The Clinical Practice Guidelineon which this Quick Reference Guide for Clinicians is based was developed by a multidisciplinary, private-sector panel comprising health care professionals and a consumer representative sponsored by the Agency for Health Care Policy and Research (AHCPR). Panel members were:
Special consultants to the panel were: Michele Batti, PT, PhD; Claire Bombardier, MD; Nortin Hadler, MD; Alf Nachemson, MD, PhD; Gordon Waddell, MD. John Holland, MD, MPH and John Webster, MD served as project directors. Project methodologists were David Schriger, MD, MPH and Paul Shekelle, MD,MPH.
An explicit, science-based methodology was employed along with expert clinical judgment to develop specific statements on patient assessment and management on acute low back problems. Extensive literature searches were conducted and critical reviews and syntheses were used to evaluate empirical evidence and significant outcomes. Peer review and pilot testing were undertaken to evaluate the validity, reliability, and utility of the guideline in clinical practice.
This Quick Reference Guide for Clinicians presents a clinical strategy for applying the statements and recommendations from the Clinical Practice Guideline.The latter provides a description of the guideline development process, thorough analysis and discussion of the available research, critical evaluation of the assumptions and knowledge of the field, more complete information for health care decisionmaking, consideration for patients with special needs, and references. Decisions to adopt particular recommendations from either publication must be made by practitioners in light of available resources and circumstances presented by the individual patient. AHCPR invites comments and suggestions from users for consideration in development and updating of future guidelines.
Please send written comments to:
This Quick Reference Guide for Clinicians contains highlights from the Clinical Practice Guideline version of Acute Low Back Problems in Adults, which was developed by a private-sector panel of health care providers and consumers. The Quick Reference Guide is an example of how a clinician might implement the panel's findings and recommendations on the management of acute low back problems in working-age adults. Topics covered include the initial assessment of patients presenting with acute low back problems, identification of red flags that may indicate the presence of a serious underlying medical condition, initial management, special studies and diagnostic considerations, and further management considerations. Instructions for clinical testing for sciatic tension, recommendations for sitting and unassisted lifting, tests for identification of clinical pathology, and algorithms for patient management are included.
This document is in the public domain and may be used and reprinted without special permission. AHCPR appreciates citation as to source and the suggested format is provided below: Bigos S, Bowyer O, Braen G, et al. Acute Low Back Problems in Adults. Clinical Practice Guideline, Quick Reference Guide Number. 14. Rockville, MD: U.S. Department of Health and Human Services, Public Health Service, Agency for Health Care Policy and Research, AHCPR Pub. No. 95-0643. December 1994.
Low back problems affect virtually everyone at some time during their life. Surveys indicate a yearly prevalence of symptoms in 50 percent of working age adults; 15-20 percent seek medical care. Low back problems rank high among the reasons for physician office visits and are costly in terms of medical treatment, lost productivity, and nonmonetary costs such as diminished ability to perform or enjoy usual activities. In fact, for persons under age 45, low back problems are the most common cause of disability.
Acute low back problems are defined as activity intolerance due to lower back or back-related leg symptoms of less than 3 months' duration. About 90 percent of patients with acute low back problems spontaneously recover activity tolerance within 1 month. The approach to a new episode in a patient with a recurrent low back problem is similar to that of a new acute episode.
The findings and recommendations included in the Clinical Practice Guideline define a paradigm shift away from focusing care exclusively on the pain and toward helping patients improve activity tolerance. The intent of this Quick Reference Guide is to bring to life this paradigm shift. The guide provides information on the detection of serious conditions that occasionally cause low back symptoms (conditions such as spinal fracture, tumor, infection, cauda equina syndrome, or non-spinal conditions). However, treatment of these conditions is beyond the scope of this guideline. In addition, the guideline does not address the care of patients younger than 18 years or those with chronic back problems (back-related activity limitations of greater than 3 months' duration).
A focused medical history and physical examination are sufficient to assess the patient with an acute or recurrent limitation due to low back symptoms of less than 4 weeks duration. Patient responses and findings on the history and physical examination, referred to as "red flags" (Table 1), raise suspicion of serious underlying spinal conditions. Their absence rules out the need for special studies during the first 4 weeks of symptoms when spontaneous recovery is expected. The medical history and physical _examination can also alert the clinician to non- spinal pathology (abdominal, pelvic, thoracic) that can present as low back symptoms. Acute low back symptoms can then be classified into one of three working categories:
Table 1. Red flags for potentially serious conditions
In addition to detecting serious conditions and categorizing back symptoms, the medical history establishes rapport between the clinician and patient. The patient's description of present symptoms and limitations, duration of symptoms, and history of previous episodes defines the problem. It also provides insight into concerns, expectations, and nonphysical (psychological and socioeconomic) issues that may alter the patient's response to treatment. Assessment tools such as pain drawings and visual analog pain-rating scales may help further document the patient's perceptions and progress.
A patient's estimate of personal activity intolerance due to low back symptoms contributes to the clinical assessment of the severity of the back problem, guides treatment, and establishes a baseline for recommending daily activities and evaluating progress.
Open-ended questions, such as those listed below, can gauge the need for further discussion or specific inquiries for more detailed information:
Guided by the medical history, the physical examination includes:
The examination is mostly subjective since patient response or interpretation is required for all parts except reflex testing and circumferential measurements for atrophy.
Physical examination evidence of severe neurologic compromise that correlates with the medical history may indicate a need for immediate consultation. The examination may further modify suspicions of tumor, infection, or significant trauma. A medical history suggestive of non- spinal pathology mimicking a back problem may warrant examination of pulses, abdomen, pelvis, or other areas.
Limping or coordination problems indicate the need for specific neurologic testing. Severe guarding of lumbar motion in all planes may support a suspected diagnosis of spinal infection, tumor, or fracture. However, given marked variations among persons with and without symptoms, range-of-motion measurements of the back are of limited value.
Vertebral point tenderness to palpation, when associated with other signs or symptoms, may be suggestive of but not specific for spinal fracture or infection. Palpable soft-tissue tenderness is, by itself, an even less specific or reliable finding.
The neurologic examination can focus on a few tests that seek evidence of nerve root impairment, peripheral neuropathy, or spinal cord dysfunction. Over 90 percent of all clinically significant lower extremity radiculopathy due to disc herniation involves the L5 or S1 nerve root at the L4-5 or L5-S1 disc level. The clinical features of nerve root compression are summarized in Figure 1.
The patient's inability to toe walk (calf muscles, mostly S1 nerve root), heel walk (ankle and toe dorsiflexor muscles, L5 and some L4 nerve roots), or do a single squat and rise (quadriceps muscles, mostly L4 nerve root) may indicate muscle weakness. Specific testing of the dorsiflexor muscles of the ankle or great toe (suggestive of L5 or some L4 nerve root dysfunction), hamstrings and ankle evertors (L5-S1), and toe flexors (S1) is also important.
Muscle atrophy can be detected by circumferential measurements of the calf and thigh bilaterally. Differences of less than 2 cm in measurements of the two limbs at the same level may be a normal variation. Symmetrical muscle bulk and strength are expected unless the patient has a neurologic impairment or a history of lower extremity muscle or joint problem.
The ankle jerk reflex tests mostly the S1 nerve root and the knee jerk reflex tests mostly the L4 nerve root; neither tests the L5 nerve root. The reliability of reflex testing can be diminished in the presence of adjacent joint or muscle problems. Up-going toes in response to stroking the plantar footpad (Babinski or plantar response) may indicate upper motor-neuron abnormalities (such as myelopathy or demyelinating disease) rather than a common low back problem.
Testing light touch or pressure in the medial (L4), dorsal (L5), and lateral (S1) aspects of the foot (Figure 1) is usually sufficient for sensory screening.
The straight leg raising (SLR) test (Figure 2) can detect tension on the L5 and/or S1 nerve root. SLR may reproduce leg pain by stretching nerve roots irritated by a disc herniation.
Pain below the knee at less than 70 degrees of straight leg raising, aggravated by dorsiflexion of the ankle and relieved by ankle plantar flexion or external limb rotation, is most suggestive of tension on the L5 or S1 nerve root related to disc herniation. Reproducing back pain alone with SLR testing does not indicate significant nerve root tension.
Crossover pain occurs when straight raising of the patient's well limb elicits pain in the leg with sciatica. Crossover pain is a stronger indication of nerve root compression than pain elicited from raising the straight painful limb.
Sitting knee extension (Figure 3) can also test sciatic tension. The patient with significant nerve root irritation tends to complain or lean backward to reduce tension on the nerve.
The patient who embellishes a medical history, exaggerates pain drawings, or provides responses on physical examination inconsistent with known physiology can be particularly challenging. A strongly positive supine straight leg raising test without complaint on sitting knee extension and inconsistent responses on examination raise a suspicion that nonphysical factors may be affecting the patient's responses. "Pain behaviors" (verbal or nonverbal communication of distress or suffering) such as amplified grimacing, distorted gait or posture, moaning, and rubbing of painful body parts may also cloud medical issues and even evoke angry responses from the clinician.
Interpreting inconsistencies or pain behaviors as malingering does not benefit the patient or the clinician. It is more useful to view such behavior and inconsistencies as the patient's attempt to enlist the practitioner as an advocate, a plea for help. The patient could be trapped in a job where activity requirements are unrealistic relative to the person's age or health. In some cases, the patient may be negotiating with an insurer or be involved in legal actions. In patients with recurrent back problems, inconsistencies and amplifications may simply be habits learned during previous medical evaluations. In working with these patients, the clinician should attempt to identify any psychological or socioeconomic pressures that might be influenced in a positive manner. The overall goal should always be to facilitate the patient's recovery and avoid the development of chronic low back disability.
Algorithm 1. Initial evaluation of acute low back problem
If the initial assessment detects no serious condition, assure the patient that there is "no hint of a dangerous problem" and that "a rapid recovery can be expected." The need for education will vary among patients and during various stages of care. An obviously apprehensive patient may require a more detailed explanation. Patients with sciatica may have a longer expected recovery time than patients with nonspecific back symptoms and thus may need more education and reassurance. Any patient who does not recover within a few weeks may need more extensive education about back problems and the reassurance that special studies may be considered if recovery is slow.
Comfort is often a patient's first concern. Nonprescription analgesics will provide sufficient pain relief for most patients with acute low back symptoms. If treatment response is inadequate, as evidenced by continued symptoms and activity limitations, prescribed pharmaceuticals or physical methods may be added. Comorbid conditions, side effects, cost, and provider/patient preference should guide the clinician's choice of recommendations. Table 2 summarizes comfort options.
The safest effective medication for acute low back problems appears to be acetaminophen. Nonsteroidal anti-inflammatory drugs (NSAIDs), including aspirin and ibuprofen, are also effective although they can cause gastrointestinal irritation/ulceration or (less commonly) renal or allergic problems. Phenylbutazone is not recommended due to risks of bone marrow suppression. Acetaminophen may be used safely in combination with NSAIDs or other pharmacologic or physical therapeutics, especially in otherwise healthy patients.
Muscle relaxants seem no more effective than NSAIDs for treating patients with low back symptoms, and using them in combination with NSAIDs has no demonstrated benefit. Side effects including drowsiness have been reported in up to 30 percent of patients taking muscle relaxants.
Opioids appear no more effective than safer analgesics for managing low back symptoms. Opioids should be avoided if possible and, when chosen, used only for a short time. Poor patient tolerance and risks of drowsiness, decreased reaction time, clouded judgment, and potential misuse/dependence have been reported in up to 35 percent of patients. Patients should be warned of these potentially debilitating problems.
Manipulation,defined as manual loading of the spine using short or long leverage methods, is safe and effective for patients in the first month of acute low back symptoms without radiculopathy. For patients with symptoms lasting longer than 1 month, manipulation is probably safe but its efficacy is unproven. If manipulation has not resulted in symptomatic and functional improvement after 4 weeks, it should be stopped and the patient reevaluated.
Traction applied to the spine has not been found effective for treating acute low back symptoms.
Physical modalities such as massage, diathermy, ultrasound, cutaneous laser treatment, biofeedback, and transcutaneous electrical nerve stimulation (TENS) also have no proven efficacy in the treatment of acute low back symptoms. If requested, the clinician may wish to provide the patient with instructions on self-application of heat or cold therapy for temporary symptom relief.
Invasive techniques such as needle acupuncture and injection procedures (injection of trigger points in the back; injection of facet joints; injection of steroids, lidocaine, or opioids in the epidural space) have no proven benefit in the treatment of acute low back symptoms.
Other miscellaneous therapies have been evaluated. No evidence indicates that shoe lifts are effective in treating acute low back symptoms or limitations, especially when the difference in lower limb length is less than 2 cm. Shoe insoles are a safe and inexpensive option if requested by patients with low back symptoms who must stand for prolonged periods. Low back corsets and back belts, however, do not appear beneficial for treating acute low back symptoms.
To avoid both undue back irritation and debilitation from inactivity, recommendations for alternate activity can be helpful. Most patients will not require bed rest. Prolonged bed rest (more than 4 days) has potential debilitating effects, and its efficacy in the treatment of acute low back problems is unproven. Two to four days of bed rest are reserved for patients with the most severe limitations (due primarily to leg pain).
Activities and postures that increase stress on the back also tend to aggravate back symptoms. Patients limited by back symptoms can minimize the stress of lifting by keeping any lifted object close to the body at the level of the navel. Twisting, bending, and reaching while lifting also increase stress on the back. Sitting, although safe, may aggravate symptoms for some patients. Advise these patients to avoid prolonged sitting and to change position often. A soft support placed at the small of the back, armrests to support some body weight, and a slight recline of the chair back may make required sitting more comfortable.
Until the patient returns to normal activity, aerobic (endurance) conditioning exercise such as walking, stationary biking, swimming, and even light jogging may be recommended to help avoid debilitation from inactivity. An incremental, gradually increasing regimen of aerobic exercise (up to 20 to 30 minutes daily) can usually be started within the first 2 weeks of symptoms. Such conditioning activities have been found to stress the back no more than sitting for an equal time period on the side of the bed. Patients should be informed that exercise may increase symptoms slightly at first. If intolerable, some exercise alteration is usually helpful.
Conditioning exercises for trunk muscles are more mechanically stressful to the back than aerobic exercise. Such exercises are not recommended during the first few weeks of symptoms, although they may later help patients regain and maintain activity tolerance.
There is no evidence to indicate that back-specific exercise machines are effective for treating acute low back problems. Neither is there evidence that stretching of the back helps patients with acute symptoms.
When requested, clinicians may choose to offer specific instructions about activity at work for patients with acute limitations due to low back symptoms. The patient's age, general health, and perceptions of safe limits of sitting, standing, walking or lifting (noted on initial history) can help provide reasonable starting points for activity recommendations. Table 3 provides a guide for recommendations about sitting and lifting. The clinician should make clear to patients and employers that:
Activity restrictions are prescribed for a short time period only, depending upon work requirements (no benefits apparent beyond 3 months).
Table 3. Guidelines for sitting and unassisted lifting 1Without getting up and moving around. 2Modification of NIOSH Lifting Guidelines, 1981, 1993. Gradually increase unassisted lifting limits to 60 lbs (men) and 35 lbs (women) by 3 months even with continued symptoms. Instruct patient to limit twisting, bending, reaching while lifting and to hold lifted object as close to navel as possible.
Algorithm 2. Treatment of acute low back problem on initial and followup visits
Routine testing (laboratory tests, plain x-rays of the lumbosacral spine) and imaging studies are not recommended during the first month of activity limitation due to back symptoms except when a red flag noted on history or examination raises suspicion of a dangerous low back or non- spinal condition. If a patient's limitations due to low back symptoms do not improve in 4 weeks, reassessment is recommended. After again reviewing the patient's activity limitations, history, and physical findings, the clinician may then consider further diagnostic studies, and discuss these with the patient.
Waiting 4 weeks before considering special tests allows 90 percent of patients to recover spontaneously and avoids unneeded procedures. This also reduces the potential confusion of falsely labeling age-related changes on imaging studies (commonly noted in patients older than 30 without back symptoms) as the cause of the acute symptoms. In the absence of either red flags or persistent activity limitations due to continuous limb symptoms, imaging studies (especially plain x-rays) rarely provide information that changes the clinical approach to the acute low back problem.
Prior to ordering imaging studies the clinician should have noted either of the following:
Physiologic evidence may be in the form of definitive nerve findings on physical examination, electrodiagnostic studies (when evaluating sciatica), and a laboratory test or bone scan (when evaluating nonspecific low back symptoms). Unquestionable findings that identify specific nerve root compromise on the neurologic examination (see Figure 1) are sufficient physiologic evidence to warrant imaging. When the neurologic examination is less clear, however, further physiologic evidence of nerve root dysfunction should be considered before ordering an imaging study. Electromyography (EMG) including H-reflex tests may be useful to identify subtle focal neurologic dysfunction in patients with leg symptoms lasting longer than 3-4 weeks. Sensory evoked potentials (SEPs) may be added to the assessment if spinal stenosis or spinal cord myelopathy is suspected.
Laboratory tests such as erythrocyte sedimentation rate (ESR), complete blood count (CBC), and urinalysis (UA) can be useful to screen for nonspecific medical diseases (especially infection and tumor) of the low back. A bone scan can detect physiologic reactions to suspected spinal tumor, infection, or occult fracture.
Should physiologic evidence indicate tissue insult or nerve impairment, discuss with a consultant selection of an imaging test to define a potential anatomic cause (CT for bone, MRI for neural or other soft tissue). Anatomic definition is commonly needed to guide surgery or specific procedures. Selection of an imaging test should also take into consideration any patient allergies to contrast media (myelogram) or concerns about claustrophobia (MRI) and costs. A discussion with a specialist on selection of the most clinically valuable study can often assist the primary care clinician to avoid duplication. Table 4 provides a general comparison of the abilities of different techniques to identify physiologic insult and define anatomic defects. Missing from the table is discography, which is not recommended for assessing patients with acute low back symptoms.
In general, an imaging study may be an appropriate consideration for the patient whose limitations due to consistent symptoms have persisted for 1 month or more:
Reliance upon imaging studies alone to evaluate the source of low back symptoms, however, carries a significant risk of diagnostic confusion, given the possibility of falsely identifying a finding that was present before symptoms began.
Definitive treatment for serious conditions (see Table 1) detected by special studies is beyond the scope of this guideline. When special studies fail to define the exact cause of symptoms, however, no patient should receive an impression that the clinician thinks "nothing is wrong" or that the problem could be "in their head." Assure the patient that a clinical workup is highly successful in detecting serious conditions, but does not reveal the precise cause of most low back symptoms.
Within the first 3 months of acute low back symptoms, surgery is considered only when serious spinal pathology or nerve root dysfunction obviously due to a herniated lumbar disc is detected. A disc herniation, characterized by protrusion of the central nucleus pulposus through a defect in the outer annulus fibrosis, may trap a nerve root causing irritation, leg symptoms and nerve root dysfunction. The presence of a herniated lumbar disc on an imaging study, however, does not necessarily imply nerve root dysfunction. Studies of asymptomatic adults commonly demonstrate intervertebral disc herniations that apparently do not entrap a nerve root or cause symptoms.
Therefore, nerve root decompression can be considered for a patient if all of the following criteria exist:
Patients with acute low back pain alone, without findings of serious conditions or significant nerve root compression, rarely benefit from a surgical consultation.
Many patients with strong clinical findings of nerve root dysfunction due to disc herniation recover activity tolerance within 1 month; no evidence indicates that delaying surgery for this period worsens outcomes. With or without an operation, more than 80 percent of patients with obvious surgical indications eventually recover. Surgery seems to be a luxury for speeding recovery of patients with obvious surgical indications but benefits fewer than 40 percent of patients with questionable physiologic findings. Moreover, surgery increases the chance of future procedures with higher complication rates. Overall, the incidence of first-time disc surgery complications, including infection and bleeding, is less than 1 percent. The figure increases dramatically with older patients or repeated procedures.
Direct methods of nerve root decompression include laminotomy (expansion of the interlaminar space for access to the nerve root and the offending disc fragments), microdiscectomy (laminotomy using a microscope), and laminectomy (total removal of laminae). Methods of indirect nerve root decompression include chemonucleolysis, the injection of chymopapain or other enzymes to dissolve the inner disc. Such chemical treatment methods are less efficacious than standard or microdiscectomy and have rare but serious complications. Any of these methods is preferable to percutaneous discectomy (indirect, mechanical disc removal through a lateral disc puncture).
Usually resulting from soft tissue and bony encroachment of the spinal canal and nerve roots, spinal stenosis typically has a gradual onset and begins in older adults. It is characterized by nonspecific limb symptoms, calledneurogenic claudicationor pseudoclaudication, that interfere with the duration of comfortable standing and walking. The symptoms are commonly bilateral and rarely associated with strong focal findings on examination. Neurogenic claudication, however, can be confused or coexist with vascular claudication,in which leg pain also limits walking. The symptoms of vascular insufficiency can be relieved by simply standing still while relief of neurogenic claudication symptoms usually require the patient to flex the lumbar spine or sit.
The surgical treatment for spinal stenosis is usually complete laminectomy for posterior decompression. Offending soft tissue and osteophytes that encroach upon nerve roots in the central spinal canal and foramen are removed. Fusion may be considered to stabilize a degenerative spondylolisthesis with motion between the slipped vertebra and adjacent vertebrae. Elderly patients with spinal stenosis who tolerate their daily activities usually need no surgery unless they develop new signs of bowel or bladder dysfunction. Decisions on treatment should take into account the patient's preference, lifestyle, other medical problems, and risks of surgery. Surgery for spinal stenosis is rarely considered in the first 3 months of symptoms.
Except for cases of trauma-related spinal fracture or dislocation, fusion alone is not usually considered in the first 3 months following onset of low back symptoms.
Algorithm 4. Surgical considerations for patients with persistent sciatica
Following diagnostic or surgical procedures, the management of most patients becomes focused on improving physical conditioning through an incrementally increased exercise program. The goal of this program is to build activity tolerance and overcome individual limitations due to back symptoms. At this point in treatment, symptom control methods are only an adjunct to making prescribed exercises more tolerable.
When patients demonstrate difficulty regaining the ability to tolerate the activities they are required (or would like) to do, the clinician may pose the following diagnostic and treatment questions:
Algorithm 5. Further management of acute low back problem
Table 5. Summary of Guideline Recommendations The ratings in parentheses indicate the scientific evidence supporting each recommendation according to the following scale: A = strong research-based evidence (multiple relevant and high-quality scientific studies). B = moderate research-based evidence (one relevant, high-quality scientific study or multiple adequate scientific studies). C = limited research-based evidence (at least one adequate scientific study in patients with low back pain). D = panel interpretation of evidence not meeting inclusion criteria for research-based evidence. The number of studies meeting panel review criteria is noted for each category.
For each clinical practice guideline developed under the sponsorship of the Agency for Health Care Policy and Research (AHCPR), several versions are produced to meet different needs.
The Clinical Practice Guideline presents recommendations for health care providers with brief supporting information, tables and figures, and pertinent references.
The Quick Reference Guide for Clinicians is a distilled version of the Clinical Practice Guideline,with summary points for ready reference on a day-to-day basis.
The Consumer Version,available in English and Spanish, is an information booklet for the general public to increase patient knowledge and involvement in health care decisionmaking.
To order single copies of guideline products or to obtain further information on their availability, call the AHCPR Publications Clearinghouse toll-free at 800-358-9295 or write to: AHCPR Publications Clearinghouse, P.O. Box 8547, Silver Spring, MD 20907.
Single copies of the Clinical Practice Guideline are available for sale from the Government Printing Office, Superintendent of Documents, Washington, DC 20402, with a 25 percent discount given for bulk orders of 100 copies of more. The Quick Reference Guide for Clinicians and the Consumer Version in English are also available for sale in bulk quantities only. Call (202) 512-1800 for price and ordering information.
The Guideline Technical Report contains complete supporting materials for the Clinical Practice Guideline, including background information, methodology, literature review, scientific evidence tables, recommendations for research, and a comprehensive bibliography. It is available from the National Technical Information Service, 5285 Port Royal Road, Springfield, VA 22161. Call (703) 487-4650 for price and ordering information.
The full text of guideline documents for online retrieval may be accessed through a free, electronic service from the National Library of Medicine called HSTAT (Health Services/Technology Assessment Text). Guideline information is also available through some of the computer- based information systems of the National Technical Information Service, professional associations, nonprofit organizations, and commercial enterprises.
A fact sheet describing Online Access for Clinical Practice Guidelines (AHCPR Publication No. 94-0075) and copies of the Quick Reference Guide for Clinicians and the Consumer Version of each guideline are available through AHCPR's InstantFAX, a fax-on-demand service that operates 24 hours a day, 7 days a week. AHCPR's InstantFAX is accessible to anyone using a facsimile machine equipped with a touchtone telephone handset: Dial (301) 594-2800, push "1," and then press the facsimile machine's start button for instructions and a list of currently available publications.
U.S. Department of Health and Human Services, Public Health Service Agency for Health Care Policy and Research, Executive Office Center, Suite 501 2101 East Jefferson Street, Rockville, MD 20852 AHCPR Publication No. 95-0643 December 1994.