Office visits and analgesic prescriptions for... : PAIN (original) (raw)
1. Introduction
Over the past two decades physicians have been encouraged to fundamentally rethink their approach to pain, from focusing on underlying causes of pain to the treatment of pain itself. This change in focus has been a response, in part, to reports on the inadequacies of pain treatment for acute medical (Marks and Sachar, 1973) inpatient post-operative (Wall, 1988; Ward, 1986), and cancer pain (Bernabei et al., 1998; Portenoy and Lesage, 1999). The endorsement of pharmacologic pain treatment has been strengthened by the analgesic pain ladder of the World Health Organization, national guidelines sponsored by the Agency for Health Care Policy and Research (AHCPR, 1992, 1994) the ‘pain as the 5th vital sign’ campaign (Sayers et al., 2000) and direct-to-consumer advertising by pharmaceutical companies. Most recently, in 2001, the Joint Commission of Hospital Accreditation began incorporating the evaluation and treatment of pain into their hospital assessment and certification process (Phillips, 2000).
A campaign promoting primary pain treatment to health professionals and to the public-at-large might be expected to increase health care visits for pain and the number and perhaps type of analgesics prescribed. We used the National Ambulatory Care Survey to see how outpatient visits for and treatment of musculoskeletal pain changed in the United States from 1980 and 2000.
2. Methods
2.1. Data source
The National Ambulatory Medical Care Survey (NAMCS) is a nationally representative, yearly survey that collects information about outpatient office visits in the United States. It is conducted by the National Center for Health Statistics (NCHS) of the Centers for Disease Control. We analyzed visits and prescriptions using NAMCS data from 1980 to 1981 and 1999 to 2000 obtained from public files on the NCHS website (2002). The combined data will be referred to as 1980 and 2000, respectively. Patients of all ages were included.
All non-Federally and non-institutionally employed medical and osteopathic physicians (excluding anesthesiologists, radiologists and pathologists) are eligible to participate in NAMCS. The physicians are randomly chosen from sampling lists of all physicians in the master files maintained by the American Medical Association and the American Osteopathic Association. The participating physicians collect patient information from a random sampling of visits (approx. 30/week) during a randomly assigned week of clinical practice. Sample visits include only face-to-face outpatient encounters (i.e. hospital visits and phone calls are excluded). At each visit, the physician aided by office staff completes a one-page data form. This form records patient, visit and physician characteristics. The average physician participation response rates were 77% in 1980–1981 and 66% in 1999–2000. Complete details about the NAMCS survey are available elsewhere (2002).
2.2. Measures
2.2.1. Musculoskeletal pain visits
In NAMCS, the patient's most important reason for visit and up to two other reasons are captured, in the patient's own words, on the data form (See Appendix, Item 13). These are then coded by the NCHS field staff using a unique classification system developed by the NCHS (1979). We defined a ‘musculoskeletal pain visit’ from those codes if the patient stated that the most important reason for the visit was pain in the back, upper and lower extremities, neck, headache and generalized non-specific pain. We focused on ‘the most important reason for visit’ response to obtain a more homogeneous reason for visit from the patient's perspective. We focused on musculoskeletal pain since analgesics are more likely to be used as the primary treatment for these complaints, as opposed to other conditions causing pain (e.g. antibiotics for ear pain).
2.2.2. Visit acuity
We created acute and chronic visit categories using the physician's recording of ‘acute problem’, ‘chronic problem, routine’ or ‘chronic problem, flare up’ for the major reason for the patient's visit. Acute visits were those where the problem was present for less than 3 months. For the chronic visit category, we combined the 2 chronic acuity ratings since they both referred to conditions present for three or more months. Post-surgery/Injury ‘follow-up’ visits were excluded in the musculoskeletal pain visit analyses because of their being neither acute, nor of ill-defined chronicity. The visit acuity variable, major, was not included through the years from 1985 until 1997 so that trends over the 2 decades could not be characterized.
2.2.3. Pharmacologic treatments
Physicians recorded medications prescribed (or continued) at the visit. Medications were then coded according to a NCHS classification system (Koch and Campbell, 1982). We looked at the following three groups of medications.
2.2.3.1. Opioids
Opioids included those medications known to bind to mu, kappa or delta receptor subtypes. Agents were coded as ‘weak’ (codeine, propoxyphene and, for 2000 only, tramadol), or ‘strong’ (hydrocodone, oxycodone, morphine, methadone, fentanyl, hydromorphone and meperidine). Opioids in cough syrup preparations and those available only in IV/IM forms were excluded.
2.2.3.2. NSAIDS
NSAIDS were all those medications identified as cyclooxygenase inhibitors. In 1980 this class included only COX I inhibitors; in 2000 COX II inhibitors were also included. Acetominophen was analyzed separately and only as medication where acetominophen was the sole active ingredient.
2.2.3.3. Other pain medications
This category included anti-seizure medications (e.g. gabapentin, tegretol), tricyclic antidepressants and muscle relaxants.
2.2.4. Non-pharmacologic treatments
In 1980, the only non-pharmacologic therapies recorded were psychotherapy/therapeutic listening and physiotherapy. In 2000, the pain-relevant options were expanded to include stress management, complementary medicine therapy, exercise, mental health and counseling/education.
2.2.5. Analysis
To provide national estimates, all analyses incorporate sampling weights which account for differential probability of selection and non-response (Woodwell, 2000). Population-based visit rates use the appropriate mid-year civilian non-institutionalized US population estimates as reported by NAMCS for the denominator (2002). Standard error calculations were corrected to account for the complex sampling design employed by NAMCS. For year 2000 data, we did so by using the -svy series of commands in STATA V.7.0 (College Station, TX) and applying the design variables supplied by NAMCS. These design variables were not issued by NCHS in the 1980 NAMCS. In this case, we derived standard errors from the generalized variance formula published by NCHS based on the estimated proportions and number of visits (Knapp and Koch, 1984; Koch, 1986).
For statistical testing, we used the weighted point estimates and corrected standard errors to calculate a z-test for proportions.
3. Results
3.1. Musculoskeletal pain visits
Annual musculoskeletal pain visit rates in the United States changed little between 1980 and 2000, increasing from 250 to 280 visits per 1000 people (Fig. 1), with visits about evenly divided between acute and chronic complaints. Because these small increases closely paralleled changes in total outpatient visit rates, musculoskeletal pain visits, as a proportion of all visits, did not change during the study period (9.6 vs. 9.9%, _P_=0.5). The characteristics of acute and chronic pain patients were similar at both times with the exception of age: chronic pain patients were typically about 10 years older (Table 1). Patient and visit characteristics changed in several respects over the 20 years (Table 2). While the median visit length was stable, the mean visit length increased by 3 min (mean visit time has increased for all office visits somewhere between 1 and 2 min (Mechanic et al., 2001)). Chronic pain patients were increasingly seen by specialists (34% in 1980 vs. 49% in 2000, P<0.001). The distribution of pain complaints also shifted somewhat between 1980 and 2000: there was a halving of visits for chronic headache (17 vs. 9% of visits, P<0.001), and an increase in both chronic lower and upper extremity pain (24 vs. 32%, and 13 vs. 18%, respectively [both P<0.01]).
Visit rates in 1980 and 2000.
Patient characteristics of visits for musculoskeletal pain complaints
Visit characteristics of musculoskeletal pain
3.2. Pharmacologic pain treatment
While visit rates remained stable during the study period, analgesic-prescribing patterns changed in several ways. First, use of opioids increased at both acute (8 vs. 11%, RR=1.38; 95% CI, 0.92–1.83) and chronic pain (8 vs. 16%, RR=2.0; 95% CI, 1.52–2.48) visits (Fig. 2). At the same time, the type of opioid prescribed also changed; use of strong opioids such as hydrocodone, oxycodone, morphine increased sixfold for acute pain and quadrupled for chronic pain, and were prescribed at 6 and 9% of visits, respectively, in 2000. These increases cannot be explained by a significant increase in visits by patients with cancer-related ICD-9CM diagnoses as these comprised only 2 and 1% of visits associated with patient complaints of musculoskeletal pain in 1980 and 2000, respectively.
Proportion of musculoskeletal pain visits at which opioids and NSAIDs were prescribed.
Second, NSAID prescriptions increased substantially at acute pain visits: in 1980 NSAIDS were prescribed at 19% of such visits and rose to 33% in 2000; RR=1.74; 95% CI, 1.52–1.9 (Fig. 2). NSAID prescriptions changed only nominally with chronic pain visits in this time frame (25 vs. 29%, RR=1.16; 95% CI, 0.97–1.35). Of note, COX-II inhibitors accounted for almost one-third of NSAID prescriptions at either acute or chronic pain visits in 2000. The use of acetominophen was consistently associated with 2% of musculoskeletal visits.
We also examined the use of other medications that have been promoted for pain treatment, particularly in chronic pain. Because the use of these medications for this purpose is a relatively recent phenomenon, we focused on the rates of use of these medications in 2000. Tricyclic antidepressants were prescribed at 2% of acute and 3% of chronic pain visits. Anti-seizure medications (e.g. gabapentin) were prescribed at 1 and 3%; and muscle relaxants at 8 and 6% of these visits, respectively.
3.3. Non-pharmacologic pain treatment
Only physiotherapy and psychotherapy were recorded across the two time periods studied. At both acute and chronic pain visits, prescriptions for physiotherapy decreased from 22 and 21%, respectively, to 16% (both P<0.01) from 1980 to 2000. During both time periods, psychotherapy was prescribed at 2 and 4% of visits for acute and chronic pain, respectively. In 2000, NAMCS also recorded prescribing of other non-pharmacologic treatments including mental health counseling, complementary medicine and stress management. Less than 3% of musculoskeletal pain visits were associated with use of these modalities.
4. Discussion
The campaign promoting primary pain treatment in the past decade has not led to an increase in patients presenting to the ambulatory care setting with musculoskeletal pain. Other than a shift in increased age, increased chronic pain visits to specialists and some changes in the distribution of pain complaints, patient and visit characteristics were quite stable between 1980 and 2000. At the same time, however, there have been some important changes in the treatment of pain in the ambulatory care setting; most notably, increases in the use of strong opioids for chronic pain and NSAIDS for acute pain.
Our findings should be interpreted in light of several potential limitations. First, only outpatient encounters where patients described pain as the most important reason for the visit were counted as pain visits. Since pain may be an important reason for the visit but might not be considered the most important reason, we undoubtedly undercount pain visits. Changes in visit rates between 1980 and 2000 should be unaffected, however, since undercounting is just as likely to occur at either time. Second, these findings do not provide explanations for the changes in visit characteristics we observed. The increased visits to specialists for chronic pain may reflect increasing pressure on primary care clinicians to provide a diagnosis for chronic pain or an increase in patients seeking out specialist opinion. The reduction in visits for chronic headaches may reflect improvement in headache treatments—specifically the availability of more effective OTC anti-inflammatory pain medications and more effective migraine therapy (i.e. sumatriptans) introduced after 1980. In addition, we did not further characterize the ICD-9CM diagnoses of musculoskeletal visits because these are not directly associated with the patient's self-reported most important reason for the visit. As would be expected, however, many of these visits had a diagnostic code of low back pain and osteoarthroses assigned to them.
Second, we also underestimate medication use since NAMCS only codes those medications actually prescribed or renewed at the visit. Undercounting is probably less of an issue with opioids compared to NSAIDS or other medications, since narcotic prescriptions are generally limited to a single month supply. Again, undercounting of opioids may affect absolute rates, but relative rates should be accurate. We are less confident about relative rates for COX-I NSAIDS. The problem here is that in 1980, NSAIDS were only available by prescription; by 2000 many NSAIDS were available over the counter. Since over the counter medications are captured but may be captured less reliably, we probably underestimate the increase in COX-I NSAID use.
The campaign for more intensive treatment of pain was largely targeted at acute, post-operative and cancer pain where the evidence of benefit and safety is strongest. Changes in the pattern of NSAID were consistent with this focus: the greatest increase in their use was seen at visits for acute pain, implying problems (and use) of shorter duration. Of interest was the substantial use of the COX II-NSAIDS despite the limited evidence of any demonstrable additional benefit in treating pain (Cannon et al., 2000). Aggressive marketing of these medications (Young, 2002) and claims of a better side-effect profile for COX II's may have influenced clinicians in choosing this class of NSAIDS, although the absolute reduction in long-term side effects is now being questioned (Mukherjee et al., 2001; Whelton et al., 2001).
Our results suggest that recommendations for the use of opioids in treating acute and cancer pain have been generalized to chronic, non-malignant pain. Such generalization is problematic given the limited data supporting the effectiveness or safety of long-term (>12 months) of opioids for these purposes. Studies about chronic opioid use in non-cancer pain, for example, have not consistently demonstrated improvements in mood or function (Ciccone et al., 2000; Harden, 2002; Schofferman, 1993). Moreover, the extent to which addiction is a problem in these settings is unknown (Harden, 2002; Savage, 1996). While the risk of addiction for a person on long-term opioids is small-reported to be somewhere between 3 and 16% (Savage, 1996), this may translate into a significant problem on a population basis given the large increase in exposure to opioids. Nevertheless, the increased use of opioids for chronic pain is understandable. These medications have been heavily marketed by pharmaceutical companies (Websource, 2002); and they often provide some immediate relief to patients with long-standing, otherwise refractory problems. Furthermore, there may have been a difference in the demands of the population in the 2000 dataset presenting to the ambulatory care setting who have already been using over-the-counter NSAIDs thereby making the use of opioids more necessary.
There is a growing consensus among pain specialists about the value of tricyclics, anti-seizure medications and non-pharmacologic therapies for chronic pain syndromes (Clinical_Practice_Guidelines_Committee, 2002). For example, tricyclic antidepressants have been shown to be beneficial for neuropathic pain (Fishbain, 2000); anti-seizure medication appear to help with some chronic neuropathies such as trigeminal neuralgia, post-herpetic neuralgia and diabetic neuropathy (Serpell and Neuropathic_Pain__Study_Group, 2002; Wiffen et al., 2003); and cognitive behavioral therapies (Morley et al., 1999) and stress management/educational programs (Superio-Cabuslay et al., 1996) are effective in chronic musculoskeletal conditions (e.g. osteoarthritis and rheumatoid arthritis). Aerobic exercise has also been found to improve physical capacity and pain symptoms in some chronic conditions including fibromyalgia (Busch et al., 2003) and osteoarthritis (O'Reilly et al., 1999). Surprisingly, we found very little use of any of the foregoing modalities. One explanation for this is that the majority of opioids for chronic musculoskeletal pain were prescribed by primary care physicians and orthopedists who may not be as aware of the potential benefits of other pharmacologic and non-pharmacologic therapies.
In summary, campaigns to increase attention to pain appear to have mostly affected how pain is treated rather than the frequency of patient visits for pain. The most substantial change has been in the use of opioids for chronic pain rather than acute pain though the evidence is strongest for the latter indication. Appropriate differentiation of acute from chronic pain by mechanisms that promote consideration of both non-pharmacologic as well as pharmacologic therapies may better serve the patient population presenting with pain as their most important reason for visit.
Acknowledgements
We would like to acknowledge Brenda Sirovich, MD, Todd MacKenzie, PhD for assistance with data analysis and the VA Outcomes Group for their helpful comments about the study design.
Dr Caudill-Slosberg was supported by a Veterans Affairs Fellowship in Ambulatory Care and is currently supported as a Veterans Affairs Quality Scholar. Drs Woloshin and Schwartz are supported by Veterans Affairs Advanced Career Development Awards in Health Services Research and Development; Dr Woloshin is also supported by a Robert Wood Johnson Generalist Faculty Scholar Award. The views expressed herein do not necessarily represent the views of the Department of Veterans Affairs or the United States Government.
Drs Woloshin and Schwartz are supported by Veterans Affairs Advanced Career Development Awards and Robert Wood Johnson Generalist Faculty Scholar Awards.
References
AHCPR, Acute pain management: operative or medical procedures and trauma. AHCPR Publication No. 92-0032; 1992.
AHCPR, Management of cancer pain. AHCPR Publication No. 94-0592; 1994.
Bernabei R, Gambassi G, Lapane K, Landi F, Gatsonis C, Dunlop R, Lipsitz L, Steel K, Mor V. Management of pain in elderly patients with cancer. JAMA. 1998;279:1877-1882.
Busch A, Schachter C, Peloso P, Bombardier C. Exercise for treating fibromyalgia, vol. 2003. The Cochrane Database of Systematic Reviews, The Cochrane Library, The Cochrane Collaboration; 2003.
Cannon GW, Caldwell JR, Holt P, McLean B, Seidenberg B, Bolognese J, Ehrich E, Mukhopadhyay S, Daniels B. Rofecoxib, a specific inhibitor of cyclooxygenease 2, with clinical efficacy comparable with that of diclofenac sodium: results of a one-year, randomized, clinical trial in patients with osteoarthritis of the knee and hip. Arthritis Rheum. 2000;43:978-987.
Ciccone DS, Just N, Bandilla EB, Reimer E, Ilbeigi MS, Wu W.-. Psychological correlates of opioid use in patients with chronic nonmalignant pain: a preliminary test of the Downhill spiral hypothesis. J Pain Symptom Manage. 2000;20:180-192.
Clinical_Practice_Guidelines_Committee, Guideline for the management of pain in osteoarthritis, rheumatoid arthritis, and juvenile chronic arthritis, Glenview, IL; 2002.
Fishbain D. Evidence-based data on pain relief with antidepressants. Ann Med. 2000;32:305-316.
Harden N. Chronic opioid therapy: another reappraisal. APS bulletin/pain and public policy, vol. 2002. American Pain Society; 2002.
Knapp D, Koch H. The management of new pain in office based ambulatory care: national ambulatory medical care survey; 1980 and 1981, 97, 1984.
Koch H. The management of chronic pain in office-based ambulatory care: national ambulatory medical care survey. DHHS Pub No. (PHS) 86-1250; 1986.
Koch H, Campbell W. The collection and processing of drug information: national ambulatory medical care survey, 1980. Vital Health Stat. 1982;2(90):1-90.
Marks R, Sachar E. Undertreatment of medical patients with narcotic analgesics. Ann Med. 1973;78:173-181.
Mechanic D, McAlpine D, Rosenthal M. Are patients' office visits with physicians getting shorter? N Engl J Med. 2001;344:198-204.
Morley S, Eccleston C, Williams A. Systematic review and meta-analysis of randomized controlled trials of cognitive behavior therapy and behavior therapy for chronic pain in adults, excluding headache. Pain. 1999;80:1-13.
Mukherjee D, Nissen S, Topol E. Risk of cardiovascular events associated with selective COX-2 inhibitors. JAMA. 2001;286:954-959.
National_Center_for_Health_Statistics. National center for health statistics: a reason for visit classification for ambulatory care, United States. Vital and health statistics. Series 2-No. 78. Hyattsville, MD: Public Health Service; 1979.
National_Center_for_Health_Statistics. National center for health statistics: ambulatory health care data; 2002.
O'Reilly S, Muir K, Doherty M. Effectiveness of home exercise on pain and disability from osteoarthritis of the knee: a randomised controlled trial. Ann Rheumatol Dis. 1999;58:15-19.
Phillips D. JCAHO pain management standards are unveiled. JAMA. 2000;1-5.
Portenoy R, Lesage P. Management of cancer pain. Lancet. 1999;15:1695-1700.
Savage S. Long term opioid therapy: assessment of consequences and risks. J Pain Symptom Manage. 1996;11:274-286.
Sayers M, Marando R, Fisher S, Aquila A, Morrison B, Dailey T. No need for pain. J Healthcare Qual. 2000;22:10-15.
Schofferman J. Long-term use of opioid analgesics for the treatment of chronic pain of nonmalignant origin. J Pain Symptom Manage. 1993;8:279-288.
Serpell M, Neuropathic_Pain__Study_Group. Gabapentin in neuropathic pain syndromes: a randomised, double-blind, placebo-controlled trial. Pain. 2002;99:557-566.
Superio-Cabuslay E, Ward M, Lorig K. Patient education interventions in osteoarthritis and rheumatoid arthritis: a meta-analytic comparison with non-steroidal anti-inflammatory drug treatment. Arthritis Care Res. 1996;9:292-301.
Wall P. The prevention of post-operative pain. Pain. 1988;33:289-290.
Ward N. Tricyclic antidepressants for chronic low-back pain. Mechanisms of action and predictors of response. Spine. 1986;11:661-665.
Websource. Pain management: blockbusters drive US pain management pharmaceuticals market, Drug Week; 2002. http:www.NewsRx.net.
Whelton A, Fort J, Puma J, Normandin D, Bello A, Verburg K. Cyclooxygenase-2-specific inhibitors and cardiorenal function: a randomized, controlled trial of celecoxib and rofecoxib in older hypertensive osteoarthritis patients. Am J Ther. 2001;8:85-95.
Wiffen P, Collins S, McQuay H, Carroll D, Jadad A, Moore A. Anticonvulsant drugs for acute and chronic pain, vol. 2003. The Cochrane Database of Systematic Reviews, The Cochrane Library, The Cochrane Collaboration; 2003.
Woodwell D. National ambulatory medical care survey: 1998 summary. Adv Data. 2000;1-26.
Young D. Studies show drug ads influence prescription decisions, drug costs. Am J Health-Syst Pharm. 2002;59:14-16.
Keywords:
Musculoskeletal pain; National Ambulatory Medical Care Survey; Opioids
© 2004 Lippincott Williams & Wilkins, Inc.