To evaluate the effectiveness and harms of pain management programs that are based on the biopsychosocial model of care, particularly in the Medicare population.
Electronic databases (Ovid MEDLINE, PsycINFO, CINAHL, Cochrane Central Register of Controlled Trials, and Cochrane Database of Systematic Reviews) from 1989 to May 24, 2021; reference lists; and a Federal Register notice.
Given lack of consensus on terminology and program definition for pain management, we defined programs as integrated (based in and integrated with primary care) and comprehensive (referral based and separate from primary care) pain management programs (IPMPs and CPMPs). Using predefined criteria and dual review, we selected randomized controlled trials (RCTs) comparing IPMPs and CPMPs with usual care or waitlist, physical activity, pharmacologic therapy, and psychological therapy in patients with complex acute/subacute pain or chronic nonactive cancer pain. Patients needed to have access to medication support/review, psychological support, and physical function support in programs. Meta-analyses were conducted to improve estimate precision. We classified the magnitude of effects as small, moderate, or large based on predefined criteria. Strength of evidence (SOE) was assessed for the primary outcomes of pain, function, and change in opioid use.
We included 57 RCTs; 8 evaluated IPMPs and 49 evaluated CPMPs. Compared with usual care or waitlist, IPMPs were associated with small improvements in pain in the short and intermediate term (SOE: low) and in function in the short term (SOE: moderate), but there were no clear differences at other time points. CPMPs were associated with small improvements in pain immediately postintervention (SOE: moderate) but no differences in the short, intermediate, and long term (SOE: low); for function, improvements were moderate immediately postintervention and in the short term; there were no differences in the intermediate or long term (SOE: low at all time points). CPMPs were associated with small to moderate improvements in function and pain versus pharmacologic treatment alone at multiple time frames (SOE: moderate for function intermediate term; low for pain and function at all other times), and with small improvements in function but no improvements in pain in the short term when compared with physical activity alone (SOE: moderate). There were no differences between CPMPs and psychological therapy alone at any time (SOE: low). Serious harms were not reported, although evidence on harms was insufficient. The mean age was 57 years across IPMP RCTs and 45 years across CPMP RCTs. None of the trials specifically enrolled Medicare beneficiaries. Evidence on factors related to program structure, delivery, coordination, and components that may impact outcomes is sparse and there was substantial variability across studies on these factors.
IPMPs and CPMPs may provide small to moderate improvements in function and small improvements in pain in patients with chronic pain compared with usual care. Formal pain management programs have not been widely implemented in the United States for general populations or the Medicare population. To the extent that programs are tailored to patients’ needs, our findings are potentially applicable to the Medicare population. Programs that address a range of biopsychosocial aspects of pain, tailor components to patient need, and coordinate care may be of particular importance in this population.
