Healthcare utilization and resource distribution before and after interdisciplinary pain rehabilitation in primary care

2.1 Study design

This was a retrospective cohort study where patient clinical data from the Swedish Quality Registry for Pain Rehabilitation for primary care (SQRP-PC) (www.ucr.uu.se/en) and HCU data from regional administrative registries were combined. One hundred forty-six participants were evaluated with respect to their HCU 1 year before and 1 year after completion of an IPRP in PC in two Swedish regions. Distribution of resources and costs was described across the two measurement intervals from a healthcare-provider perspective using within-group analysis.

2.2 Data sources

Patient clinical data were collected via comprehensive questionnaires including patient-reported outcome measures (PROMs) concerning pain, physical and emotional functioning, coping, health-related quality of life, life and work satisfaction, etc. Questionnaires were filled in by the patients at baseline, immediately after IPRP and at 1-year follow-up. The standardized and validated instruments used in the questionnaires are described in detail in Table S1 (Supplementary material) [31–40]. The clinical data subsequently formed the start of the SQRP-PC at the end of 2015.

Individual data on HCU in terms of physical visits, telephone contacts, and administrative items were retrieved by linking personal identification numbers of the participants to data from medical registers of the region. Data were retrieved for two time intervals: 1 year before IPRP (between October 2011 and August 2014) and 1 year after the termination of IPRP (between January 2013 and November 2015). One-year intervals were chosen since register data were available for a 1-year follow-up. All outpatient healthcare contacts, irrespective of the reason for seeking care, were included. IPRP duration was approximated to be 3 months for all patients. HCU data were not collected during this time. HCU data were retrieved for the following professionals: GPs, specialists, PC nurses, specialist nurses, physiotherapists, occupational therapists, psychologists, and social workers. Contacts with dental and maternal care, as well as dietitian and vaccination services, were considered irrelevant and were therefore excluded. Visits to orthoptists, speech therapists, audiologists, assistant nurses, hearing care technicians, and orthopedic technologists were omitted because they were not compatible with both datasets of the regions. Data concerning visits to the emergency ward and hospitalizations were not available.

Unit costs for year 2022, based on the cost per patient (CPP) method, were retrieved (September 9, 2023) from the Swedish Association of Local Authorities and Regions (SALAR) in SEK. These unit costs were the latest available at the time. The SALAR works with gathering CPP data to unify the estimation of unit costs nationally and improve comparability between regions. Nevertheless, the manner of calculating in the regions currently varies. Therefore, the unit costs for the region of Västerbotten were used to make the cost description uniform. The retrieved unit costs entailed two professional categories – physicians and other caregivers – separated into primary and specialist care levels. The unit costs are estimates of the average cost for the professional category, independent of specialty. The unit cost template included personnel costs and education, medical materials, medicines, laboratory, x-ray and other medical services, premises costs, IT and telecommunication costs, general costs such as transport, and overhead costs. Telephone calls and administrative items were included in the retrieved unit cost.

The outcome measure of this study was the number of outpatient healthcare contacts.

2.3 Participants

The participants in this study completed IPRPs in PC, including 1-year follow-up, between August 29, 2012, and September 19, 2015. All participants who completed and returned questionnaires at both the baseline and 1-year follow-up were included in the study. The inclusion criteria were (1) disabling chronic pain for >3 months, (2) age between 18 and 65 years, (3) potential for an active life change, (4) no further medical assessments needed, (5) sufficient knowledge of the Swedish language, and (6) agreement not to participate in other parallel treatments during the program. Exclusion criteria were major somatic or psychiatric disease, a history of significant substance abuse, and/or state of acute crisis. Patients were consecutively referred to the IPRP by a GP, a physiotherapist, or an occupational therapist in PC.

2.4 Settings

IPRPs were conducted at 11 PC centers in two councils (today’s regions) across Sweden, one in the north (Västerbotten) and one in the south (Östergötland). The IPRPs were realized as part of the so-called Rehabilitation Guarantee, a national initiative to boost return to work and decrease and prevent sickness absence among patients with chronic musculoskeletal pain, by providing the PC center with a financial compensation of 25,000 SEK per patient who completed IPRP. The IPRPs were 6–10 weeks long and entailed 1–3 patient visits at the PC center on a weekly basis. One visit ranged from 1.5 to 3.5 h, and the total average duration was 27.8 h per program (minimum 14 h to maximum 49 h). The IPRP teams consistently involved at least one physiotherapist and one occupational therapist, but other professionals were also involved, including GPs, PC nurses, psychologists, and social workers, depending on the patient’s needs. At least one team member was trained in cognitive behavioral therapy. The IPRP consisted of group intervention or a combination of individual and group activities. Most of the team professionals had earlier experience with IPRP or received training in interdisciplinary treatment by professionals from specialist care beforehand. Detailed data regarding the contents and dosages of the IPRP were not available.

2.5 Cost description

A healthcare-provider perspective was used to evaluate the costs of HCU. The providers are defined as regional PC centers, and one private PC unit was also included. Based on the retrieved unit costs received from the SALAR, four different unit costs were outlined in the study: (1) specialist, (2) GP, (3) specialist nurse, and (4) other professionals (Table 1). One unit is equal to one visit. Costs from the two measurement intervals were compared in total and per caregiver category. All costs were accounted for in SEK and converted into EUR using an average exchange rate (EUR 1 = SEK 11.9171) provided by the Bank of Sweden (October 13, 2023). The time horizon of 1 year before and 1 year after terminated IPRP was estimated to be long enough to distinguish trends of change in HCU within the participant cohort.

2.6 Statistical analysis

Statistical analyses were conducted according to the per-protocol principle and carried out in IBM SPSS Statistics Version 28.0.1.0. HCU data were compared with paired t-tests, 1 year before and after IPRP, in SPSS. The paired t-test was also applied when comparing PROMs at baseline and at 1-year follow-up in SPSS. A confidence interval of 95% with a level of significance of p < 0.05 was applied. Effect sizes were calculated for differences in PROMs at baseline and 1-year follow-up using Cohen’s d.

3.1 Patient characteristics

In total, 146 patients were included in the study. In Figure 1, a flow diagram of the entire study population is displayed. There were no statistically significant differences at baseline between those who filled in questionnaires at follow-up and those who did not answer and were thereby excluded (65%). Descriptive data on the participants are presented in Table 2. Most of the participants were women (85%), and the mean age was 44 years. The mean pain duration was approximately 9 years (median 6 years).

Figure 1

Study flow diagram.

3.2 Health status

There were significant improvements in nearly all PROMs (p < 0.001) except satisfaction with vocation (p < 0.072), current pain intensity (p < 0.237), and pain intensity last week (p < 0.052), presented in Table 2.

3.3 HCU

HCU decreased from 4,710 to 3,961 contacts (Table 3), corresponding to a 16% (p = 0.013) decrease 1 year after IPRP compared to 1 year before. The proportions of visits versus administrative items and telephone contacts remained the same (59 vs 41%) before and after IPRP. The total number of visits decreased by 15% (p = 0.055) and administrative items/telephone contacts by 17% (p = 0.018) during the year after IPRP.

All healthcare contacts during the year before and the year after IPRP are presented per profession in Table S2 (Supplementary material). Visits to physiotherapists (34%) and GPs (22%) were the most frequent during the year before IPRP. These professions also had the greatest drop in visits, 31% (p = 0.048) versus 23% (p < 0.001).

Telephone contacts and administrative items decreased 61% for physiotherapists (p = 0.007) and 35% for PC nurses (p < 0.001), while they decreased 15% for GPs (p = 0.134). Contrary to the drop in visits to the physiotherapists and GPs, visits to PC nurses increased by 6% (p = 0.779), as did visits to occupational therapists (p = 0.818) and psychologists/social workers (p = 0.630) (5 vs 10%). In total, 60% of participants reduced the number of visits by at least one visit (n = 88), and 40% increased the number of visits by at least one visit (n = 52). Six patients had the same number of visits before and after.

3.4 Cost description

The distribution of costs is presented in Table 4. The cost of HCU in terms of physical visits for all participants was EUR 519,613 the year before IPRP and EUR 456,327 the year after, excluding costs for telephone contacts and administrative items. The mean annual HCU cost per patient with chronic pain amounted to EUR 3,561 (minimum EUR 233 to maximum EUR 9053) 1 year before and EUR 3,128 (minimum EUR 0 to maximum EUR 12,424) 1 year after IPRP. The total decrease in costs was 12% (EUR 63,286) or EUR 434 per participant. The most significant costs consisted of visits to GPs (26%), specialists (37%), and physiotherapists (17%), constituting 80 vs 75% of the costs before and after IPRP. Averted visits to GPs and physiotherapists contributed to the largest percentage (93%) of the cost savings.

4.1 Strengths and limitations

The strength of this study is its application of data from real-life settings and regional administrative registers. Previous HCU studies concerning chronic pain are most often based on self-reported data and are thus accompanied by the risk of recall bias. Further, HCU data are often limited to a few professionals, e.g., GPs and physiotherapists [16,24]. In our study, we included the most frequent professionals in the IPRP, which makes it possible to see its organizational impact.

This was a pragmatic clinical study, and achieving a control group was not feasible. The lack of a control group is a limitation of this study, which makes it impossible to definitely attribute the decline in HCU to the participation in the IPRP. Nevertheless, the average pain duration of the participants was 9 years, which makes it less likely that it was a natural decrease in HCU. The lack of a control group is a common challenge when evaluating real-world data [22,48]. The creation of an objection-free control group is almost impossible to attain ethically, economically, and practically [22,49].

Since the study was not initially planned to include an economic evaluation, there were limitations in the study consent and the extraction of HCU data. For example, telephone visits, which cost wise may be considered equivalent to a physical visit, were not distinguishable from reported telephone contacts, and possible telephone visits were therefore not included in the cost description. In addition, costs for medications, emergency ward visits, and hospitalizations were not available. Furthermore, due to the complexity of chronic pain conditions, this study is limited in not being able to account for comorbid illness explicitly related to chronic pain.

Unit costs were not discounted nor adjusted for inflation in the cost description, which is also a limitation. Since our aim was to compare HCU burden and distribution before and after IPRP, we focused on the percental changes over time. Another limitation is the relatively aged data used in this study. Nonetheless, chronic pain remains a prevailing societal problem, and the results would most likely be unchanged today. Finally, the results of this study should be regarded with caution due to the relatively small sample size and large drop-out, which limits their generalizability.