007). The study protocol was approved by the Institutional Review Board of the Hospital Clinico San Carlos (Ethics Committee).
Since 1998 our health district has provided an early specialized care program to subjects with work disability related to MSDs (International Classification of Diseases, Ninth Revision coding system) (12).
The inclusion criteria for the present study were having MSD-related temporary work disability episodes lasting for ≥4 weeks and having been followed in our early-attention rheumatologic care program (12). Subjects with MSD-related temporary work disability episodes lasting for >8 weeks were excluded. All eligible patients were invited to participate in the study, and those who gave their verbal consent were randomly assigned to either the intervention group, which received a cognitive–behavioral treatment complementary to the rheumatologic care program, or to the control group, which continued with the rheumatologic care program.
We considered that a 15% reduction in the average duration of temporary work disability in the intervention group would be important from the socioeconomic point of view. To evaluate the difference in the average duration of temporary work disability between groups, we calculated that 120 patients were needed for the intervention group and that ≥120 patients were needed for the control group (power 0.8%, error 0.05%, by bilateral test). However, because temporary work disability duration did not follow a normal distribution and we performed the calculation on the basis of a nonparametric test, we increased the sample size to 10%. The size of the theoretical sample was therefore 264 patients for both groups combined.
Computer-generated randomization lists of numbers were produced (with an intervention:control ratio of 3:2 in order to maximize the number of patients treated and to utilize our resources) and allocated to the 2 specialized care centers by means of sealed envelopes. In every specialized care center, 2 rheumatologists attending the routine MSD-related temporary work disability program opened an envelope when patients gave their verbal informed consent. Patients maintained their group assignments in successive episodes of MSD-related temporary work disability during the followup.
The intervention group received a cognitive–behavioral treatment program in an individual form that was distributed weekly. The design of the program was based on the characteristic of disability instead of on the diagnosis, and this was adapted to the psychological and sociodemographic conditions of the patient. We conducted our program in accordance with the Beck cognitive therapy model (24) and the Activating Event, Belief, and Consequent Emotion model proposed by Ellis and MacLaren (25), and used verified techniques for musculoskeletal pain (26–28). The cognitive–behavioral treatment was administered by a trained psychologist (LL) with experience in musculoskeletal diseases and whose telephone was used for consultations or appointment requests.
The cognitive–behavioral treatment was provided weekly in 3 levels, according to the patients' evolution. All patients received the first level, consisting of 2 60-minute sessions. The first session included education on pain and ergonomics and training in abdominal breathing. Furthermore, patients received a booklet with the objectives and methodology of the cognitive–behavioral treatment with examples to practice. The booklet included a series of chapters about frequent problems in pain and disability conditions and some cognitive–behavioral strategies designed to help patients cope with them. In the second session, doubts and difficulties with the booklet in the first session were resolved. After 2 weeks, patients who continued on sick leave went to the second level, which was composed of 3 sessions. Patients who did not return to work after that started in the third level of the cognitive–behavioral treatment, which had an indeterminate number of sessions depending on the patient's evolution. In this level, a revision of the previous techniques was performed and patients were trained in coping skills for interpersonal and work issues.
The following variables were analyzed in this study: sociodemographic (age, sex, and job), clinical (cause of temporary work disability episode), temporary work disability duration (defined as days of sick leave per episode), and costs (direct and indirect). MSD-related temporary work disability episodes were defined by a beginning (the day the initiation form was issued by the rheumatologist) and an end (the day the end form was issued by the rheumatologist), irrespective of the date of the entry to the cognitive–behavioral treatment.
Efficacy was defined as the difference between groups in the number of days spent on sick leave per temporary work disability episode. We expressed the efficacy in terms of 1) the differences in mean and median duration of the temporary work disability episodes, 2) the total number of days saved in the intervention group, obtained by multiplying the total number of episodes in the intervention group by the days saved per patient, 3) the relative efficacy of the intervention expressed as the percentage of days of sick leave saved per process in the intervention group, and 4) the relative rate to return to work (see the statistical analysis below). We analyzed the efficacy of the treatments in all episodes and also in the relapse episodes alone.
Costs were classified as direct (health system costs) and indirect (losses of productivity). Cost efficacy was defined as the amount of money required to save 1 day of temporary work disability. The cost–benefit was defined as dollars invested divided by dollars saved. The net benefit was defined as dollars saved minus dollars invested.
Patient age, sex, job data, and cause and dates of temporary work disability episodes were recorded by the rheumatologist at the time of the randomization. Costs were obtained as follows. First, the receipt of medical and surgical specialized care related to MSDs was considered, including diagnostic tests, physiotherapy, and rehabilitation. This information was obtained for both groups from the Information System of Specialized Care. Second, the information on the patients' medications, joint aspirations, or joint injections was obtained from their clinical records. Costs related to diseases other than MSDs or complications during followup were not included in the study. Monetary values were assigned to each cost item on the basis of the Instituto Madrileño de Salud reference prices, the National Institute of Statistics, the Instituto Nacional de la Seguridad Social, and the National Pharmacotherapeutical Catalogue, and were converted into 2007 US dollars. To estimate the indirect costs, we multiplied the number of sick leave days by the average daily wage in Madrid in 2007 ($57.79). Intervention costs were $56,016 (for a part-time psychologist for 2.5 years and book-keeping costs including administrative staff, photocopies, and phone calls).
All patients who were randomly assigned to a study group received followup contacts. Differences between groups related to baseline characteristics were tested using the Mann-Whitney U test and contingency tables. The number of episodes of temporary work disability was assessed with the Mann-Whitney U test. The differences in duration of temporary work disability episodes were tested by t-test and survival techniques (the log rank test). Kaplan-Meier curves were set to account for duration of temporary work disability within patients, as in a Poisson model. Cox regression analyses were adjusted by variables that were unevenly distributed between groups at baseline and had an association with outcome. In all models, the dependent variable was the number of days spent off work, and the results were expressed as the hazard ratio (HR; or relative rate to return to work) in the intervention group compared with the control group (with a 95% confidence interval [95% CI]). Differences between groups in direct and indirect costs were estimated using the Student's t-test and contingency tables. A sensitivity analysis including all costs analyzed was performed to investigate the robustness of the model, and therefore to verify the validity of our results. In this analysis, the efficacy of the cognitive–behavioral intervention in terms of reduction of days of temporary work disability and the totality of costs was reduced and increased 15% in each group (in opposite directions, depending on whether the best- or worst-case scenario was used).
Analyses were performed with Stata software, version 9.0 (StataCorp, College Station, TX). For all comparisons, 2-tailed P values less than 0.05 were considered statistically significant.
A total of 181 patients were invited to participate (all accepted the invitation to be included in the study), were consecutively recruited, and generated 222 episodes of MSD-related temporary work disability during the study period. We found 66 patients allocated to the control group and 115 to the intervention group, generating 82 and 141 temporary work disability episodes, respectively. Approximately two-thirds of the patients (75.7% in the control group and 77.3% in the intervention group) had only 1 episode; the rest of them had ≥2 episodes (24.2% in the control group and 22.6% in the intervention group). On average, patients were randomly assigned to a study group 30 days after starting their first episode of temporary work disability.
The mean age of the patients was 45.5 years, and most were women (77.3%). In all patients, back pain, neck pain, and soft tissue problems were the most common causes of MSD-related temporary work disability episodes. Most patients (44.6%) were white-collar workers (including management, intellectual, and secretarial jobs), 37.8% were blue-collar workers (including all manual jobs), and 17.5% were service workers (Table 1).
|Control (n = 66)||Intervention (n = 115)||P†|
|Age, mean ± SD years||46 ± 10||44 ± 10||ns|
|Cause of TWD|
|Type of work|
The mean ± SD duration of MSD-related temporary work disability episodes was 108.64 ± 108.73 days, with a median of 67 days. Episodes tended to be shorter in the intervention group (mean 97.94 days versus 127.03 days in the control group; P = 0.053) (Table 2). The program achieved a relative efficacy of 22.9%, saving 29.09 days per episode, and saving a total of 4,101.69 days during the study period. The relative rate to return to work did not reach statistical significance between groups (HR 1.28, 95% CI 0.96–1.71).
|Patients with only 1 TWD episode||50||89|
|Patients with TWD relapse episodes||15||26|
|Total TWD episodes||81||141|
|Duration of TWD episodes, mean ± SD days||127 ± 116||97.9 ± 102|
|Duration of TWD episodes, median (25th, 75th percentiles) days||82 (46, 179)||64 (43, 107)||0.053|
|Total TWD days||10,343||13,810|
|Days of TWD per 100 patients||15,671||12,008|
|Days of TWD saved||4,102|
|Relative efficacy of the program, %||22.9|
|Relative rate to return to work, HR (95% CI)||1.28 (0.96–1.71)|
When we analyzed only the first episodes of temporary work disability, there were no differences between groups related to duration (mean 105.68 days in the intervention group versus 109.92 days in the control group; P = 0.79). Nevertheless, there were differences between the relapse episodes in each group. The program achieved the greatest effect related to the relapse episodes, which were significantly shorter in the intervention group (mean 63.69 days versus 197.62 days in the control group; P = 0.002), with a relative efficacy of 67.7% (Table 3). The relative rate to return to work in the relapse episodes was also significantly higher in the intervention group (HR 3.84, 95% CI 1.78–8.26; P = 0.001). However, there were no statistical differences between the groups in the relative rate to relapse (per year), and patients in both groups had a similar probability of relapse (HR 0.63, 95% CI 0.32–1.23; P = 0.17).
|Duration, median (25th, 75th percentiles)||68 (45, 147)||66 (44, 115)|
|Relative efficacy, %||3.8|
|Relative rate to return to work, HR (95% CI)||1.03 (0.74–1.40)|
|Duration, median (25th, 75th percentiles)||136 (81, 330)||48 (23, 86)||0.002|
|Relative efficacy, %||67.7|
|Relative rate to return to work, HR (95% CI)||0.001||3.84 (1.78–8.26)|
The survival curves for return to work also show the efficacy of the program. The effect of the program or the difference in duration of temporary work disability episodes between groups started at 10 weeks, with the period of major efficacy lasting until the seventh month (Figure 1). The effect of the program with regard to relapse episodes began early, around the fourth week, and the most efficacy was achieved between the second and the third month, but the effect continued throughout the year.
Direct and indirect costs were significantly lower in the intervention group (Table 4), saving $1,796 per patient. Important savings were also obtained in pharmaceuticals ($66 per patient) and rehabilitation ($157 per patient), and the highest savings were related to productivity loss ($1,546 per patient). The program achieved a net benefit of $172,607. Moreover, to save 1 day of temporary work disability, a maximum of $13.50 had to be invested in the program. In terms of cost–benefit, every dollar invested produced a savings of $4.08 at the end of the second year.
|Variable||Control group (n = 66)||Intervention group (n = 115)||Economic results|
|Mean cost per patient|
|Cost in all patients|
|Savings per patient‡|
In the worst-case scenario, the sensitivity analysis showed a cost efficacy of $15.70 invested to save 1 day of MSD-related temporary work disability, a cost–benefit of $3.01 returned per $1 invested, and a net benefit of $129,911 (Table 5). In the best scenario, it showed a cost efficacy of $11.60 invested to save 1 day of MSD-related temporary work disability, a cost–benefit of $5.52 returned per $1 invested, and a net benefit of $215,308.
|Variable||Worst-case scenario†||Base-case costs||Best-case scenario‡|
|Direct cost per patient||647||586||761||510||875||434|
|Indirect cost per patient||7,092||7,818||8,343||6,798||9,594||5,778|
|Duration of TWD, days||121||104||127||98||133||92|
The results of our study show that an early cognitive–behavioral treatment coupled with our routine MSD-related temporary work disability program, offered to patients with work disability due to MSD, is effective, achieving a faster return to work, shorter relapse episodes, and a very positive economic result. To our knowledge, this is the first time that an early cognitive–behavioral treatment has been analyzed with an economic evaluation in a nonselected population of workers with MSD-related temporary work disability episodes.
As expected, patients in the study were mostly middle-aged, and back pain was the most prevalent symptom (29, 30). Especially noteworthy is the high percentage of women. As we reported in our previous study (12), women and men enter the MSD-related temporary work disability program in equal proportions, but women's episodes were longer, and therefore more women were attended to in our program. The impact of female sex on pain and disability episodes, and especially as a prognostic factor for duration of sickness-related work absence, has been persistently shown in previous studies (14, 31–35). The longer duration of women's pain and disability probably indicates a complex process with a large set of associated factors.
Early cognitive–behavioral treatment coupled with a routine MSD-related temporary work disability program achieved a reduction in days of work disability. We would like to emphasize that the MSD-related temporary work disability program is already effective, obtaining savings of ∼40% in days of temporary work disability (12); therefore, the cognitive–behavioral treatment was offered to patients who were already being attended to in an effective way. The additional 20% reduction in days of temporary work disability that the cognitive–behavioral treatment added to the current program supports the implementation of this intervention in order to save additional sick leave days and economic resources.
As revealed by survival curves, the efficacy of the program was mainly achieved due to shortening the duration of relapse episodes. In contrast to medical treatments, the learning of the cognitive–behavior techniques, similar to most psychological therapies, requires more time to be effective, and this could be a possible reason for the lack of differences between the groups in the duration of the first episodes. Previous studies have found that patients with sick leave receiving a cognitive–behavioral treatment continue improving with time, but only when the patients are treated early (21). Other studies have demonstrated that the probability of relapse in active workers with musculoskeletal disease is very high (21, 36, 37), so reducing the duration of relapses is an important goal.
When we compared our program with other cognitive–behavioral treatment programs that have not produced significant improvement (22, 23), we found some differences and believe that the success of our therapy was based on the following aspects. First and foremost, patients in our study had direct and faster access to the psychologist during the initial episode and during every relapse. Second, our patients had the possibility of receiving personal support and/or support via telephone at all times, including after the end of the temporary work disability episode.
We thought that having support via telephone from the psychologist available when the patient needed it, in order to receive more advice or to ask for an appointment, would allow the patient to not feel helpless. Another reason for the efficacy we found might be our individualized application of the cognitive–behavioral treatment. Our program, adapted to the individual needs of the patients, differs from other, structurally rigid cognitive–behavioral treatment intervention programs.
Furthermore, the reduction in loss of productivity was accompanied by a decrease in use of the health system, thereby reducing health costs, especially in terms of pharmaceuticals and rehabilitation. As a consequence, direct and indirect costs were significantly lower in the intervention group. Data from the full economic evaluation are complex to compare with cognitive–behavioral treatments applied to MSD-related temporary work disability in other studies because other studies have differed in terms of intervention programs, outcome measures, and health and social security systems. However, the efficacy of our study method, demonstrated in terms of costs and supported in the sensitivity analysis, is clear in spite of these differences.
Our results demonstrate that a cognitive–behavioral treatment performed in the subacute moment in the disability process helps subjects manage pain and disability, and could contribute to avoiding the transition to long-term work disability in some cases. The importance of an early intervention is also supported by the fact that returning to work becomes unlikely if the time taken off of work exceeds 12 weeks (12, 38–40). In this sense, our intervention might represent an intermediate approach prior to the use of more intensive programs in rehabilitation clinics. We are in agreement with previous studies (41, 42) that a brief intervention may be all that is required for the majority of patients, but that others will need a more extensive intervention. Our aim was for a non-time based intervention using specific effective techniques, such as a stepped approach.
The present study has several limitations. First, our inability to reveal statistical differences in bivariate analyses might be related to the sample size, and to sizeable dispersion in the SDs. As a consequence of the positive results obtained in our temporary work disability program (12), after 2 years it was integrated into the daily routine of our rheumatology practice. As a result of this integration, there was a progressive reduction in the total number of attended patients, and therefore the total number of patients included in the study diminished. However, the sample size here was larger than in other previous studies. Another limitation is that the treatment was conducted by one psychologist, and although this avoids experimental contamination, it could reduce the generalization of the results to treatment conducted by other psychologists.
Several strengths are found in this study. First of all, it is necessary to consider that the study was undertaken in real-life conditions, taking into account all workers without exclusion for age or sex. Neither patients with specific pathologies (inflammatory diseases, lumbar hernias, or the like) nor those with work litigations were excluded from our sample. It is also notable that high-quality measures such as returning to work and direct and indirect costs were used as outcome variables.
In summary, our findings support that cognitive–behavioral treatment offered to patients with 4 weeks of work disability caused by MSDs results in an efficient complement for a rheumatologic care program, increasing the benefits both for the patient and for the economy of the health and social security system. Further complementary studies should be performed in order to verify our results in different settings and health systems, and to demonstrate the effectiveness of the multidisciplinary approach in working-age adults with musculoskeletal pain (43, 44).
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be submitted for publication. Dr. Abasolo had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. Leon, Jover, Candelas, Lajas, Vadillo, Blanco, Loza, Perez, Redondo, Abasolo.
Acquisition of data. Leon, Jover, Candelas, Lajas, Vadillo, Blanco, Loza, Perez, Redondo, Abasolo.
Analysis and interpretation of data. Leon, Jover, Candelas, Lajas, Vadillo, Blanco, Loza, Perez, Redondo, Abasolo.