This protocol supplement contains the following items:

Details of the final protocol and recruitment outcomes are in the following publication: Krebs EE, Jensen AC, Nugent S, DeRonne B, Rutks I, Leverty D, Gravely A, Noorbaloochi S, Bair MJ, Kroenke K. Design, recruitment outcomes, and sample characteristics of the Strategies for Prescribing Analgesics Comparative Effectiveness (SPACE) trial. Contemporary Clinical Trials. 2017 Nov;62:130-139. PMID: 28893675 DOI 10.1016/j.cct.2017.09.003

Chronic musculoskeletal pain conditions are among the most common problems seen in primary care. As the importance of these conditions for the health of individuals and society has been increasingly recognized, use of long-term opioid therapy for chronic musculoskeletal pain has grown exponentially. Unfortunately, research has not kept pace with this change in opioid prescribing practice.

The balance of benefits and risks for patients receiving opioids for chronic pain is currently unclear. Available evidence for effects of opioid treatment on chronic pain is primarily from short-term placebocontrolled trials with limited applicability to clinical practice. Guidelines support the use of opioid therapy when other treatments have not improved pain intensity and function, but these recommendations are based primarily on expert opinion due to limitations in the quantity and quality of evidence.

The long-term medical safety of opioids is poorly described in the literature. Clinical trials of opioids have been designed to primarily assess short-term side effects, such as constipation. Preliminary investigations suggest that long-term opioid therapy may adversely affect pain sensitivity, endocrine function, fall risk, and sleep, but the incidence and severity of these potential effects have not been well-quantified. Additionally, despite widespread concerns about opioid misuse and iatrogenic addiction, the magnitude of these risks is currently unknown. Estimates of for misuse and for addiction among patients receiving opioids are based on research with important methodological limitations.

As a result of these uncertainties, the proper place of opioids in chronic pain is controversial. Some experts argue for a reduced role, pending evidence of opioids' long-term effectiveness and safety; whereas others argue for expanded use of strong opioids, such as morphine, as important weapons in the fight against chronic pain. Primary care clinicians are caught in a double-bind-concerned that they may cause more harm than benefit, but not wanting to under-treat pain.

We propose to fill a critical gap in the evidence by conducting a pragmatic randomized trial comparing two clinically relevant prescribing strategies-opioid-intensive versus opioid-avoidant-for chronic musculoskeletal pain. The Strategies for Prescribing Analgesics Comparative Effectiveness (SPACE) trial will evaluate multiple patient-centered pain and safely outcomes over 12 months. It will compare two prescribing strategies that are both flexible algorithms, but differ strongly in their emphasis; the opioid-intensive strategy will use strong opioids early in treatment, whereas the opioid-avoidant strategy will optimize non-opioid medications while delaying and minimizing opioid use. Participants will be Veterans seen in primary care for chronic back or lower-extremity arthritis pain who have moderate-severe pain refractory to daily analgesics.

Aim 1: To compare 12-month effects of opioid-intensive and opioid-avoidant prescribing strategies on pain-related function and pain intensity.

Hypothesis 1: The opioid-intensive strategy will improve pain-related function and pain intensity more than the opioid-avoidant strategy.

Aim 2: To compare harms of opioid-intensive and opioid-avoidant prescribing strategies over 12 months.

Hypothesis 2: The opioid-intensive strategy will cause more adverse medication-related symptoms, more clinically important adverse events, and greater decrements in physical and cognitive performance than the opioid-avoidant strategy.
Aim 3: To compare effects of opioid-intensive and opioid-avoidant prescribing strategies on secondary outcomes, including health-related quality of life, pain sensitivity, and aberrant drugrelated behaviors.
Aim 4: To elicit participants' perceptions of their response to the intervention and of the value of intervention components.

As noted previously, chronic pain management is often complex and time-consuming. Many Veterans with chronic pain are dissatisfied with pain management and many VA primary care providers feel unsupported in their efforts to treat pain. The SPACE trial will compare two different analgesic prescribing strategies, both delivered through an intervention model that is designed to address key barriers to effective pain treatment in primary care. The SPACE intervention design draws on the Three Component Model (TCM), which was initially developed and validated by Roudebush (Indianapolis) VA Co-investigator Dr. Kroenke and colleagues in a multisite depression implementation trial. The TCM was designed to provide a structured approach to depression interventions, addressing the main barriers to effective depression care while maintaining feasibility in community practices. The barriers that it was designed to address are equally applicable in pain care-primary care time pressures, competing demands of acute concerns and medical conditions, suboptimal access to specialty care, and lack of structured reassessment and action in response to failure of interventions that are undertaken. PI Dr. Erin Krebs, previously of the Roudebush (Indianapolis) VA and colleagues have successfully used the TCM as the basis for two trials focused on management of pain symptoms, INCPAD (an NCI-funded trial involving cancer patients with pain or depression) and SCOPE. By enhancing primary care and creating a bridge to specialty consultation, the TCM complements the VHA stepped care model of pain care described in the current VHA Pain Management Directive.
The TCM is compatible with the Chronic Care Model, but is focused on the relationship among three care providers collaborating through structured communication to optimize symptom care processes. The three components of the TCM are primary care, care management, and enhanced specialty support. Figure 1 shows the relationships among the three types of providers and the central role of the care manager. In SPACE, care management will be delivered by a clinical pharmacist care manager and specialty supervision will be provided by clinician investigators. The four key process steps addressed by the TCM are symptom detection, treatment initiation, monitoring of response and adverse effects, and adjustment of therapy when needed. Within the VA, the first step (symptom detection) has been widely implemented through the Pain as the Fifth Vital Sign Initiative. Participants in SPACE will have already had treatment initiated, but will not have experienced a meaningful response, so SPACE will focus on the monitoring and treatment adjustment processes, which will occur as a series of structured treatment trials. Use of measurement tools for interval reassessment is a key feature of the TCM that will be fully integrated into the SPACE interventions.

SPACE is a pragmatic randomized clinical trial designed to compare the effectiveness at 12 months of two clinically-relevant prescribing strategies delivered within a care management model for chronic musculoskeletal pain. Rather than comparing simple uniform treatments, this study will compare two flexible prescribing algorithms that can each be tailored to individual patient preferences and responses. Patients will be assigned to either 1) the opioid-avoidant arm, which emphasizes non-opioid medications from several drug classes, or 2 ) the opioid-intensive arm, which emphasizes early use of strong opioid analgesics. SPACE will randomize 276 Veterans with chronic musculoskeletal pain of at least moderate severity to either the opioid-avoidant ( ) or the opioid-intensive ( ) arm (see Figure 2). In both arms, treatments will be titrated, added, or rotated as needed to achieve improvement in pain. The intervention structure and data collection protocol will be the same for both arms; only the prescribing strategy assignment will differ between them.

Veterans will be eligible for participation if they have chronic back or lower extremity osteoarthritis pain with moderate-severe intensity and interference with function despite analgesic therapy. Inclusion criteria are further defined as follows:
(a) Chronic back or lower extremity osteoarthritis pain: For this study, we define chronic pain as pain that is present nearly every day for at least 6 months. A standard definition of chronic pain is not agreed upon, but most studies have used months as the required duration. Patients with a primary pain complaint of back pain or hip or knee arthritis pain are eligible. These two types of pain account for the majority of pain complaints and long-term opioid therapy. Patients with additional pain locations will not be excluded unless they have a primary location other than back, hip, or knee or meet criteria for fibromyalgia.
(b) Moderate-severe intensity and interference with function: Eligible patients will have both moderatesevere pain intensity nearly every day, defined using standard cutoffs as Brief Pain Inventory (BPI) average pain item (on a scale of 0-10), and interference with function, as defined by BPI interference score
(c) Despite analgesic therapy: To ensure that potential participants are appropriate for opioid therapy, we will require that their pain has persisted at a moderate-severe level despite analgesic use at least nearly every day for the past 6 months. Veterans already receiving current chronic opioid therapy will be excluded (see C2 below).

As a pragmatic clinical trial, this study will enroll a generalizable sample of primary care patients who would be considered eligible for either opioid or non-opioid analgesic therapy. We have therefore kept exclusion criteria to the minimum necessary to ensure both patient safety and internal validity. Veterans who meet any of the following exclusion criteria that may interfere with outcome assessment will be ineligible: a) schizophrenia, bipolar disorder, or other psychosis; b) moderately severe cognitive impairment, defined as errors on a brief cognitive screener; ) anticipated back, knee, or hip surgery within 12 months; and d) anticipated life expectancy of less than 12 months. We will also exclude patients receiving current chronic opioid therapy and those with absolute contraindications to either prescribing strategy.
For exclusion purposes, we define current chronic opioid therapy as any long-acting opioid (i.e., methadone, transdermal fentanyl, or any sustained-release opioid) dispensed in the past 3 months or any prescription for tablets/capsules of short-acting schedule 2 or 3 opioids dispensed in the past 3 months. Under this definition, patients who received a small number of short-acting opioids per month, but less than required to use them on a daily round-the-clock basis, would be eligible. Patients receiving current chronic opioid therapy will be excluded for two main reasons. First, whether to discontinue or rotate/escalate opioid therapy in patients without an adequate response to opioids is a different research question from those addressed by this study. Second, insufficient "washout" of preexisting opioid therapy could confound study results because chronic opioid therapy can induce physical dependence and because discontinuation of opioids may lead to changes in pain sensitivity of uncertain duration.

Finally, because participants should be considered eligible in clinical practice for either opioid or nonopioid analgesic therapy, patients with contraindications to all drugs in either arm will be excluded. In general, contraindications for specific medications include known allergy, previous serious adverse effect, or failure of a previous adequate trial of that medication. Specifically, in the opioid-avoidant arm, a patient who previously failed adequate trials of all three drug class steps would meet exclusion criteria; in the opioid-intensive arm, a patient who previously failed high-dose sustained-release opioid therapy would meet exclusion criteria.

To determine adequacy of prior medication trials, the pharmacist care manager will conduct a medication history as described in Section E1. For each drug used, patients will be queried about adverse effects, pain response, and dose/duration of use. We will consider a patient to have "failed" a drug if they did not tolerate it or did not respond to an adequate trial. Adequacy of a drug trial is defined by both dose and duration of use. Figure 3 includes a flow chart for this assessment and a preliminary
table of dose/duration specifications for key drugs; these specifications will be expanded and finalized by the study team during the start-up phase of the trial.

For opioid therapy, additional contraindications include those outlined by VA/DoD clinical practice guidelines, as follows: a) acute psychiatric instability, defined as current uncontrolled severe depression, severe PTSD, or suicidal ideation; b) substance use disorder not in remission or treatment; and c) documented or suspected diversion of controlled substances.

*Combinations of a "weak" opioid with acetaminophen (e.g., codeine/acetaminophen)

Primary care providers (PCPs) will be informed of the study and asked to provide written consent to approach their patients for participation. The Minneapolis VA has 47 full-time equivalent primary care physicians. We have an excellent track record of recruiting PCPs in this manner; in previous trials at the Roudebush VA, of PCPs have consented.

The primary enrollment method will be a mail and phone strategy as used in our trials conducted at the Roudebush VA. Potentially eligible patients of participating PCPs will be identified through local VistA searches, using the criteria in Table 1. Searches will be updated every month during the enrollment period. Potentially eligible Veterans will be mailed a recruitment letter describing the study and cosigned by study Principal Investigator Dr. Krebs and their PCP. Potential participants will then be contacted by phone within a week after receipt of the letter to determine interest in participating and to assess eligibility. If the Veteran is eligible and interested in participating, an appointment will be scheduled for the informed consent process. After participants provide written informed consent and authorization, the baseline interview will be conducted.

If enrollment falls below targets, a second method of enrollment will be in-clinic contact of potential subjects by cross-referencing of the potentially eligible list with the appointment schedule for each participating PCP.

This study will be performed at the Minneapolis VAMC, where we expect approximately of the potentially eligible population to be women. Eligible women and racial/ethnic minorities will be encouraged to enroll. This study will not be powered for subgroup analyses with women or minority participants, but we will examine results for evidence of trends that can be prospectively evaluated in a future multisite trial. Gender and race/ethnicity will be considered in all multivariable analyses. Additionally, the purposive sampling technique we will use for the qualitative evaluation will specifically "oversample" female and minority participants in order to adequately represent their views.

After providing informed consent and completing the baseline assessment, participants will be randomized to the opioid-intensive arm ( ) or the opioid-avoidant ( ) arm. Randomization will be stratified by primary pain location (back or hip/knee) to assure balanced numbers of participants with LBP or OA. Randomization will occur in randomly varying block sizes of 2 and 4.

Participants will not be masked to treatment arm assignment due to the complexity of the medication prescribing strategies. The pharmacist care manager and supervising clinical investigators who implement the interventions will also be aware of treatment assignment. To maintain allocation concealment, randomization will be conducted by the pharmacist care manager at the first treatment visit, using sealed envelopes prepared by the statistician. Outcome assessors will be masked to treatment assignment. The masking of outcome assessors and the structured nature of outcome measures are expected to minimize potential biased ascertainment.

A clinical pharmacist care manager will be the central intervention provider for both arms. Clinical pharmacists have an established role in Minneapolis VA primary care clinics that includes protocoldriven adjustment of medication therapy. Clinical pharmacists assigned to each primary care clinic can currently be consulted to see patients with difficult to control hypertension or hyperlipidemia. They adjust medications, monitor response, and return disease management back to the primary care provider once blood pressure or lipid medications are optimized. This program has been both wellreceived by primary care providers and associated with improvement in outcomes. The SPACE intervention protocol echoes this familiar and successful program.
Participants will see the pharmacist care manager in a study clinic, with the first intervention visit scheduled as soon as possible after study enrollment. At this visit, the care manager will obtain a detailed history of current and past pain medications. The medication history will include review of current and past medications, including examination of patients' medication packages they will bring to the visit and CPRS medication lists. Information from these sources will be supplemented by review of a checklist of analgesic and adjuvant medications, including photographs of drugs to stimulate patient
recall. The care manager will confirm eligibility as described in Sections C, randomize the participant as described in Section D, and recommend initial medication adjustment according to the prescribing strategy for the assigned treatment arm.
We expect that most adjustments in pain medications will occur in the beginning of the study period for most patients, so intervention visits will be scheduled at baseline and approximately monthly for the following two visits. Additional visits will be scheduled according to individual participants' needs during the remainder of the 12 month intervention period, with a maximum of 6 months between visits for participants who achieve adequate pain control on a stable medication regimen. Because the extent of care manager contact may differ among patients over the course of the trial, we will track these contacts and examine between-group differences in care manager contact as a potential confounder. Before and between intervention clinic visits, patients will have their symptoms reassessed using automated symptom monitoring (see Section F).
Case review meetings with the pharmacist care manager and supervising clinician investigators will be held each week to review progress with pain management, adverse effects, and any clinical problems. Clinical pharmacists are able to write prescriptions for non-scheduled medications. A physician investigator will sign paper prescriptions for DEA scheduled medications (i.e., opioids). All study medications will be dispensed through the Minneapolis VA research pharmacy and mailed to participants; from the participants' perspective, this dispensing process will not differ from the usual non-research medication dispensing experience.
The primary method of communicating with primary care about routine patient progress and pain treatment decisions will be care management notes entered into CPRS; this is consistent with usual consultant practice at our VAMC. If indicated by emergent adherence or safety issues (e.g., repeated attempts by a participant to receive pain medication outside of the study, important adverse effects), a care manager or investigator will communicate directly with the patient's primary care provider. At the end of the trial period, we will provide primary care providers with a synopsis of each patient's individual response to therapy, including change in pain and function and any adverse effects. A plan for continuation or discontinuation of study therapy will be developed for each patient on an individual basis. The clinical pharmacist will facilitate tapering of opioids at the end of the study if, based on individual patient outcomes, discontinuation of study opioids is determined to be the appropriate course.

Interval reassessment of pain, adverse effects, and adherence will be done by automated symptom monitoring, through interactive voice recorded (IVR) phone calls. Automated symptom monitoring will be scheduled to occur prior to in-person study clinic visits and at the time of medication refill or renewal. It can also be scheduled at interval time points as customized by the care manager, such as weeks after a medication change. The automated symptom monitoring assessment will include the following abbreviated measures: (a) the 3 -item PEG, a brief measure of pain intensity and interference that is responsive to change; (b) a patient global impression of change rating; and (c) the PHQ-4, which includes the depression screening items of the widely used PHQ-2 plus two anxiety screening items. Also included will be single questions about medication underuse, overuse, and side effects, as well as the option to send a request for contact by the care manager. Participants who do not complete their scheduled automated assessments will receive a reminder; if this fails, they will have the option of completing a paper version of the assessment at the time of their next visit.
Automated symptom monitoring by IVR will be used solely as a clinical tool for the pharmacist care manager and supervising clinician investigators to monitor and adjust therapy. It will not be used to assess research outcomes. The Minneapolis VA has an existing contract with Audiocare for clinical and research IVR services. Appropriate VA data security, information privacy, and research oversight bodies have approved IVR use in prior studies.

The analgesic prescribing strategies in this proposal were developed based on evidence-based recommendations from published disease and drug class-specific guidelines, systematic reviews, and our own published reviews and refined through our experience with algorithmbased pain medication adjustment in the clinical trials at the Roudebush VA. Each medication has demonstrated efficacy for relevant indications from randomized clinical trials and, in most cases, from meta-analyses or systematic reviews of multiple RCTs. Direct comparative data are not available for most medications, so sequencing is based on other factors, such as safety, tolerability, and VA formulary availability. As is the case with depression, regular monitoring and adjustment of pain therapy is likely as important in achieving good outcomes as the specific initial treatment chosen. Both prescribing strategies will be operationalized as a series of individual medication trials undertaken by the patient and pharmacist care manager using a structured shared decision-making approach.

Patients will enter the prescribing strategies at different levels, depending on their baseline medications and past medication history. Their baseline medications may be discontinued, adjusted to achieve an adequate therapeutic trial, or continued with the addition of adjunctive drugs, depending on their individual history of medication use (including dosing, scheduling, and adherence), therapeutic response, and adverse effects. In general, each level in each step of the prescribing strategies will be tried before moving to the next higher step. Importantly, both the opioid-intensive and opioid-avoidant prescribing strategies are titrated to clinical response rather than a specific type or dose of treatment. This pragmatic treat-to-target approach more closely mirrors real-world practice, which tailors evidencebased treatments to patient-specific outcomes.

Decisions to change pain medications must be made in a shared process because, unlike assessment of conditions that are evaluated independent of patient report (e.g., hypertension), pain control is inherently subjective. Decisions about whether to continue a current medication or change to another drug or dose are necessarily influenced by factors unique to the individual patient with pain, such as internal weighing of pain and medication effects on valued roles and activities, beliefs and values about pain and pain medications, and personal coping styles. SPACE will use a structured decision-making approach to consider patient preferences along with results of symptom monitoring. The two main considerations included in the algorithm for medication change decisions (Figure 4) are the following: 1) presence of pain response, defined as improvement in both PEG score and patient global impression of change, and 2 ) whether the patient desires change in medications.

Evidence for relative effectiveness of various opioid regimens, including different drugs and dosing schedules, is extremely limited; therefore, opioid initiation and adjustment will be guided by VA/DoD guideline-concordant principles, applied in a personalized manner. Figure 5 outlines principles of opioid adjustment and stepped medication options.
Opioid medications will be titrated gradually, with frequent reassessment, until treatment goals are met, dosing is limited by adverse effects, or a morphine equivalent (MEq) daily dose of 200 mg is reached. The issue of opioid dosing limits is highly controversial and no well-accepted upper dosing limit has been established; however, evidence for use of high-dose opioids is limited and reports suggest a positive relationship between dose and serious adverse events. Current VA/DoD guidelines recommend dosing based on patient need, not on predetermined maximal dose. In contrast, opioid prescribing guidelines from a Washington State interagency group recommend specialty consultation for doses day. Given these data and remaining uncertainties, we will not increase doses beyond day in this study. Participants reaching this limit may be rotated to another opioid, which can allow a decrease in the total daily dose.

Figure 5: Opioid-intensive Prescribing Strategy

Add on for breakthrough

As in the opioid-intensive arm, medication initiation and adjustment in the opioid-avoidant arm will be guided by general principles applied in a personalized treat-to-target manner. Figure 6 outlines the opioid-avoidant principles and stepped medication options. Adjuvant medications (designated Steps A and B) are included as analgesia-enhancing steps that may occur in parallel with prescribing of analgesics in Steps 1-3. Tramadol, included in Step 2, is a centrally acting analgesic that affects mu opioid receptors, as well as serotonin and norepinephrine reuptake. Tramadol is not a DEA controlled substance due to its limited abuse potential. Some patients in the opioid-avoidant arm will advance to low-dose therapy with typical opioids (Step 3), but the nature of the prescribing strategies will create a clear separation between the arms in the incidence and intensity of opioid use.

Figure 6: Opioid-avoidant Prescribing Strategy

If painful area is limited in size

If pain has neuropathic aspects

Treatment agreements are widely advocated for use among patients receiving opioid medications, although evidence for their effectiveness in preventing opioid misuse is limited. VA/DoD guidelines recommend reviewing a written Opioid Pain Care Agreement before initiating opioid therapy. In the proposed trial, a Pain Care Agreement will be employed in both intervention arms (regardless of medication type prescribed) as a way to structure medication education and clarify roles and responsibilities. At the time of study enrollment, the SPACE Pain Care Agreement will be entered into the patient's CPRS medical record and copies will be provided to the patient and his/her primary care provider. Violations of the agreement will be handled according to the nature of the behavior, as described in Section J.

At each intervention visit, use of non-study medications, over-the-counter medications, and herbal therapies or supplements will be evaluated, along with risk of drug interactions. Participants will be informed of potential interactions and study therapy will be adjusted as needed to avoid or minimize the likelihood of adverse drug interactions and maximize therapeutic response. A number of medication or class-specific safety rules will also be employed, including but not limited to the following: 1) Acetaminophen will be limited to day in liver disease or regular alcohol use ( drinks per day or drinks episodically) and day in other patients. 2) Patients considered for an NSAID must have had a creatinine within 6 months of initiation. Patients with estimated GFR , heart failure, peptic ulcer disease, or liver failure will not be started on an NSAID. Patients initiating an NSAID who are older than 65, have diabetes, or use a diuretic will have a follow-up creatinine within 4-8 weeks. Gastrointestinal prophylaxis will be prescribed in concordance with current guidelines. 3) Tramadol will not be used in patients with a history of seizures. 4) Patients with a history of heart disease, arrhythmia, or syncope will not be started on methadone. Patients considered for methadone must have had an electrocardiogram (ECG) within 3 months. If the baseline QTc interval is , methadone will not be started. Patients receiving methadone will receive a follow-up ECG 3-6 weeks
after starting therapy; if the follow-up QTc is and , the potential risks and benefits of therapy will be reevaluated with the patient and the ECG will be monitored at least every 2 months. If any follow-up QTc is , methadone will be discontinued or tapered.
The American Geriatrics Society (AGS) recently published updated guidelines for pain treatment in patients who are frail or years old. For older adults, as well as participants with higher risk for adverse effects (e.g., those with chronic medical conditions), we will follow safety recommendations outlined in the AGS guidelines as well as the classic geriatrics maxim to "start low, go slow." Examples of specific safety rules that will be employed include, but are not limited to, the following: 1) The lowest possible dose of a given drug will be used as the starting dose; 2 ) Nortriptyline will be titrated to a maximum dose of 50 mg ; amitriptyline will not be used; 3) Tramadol will be titrated to a maximum dose of 300 mg ; 4) An immediate-release opioid will be used prior to any long-acting opioid to determine the appropriate starting dose; 5) NSAIDs with high cyclooxygenase-2 inhibitor selectivity (e.g., diclofenac) will not be used. All participants will receive closer monitoring than is available in usual practice, with reassessment of pain and adverse effects after each medication change, and dose adjustments no more often than once a week.

Reports of side effects from automated symptom monitoring (Section F) will be followed up by the pharmacist care manager. In response to patient reports of adverse effects (AE), medications may be adjusted or clinical strategies may be employed to address ameliorable or preventable minor AEs (e.g., itching). Recommendations for prevention of constipation will be provided to all participants in both study arms at the outset. Important AEs (e.g., rectal bleeding) will also be evaluated by a study physician and reported to the patient's primary care provider. Potentially urgent AEs (e.g., chest pain) will be referred for immediate evaluation in the medicine clinic or emergency department.
Patients with mood symptoms will be further evaluated with the PHQ-9 for depression or GAD-7 for anxiety. Clinically significant new or worsening psychiatric symptoms will be reported to the primary care provider and patients will be offered referral to their primary care team psychologist. Veterans who endorse thoughts of self-harm on item 9 of the PHQ-9 depression scale will be assessed with a suicide risk algorithm.

Problems with medication overuse or underuse have many potential causes, including those related to the therapy (e.g., inadequate efficacy, side effects) and those related to patient (e.g., concerns about medication). When medication adherence issues are identified, they will be evaluated and addressed by the study care team on an individual basis with the goals of improving pain and tolerability of pain medications.
Aberrant drug-related behaviors (ADRB) are a particular class of nonadherence behaviors that have been associated with opioid misuse and substance use disorders. Although ADRB are often discussed in relation to opioid analgesic therapy, they may occur in any patient with pain. Indeed, ADRB may represent "pseudo-addiction" in the context of pain under-treatment, which may be more likely to occur among patients receiving only non-opioids or low-dose opioids. For these reasons, the clinical response to ADRB will be determined by the nature of the observed behaviors, rather than the medication prescribed or the study arm assigned.
A validated checklist of potentially concerning ADRB will be completed by the care manager at each visit for participants in both arms. Any emerging pattern of ADRB will be reviewed at weekly case meetings and the study team will prescribe an individual monitoring plan to aid in diagnosing and ameliorating the behavior. This plan will be reassessed at each visit and may include any combination of the following actions: a) medication adjustment; b) more frequent study visits; c) adherence counseling or education; d) review of state prescription monitoring database to identify duplicate or unreported external opioid prescriptions; e) referral for evaluation by a mental health or addiction
clinician; f) shorter prescription renewal intervals; g) urine drug testing; and h) pill counts. Serious ADRB, such as illicit drug use or potential diversion of prescribed medication, may result in medication changes, including tapering or discontinuation of opioids. These behaviors will not be considered reasons for withdrawal from the trial and all participants will be evaluated in their assigned arm.

Study measures include both patient-reported outcomes that will be collected in interviews at baseline and , and 12 months and test-based outcomes that will be completed in person at baseline, 6 , and 12 months. Interviews will require approximately 50 minutes at 0,6 , and 12 months and approximately 15 minutes at 3 and 9 months. In-person outcome assessment visits will require 50-60 minutes. A small incentive ( ) will be provided for each outcome assessment call and visit. Incentives of this magnitude offset costs of participation and are in the standard range used in our previous VA and non-VA effectiveness trials. No incentives will be provided for study clinical visits nor for automated symptom monitoring, which is conducted for clinical intervention purposes.
Patient-reported outcome measures and timing of administration are displayed in Table 2 and described in Section K1. Although participants will be asked to complete a substantial battery of patientreported measures in this proposed study, we believe this outcome assessment protocol is both appropriately comprehensive and reasonable in terms of respondent burden. Outcome assessment protocols of similar length have been well tolerated in multiple symptom management trials involving thousands of patients with depression or pain. Indeed, many participants in these trials have reported that they appreciated the sustained attention to their symptoms and perceived benefit from therapeutic disclosure.

QOL= quality of life; PTSD = Posttraumatic Stress Disorder
*Items completed for repeat assessment are in parentheses when they differ from those completed at baseline.

The Initiative on Methods, Measurement, and Pain Assessment in Clinical Trials (IMMPACT) guidelines recommend assessment of multiple core outcome domains in pain clinical trials, including pain intensity, function, patient ratings of improvement, and symptoms and adverse events. We will assess each of these pain outcome domains with validated patient-reported measures.

At 0,6 , and 12 months, participants will undergo an in-person assessment of domains that cannot be adequately captured with self-reported measures. A research assistant will administer tests of pain sensitivity, physical performance, and cognitive function.

Painful conditions are associated with risk of falling among both middle-aged and older patients; this risk may be exacerbated by medications with psychomotor effects. We will test 3 physical performance factors that may be especially important in determining risk for falls-balance, lower extremity strength, and gait speed. The Fullerton Advanced Balance scale is a 10 -item test that evaluates multiple dimensions of balance in independent adults. It was designed to be more challenging than balance measures developed for frail elderly persons and has been validated as a predictor of falls risk among middle-aged and older adults with chronic pain. Participants will also complete the Chair Stand test, a brief assessment of lower extremity strength. The Chair Stand measure assessed healthy adults' (age 20-85) time to stand from a seated position ten times, but subsequent studies found substantial floor effects among older adults and arthritis patients, who were often unable to complete ten stands. In our population of patients with functional limitations due to back and lower-extremity arthritis pain, we will use a modified version that counts the number of stands completed in 30 seconds. Gait speed tests are commonly used as performance measures in osteoarthritis. We will use a timed 8 -meter walk protocol that has been shown to reliably assess normal and fast-paced walking speeds among moderately-impaired osteoarthritis patients.
Cognitive function will be assessed with a brief battery of tests, the Indiana University Telephone-Based Assessment of Neuropsychological Status (IU-TBANS), which includes four standard clinical instruments covering the following domains: attention, learning and recall, information processing speed, and verbal fluency. This test battery was developed as a sensitive measure of subtle cognitive dysfunction among community-dwelling cancer survivors (at least one year post-treatment) so we
believe it will be sensitive to small deficits in cognitive function among patients with pain. IU-TBANS has been validated in both telephone and in-person forms and can be completed in 15-20 minutes.
Participants will also have blood collected by VA clinical staff according to their routine procedures. Urine for drug testing will be collected at 0,6 , and 12 months. All participants will have total testosterone measured at 0 and 12 months; women will also have estradiol measured at 0 and 12 months. We will assess pain sensitivity at baseline and follow-up visits with the Cold Pressor task, a method that has been used in clinical research with adults and children since the 1960s and is the most sensitive measure of opioid-associated pain hypersensitivity. Following recommended protocols for testing, participants will submerge their non-dominant hand up to the wrist in water continuously circulated by a commercial cooling bath. They will be asked to keep their hand submerged until it becomes too uncomfortable to continue, at which time they are instructed to withdraw it from the water. Pain threshold is time until the stimulus is first perceived as painful; pain tolerance is time until withdrawal of the hand.

Adverse events will be ascertained using a combination of patient report and medical record review. Participants will also be asked to report occurence of any falls, hospitalizations, or emergency visits at each outcome assessment call.
Additionally, falls will be assessed using a prospective calendar and postcard monitoring protocol. Participants will track falls on monthly falls calendars and will be asked to complete a postcard and mail it to the research team if a fall occurs. When postcards are received, a research assistant will follow-up with a phone call to obtain information about the cause and outcome of the fall. To ensure no falls are missed, the calendars will be collected every three months and outcome assessors will follow-up on any additional falls recorded on the calendars but not previously reported by postcard. This method of fall ascertainment is consistent with recommendations based on a systematic review of fall outcome assessment.

Hospitalizations and emergency visits will be identified and confirmed using medical record review at the end of the study. VA hospitalizations and emergency visits will be ascertained using complete data available in VistA; non-VA hospitalizations and emergency visits will be ascertained from patient report and confirmed by medical records. These events will be considered as adverse outcomes regardless of where they occurred. To determine whether the events were analgesic-related, two physicians who are masked to study arm assignment will review records of emergency visits and hospital admissions. The World Health Organization system for standardized case causality assessment will be used to determine whether the relationship between an event and analgesic use is probable or greater. Initial reviews will be conducted independently to allow for calculation of inter-rater reliability. Subsequently, reviewers will meet and attempt to reach consensus; if the two reviewers disagree and cannot reach consensus, a third reviewer will break the tie.
No single measure is adequately predictive or diagnostic of opioid misuse, so data from patients, clinicians, records, and laboratory tests will be use to assess the occurrence of behaviors suggestive of opioid misuse or addiction. In addition to previously described measures, patient-reported substance use history will be evaluated at baseline. The Addiction Behavior Checklist (ABC) will be completed by the clinical pharmacist at each intervention visit. The ABC was developed within a VA chronic pain setting to provide a brief measure for longitudinal clinician assessment of ADRB. Urine drug testing will be obtained at 0,6 , and 12 months as part of outcome assessments for all participants.
Additionally, medical record reviews will be completed at the end of the study. Research assistants trained in chart abstraction will review the electronic medical record over the 12 month study period for documented ADRB and evidence of potential opioid misuse.

In pain clinical studies, dropouts often occur due to intervention-related factors (e.g., adverse effects, lack of treatment efficacy) and missing data can pose a substantial threat to internal validity. SPACE has several design features meant to enhance participant retention and reduce missing data. First, the interventions are flexible treat-to-target prescribing strategies that allow tailoring to each patient's needs. Medications will be adjusted individually, reducing dropouts due to lack of efficacy or intolerance. Second, both study arms are active interventions, which should reduce differential dropout that can occur among patients assigned to a control arm. Third, all randomized participants will be strongly encouraged to complete all outcome assessments, whether or not they continue in the intervention. Fourth, to reinforce participants' sense of commitment to the study, we will send a quarterly SPACE newsletter to participants including tips on pain self-management and positive news from pain research. Finally, if participants are at risk of dropping out due to assessment burden, they will be given the option of completing a minimum core assessment comprising primary pain (i.e., BPI) and adverse effect (e.g., symptom checklist) measures.

We will recruit four subgroups of SPACE participants for qualitative interviews-those who respond to paint treatment and non-responders from each of the two arms. Within each of the subgroups, we will purposefully sample participants with a broad range of intervention adherence behavior (e.g., number of completed care manager contacts). Experience with conduct of the trial will inform our sampling strategy; for example, we may identify patterns of non-adherence that merit further exploration. During the initial consent process for the SPACE trial, all participants will be asked for permission to contact them after intervention completion. The data analyst will provide names of potential participants for the qualitative phase and the study coordinator will contact them within four weeks of completing the SPACE trial.

Throughout data collection and analysis, the research team will remain masked to participants' responder and adherence status. We will use individual semi-structured interviews to address our qualitative aims. Development of the semi-structured interview guide will be informed by the collaborative management framework and TCM as well as questions that arise during the conduct and analysis of the main SPACE trial. Based on experience with the main trial, questions will be further developed, revised, and piloted prior to the beginning of the qualitative phase. The project coordinator, will conduct interviews in a private VA office within 2 months of each participant's trial completion. Interviews will be digitally audio-recorded and transcribed for analysis as they become available. We will continue to conduct interviews with patients until theoretical saturation is reached. Based on our experience and recommendations for sampling subgroups, we expect to conduct up to 10 interviews for each group (responders and non-responders in each study arm), for a total of 40 participant interviews.

Results of the SPACE trial will be interpreted not in isolation, but in the context of results from complementary VA-funded pain clinical trials designed to address different gaps in our knowledge of chronic pain management. Together, these distinct but complementary studies will provide necessary data for clinicians and policy-makers working to improve VA pain care.
The primary knowledge gap SPACE addresses is the long-term benefit of an opioid-intensive prescribing strategy compared with an opioid-avoidant strategy. Information about the magnitude of benefit, along with additional information about the likelihood of harm, is necessary for providers to make the best decisions with individual patients about their pain treatment options. Meanwhile, results of complementary studies, including SCOPE and ESCAPE at the Roudebush VA, will inform VA on how to address provider and structural barriers to pain care.
To disseminate pain research findings, we will take full advantage of the robust pain care infrastructure that exists within the VA. The National Pain Management Strategy Coordinating Committee is an
important channel for disseminating findings to VA providers, administrators, and policy makers. Dr. Bair serves on this committee and will facilitate reporting of findings in the form of reports and presentations at their meetings. Findings will be further disseminated through VISN Pain Points of Contact, the Office of Quality and Performance, and VA Center for Medication Safety. We will also present to the large clinically-oriented audience of the VA pain list-serve and monthly national VA pain educational teleconferences.

Findings will be disseminated to our research audiences through scientific presentations and publications, HSR&D cyber seminars, and presentations to VA and non-VA interest groups. Principal Investigator Dr. Krebs and Co-Investigator Dr. Bair are both active members of the VA Pain Research Working Group and Society of General Internal Medicine (SGIM) Pain Interest Group. Both groups have active list-serves and annual face-to-face meetings. The VA Pain Research Working Group has monthly telephone conferences to promote coordination and dissemination of VA-relevant pain research. In addition, we will disseminate study findings to the Substance Use Disorder and Polytrauma QUERIs and the HSR&D Center for Information Dissemination and Education Resources (CIDER).
(1) Boudreau D, Von KM, Rutter CM et al. Trends in long-term opioid therapy for chronic noncancer pain. Pharmacoepidemiol Drug Saf 2009;18:1166-1175.
(2) Caudill-Slosberg MA, Schwartz LM, Woloshin S. Office visits and analgesic prescriptions for musculoskeletal pain in US: 1980 vs. 2000. Pain 2004; 109:514-9.
(3) Martell BA, O'Connor PG, Kerns RD et al. Systematic review: opioid treatment for chronic back pain: prevalence, efficacy, and association with addiction. Ann Intern Med 2007;146:116127.
(4) Nüesch E, Rutjes AWS, Husni E, Welch V, Jüni P. Oral or transdermal opioids for osteoarthritis of the knee or hip. Cochrane Database of Systematic Reviews 2009.
(5) Chou R, Fanciullo GJ, Fine PG et al. Clinical Guidelines for the Use of Chronic Opioid Therapy in Chronic Noncancer Pain. J Pain 2009; 10:113-130.
(6) VA/DoD Opioid Therapy Guideline Working Group. Clinical Practice Guideline for Management of Opioid Therapy for Chronic Pain. 2010. Washington, D.C., Department of Veterans Affairs; Department of Defense.
(7) Chu LF, Angst MS, Clark D. Opioid-induced hyperalgesia in humans: molecular mechanisms and clinical considerations. Clin J Pain 2008;24:479-496.
(8) Katz N, Mazer NA. The impact of opioids on the endocrine system. Clin J Pain 2009;25:170175.
(9) Wang D, Teichtahl H. Opioids, sleep architecture and sleep-disordered breathing. Sleep Med Rev 2007;11:35-46.
(10) Ballantyne JC, Laforge KS. Opioid dependence and addiction during opioid treatment of chronic pain. Pain 2007;129:235-255.
(11) Oxman TE, Dietrich AJ, Williams JW, Jr., Kroenke K. A three-component model for reengineering systems for the treatment of depression in primary care. Psychosomatics 2002;43:441-450.
(12) Kroenke K, Theobald D, Wu J et al. Effect of telecare management on pain and depression in patients with cancer: a randomized trial. JAMA 2010;304:163-171.
(13) Pain Management: VHA Directive 2009-053. Veterans Health Administration 2009; Available from: Veterans Health Administration. Accessed November 17, 2010.
(14) Verhaak PF, Kerssens JJ, Dekker J, Sorbi MJ, Bensing JM. Prevalence of chronic benign pain disorder among adults: a review of the literature. Pain 1998;77:231-239.
(15) Von Korff M, Saunders K, Thomas RG et al. De facto long-term opioid therapy for noncancer pain. Clin J Pain 2008;24:521-527.
(16) Keller S, Bann CM, Dodd SL, Schein J, Mendoza TR, Cleeland CS. Validity of the Brief Pain Inventory for use in documenting the outcomes of patients with noncancer pain. Clin J Pain 2004;20:309-18.
(17) Tan G, Jensen MP, Thornby JI, Shanti BF. Validation of the Brief Pain Inventory for chronic nonmalignant pain. J Pain 2004;5:133-7.
(18) Callahan CM, Unverzagt FW, Hui SL, Perkins AJ, Hendrie HC. Six-item screener to identify cognitive impairment among potential subjects for clinical research. Med Care 2002;40:771781.
(19) Kroenke K, Bair MJ, Damush TM et al. Optimized antidepressant therapy and pain selfmanagement in primary care patients with depression and musculoskeletal pain: a randomized controlled trial. JAMA 2009;301:2099-2110.
(20) Bex SD, Boldt AS, Needham SB, Bolf SM, Walston CM, Ramsey DC et al. Effectiveness of a Hypertension Care Management Program Provided by Clinical Pharmacists for Veterans. Pharmacotherapy, In press. 2010.
(21) Krebs EE, Lorenz KA, Bair MJ et al. Development and initial validation of the PEG, a threeitem scale assessing pain intensity and interference. J Gen Intern Med 2009;24:733-738.
(22) Krebs EE, Bair MJ, Damush TM, Tu W, Wu J, Kroenke K. Comparative responsiveness of pain outcome measures among primary care patients with musculoskeletal pain. Med Care 2010;48:1007-1014.
(23) Fischer D, Stewart AL, Bloch DA, Lorig K, Laurent D, Holman H. Capturing the patient's view of change as a clinical outcome measure. JAMA 1999;282:1157-1162.
(24) Kroenke K, Spitzer RL, Williams JB, Lowe B. An ultra-brief screening scale for anxiety and depression: the PHQ-4. Psychosomatics 2009;50:613-621.
(25) American Geriatrics Society Panel on the Pharmacological Management of Persistent Pain in Older Persons. Pharmacological management of persistent pain in older persons. J Am Geriatr Soc 2009;57:1331-1346.
(26) Zhang W, Moskowitz RW, Nuki G et al. OARSI recommendations for the management of hip and knee osteoarthritis, Part II: OARSI evidence-based, expert consensus guidelines.
Osteoarthritis Cartilage 2008;16:137-162.
(27) Chou R, Huffman LH. Medications for acute and chronic low back pain: a review of the evidence for an American Pain Society/American College of Physicians clinical practice guideline. Ann Intern Med 2007;147:505-514.
(28) Cepeda MS, Camargo F, Zea C, Valencia L. Tramadol for osteoarthritis: a systematic review and metaanalysis. J Rheumatol 2007;34:543-555.
(29) Mason L, Moore RA, Derry S, Edwards JE, McQuay HJ. Systematic review of topical capsaicin for the treatment of chronic pain.[see comment]. [Review] [33 refs]. BMJ 2004;328:991.
(30) Salerno SM, Browning R, Jackson JL. The effect of antidepressant treatment on chronic back pain: a meta-analysis. Arch Intern Med 2002;162:19-24.
(31) Kroenke K, Krebs EE, Bair MJ. Treatment of Chronic Pain Syndromes. In: Schatzberg AF, Nemeroff CB, eds. Textbook of Psychopharmacology. Fourth edition ed. Washington D.C.: American Psychiatric Publishing, Inc.; 2009; 1441-1474.
(32) Kroenke K, Krebs EE, Bair MJ. Pharmacotherapy of chronic pain: a synthesis of recommendations from systematic reviews. Gen Hosp Psychiatry 2009;31:206-219.
(33) Adli M, Bauer M, Rush AJ. Algorithms and collaborative-care systems for depression: are they effective and why? A systematic review. Biol Psychiatry 2006;59:1029-1038.
(34) Kroenke K, Bair MJ, Damush TM et al. Optimized antidepressant therapy and pain selfmanagement in primary care patients with depression and musculoskeletal pain: a randomized controlled trial. JAMA 2009;301:2099-2110.
(35) Von KM, Tiemens B. Individualized stepped care of chronic illness. West J Med 2000;172:133-137
(36) Chou R. Drug class review on long-acting opioid analgesics. Final report, update #5. Drug Effectiveness Review Project, Oregon Evidence-based Practice Center 2008; Available from: Drug Effectiveness Review Project, Oregon Evidence-based Practice Center. Accessed January 24, 2010.
(37) Dunn KM, Saunders KW, Rutter CM et al. Opioid prescriptions for chronic pain and overdose: a cohort study. Ann Intern Med 2010;152:85-92.
(38) Agency Medical Directors' Group. Interagency Guideline on Opioid Dosing for Chronic Noncancer Pain: An educational aid to improve care and safety with opioid therapy . www agencymeddirectors wa gov 2010; Accessed November 15, 2010.
(39) Starrels JL, Becker WC, Alford DP, Kapoor A, Williams AR, Turner BJ. Systematic review: treatment agreements and urine drug testing to reduce opioid misuse in patients with chronic pain. Ann Intern Med 2010;152:712-720.
(40) Lanza FL, Chan FKL, Quigley EMM. Guidelines for Prevention of NSAID-Related Ulcer Complications. Am J Gastroenterol 2009;104:728-738.
(41) Krantz MJ, Martin J, Stimmel B, Mehta D, Haigney MC. QTc interval screening in methadone treatment. Ann Intern Med 2009;150:387-395.
(42) American Geriatrics Society Panel on the Pharmacological Management of Persistent Pain in Older Persons. Pharmacological management of persistent pain in older persons. J Am Geriatr Soc 2009;57:1331-1346.
(43) Kroenke K, Spitzer RL, Williams JB. The PHQ-9: validity of a brief depression severity measure. J Gen Intern Med 2001;16:606-13.
(44) Spitzer RL, Kroenke K, Williams JB, Lowe B. A brief measure for assessing generalized anxiety disorder: the GAD-7. Arch Intern Med 2006;166:1092-1097.
(45) Wu SM, Compton P, Bolus R et al. The addiction behaviors checklist: validation of a new clinician-based measure of inappropriate opioid use in chronic pain. J Pain Symptom Manage 2006;32:342-351.
(46) Williams LS, Kroenke K, Bakas T et al. Care management of poststroke depression: a randomized, controlled trial. Stroke 2007;38:998-1003.
(47) Dworkin RH, Turk DC, Farrar JT et al. Core outcome measures for chronic pain clinical trials: IMMPACT recommendations. Pain 2005;113:9-19.
(48) Keller S, Bann CM, Dodd SL, Schein J, Mendoza TR, Cleeland CS. Validity of the Brief Pain Inventory for use in documenting the outcomes of patients with noncancer pain. Clin J Pain 2004;20:309-18.
(49) Tan G, Jensen MP, Thornby JI, Shanti BF. Validation of the Brief Pain Inventory for chronic nonmalignant pain. J Pain 2004;5:133-7.
(50) Roland M, Fairbank J. The Roland-Morris Disability Questionnaire and the Oswestry Disability Questionnaire. Spine 2000;25:3115-3124.
(51) Stroud MW, McKnight PE, Jensen MP. Assessment of self-reported physical activity in patients with chronic pain: development of an abbreviated Roland-Morris disability scale. Pain 2004;5:257-263.
(52) Fischer D, Stewart AL, Bloch DA, Lorig K, Laurent D, Holman H. Capturing the patient's view of change as a clinical outcome measure. JAMA 1999;282:1157-1162.
(53) McHorney CA, Ware JE, Jr., Raczek AE. The MOS 36-Item Short-Form Health Survey (SF36): II. Psychometric and clinical tests of validity in measuring physical and mental health constructs. Med Care 1993;31:247-263.
(54) Weingart SN, Gandhi TK, Seger AC et al. Patient-reported medication symptoms in primary care. Arch Intern Med 2005;165:234-240.
(55) Jonsson T, Christrup LL, Hojsted J et al. Symptoms and side effects in chronic non-cancer pain: patient report vs. systematic assessment. Acta Anaesthesiol Scand 2011;55:69-74.
(56) Lin JM, Brimmer DJ, Maloney EM, Nyarko E, Belue R, Reeves WC. Further validation of the Multidimensional Fatigue Inventory in a US adult population sample. Popul Health Metr 2009;7:18.
(57) Wells GA, Li T, Kirwan JR et al. Assessing quality of sleep in patients with rheumatoid arthritis. J Rheumatol 2009;36:2077-2086.
(58) Kosinski M, Bayliss MS, Bjorner JB et al. A six-item short-form survey for measuring headache impact: the HIT-6. Qual Life Res 2003;12:963-974.
(59) McGahuey CA, Gelenberg AJ, Laukes CA et al. The Arizona Sexual Experience Scale (ASEX): reliability and validity. J Sex Marital Ther 2000;26:25-40.
(60) Prins A, Ouimette P, Kimerlin R et al. The primary care PTSD screen (PC-PTSD): development and operating characteristics. Primary Care Psychiatry 2003;9:9-14.
(61) Fiellin DA, Reid MC, O'Connor PG. Screening for alcohol problems in primary care: a systematic review. Arch Intern Med 2000;160:1977-1989.
(62) Smith PC, Schmidt SM, Allensworth-Davies D, Saitz R. A single-question screening test for drug use in primary care. Arch Intern Med 2010;170:1155-1160.
(63) Substance Abuse and Mental Health Services Administration. Results from the 2008 National Survey on Drug Use and Health: National Findings (Office of Applied Studies, NSDUH Series H-36, HHS Publication No. SMA 09-4434). 2009. Rockville, MD.
(64) Schieffer BM, Pham Q, Labus J et al. Pain medication beliefs and medication misuse in chronic pain. J Pain 2005;6:620-629.
(65) Linde K, Witt CM, Streng A et al. The impact of patient expectations on outcomes in four randomized controlled trials of acupuncture in patients with chronic pain. Pain 2007;128:264271.
(66) Rose DJ, Lucchese N, Wiersma LD. Development of a multidimensional balance scale for use with functionally independent older adults. Arch Phys Med Rehabil 2006;87:1478-1485.
(67) Rutledge DN, Cherry BJ, Rose DJ, Rakovski C, Jones CJ. Do fall predictors in middle aged and older adults predict fall status in persons with fibromyalgia? An exploratory study. Res Nurs Health 2010;33:192-206.
(68) Jones CJ, Rikli RE, Beam WC. A 30-s chair-stand test as a measure of lower body strength in community-residing older adults. Res Q Exerc Sport 1999;70:113-119.
(69) Terwee CB, Mokkink LB, Steultjens MP, Dekker J. Performance-based methods for measuring the physical function of patients with osteoarthritis of the hip or knee: a systematic review of measurement properties. Rheumatology (Oxford) 2006;45:890-902.
(70) Fransen M, Crosbie J, Edmonds J. Reliability of gait measurements in people with osteoarthritis of the knee. Phys Ther 1997;77:944-953.
(71) Unverzagt FW, Monahan PO, Moser LR et al. The Indiana University telephone-based assessment of neuropsychological status: a new method for large scale neuropsychological assessment. J Int Neuropsychol Soc 2007;13:799-806.
(72) Mitchell LA, MacDonald RA, Brodie EE. Temperature and the cold pressor test. J Pain 2004;5:233-237.
(73) Ganz DA, Higashi T, Rubenstein LZ. Monitoring falls in cohort studies of communitydwelling older people: effect of the recall interval. J Am Geriatr Soc 2005;53:2190-2194.
(74) Uppsala Monitoring Centre. WHO-UMC system for standardised case causality assessment. 11. Available at http://who-umc.org/Graphics/24734.pdf. Accessed 11/16/11.
(75) Michna E, Ross EL, Hynes WL et al. Predicting aberrant drug behavior in patients treated for chronic pain: importance of abuse history. J Pain Symptom Manage 2004;28:250-258.
(76) National Research Council. The Prevention and Treatment of Missing Data in Clinical Trials. Washington, D.C.: National Academies Press, 2010.
(77) Crabtree BF, Miller WLE. Doing qualitative research. Second ed. Thousand Oaks, California: Sage Publications, 1999.