Coralyn W. Whitney

Regression to the mean in treated versus untreated chronic pain

&NA; The course of pain associated with temporomandibular disorders (TMD) and other chronic pain conditions is typically episodic. Its expression may influence when a person seeks treatment, for example, when the level of pain flares up or exceeds its characteristic severity. Improvement in pain status subsequent to entering treatment may be due to: (1) specific effects of treatment; (2) non‐specific effects of treatment (‘placebo effects’); or (3) regression to the mean. Due to regression to the mean, uncontrolled evaluation of treatment in persons self‐selected by a pain flare‐up may lead to erroneous conclusions concerning effects of treatment by patients, providers, and/or researchers. For this report, the magnitude of regression to the mean due to self‐selection for treatment is estimated by comparing subjects who sought treatment for TMD pain (n = 147) to a random sample of subjects with TMD pain not seeking treatment (n = 95. Among subjects seeking treatment, a significant 14.7‐point reduction in VAS pain intensity was observed at 1‐year follow‐up. A control group of TMD subjects not seeking treatment showed no mean reduction in pain intensity but reported lower pain intensity at baseline than the group seeking care. When both groups of subjects were stratified on baseline VAS pain values, the reduction in pain increased as the baseline pain level increased, but no differences between comparable treated and untreated cases in the extent of improvement were observed. The before‐after differences in both groups may be attributed to regression to the mean. We conclude that before‐after differences in pain intensity can be large and that such improvement may be largely due to regression to the mean. This suggests the need for research which differentiates change due to regression to the mean (due to homeostatic processes, random within‐subject variation, or measurement error) from change due to specific and non‐specific effects of treatment. In clinical practice, the personal experience of patients and clinicians who observe improvement after initiation of treatment should be regarded as an unreliable guide to treatment efficacy due to regression to the mean. This phenomenon may contribute to the proliferation and continued use of treatments of unproven efficacy for pain management and suggests caution in the use of costly or risky pain treatments the efficacy of which is unknown.

Brief group cognitive-behavioral intervention for temporomandibular disorders

&NA; Temporomandibular disorders (TMD) are currently viewed as an interrelated set of clinical conditions presenting with signs and symptoms in masticatory and related muscles of the head and neck, and the soft tissue and bony components of the temporomandibular joint. Epidemiologie and clinical studies of TMD confirm its status as a chronic pain problem. In this report we present results from a randomized clinical trial which compared, at 3‐ and 12‐month follow‐ups, the effects of usual TMD treatment on TMD pain and related physical and psychological variables with the effects of a cognitive‐behavioral (CB) intervention delivered to small groups of patients before usual TMD treatment began. The purpose of this study was to determine whether a minimal CB intervention followed by dental TMD treatment enhanced the effects of usual clinical dental treatment. A second purpose of the study was to determine whether patients classified as high in somatization and psychosocial dysfunction would respond less favorably to this minimal intervention than would those low in somatization and dysfunction. Patients who participated in the CB intervention followed by usual treatment showed greater long‐term decreases in reported pain level and pain interference in daily activities than did patients who received only usual treatment. The benefits of CB intervention were not seen when the CB and UT groups were compared at 3‐month follow‐up. During the 3–12‐month follow‐up interval, however, the UT group maintained essentially the same level of improvement in characteristic pain while the CB group continued to improve, as hypothesized. During this same follow‐up interval, the CB group also showed a strong trend toward continued improvement in pain interference. Such effects were not observed for depression, somatization, or clinical measures of jaw range of motion. Additionally, as hypothesized, dysfunctional chronic pain patients did not appear to benefit from the brief CB intervention. Intent to treat analyses were also performed to assess generalizability of the results.

The analgesic effects of caffeine in headache

&NA; Caffeine is frequently added to mild analgesic preparations but its effect when used alone on pain has never been studied in humans. Using a double‐blind placebo‐controlled multiple crossover design, 53 patients with non‐migrainous headaches were given placebo, acetaminophen, 2 doses of caffeine and 2 combinations of caffeine with acetaminophen. Caffeine appeared to have independent analgesic effects that were equivalent to acetaminophen and were still significant when statistical adjustments were made for prior caffeine consumption and caffeines effects on mood.

Somatization and pain dispersion in chronic temporomandibular disorder pain

&NA; We investigated the relationship between somatic and psychological symptoms and pain reported during a clinical examination for 220 patients with chronic temporomandibular disorder (TMD) pain, The clinical examination involved palpation of the muscles of the face and neck, as well as intraoral sites and non‐TMD‐related placebo sites. A distinction was drawn between somatization — the tendency to report numerous somatic symptoms — and psychological distress manifested by report of numerous affective and cognitive symptoms. Somatization was assessed with the Somatization scale of the SCL‐90‐R; cognitive/ affective distress was assessed with the non‐somatic items of the Anxiety and Depression scales. Heightened somatization and high‐intensity pain were strong predictors of widely dispersed muscle palpation pain during the clinical examination. High‐somatization patients were 3 times more likely than low‐somatization subjects to report having a painful placebo site. Pain dispersion was more closely linked to report of number of somatic symptoms than to report of affective and cognitive symptoms of psychological distress.

THE INTRAUTERINE DEVICE AND PELVIC INFLAMMATORY DISEASE: THE WOMEN'S HEALTH STUDY REANALYZED

The Womens Health Study (WHS) was a large, widely accepted and influential case-control study of the relationship between the use of intrauterine contraceptive devices (IUDs) and pelvic inflammatory disease (PID). The data were collected at 16 hospitals in 9 cities across the U.S.A. from October 1976 through August 1978. The first paper on this research was published in 1981 and concluded that IUDs increase the risk of PID. The report cited an estimated RR (relative risk) of PID for current IUD users vs nonIUD users of 1.6 with a 95% confidence interval of (1.4, 1.9). However, careful examination of the report reveals that the data support conclusions antithetical to those at which the author arrived. When the second report on the WHS was published in 1983, it was anticipated that many of the shortcomings of the first report would be corrected, but they were not. In 1983 we undertook a complete reanalysis of the same WHS data using more appropriate criteria and the results were compared to the first two published reports. The reanalysis revealed an RR of 1.02 (0.86, 1.21) for current IUD users compared to noncontraceptors. The conclusion of the WHS should have been that IUDs do not increase the risk of PID.

Measurement of characteristic pain intensity in field research

,METH61DS: Questionnaires and a ? elective surgery. 0 cm visual analog scale (VAS) were used within 24 hrs. prior to surgery to obtain information about: surgery (VAS), pre-surgical pain (VPS), general attitudes a g ain expected postout pain speclflc attitudes about pain with this operation (7 items), and pain i 7 items), his ory with each past surgery, global surgical pain, pain with each past trauma,. \ pain

Paradoxical effects of nitrous oxide on human memory

Using the method of adjusted learning, subjects learned number-noun pairs while breathing either placebo or 30% nitrous oxide. Subjects breathing nitrous oxide required more acquisition trials to attain a learning criterion than did subjects breathing placebo. Two weeks later, half of the subjects from each group were administered either placebo or nitrous oxide and were asked to recall the noun that had accompanied each number cue. Results showed that: 1) nitrous oxide inhalation can decrease the accessibility of to-be-recalled material and 2) nitrous oxide administered during the acquisition of material can paradoxically improve the recall of that material 2 weeks later. The additional number of acquisition trials subjects received during nitrous oxide inhalation could potentially account for this paradoxical enhancement of delayed recall; however, correlational analyses suggest this was not the case. No evidence for any state-dependent effects of nitrous oxide on cued recall were found.

Tort reform and malpractice liability insurance

Legal factors related to the malpractice liability insurance purchased by general dentists in the United States were investigated using a comprehensive multivariate model that assessed the contribution of legal provisions affecting the probability of a malpractice claim, the probability of a payment, and the average size of the payment. General practice dentists in the United States were selected randomly, and 3,048 dentists were studied by mail survey. A number of legal statutes (periodic payment allowed, percentage fault liability informed consent limits, limits on res ipsa loquitor, attorney fee control, some statute of limitations provisions) had the intended effect of reducing the malpractice insurance sought by dentists. Other provisions, such as binding arbitration, may have unintended cost-raising effects. Previous malpractice claims were associated with purchasing greater amounts of insurance. These findings have implications for future changes in the legal system as part of health care reform.

Perceived ability to control pain & continued TMD pain

AIM OF INVESTIGATION: Little data exists regarding the influence of perceived ability to control pain on the course of pain and associated impairment. This study examined the relationship between R atients’ baseline confidence in their ability to relieve pain and the presence of pain one year later, a er adjusting for baseline measures of pain intensity, interference with daily activities, years since onset of pain and psychological status.

Response to Drs. J. Nick, K. Mann and A. Temple

We thank Drs. Nick, Mann and Temple from McNeil Products (producers of Tylenol-acetaminophen) for a chance to clarify aspects of interpretation and design in our study comparing caffeine alone with placebo, acetaminophen and acetaminophen with caffeine in the treatment of recurrent non-migrainous headaches. First, they have quoted our conclusions out of context and then criticized our study based on their inaccurate quote. In our abstract we state, “Caffeine appeared to have independent analgesic effects . . ” and the full quote that they partially quoted begins, “Data from this study support the hypothesis that caffeine has a direct analgesic effect “. We never state that we have proven the hypothesis or that our conclusions are definite. Our cautious, scientific presentation of the study’s results were thus neither unwarranted, overstated or premature. We started the study with a specific hypothesis, studied a carefully defined patient population, employed a placebo condition and used patients as their own controls to maximize the power of the study, and used what was then state-of-the-art scales to maximize sensitivity and reliability. Thus, it would be hard to characterize our study as “broad-ranging, loosely controlled and hypothesis generating”. The following are specific responses numbered to match the numbered specific criticisms in their letter. 1. Limiting cross-over designs to chronic illnesses is not universally accepted. Carryover effects even with washouts are often a problem in these situations. For example, cross-over designs in depression research are usually not acceptable because some mood improvement may continue into the next condition. The appropriateness of cross-over trials is not based on acuity or chronicity. Acceptable conditions for doing a cross-over trial include the ability to design a trial that has no significant differences in baseline conditions between treatments. In our study there were no significant baseline differences in pain, or prior caffeine consumption between treatment conditions. Possible variations in duration of headaches may decrease the sensitivity of the study but does not invalidate the results. To imply that each headache is unique and therefore cannot or should not be compared with others either within or between subjects is a nihilistic absurdity. To insure some comparability and uniformity in our subjects, each had to have had headaches without migraine features at least 6 times a month with a severity of 2 or more on a l-5 pam scale. These entry criteria were more specific than most of the headache analgesia studies published to date. We agree with the criticism that a 6-way cross-over design was excessively complicated and increased the chances of finding sequence effects that complicate interpretation of results. We recommend that future studies be limited to 2or possibly 3-way cross-over designs comparing 130 mg of caffeine with placebo and possibly a standard dose of a reference analgesic, e.g., acetaminophen or aspirin. Because caffeine effects mood which can then effect pain reports and because caffeine withdrawal can cause headache, studies cannot responsibly ignore these confounding variables. Instead they should, in examining the effects of caffeine on headache, measure these variables and account for them in the data analysis. Because mood is also a dependant variable, accounting for this in the design and analysis will make the outcome of any such study more difficult to interpret. Secondly, this study was planned and conducted as a series of 2-h trials with measurement at 30, 60 and 120 min. The reasons include: (1) better subject compliance than longer trials; (2) clinically realistic parameters onsets of clinically significant action beyond 2 h has limited clinical applicability for acute pain; and (3) a comparison with a known standard. All trials of acetaminophen (and aspirin) demonstrate clinically significant and near peak effects at 2 h. The purpose of this study was to determine if caffeine alone has analgesic effects in a 2-h clinical trial. A finer analysis involving a longer trial measuring peak effect, duration of effect or total effect is not necessary for this purpose. However, we do agree that because caffeine appears to have analgesic effects, a future study should more precisely delineate its effects by including these measures. The only reason for patients to take medication at least 6 h before bedtime was to avoid caffeine-induced insomnia. When interaction effects are found, there are several alternative analyses that can be done, all of which risk type-11 (inaccurately accepting the null hypothesis) errors because the Ns in each condition are very small (N = 6 to 101, thus resulting in a very low power design. This does not mean that we had an insensitive study but rather that it lost sensitivity when the data was divided into individual groups for each period. The two most commonly used approaches when such interaction effects are found are to analyze only the first headache or to treat each sequence as a separate experiment. We did analyze the first headache data and not unexpectedly found no statistically significant differences among the treatments. It is well known in analgesia research that the effect size of acetaminophen requires a substantially larger N (depending in part on the sensitivity of outcome measures) to demonstrate acetaminophen’s analgesic effects. Somewhat to our surprise because we do not believe caffeine’s effect size is greater than acetaminophen’s, three of the sequences showed caffeine alone (65 mg in one and 130 mg in two others) to be significantly better than placebo at 2 h. The drug-by-period graph that they present to a large part reflects the extreme fluctuation that can be expected from too small cells where only one subject can significantly skew the mean. We believe they are overstating the results of the first period analysis.